Abstract
This dissertation collects three papers on the real exchange rate.
The fisrt paper of this disertation develops a classical political economy framework
put forth by Anwar Shaikh. Unlike mainstream theories which focus on relative
consumer or producer prices, we argue that the real relative unit labor cost is the center of
gravity or main force that explains the long-run behavior of the real exchange rate.
Therefore, in our view, international competition is driven mainly by real cost differences
between countries. The paper further argues that capital inflows, triggered by positive
real interest rate differentials, could create short-run deviations of the real exchange rate
from its center of gravity. Our econometric results using the ARDL-ECM framework
confirmed the main hypothesis of this paper for 16 OECD countries, Taiwan, and 3
developing countries.
The second paper presents a case study of the Mexican-U.S. real exchange rate. In
this paper we argue that the relative unit labor cost of the Mexican and US manufacturing
sectors is a good indicator of the real exchange rate. Our econometric models show
evidence that these two variables, as well as the real net capital inflows to Mexico and the
government final consumption expenditures, are structurally related. However, for the
period 1983-2011, we showed that the real relative unit labor cost ratio is the most
important variable in explaining the long-run bahavior of the Mexican real exchange rate.
The third paper assesses the effects of an undervalued currency on economic growth.
Here, I revert to the standard theory of real exchange rates (as opposed to the classical
one adopted in the first two chapters), namely that the real exchange rate is in equilibrium
when the balance of trade is zero. This is the foundation for Rodrik’s PPP undervaluation
index. Following him, I extend and apply his approach to a variety of countries. Our
econometric results suggest that real exchange rate undervaluation has, to differing
degrees, been able to enhance the economic growth of developed and developing
countries. Nevertheless, when we disaggregate the main components of aggregate
demand for different clusters of developed and developing countries using the Stock
Flow Consistent approach (SFC), we find that in general, an undervalued currency has
expansionary and contractionary effects in the short-run, specifically via the export sector
and the level of aggregate consumption, respectively. This paper also estimates the
effects of an undervalued currency on the level of investment and the trade balance.
The Political Economy of Real Exchange Rate Behavior:
Theory and Empirical Evidence for Developed and
Developing Countries
by
Francisco A. Martínez Hernández
January 2016
Submitted to The New School for Social Research of The New School in partial
fulfillment of the requirements for the Degree of Doctor of Philosophy.
Dissertation Committee:
Dr. Anwar Shaikh
Dr. William Milberg
Dr. Jamee Moudud
Dr. James Dodd
All rights reserved
INFORMATION TO ALL USERSThe quality of this reproduction is dependent upon the quality of the copy submitted.
In the unlikely event that the author did not send a complete manuscriptand there are missing pages, these will be noted. Also, if material had to be removed,
a note will indicate the deletion.
All rights reserved.
This work is protected against unauthorized copying under Title 17, United States CodeMicroform Edition © ProQuest LLC.
ProQuest LLC.789 East Eisenhower Parkway
P.O. Box 1346Ann Arbor, MI 48106 - 1346
ProQuest 10126008
Published by ProQuest LLC (2016). Copyright of the Dissertation is held by the Author.
ProQuest Number: 10126008
Ⓒ2016 Francisco A. Martínez Hernández
All rights reserved
To Camila
v
Acknowledgments
The papers collected in this dissertation own a great deal of help from several people and
institutions in the U.S. and Mexico. I would like to start thanking to the Consejo Nacional de
Ciencia y Tecnología (CONACyT), el Banco de México, la UNAM, and the New School for
Social Research for having granted me the financial support to conduct my doctoral studies in
New York City for several years. I am also grateful to Anwar Shaikh, Jamee Moudud, William
Milberg, and Alec Gershberg for having granted me several teaching and research assistantships
at the New School; these academic experiences contributed so much to my knowledge and
enriched enormously my academic experience. I also want to thank to Robert Kostrzewa and
Tsuya Yee for granting me a dissertation fellow award in 2012, this scholarship helped me to
return to the U.S. in order to continue developing this dissertation.
I am intellectually indebted with the Chair of my Dissertation Committee, Anwar Shaikh.
The first two chapters of this dissertation were inspired on his papers on the real exchange rate.
He kindly helped me to construct the data set used in the first chapter, and always he was there to
give me his wise opinion and to guide me in the right direction. I also thank him for teaching me
many things while I was his RA at the SCEPA, especially how to use economic theory to develop
and to identify meaningful long-term patterns between economic variables.
Through my years at the New School, I have cultivated many friendships that in many ways
have made my stay at New York very pleasant. I want to mention first to Barbara Herbst
(Secretary of the Economics Department), who has helped me countless times with her advises
and administrative paperwork. Also I want to thank to Mishan Hing and Francisco Aldape for
expanding my knowledge of many topics in our casual talks in several places around the New
School. Other friends to whom I want to thank for their friendship are Didier Ntoko, Xiao Jiang,
Daniela Arias, Nancy Hernández, Percival Pineda, Cameron Weber, Sedar Fred Pougaza, Pedro
Cadenas, Kurt Birson, Cleiton Silva, Luis Torres, and Katherine Moss.
Professors and friends from the UNAM were also very helpful in different stages of my
doctoral studies. I want to thank to Professor Angel de la Vega, Julio López, and Luis Quintana
for their comments to my dissertation and the opportunity to make several presentations of this
dissertation at the UNAM. I am in great debt with Noemi Medina and Saul Mendoza for their
help and time to overcome difficulties I encountered in some highly technical classes. I would
like to give special thanks to Rafael Valencia, Saul Herrera, and Omar Contreras. Saul and Omar
contributed so much in the development of this dissertation as they provided me with several
vi
sources of data and Rafael helped me with the final edition of this dissertation. Last but not least,
I want to thank to professor Jaime Ros, Roberto Gochez, and Martin Rapetti for their constructive
comments to this dissertation.
Finally, I want to thank to my family because without their constant support and
encouragement I would not have been able to accomplish anything at all. I want to thank Aida
Villalobos Sosa for the happy years we spent together as a couple and for taking care of my little
daughter Camila in my absence. Thanks so much to Conchita (my mom) and to Quique (my
brother) for taking care of me after my foot’s operation (December 2015). Although I lost my
father Luis Antonio since 2011, I want to tell him that my love for him will always live in my
memory.
Francisco A. Martínez Hernández
The New School for Social Research
January 2016
vii
Contents
Acknowledgments ................................................................................................... v
List of Figures ........................................................................................................ ix
List of Tables.......................................................................................................... ix
I. The Political Economy of Real Exchange Rate Behavior: Theory and
Empirical Evidence for Developed and Developing Countries, 1960-2010 ..... 1
I.1 Introduction ................................................................................................. 1
I.2 Conventional Theories of Real Exchange Rate Determination and Their
Empirical Results ........................................................................................ 4
I.3 A Classical Framework for the Analysis of Real Exchange Rate ............. 13
I.4 The Role of Interest Rate Differential ....................................................... 26
I.5 Statistical Analysis: RXR, RULC, Real Interest Rate Differential, and
Trade Balance ............................................................................................ 32
I.6 Empirical Evidence of Alternative Real Exchange Rate Determination
Based on ARDL-ECM Modelling............................................................. 41
I.7 Concluding Remarks ................................................................................. 48
References ........................................................................................................ 50
II. An Alternative Theory of Real Exchange Rate Determination: Theory and
Empirical Evidence for the Mexican Economy, 1970-2011 ............................. 52
II.1 Introduction ............................................................................................... 52
viii
II.2 Conventional Models of Exchange Rate Determination ........................... 56
II.3 An Alternative Theory of Real Exchange Rate Determination ................ 61
II.4 Statistical Analysis: RULCR US-MEX .................................................... 67
II.5 Empirical Evidence of Alternative Real Exchange Rate Determination
Using an ARDL-ECM Model ................................................................... 66
II.6 Concluding Remarks ................................................................................. 72
References ........................................................................................................ 74
III. Real Exchange Rate, Effective Demand, and Economic Growth: Theory and
Empirical Evidence for Developed and Developing Countries, 1960-2010 ... 76
III.1 Introduction ............................................................................................... 76
III.2 Undervaluation and GDP Per Capita: the Cross-Country Evidence by
World Regions through Time .................................................................... 80
III.3 Undervaluation and Economic Growth .................................................... 87
III.4 Undervaluation, Effective Demand, and Economic Growth in the Short
Run ............................................................................................................ 96
III.5 Undervaluation, Investment, and External Sector in the Long-Run ...... 105
III.6 Concluding Remarks .............................................................................. 110
References ...................................................................................................... 113
Appendix to Chapter I-A ..................................................................................... 115
Appendix to Chapter I-B ..................................................................................... 140
Appendix to Chapter II-A ................................................................................... 150
Appendix to Chapter II-B ................................................................................... 162
Appendix to Chapter III ...................................................................................... 166
ix
List of Figures
Figure 1: Real Effective Exchange Rate, Adjusted Real Effective ULC Ratio, Trade Balance
(X/M), and Real Interest Rate Differential .................................................................................... 34
Figure 2: Real Exchange Rate, Real ULC Ratio, Trade Balance (X/M), and Real Interest Rate
Differential ..................................................................................................................................... 39
Figure 3: Real Exchange Rate Index and Real Net Capital Inflows, 1970-2011 ........................... 73
Figure 4: Real Exchange Rate Index and Real Mexican Government Expenditure, 1970-2011 ... 75
Figure 5: Real Exchange Rate Index and Real Unit Labor Cost Ratio Index (US-Mex) .............. 76
Figure 6: Real Exchange Rate Index and Real Unit Labor Cost Ratio Index (US-Mex) .............. 66
Figure 7: Real Exchange Rate and GDP Per Capita (PWT 7.1), 1960-2010 ................................. 82
Figure 8: Currency Under-Over/Valuation and Economic Growth ............................................... 88
Figure 9: Currency Under-Over/Valuation and Economic Growth ............................................... 89
Figure 10: Undervaluation and Effective Demand: China, Argentina, and Mexico .................... 104
List of Tables
Table 1: Panel Unit Root Test, Real Effective Exchange Rate ...................................................... 11
Table 2: Panel Cointegration Test .................................................................................................. 12
Table 3: Econometric Results ........................................................................................................ 43
Table 4: Econometric Results, Trade Balance Model .................................................................... 46
Table 5: Unit Root Test.................................................................................................................. 60
Table 6: ECM Results For Mexico-US, 1971-2011 ...................................................................... 68
Table 7: ECM Results For Mexico-US, 1976-2011 ...................................................................... 69
Table 8: ECM Results For Mexico-US, 1983-2011 ...................................................................... 70
Table 9: ECM Results For Mexico-US, 1983-2011 ...................................................................... 71
Table 10: ECM Results For Mexico-US, 1983-2011 .................................................................... 72
Table 11: Average Economic Growth By Regions ........................................................................ 77
Table 12: Balassa-Samuelson Effect.............................................................................................. 84
Table 13: Undervaluation And Economic Growth, Fixed Effect Models ..................................... 93
Table 14: Undervaluation And Economic Growth, Dynamic Panel Data ..................................... 95
Table 15: Aggregate Demand Components: Short-Run Multipliers, 1990-2010 ........................ 101
1
Chapter I
The Political Economy of Real Exchange Rate
Behavior: Theory and Empirical Evidence for
Developed and Developing Countries, 1960-2010
I.1 Introduction
World trade and international finance have evolved since the time of the classical
political economists. Yet mainstream economists have ignored some key lessons of Karl
Marx, Roy Harrod and Keynes with regard to the understanding of the uneven national
and international competitive conditions under capitalism (e.g. the role of differences in
technology, wage rate differentials, price rigidities, and the role of money in production).
Overall, the consequences of an unequal international trading system have been
unbalanced trade and high international levels of indebtedness, which have imposed an
expansionary bias for trade surplus countries and a deflationary bias for trade deficit
countries (especially for small economies which use a kind of hard currency to undertake
their international transactions).
2
Standard monetary and trade theories insist that unbalance trade situations are
temporarily in so far as they assume that domestic and external prices will have to move
in such a way as to make countries equally competitive at the international level (see for
instance Krugman 1988; 1991). However, we argue these mainstream approaches1 are
flawed in their understanding of international competitiveness as they rely on the neo-
classical price mechanism and on the theory of comparative advantage, with the
expectation that in the long-run, exports will equal imports. That is, conventional
analyses assume that, in the long-run, trade between countries will be roughly in balance.
This paper adopts a Classical grounded theory of price, which takes into
consideration the role of intra-and-inter industrial competition (real competition),
productivity, and real wages in the tradable sector. We argue that this alternative theory
of price is capable of explaining the level of national and international terms of trade, and
thus, the changes in the real exchange rate and the degree of international
competitiveness.
Our approach is based on Shaikh (1980; 1991; 1999b). According to this alternative
approach, the center of regulation or gravity of the real exchange rate is the real adjusted
unit labor cost ratio between two or different trading partner countries, so the evolution of
this latter ratio is proposed as an indicator of international competitiveness. The basic
idea of this approach is that in the realm of manufacturing production at the national and
international scale, to the extent that capital and intermediate goods are traded in
international markets whereas labor remains largely immobile internationally, labor cost
1 Recently, extensions of the mainstream theories argue that “market imperfections” have prevented trade
balances of being balanced (see Shaikh 2012 for a discussion about this point).
3
are likely to diverge much more across countries than other costs of production, and
therefore play a disproportionately important role in competitiveness. Hence, real unit
labor costs in manufacturing (labor cost per unit of output or equivalently labor cost
divided by output per worker) capture a key underlying determinant of competitiveness in
traded goods.
After this introduction, in the second section of this paper, we analyze and discuss
the underlying problems of the Purchasing Power Parity hypothesis [hereafter PPP
hypothesis]. The third section develops our alternative approach. We begin with a close
economy and we move on to the open economy. In the fourth section, drawing on Marx,
Harrod and Keynes, we explain why trade imbalances ultimately have an impact on the
internal liquidity of the economies and thus on domestic interest rates, and real exchange
rates, rather than on relative prices. In the fifth section, we present a graphical analysis of
the interrelationship between the indexes of the real effective exchange rate, the real
adjusted unit labor cost ratio, the short-run real interest rate differential, and the trade
balance for sixteen OECD countries, Taiwan, and three developing countries mostly for
the period 1960-2010. The sixth section investigates the nature of the long-run
relationship between the three aforementioned indexes and the trade balance through
cointegrating and error correction models using the autoregressive distributed lag
framework (ARDL-ECM). The last section provides some concluding remarks.
4
I.2 Conventional Theories of Real Exchange Rate Determination and
Their Empirical Results
As is well known, the PPP hypothesis has its foundation in the Law of One Price
(hereafter, LOP), whose main argument claims that if one abstracts from tariffs and
transportation costs, unfettered trade in goods should ensure identical prices across all
countries. This relation can be seen in equation 1, where the price of good i in the home
country (𝑃𝑡𝑖) multiplied by the exchange rate (𝑆𝑡) must equal its price in the foreign
country (𝑃𝑡𝑖∗). Therefore, if this law holds for every individual good, then it follows
immediately that it must hold for any identical basket of goods (equation 2); From this
relationship, according to Gustav Cassel’s original PPP hypothesis (1921:37), the
nominal exchange rate of a country is related to (caused by the changes in) the relative
price levels of all i = 1,,…,n goods between two countries (equation 3).
𝑃𝑡𝑖𝑆𝑡 = 𝑃𝑡
𝑖∗ ∑ 𝛼𝑖𝑛𝑖=1 𝑃𝑡
𝑖𝑆𝑡 = ∑ 𝛼𝑖𝑛𝑖=1 𝑃𝑡
𝑖∗ 𝑆𝑡 =∑ 𝛼𝑖𝑛
𝑖=1 𝑃𝑡𝑖∗
∑ 𝛼𝑖𝑃𝑡𝑖𝑛
𝑖=1
(1, 2 and 3)
Equation 1 reflects the LOP where subscript t stands for time, the asterisk denotes the
foreign country and the exchange rate (𝑆𝑡) is defined as foreign currency required to buy
one unit of home currency2. Equation 3 reflects the so-called absolute PPP hypothesis,
where the weights in the summation satisfy ∑ 𝛼𝑖 = 1𝑛𝑖=0 .
At the empirical level, only with the exception of the interwar periods and periods of
hyperinflations, it is widely accepted that equation 3 does not hold for any country and
2 In this definition of the exchange rate, a rise means an appreciation of home currency, a reduction a
depreciation.
5
period (Stein et al, 1997:235). According to the adherents of the absolute PPP hypothesis,
this hypothesis does not hold basically because different countries use different price
index weights to calculate price levels; other factors that have been argued are the
existence of restrictions on trade such as tariffs or quotas, transportation costs, taxes, and
imperfect information about prices in the two countries. So, in order to overcome these
factors that prevent the PPP to hold, the adherents to the PPP hypothesis have proposed
the relative version of the PPP hypothesis, which states that “the percentage change in the
exchange rate over a given period just offset the difference in inflation rates in the
countries concerned over the same period” (Taylor and Taylor 2004:137). We can
express this statement formally if we write equation 1 in natural logarithms and drop the i
superscript (the Ps now have the interpretation of the overall price level), then we obtain
equation 4.
ln 𝑆𝑡 = ln 𝑃𝑡∗ − ln 𝑃𝑡 Δ𝑠𝑡 = Δ𝑝𝑡
∗ − Δ𝑝𝑡 (4 and 5)
Now if we take first differences to equation 4, we get the proportionate change in the
exchange rate, Δ𝑠𝑡, as a function of the difference in the proportionate change in foreign
and home prices (equation 5).
Equation 5 implies that in the short-run the exchange rate could deviate from its
long-run equilibrium value (usually an arbitrarily base year), but converges eventually to
that equilibrium value. “Frenkel (1976) for example, assumes continuous adjustment to
PPP, while the Dornbush (1976) over-shooting model assumes that prices are ‘sticky’ in
the short-run, but that in the long-run prices will be equated in a common currency”
(Fraser et al., 1991:1749). In econometric terms, the applicability of the relative PPP
6
hypothesis means that the variables in equation 4 hold a cointegrating long-run
relationship, or that the real exchange rate index (see equation 6 below) follows a
stationary process over time (mean reverting process).
Multiple empirical investigations for different countries with different level of
development, different periods, different price indexes (e.g. CPI, PPI or disaggregate
price indexes) and, exchange rate regimes have widely reported evidence against the
relative PPP hypothesis (see inter alia Enders, 1988:507; Fraser et al., 1991:1757;
Harvey, 1996:580; Stein et al, 1997:236; Engel, 1996:20; Martinez, 2010:62). These
empirical results mean that the real exchange rates are not a stationary series, or that the
nominal exchange rates and the level of prices among countries do not tend to converge
over the long-run, although the latter result could be an expected outcome between
developed and developing countries, empirical results against the relative PPP also tends
to happen even among developed countries (see for instance our own results below).
As of early 1990s, different empirical investigations started reporting evidence in
favor of the relative version of PPP hypothesis by rejecting either the null hypothesis of
unit root in the real exchange rates or the null hypothesis of no cointegration between
nominal exchange rates and relative prices (Grilli and Kaminski, 1991:192; Lothian and
Taylor, 1995:488). The research efforts to show that real exchange rates converge to their
PPP level in the long-run have basically taken three different research agendas: 1)
increasing the sample to 100 years or more; 2) taking into account the difference between
tradable and not tradable goods (the so-called Balassa-Samuelson effect (hereafter B-S
7
thesis)); and 3) more recently the use of econometric panel data models in order to
increase the power of the statistic tests.
Notwithstanding these findings in favor of the relative PPP hypothesis, several
empirical investigations have seriously questioned these findings on the basis of the
inconsistency in the construction of price indexes over 100 years or more; the lack of
power in standard unit root tests; and on the empirical results, for different countries, of a
very weak relationship between non-tradable goods price indexes and the real exchange
rate.
For instance, in their extended review of the PPP literature, MacDonald and Stein
(1999:139-140) mention that “although studies which extend the span by increasing T are
interesting, they are not without their own specific problems in that the basket used to
construct the price indices is likely to be very different at the beginning and end of the
sample… … Also, such studies suffer from spanning both fixed and flexible rates
regimes and are therefore not regime invariant. For these reasons, attention has turned
from expanding T to extending N, the cross-sectional dimension”.
With regard to the lack of power in standard unit root test, Engel (1996:5), using the
decomposition of the real exchange rate through domestic and external prices of tradable
and nontradable price indexes (which is the core relationship of the B-S thesis), shows
that the recent works which use longer sample periods (100 years or more) may have
8
reached the wrong conclusion because there may be a serious size bias in the their unit
root test3.
In order to analyze in more detail Engel’s results, we first show the definition of the
real exchange rate as a price of foreign goods to domestic goods corrected for the
nominal exchange rate. In log terms,
𝑞𝑡 = 𝑠𝑡 + 𝑝𝑡∗ − 𝑝𝑡 (6)
If price indexes are geometric means of traded goods prices and non-traded goods prices,
we can write
𝑝𝑡 = (1 − 𝛼)𝑝𝑡𝑇 + 𝛼𝑝𝑡
𝑁 and 𝑝𝑡∗ = (1 − 𝛽)𝑝𝑡
𝑇∗ + 𝛽𝑝𝑡𝑁∗
Where 𝑞𝑡 is the log of the real exchange rate, 𝑠𝑡 is the log of the nominal exchange rate,
𝑝𝑡 is the log of the price index, 𝑝𝑡𝑇 is the log of the traded-goods price index, 𝑝𝑡
𝑁 is the log
of the nontraded-goods price index, 𝛼 is the share that nontraded goods take in the price
index, whereas 𝛽 is nontraded goods’ share in the foreign price index. Then the real
exchange rate can be written as
𝑞𝑡 = 𝑥𝑡 + 𝑦𝑡 (7)
Where
𝑥𝑡 = 𝑠𝑡 + 𝑝𝑡𝑇∗ − 𝑝𝑡
𝑇
𝑦𝑡 = 𝛽(𝑝𝑡𝑁∗ − 𝑝𝑡
𝑇∗) − 𝛼(𝑝𝑡𝑁 − 𝑝𝑡
𝑇)
𝑞𝑡 = (𝑠𝑡 + 𝑝𝑡𝑇∗ − 𝑝𝑡
𝑇) + 𝛽(𝑝𝑡𝑁∗ − 𝑝𝑡
𝑇∗) − 𝛼(𝑝𝑡𝑁 − 𝑝𝑡
𝑇)
3 Besides Engel’s results (1996), “Campbell and Perron (1991), and others, have noted univariate unit root
tests have relatively low power to reject the null when it is in fact false, especially when the autoregressive
component is close to unity (which it often is with time-series data)” (MacDonald and Stein 1999:138).
9
Engel (1996:15) shows that when a random variable like the real exchange rate (𝑞𝑡)
evolves according to the sum of two processes –a stationary but persistent component
(the relative price of traded goods between the countries, 𝑥𝑡) and non-stationary
component (the difference of the relative price of nontraded- to traded-goods prices in
each country, 𝑦𝑡)– the unit root tests, and also the tests of cointegration, suffer a large
size bias and one will either tend to reject the null hypothesis of unit roots in the real
exchange rates or the null hypothesis of no cointegration between nominal exchange rates
and relative prices. Therefore, his work concludes that, the recent works which use longer
sample periods (100 years or more) may have reached the wrong conclusion4.
With regard to the relationship between nontraded goods price indexes and the real
exchange rate, also Engel (1999:532), using the same decomposition in equation 7, found
out that for the bilateral movements of the U.S. real exchange rate with regard to a
number of other high-income countries for the period 1962/1970-1990/1995 (Canada,
Denmark, Finland, France, Germany, Italy, Japan, and Norway), the different measures
of the relative prices of nontraded to traded goods prices in each country (𝑦𝑡)5, appear to
4 Engel (1996:6-28) study uses quarterly data for the years 1970 to 1995 and a variety of price indexes for
which data is available on sub-components that can be identified as traded and non-traded goods. With the
25-year time series of the disaggregate prices, Engel modeled the US dollar/pound real exchange rate and
assumes, due to theoretical reasons, that 𝑥𝑡 component should be a stationary process (he assumes that LOP
applies for traded goods in the long-run), so 𝑥𝑡 is modeled as an AR (1) process (stationary series).
Meanwhile 𝑦𝑡 is modeled as a random wall process (non-stationary series). Then, 𝑞𝑡 is a non-stationary
series. With the parameters of his real exchange rate model, he then uses Monte Carlo exercises to simulate
the behavior of the real exchange rate, nominal exchange rate and relative prices in 100-year samples (400
observations). For these latter set of data, Engels’ results show that the unit root (DF and ADF) and
cointegration tests (ECM and Horvath-Watson test) suffer a high size bias because they fail to detect a unit
root process, even when the unit root component accounts for over 85% of the long-run movements in the
real exchange rate. 5 Engel (1999:509) considers four measures of nontraded-goods prices: The first is based on consumer
price indexes (CPIs), where it classifies services and housing as nontradable and commodities as tradable;
the second is constructed from an OECD database of output prices; the third is based on price deflactors for
10
account for almost none of the movements of U.S. real exchange rates. That is, for
different time horizons (as short as one month to as long as 30 years), Engel’s results
show that the mean-squared error (MSE) of the 𝑥𝑡 component (the traded-goods
component) accounts for nearly 100 percent of the MSE of the U.S. real exchange rate
changes at all horizons (except for the Canadian-U.S. rate).
The foregoing results are important because they, by and large, cast serious doubts
about the alleged causes of failure of the PPP on the basis of the B-S thesis. That is,
according to the B-S thesis, even if the PPP for tradable holds (𝑆𝑡 = 𝑃𝑡𝑇∗ 𝑃𝑡
𝑇⁄ ), if the ratio
of a country’s non-tradables goods prices relative to its traded goods rises faster than the
same relation in another country, then the real exchange rate will tend to appreciate when
measured in terms of a general price index, that is, there would be a systematic deviation
of the real exchange rate from its PPP long term equilibrium level6.
Engel’s results (1999:511) also suggest that if the PPP does not hold in the long run,
it may not be due to the 𝑦𝑡 component, but because the nominal exchange rates and
relative tradable good price indexes among countries might not follow a similar trend
over time. However, the B-S thesis, the balance of payment approach to the exchange
rate determination, as well as other exchange rate approaches (i.e. the elasticity
approach), firmly state that the PPP hypothesis for traded goods applies based on the
assumptions that: 1) wages rise in step with accelerated productivity growth in the
personal consumption expenditures; the fourth measure uses the aggregate producer price index (PPI) as a
measure of trade-good prices, and the nontraded component, 𝑦𝑡 , is constructed from aggregate CPIs relative
to aggregate PPIs. 6 Although the B-S thesis was originally proposed to explain the deviation of the bilateral exchange rate
from its PPP long run equilibrium level between a developed country and a developing country, many
economists have also used this framework to explain the deviation of the PPP among developed countries.
11
tradable sectors (perfect competition assumption); and 2) wages equalize across sectors
within a nation because of labor mobility (but no between countries) (see Botwinick 1993
and Canzoneri et al. 1999 for a critical discussion of these assumptions). The purpose of
this section is to investigate if the relative PPP hypothesis applies for traded goods. Our
calculations revolve around the “real effective exchange rate” for 16 OECD countries for
the period 1960-2010.
Two annually balanced panel data models were estimated: 1) a panel unit root test
for the index of the real effective exchange rates and; 2) a panel cointegration model
between the indexes of the nominal exchange rates, the manufacturing price index of
each country and a geometric index of the manufacturing price index from the 16 OECD
countries. The selected base year for all these indexes is 2002 and the methodology about
the construction of these indexes is presented in section 5. The econometric results for
both models are shown in table 1 and 2.
Table 1: Panel Unit Root Test, Real Effective Exchange Rate
Number of Obs: 793 794 796 777 816 816
Breitung t-stat
A B C A
Common Unit Root Process 0.042 0.341 0.844 0.109 0.000 0.000
Notes: Model A adds a constant and a trend, model B adds only a constant and model C does not include anything.
and represent the Handri test statistics, where the null hypothesis considers that the series are stationary in levels or
around a deterministic trend, respectively. The bold squares indicate the rejection of the null hipotesis at 5% significance level.
Source: Own elaboration base on data from the Bureau of Labor Statistics.
Hadri Z-statLevin, Lin & Chu t*
12
Table 2: Panel Cointegration Test
The panel unit root tests on table 1, based on the assumption of a common unit root test,
show in general evidence of a common unit root for the 16 OECD countries. The test by
Levin, Lin, and Chu under the assumption of a constant and a trend, reject the null
hypothesis of presence of a common unit root for these countries. However, under the
similar assumption, the test by Breitung and Hadri do not reject the null hypothesis of a
common unit root.
The cointegrated model test in table 2 reports the result of the test by Pedroni and
Kao. In the case of Pedroni’s test, models A and C with the assumptions of common
1) Pedroni's test (816 Observations)
Alternative hypothesis: common AR coefs. (within-dimension)
Weighted Weighted Weighted
Statistic Prob. Statistic Prob. Statistic Prob.
Panel v-Statistic 0.01 0.50 2.94 0.00 -2.15 0.98
Panel rho-Statistic 0.35 0.64 -1.48 0.07 1.27 0.90
Panel PP-Statistic -0.76 0.23 -2.00 0.02 0.03 0.51
Panel ADF-Statistic -7.10 0.00 -7.30 0.00 -0.84 0.20
Alternative hypothesis: individual AR coefs. (between-dimension)
Statistic Prob. Statistic Prob. Statistic Prob.
Group rho-Statistic 1.14 0.87 -0.28 0.39 2.03 0.98
Group PP-Statistic -0.73 0.23 -1.94 0.03 -0.62 0.27
Group ADF-Statistic -6.29 0.00 -7.54 0.00 -2.51 0.01
2) Kao's test (816 Observations)
Null Hypothesis: No cointegration
t-Statistic Prob.
-0.69 0.25
Notes: Model A adds a constant and a trend, model B adds only a constant and model C does not include anything.
The bold squares indicate the rejection of the null hypothesis at 5% significance level.
Source: Own elaboration base on data from the Bureau of Labor Statistics.
Model B
Model A Model B Model C
Model A Model B Model C
13
AR’s coefficients and individual AR coefficients show that the indexes of the nominal
exchange rates, the manufacturing price index of each country and a geometric index of
the manufacturing price index from the 16 OECD countries do not hold a long run
cointegrating relation. Model B under Pedroni’s test, rejects the null hypothesis of no
cointegration, however, like the case of the panel unit root test, model B under Kao’ test
does not reject the null hypothesis of no cointegration.
The upshot of the foregoing results is that even with the expansion of the sample
offered by the panel data model, it is not clear that the relative PPP hypothesis with
tradable goods hold in the long-run. The results in table 1 and 2 suggest that the opposite
is more likely.
I.3 A Classical Framework for the Analysis of Real Exchange Rate
Theory of Competition, Profit Rate and Industrial Price Formation
The point of departure for Shaikh’s real exchange rate model is the classical theory of
competition, which can be traced back to the writings of Smith, Ricardo, and Marx. This
Classical approach considers competition as rivalry among firms where producers try to
obtain a higher share of their market by lowering costs and cutting prices. Thus, firms are
seen as constantly trying to reduce costs, mainly through suppression of the growth of
14
real wages and via the introduction, at intervals, of better techniques of production (i.e.
increased productivity due to technological changes).
From this Classical point of view, real competition implies that firms within an
industry do not necessarily face a similar cost structure (i.e., there are technology
differences among firms). In other words, the ‘Law of One Price’ (LOP) or what Adam
Smith and David Ricardo called ‘natural price’ and Karl Marx ‘price of production’,
forces firms within one industry to sell their products in the market at one price, which
eventually, due to the differences in the conditions of production, will bring about
different profit margins and profit rates to each firm in a given industry.
Therefore, the operation of the LOP (ignoring transportation cost, taxes, etc.) would
create a profit differential among firms within an industry (intra-industrial competition),
where the most favorable condition(s) of production (lower cost structure) will get a
greater profit rate, because the possession of the best productive technique can allow the
most competitive producer to enforce the market-selling price. That is, the price that
prevails in one particular market is not the average price of the industry but the least cost-
price determined by the most efficient producer(s) in that industry. This price is called the
regulating price and the producer is the regulating capital, as distinguished from the
average price and the average capital (Ruiz-Nápoles 1996; Shaikh 1999b). In turn, non-
regulating capitals will be forced by competition to sell at the same price, and will
therefore have a variety of profit rates determined by their own various conditions of
production (Shaikh 1999b, 2).
15
On the other hand, this ‘profit rate heterogeneity’ within one industry will not pass
unnoticed for long. The profit differential can easily call the attention of other capitalists
in other industries, i.e., the other regulating capitals, who are themselves eager for profits.
In this regard, the Classical tradition presupposes that the ‘free mobility of capital’ among
different industrial sectors (inter-industry competition) produces the tendency for a rough
equalization of profit rates between the previous unequal profit sectors, that is, between
the regulating capitals of different industries (Foley 1986; Ge 1993; Guerrero 1995;
Antonopoulos 1997; Sarich 2007).
Finally, for the Classical perspective, the dynamic and turbulent process of intra-and-
inter industrial competition does not require the corresponding equalization of wage rates
in order to reach a rough equalization of profit rates across industries (Shaikh 1991, 2). In
other words, flows of new capitalist investments into an industry can have a significant
impact on supply and prices without necessarily affecting sector wage rates (no full
employment is assumed). Where the mobility of labor is for any reason restricted, wage
differentials can persist even though profit rates may be equalized (Botwinick 1988, 250;
Shaikh 1991, 2).
Competition and Industrial Relative Prices at National Level
In this section we start by assuming that at national level intra-industry competition
makes regional markets integrated for any given commodity, so that the market prices of
either commodity i or j will be roughly the same between regional markets; here, we
assume that in general, LOP applies. On the other hand, inter-industry competition will
16
make the above common market prices of 𝑃𝑖 and 𝑃𝑗 themselves gravitate around their
respective regulating prices of production, 𝑃𝑖∗ and 𝑃𝑗
∗. Therefore, in the long run we have
equations 8 and 9:
𝑃𝑖 ≅ 𝑃𝑖∗ ; 𝑃𝑗 ≅ 𝑃𝑗
∗ ; 𝑃𝑖
𝑃𝑗≅
𝑃𝑖∗
𝑃𝑗∗ (8, 9 and 10)
By combining equations 8 and 9 we can get 10. From 10 follows that in the long run, the
ratio of any exchange between commodities i and j (i.e., the domestic terms of trade in
the economy) will be approximately equal to the relative regulating prices of those
bundles of commodities.
Equation 10 above also shows that within a nation, due to inter-industry competition,
the relative prices of production of products i and j are driven by the best-practice
producer, i.e., the regulating producer (Ge 1993, 254). The important issue with regard to
equation 10 is to understand what the main determinants of the level of these prices are
and to understand their implications with regard to the explanation of the long run
national terms of trade. To settle these issues, Shaikh (1984, 1998) reformulates
Pasinetti’s presentation (1977) of Sraffa’s (1963) price system to show that the relative
prices in question can be linked to unit costs, specifically, to total vertically integrated
unit labor costs, that is, the quantities of labor directly and indirectly embodied in each
physical unit of the commodities system (i.e., what Karl Marx called living and dead
labor, or labor values).
On the one hand, Passinetti (1977, 73) viewed the system of prices as one where
each commodity is produced via the use of a certain physical quantity of labor and a
17
given physical quantity of commodities required as means of production. Accordingly,
the value added in the economic system is assumed to be divided into two categories:
wages and profits. Thus, Passinetti’s representation of the price system can be depicted as
follows:
𝑝 = 𝑎𝑛𝑤 + 𝑝𝐴(1 + 𝜋) (11)
Where p denotes the (row) vector of prices, w (a scalar) the wage rate and 𝜋 (also a
scalar) the rate of profit. A denotes a matrix of inter-industry coefficients (or an Input-
Output matrix) and 𝑎𝑛 a row vector of direct labor coefficients.
Based on equation 11, after some algebraic manipulation, Passineti (1977, 76-77)
reached his general price system (equation 12), where the profit rate lies at some point
between its maximum value (), which corresponds to a theoretical possibility, where the
wage rate is zero, and a non-zero profit rate (�̅�) and a positive wage rate (𝑤).
𝑝 =1
1 + �̅�𝑎𝑛 [
1
1 + �̅�𝐼 − 𝐴]
−1
𝑤; (12) 𝑣 = 𝑎𝑛(𝐼 − 𝐴)−1𝑤 (13)
With equation 12, Passinetti (1977, 81) reinterpreted Leontief’s input-output model to
show Ricardo’s theory of value when in the equation above �̅� = 0 (the entire net product,
or surplus of the economic system, goes to wages); in this case prices become
proportional to the quantities of direct and indirect labor (vertically integrated labor
coefficients, 𝑣 in equation 13 above). However, as soon as �̅� > 0, the latter result is no
longer so. The quantities of indirect labor ((𝐼 − 𝐴)−1) come to acquire a greater ‘weight’
relative to the quantities of direct labor (𝑎𝑛). Therefore, according to Passinetti, the
18
proportionality between prices and quantities of embodied labor (i.e., the pure labor
theory of value) breaks down under conditions of positive profit rates.
On the other hand, Shaikh (1984) reformulated Passinetti’s system of prices using
Adam Smith’s long run competitive decomposition of price in order to show that relative
prices can be linked to relative total vertically integrated unit labor costs. In formal terms,
this decomposition of price can be illustrated as follows: let p, u, 𝜋, and m be the per unit
price, labor costs, gross profits, and material costs, respectively, of some given
commodity. Then by definition we may write 𝑝 = 𝑢 + 𝜋 + 𝑚. However, the materials
costs are simply the price of some bundle of materials, which in turn may be decomposed
into unit labor costs, profits, and their own material costs one (conceptual) stage back.
This decomposition can be repeated on the material costs of the materials bundle itself,
and so on, so that without any loss of generality Adam’s Smith long run competitive
decomposition of price can be represented as (Shaikh 1984, 1998):
𝑝 = 𝑢 + 𝜋 + 𝑚 = 𝑢 + 𝜋 + 𝑢(1) + 𝜋(1) + 𝑚(1)
= 𝑢 + 𝜋 + 𝑢(1) + 𝜋(1) + 𝑢(2) + 𝜋(2) + 𝑚(2) + ⋯ (14)
In equation 14, the new (residual) material cost 𝑚(1) is smaller than the original material
cost 𝑚. Furthermore, if we repeat the above process we can reduce 𝑚(1) to its wages,
profits and material costs, so that 𝑚(1) = 𝑢(2) + 𝜋(2) + 𝑚(2), and then in turn reduce this
remaining material cost to its components, and so on, until in the limit there is no residual
materials cost at all. In this way, regardless of how the price is actually determined, we
19
can always express it as an infinite series of wages and profits in conceptually receding
stages of production (Shaikh 1984, 66).
Along these lines, we can denote the sum of all the direct and indirect (vertically
integrated) unit labor costs by 𝑣 = 𝑢 + 𝑢(1) + 𝑢(2) + 𝑢(3) + ⋯ and that of all the direct
and indirect (vertically integrated) unit gross profits by 𝜋𝑇 = 𝜋 + 𝜋(1) + 𝜋(2) + 𝜋(3) + ⋯ .
Then equation 14 can be re-expressed as:
𝑝 = 𝑣 + 𝜋𝑇 = 𝑣(1 + 𝜌) (15)
Where 𝜌 = 𝜋𝑇 𝑣⁄ the average direct and indirect (i.e. the vertically integrated) profit-
wage ratio.
Here we have to bear in mind that this decomposition of price can be applied to any
price whatsoever, since it follows from an accounting identity. Hence, it follows that for
any two industries i and j, respectively, we can always express their relative prices in the
form expressed by equation 16, which Shaikh (1984) calls the Fundamental Equation of
Prices.
𝑝𝑖 𝑝𝑗⁄ ≡ (𝑣𝑖 𝑣𝑗⁄ ) ∗ (𝑧𝑖𝑗) (16)
Where 𝑧𝑖𝑗 = (1 + 𝜌𝑖) (1 + 𝜌
𝑗)⁄ = the ratio of the vertically integrated profit-wage ratios.
The Fundamental Equation of Prices shows that the relative prices of any two
commodities depends only on two terms: their relative vertically integrated unit labor
costs, and their relative vertically integrated gross profit margins. At this point, Shaikh
(1984), following Ricardo (1817), intertwines theoretical and empirical results from the
20
analysis of vertically integrated structures to assume that the second term of equation 16,
the ratio of the vertically integrated profit-wage ratios (𝑧𝑖𝑗), can be viewed as a
disturbance term with a stable value around 1. That is, Shaikh assumes that, due to the
high interconnectedness among industrial sectors, even large variations between 𝜌𝑖 and 𝜌𝑗
would induce only small variations in relative prices vis-à-vis relative vertically
integrated unit labor costs, so that 𝑧𝑖𝑗 is not structurally relevant in explaining the long-
run level of relative prices.
Hence, equation 17 below can be considered as an excellent approximation to equation
16.
𝑃𝑖
𝑃𝑗≅
𝑣𝑖
𝑣𝑗 (17)
Therefore, equation 17, which reflects Ricardo’s theory of value, where the relative
amounts of direct and indirect labor used in the production of two goods regulates the
exchange value of those goods over time, could be an excellent approximation to measure
industrial national terms of trade. In fact, empirical research at the national and
international level based on input-output data have shown that the vertically integrated
unit labor costs provide an excellent approximation (on the order of 85-90%) of relative
prices (see Shaikh 1984; Ochoa 1988; Bienenfeld 1988; Milberg and Elmslie 1992;
Chilcote 1997; Roman 1997; Ruiz-Nápoles 1996, 2010).
Finally, we have to recall that the development of equation 17 was carried out in real
terms. However, due to the fact that even within a single region, the consumer price index
(cpi) may differ because not all goods are tradable, then equation 18 takes into account
21
this difference between tradables and nontradables as follows: 𝒗𝑟 = 𝒗 𝑐𝑝𝑖⁄ = the real
vertically integrated unit labor costs, 𝑝𝑇𝑖= the price of some bundle of nationally traded
goods and 𝑐𝑝𝑖 𝑝𝑇⁄ = the adjustment for regional differences in tradable/nontradable
prices (Shaikh and Antonopoulos 2013, 209).
𝑃𝑖
𝑃𝑗≅ (
𝒗𝑟𝑖
𝒗𝑟𝑗) (
(𝑐𝑝𝑖𝑖 𝑝𝑇𝑖⁄ )
(𝑐𝑝𝑖𝑗 𝑝𝑇𝑗⁄ )
) (18)
Competition, Industrial Relative Prices and International Terms of Trade
Among economic theories, it is conventionally accepted that competition within a
country is regulated by the law of absolute advantage in costs (Guerrero 1995, 20; Félix
and Sorokin 2008, 290). However, as has been noted by Harrod (1957, chapter VI) and
Shaikh (1980, 1991), when it comes to international trade, neoclassical theory abandons
the principle of absolute costs as the main driving force of international competition and
replaces it with the principle of comparative costs. That is, for the theory of comparative
costs, countries will tend to specialize in the production of those goods that they produce
relatively more cheaply. According to this logic, backward nations will tend to specialize
in those goods where they have a lower relative cost/price relation, even if they produce
those goods inefficiently, or if they produce them more expensively than other advanced
nations (Salgado et al. 2010).
Moreover, for comparative advantage theory, regardless of the absolute competitive
position of each country, once international trade takes place, trade imbalances, that may
22
initially occur, would eventually tend to disappear via adjustments in domestic and
external prices (i.e., adjustment via the real exchange rate). That is, when neoclassical
trade models accept Hume’s price-specie-flow mechanism (which relies on the quantity
theory of money) and J. S. Mill’s ‘reciprocal national demand’ theory, they assume that
each country would always be forced to maintain a roughly constant real exchange rate
and a balanced trade balance. In other words, the often implicit assumption of
neoclassical trade models is that countries with net capital inflows (e.g., due to trade
surplus) will have a tendency to see increased domestic prices, while those countries with
net capital outflows (e.g., due to trade deficit) will have a tendency to experience a
reduction of domestic prices; the net result is that this ‘automatic mechanism’ would tend
to leave the real exchange rate constant over time (see Shaikh 1980, 216; Sarich 2006,
473; Martinez 2010, 60; Hunt and Lautzenheiser 2011, 190).
For our alternative real exchange rate model, if free trade is assumed to prevail,
absolute advantage would be reflected in international markets by the dominance of
regulating prices (from the regulating producer) for each and every tradable commodity.
That is, tradable goods are expected to sell for approximately the same international
market price, when expressed in a common currency, in every country after accounting
for disparities which arise due to differences in transportation costs, indirect taxes and so
on (we assume that LOP applies)7. Thus, market prices of regulating capitals are expected
to conform to the relative, vertically integrated, real unit labor costs of the regulating
firms (𝒗𝒓). So, under these circumstances, international competitiveness, measured by
7 It is worth noting that, since in itself, the LOP does not imply a long-run equilibrium real exchange rate
(at which balance of trade would be equal to zero), it is possible that the LOP prevails even when there is a
trade surplus or trade deficit (see Antonopoulos 1997, 11; Sarich 2007, 472).
23
changes in vertically integrated real unit labor costs of the respective tradable sectors,
leads to changes in international terms of trade. That is, if the national country’s
regulating capitals manage to reduce their costs of production and so their prices, then,
ceteris paribus, its terms of trade, would decline, depreciating its real exchange rate and
increasing its international competitiveness vis-à-vis its international competitor(s).
In order to bring our alternative framework to the analysis of international
competition, equation 18 above would have to be re-arranged in order to get a proper
measure of a country’s terms of trade, that is, the nominal exchange rate has to be taken
into account in order to proper reflect in one common currency changes in international
terms of trade. This step can be represented by equation 19:
𝒓𝒙𝒓𝒊𝒋 ≡ 𝒆𝒊𝒋 ∗𝑷
𝑷∗≈ (
𝒗𝒓
𝒗𝒓∗) (
(𝑐𝑝𝑖 𝑝𝑇⁄ )
(𝑐𝑝𝑖∗ 𝑝𝑇∗⁄ )
) 𝑜𝑟 𝒓𝒙𝒓𝒊𝒋 ≡ 𝒆𝒊𝒋 ∗𝑷
𝑷∗≈ (
𝒗𝒓
𝒗𝒓∗) (
𝜏
𝜏∗) (19)
In equation 19, the * represent foreign variables; whereas 𝒆𝒊𝒋 represents the nominal
exchange rate (foreign currency/domestic currency); 𝑷 and 𝑷∗ represents prices of
domestic and foreign tradable goods, respectively; and 𝜏 represents the adjustment for the
tradable/nontradable content (i.e. the openness) of the consumption bundle for each
country.
Equation 19 clearly shows that the long-term movements of real exchange rates are
roughly determined by the productive conditions of each country. That is, the
international competitiveness of a country is primarily based on its absolute advantage in
terms of product technology, labor productivity and real wages of its tradable sectors
(components embedded in 𝒗𝒓). Therefore, differences in these components of the real
24
cost of production (and regulating prices of production) between nations, would tend to
determine the changes in their international terms of trade and thus in their long run real
exchange rate.
It is worth pointing out that international regulating firms are not located in just one
country but they are spread out across different countries in the world. So, here we
assume that 𝒗𝒓 and 𝒗𝒓∗ are the overall best-practice costs of the tradable bundle in
question for each country. For this reason, those nations that manage ‘to create’ or ‘to
keep up’ a large number of regulating capitals in a huge range of industrial tradables
could, in general, be capable of getting favorable external conditions like a lower real
exchange rate; growing exports (and a higher market share) and thus growing profits
relative to other international firms producing similar goods; to be able to engage much
faster in more research and development; and to be capable of maintaining a sustained
trade surplus (e.g., China, Germany, and Japan).
Conversely, those nations that are not able to maintain a sufficient number of
regulating capitals in their industrial tradable sectors would, if unprotected from
competition, be vulnerable to unfavorable external conditions like a high (uncompetitive)
real exchange rate; decreasing exports (and a lower market share) and thus shrinking
profits; to be less able to do research and development; and to maintain persistent trade
deficits, which would have to be financed by growing external indebtedness.
Having mentioned the key elements of our alternative framework, it is now worth
mentioning that in order to undertake an empirical estimation of this theory for a myriad
of developed and developing countries, we will use the real unit labor cost (𝑈𝐿𝐶𝑟) from
25
the manufacturing sector as a proxy for the real vertically integrated unit labor costs and
producer price indexes (PPI) as a proxy for tradable prices. The first reason of using
direct unit labor cost and producer price indexes is due to their availability for the major
OECD countries over a sufficiently long time span. In order to estimate vertically
integrated costs, one would need input-output tables for all of the countries involved,
over a sufficient time span to permit the creation of an adequate time series. This is
beyond the scope of this study. The second reason is that other similar investigations
have used direct unit labor cost with good empirical results (see Roman 1997; Martinez
2010; Shaikh and Antonopoulos 2013). The latter reformulation is reflected in equation
20.
𝒓𝒙𝒓𝒊𝒋 ≡ 𝒆𝒊𝒋 ∗𝑷𝒊
𝑷𝒋∗ ≈
𝑼𝑳𝑪𝒓
𝑼𝑳𝑪𝒓∗
(𝜏
𝜏∗) (20)
The conclusions from this framework are the following:
A) The international competitive position of a country, measured by the real unit
costs of its tradable, pins its real exchange rate, so the real exchange rate is not
free to eliminate trade imbalances.
B) Neither flexible nor quasi-flexible exchange rate regimes will be able to
correct structural trade imbalances induced by international competition.
C) Trade surplus and trade deficits are direct consequences of the relative
competitive position of a nation. So exchange rate devaluation will only have
a temporary effect on national competitiveness if the general conditions of
production are not improved.
D) Exchange rate devaluations could be successful only to the extent that they
affect the real unit costs (via the real wage) and/or the tradable/nontradables
26
price ratio of consumer goods (𝜏), and those changes depend on the ability of
workers and consumers to resist such effects (Shaikh 1995; Shaikh and
Antonopoulos 2013).
I.4 The Role of Interest Rate Differential
Our fundamental equation 20 developed above, which shows that in the long-run, the real
exchange rate is structurally linked (through the intra-and-inter industrial competition
processes between international capitalists) to the relative real unit labor costs ratio
between two countries, does not necessary presuppose a strict rigid relationship inasmuch
as there can be some elements that are not in the equation 20 that could create a short-run
deviation between both variables (exchange rate misalignment: rxr><rulcr). However,
contrary to the ‘automatic’ price adjustment mechanism implicit in neoclassical and
monetarist trade models which arise due to the inflow and outflow of capital, leaving all
trading nations with balanced trade (and a roughly constant real exchange rate over time),
the Classical and Keynesian traditions use different mechanisms with regard to analyzing
capital inflows/outflows, that do not necessarily create a change in internal prices. That
is, for the latter traditions, trade imbalances ultimately have an impact on the internal
liquidity of the economies and thus on the internal interest rate rather than on relative
prices. Therefore, if the level of the national interest rate is high enough to create an
attractive interest rate differential (e.g., due to a high trade deficit), eventually, this
positive interest rate differential could trigger an important capital inflow into one nation
which could create an exchange rate misalignment, and as a consequence, a persistent
trade imbalance and external indebtedness.
27
Drawing on the Classical-Marxian and Keynesian monetary theories of production,
this section seeks to show the similarities between these two approaches and their
differences with the Quantitative Theory of Money (QTM) and the Post Keynesian
determination of the exchange rate. In so doing, we can understand why trade imbalances
ultimately have an impact on the internal liquidity of economies and on interest rate
differentials but not (at least not automatically) on relative prices as is posited by the PPP
hypothesis and the QTM.
Monetary Theory of Production, Interest Rate Differential, and Real Exchange Rate
Karl Marx was among the first in clarifying the monetary consequences of the
exports/imports of gold-money due to trade imbalances for the functioning of the banking
and production systems as a whole. In Capital volume III he shows that changes in gold-
money within an economy (e.g., due to unbalanced trade) lead solely to changes in bank
reserves rather than to changes in the price level (Marx 1981, Vol. III, 683). Marx also
shows that an increase in bank reserves is generally accompanied by a decline in the rate
of interest as the banks strive to convert reserves into capital. Conversely, a drop in bank
reserves below the legal minimum tends to lead to a rise in the rate of interest. Therefore
a decrease in the quantity of gold raises only the interest rate, whereas an increase in the
quantity of gold lowers the interest rate; and if not for the fact that the fluctuations in the
interest rate enter into the determination of cost-prices, or in the determination of demand
and supply, commodity prices would be wholly unaffected by them (Marx 1981, Vol. III,
685).
28
As we could observe in the foregoing paragraph, Marx was also fully aware that
changes in the rate of interest will eventually lead to changes in effective demand and
will affect the level of production (supply of commodities) via the credit system which
itself could help expand/lessen the production and circulation of commodities and
money-capital. However, Marx’s second criticism of the theory of money refers to the
fact that even an ongoing stimulus to effective demand (e.g., due to the extra money in
circulation and a lower national interest rate) does not have an increase pari passu on the
price level. That is, according to Marx (Marx 1981, Vol. III, 580) ‘in periods of
predominant credit, the velocity of the money increases faster than commodity-prices,
whereas in times of declining credit commodity prices fall slower than the velocity of
circulation’. Therefore, under these circumstances, although an increase in effective
demand may temporarily increase prices in some sectors, and hence raise profits in some
sectors, this increase must eventually lead to an expansion of production to meet the new
demand. And as production expands prices will fall until (all other things being equal)
they regain their original level (Shaikh, 1980).
According to Milberg (1994, 2002), Keynes argues against the supposed automatic
equilibrating forces of relative price adjustments that tend to maintain balanced trade and
full employment. That is, Keynes rejected the likelihood and efficiency of each of the
‘classical’ adjustment mechanisms –wages and exchange rates– when persistent
unemployment characterizes the economy. In this regard, Milberg (2002) has shown that
Keynes’ writings from 1929 to 1933 dealt with the issue of free trade, tariffs, and
29
protectionism and the argument that under conditions of less than full employment, a
policy of free trade could actually worsen economic conditions.
For instance, in the Macmillan Committee Report of 1930, Keynes wrote the following:
The fundamental ground of the free trade argument is that we ought to take the
McKenna Duties off in order that we should stop the making of cars and make
something else for which we are better suited. And the logical link between one and the
other is through this chain, and no other. Just like the Bank rate argument, it works
beautifully in a fluid system. But supposing we get jammed at the point of
unemployment, the alternative for a time may be between producing motor cars or
producing nothing (CW XX: 114).
In July of 1930, Keynes wrote to Prime Minister J. Ramsay MacDonald the following:
Free trade is profoundly based on the assumption of equilibrium conditions, and in
particular that wages always fall to their strict economic level. If they do not, and if for
several reasons we do not desire them to, then it is only by means of a tariff that the
ideal distribution of resources between different uses, which free trade aims at, can be
achieved; and there is an unanswerable theoretical case for a countervailing import duty
(and also for an export bounty) equivalent to the difference between the actual wage and
the economic wage…
I am no longer a free trader –and I believe that practically no-one else is – in the
old sense of the term to the extent of believing in a very high degree of national
specialization and in abandoning any industry which is unable for the time being to hold
its own. Where wages are immobile, this would be an extraordinarily dangerous
doctrine to follow (CW XX: 379-80).
According to Milberg (2002, 240), in the Macmillan Committee Report of 1930, one can
clearly see Keynes’ belief that under conditions of persistent unemployment, the
mechanisms which would otherwise transform a situation of differential comparative
costs into one of differences in absolute money costs and prices no longer operates. That
is, the wage adjustment simply does not take place to a sufficient degree to guarantee that
the ‘law’ of comparative advantage will dictate the commodity composition and the
balance of trade (Milberg 2002).
30
Roy Harrod, in his International Economics (Harrod 1957, chapter VI), also argues
against the self-equilibrating trade balance assumed by the QTM due to the effects of
gold inflows/outflows onto the level of prices. According to Harrod (1957, 113), the
competitiveness of a country will only be reduced (or increased) if the costs of production
are raised (or lowered). Therefore, to Harrod, if the alleged force of the gold
inflows/outflows is to be effective in this regard, it must have success in affecting the
following factors:
(a) The general level of wages and the general prices of production.
(b) The level of economic activity and employment in the country.
Harrod points out that (a) was the real basis for the ‘classical mechanism’, in the sense
that this theory assumes that an inflow of gold tends to raise prices (and an outflow to
reduce them), it being implicitly assumed that factors were fully employed. However,
Harrod (1957, 114) believed that the flows of gold do not have a direct impact on wages
due to the ‘notoriously somewhat sticky general level of wages’.
With regard to point (b), Harrod believed that short-term capital movements,
triggered by trade imbalances, exchange rate movements, and interest rate differentials,
could have an impact on the level of activity and employment. Nonetheless, the potential
impact of gold inflows/outflows on domestic prices and the trade balance would hinge on
whether or not the factors of production are fully employed. That is, if there is initially
considerable unemployment and low capacity utilization, an inflow of gold is not
expected to have an effect on prices or the general rate of wages. Conversely, if gold
31
inflows lead to a considerable stimulus of economic activity, to the extent that
employment and profits also increase substantially, then an increase of the general rate of
wages and prices can be expected. According to Harrod (1957, 134), only at this point of
higher growth of wages and prices, does the ‘classical mechanism’ come into play.
Finally, for Keynes, like Marx and Harrod, trade imbalances among countries will
tend to operate through changes in internal liquidity, from which changes in interest rate
differentials, investment and income will be derived, but not changes in relative prices
precisely because money-wage flexibility, inter alia, fails to bring about balanced trade.
Therefore, persistent trade imbalances –not balanced trade– are the likely outcomes.
𝑟𝑥𝑟𝑖𝑗 = 𝑓(𝑟𝑢𝑙𝑐𝑟𝑖𝑗∗ , 𝑖𝑖 − 𝑖∗) (21)
The central hypothesis of this paper, following Shaikh and Antonopoulos (2013) and
equation 21 above, is that the structural component of long-run real exchange rates is
explained by the relative adjusted real unit labor cost ratio (𝑟𝑢𝑙𝑐𝑟𝑖𝑗∗ ); meanwhile interest
rate differentials (𝒊𝒊 − 𝒊∗) caused chiefly by trade and payment imbalances may create
exchange rate misalignments in the short-to-medium term (around the long-run
condition). That is, a positive interest rate differential might induce foreign capital
inflows, put the capital account into surplus, raise the nominal exchange rate and hence
raise the real exchange rate (𝑒 (𝑃
𝑃∗)) –i.e. an appreciation of the real exchange rate. The
opposite outcome would be expected if a country with trade surplus maintains a zero or
negative interest rate differential.
32
At this point it is worth mentioning that for the Post Keynesian determination of the
‘nominal’ exchange rate, the interest rate differential is the main factor that determines
the changes of the exchange rate in the short-to-medium-to-long term (Harvey 2005,
2013). For the Post Keynesian theory, the breakdown of the Bretton Woods agreements
triggered a massive mobility of capitals mainly for speculative purposes rather than to
finance international trade transactions, so that the great volatility of the nominal
exchange rates of the recent years is mainly explained by capital inflows/outflows
pursuing positive profits derived by nominal exchange rate appreciation and positive
interest rate differentials.
While our approach agrees with the Post Keynesian explanation of the effect of the
interest rate differential (specially the real ones) on the nominal exchange rate, we believe
that such effect pertains to the short-to-medium term. As we argued above, the main
determinant of the real exchange rate in the long term is the relative adjusted real unit
labor cost ratio. The difference between both theories must be settled by robust empirical
analysis.
I.5 Statistical Analysis: RXR, RULC, Real Interest Rate Differential,
and Trade Balance
In this section, we include a graphical analysis, and we briefly describe the variables used
to compute the indexes of the real effective exchange rate, the adjusted real unit labor
cost ratio, and the real interest rate differential for 16 OECD countries for the period
33
1960-2010. In the appendix the reader can see the full methodology used for the
construction of these three indexes:
𝑟𝑥𝑟𝑐,𝑡 =𝑝𝑚𝑓𝑔𝑖,𝑡 ∗ 𝑒𝑖,𝑡 ∗ 100
((𝑝𝑚𝑔 ∗ 𝑒)𝑜𝑒𝑐𝑑)𝑡
(I)
𝑟𝑥𝑟 stands for the index of the real effective exchange rate, 𝑝𝑚𝑓𝑔 stands for the index of
manufacturing prices, 𝑒 stands for the nominal exchange rate (foreign currency relative to
U.S. dollar), (𝑝𝑚𝑓𝑔)𝑜𝑒𝑐𝑑 stands for a manufacturing price geometric traded weighted
average index for 16 OECD countries.
𝑟𝑢𝑙𝑐𝑎𝑑𝑗𝑟𝑎𝑡𝑖𝑜𝑖,𝑡 =𝑅𝑈𝐿𝐶𝑎𝑑𝑗𝑖,𝑡 ∗ 100
𝑅𝑈𝐿𝐶𝑎𝑑𝑗𝑂𝐸𝐶𝐷𝑡 (II)
𝑟𝑢𝑙𝑐𝑎𝑑𝑗𝑟𝑎𝑡𝑖𝑜 stands for the index of the adjusted real unit labor cost ratio, 𝑅𝑈𝐿𝐶𝑎𝑑𝑗
stands for the adjusted real unit labor cost, and 𝑅𝑈𝐿𝐶𝑎𝑑𝑗𝑂𝐸𝐶𝐷 denotes the adjusted real
unit labor cost geometric traded weighted average index for 16 OECD countries.
𝑟𝑒𝑎𝑙𝑖𝑛𝑡𝑟𝑎𝑡𝑒𝑑𝑖𝑓𝑓𝑡 = 𝑖𝑛𝑡𝑟𝑎𝑡𝑒𝑑𝑖𝑓𝑓𝑡 − 𝑔𝑝𝑝𝑖𝑟𝑎𝑡𝑖𝑜𝑡 (III)
𝑟𝑒𝑎𝑙𝑖𝑛𝑡𝑟𝑎𝑡𝑒𝑑𝑖𝑓𝑓 is the index of the real interest rate differential, 𝑖𝑛𝑡𝑟𝑎𝑡𝑒𝑑𝑖𝑓𝑓 stands for
nominal interest rate differential, and 𝑔𝑝𝑝𝑖𝑟𝑎𝑡𝑖𝑜 is a growth rate of a geometric traded-
weighted average of the producer price index for 16 OECD countries.
On the left hand side of figure 1, for each of the 16 OECD countries and Taiwan, we
plot the real effective exchange rate, the adjusted real unit labor cost ratio, and the trade
balance (X/M, goods) for the period 1960 to 2010. For the same period, on the right hand
side, we plot the deviation of the adjusted real unit labor cost ratio from the real effective
34
exchange rate (deviation) and the short-run real interest rate differential (measured in
percentage). Finally, it is worth mentioning that an increase of real effective exchange
rates means an appreciation, while a decrease a depreciation, in both cases with regard to
its trading partners; the same applies for the adjusted real unit labor cost ratio.
Figure 1: Real Effective Exchange Rate, Adjusted Real Effective ULC Ratio, Trade
Balance (X/M), and Real Interest Rate Differential
0.13
0.33
0.53
0.73
0.93
1.13
60
80
100
120
140
160
180
1960 1965 1970 1975 1980 1985 1990 1995 2000 2005 2010
Australia
Real Effective Exchange Rate Adjusted Real Effective ULC Exports/Imports
-20
-10
0
10
20
30
40
50
60
-12
-10
-8
-6
-4
-2
0
2
4
6
1961 1966 1971 1976 1981 1986 1991 1996 2001 2006
%
Australia
real.int.rate.diff Deviation
0.85
0.9
0.95
1
1.05
1.1
70
80
90
100
110
120
130
140
150
1960 1965 1970 1975 1980 1985 1990 1995 2000 2005 2010
BelgiumReal Effective Exchange Rate Adjusted Real Effective ULC Exports/Imports
-15
-10
-5
0
5
10
15
20
25
30
35
-10
-5
0
5
10
15
20
1961 1966 1971 1976 1981 1986 1991 1996 2001 2006
%
Belgiumreal.int.rate.diff Deviation
0.5
0.6
0.7
0.8
0.9
1
1.1
1.2
60
80
100
120
140
160
1960 1965 1970 1975 1980 1985 1990 1995 2000 2005 2010
CanadaReal Effective Exchange Rate Adjusted Real Effective ULC Exports/Imports
-35
-30
-25
-20
-15
-10
-5
0
5
10
-6
-5
-4
-3
-2
-1
0
1
2
3
4
5
1961 1966 1971 1976 1981 1986 1991 1996 2001 2006
%
Canada
real.int.rate.diff Deviation
35
0.15
0.35
0.55
0.75
0.95
1.15
55
65
75
85
95
105
115
125
135
1960 1965 1970 1975 1980 1985 1990 1995 2000 2005 2010
DenmarkReal Effective Exchange Rate Adjusted Real Effective ULC Exports/Imports
-25
-20
-15
-10
-5
0
5
-2
0
2
4
6
8
10
12
1961 1966 1971 1976 1981 1986 1991 1996 2001 2006
%
Denmark
real.int.rate.diff Deviation
-0.5
-0.3
-0.1
0.1
0.3
0.5
0.7
0.9
1.1
1.3
1.5
70
90
110
130
150
170
1960 1965 1970 1975 1980 1985 1990 1995 2000 2005 2010
FinlandReal Effective Exchange Rate Adjusted Real Effective ULC Exports/Imports
-40
-30
-20
-10
0
10
20
30
-10
-8
-6
-4
-2
0
2
4
6
1961 1966 1971 1976 1981 1986 1991 1996 2001 2006%
Finland
real.int.rate.diff Deviation
0.2
0.4
0.6
0.8
1
60
70
80
90
100
110
120
130
140
1960 1965 1970 1975 1980 1985 1990 1995 2000 2005 2010
FranceReal Effective Exchange Rate Adjusted Real Effective ULC Exports/Imports
-15
-10
-5
0
5
10
15
20
25
30
35
40
-10
-5
0
5
10
15
1961 1966 1971 1976 1981 1986 1991 1996 2001 2006
%
Francereal.int.rate.diff Deviation
0.1
0.3
0.5
0.7
0.9
1.1
1.3
45
55
65
75
85
95
105
115
125
1960 1965 1970 1975 1980 1985 1990 1995 2000 2005 2010
GermanyReal Effective Exchange Rate Adjusted Real Effective ULC Exports/Imports
-10
-5
0
5
10
15
20
25
-4
-2
0
2
4
6
8
1961 1966 1971 1976 1981 1986 1991 1996 2001 2006
%
Germany
real.int.rate.diff Deviation
36
-0.2
0
0.2
0.4
0.6
0.8
1
1.2
70
80
90
100
110
120
130
140
1960 1965 1970 1975 1980 1985 1990 1995 2000 2005 2010
ItalyReal Effective Exchange Rate Adjusted Real Effective ULC Exports/Imports
-15
-10
-5
0
5
10
15
20
-20
-17
-14
-11
-8
-5
-2
1
4
7
10
1961 1966 1971 1976 1981 1986 1991 1996 2001 2006
%
Italy
real.int.rate.diff Deviation
-0.1
0.1
0.3
0.5
0.7
0.9
1.1
1.3
1.5
50
60
70
80
90
100
110
120
130
140
1960 1965 1970 1975 1980 1985 1990 1995 2000 2005 2010
JapanReal Effective Exchange Rate Adjusted Real Effective ULC Exports/Imports
-20
-15
-10
-5
0
5
10
15
20
25
30
-10
-8
-6
-4
-2
0
2
4
6
1961 1966 1971 1976 1981 1986 1991 1996 2001 2006%
Japan
real.int.rate.diff Deviation
0.05
0.25
0.45
0.65
0.85
1.05
1.25
1.45
20
70
120
170
220
270
320
1960 1965 1970 1975 1980 1985 1990 1995 2000 2005 2010
KoreaReal Effective Exchange Rate Adjusted Real Effective ULC Exports/Imports
-140
-120
-100
-80
-60
-40
-20
0
20
40
60
80
-40
-30
-20
-10
0
10
20
30
1961 1966 1971 1976 1981 1986 1991 1996 2001 2006
%
Koreareal.int.rate.diff Deviation
0.2
0.3
0.4
0.5
0.6
0.7
0.8
0.9
1
1.1
1.2
50
70
90
110
130
150
1960 1965 1970 1975 1980 1985 1990 1995 2000 2005 2010
NetherlandsReal Effective Exchange Rate Adjusted Real Effective ULC Exports/Imports
-30
-20
-10
0
10
20
30
40
-8
-6
-4
-2
0
2
4
6
8
10
12
1961 1966 1971 1976 1981 1986 1991 1996 2001 2006
%
Netherlands
real.int.rate.diff Deviation
37
-0.2
0.3
0.8
1.3
1.8
20
40
60
80
100
120
140
1960 1965 1970 1975 1980 1985 1990 1995 2000 2005 2010
NorwayReal Effective Exchange Rate Adjusted Real Effective ULC Exports/Imports
-25
-20
-15
-10
-5
0
5
10
15
20
25
30
-12
-10
-8
-6
-4
-2
0
2
4
6
8
10
1961 1966 1971 1976 1981 1986 1991 1996 2001 2006
%
Norway
real.int.rate.diff Deviation
0.3
0.4
0.5
0.6
0.7
0.8
0.9
20
40
60
80
100
120
140
160
1960 1965 1970 1975 1980 1985 1990 1995 2000 2005 2010
SpainReal Effective Exchange Rate Adjusted Real Effective ULC Exports/Imports
-20
-10
0
10
20
30
40
50
60
-14
-12
-10
-8
-6
-4
-2
0
2
4
6
8
1961 1966 1971 1976 1981 1986 1991 1996 2001 2006%
Spain
real.int.rate.diff Deviation
0
0.2
0.4
0.6
0.8
1
1.2
70
90
110
130
150
170
190
1960 1965 1970 1975 1980 1985 1990 1995 2000 2005 2010
Sweden
Real Effective Exchange Rate Adjusted Real Effective ULC Exports/Imports
-30
-20
-10
0
10
20
30
-8
-6
-4
-2
0
2
4
6
8
1961 1966 1971 1976 1981 1986 1991 1996 2001 2006
%
Swedenreal.int.rate.diff Deviation
0.00
0.20
0.40
0.60
0.80
1.00
1.20
1.40
1.60
30
50
70
90
110
130
150
170
1976 1979 1982 1985 1988 1991 1994 1997 2000 2003 2006 2009
TaiwanReal Effective Exchange Rate Adjusted Real Effective ULC Exports/Imports
-20
-10
0
10
20
30
40
50
60
70
80
90
1976 1979 1982 1985 1988 1991 1994 1997 2000 2003 2006 2009
TaiwanDeviation
38
To a large extent, the graphs displayed above seem to corroborate the main hypotheses of
this paper. On the one hand, the adjusted real effective ULC ratio seems to determine the
level and the long-term trajectory of the real effective exchange rate. That is, for the 16
OECD countries and Taiwan, both measures move very closely together on their long-run
path. Furthermore, we can also see that for all these countries the adjusted real effective
ULC ratio and the real effective exchange rate follow a strong opposite short and long-
term path with the trade balance, suggesting that the adjusted real effective ULC ratio and
the real effective exchange rate capture an important degree of international
competitiveness.
On the other hand, on the right hand side graphs, the short run fluctuations or
deviations of the real effective exchange rate from its center of gravity (i.e., adjusted
𝑼𝑳𝑪∗ 𝑼𝑳𝑪⁄ ) seem to be explained by the real effective interest rate differential; that is,
-0.2
0
0.2
0.4
0.6
0.8
1
30
40
50
60
70
80
90
100
110
120
130
1960 1965 1970 1975 1980 1985 1990 1995 2000 2005 2010
United KingdomReal Effective Exchange Rate Adjusted Real Effective ULC Exports/Imports
-40
-30
-20
-10
0
10
20
-15
-10
-5
0
5
10
15
20
1961 1966 1971 1976 1981 1986 1991 1996 2001 2006
%
United Kingdomreal.int.rate.diff Deviation
0.5
0.7
0.9
1.1
1.3
1.5
1.7
30
50
70
90
110
130
150
170
190
1960 1965 1970 1975 1980 1985 1990 1995 2000 2005 2010
United StatesReal Effective Exchange Rate Adjusted Real Effective ULC Exports/Imports
-35
-30
-25
-20
-15
-10
-5
0
5
10
-5
-4
-3
-2
-1
0
1
2
3
4
5
1961 1966 1971 1976 1981 1986 1991 1996 2001 2006
%
United Statesreal.int.rate.diff Deviation
39
these deviations seem to be caused by positive and negative changes in real interest rate
differentials. However, for a few countries, the correlation between the deviations and the
real effective interest rate differentials seem not so strong, and for a few years, the latter
correlation seems to be negative instead of positive (e.g., Finland and Norway).
Nonetheless, by and large, the correlation between the deviations and the real effective
interest rate differentials seem to be strong and positive.
On the left hand side of figure 2, for the case of Argentina, El Salvador, and Mexico,
we plot the bilateral real exchange rate, the real unit labor cost ratio with respect to the
U.S., and the trade balance (X/M, goods) mainly from the period 1960 to 2010. On the
right hand side, for different periods according to the availability of data, we plot the
deviation of the real unit labor cost ratio from the bilateral real exchange rate (deviation)
and the real interest rate differential (measure in percentage).
Figure 2: Real Exchange Rate, Real ULC Ratio, Trade Balance (X/M), and Real Interest
Rate Differential
-1.5
-1
-0.5
0
0.5
1
1.5
2
2.5
3
3.5
20
70
120
170
220
270
1960 1965 1970 1975 1980 1985 1990 1995 2000 2005
Argentina-US
Real Effective Exchange Rate Adjusted Real Effective ULC Exports/Imports
-100
-50
0
50
100
150
1960 1965 1970 1975 1980 1985 1990 1995 2000 2005
-15
-10
-5
0
5
10
15
20
25
30
%
Argentina-US
real.int.rate.diff Deviation
40
The graphs displayed in figure 2 also seem to corroborate the main hypotheses of this
paper. On the one hand, the real ULC ratio seems to determine the level and the long-
term trajectory of the bilateral real exchange rate, that is, for the case of Argentina, El
Salvador, and Mexico, the latter both measures appear to move very closely each other on
their long-run path. Furthermore, we can see that for these three countries the adjusted
real effective ULC ratio and the real effective exchange rate follow a strong opposite
short and long-term path with the trade balance, which means that the adjusted real
effective ULC ratio and the real effective exchange rate capture an important degree of
international competitiveness.
On the other hand, despite the limited interest rate data for Argentina and Mexico (no
data at all for the case of El Salvador), the short run fluctuations or deviations of the
bilateral real exchange rate from its center of gravity (i.e. 𝑼𝑳𝑪𝒓 𝑼𝑳𝑪𝑟∗⁄ ) seem to be
0.1
0.2
0.3
0.4
0.5
0.6
15
35
55
75
95
115
135
1960 1965 1970 1975 1980 1985 1990 1995 2000 2005 2010
El Salvador-US
Real Effective Exchange Rate Adjusted Real Effective ULC Exports/Imports
-50
-40
-30
-20
-10
0
10
20
30
40
1960 1965 1970 1975 1980 1985 1990 1995 2000 2005 2010
%
El Salvador-US
Deviation
-0.3
0.1
0.5
0.9
1.3
1.7
2.1
30
70
110
150
190
230
270
1970 1975 1980 1985 1990 1995 2000 2005 2010
Mexico-US
Real Effective Exchange Rate Adjusted Real Effective ULC Exports/Imports
-200
-150
-100
-50
0
50
100
-50
-40
-30
-20
-10
0
10
20
30
1976 1981 1986 1991 1996 2001 2006 2011%
Mexico-USreal.int.rate.diff Deviation
41
explained by the real interest rate differential, that is, these deviations seem to be caused
by positive and negative changes in real interest rate differentials.
I.6 Empirical Evidence of Alternative Real Exchange Rate
Determination Based on ARDL-ECM Modelling
Given that our variables may be potentially I(1), the nature of their long-run trend
relationship can be studied via the use of cointegration analysis and error correction
models (ECM). The use of the ECM framework has three advantages. First, an ECM
incorporates both short-run and long-run (or trend) relationships. Second, the sign and
significance of the error correction coefficient (ECC) provide an indication of Granger
causality in a non-stationary context (Enders 1995:367). For an error correction model to
be stable the ECC has to satisfy the following stability criterion: -1 < ECC ≤ 0 (Hill et al.
2011, 500). Third, for a stable ECM, the absolute value of the ECC provides an
indication of the time that it takes for the variables to reach an approximate equilibrium
relationship.
We deployed the autoregressive distributed lag (ARDL) modelling using Microfit
5.0 (Pesaran and Pesaran 2009). The ARDL framework does not require prior unit root
testing, thereby eliminating one source of error involved with such tests on relatively
small sample sizes.
Table 3 lists the long-run parameters and the ECC of the seventeen OECD countries
(including Taiwan) and three developing countries (Argentina, El Salvador, and Mexico)
that showed evidence of being cointegrated, with the log of the adjusted real unit labor
42
cost ratio and the real interest rate differential acting as the long-run forcing variables.
The cointegrating results follow from the fact that the F statistic is above the upper
bound. The ECC for each country is found to be negative, below one and statistically
significant when the first difference of the log of the real effective exchange rate is the
dependent variable, suggesting that the log of the adjusted real unit labor cost ratio and
the real interest rate differential for any country Granger-causes the log of the real
effective exchange rate.
Moreover, the last column of table 3 lists the correlation coefficients of the Trade
Balance (TB) and the adjusted real unit labor cost ratio (RULCR). This correlation
coefficient turned out to be negative (although with different degrees) for sixteen
countries, which suggests that a relative reduction of the real unit labor cost tends to
improve the international competitiveness of these nations. However, this correlation
coefficient turned out to be positive for Germany, Japan, Norway, and the US, which
suggests that a relative reduction of the real unit labor cost tends to worsen the
international competitiveness of these nations (more on this below).
43
Table 3: Econometric Results
It is worth mentioning that we always found a cointegrating relation between the log of
the real effective exchange rate and the log of the real adjusted unit labor cost ratio, and
in most cases, also with the real interest rate differential, that is, only in one case for the
estimated period, the real interest rate differential did not cointegrate with the other two
variables in our ECM framework (this was the case of Sweden), which suggests that this
Time Trend LRULCR Real.Int.Rate.Diff ECM Regression Date F Statistic
Intercept Coefficient Coefficient Coefficient Coefficient & ARDL Order Correlations
(t-Ratio) (t-Ratio) (t-Ratio) (t-Ratio) (t-Ratio) Lower And TB & RULCR
[p-value] [p-value] [p-value] [p-value] [p-value]rulc → rxr or rxr →
rulc Upper Bounds
1.077 -0.7580 1966-2007 5.33
[120.20] [-3.22] ARDL (6, 5) -0.157
[.000] [0.003] 3.31 [I(0)]* 1991-2007
Both 4.32 [I(1)]*
-5.014 2.116 -0.0439 -0.518 1967-2010 6.66
[ -2.761] [5.348] [ -2.515] [-4.104] ARDL (2, 3, 5) -0.191
[.010] [.000] [.017] [.000] 4.09 [I(0)]* 1960-2010
rulc → rxr 5.26 [I(1)]*
0.011 0.911 0.0107 -0.527 1980-2010 17.73
[5.481] [53.727] [1.992] [ -6.393] ARDL (1, 0, 2) -0.006
[0.000] [.000] [.057] [.000] 4.25 [I(0)]* 1997-2010
rulc → rxr 5.41 [I(1)]*
1.006 -0.229 -0.302 1984-2010 10.23
[77.007] [-2.248] [-2.715] ARDL (2, 4, 6) -0.378
[.000] [.042] [.016] 3.10 [I(0)]* 1960-2010
Both 4.31 [I(1)]*
1.01 -0.026 -0.5 1973-2010 9.35
[150.044] [-3.267] [-6.730] ARDL (1, 0, 6) -0.207
[.000] [.003] [.000] 2.95 [I(0)]* 1988-2010
rulc → rxr 4.17 [I(1)]*
1.063 -0.0773 1961-2010 13.546
[32.529] [-3.436] ARDL (1, 1, 1) -0.292
[.000] [.001] 2.87 [I(0)]* 1994-2010
rulc → rxr 4.03 [I(1)]*
1.003 -0.027 -0.474 1971-2010 4.38
[242.510] [-2.297] [-2.773] ARDL (5, 3, 6) -0.549
[.000] [.031] [.010] 2.94 [I(0)]* 1960-2010
Both 4.12 [I(1)]*
-0.009 1.097 0.027 -0.533 1971-2010 4.879
[-8.414] [127.963] [2.066] [-3.451] ARDL (2, 6, 6) -0.201
[.000] [.000] [.050] [.002] 3.41 [I(0)]** 1960-2010
Both 4.44 [I(1)]**
1.021 0.016 -0.51 1970-2010 6.23
[239.79] [1.886] [-4,69] D86 ARDL (6, 6, 0) 0.294
[.000] [0.071] [.000] D00 2.953 [I(0)]* 1960-2010
rulc → rxr 4.091 [I(1)]*
1.105 0.758 0.018 -0.867 1981-2010 5.496
[1.917] [5.994] [2.167] [-4.401] ARDL (5, 0, 1) -0.127
[.069] [.000] [.042] [.000] 4.27 [I(0)]* 1960-2010
rulc → rxr 5.44 [I(1)]*
Dummies
Argentina N. A. N. A. N. A. N. A.
Australia N. A. N. A.
Belgium N. A. N. A.
Finland N. A. N. A. N. A.
Canada N. A. N. A. N. A.
Denmark N. A. N. A. N. A.
El
Salvador N. A. N. A. N. A. D86
France N. A. N. A.
Germany N. A. N. A.
Italy N. A. N. A.
44
variable was not relevant to explain the changes in the log of the real effective exchange
rate. In seven other cases, although the real interest rate differential held a cointegrating
long-run relationship with the other two variables, its sign was negative (this was the case
of Australia, Canada, Denmark, Finland, Japan, Norway, and Spain). This result is a very
well-known puzzle in international finance, which means that positive changes in real
interest rate differentials tend to depreciate the real exchange rate. Multiple
investigations, in the tradition of covered and uncovered interest parity have also
encountered this counterintuitive result for several countries and periods (Engel 2013).
However, for the nine remaining countries, the sign of the real interest rate differential
was the expected one, suggesting that positive interest rate differentials tend to appreciate
the real exchange rate (see table 3).
45
Table 3 (Cont´d): Econometric Results
It is quite likely that changes in the real exchange rate could lead to corresponding
variations of the adjusted real unit labor cost ratio and the real interest rate differential,
for example by altering the growth rate of all domestic prices. We therefore made the log
of the adjusted real unit labor cost ratio and the real interest rate differential the
Time Trend LRULCR Real.Int.Rate.Diff ECM Regression Date F Statistic
Intercept Coefficient Coefficient Coefficient Coefficient & ARDL Order Correlations
(t-Ratio) (t-Ratio) (t-Ratio) (t-Ratio) (t-Ratio) Lower And TB & RULCR
[p-value] [p-value] [p-value] [p-value] [p-value]rulc → rxr or rxr →
rulc Upper Bounds
1.012 -0.067 -0.835 1977-2010 7.47
[301.387] [-4.362] [-4.705] ARDL (3, 3, 2) 0.731
[.000] [.000] [.000] 3.01 [I(0)]* 1977-2010
rulc → rxr 4.23 [I(1)]*
3.691 -0.008 0.266 0.0107 -0.629 1965-2010 7.36
[7.623] [-3.634] [2.210] [2.108] [-4.497] ARDL (1, 4, 1) -0.465
[.000] [.001] [.034] [.042] [.000] 5.29 [I(0)]* 1960-2010
rulc → rxr 6.35 [I(1)]*
3.902 0.179 0.0031 -0.676 1982-2011 7.32
[14.706] [3.397] [1.893] [-5.904] D86 ARDL (1, 0, 0) -0.312
[.000] [.002] [.070] [.000] D95 4.24 [I(0)]* 1982-2011
rulc → rxr 5.42 [I(1)]*
0.009 0.918 0.0381 -0.964 1990-2010 6.78
[5.896] [63.585] [2.745] [-4.402] ARDL (4, 4, 4) -0.814
[0.001] [.000] [0.033] [.002] 4.59 [I(0)]* 1971-2010
Both 5.92 [I(1)]*
0.008 0.955 -0.094 -0.152 1968-2010 5.11
[2.42] [35.75] [-2.53] [-2.51] ARDL (1, 3, 4) 0.91
[0.021] [.000] [.016] [.017] 3.38 [I(0)]** 1960-2010
Both 4.41 [I(1)]**
1.01 -0.0711 -0.32 1982-2010 7.33
[164.1] [-2.13] [-3.22] ARDL (4, 3, 4) -0.086
[.000] [0.048] [.005] 3.04 [I(0)]* 1984-2010
rulc → rxr 4.26 [I(1)]*
0.992 -0.187 1963-2010 3.6
[117.1] [-2.445] ARDL (3, 2) -0.807
[.000] [.019] 2.47 [I(0)]** 1960-2010
rulc → rxr 3.36 [I(1)]**
0.986 -0.363 1986-2009 21.017
[160.3] [-5.76] ARDL (2, 2) -0.503
[.000] [.000] 3.56 [I(0)]* 1986-2010
Both 4.62 [I(1)]*
0.972 0.043 -0.409 1975-2010 9.77
[200.2] [5.47] [-6.3] D97 ARDL (1, 0, 1, 1) -0.21
[.000] [.000] [.000] 2.68 [I(0)]* 1960-2010
rulc → rxr 4.06 [I(1)]*
0.972 0.0271 -0.411 1963-2010 4.959
[210.9] [2.42] [-4.6] ARDL (2, 0, 0) 0.881
[.000] [.019] [.000] 2.85 [I(0)]* 1960-2010
Both 4.05 [I(1)]*
*95% confidence level
**90% confidence level
A dummy such as D95 takes on the value of 1 in 1995 and is zero in other periods. The same applies to other dummy variables.
N. A. stands for Not Applicable
Dummies
Japan N. A. N. A. N. A.
Korea N. A.
Mexico N. A.
Netherlan
dsN. A. N. A.
Norway N. A. N. A.
Spain N. A. N. A. N. A.
Taiwan N. A. N. A. N. A. N. A.
Sweden N. A. N. A. N. A. N. A.
N. A.
UK N. A. N. A.
US N. A. N. A.
46
dependent variables in two ECM for each country. When we took the log of the adjusted
real unit labor cost ratio as the dependent variable, we found that in the cases of
Argentina, Canada, Finland, France, Netherlands, Norway, Taiwan, and the US, the F-
statistic and the ECC were statistically significant, which means that for these eight
countries there are multiple feedbacks between the adjusted real unit labor cost ratio and
the real effective exchange rate. In the case of the other twelve countries, the direction of
the causality goes from the adjusted real unit labor cost ratio to the real effective
exchange rate (see table 3). When we took the real interest rate differential as the
dependent variables, in none of the cases were both the F statistic and the ECC
statistically significant.
Table 4: Econometric Results, Trade Balance Model
Finally, for those countries where we found a positive correlation between the Trade
Balance (TB) and the RULCR, we decided, using the ARDL-ECM framework, to do a
full model for the log of the Trade Balance (LTB) in order to estimate the long-run
Time Trend LRULCR LGDP LWGDP ECM Regression Date F Statistic
Intercept Coefficient Coefficient Coefficient Coefficient Coefficient & ARDL Order
(t-Ratio) (t-Ratio) (t-Ratio) (t-Ratio) (t-Ratio) (t-Ratio) Lower And
[p-value] [p-value] [p-value] [p-value] [p-value] [p-value] Upper Bounds
-0.452 -0.405 0.4606 -0.65000 1982-2010 5.46
[-2.05] [-2.18] [2.85] [-3.86] ARDL (2, 3, 1, 3)
[.055] [0.043] [0.011] [0.001] 2.78 [I(0)]*
4.10 [I(1)]*
-1.989 0.604 -0.478 1976-2010 5.12
[-2.831] [ 2.89] [-4.00] ARDL (2, 6, 0)
[.009] [.008] [.000] 2.99 [I(0)]*
4.23 [I(1)]*
3.107 -1.05 -0.257 1968-2010 5.04
[2.09] [-1.95] [ -3.59] ARDL (1, 3, 3)
[.044] [.058] [.001] 2.90 [I(0)]*
4.08 [I(1)]*
19.78 -1.019 -0.963 -0.353 1964-2010 5.98
[2.98] [-2.02] [-3.57] [-3.79] D73 ARDL (4, 0, 4)
[0.024] [.051] [.001] [.001] 4.09 [I(0)]*
5.19 [I(1)]*
Dummies
Germany N. A. N. A. N. A.
N. A.
Norway N. A. N. A. N. A. N. A.
US N. A. N. A.
Japan N. A. N. A. N. A.
47
relationship between the TB and the RULCR, so we took as independent variables the log
of the RULCR, the log of the real GDP, and the log of the World real GDP. The results
on table 4 show that for the cases of Germany, Japan, and the US, there is an equilibrium
negative long-run relationship (elasticity) between the log of the Trade Balance and the
log of the RULCR, which suggests that a relative reduction of the real unit labor cost
tends to improve the international competitiveness of these nations. For the case of
Norway, the long-run relationship (elasticity) between the log of the TB and the log of
the RULCR turned out to be positive, which suggests that for this country, a relative
increase of the real unit labor cost does not have a negative impact on its trade balance
(perhaps this is a result of the composition of Norway’s main exports: oil and technology
associated to the extraction of oil). For the four countries, the log of their real GDP turned
out to be statistically significant and negative; only for the case of Japan the latter
relationship turned out to be positive. Finally, only for the case of Germany the log of the
World real GDP turned out to be significant and positive, which suggests that to an
important degree Germany’s trade balance depends on the world real GDP performance.
48
I.7 Concluding Remarks
An important conclusion of our alternative approach is that neither flexible nor quasi-
flexible exchange rate regimes will be able to correct structural trade imbalances induced
by international competition. That is, trade surpluses and trade deficits are direct
consequences of the relative competitive positions of nations. So exchange rate
devaluations will only have a temporary effect on national competitiveness if the general
conditions of production are not improved.
Drawing on Marx, Harrod, and Keynes, we presented theoretical reasons for why
trade imbalances ultimately have an impact on the internal liquidity of the economies and
thus on the internal interest rate rather than on relative prices. Therefore, if the level of
the national interest rate is high enough to create an attractive interest rate differential
(e.g., due to a high trade deficit), eventually this positive interest rate differential could
trigger an important capital inflow into one nation which could create an exchange rate
misalignment, and as a consequence, a persistent trade imbalance and external
indebtedness.
Our econometric results using the ARDL-ECM framework confirmed the main
hypothesis of this paper, namely that there is an equilibrium long-run relationship
between the real effective exchange rate and the adjusted real unit labor cost ratio for 16
OECD countries, Taiwan, and 3 developing countries. However, when we investigated
the possibility of a long-run relationship among the real effective exchange rate, the
adjusted real unit labor cost ratio and the short-run real interest rate differential, we could
only find a meaningful cointegrating vector with correct signs for the cases of Belgium,
49
France, Germany, Italy, Korea, Mexico, Netherlands, UK, and the US, which indicates
that for these countries a positive real interest rate differential tends to appreciate their
real exchange rates.
Our correlation and cointegration analysis suggests that for all the countries under
analysis but Norway, there is a negative relationship between the Trade Balance and the
RULCR, which hints that a relative reduction of the real adjusted unit labor cost tends to
improve the international competitiveness of these nations.
In short, the generalized modernization of technology to raise productivity and to
lower unit labor costs, is the only long-term solution to the problem of competitive
disadvantage. More precisely, cost reductions can only be produced by the introduction
of more efficient technologies, or, in the short-term, by the reduction of the real wage
rate. Therefore, as long as the least competitive economies at the international level do
not improve their general technical conditions of production, these countries national
industries will be structurally uncompetitive and as a result, these countries could see
permanent trade deficits.
50
References
Antonopoulos, R. (1997), "An Alternative Theory of Real Exchange Rate Determination
for the Greek Economy. New York: Unpublished PhD dissertation, New School for
Social Research.
Botwinick, M. (1988), Regularities in price changes as an effect of change in distribution.
Cambridge Journal of Economics.
Foley, D. (1986). Understanding Capital: Marx's Economic Theory. Cambridge MA,
Harvard University Press.
Enders, Walters (1995), Applied Econometric Times Series. John& Sons, Inc.
Feliz y Sorokin (2008), Rigidez estructural del tipo de cambio? El caso de Argentina a la
luz de un enfoque marxista. Interpretaciones heterodoxas de las crisis económicas en
Argentina y sus efectos sociales. Compiladores: Fernando Toledo y Julio Neffa. Eds.
Trabajo & Sociedad y Conicet.
Ge, Qi (1993), An Alternative Theory of the Terms of Trade. New York: Unpublished
PhD dissertation, New School for Social Research.
Guerrero, Diego (1995), Competitividad teoria y practica. Editorial Ariel.
Harrod, Roy (1957), International Economics. Chicago: The University of Chicago Press.
Harvey, John. Exchange Rates and Trade Flows: A Post Keynesian Analysis, 2013.
Online document, accessed May 11:
http://www.economicsnetwork.ac.uk/heterodox/lecturenotes
— (2005) Post Keynesian versus Neoclassical Explanations of Exchange Rate
Movements: A Short Look at the Long Run, Texas Christian University, Working
Paper Nr. 05-01.
Hill, R. Carter, Griffiths, Willian E., and Lim, Guay (2011), Principles of Econometrics.
4th edition . John Wiley & Sons.
Hunt, E. K. and Lautzenheiser (2011), M. History of Economic Thought: A Critical
Perspective. M.E Sharpe, Inc.
Martinez-Hernandez, Francisco (2010), An Alternative Theory of Real Exchange Rate
Determination: Theory and Empirical Evidence for the Mexican Economy, 1970-
2004. Investigacion Economica.
Marx, Karl (1981), Capital: A Critical of Political Economy. London: Penguin Books.
— (1973), Grundrisse: Outlines of the Critique of Political Economy. Penguin, online
version by Andrew Lannan:
http://www.marxists.org/archive/marx/works/1857/grundrisse/index.html
Milberg, W. (1994), Is Absolute Advantage Passe? Toward a Kenynesian /Marxian
Theory of International Trade. Great Britian: In., Glick, Mark.
51
Milberg, W. (2002), Say's Law in the Open Economy: Keynes's Rejection of the Theory
of Comparative Advantage. In Keynes, Uncertainty, amd the Global Economy. Eds:
Dow and Hillard, Massachusetts: Edward Elgar Publishing Limite.
Pasinetti, L. (1971), Lectures on the Theory of Production. New York: Columbia
University Press.
Pesaran Bahram and Hashem Pesaran (2009), Time Series Econometrics using Microfit
5.0: A User's Manual . Oxford University Press.
Ricardo, David (2004), On the Principles of Political Economy and Taxation. Indiana:
Liberty Fund. Editied Vol.I.
Ruiz-Napoles, Pablo (2010), Costos unitario laborales verticalmente integrados por rama
en Mexico y Estados Unidos 1970-2000. Investigacion Economica.
— (1996), Alternative Theories of Real Exchange Rate Determination , A Case Study:
The Mexican Peso and the US Dollar. New York: Unpublished PhD dissertation, New
School for School of Research.
Salgado, Melissa, Gochez, Roberto, Bolanos, Francisco (2010), Los determinates
estructurales de la evolucion de los flujos comerciales entre El Salvador (ES) Y la
Union Europea (UE). Friendrich Ebert Stiftung.
Sarich, John (2006), What Do We Know About The Real Exchange Rate? A Classical
Cost of Production Story. Review of Political Economy.
Shaikh and Antonopoulos (2013), Explaining long term exhange reate behavior in the
United States and Japan. Alternative Theories of Competition: Challenges to the
Orthodoxy. Eds: Moudud, Bina, and Mason, Routledge.
— (1998), Explaining long term exhange reate behavior in the United States and Japan.
The Jerome Levy Economics Institute.
Shaikh, Anwar (2012), Competition Matters: China's Exchange Rate and Balance of
Trade. The Forum, Summer.
— (1999a), Explaining the U.S. trade deficit. Washington D.C.: Testimony before the
Trade Deficit Review Commission.
— (1999b), Real exachange rates and international mobility of capital. The Jerome Levy
Economics Insititute.
— (1991), Competition and Exchange Rates: Theory and emprirical Evidence. New
York: New School for Social Reserach.
— (1980), On the laws of international exchange. Cambridge: Cambridge University
Press.
Smith, Adam (1999), The Wealth of Nations Books I-III. England: Penguin Books.
52
Chapter II
An Alternative Theory of Real Exchange Rate
Determination: Theory and Empirical Evidence for the
Mexican Economy, 1970-2011
II.1 Introduction
The recent history of the Mexican economy has shown that its worst economic crises
have been due to balance of payments problems, which eventually lead to foreign
exchange rate crises (1976-1977, 1982, 1986-1988 and, 1994-1995). Although
conventional exchange rate models hold that in the long-run real exchange rates will
move in such a way as to make countries equally competitive, such an argument is far
from being true because in reality countries are unequally competitive. In the case of
Mexico (Mex), a clear and thorough assessment of real exchange rate determination and
its relationship with the balance of payment, especially with the current account, which
has been negative since the late forties despite currency devaluations, is necessary.
53
A serious problem with conventional economic analyses is its reliance on price
mechanics [Purchasing Power Parity (PPP) and related theories] and comparative
advantage theory, with the aim of expecting that in the long-run exports will equal
imports. In other words, conventional analyses assume that, in the long-run, trade
between countries will be roughly balanced.
On the one hand, international trade theory postulates that if one abstracts from
various sources of financial flows and government intervention in the foreign exchange
market, exchange rates will move toward their equilibria levels when they reach their
market-clearing values. That is, at an equilibrium exchange rate that reflects the relative
price levels of the trading-partners and the domestic economy. This then leads to the
central proposition of foreign trade theory, namely, that under these circumstances,
nominal exchange rates move automatically to make the balance of trade equal to zero.
Along this line of thought, it follows that trade deficits and surpluses are the outcome of
short-run deviations of exchange rates from their equilibrium levels (Antonopoulos,
1997; Ruiz-Nápoles, 1996).
On the other hand, neoclassical trade theory assumes that competitiveness between
countries is determined by the comparative cost principle. Thus, according to this
principle, any country would always find at least one industry in which it is competitive.
Hence, if the exchange rate is adequately managed to achieve and maintain such
competitiveness, foreign trade will tend to be balanced (Ruiz-Nápoles, 2010). In other
words, this standpoint assumes that long-run real exchange rates will eventually do away
54
with competitive differences, without requiring any change in wages, productivity, and
technical change.
It is, nevertheless, important to point out that despite the fact that the two foregoing
principles are too often embraced by academic analyses and economic policy makers;
historical data have provided ample testimony to the persistence of trade imbalances
(even under managed (dirty floats), fixed and flexible exchange rate regimes, across
countries and across time). Importantly, current models of the exchange rate perform
quite poorly at the empirical level (Harvey, 1996; Kruger, 1983; Stein et al., 1995).
Hence, mainstream models may be unreliable guides to economy policy. This paper aims
to put forth an alternative theory of real exchange rate determination of the Mexican peso
with respect to the United States dollar (US dollar). Our model is based upon a classical
approach to the theory of competition developed in Shaikh (1980, 1991, 1999b, and
2013).
According to this theory of competition, which has its origins in the works of Marx
and Keynes (Milberg, 1994), the international competitiveness of a country, or industry,
is primarily based on its absolute advantage in terms of product technology and labor
productivity. This framework argues that it is a country’s competitive position, measured
by the real unit labor cost of its tradable sector, which determines the center of gravity of
the real exchange rate. That is, differences among the real production costs of nations
determine their international terms of trade and hence their long-run real exchange rates.
Our alternative approach also argues that the international money flows occasioned by
balance of trade imbalances do not change price levels as the quantity theory of money
55
claims, but rather change interest rates as Marx, Keynes and Harrod claim. This means
that absolute cost advantages are not eliminated by the money flows, so they continue to
rule. It also means that free trade will give rise to trade imbalances which will be
automatically covered by corresponding capital flows, so that a country with a balance of
trade deficit could end up as an international debtor.
Three key proposals follow from our alternative approach. First, real exchange rates
can be pinned down by the vertically integrated real unit labor cost ratios of the tradable
sectors of the transacting countries. Second, trade surpluses and deficits are not anomalies
of a competitively functioning international world market system, nor need they be
temporary. Third, devaluations will not have a lasting effect on trade balances, unless
accompanied by fundamental changes in national real wages or productivity.
In order to test the main hypotheses of Shaikh’s model for the Mexican economy, the
second section of this paper reviews the principal models of exchange rate determination,
putting special emphasis on their point of agreement (PPP). In the third section we
develop the main points of Shaikh’s works and we incorporate some minor additions to
his formal model of long-run real exchange rate (net capital inflows and government
expenditure). The fourth section presents the methodology used to build the relative unit
labor cost time series (RULC US-Mex). The fifth section presents an econometric model
of the long-run real exchange rate determination. Final remarks are included in the sixth
section.
56
II.2 Conventional Models of Exchange Rate Determination
PPP and Related Theories
As is well known, PPP hypothesis has its foundation in the Law of One Price (hereafter,
LOP), whose main argument claims that if one abstract from tariffs and transportation
costs, unfettered trade in goods should ensure identical prices across countries. Therefore,
if this law holds for every individual good, then it follows immediately that it must hold
for any identical basket of goods. In other words, the LOP is not a theory of the exchange
rate8, but rather a test of market efficiency inasmuch as independently of the local
conditions of production and individual producer’s cost, their selling prices must be
approximately equal (Antonopoulos, 1997; Ruiz-Nápoles, 1996).
PPP is a theory of exchange rate determination as it asserts that nominal exchange
rates in general, move in the appropriate direction so as to equalize the relative price
levels between two countries. Thus, although it is often not explicit which underlying
mechanism would be necessary in order to create a particular common level of prices, for
the adherents of the PPP hypothesis, price level movements are dominated by monetary
factors in the sense that if money supply increases, then also the price level would do it in
the same proportion (Dornbusch, 1988; Froot and Rogoff, 1994; Rogoff et al., 2001).
More specifically, for the trade theory that underpins the PPP hypothesis, the mechanism
through which exports match imports in the long-run is the same mechanism that
guarantees that the price levels will be equalized between two countries that trade with
8 It is worth noting that, since in itself, the LOP does not imply a long-run equilibrium real exchange rate
(at which balance of trade would be equal to zero), it is possible to have the LOP prevail even when there is
a trade surplus or trade deficit (Antonopoulos, 1997).
57
each other. This principle is known as Hume’s price-specie-flow mechanism
(Antonopoulos, 1997; Ruiz-Nápoles, 1996; Shaikh, 1980).
According to this principle, it is the amount of money in circulation which varies
with the trade balance that causes the level of prices to change (Ruiz-Nápoles, 2004).
That is, for the mainstream trade theorists (and Ricardo’s theory of trade) in a two
commodities, two countries model, trade can only take place in terms of money prices.
So, departing from a situation in which one country has absolute advantage in producing
both commodities (due to higher productivity and better technology), it would be paid by
its exports with money. Then the net inflow of money makes its price go up until one of
the two absolute advantages disappears, via the quantity theory of money.
Simultaneously, the net outflow of money in the less efficient country makes its prices go
down until one of them is relatively lower so as to make the good attractive for
importation from abroad. Here it is the money flow which does a sort of transformation
of absolute into relative advantages (Shaikh, 1980).
Whether the model is the Ricardian one or that of Heckscher-Ohlin, and
notwithstanding their differences regarding the source of absolute (dis)advantage, these
models predict that these absolute advantages will turn into comparative ones. Therefore,
both models come to the same conclusion: trade is desired by both nations because it
improves their general economic welfare; money inflows and outflows, eventually
change the price ratios of the two countries and in so doing they bring about balanced
trade.
58
Along these lines and putting aside sterilization policies, the standard theory expects
the long-run real exchange rate to gravitate around that level which balances trade. In this
regard, the formal structure of the PPP hypothesis proposes that the nominal exchange
rate between the currencies of two countries is the price ratio of the two countries (see
equation 1):
𝑒 = 𝑃 𝑃∗⁄ (1)
Where, e is the nominal exchange rate, P is the price level of country A and 𝑃∗ of country
B. This is the absolute (or strong) PPP hypothesis. The relative version of this statement,
known as the relative (or weak) PPP hypothesis, states that the nominal exchange rate,
instead of being equal to, has a constant proportional relationship to the price ratio of the
two countries (see equation 2):
𝑒 = 𝑘(𝑃 𝑃∗⁄ ) (2)
Where, k is a constant parameter that reflects the given obstacles to trade. Nonetheless, to
the extent that there is a change in the price ratio, the nominal exchange rate will change
as well. Thus, we can re-write the real exchange rate as follows (see equation 3 and 4):
𝑒𝑟 = 𝑒(𝑃∗ 𝑃⁄ ) 𝑜𝑟 𝑒𝑟 = 𝑒(𝑃∗ 𝑘𝑃⁄ ) (3 𝑎𝑛𝑑 4)
Equation 4, implies that the real exchange rate, 𝑒𝑟, is invariant through time, since an
opposite and equivalent change in the nominal exchange rate, e, always matches a change
in 𝑃∗ 𝑃⁄ as suggested in equations [1] and [2] (Stein et al., 1995). As a result of this
monetary mechanism, the PPP hypothesis asserts that real exchange rates are expected to
be stationary over the short and the long-run.
59
In effect, both versions of the PPP hypothesis (strong and weak) expect that in the
short-run and the long-run the rate of change of the nominal exchange rate offsets the
relative rate of inflation. Hence, from this perspective, real exchange rates remain
roughly unaltered through time. However, for different countries and different time
spans, empirical data and econometric tests have shown that real exchange rates are
simply not-stationary in either the short-run or the long-run (Antonopoulos, 1997;
Harvey, 1996; Stein et al., 1995; Shaikh et al., 2012).
On the one hand, PPP is not accepted in the short-run, as prices are assumed to be
sticky; hence overshooting is not only possible, but predictable (Antonopoulos, 1997).
Besides, due perhaps to the growth of capital flows via financial markets and speculation,
and the volatility of the nominal exchange rate; these models have tended to accept that
the PPP does not apply in the short-run. Nonetheless, adherents continue to believe that
PPP applies in the long-run as a natural result of floating exchange rates (Stein et al.,
1995).
On the other hand, empirical tests conducted over a 50-year span of the postwar
period, also confirm that under floating exchange rates the PPP hypothesis is rejected
(Froot and Rogoff, 1994). This latter difficulty has forced supporters of the PPP
hypothesis to argue that any convergence which might exist must be extremely slow
(Rogoff et al., 2001), requiring perhaps 75 or even 100 years of data in order to become
evident (Froot and Rogoff, 1994).
60
In this regard, one must keep in mind that despite the notable differences between the
classic exchange rate models9 and the monetarist models, namely asset-market
approaches to the exchange rate with rational expectations and intertemporal
optimization, both groups of models rely heavily on the assumption of stationarity of the
real exchange rate series (weak version of the PPP hypothesis). Therefore, the existence
of a non-stationary series of the real exchange rate would invalidate all of them (Stein et
al., 1995).
For the Mexican case, our own estimations show that the real exchange rate of the
Mexican peso with respect to the US dollar, for annually data for the period 1970-2011,
is also a non-stationary process. Unit root tests for the level of the log of the Mexican real
exchange rate [ADF, PP and, the KPSS test], by and large, do not reject the null
hypothesis of the presence of a unit root. Therefore, for this period, this series follows a
non-stationary process, I(1) (see table 3 and figure 2).
Table 5: Unit Root Test
9 Here we refer to the elasticity approach, absorption approach and, the classical balance of payment
approach.
Variables A B C A B C
LRXRI -2.68 -2.70 -0.48 -2.69 -2.75 -0.35 0.13 0.14
∆ LRXRI -6.44 -6.53 -6.62 -7.16 -7.08 -7.24 0.14 0.15
∆∆ LRXRI -8.89 -9.02 -9.14 -19.23 -19.68 -20.10 0.25 0.25
Notes: ∆ indicates differences, LRXRI stands for the log of the real exchange rate index, 1988=100.
Conclusions: LRXRI ~ I(1)
Model A adds a constant and a trend, model B adds only a constant and modelo C does not include nothing.
The bold squares indicate the rejection of the null hipotesis at 5% significance level.
and represent the KPSS test statistics, where the null hypothesis considers that the series are stationary
in levels or around a deterministic trend, respectively.
ADF PP KPSS
61
In short, table 3 suggests that the PPP hypothesis does not apply to the Mexican case, and
moreover, also suggests that the inflation rates between Mexico and US do not follow a
common path in the short and long-run. Consequently, the PPP hypothesis, which
involves the use of price indexes both in its strong and weak versions, does not
necessarily reflect the degree of competitiveness of the economy, since it emphasizes
more the general price level.
II.3 An Alternative Theory of Real Exchange Rate Determination
Theory of Competition and Real Exchange Rate
The aim of this section is to develop the crucial points of Shaikh’s alternative theory of
real exchange rate determination (1980, 1991, 1999b, and 2013), which is based upon a
non-neoclassical theory of competition and the principle of absolute advantage as the
main determinant of international competitiveness. Thus, the point of departure for
Shaikh’s model is the classical theory of competition, which can be traced back to the
writings of Smith, Ricardo, and Marx. This approach considers competition as rivalry
among firms, in the classical sense where all producers try to obtain a share of their
market by lowering costs.
With regard to the domestic competition, rivalry takes primarily the form of price
competition where each firm attempts to undercut competitors’ prices. This rivalry is
carried out, as a rule of thumb, through the introduction (at intervals) of better techniques
of production with the clear objective of reducing the unit cost of production
62
(investments). Shaikh assumes that any industrial economy’s prices are determined by a
dual, intra-industry and inter-industry, competition process.
On the one hand, competition in a generalized market economy mainly refers to
competition of different capitals. Once production has taken place, producers of
individual goods are disciplined by the market not to charge an arbitrary price, but rather
the selling price determined by the better conditions of production. That is, the price that
prevails in one particular market is not the average price of the industry but the least cost
price determined by the most efficient producer in that industry. This price is called the
regulating price and the producer is the regulating capital, as distinguished from the
average price and the average capital (Ruiz-Nápoles, 1996; Shaikh, 1999b). In turn, “non-
regulating capitals will be forced by competition to sell at the same price, and will
therefore have a variety of profit rates determined by their own various conditions of
production” (Shaikh, 1999b:2).
On the other hand, competition between industries means that, to the degree that one
industry is capable of realizing higher rates of return than the average prevailing rate of
other industries, the more capital it will be able to attract either because other envious
capitals will enter that particular market, or because the profitable enterprise will be able
to expand faster by re-investing and, hence, enlarging their own capital formation. Under
such circumstances, it is the free mobility of a factor(s) of production that produces the
tendency for a rough equalization of profit rates between the previously unequal profit
sectors (Antonopoulos, 1997). Thus, the rates of profit that are equalized by capital flows
are the profit rates of new investments in the regulating conditions of production (Shaikh,
63
1999b). In other words, regulating prices of production in each industry are nothing else
but the embodiments of the regulating techniques of production. As such, they
incorporate the prevailing rate of profit and act as the center of gravity of selling prices.
Hence, it is the best technology generally available for a new investment that forms the
regulating conditions (Shaikh, 1999b).
With regard to international competition, Shaikh’s model assumes that since
production techniques used by firms within one nation differ, one would expect that
techniques of production of any World Industry, where individual firms are spread out
through various countries, will vary from one nation to another as well. Real wages,
generally, will differ among countries as well (especially among developed and
developing countries). So, the resulting lower unit cost of production, which can be
measured as the total (vertically integrated) unit labor cost (ULC), would allow
international regulating firm(s) to either lower their own market price and thus, enlarge
their market share or, perhaps, to temporarily sell at the prevailing market price and
capture a higher profit per unit sold. In either case, the result is the same: the international
regulating firm(s) will be in a position to make more profits relative to other international
firms producing similar goods and thus faster engage in more R&D. The winners are
those international firms capable of maintaining an absolute cost advantage vis-à-vis their
competitor. Conversely, those firms suffering a loss of market share and shrinking profits
would have an absolute disadvantage (Antonopoulos, 1997).
64
Formal Theoretical Model
In his model, Shaikh follows the Ricardo-Marx tradition of relating prices to relative
labor inputs costs (Shaikh, 1980), adopting Pasinetti’s model (1977). Thus, Shaikh starts
from a closed economy, where relative prices of any two commodities i,j are dominated
by the relative prices of the regulating capitals (𝑝𝑖∗, 𝑝𝑗
∗), which themselves are subject to
their own vertically integrated unit labor costs (𝑣𝑖∗, 𝑣𝑗
∗). This interrelationship between
relative prices and relative vertically integrated unit labor costs is expressed in equation
5:
𝑃𝑖
𝑃𝑗≅
𝑃𝑖∗
𝑃𝑗∗ ≅
𝑣𝑖∗
𝑣𝑗∗ (5)
In any industry, the prevailing market price of a commodity (𝑝∗) is regulated by the
regulating firm’s cost, which can be expressed as the vertically integrated nominal wage
(𝑤∗) the regulating firm is subject to, and the vertically integrated labor requirement
(𝜆∗), dictated by the technology this regulating firm uses, in the sense of Sraffa and
Pasinetti.10 If the nominal wage (𝑤∗) is divided by the consumer price index (cpi), and
letting subscript r designate real instead of nominal measures, then the real wage (𝑤𝑟∗) is
equal to 𝑤∗ 𝑐𝑝𝑖⁄ . It follows that the real vertically integrated unit labor cost will be equal
to 𝑤𝑟∗𝜆∗, which in equation 6 will be reflected as 𝑣𝑟
∗.
10 For Sraffa and Passinetti, there is a unique set of rates of exchange among commodities that is
determined by technology alone and that must be adopted in order to keep the system in a self-replacing
state. Sraffa especially points out that these rates of exchange might indifferently be called ‘natural prices’,
or ‘prices of production’, or ‘values’. In a precisely parallel way, the relation in the price system does not
go —as traditional marginal analysis would have it— from final consumers’ preferences to ‘imputed’ costs.
As classical analysis has always claimed, it goes from costs of production to ‘natural’ prices (see Pasinetti,
1992).
65
𝑃𝑖
𝑃𝑗≅ (
𝑤𝑖∗ 𝑐𝑝𝑖𝑖⁄
𝑤𝑗∗ 𝑐𝑝𝑖𝑗⁄
) (𝜆𝑖
∗
𝜆𝑗∗) ≡ (
𝑤𝑟𝑖∗
𝑤𝑟𝑗∗ ) (
𝜆𝑖∗
𝜆𝑗∗) ≡
𝑣𝑟𝑖∗
𝑣𝑟𝑗∗ (6)
Competition drives firms towards introducing more effective technologies. As a sector’s
regulating capital lowers its real relative cost, aggressive competition will drive down the
sector’s relative price as well. Thus, its own purchasing power, vis-à-vis the other goods,
is expected to depreciate when its competitive position improves.
According to Shaikh’s model, when the ULC of one of the two goods declines, the
competitive position of the country producing that good improves, and thus there is a real
depreciation of its currency. Shaikh’s model also assumed a two-country, two-good
model, under complete specialization, so the exports of each country must be equal to the
imports of the other country. In addition, specialization implies that each country contains
exclusively one of the two regulating capitals, as each country is the sole producer of one
of the two goods being traded.
Now, in equation 7, the nominal exchange rate of a country 𝑒𝑎𝑏 is defined as the
number of units of currency a per one unit of currency b. Thus, a rise in the exchange rate
corresponds to a depreciation of currency a, as more units are needed for one unit of the
foreign currency:
𝑒𝑎𝑏 =𝑐𝑢𝑟𝑟𝑒𝑛𝑐𝑦 𝑎
𝑐𝑢𝑟𝑟𝑒𝑛𝑐𝑦 𝑏 (7)
Finally, equation 8 gives a general definition of the terms of trade of country a relative to
country b as follows:
𝑡. 𝑜. 𝑡𝑎𝑏 =𝑃𝑥𝑎
𝑃𝑚𝑎𝑒𝑎𝑏
=𝑃𝑥𝑎
𝑃𝑥𝑏𝑒𝑎𝑏
(8)
66
Combining equations [6] and [8], the terms of trade can be re-written as follows:
𝑡. 𝑜. 𝑡𝑎𝑏 =𝑃𝑥𝑎
𝑃𝑥𝑏𝑒𝑎𝑏
≅𝑣𝑥𝑎
𝑣𝑥𝑏𝑒𝑎𝑏
≡ (𝑤𝑎
𝑤𝑏𝑒𝑎𝑏) (
𝜆𝑥𝑎
𝜆𝑥𝑏𝑒𝑎𝑏
) (9)
At this point, Shaikh makes the simplifying assumption that both countries consume
similar baskets of tradable consumption goods. Then according to the law of one price,
𝑝𝑐𝑇𝑎= 𝑝𝑐𝑇𝑏
𝑒𝑎𝑏, where 𝑝𝑐 is the price of consumption goods, and the subscript (𝑇)
stands for tradable. Thus, equation [9] can be transformed into equation 10 as follows:
𝑃𝑥𝑎
𝑃𝑥𝑏𝑒𝑎𝑏
≅ (𝑤𝑎 𝑃𝑐𝑎
⁄
𝑤𝑏 𝑃𝑐𝑏⁄
) (𝜆𝑥𝑎
𝜆𝑥𝑏
) (𝑃𝑐𝑎
𝑃𝑐𝑇𝑎⁄
𝑃𝑐𝑏𝑃𝑐𝑇𝑏
⁄) (
𝑃𝑐𝑇𝑎
𝑃𝑐𝑇𝑏𝑒𝑎𝑏
) (10)
From equation [6], the first two ratios on the right hand side of equation [10] are
equivalent to the familiar 𝑣𝑟 ratio of two countries (ULC), the third ratio is the
relationship of non-tradables to tradables of consumption goods in each country
respectively, and the fourth ratio is equal to unity. Thus, equation 10 can be re-written as
indicated in equation 11.
𝑃𝑥𝑎
𝑃𝑥𝑏𝑒𝑎𝑏
≅ (𝑣𝑟𝑥𝑎
𝑣𝑟𝑥𝑏
) (𝑃𝑐𝑎
𝑃𝑐𝑇𝑎⁄
𝑃𝑐𝑏𝑃𝑐𝑇𝑏
⁄) (11)
Here, if we take the inverse of equation [11], then we end up with a definition of real
exchange rate (deflated by the export prices), which according to equation 12, tends to
follow the long-term trajectory of the real vertically integrated ULC ratio adjusted by the
non-tradable to tradable goods ratio:
𝑒𝑟𝑎𝑏= 𝑒𝑎𝑏 (
𝑃𝑥𝑏
𝑃𝑥𝑎
) ≅ (𝑣𝑟𝑥𝑏
𝑣𝑟𝑥𝑎
) (𝑃𝑐𝑏
𝑃𝑐𝑇𝑏⁄
𝑃𝑐𝑎𝑃𝑐𝑇𝑎
⁄) (12)
In simply terms, we can re-write equation [12] as indicated in equation [12b]
67
𝑒𝑟𝑎𝑏≅
𝑣𝑟𝑥𝑏
𝑣𝑟𝑥𝑎
∗ 𝑇𝑏 𝑇𝑎⁄ (12𝑏)
Where 𝑇 = 𝑃𝑐 𝑃𝑐𝑇⁄
Non-tradable to tradable adjustment term (𝑇) is defined as the ratio of CPI to PPI, since
the former includes both tradable and non-tradable goods and services, and the latter is
proxy for tradable goods only.
In natural log form, equation [12b] is reformulated as indicated in equation [13].
ln 𝑒𝑟𝑎𝑏≅ ln (
𝑣𝑟𝑥𝑏
𝑣𝑟𝑥𝑎
) + ln (𝑇𝑏
𝑇𝑎) (13)
Finally, Shaikh’s position is that, in any case, neither flexible nor fixed exchange rate
variations will correct the structural trade imbalances induced by international
competition because the determinants of the terms of trade (𝑤𝑟∗ and 𝜆∗) are not free to
move so as to bring about automatic balance of trade adjustment. Hence, the generalized
modernization of technology, to raise productivity and lower unit labor costs, is the only
long-term solution to the problem of competitive disadvantage. More precisely, a cost
reduction can only be produced by the introduction of more efficient technologies, or, in
the short-term, by the reduction of the real wage rate.
II.4 Statistical Analysis: RULCR US-MEX
This section describes the methodology and construction of the variables needed to
estimate Shaikh’s alternative model of real exchange rate determination. Furthermore, we
68
discuss why in addition to the relative unit labor cost ratio US-Mex (which we claim is
the main determinant of the long-run path of the real exchange rate), we added to our
model the following variables: (1) net capital inflows to Mexico and (2) Mexican real
government final consumption expenditure.
The present empirical work relies mostly on data from the manufacturing sector
because there are numerous problems in the availability of data from the manufacturing
export sector, especially for the Mexican economy. In other words, since not all of the
required data were available for the period under study, we instead calculated the direct
unit labor costs in the manufacturing sector as a proxy for the unit labor cost in the
manufacturing export sector.
Although our empirical analysis covers the period 1970-2011, it is worth mentioning
that in order to construct the 1970-2011 series for the Mexican real ULC, we used four
manufacturing surveys with different base years calculated by the Mexican statistical
authority (INEGI). The first one runs from 1970 to 1982 (1970=100), the second one runs
from 1980 to 1993 (1980=100), the third one runs from 1988 to 2004 (1993=100), and
the last one runs from 2003 to 2011 (2008=100). Although these four surveys calculate
the same data of output, wages, and employment from the manufacturing sector, these
four surveys are not homogenous because the last two surveys take into account a large
number of productive sectors. In terms of an index number, we might expect a large jump
from 1987 to 1988, due solely to the change of one survey to another.
In the next section, we estimated some Error Correction Models (ECM) under the
autoregressive-distributed lag (ARDL) modelling framework. In estimating our
69
econometric models, we used the following two functional relationships for the period
1970-2011:
𝑟𝑥𝑟𝑖𝑗 = 𝑓(𝑟𝑢𝑙𝑐𝑟𝑖𝑗∗ ) (14)
𝑟𝑥𝑟𝑖𝑗 = 𝑓(𝑟𝑢𝑙𝑐𝑟𝑖𝑗∗ , 𝑛𝑐𝑖𝑖 , 𝑔𝑖) (15)
Where,
1) The real exchange rate, 𝑟𝑥𝑟, is equal to the nominal exchange rate deflated by the
price ratio of the foreign country (j) to the domestic country (i):
𝑟𝑥𝑟𝑖𝑗 = 𝑒𝑖𝑗 ∗𝑝𝑗
𝑝𝑖 (16)
2) 𝑟𝑢𝑙𝑐𝑟𝑖𝑗∗ stands for a real unit labor cost ratio*, and is equal to the 𝑟𝑢𝑙𝑐𝑗 of the
foreign country (j) divided by the 𝑟𝑢𝑙𝑐𝑖 of the domestic country (i)
𝑟𝑢𝑙𝑐𝑟𝑖𝑗∗ ≡
𝑟𝑢𝑙𝑐𝑗
𝑟𝑢𝑙𝑐𝑖 (17)
3) 𝑟𝑢𝑙𝑐𝑖,𝑗 stands for real total remuneration in manufacturing in local currency
divided by productivity. That is, real unit labor cost is defined as total labor cost
in manufacturing divided by productivity:
𝑟𝑢𝑙𝑐𝑖,𝑗 =𝑡𝑜𝑡𝑎𝑙 𝑟𝑒𝑎𝑙 𝑙𝑎𝑏𝑜𝑟 𝑐𝑜𝑠𝑡
𝑝𝑟𝑜𝑑𝑢𝑐𝑡𝑖𝑣𝑖𝑡𝑦 (18)
4) At the most general level, the two series that need to be developed are the real
unit labor cost (RULC) of the manufacturing sector of Mexico and that of the US.
Thus, if we substitute the equation that defines the real unit labor cost ratio*
70
(𝑟𝑢𝑙𝑐𝑖𝑗∗ ) into equation [16], we get equation [19], which proposes a similar
relationship as equation [12]:
𝑟𝑥𝑟 ≡𝑀𝑒𝑥𝑃𝑒𝑠𝑜 𝑈𝑆𝑑𝑜𝑙𝑙𝑎𝑟⁄
𝑐𝑝𝑖𝑀𝑒𝑥 𝑐𝑝𝑖𝑈𝑆⁄
≈(ℜ 𝑤𝑎𝑔𝑒𝑠𝑢𝑠 + ℜ 𝑠𝑎𝑙𝑎𝑟𝑖𝑒𝑠𝑢𝑠) (ℜ 𝑜𝑢𝑡𝑝𝑢𝑡𝑢𝑠 𝑒𝑚𝑝𝑙𝑜𝑦𝑚𝑒𝑛𝑡𝑢𝑠⁄ )⁄
(ℜ 𝑤𝑎𝑔𝑒𝑠𝑀𝑒𝑥 + ℜ 𝑠𝑎𝑙𝑎𝑟𝑖𝑒𝑠𝑀𝑒𝑥) (ℜ 𝑜𝑢𝑡𝑝𝑢𝑡𝑀𝑒𝑥 𝑒𝑚𝑝𝑙𝑜𝑦𝑚𝑒𝑛𝑡𝑀𝑒𝑥⁄ )⁄ (19)
5) 𝑛𝑐𝑓𝑖 stands for real net capital inflows to Mexico and it considers deposits, loans
and credits to commercial and public banks in Mexico, as well as non-banking
public and private sectors; foreign investment, which includes direct and indirect
investment; securities issued abroad, both public and private; and the net errors
and omissions in the balance of payments.
6) 𝑔𝑖 stands for the real general government final consumption expenditure of the
Mexican economy.
Data Description
The period of the present study spans from 1970-2011. All price deflators are 1988=100
and each money variable is measured in the corresponding local currency. As we
mentioned above, to carry out the estimation of this alternative model, we assumed that
the manufacturing sector represents the majority of tradable goods, which is correct for
all industrialized and some semi-industrialized countries. In the case of Mexico,
manufacturing trade has grown in importance and it currently represents 90 percent of
total exports and 87 percent of total imports (Fujii, 2000; Martínez, 2003; Ruiz-Nápoles,
71
1996). In addition, the United States is Mexico’s major trading partner. Exports to and
imports from the US account for 75 percent of Mexico’s total foreign trade.
Mexico’s Real Unit Labor Cost
The construction of the RULC of the Mexican manufacturing sector was carried out with
the data provided by INEGI and by Mexico’s Central Bank. Gross domestic product in
manufacturing was deflated by using the implicit prices of the manufacturing sector.
Total wages and salaries were deflated by the consumer price index (1988=100). We also
consider the total number of workers in manufacturing. Thus, real unit labor costs are
wages and salaries paid in manufacturing multiplied by the number of workers in
manufacturing and divided by real gross domestic product in manufacturing:
𝑟𝑢𝑙𝑐𝑀𝑒𝑥 = (ℜ 𝑤𝑎𝑔𝑒𝑠𝑀𝑒𝑥 + ℜ 𝑠𝑎𝑙𝑎𝑟𝑖𝑒𝑠𝑀𝑒𝑥)
∗ [(𝑒𝑚𝑝𝑙𝑜𝑦𝑚𝑒𝑛𝑡𝑀𝑒𝑥) (ℜ 𝑜𝑢𝑡𝑝𝑢𝑡𝑀𝑒𝑥)⁄ ] (20)
US Real Unit Labor Cost
The construction of the RULC of the United States manufacturing sector was carried out
with data provided by the Department of Commerce (BEA). Gross domestic product in
manufacturing was deflated by using the implicit prices of durable and non-durable goods
indexes (average). Total wages and salaries were deflated by the consumer price index
(1988=100). We also consider the total number of workers in manufacturing. Thus, for
the US economy, the real unit labor cost is also equal to real wages and salaries paid in
72
manufacturing multiplied by the number of workers in manufacturing and divided by real
gross domestic product in manufacturing:
𝑟𝑢𝑙𝑐𝑈𝑆 = (ℜ 𝑤𝑎𝑔𝑒𝑠𝑈𝑆 + ℜ 𝑠𝑎𝑙𝑎𝑟𝑖𝑒𝑠𝑈𝑆) ∗ [(𝑒𝑚𝑝𝑙𝑜𝑦𝑚𝑒𝑛𝑡𝑈𝑆) (ℜ 𝑜𝑢𝑡𝑝𝑢𝑡𝑈𝑆)⁄ ] (21)
Net Capital Inflows
The construction of the net capital inflows to Mexico was carried out with data provided
by Banco de México. Once constructed, this series was deflated by the US consumer
price index (base 1988=100). We considered this variable inasmuch as the Mexican
government has since the late 1980s implemented several policies to attract foreign
capital in order to stabilize the exchange rate and to finance the current account deficit
(Martínez, 2003). Hence, we assume that an ongoing supply of foreign exchange (dollars)
could, sooner or later, decrease the price of the nominal exchange rate (𝑒), which itself
could end up appreciating the real exchange rate. Conversely, a significant reduction of
foreign exchange could increase the price of the nominal exchange rate (𝑒), which itself
could end up depreciating the real exchange rate. Nonetheless, it is worth mentioning that
the latter situations could occur without necessarily having to have a strong impact on
domestic prices and costs. More precisely, although the nominal exchange rate could
have an impact on domestic prices and costs, our theoretical framework claims that the
main direction of causality in the long-run goes from real unit labor costs to prices, not
the other way around. We will go back to this point when we discuss our econometric
results in the next section. The relationship between the real exchange rate and the real
net capital inflows is shown in figure 3.
73
In figure 3, the real exchange rate index indicates a real depreciation when its value
is above 100 and a real appreciation when its value is below 100. There is a negative co-
variation between the two variables, allowing for the possibility that net capital flows
(supply of dollars) could explain to some degree the deviation of the real exchange rate of
the Mexican Peso from its theoretical proposed primary determinant (center of gravity),
namely, the real unit labor cost ratio. Finally, as a result of this negative correlation, we
would expect a negative sign between these variables in our econometric model.
Figure 3: Real Exchange Rate Index and Real Net Capital Inflows, 1970-2011
Source: own elaboration based on data from Banco de Mexico.
Mexican Government Expenditure
The construction of the general government final consumption expenditure (𝑔) in
constant US dollars was carried out with data provided by the World Bank (WDI). Once
we obtained the Mexican real GDP in US dollars of 1988, we multiply this real GDP by
-10000
-5000
0
5000
10000
15000
20000
25000
30000
0
20
40
60
80
100
120
140
1970 1974 1978 1982 1986 1990 1994 1998 2002 2006 2010
Millio
ns o
f Dao
llars
Ind
ex
Real Exchange Rate Real Net Capital Inflows
74
the general government final consumption expenditure as a percentage of the Mexican
GDP.
We included the real government expenditure in our model of real exchange rate
determination due also to its likely impact on the nominal exchange rate (𝑒). That is, the
government expenditure has been financed, to an important degree, by the oil revenues in
the fiscal accounts, which by themselves have represented around 34-37% of total public
sector revenues, or about 6.8% points of GDP in the last 20 years. (Martínez and Herrera,
2006). Thus, we assume that the amount of foreign exchange earnings from the oil
exports has a double effect upon the level of the nominal exchange rate. On the one hand,
a higher level of oil exports (which is correlated with a higher level of 𝑔) entails an
increase in the supply of foreign exchange that could put a downward pressure on the
nominal exchange rate (𝑒), that is, an appreciation of the real exchange rate, ceteris
paribus. Conversely, a decrease in the level of oil exports (which is correlated with a
lower level of 𝑔) entails a decrease in the supply of foreign exchange that could put an
upward pressure on the nominal exchange rate (𝑒), that is, a depreciation of the real
exchange rate, ceteris paribus. In other words, the government expenditure is an
“instrumental variable” that helped us deal with the problem of contemporaneous bi-
directional causality between the oil exports and the nominal and real exchange rate.
Finally, another channel of transmission from the real government expenditure to the
nominal exchange rate could come from the changes in the nominal interest rate. That is,
if the fiscal revenue is not enough to finance the government budget or if a given level of
government expenditure contributes to generate an unexpected higher level of aggregate
75
demand, then any of these two scenarios could lead to the issue of more Treasury bonds
at a higher interest rate. Thus, if the increase in the domestic interest rate creates an
attractive interest rate differential, then this interest rate differential could trigger an
important capital inflow into the country, which would contribute to create a higher
supply of foreign exchange, which sooner or later could end up putting a downward
pressure on the nominal exchange rate (𝑒), that is, appreciating the real exchange rate,
ceteris paribus.
Figure 4: Real Exchange Rate Index Real Mexican Government Expenditure, 1970-2011
Source: own elaboration based on data from Banco de Mexico and World Bank (WDI).
The relationship between the real exchange rate and the real government expenditure is
shown in figure 4. In figure 4, although the government expenditure follows an upward
trend almost during the whole period, one can see a negative correlation between these
two variables, so we would expect a negative sign between the relationship of these
variables in our econometric model.
0
5000
10000
15000
20000
25000
30000
35000
40000
45000
0
20
40
60
80
100
120
140
1970 1974 1978 1982 1986 1990 1994 1998 2002 2006 2010
Millio
ns o
f Do
llars
Ind
ex
Real Exchange Rate Government Expenditure
76
Center of Gravity: Real Unit Labor Cost Ratio (US-Mex)
As was mentioned above, we used four manufacturing surveys to estimate the real unit
labor cost (RULC) for the Mexican economy. From there, we created three calculations
of the real unit labor cost ratio (US-Mexico) with different base years and we compared
them graphically with the real exchange rate in order to show that the real exchange rate
trend is closely related to the movements of the real unit labor cost ratio (US-Mex) (see
figure 5).
Figure 5: Real Exchange Rate Index and Real Unit Labor Cost Ratio Index (US-Mex)
Source: own elaboration based on data from Banco de Mexico, INEGI, and US Department of Commerce
(BEA).
Calculation for both variables for the whole period 1970-2011 (see figure 6) shows the
same long-term correlation between the real exchange rate and the real unit labor cost
ratio (US-Mex). But in this last plot (figure 6), we can also observe, as we anticipated
0
20
40
60
80
100
120
140
160
1970 1972 1974 1976 1978 1980 1982 1984 1986
1970-1987: Indexes Base Year 1977=100
Real ULCR Real Exchange Rate
0
20
40
60
80
100
120
1988 1990 1992 1994 1996 1998 2000 2002 2004
1988-2004: Indexes Base Year 1988=100
Real ULCR Real Exchange Rate
40
50
60
70
80
90
100
110
120
2003 2004 2005 2006 2007 2008 2009 2010 2011
2003-2011: Indexes Base Year 2003=100
Real ULCR Real Exchange Rate
66
above, a relatively large downward fall of the RULCR series between 1987 and 1988,
due to the use of two similar manufacturing surveys but with different elements between
them. Nonetheless, both series appear to be strongly and positively correlated through the
whole period.
Figure 6: Real Exchange Rate Index and Real Unit Labor Cost Ratio Index (US-Mex)
Source: own elaboration based on data from Banco de Mexico, INEGI, and US Department of Commerce
(BEA).
II. 5 Empirical Evidence of Alternative Real Exchange Rate
Determination Using an ARDL-ECM Model
In this section we test two hypothesis through econometric techniques: 1) our main
hypothesis is that the long-run trend of the real exchange rate of the Mexican peso with
respect to the US dollar is mainly determined by the relative real unit labor costs of the
US and Mexican manufacturing sectors; 2) our second hypothesis is that the real net
capital flows to Mexico and the Mexican government final consumption expenditures,
also contribute to explain, but to a lesser extent, some part of the long-run path of the
0
50
100
150
200
250
1970 1974 1978 1982 1986 1990 1994 1998 2002 2006 2010
1970-2011: Indexes Base Year 1988=100
Real Exchange Rate Real ULCR
67
Mexican real exchange rate. To pursue this analysis, we first rewrite equations 14 and 15,
respectively, in an appropriate form for econometric analysis as follows:
𝑙𝑟𝑥𝑟𝑖𝑡 = 𝛼 + 𝛽1𝑇 + 𝛽2𝑙𝑟𝑢𝑙𝑐𝑟𝑖𝑡 + 𝑢1𝑡 (22)
𝑙𝑟𝑥𝑟𝑖𝑡 = 𝛼 + 𝛽1𝑇 + 𝛽2𝑙𝑟𝑢𝑙𝑐𝑟𝑖𝑡 + 𝛽3𝑛𝑐𝑖𝑀𝑒𝑥 𝑡 + 𝛽4𝑙𝑔𝑀𝑒𝑥 𝑡 + 𝑢2𝑡 (23)
where: 𝑙𝑟𝑥𝑟𝑖𝑡, the log of the index of the real exchange rate of pesos/dollar; 𝛼, a constant
term; 𝑇, a time trend; 𝑙𝑟𝑢𝑙𝑐𝑟𝑖𝑡, the log of the index of real unit labor cost ratio (US-Mex);
𝑛𝑐𝑖𝑀𝑒𝑥 𝑡, net capital inflows to Mexico in constant US dollars11; 𝑙𝑔𝑀𝑒𝑥 𝑡, the log of the
general government final consumption expenditure in constant US dollars; 𝑢1𝑡 and 𝑢2𝑡,
error terms.
The sample used for the econometric analyses considered the period 1970-2011,
where all the series have 1988=100 as the base year. We considered two econometric
specifications: the first one tests for the existence of a long-run relationship between the
Mexican real exchange rate and the real unit labor cost ratio (US-Mex). The second
model intends to measure the long-run impacts of 𝑛𝑐𝑖𝑀𝑒𝑥 𝑡 and 𝑙𝑔𝑀𝑒𝑥 𝑡 upon the Mexican
real exchange rate.
We use an ARDL model using Microfit 5 for two reasons. First, this testing and
estimation strategy can be applied irrespective of whether the regressors are I(0) or I(1),
and can avoid the pre-testing problems associated with the standard cointegration
analysis which requires the classification of the variables into I(1) and I(0) (Pesaran and
11 The coefficients of the log-linear specification, being elasticities, and hence scale-invariant, are much
easier to interpret. However, the series for ncf was not calculated in log form due to the presence of
negative numbers.
68
Pesaran, 2009:308). Second, the sign and significance of the error correction coefficient
(ECC) provides an indication of Granger causality in a non-stationary context. For an
error correction model (ECM) to be stable, the ECC has to satisfy the following stability
criterion: -1 < ECC ≤ 0 (Hill et al., 2011:500).
Table 6: ECM Results for Mexico-US, 1971-2011
Table 6 shows the long-run or cointegrating relationship along with the ECC (speed of
adjustment)12 associated to equation 22. The F test indicates the existence of an
equilibrium long-run relationship between the variables (𝑙𝑟𝑥𝑟𝑖, 𝑙𝑟𝑢𝑙𝑐𝑟𝑖, and a small
trend), as it lies above the lower and upper bounds. The log of the real exchange rate
index (𝑙𝑟𝑥𝑟𝑖) is the dependent variable. The positive sign and statistical significance of
the cointegrating parameters indicate a strong long-run correlation (elasticity) between
the real exchange rate (𝑙𝑟𝑥𝑟𝑖) and the real unit labor cost ratio US-Mex (𝑙𝑟𝑢𝑙𝑐𝑟𝑖), that is,
according to the cointegrating vector a 1% increase in the real unit labor cost ratio (US-
Mex) tends to increase in 0.8% the real exchange rate. Finally, the stability condition of
the ECM is fulfilled as the ECC is lower than 1 and negative, which suggests that the
𝑙𝑟𝑢𝑙𝑐𝑟𝑖 Granger cause the 𝑙𝑟𝑥𝑟𝑖.
12 The larger the error correction coefficient (in absolute value) the faster will be the economy’s return to its
equilibrium, once shocked (Pesaran and Pesaran, 2009:311). The ECC from table 2, estimated at
-0.62365(0.13247) suggests a moderate speed of convergence to equilibrium.
Regressor Coefficient T-ratio [prob] F-Stat ARDL/DW/R-Bar-sq
LRULC 0.82347 82.0295[.000] Lower bound 2.695 ARDL (1, 0)
Trend 0.03316 18.3285[.000] Upper bound 3.837 DW= 1.71
Speed of Adjustment -0.62365 -4.7080[.000] F-Stat 5.5731[.008] Adj-R-sq= 0.33
69
Besides, with equation 22, through two different methods, we determined that
𝑙𝑟𝑢𝑙𝑐𝑟𝑖 acts as the ‘long-run forcing’ variable. The first method is represented by the
ECM in table 6, where the significance of the F-statistic when the 𝑙𝑟𝑥𝑟𝑖 is the dependent
variable indicates that 𝑙𝑟𝑢𝑙𝑐𝑟𝑖 Granger cause 𝑙𝑟𝑥𝑟𝑖. However, when 𝑙𝑟𝑢𝑙𝑐𝑟𝑖 acts as the
dependent variable in an ECM, the F-statistics fall well below the lower bound, which
indicates that 𝑙𝑟𝑥𝑟𝑖 is an endogenous variable. The second method consisted in the
application of two Wu-Hausman exogeneity tests.13 The first exogeneity test indicates
that 𝑙𝑟𝑢𝑙𝑐𝑟𝑖 is an exogenous variable, while the second exogeneity test indicates that
𝑙𝑟𝑥𝑟𝑖 is an endogenous variable. The conclusion of these econometric tests is that for the
period 1971-2011, the long-run path of the real exchange rate was explained by the
changes in the real unit labor cost ratio (US-Mex).
Table 7: ECM Results for Mexico-US, 1976-2011
Table 7 shows the long-run or cointegrating relationship along with the ECC (speed of
adjustment) associated to equation 23. In table 7 we present a different period (1976-
2011) as we could not find a meaningful cointegrating vector for previous periods. Here
again the log of the real exchange rate index (𝑙𝑟𝑥𝑟𝑖) is the dependent variable. However,
13 The exogeneity Wu-Hausman test is a F-statistics under the null hypothesis of exogeneity. In the case of
the present application, the first test accepts the null hypothesis, that is, the F-statistic (0.279) indicates that
the 𝑙𝑟𝑢𝑙𝑐𝑟𝑖 is an exogenous variable. Whereas in the second test the F-statistic (7.65) indicates that the
𝑙𝑟𝑥𝑟𝑖 is an endogenous variable.
Regressor Coefficient T-ratio [prob] F-Stat ARDL/DW/R-Bar-sq
LRULC 0.553 4.8327[.000] Lower bound 3.539 ARDL (3, 0, 0, 3)
NCF -1.53E-05 -2.8121[.009] Upper bound 4.667 DW= 1.96
LG -0.362 -2.1724[.040] F-Stat 5.3827[.002] Adj-R-sq= 0.728
INPT 4.881 3.0468[.005]
Trend 0.036 4.5818[.000]
Speed of Adjustment -0.754 -4.7601[.000]
70
this is a more general model, where the expected signs and the statistical significance of
the parameters in the cointegrating vector indicate a strong long-run correlation
(elasticity) between the log of the real exchange rate (𝑙𝑟𝑥𝑟𝑖), the log of the real unit labor
cost ratio US-Mex (𝑙𝑟𝑢𝑙𝑐𝑟𝑖), the net capital inflows to Mexico (𝑛𝑐𝑖), and the log of the
general government final consumption expenditure (𝑙𝑔). It is worth noting two important
differences of this last model with the former model from table 6, that is, not only the
coefficient of the 𝑙𝑟𝑢𝑙𝑐𝑟𝑖 reduces a little bit (from 0.82 to 0.55) but also the adjusted R-
square increases significantly (from 0.33 to 0.73) when we added the log of the
government expenditure and the net capital inflows to Mexico into the model, which
means that these variables also contribute to explain the long-term behavior of the real
exchange rate.
Table 8: ECM Results for Mexico-US, 1983-2011
Table 8 shows another ARDL-ECM model associated to equation 23 (similar to that
model from table 7) but with a different period (1983-211) inasmuch as we could not find
more meaningful cointegrating vectors for previous or forward periods. Here it is
important to note two things. First, in table 7 and 8 the coefficients are nearly the same.
However, the adjusted R-square in table 8 increases a considerable 9% (from 0.728 to
0.818). Second, the signs of the coefficients are the same in both tables, indicating that
Regressor Coefficient T-ratio [prob] F-Stat ARDL/DW/R-Bar-sq
LRULC 0.617 4.1381[.001] Lower bound 3.539 ARDL (3, 0, 1, 3)
NCF -1.76E-05 -2.9285[.009] Upper bound 4.667 DW= 2.02
LG -0.413 -2.3352[.032] F-Stat 4.8364[.007] Adj-R-sq= 0.818
INPT 4.9696 3.2068[.005]
Trend 0.042 3.5433[.002]
Speed of Adjustment -0.659 -4.6466[.000]
71
the 𝑙𝑟𝑥𝑟𝑖 and the 𝑙𝑟𝑢𝑙𝑐𝑟𝑖 have a positive long-run relationship, and a negative one with
the 𝑙𝑔 and the 𝑛𝑐𝑖. In this model of table 8, the cointegrating vector suggests that a 1%
increase in the 𝑙𝑟𝑢𝑙𝑐𝑟𝑖 increases in 0.6% the 𝑙𝑟𝑥𝑟𝑖; also suggests that a 1% increase in the
𝑙𝑔 decreases in 0.413% the 𝑙𝑟𝑥𝑟𝑖; with respect to the 𝑛𝑐𝑖, the cointegrating vector
suggests that an increase in one thousand millions of dollars in one year tends to decrease
in 1.76% annually the 𝑟𝑥𝑟.14
Finally, for the period 1983-2011, we decided to run two additional ECM models in
order to estimate, through a comparison method among the adjusted R-squares from table
8, 9, and 10, the individual contribution of the 𝑙𝑟𝑢𝑙𝑐𝑟𝑖, the 𝑛𝑐𝑖, and the 𝑙𝑔 to the short
and long-run explanation of the 𝑙𝑟𝑥𝑟𝑖. Table 9 shows an ECM-ARDL model where the
𝑙𝑟𝑥𝑟𝑖 is the dependent variable and the 𝑙𝑟𝑢𝑙𝑐𝑟𝑖 is the independent variables. The adjusted
R-square from table 9 indicates that the 𝑙𝑟𝑢𝑙𝑐𝑟𝑖 explains in 41.8% the changes in the
𝑙𝑟𝑥𝑟𝑖.
Table 9: ECM Results for Mexico-Us, 1983-2011
Table 10 shows also an ECM-ARDL model where the 𝑙𝑟𝑥𝑟𝑖 is the dependent variable
and the 𝑛𝑐𝑖 is the independent variables. The adjusted R-square from table 6 indicates
that the 𝑛𝑐𝑖 explains in 28.9% the changes in the 𝑙𝑟𝑥𝑟𝑖. So, by a difference method
14 The coefficient of the ncf is a semielasticity, so if we multiply this coefficient per 100, then we get the
growth rate of the rxr.
Regressor Coefficient T-ratio [prob] F-Stat ARDL/DW/R-Bar-sq
LRULC 0.83037 48.1687[.000] Lower bound 4.066 ARDL (1, 0)
Trend -0.03217 12.7115[.000] Upper bound 5.119 DW= 1.874
Speed of Adjustment -0.57913 -4.7115[.000] F-Stat 5.63[.005] Adj-R-sq= 0.418
72
between the adjusted R-squares from table 8, 9, and 10, we can estimate that the 𝑙𝑔
explains only 11.1% of the changes in the 𝑙𝑟𝑥𝑟𝑖. Hence, from this analysis we can
conclude that for the period 1983-2011, the 𝑙𝑟𝑢𝑙𝑐𝑟𝑖 has been the most important variable
in the determination of the Mexican real exchange rate as it alone has explained 41.8% of
the changes of the 𝑙𝑟𝑥𝑟𝑖.
Table 10: ECM Results for Mexico-US, 1983-2011
II.6 Concluding Remarks
This paper presents an alternative approach to the determination of the real exchange rate.
Our method relies on the classical approach to the theory of competition developed in
Shaikh (1980, 1991, 1999b, and 2013). From our theoretical framework and empirical
evidence we can conclude two things: first, that neither the absolute nor relative versions
of the PPP will generally hold; second, that devaluations will not have a lasting effect on
the trade balances, unless they are accompanied by fundamental changes in the real
production costs of nations (i.e., in national real wages and productivity).
The empirical results of our alternative model show that, in contrast to PPP
hypothesis, the real unit labor cost of the manufacturing sector between the US and
Mexican economies is a good approximation to estimate the effective real exchange rate.
Our econometric models show evidence that these two variables, as well as the real net
Regressor Coefficient T-ratio [prob] F-Stat ARDL/DW/R-Bar-sq
NCF -1.66E-05 -3.4455[.002] Lower bound 4.066 ARDL (1, 0)
Trend 4.5996 79.5526[.000] Upper bound 5.119 DW= 2.0522
Speed of Adjustment -0.5166 -3.5544[.001] F-Stat 5.2508[.007] Adj-R-sq= 0.28998
73
capital inflows to Mexico and the government final consumption expenditures, are
structurally related. However, for the period 1983-2011, we showed that the real relative
unit labor cost ratio (US-Mex) is the most important variable in explaining the long-run
behavior of the Mexican real exchange rate.
From our ARDL-ECM models and statistical analyses we can also draw the
following conclusions. First, the 𝑙𝑟𝑢𝑙𝑐𝑟𝑖 and the 𝑙𝑟𝑥𝑟𝑖 are positively correlated but the
𝑙𝑟𝑢𝑙𝑐𝑟𝑖 acts as an exogenous variable (‘long-run forcing’ variable), which tends to
determine the long-run trend of the 𝑙𝑟𝑥𝑟𝑖. Therefore, we can suggest that the appreciation
of the real exchange rate that the Mexican economy has systematically observed in the
last twenty years has been, to an important extent, due to the lack of enough capital
(investment) and productivity in its manufacturing sector vis-à-vis the manufacturing
sector of the US. Second, the econometric results also indicate that an increase (decrease)
of the government expenditure (which is highly correlated with the revenue from the oil
exports) and the real net capital inflows to Mexico tend to appreciate (depreciate) the real
exchange rate in the long-run.
The empirical validation of this alternative model explains why, after the multiple
dramatic devaluations of the Mexican currency (1976, 1982, 1986-88, and 1994), this
country, regardless of the exchange rate regime, has not been able to maintain an
adequate level of competitiveness in order to balance its international trade. Our
alternative theory implies that, in general, the competitive position of firms in the
Mexican manufacturing industry has been far away from those internationally regulating
firms in the global manufacturing industry (especially those in China). Therefore, as long
74
as the Mexican economy does not improve its general technical conditions of production,
Mexico will be structurally disadvantaged regarding the real production costs of its
national industries vis-à-vis their competitor and, as a result, Mexico will run permanent
trade deficits.
References
Antonopoulos, R. (1997), “An Alternative Theory of Real Exchange Rate Determination
for the Greek Economy” Unpublished PhD dissertation, New School for Social
Research, New York.
Banco de México, Indicadores Económicos y Financieros.
<http://www.banxico.org.mx/> [31 Jul. 2013].
Dornbusch, R. (1988), “Exchange Rates and Inflation” Cambridge: MIT Press.
Froot, K. and Rogoff, K. (1994), “Perspective on PPP and Long-Run Real Exchange
Rates” NBER, Working Paper No. 4952.
Fujii, G. (2000), “Apertura Externa y Empleo Manufacturero en México” Momento
Económico No. 109.
Harvey J. (1996), “Orthodox approach to exchange rate determination: a survey”
Journal of Post-Keynesian Economics, Vol. 18, No. 4. Summer.
Hill, R. Carter, Griffiths, William E., and Lim, Guay. (2011), Principles of Econometrics.
4th edition, John Wiley & Sons.
Instituto Nacional de Geografía e Informática (INEGI), Banco de Información
Económica. <http://www.inegi.org.mx/>. [5 Jun. 2013].
Krueger, A. (1983), “Exchange Rate Determination” Cambridge Surveys of Economic
Literature.
López J. (1998), “La Macroeconomía de México: El Pasado Reciente y el Futuro
Posible” Miguel Ángel Porrúa y UNAM.
Martínez, F. and Herrera, S. (2006), “Debilidades y Fortalezas de la Industria Petrolera
en México, 1985-2005” Economía Informa, UNAM, Número 340.
Martínez, F. (2003), “Efectos de la Liberalización Financiera sobre el Crecimiento
Económico de México” Unpublised Bachelor Tesis. UNAM, México.
75
Milberg, W. (1994), “Is Absolute Advantage Passe? Toward a Keynesian/Marxian
Theory of International Trade” In, Glick, Mark (ed.). Competition, Technology and
Money, Classical and Post-Keynesian Perspective. Great Britain. Edward Elgar.
Pasinetti, L. (1992), “Sraffa’s Circular Process and the Concepts of Vertical Integration”
in Essays on Pierro Sraffa: critical perspectives on the revival of classical theory.
Edited by Krishna Bharadwaj and Bertram Schefold. Routledga.
Pesaran Bahram and Hashem Pesaran M. (2009), “Time Series Econometrics using
Microfit 5.0: A User’s Manual” Oxford University Press.
Rogoff, K., Froot, K., and Kim M. (2001), “The Law of One Price Over 700 Years” IMF,
Working Paper No. 01/174.
Ros, J. and Lustig, N. (2000), “Trade and Financial Liberalization with Volatile Capital
Inflows: Macroeconomics Consequences and Social Impacts in Mexico during the
1990s” Center for Economic Policy Analysis, Working Paper Series I, No. 18.
Ruiz-Nápoles, P. (2010), “Costos unitarios laborales verticalmente integrados por rama
en México y Estados Unidos, 1970-2000” Investigación Económica. Num. 273.
— (2004), “The Purchasing Power Parity Theory and Ricardo’s Theory of Value”
Contributions to Political economy, Vol. 23.
— (1996). “Alternative Theories of Real Exchange Rate Determination, A Case Study:
The Mexican Peso and the US Dollar” Unpublished PhD dissertation, New School for
Social Research, New York.
Shaikh, Anwar M. and Antonopoulos, Rania. (2013), ‘Explaining Long-Run Exchange
Rate Behavior in the United States and Japan.’ In Alternative Theories of
Competition: Challenges to the Orthodoxy, ed. Jamee Moudud: Routledge, pp. 201–
228.
Shaikh, Anwar. (1999b) “Real Exchange Rates and International Mobility of Capital”
The Jerome Levy Economics Institute, Working Paper No. 265.
— (1999a), “Explaining the U.S. Trade Deficit” Testimony before the Trade Deficit
Review Commission, Washington, D.C.
— (1991), “Competition and Exchange Rates: Theory and empirical Evidence”
Department of Economics, New School for Social Research.
— (1980), “On the Laws of International Exchange” in Growth, Profits, and Property:
Essays in the Revival of Political Economy, E. Nell (ed.), Cambridge University
Press.
Stein, J., A. and Associates (1995), “Fundamental Determinants of the Exchange Rates”
Oxford: Clarendon Press.
U. S. Department of Commerce, Bureau of Economic Analysis (BEA).
<http://www.bea.gov/>. [15 Aug. 2013].
76
Chapter III
Real Exchange Rate, Effective Demand, and Economic
Growth: Theory and Empirical Evidence for Developed
and Developing Countries, 1960-2010
III.1 Introduction
In the current context of slow economic growth experienced by many developed and
developing countries and in light of the well-documented set of economic policies
undertaken by several East Asian economies leading to the so-called Asian ‘miracle’
(e.g., UNCTAD 1998; Amsden 2001; Chang 2006), some economists have proposed the
use of the nominal exchange rate as a ‘policy variable’ with the purpose of maintaining
an undervalued real exchange rate, in order to boost national exports and investment in
the tradable sector, which would lift productivity, employment, and economic growth
(e.g., Rodrik 2008; Razmi et al. 2009; Bhalla 2012).
77
Given that the exchange rate is a macroeconomic-price capable of having
considerable influence on the allocation of resources (mainly financial resources e.g.,
foreign direct investment and industrial employment), it is likely that the lack of
economic growth that many economies have experienced is due, to some extent, to
recurring periods of exchange rate overvaluation, which has limited the process of
industrial upgrade, created permanent trade deficits and maintained low levels of
domestic savings and foreign exchange reserves. This diagnosis could be especially true
for African and Latin American economies, which on average, during the years (1981-
2010), have been performing worse than in previous periods (1960-1980) and other
developing countries (see table 11).
Table 11: Average Economic Growth by Regions
For the purpose of this chapter 3, I adopt the standard measure of exchange rate
undervaluation, in contradistinction to my methodology in the first two chapters. Thus,
the focus of this investigation is to make a critical analysis of the effects and
consequences of maintaining an undervalued currency upon the economic growth of
developed and developing countries for the periods 1960-2010, 1960-1980, and 1981-
2010. We begin by reformulating an undervaluation index proposed by Johnson et al.
Countries Growth Growth S.D. Growth Growth S.D. Growth Growth S.D.
All countries (96) 4 4.74 4.98 5.32 3.45 4.27
Developed (25) 3.87 3.41 5.27 3.53 3.07 3.06
Developing (71) 4.05 5.13 4.87 5.84 3.58 4.62
Africa (35) 3.83 5.62 4.61 6.73 3.44 4.9
Asia (18) 5.49 4.52 5.90 5.34 5.23 3.91
L.A. (20) 3.68 4.4 4.87 4.53 2.92 4.16
The numbers in parenthesis indicate the number of countries. S.D. stands for standard deviation.
1960-2010 1960-1980 1981-2010
78
(2007) and Rodrik (2008). This standard methodology for the construction of the real
exchange rate (RER) and an undervaluation index uses the Purchasing Power Parity
(PPP) conversion factor calculated by the Pen World Tables 7.1, which assumed
2005=100 as the base year or equilibrium reference for prices, thus the deviation of the
PPP conversion factor from the market exchange rate (XRAT) is considered a measure of
currency misalignment (under/over-valuation).15
𝑃𝑃𝑃 𝑐𝑜𝑛𝑣𝑒𝑟𝑠𝑖𝑜𝑛 𝑓𝑎𝑐𝑡𝑜𝑟 = 𝐺𝐷𝑃 𝐷𝑒𝑓𝑙𝑎𝑡𝑜𝑟𝐶𝑜𝑢𝑛𝑟𝑦 𝑥 𝐺𝐷𝑃 𝐷𝑒𝑓𝑙𝑎𝑡𝑜𝑟𝑈𝑆⁄
𝑅𝐸𝑅 = 𝑋𝑅𝐴𝑇 𝑃𝑃𝑃⁄ (1)
It is worth mentioning that this measure of under/over-valuation is based on comparison
of countries’ price levels relative to the U.S. and differs substantially from an estimated
level of the real exchange rate that would achieve balance of payment equilibrium. Also,
this standard measure of under/over-valuation differs substantially from the estimated
measure of under/over-valuation according to the classical perspective presented in
Chapter 1 and 2 of this dissertation. For the calculation of the real exchange rate and the
RULCadj in chapter 1 and 2, we used price indexes and assumed that the system was
close to equilibrium in 2002=100 (i.e. we set the level of rxr equal to the level of rulcadj
in that period). This does not affect the trends of either variable, but it does determine the
level of their deviations and hence the periods in which the deviations are positive or
negative –i.e. the periods in which the rxr is under/over-valued. The most important point
15 The ratio in equation 1, also referred to as the national price level, indicates the number of units of a country’s currency required to buy the same amount of goods and services in the domestic market as a U.S. dollar would buy in the United States, so it makes possible to compare the cost of the bundle of goods that make up gross domestic product (GDP) across countries.
79
is that the two (classical and neoclassical) approaches are based on opposing arguments
and will generally give different measures of over- and under-valuation.16
In this paper, we test econometrically the impact of an undervalued currency “as
defined by the standard approach” upon the economic growth of different set of countries
and periods. Our results confirm what other empirical investigations have previously
concluded (Rodrik 2008; Razmi et al. 2009). Overall, in the medium-to-long-run an
undervalued real exchange rate has a positive effect on economic growth mainly via the
size of the export sector and the maintained period of undervaluation. Contrary to
Rodrik’s results, our results suggest that this positive effect applies not only to
developing countries but also to developed countries.
Nevertheless, when we disaggregate the main components of aggregate demand for
different clusters of developed and developing countries, we found that in general, an
undervalued currency has expansionary and contractionary effects in the short-run,
specifically via the export sector and the level of aggregate consumption, respectively.
Therefore, we believe that the analysis of the effects of an undervalued currency upon
economic growth should also be carried out on a case-by-case basis, in order to try to
evaluate ‘correctly’ the structural parameters of each economy (types of exports, degree
of trade openness, level of foreign indebtedness (public and private), etc.) and ‘all the
channels’ through which a currency depreciation could affect the level of economic
activity.
16 The author of this dissertation thanks Professor Anwar Shaikh for clarifying this point to me.
80
After this introduction, the second section develops a reformulation of Rodrick’s
(2008) exchange rate undervaluation index, whose main sample considers 96 countries,
and different sub-samples, makes distinction between developed and developing
countries, countries from Africa, Asia, and Latin America. This section shows different
Balassa-Samuelson effects among developing countries, something other researchers
have overlooked. The third section estimates the long-run effect of the real exchange rate
(rxr) on the growth rate of the GDP and the growth rate of the GDP per capita. The fourth
section, drawing on the stock flow consistent (SFC) approach, for different clusters of
developed and developing countries, we compare the degree of exchange rate over/under-
valuation with the three main components of aggregate demand and the national wage-
share in order to identify shifts in any of these components of aggregate demand and
possible variations in income distribution associated to changes in the value of the
currency. The fifth section estimates the effect of the rxr on the level of investment with
respect to GDP (I/GDP) and the trade balance of goods (X/M) for our different clusters of
developed and developing countries.
III.2 Undervaluation and GDP Per Capita: the Cross-Country Evidence
by World Regions through Time
Building on the work by Johnson, Ostry, and Subramanian (2007), Rodrik (2008) and
Bhalla (2012), we computed an index of exchange rate overvaluation in three steps. First,
81
we define the real exchange rate (RER) in its natural log form as the log of the ratio of the
data on exchange rates (XRAT) and Purchasing Power Parity conversion factors (PPP)17.
ln 𝑅𝐸𝑅𝑖𝑡 = ln(𝑋𝑅𝐴𝑇𝑖𝑡 𝑃𝑃𝑃𝑖𝑡⁄ ) (1B)
where i is an index for countries and t is an index for (1-year) time periods. XRAT and
PPP are expressed as national units per U.S. dollar. When ln RER is greater than ZERO
it indicates that the value of the currency is lower (more depreciated) than is indicated by
purchasing-power parity (PPP) conversion factor, which is considered the ‘equilibrium’
level of exchange rate, so that the deviation of XRAT from PPP measures the level of
currency misalignment (under/over-valuation).
A second step in the construction of the rxr is to take into consideration the price
difference (due to unequal productivities) between tradable and non-tradable goods
among developed and developing countries. That is, according to the Balassa-Samuelson
effect, higher productivity in the tradable sector of rich countries, pushes up the general
level of prices and the real exchange rates; while low productivity in the tradable sector
of poor countries tends to maintain or lower the general level of prices and more
devaluated/depreciated exchange rates. So to discount this income effect over the real
exchange rate, equation 2 regresses ln RER on the log of per-capita GDP (LRGDP_PC):
ln 𝑅𝐸𝑅𝑖𝑡 = 𝛼 + 𝛽 ln 𝑅𝐺𝐷𝑃_𝑃𝐶𝑖𝑡 + 𝑓𝑡 + 𝑢𝑖𝑡 (2)
Where 𝒇𝒕 is a fixed effect (hereafter, FE) for time period and 𝒖𝒊𝒕 is the error term. Using
equation 2, for a sample of 184 countries for the period 1950-2004 and with data from the
Penn World Tables 6.2, Rodrik (2008) and Razmi et al. (2009) found a �̂� = −𝟎. 𝟐𝟓,
17 Bhalla (2012) used the inverse definition of real exchange rate (ln PPP/XRAT).
82
which means that when income per capita increases by 10%, the real exchange rate for
developed and developing countries appreciates around 2.5%. When we tried to replicate
this exercise and expand the time span by using the “revised” Penn World Tables (PWT)
7.1 for the period 1960-2010 and for a sample of 94 countries (which were selected due
to the availability of the data for XRAT and PPP for this used period), we identified a
dissimilar pattern for the relationship of the unadjusted LRXR (from equation 1B) and
the log of the income per capita of developed and developing countries as is shown in
Figure 7.
Figure 7: Real Exchange Rate and GDP Per Capita (PWT 7.1), 1960-2010
-2
-1
0
1
2
3
5 6 7 8 9 10 11
LRGDP_PC
LRX
R
96 Countries, 1960-2010
-2
-1
0
1
2
3
5 6 7 8 9 10 11
LRGDP_PC
LRX
R
71 Developing Countries, 1960-2012
-0.8
-0.4
0.0
0.4
0.8
1.2
1.6
7 8 9 10 11
LRGDP_PC
LRX
R
25 Developed Countries, 1960-2010
-2.5
-2.0
-1.5
-1.0
-0.5
0.0
0.5
1.0
1.5
5 6 7 8 9 10
LRGDP_PC
LRX
R
35 African Countries, 1960-2010
-0.5
0.0
0.5
1.0
1.5
5 6 7 8 9 10 11
LRGDP_PC
LR
XR
18 Asian Countries, 1960-2010
-1.0
-0.5
0.0
0.5
1.0
1.5
2.0
2.5
3.0
7.0 7.5 8.0 8.5 9.0 9.5
LRGDP_PC
LR
XR
20 L. A. Countries, 1960-2010
Figure 7 clearly shows that developed countries follow completely the relationship first
pointed out by Balassa and Samuelson. Meanwhile the (unadjusted) rxr of developing
countries as a whole seems to follow a lower rate of change (i.e. overvaluation) as
83
income per capita increases. When we split up our sample of countries (as in figure 7)
into two periods (1960-1980 and 1981-2010), the same patterns remain (Please see
Figure 11 at the appendix)18. The upshot is that trying to estimate equation 2 for a
combined sample of developed and developing countries might be misleading and biased.
Therefore, we basically decided to estimate the adjusted (for income) rxr taking up
different blocks of countries (as in Figure 7) using equation 3, which is similar to
equation 2 but which adds a fixed effect for cross-sections, to consider different initial
conditions between countries.
ln 𝑅𝐸𝑅𝑖𝑡 = 𝛼 + 𝛽 ln 𝑅𝐺𝐷𝑃_𝑃𝐶𝑖𝑡 + 𝑓𝑖 + 𝑓𝑡 + 𝑢𝑖𝑡 (3)
Where 𝑓𝑖 is a FE for cross-section and 𝑓𝑡 is a FE for time period and 𝑢𝑖𝑡 is the error term.
Using equation 3, we estimated panel (two-way) FE models for different periods
(1960-1980, 1981-2010, and 1960-2010) and blocks of countries in order to estimate the
parameter beta in equation 3 (the Balassa-Samuelson effect). Our main sample considers
96 countries (developed and developing), and in order to take into account different
stages of development (measured by income per capita) and geographic regions through
time, we split up this main sample into different groups of countries as follows: 25
developed countries, 72 developing countries, 35 countries from Africa, 18 countries
from Asia, and 20 countries from Latin America. The list of countries for each group can
be found at the appendix. Using these more homogeneous groups of countries allows for
18 It is worth mentioning that Polterovich and Popov (2002) also reported this different long-run path of the
real exchange rate between developed and developing countries. According to them, this difference path is
the result that developing countries as a group were not catching up with rich countries in productivity
levels during the 1975-1999 period.
84
more precise estimates and reveals interesting differences in the estimated coefficients
based on different structural characteristics.19
The econometric results of these panel regressions for different time periods and
samples are shown in table 12.20 Our investigation uses annual data, and a priori, we did
not impose any restriction on our econometric panel models (i.e., pooled model, one-way,
or two-way FE), that is, we decided which sort of panel model to estimate based on a
Chow test (Baltagi 2005:13).
Table 12: Balassa-Samuelson Effect
Performing a Chow test to our different block of countries and time-periods allowed us to
consider the heterogeneity between countries and possible changes over time. Thus our
strategy was to estimate two Chow (F) tests for each block of countries. The first F-test
compared a pooled (restricted) model against a cross-section FE (unrestricted) model.
The second F-test compared a pooled (restricted) model against a period FE
(unrestricted) model. The results of these F-tests showed that in all the cases but two, the
19 It is worth mentioning that perhaps this lack of convergence between the prices of developed and
developing countries explains the failure of PPP theory. 20 It is worth mentioning that our FE panel regressions for equation 1 differs in several aspects to those
estimated by Johnson et al. (2007) and Rodrik (2008). While Johnson et al. (2007) estimated panel
regressions for each year; Rodrik (2008) used five-year averages to calculate his FE (one-way for time-
period) panel model, using data from the Penn World Table 6.1.
β Adj-R^2 FE β Adj-R^2 FE β Adj-R^2 FE
All countries (96) -0.09 [-4.7] 0.54 2w -0.13 [-5.2] 0.79 1w, C -0.24 [-8.8] 0.65 2w
Developed (25) -0.32 [-18] 0.67 1w, C -0.32 [-7.8] 0.60 1w, P -0.30 [-4.9] 0.64 1w, C
Developing (71) 0.13 [5.3] 0.56 2w 0.09 [3.8] 0.81 1w, C -0.10 [-3.2] 0.50 2w
Africa (35) 0.16 [7.2] 0.32 1w, C 0.37 [6.2] 0.83 2w 0.09 [3.2] 0.47 2w
Asia (18) -0.13 [-7] 0.46 2w -0.15 [-5] 0.64 1w, C -0.20 [-5] 0.79 2w
L.A. (20) -0.31 [-3.6] 0.10 1w, P -0.20 [-3.7] 0.93 1w, C -0.39 [-5.2] 0.53 1w, C
Note: The numbers in brackets stand for the t-statistic (based on White cross-section or White period standard errors & covariance).
The numbers in parenthesis indicate the number of countries in each panel regression.
1960-2010 1960-1980 1981-2010
85
cross-section FE models (1w, C) captured most of the variability of the dependent
variable (ln RER) due to the fact that these F-tests were highly statistically significant
with high r-squares. In other instances, the second F-test showed that the period FE
models (1w, P) captured some part of the variability of the dependent variable (ln RER),
due to the fact that some F-tests were statistically significant with high r-squares. In one-
third of the cases the combination of cross-section and period FE models resulted in
highly statistically significant parameters and high r-squares, which paves the way for
well-estimated two-way (2w) FE models.
The results on Table 11 show that there has been a differentiated B-S effect between
countries through time, since the estimated �̂� parameter (the B-S effect) from equation 3
is different for each block of countries. However, the estimated �̂�s for each block are
relatively similar for the three estimated periods. The only exceptions were the
estimations for the main sample of 96 countries and that for developing countries, whose
averages for the first two periods (1960-1980 and 1981-2010) do not match the estimated
parameter for the total sample (1960-2010).
The results on Table 11 also show that the developed countries have maintained a
relatively high but stable B-S effect for the whole period (�̂� = -0.32). We would highlight
that Balassa (1964) found the same value for developed countries. The second block of
countries that reported a relatively high B-S effect were the Latin American countries,
which for the whole period showed an almost similar B-S effect to the developed
countries (�̂� = -0.31). However, Table 11 shows that Latin American economies’
exchange rate appreciation on average has been growing over time, since in the first
86
period they had a much lower B-S effect (�̂� = -0.20) than in the second estimated period
(�̂� = -0.39).
The Asian countries maintained a relatively low B-S effect for the whole period (�̂� =
-0.13), although from the first to the second period the B-S effect increased in 0.5 per
cent points (from �̂� = -0.15 to �̂� = -0.20). Finally, for the three estimated periods, the
sample of African countries reported an unusual positive B-S effect (�̂� = 0.16), which
suggests that the level of income per capita tends to increase with the level of
undervaluation, or vice versa, depending on the true process of causality between these
variables.
Finally, to arrive at our index of undervaluation for each block of countries and
period, we take the difference between the actual real exchange rate (from equation 1)
and the Balassa-Samuelson-adjusted rate (from equation 3) as follows:
ln 𝑈𝑁𝐷𝐸𝑅𝑉𝐴𝐿𝑖𝑡 = ln 𝑅𝐸𝑅𝑖𝑡 − ln 𝑅𝐸�̂�𝑖𝑡 (4)
Where 𝐥𝐧 𝑹𝑬�̂�𝒊𝒕 is the predicted values from equation 3.
Defined in this way, ln UNDERVAL is comparable across countries ‘within’ each
block of countries and over time. Whenever ln UNDERVAL exceeds ZERO, it indicates
that the currency in dollar terms is undervalued. When ln UNDERVAL is below ZERO,
the currency is overvalued. Thus, this measure of undervaluation is centered at 0.
87
III.3 Undervaluation and Economic Growth
Having estimated our index of undervaluation based on conventional methodology, the
aim of this section is threefold: (i) to analyze the statistical relationship between the rxr
and economic growth for the blocks of countries that we defined in the previous section,
(ii) to draw some conclusion about the rate of real exchange rate that helps in the
medium-to-long-run to stimulate economic growth, and (iii) to estimate through
econometric panel models the effects of an undervalued currency on economic growth.
Statistical Analysis by Regions
Figures 8 and 9 show six scatterplots that relate our index of under/over-valuation and
economic growth for each block of countries. Each scatterplot has the measure of
over/under-valuation on the y axis (centered at zero) and economic growth on the x axis
(almost centered at cero). At the upper and bottom corners of each scatterplot lies a chart
that indicates the number of observations within each quadrant. We also added the default
trend line estimated by Eviews 9.
88
Figure 8: Currency Under-Over/Valuation and Economic Growth
-1.5
-1.0
-0.5
0.0
0.5
1.0
1.5
-30 -20 -10 0 10 20 30 40
GROWTH
Overv
alu
ation / U
nderv
alu
ation
2, 501
1, 781
357
257
(I) 96 Countries, 1960-2010
-.4
-.2
.0
.2
.4
-5 0 5 10 15 20
GROWTHO
verv
alu
ation / U
nderv
alu
ation
29 525
80 545
(II) 25 Developed Countries, 1960-2010
-1.6
-1.2
-0.8
-0.4
0.0
0.4
0.8
1.2
1.6
-30 -20 -10 0 10 20 30 40
GROWTH
Overv
alu
atio
n / U
nd
erv
alu
ation
(III) 71 Developing Countries, 1960-2010
221 1,491
284 1,261
Scatterplot (I), which encompasses the full sample of countries (96), shows that these
economies have experienced more periods of economic growth with an undervalued
currency (Quadrant I). Quadrant number 2, however, indicates us that undervaluation
could also have contractionary effects for some economies. Quadrant 4 shows that some
countries can also grow with an overvalued currency.
Scatterplot (II), which contains the sample of developed countries (26), shows that
developed economies have experienced almost the same periods of economic growth
with an undervalued and overvalued currency. It also shows that developed countries had
had the lowest relative number of contractionary-devaluating periods (29/554) for the
whole sample of 96 countries. Scatterplot (III), which plots the sample of developing
countries (71), resembles almost entirely the same conclusion from Scatterplot (I). The
numbers (in the charts) for the developing countries indicate that having an undervalued
currency has the highest probability of ending up in a context of positive economic
growth (0.457%-0.06%=0.39%):
(𝐼)1,491
3,257= 0.457 + (𝐼𝐼)
221
3,257= 0.06 + (𝐼𝐼𝐼)
284
3,257= 0.08 + (𝐼𝑉)
1,261
3,257= 0.38 = 1
89
Figure 9: Currency Under-Over/Valuation and Economic Growth
-1.5
-1.0
-0.5
0.0
0.5
1.0
1.5
-20 -10 0 10 20 30 40
GROWTH
Overv
alu
ation / U
nd
erv
alu
ation
(IV) 35 African Countries, 1960-2010
101
172 601
688
-.75
-.50
-.25
.00
.25
.50
.75
-10 0 10 20
GROWTHO
ve
rvalu
ation / U
nd
erv
alu
ation
(V) 18 Asian Countries, 1960-2010
30 415
49 374
-1.0
-0.5
0.0
0.5
1.0
1.5
-10 0 10 20
GROWTH
Overv
alu
ation / U
nderv
alu
ation
(VI) 20 Latin American Countries, 1960-2010
55
92
386
413
Scatterplot (IV), which depicts the sample of African countries (35), indicates that these
countries have experienced more periods of economic growth with an undervalued
currency (Quadrant I) than with an overvalued currency (Quadrant IV). Scatterplot (V),
which features the sample of Asian countries (18), shows that these countries had had the
second relative lowest number of contractionary-devaluating periods (30/445) and that
high levels of economic growth are associated positively with an undervalued currency
and lower levels of overvaluation.
Scatterplot (VI), which shows the sample of Latin American countries (20), indicates
that economic growth does not seem to have a particular relationship with an under/over-
valuated currency. In addition, the numbers in the charts indicate that the countries of the
Latin American economies’ on average, had had the relative largest periods of
overvaluation (505/946) in comparison with other economic areas.
Based on the visual inspection of scatterplots (I) and (IV), but especially on (III), we
infer that, by and large, the rate of real exchange rate that has been able to boost
economic growth in the medium-to-long-run within our group of 96 countries, and in
90
particular within our group of 71 developing countries, is the rate around the
‘equilibrium’ real exchange rate (around zero with an undervaluation rate of around 25%
and overvaluation rate of around 15%). Furthermore, the scatterplots (I, IV, and III) also
show that higher levels of under-and-overvaluation (above 40% and 30% respectively)
are associated with lower levels of economic growth. Hence, it seems that the best policy
action that national authorities could undertake to stimulate economic growth in the
medium-to-long run is to try to avoid large real exchange rate misalignments (under-and-
over-valuation) (Johnson et al. 2007; Berg and Miao 2009) and to try to avoid exchange
rate volatility, which tends to discourage trade and investment (Eichengreen 2008:3).
In short, the conclusion drawn from Scatterplots (I, III, and IV) suggests that a
growth rate of the real exchange rate ‘around’ zero in the long run is the most beneficial
rate to reach higher economic growth, especially for the case of developing countries.
However, the exception is the case of our sample of 18 Asian countries, which in general
for the period 1960-2010 seems to maintain a positive association between an
undervalued currency and economic growth.
Econometric Evidence
Based on our index of undervaluation, we estimated through different econometric
specifications and panel data models, the effect of an undervalued currency on economic
growth for our classification of countries by economic regions defined above. It is
important to beer in mind that for each block of countries we estimated a particular
undervaluation index, due to the fact that we identify different Balassa-Samuelson effects
91
between developed and developing countries and among developing countries. Therefore,
we start by estimating a panel model with 96 countries and then we estimated smaller
panel models consisting of more-or-less homogeneous groups of countries. Before we
explain our econometric results in depth, we briefly review the results obtained by other
empirical analyses that have tested the relation between economic growth and
undervaluation.
In his empirical investigation, Rodrik (2008) used a real exchange rate adjusted for
the Balassa-Samuelson effect (similar to equation 2 above) for a sample of 184 countries
and eleven 5-year time periods from 1950-1954 through 2000-04. He found (using
equation 6 below which is a 2w FE model) that an undervalued currency has a positive
effect on economic growth (growth). However, this positive effect operates only for
developing countries (with a 𝛿∗ = 0.027), as the estimates for developed countries were
not statistically significant different from zero. According to Rodrik, the effect of
undervaluation on growth operates through its positive impact on the relative size of
industry measures as the ratio of industry to GDP.
In another study with the same methodology, number of countries and periodicity of
data, Razmi et al. (2009) found also that an undervalued currency has a positive effect on
economic growth for a sample of 184 countries. However, when this sample of 184 is
divided and classified by developed and developing countries, the authors found
econometric evidence indicating that although the results are more robust for developing
countries (in the range between 𝛿∗ = 0.017 and 𝛿∗ = 0.026), there is evidence that
undervaluation affects growth positively in developed countries as well (in the range
92
between 𝛿∗ = 0.014 and 𝛿∗ = 0.017). According to the authors, the effect of
undervaluation on growth operates through its positive impact on investment, which only
operates in developing countries according to different specifications and econometric
methods.
Using a series of panel data regressions with six different samples and three different
periods, we use the following two econometric specifications in order to test the effects
of undervaluation on economic growth:
𝑔𝑟𝑜𝑤𝑡ℎ𝑖𝑡 = 𝛼 + 𝛽 ln 𝑅𝐺𝐷𝑃𝐶𝐻𝑖𝑡−1 + 𝛿 ln 𝑈𝑁𝐷𝐸𝑅𝑉𝐴𝐿𝑖𝑡 + 𝑓𝑖 + 𝑓𝑡 + 𝑢𝑖𝑡 (5)
𝑔𝑟𝑜𝑤𝑡ℎ𝑖𝑡∗ = 𝛼∗ + 𝛽∗ ln 𝑅𝐺𝐷𝑃𝐶𝐻𝑖𝑡−1 + 𝛿∗ ln 𝑈𝑁𝐷𝐸𝑅𝑉𝐴𝐿𝑖𝑡 + 𝑓𝑖 + 𝑓𝑡 + 𝑢𝑖𝑡
∗ (6)
Equations 5 and 6 depict two two-way FE models where the dependent variable in
equation 5 is the annual growth rate of real GDP (growth), the dependent variable in
equation 6 is the annual growth rate of real GDP per capita (growth*), 𝑈𝑁𝐷𝐸𝑅𝑉𝐴𝐿 is
our measure of undervaluation, 𝑅𝐺𝐷𝑃𝐶𝐻𝑖𝑡−1 is the initial income per capita level which
captures the converge term, 𝑓𝑡 is a time specific effect, 𝑓𝑖 is a country specific effect, 𝑢𝑖𝑡
and 𝑢𝑖𝑡∗ are error terms. The estimated values using Eviews 9 for the parameters 𝛿 and 𝛿∗
from equation 5 and 6, respectively, are shown in Table 13.
93
Table 13: Undervaluation and Economic Growth, Fixed Effect Models
The estimated results from equation 5 and 6 show that for our six different samples of
countries for the whole estimated period 1960-2010, an undervalued currency has been
associated positively with growth and growth*. An outstanding result for these
estimations suggests that developed countries have had a larger positive impact of an
undervalued currency on growth and growth* (𝛿 = 3.82 and 𝛿∗ = 2.45) than the
developing countries (𝛿 = 2.12 and 𝛿∗ = 1.84). For the case of developing countries, the
estimated elasticities (𝛿 and 𝛿∗) suggest that, ceteris paribus, an annual increase of 20%
in undervaluation would boost growth and growth* by 0.42% and 0.36% annually,
respectively (2.12*0.20=0.424% and 1.84*0.20=0.368%), a non-negligible amount by
any measure.
When we split up each of our six samples into two sub-periods (1960-1980 and
1981-2010), the results on table 13 show that for all the sample of countries the effect of
undervaluation on growth and growth*, although positive, had a little larger effect in the
first period than in the second one. The only exception to this positive pattern was the
lag(s) lag(s) lag(s) lag(s) lag(s) lag(s)
All countries (96) 2.18 1 1.77 1 1.90 0 2.67 0 1.91 1 1.70 1
[7.28] [6.37] [2.74] [3.82] [5.50] [4.85]
Developed (25) 3.82 0 2.45 0 4.75 0 4.22 0 4.10 0 2.94 0
[5.18] [2.75] [4.02] [3.31] [3.02] [3.82]
Developing (71) 2.12 1 1.84 1 2.41 2 2.64 2 1.57 1 1.5 1
[6.48] [5.67] [2.42] [2.8] [4.1] [3.65]
Africa (35) 1.52 0 2.36 0 -2.54 0 -2.76 0 2.46 0 2.09 0
[2.68] [4.42] [-1.71] [-1.88] [4.03] [3.18]
Asia (18) 3.2 0 2.25 0 2.7 1 2.59 1 2.11 3 1.62 3
[4.66] [4.13] [2.12] [1.89] [1.85] [1.83]
L.A. (20) 1.91 1 2.35 1 3.7 1 3.43 1 2.64 1 2.65 1
[3.55] [4.11] [3.23] [2.72] [4.83] [3.7]
Note: The numbers in brackets stand for the t-statistic (based on White cross-section or White period standard errors & covariance).
The numbers in parenthesis indicate the number of countries in each panel regression.
1960-2010 1960-1980 1981-2010
�̂�∗ �̂�∗ �̂�∗
94
case of the African countries, which according to its estimated elasticities, an
undervalued currency had a considerably large negative impact on growth and growth*
during the first period.
As is very well-known, it is possible that the estimated parameters using panel FE
models from equation 5 and 6 suffer a problem of endogeneity bias as there could be a
problem of contemporaneous simultaneity effect between the dependent variables
(growth and growth*) and the independent variable (ln 𝑈𝑁𝐷𝐸𝑅𝑉𝐴𝐿), which would lead
to the underestimation of 𝛿 and 𝛿∗. However, a dynamic specification with a lagged
dependent variable is likely to improve our estimates considerably in the presence of
persistence effects and omitted supply-side factors (such as institutional variables). Using
the General Method of Moments (GMM) corrects for both of these problems, because it
allows the inclusion of a lagged dependent variable (in levels and in different forms) and
also controls for endogeneity through the use of instrumental variables (Arellano and
Bond 1991).
𝑔𝑟𝑜𝑤𝑡ℎ𝑖𝑡 = 𝛽∗∗𝑔𝑟𝑜𝑤𝑡ℎ𝑖𝑡−1 + 𝛿∗∗ ln 𝑈𝑁𝐷𝐸𝑅𝑉𝐴𝐿𝑖𝑡 + 𝑓𝑡 + 𝑢𝑖𝑡∗∗ (7)
Equations 7 depicts a GMM specified equation where the dependent variable is the
annual growth rate of real GDP (growth), 𝑔𝑟𝑜𝑤𝑡ℎ𝑖𝑡−1 is a lagged variable of the
dependent variable, 𝑈𝑁𝐷𝐸𝑅𝑉𝐴𝐿 is our measure of undervaluation, 𝑓𝑡 is a time specific
effect, and 𝑢𝑖𝑡∗∗ is an error term. The estimated results for equation 7 using Eviews 9 are
shown in Table 14.
95
Table 14: Undervaluation and Economic Growth, Dynamic Panel Data
The estimated results using the GMM method on table 4 suggest similar conclusions
drawn from the results on table 13 using panel FE models. That is, for our six samples of
countries for the period 1960-2010, there has been a positive long-run effect of an
undervalued currency on growth. Moreover, when we split up each of our six samples
into two sub-periods (1960-1980 and 1981-2010), the GMM results show almost similar
results to those estimated with FE panel models. There are, however, two important
differences to be noted. First, for the first estimated period (1960-1980), the GMM results
suggest that an undervalued currency had a considerably larger negative impact on
growth for the African countries in our sample than the results estimated with the FE
model. Second, the GMM results indicate that the Asian countries were able to achieve
faster economic growth (growth) through an undervalued currency as the positive impact
(elasticity) almost doubled from the first (𝛿∗∗ = 1.92) to the second period (𝛿∗∗ = 3.64).
In short, Tables 13 and 14 confirm what other empirical investigations have found in
regard to the relationship between undervaluation and economic growth, that is, these
results mainly suggest that in the long-run the real exchange rate has not been neutral in
stimulating the economic activity. The purpose of the following sections is to identify the
lag(s) PD lag(s) PD lag(s) PD
All countries (96) 2.33 [4.44] 1 Yes 2.90 [3.07] 0 Yes 2.07 [2.27] 1 No
Developed (25) 3.23 [2.86] 0 Yes 3.10 [2.20] 0 Yes 3.55 [2.40] 0 Yes
Developing (71) 2.77 [4.34] 1 Yes 3.34 [47] 1 No 1.87 [2.06] 1 Yes
Africa (35) 1.44 [2] 0 No -3.27 [-475] 0 No 2.16 [2.32] 1 Yes
Asia (18) 2.75 [3.28] 1 No 1.92 [5.82] 2 No 3.64 [2.68] 3 No
L.A. (20) 2.8 [2.5] 1 No 3.51 [2.24] 1 No 3.5 [4.23] 1 Yes
Note: The numbers in brackets stand for the t-statistic (based on White cross-section or White period standard errors & covariance).
The numbers in parenthesis indicate the number of countries in each panel regression. PD stands for period dummies.
1960-2010 1960-1980 1981-2010
96
channels through which the positive stimulus of undervaluation to economic growth
operates and to investigate if this positive stimulus also holds in the short-run.
III.4 Undervaluation, Effective Demand, and Economic Growth in the
Short Run
According to standard open economy models, a real depreciation of the exchange rate has
an expansionary short-run effect via aggregate demand, provided that price elasticities
satisfy the Marshall-Lerner (M-L) condition (so that the trade balance improves) and
there are unemployed resources in the devaluing country (so that output can increase). A
real depreciation can also yield benefits on the supply side by increasing the relative price
of tradable goods, which creates incentives to shift domestic production towards tradables
and demand towards non-tradables, thereby freeing up a greater surplus for exports
(Blecker and Razmi 2008:87).
However, in two cases, a real depreciation of the exchange rate could have a
contractionary short-run effect even if the M-L condition is fulfilled: (i) if the devaluation
leads to higher domestic prices due to higher costs (mainly imports), which may create a
shift in the distribution of income in favor of capital and against labor, by enabling firms
to increase price-cost margins (i.e., lower real wages). If capital owners have a higher
propensity to save than workers, then overall aggregate demand may fall in spite of
increased exports; and (ii) if the devaluation increases the indebtedness ratio of firms and
governments indebted in foreign currency (Diaz-Alejandro 1963; Blecker and Razmi
2008). Furthermore, if the response of exports (and import substitution) to the change in
97
relatively prices is slow, then the currency depreciation may result in the short-run in a
worsening of the trade balance, the terms of trade and of profits (López and Perrotini
2006).
Most of the studies that have estimated empirically the effects of the changes of the
real exchange rate on economic growth, focus mainly on the effects of exchange rate
devaluations on the external sector, specifically on the performance of the national export
sector; that is, for these analyses, an undervalued currency leads to a favorable change(s)
in the relative prices of commercial goods, which tend to raise exports and investment in
the tradable sector, employment and economic growth (e.g., Rodrik 2008; Razmi et al.
2009; Bhalla 2012). However, we believe that the foregoing chain of causation, although
probably correct, does not describe the full picture associated with exchange rate
devaluations. That is, in developing countries, and even in some developed countries,
changes in the rxr also see important changes in the other components of aggregate
demand (aggregate consumption, investment, government expenditure), and even
changes in the stance of monetary policy, and thus on economic growth in the short-run.
In this section, drawing mainly on work by Taylor (2004) and Shaikh (2012), we
make use of the GDP national accounting identity in terms of the three main sectoral
balances (private sector, government, and the rest of the world) under the framework of
the Stock-Flow Consistent (SFC) model in order to identify shifts in any of these three
components of aggregate demand due to periods of exchange rate undervaluation.
Furthermore, we also contrast the degree of exchange rate over/under-valuation with the
level of the national wage-share (Wages+Salaries/GDP) in order to detect possible
98
variations in income distribution due to changes in the value of the currency. We carried
out these analyses for different clusters of developed and developing countries. Before
showing and describing our empirical findings, we briefly refer to the underlying points
of the SFC model.
For the SFC theorists, in any national economy, the level of economic activity is
influenced not just by a country’s income distribution but also by the outcome of the
balance between demand “injections” – private investment in fixed capital and
inventories, public spending, and exports – and “leakages” – private saving, taxes, and
imports (Taylor 2004:13). That is, the level of economic activity ‘over any time period’ is
determined by the difference between domestically available aggregate demand (D) and
domestically available aggregate supply (Q):
𝐸 ≡ 𝐷 − 𝑄 = (𝐶 + 𝐼 + 𝐺 + 𝑋) − (𝑌 + 𝑀) (8)
Where E stands for excess demand, C consumption, I investment in desired stocks of
fixed capital and inventories, G government spending, X export demand, Y domestic
supply, and M imports. Now, let T=total private sector (households and business) taxes,
so equation 8 can be written in terms of three sectoral contributions to excess demand:
the private sector deficit, which is the excess of its expenditures over its disposable
income [(𝐶 + 𝐼) − (𝑌 − 𝑇)]; the government deficit [𝐺 − 𝑇]; and the foreign trade
surplus [𝑋 − 𝑀] (Shaikh 2012:126):
𝐸 ≡ 𝐷 − 𝑄 = [(𝐶 + 𝐼) − (𝑌 − 𝑇)] + [𝐺 − 𝑇] + [𝑋 − 𝑀] (9)
= [𝐼 − 𝑠𝑌] + [𝐺 − 𝑡𝑌] + [𝑋 − 𝑚𝑌]
99
Where s, t, and m stand respectively for saving rate, the import propensity, and the tax
rate. Moudud (2010:92) and Shaikh (2012:127) emphasize that there is nothing in this ex
ante relation which requires that the three balances add up to zero. It is, however, the
incorporation of the undesired inventory change into investment over some putative
‘short run’ period which converts the ex-ante non-zero balance into an ex-post zero-
balance identity which reflects the real financial balances of the economy:
�̇� + �̇� + �̇� = [𝐼 − 𝑠𝑌] + [𝐺 − 𝑡𝑌] + [𝑋 − 𝑚𝑌] = 0 (10)
Where �̇� (= 𝑑𝐷 𝑑𝑡⁄ ), �̇�, and �̇� stand respectively for the net change per unit time in
financial claims against the private sector, in government debt, and in foreign assets.
Equation (10) shows how claims against an institutional entity must be growing when its
demand contribution to Y exceeds Y itself. So when 𝑋 < 𝑚𝑌, net foreign assets of the
home economy are declining, while 𝐺 > 𝑡𝑌 means that its government is running up debt
(Taylor 2004:14). Due to the nature of this budget constrained equation 10, any excess
demand by one (or two) institutional entity(ies) must be exactly offset by the other (two)
institutional entity(ies). In words of Godley and Lavoie (2007:38), “everything comes
from somewhere and everything goes somewhere… … Within this framework, ‘there
are no black holes’ (Godley 1996:6)”. Therefore, it is true from (10) that �̇� + �̇� + �̇� = 0.
Explicitly we are not presenting (modeling) any behavioral equation as the purpose
of this investigation is only to analyze the changes (if any) of the main components of
aggregate demand due to changes of the real exchange rate. Along these lines, figures 12,
13, 14, and 15 in the appendix show the three main macroeconomic sectoral balances
with respect to GDP (for the period 1990-2012) along with the degree (percentage) of
100
undervaluation (in a shaded area only for the period 1990-2010). In a second graph these
figures also show the economic growth rate and the degree (percentage) of over/under-
valuation. In a third graph the figures also show the degree (percentage) of over/under-
valuation along with the wage-share (as percentage of the GDP) for 6 developed
countries, 6 African countries, 9 Asian countries, and 9 Latin America countries. The
source of data is mentioned in the appendix.
In general terms, for our sample of countries (developed and developing), we can
summarize our main findings as follows:
With the exception of countries like Greece, Tanzania (in the 2000s), Senegal,
and Guatemala, it seems that the M-L conditions are fulfilled by our sample of
countries in the short-run.
With the exception of countries like Senegal, Japan, South Korea, Malaysia,
Singapore, Thailand, and Chile, it seems that an undervalued currency affects in a
negative way the income distribution of developed and developing countries.
Undervaluation has mostly helped to propel the economic growth of East Asian
countries, and in a few cases, other countries (Tanzania, Argentina, Mexico, and
Panama).
The foregoing results were drawn based on the visual inspections of the Figures 12-15
(from the appendix) and the average short-run multipliers in Table 15. For each country,
these multipliers were measured as the average (for the period 1990-2010) of the ratios of
the annual sectoral balance with respect to GDP divide by the corresponding degree
(percentage) of currency undervaluation21. Hence, these multipliers measure the change
in the financial position of the three main sectoral balances as a result of a change in the
degree of undervaluation. Due to the nature of the budget-constrained equation 10, in
21 It is worth mentioning that we did not take into consideration years of exchange rate over-valuation. The
estimation of these multipliers corresponds only to years with real exchange rate undervaluation.
101
Table 15 for each row the sum of the first three columns (the last one with an opposite
sign) adds up to zero. In Table 15 columns four and five measure the average changes in
income distribution and economic growth as a result of a change in the degree of
undervaluation, respectively.
Table 15: Aggregate Demand Components: Short-Run Multipliers, 1990-2010
Countries
Belgium 1.39 1.62 3.01 -0.16 2.59
Canada 0.137 0.290 0.427 -0.118 0.476
France -0.211 0.396 0.185 0.007 0.224
Greece -0.934 0.265 -0.669 0.059 0.226
UK -7.02 -0.01 -7.03 -0.31 0.84
US -4.869 -1.228 -6.098 -0.110 2.000
Average by Region -1.92 0.22 -1.70 -0.11 1.06
Camerun 0.204 -0.019 0.185 -0.033 0.438
Cote d' Ivore 0.00 2.80
Egypt -0.188 -0.054 -0.242 -0.021 0.267
Nigeria 0.045 0.242
Senegal -3.089 0.045 -3.044 -0.374 0.476
South Africa -0.13 1.77
Tanzania -0.989 -0.10 0.84
Average by Region -0.089 0.98
China 0.29 0.20 0.50 -0.10 1.26
Hong Kong 5.02 -1.25 3.76 -0.49 3.18
India -1.07 0.61 -0.46 -0.27 1.55
Indonesia -0.88 0.64 -0.24 -0.08 1.63
Japan 1.085 -0.694 0.392 0.176 0.626
Korea 0.45 -0.77 -0.32 -0.04 3.58
Malaysia 4.17 -0.98 3.19 0.00 1.76
Singapore 22.03 -6.30 15.73 0.42 5.59
Thailand 6.41 -3.59 2.82 0.31 0.68
Average by Region 4.17 -1.35 2.82 -0.009 2.21
Argentina 0.101 0.007 0.108 0.001 0.100
Brazil -0.286 0.408 0.122 -0.060 0.210
Chile -0.013 -0.078 -0.091 0.160 0.424
Colombia 0.157 -0.213 -0.056 -0.021 0.276
Guatemala -2.811 0.019 -2.791 -0.135 0.972
Mexico 0.089 0.189 0.277 0.369 2.090
Panama 0.356 0.158 0.515 -0.043 0.994
Peru 4.156 -1.078 3.078 -0.383 3.169
Uruguay 0.109 -0.059 0.050 -0.018 0.299
Average by Region 0.207 -0.072 0.135 -0.015 0.948
102
The main findings of this short-run analysis can be understood by looking at the average
estimation for these ratios for each region, where we can see that for our sample of
African and Latin American countries, on average, an undervalued currency is correlated
with low economic growth (as their average ratios (𝑔𝑟𝑜𝑤𝑡ℎ %𝑟𝑥𝑟⁄ ) < 1), that is, real
devaluations have tended to be contractionary with a regressive distribution of income (as
the average ratio (∆(𝑊 𝐺𝐷𝑃⁄ ) %𝑟𝑥𝑟⁄ ) < 0), especially for African countries. For the case
of the Latin American countries, although real exchange rate devaluations have
stimulated economic growth via their stimulative effects on private investment (0.207)
and external sector (0.135), these multiplier effects have been week, especially if we
compare these multiplier effects with those calculated for the Asian countries (4.17 and
2.82, respectively).
For the case of developed countries, although our sample is very small, we can
surmise that only for these countries an undervalued currency has had very negligible
effects on economic growth ((𝑔𝑟𝑜𝑤𝑡ℎ %𝑟𝑥𝑟⁄ ) = 1.06) as well as having a negative
effect on the income distribution for these countries ((∆(𝑊 𝐺𝐷𝑃⁄ ) %𝑟𝑥𝑟⁄ ) < 0). For these
developed countries, the driving force of economic growth, on average, has been the
excess of government expenditure over taxes, that is, public debt.
For the case of the Asian countries, our short-run multipliers (correlations) in Table
15 show that this region has outperformed the other regions of the world. That is, for
these Asian countries, an undervalued currency has tended to boost aggregate demand via
a higher level of exports over imports, which is connected or explained by an increasing
difference between total investment and national total savings, that is, increasing levels of
103
FDI. Although these countries have also seen a regressive distribution of income as a
consequence of the undervaluation of the currency, on average, this reduction in equity
has been lower than the other regions of the world.
The results in Table 15 explain, to some degree, why the Latin American countries
vis-à-vis the Asian countries have been unable to create more economic growth through
an undervalued currency. Other factors that help to explain these poor results are the
structural economic problems of Latin America such as low levels of productive
investment, structural trade imbalances, and low productivity growth (Palma, G. 2010);
all of these traits are associated with the worst income distribution in the world (ECLAC
2012).
Finally, Figure 10 shows the changes in total effective demand due to changes in the
degree of undervaluation for the contrasting cases of China, Argentina, and Mexico. On
the one hand, we observe that the changes in the degree of undervaluation of the renminbi
are closely associated with the changes in the growth rate, and a positive increase in the
level of investment over the saving, and the level of exports over the imports. An
undervalued renminbi also seems to be correlated negatively with the China’s wage-
share. On the other hand, in the cases of Argentina (2002) and Mexico (1995), we
observe the classical effects of a real devaluation on the economy: i.e. the fulfillment of
the M-L condition, an initial strong drop of the real GDP (despite the strong increase in
the level of investment over saving) and a strong drop in the wage-share.
104
Figure 10: Undervaluation and Effective Demand: China, Argentina, and Mexico
Although the real devaluation was much stronger and longer maintained in Argentina
than in Mexico, which allowed the former to maintain trade surpluses for several years, it
is likely that the quicker reversion of the drop in the wage-share in Argentina compared
with Mexico was due to quick and socially targeted government intervention. In
Argentina, after the devaluation of 2001-2, the conversion of a fiscal surplus into a fiscal
deficit took only one year, whereas for Mexico, an effective fiscal response came three
years after the December 1994 devaluation. This is a working hypothesis.
-4
-2
0
2
4
6
8
10
90 92 94 96 98 00 02 04 06 08 10 12
G - T X - M I - S
China1990 - 2012
% o
f G
DP
0
10
20
30
40
50
2
4
6
8
10
12
14
16
90 92 94 96 98 00 02 04 06 08 10 12
UNDERVAL GROWTH
China1990 - 2012
%
%
0
10
20
30
40
50
40
44
48
52
56
90 92 94 96 98 00 02 04 06 08 10 12
UNDERVAL WS
China1990 - 2012
%
% o
f GD
P
-8
-4
0
4
8
12
16
90 92 94 96 98 00 02 04 06 08 10 12
G - T X - M I - S
Argentina1990 - 2012
% o
f G
DP
-40
0
40
80
120
-15
-10
-5
0
5
10
15
90 92 94 96 98 00 02 04 06 08 10 12
UNDERVAL GROWTH
Argentina1990 - 2012
%
%
-40
0
40
80
120
32
36
40
44
48
52
90 92 94 96 98 00 02 04 06 08 10 12
UNDERVAL WS
Argentina1990 - 2012
%
% o
f GD
P
-6
-4
-2
0
2
4
6
90 92 94 96 98 00 02 04 06 08 10 12
G - T X - M I - S
Mexico1990 - 2012
% o
f G
DP
-20
-10
0
10
20
-8
-4
0
4
8
90 92 94 96 98 00 02 04 06 08 10 12
UNDERVAL GROWTH
Mexico1990 - 2012%
%
-20
-10
0
10
20
36
38
40
42
44
46
90 92 94 96 98 00 02 04 06 08 10 12
UNDERVAL WS
Mexico1990 - 2012
%
% o
f GD
P
105
III.5 Undervaluation, Investment, and External Sector in the Long-Run
According to the standard theory, an undervalued currency can have positive effects on
the level of investment and on the trade balance (see Rodrik 2008 and Razmi et al. 2009),
so this section seeks to assess the effects (if any) of the rxr (measured according to the
standard methodology based on equation 3) on the level of investment with respect to
GDP (𝐼 𝐺𝐷𝑃⁄ ) and the trade balance of goods (𝑋 𝑀⁄ ) for the different clusters of
developed and developing countries defined above mainly for the period 1981-2010.
Using a series of panel data regression models with six different samples, we use the
following two econometric specifications (equation 11 & 12) in order to test the
aforementioned effects:
ln𝐼
𝐺𝐷𝑃= 𝛾1 + 𝛾2𝐺𝑟𝑜𝑤𝑡ℎ + 𝛾3Ln 𝑟𝑥𝑟 + 𝑓𝑖 + 𝑓𝑡 + 𝑢𝑖 (11)
ln𝑋
𝑀= 𝜃1 + 𝜃2 ln 𝐺𝐷𝑃∗ + 𝜃3 ln 𝐺𝐷𝑃 + 𝜃4Ln 𝑟𝑥𝑟 + 𝑓𝑖 + 𝑓𝑡 + 𝑒𝑖 (12)
Equations 11 and 12 depict two two-way FE models where the dependent variable in
equation 11 is the rate of investment as a percentage of the GDP, growth is the economic
growth rate and rxr is our calculated index of real exchange rate, 𝑓𝑖 is a country specific
effect, 𝑓𝑡 is a time specific effect, and 𝑢𝑖 is an error term. In equation 12, the dependent variable
is the trade balance of goods, 𝐺𝐷𝑃∗ is the world GDP which we incorporated in order to calculate
an external demand effect, 𝐺𝐷𝑃 is the national gross domestic product, and 𝑒𝑖 is an error term.
We first refer to the method to calculate the estimates of equation 11 and its results. We
subsequently do the same for equation 12.
106
From the results showed in Table 14, we know that the rxr has a direct impact on the
rate of economic growth (growth), so the estimates of equation 11 would be biased and
thus misleading. In order to avoid this collinearity problem, we decided to apply a Two-
Stage Least Square (2SLS) method using two regressions to estimate equation 11:
𝐺𝑟𝑜𝑤𝑡ℎ = 𝜖1 + 𝜖2Ln 𝑟𝑥𝑟 + 𝑓𝑖 + 𝑓𝑡 + 𝑢𝑖∗ (11𝐴)
ln𝐼
𝐺𝐷𝑃= 𝜗1 + 𝜗2𝑢𝑖
∗ + 𝜗3Ln 𝑟𝑥𝑟 + 𝑓𝑖 + 𝑓𝑡 + 𝑢𝑖∗∗ (11𝐵)
Equation 11A ‘cleans’ or discounts the effect of the rxr on economic growth, so the
variable 𝑢𝑖∗ in equation 11B (error term in equation 11A) represents economic growth without
the effect of the rxr on it. Hence the parameter 𝜗2 measures the accelerator effect of
economic growth (demand growth) over the rate of investment, whereas the parameter 𝜗3
measures the effect of rxr on the rate of total investment. The estimated results for equation 11B
(using Eviews 9) are shown in table 16.
Table 16: Undervaluation and Investment Rate by Economic Region
The estimated results in Table 16 indicate that all the economies in our six different
samples have a positive accelerator effect, as the parameter 𝜗2 is positive for all of them.
However we can see that the accelerator effect of demand on the investment rate on
Adj R-Square FE Period
All Countries (96) 0.011 [10] 0.14 [7.2] 0.51 1w, C 1982-2010
Developed (25) 0.008 [3.9] -0.137 [-4.2] 0.55 1w, C 1982-2010
Developing (71) 0.014 [9.3] 0.11 [4.1] 0.49 1w, C 1982-2010
Africa (35) 0.014 [5.5] 0.22 [4.7] 0.44 1w, C 1982-2010
Asia (18) 0.012 [5.4] -0.11 [-2.3] 0.55 1w, C 1983-2010
L.A. (20) 0.012 [5.1] 0.09 [2.7] 0.47 2w 1982-2010
Note: The numbers in brackets stand for the t-statistic.
107
average is much higher in developing economies (0.014*100=1.4%) than in developed
economies (0.008*100=0.8%)22. Note that Africa is the region with the highest value for
this parameter (0.014). Asia and Latin American economies have similar accelerator
effects (0.012).
With regard to the effect of the rxr on the investment rate (𝜗3), the first thing to note
is that the rxr has a negative elasticity-effect on the investment rate for developed
countries but a positive elasticity-effect on the investment rate for developing countries
on average, which means that an increase (decrease) in the level of undervaluation
(overvaluation) tends to increase (decrease) the investment rate in developing countries,
whereas undervaluation (overvaluation) tends to decrease (increase) the investment rate
in developed countries (similar conclusions are reported also by Razmi et al., 2009).
African and Latin American countries show a positive elasticity-effect of the rxr on the
investment rate, although the elasticity-effect of the former (0.22) more than doubles the
effect on the latter (0.09). However, it is worth mentioning that the final effect on the rate
of investment would hinge on the degree of over-under/valuation of the currency.
The effect of the rxr on the investment rate in the case of Asian economies appears to
be negative, which we believe could be due to two reasons: (i) the sample of Asian
economies is a mix of developed (6) and developing (12) countries, so it is likely that this
result is been driving by the former set of countries rather than the latter; and (ii) it is also
likely that rxr undervaluation only is capable of enhancing the level of FDI (as we could
see with the results in Table 15) but not necessarily the total rate of investment (𝐼 𝐺𝐷𝑃⁄ ).
22 The parameter for growth is a semielasticity, as it was calculated from a log-linear regression, for that
reason we multiply it by 100.
108
In any case, the average result for this sample of Asian countries is that overvaluation of
the rxr tends to increase the total investment rate, whereas undervaluation of the rxr tends
to decrease the total investment rate.
Table 17: Marshall-Lerner Condition by World Economic Region
The estimated results from equation 12 in Table 17, where all the estimated parameters
are elasticities, show that the Marshall-Lerner (M-L) condition is fulfilled by these
clusters of developed and developing countries as the parameter (elasticity) 𝜃4 is positive
for all of them, which means that a real devaluation tends to improve the trade balance
(of goods). These results also indicate that in the medium-to-long term a real devaluation
has stronger expansionary demand effects in developing countries than in developed
countries. For instance, a 10% real devaluation would improve the trade balance in 1.89%
(=0.189*10%) in developing countries, whereas a similar real devaluation would only
improve the trade balance in 1.1% (=0.11*10%) in developed countries.
By economic region, Latin American countries show the highest response (elasticity)
of a real devaluation to the trade balance of goods (0.312). For the Asian countries this
same response (elasticity) was also considerably high (0.25), and for the African
countries this response was a little lower (0.13) but higher than the developed countries
Adj R-Square FE Period
All Countries (96) -0.13 [-2.83] 0.12 [3.36] 0.14 [5.68] 0.75 1w, C 1981-2010
Developed (25) -0.13 [-2.74] 0.15 [3.73] 0.11 [3.34] 0.86 1w, C 1986-2010
Developing (71) -0.208 [-3.3] 0.15 [3.2] 0.189 [6.16] 0.73 1w, C 1981-2010
Africa (35) -0.57 [-4.92] 0.40 [4.55] 0.13 [2.29] 0.73 1w, C 1984-2010
Asia (18) 0.405 [4.75] -0.09 [-2] 0.25 [5.30] 0.80 1w, C 1981-2010
L.A. (20) 0.46 [3.07] -0.71 [5.9] 0.312 [6.2] 0.83 1w, C 1984-2010
Note: The numbers in brackets stand for the t-statistic.
109
(0.11). However, as we pointed out before, the final effect on the trade balance would
depend on the degree of over-under/valuation of the currency as well as the maintained
period of the under-over/valuation of the currency. Connecting these results in Table 17
with the results obtained in section III, we can conclude that the Latin American
countries have not taken full advantage of this high response of the trade balance to the
change of the real exchange rate. One observes a persistent tendency towards real
exchange rate appreciation in Latin America. The opposite is true for African economies
countries but especially for the case of the Asian countries, as these latter economies have
managed to maintain more periods of currency undervaluation (see Figure 9).
With regard to the parameter 𝜃3 in Table 16, which measures the effect of the
changes in GDP on the trade balance of goods, our results show that for the whole sample
of countries (96), the sample of developed countries (25), the sample of developing
countries (71), and the sample of African countries (35), an increase in the GDP tends to
improve the trade balance of goods. This result is especially true for the African
economies as they reported the highest elasticity (0.40) for the relationship between the
GDP and the trade balance of goods, which indicates that an important part of the
domestic production is exported to other countries.
For the cases of the Asian and Latin American countries, the parameter 𝜃3 turned out
to be negative. However, the magnitude of the difference in this parameter 𝜃3 between
both regions indicates that Latin American countries, on average, suffer a serious
structural trade deficit when income increases. For instance, an increase in 3% in the
GDP leads to an increase in the trade deficit in the Asian countries of -0.27% (=3%*-
110
0.09), whereas the same increase in the GDP leads to an increase in the trade deficit in
the Latin American countries of -2.13% (=3%*-0.71).
The parameter 𝜃2 in Table 16, which measures the effect of the external demand on
the trade balance of goods for our sample of countries, indicates that only the Latin
American and Asian countries, on average, have been able to improve their trade balance
(of goods) when the world GDP increases. The results also indicate that the other set of
countries have a negative relationship between their trade balance (of goods) and the
world GDP.
III.6 Concluding Remarks
Based on our own reformulation of Rodrik’s (2008) undervaluation index “based on
standard theory” using the PWT 7.1, this paper presents evidence of different long-run
patterns in the relationship between the real exchange rate and income per capita for
developed and developing countries. With these differences between countries in mind,
we subsequently estimated different Balassa-Samuelson effects between countries. We
calculated possibly better estimations of real exchange rate under/over-valuations.
This paper also finds, like other previous research, that real exchange rate
undervaluation tends to enhance, although with different degrees and intensity, economic
growth (GDP and GDP per capita) in the medium-to-long run for developed and
developing countries for the periods 1960-1981 and 1982-2010. The only exception for
this positive relationship was our sample of African countries for the period 1960-1981.
111
By and large, the real exchange rate has not been neutral in stimulating economic growth
in the long-run.
Using the Stock Flow Consistent approach and our own index of undervaluation for
the period 1990-2010, we could conclude that in the short-run, on average, the sample of
Asian economies see positive and larger effects of an undervalued currency on economic
growth. The calculated multipliers show that an undervalued currency in our sample of
Asian countries boost aggregate demand via a higher level of exports over imports, which
is connected or explained by an increasing difference between total investment and
national total savings, that is, increasing levels of FDI. Although the Latin American
countries in our sample also see a similar chain of effects, the average multiplier effects
for this region were weak in comparison to those obtained for the Asian economies. The
estimated multiplier effects for developed countries suggest that an undervalued currency
also has week effects on economic growth. Due to the lack of data, we could not
undertake a short-run analysis for our sample of African countries.
Another important finding from our short run analysis is that periods of exchange
rate undervaluation are associated with negative changes in the participation of total
wages in the GDP, that is, real devaluation tends to be regressive in all the economic
areas analyzed in this paper, but specially in our small sample of developed economies, in
the sample for African countries, and to a lesser extent, in Latin American and Asian
countries.
Based on a visual inspection of the main components of the aggregate demand with
respect to GDP, the growth rate, and the wage-share, we can observe that strong currency
112
devaluations in developing countries are contractionary and regressive (see Figures 12-15
in the appendix). That is, despite the increasing stimulus of the level of exports over
imports and the level of investment over national savings, which are associated with the
devaluation of the currency, the level of output tends to fall considerably, and with it, the
participation of total wages in the GDP and the general level of consumption. Although
in the medium-to-long-run, developing economies tend to improve their economic
performance due to the positive effects of a devaluated currency on the level of
investment and exports over imports (see the results in Table 16 and 17), for many
developing countries a prolonged (permanent) currency devaluation tends to maintain
lingering negative effects on the level of the wage-share, which despite the positive effect
of the weakened currency on growth, could require many years to return to the pre-
devaluation level of the wage share.
The bottom line of this paper is that the real exchange rate is a double-edged sword.
On the one hand, an undervalued currency could improve international competitiveness
in the short-to-medium term but could decrease the wage-share over time. Conversely, an
overvalued currency could improve the wage-share in the short-to-medium term, but
reduce the general level of competitiveness through time. Therefore, if a country wishes
to improve its international competitiveness without significantly affecting its national
distribution of income, it should take efforts to increase its international competitiveness
through the development of new technologies and more differentiated products. Of
course, such countries should also seek lower production costs. In the absence of a strong
national bourgeoisie, the response could come from a vigorous public sector capable of
developing relevant and dynamic economic sectors.
113
References
Amsden, Alice (2001), “The Rise of the “The Rest”: Challenges to the West from Late-
Industrialing Economies” Oxford University Press.
Arellano, M. and Bond, S. (1991), “Some tests of specification for panel data: Monte
Carlo evidence and an application to employment equations” Review of Economic
Studies 58 (2).
Baltagi (2005), “Econometric Analysis of Panel Data” Third Edition.
Bhalla Surjit (2013), “Devaluing to Prosperity: Misaligned currencies and their Growth
Consequences” Peterson Institute for International Economics, Washington, DC.
Blecker R. and Razmi A. “The fallacy of composition and contractionary devaluation:
output effects of real exchange rate shocks in semi-industrialised countries”
Cambridge Journal of Economics 2008, 32, 83-109.
Chang, Ha-Joon (2013), "Britain, a nation in decay". The Guardian, March 8. Available
at [http://www.theguardian.com/commentisfree/2013/mar/08/britain-economy-long-
term-fix].
Diaz Alejandro, Carlos (1963), “A note on the impact of Devaluation and the
Redistributive Effect” Journal of Political Economy, Vol. 71, No. 6.
Frenkel, Jeffrey (2005), “Mundell-Fleming Lecture: Contractionary Currency Crashes in
Developing Countries” IMF Staff papers, Vol. 52, Number 2.
Frenkel, R. and Taylor, L (2006), Real Exchange Rate, Monetary Policy and
Employment: Economic Development in a Garden of Forking Paths. Political
economy Research Institute.
Godley, Wynne and Lavoie, Marc (2007), “Monetary Economics: An Integrated
Approach to Credit, Money, Income, Production and Wealth” Palgrave Macmillan.
Johnson, S., Ostry, J., Subramanian, A. (2007), “The Prospects for Sustained Growth in
Africa: Benchmarking the Constraints”, IMF Working Paper No. 07/52.
Kalyoncu, H., Artan, S., Tezekici, S. (2008), “Currency Devaluation and Output Growth:
An Empirical Evidence from OECD Countries” International Research Journal of
Finance and Economics, issue 14.
114
Karshenas, Massoud (2007), "Real Exchanges Rates, Labor Markets, and Manufacturing
Exports in a global perspective" in Shaikh, A. (2007) "Globalization and The Myths of
Free Trade: History, Theory, and Empirical Evidence" Routledge.
Lopez and Perrotini (2006), “On floating exchange rates, currency depreciation and
effective demand” Banca Nazionale del Lavoro, vol. LIX, no 238, September, pp.
221-42.
Moudud, Jamee (2010), “Strategic Competition, Dynamics, and the Role of the State. A
New Perspective” Edward Elgar.
Palma, G. (2010), “Why has productivity growth stagnated in most Latin American
countries since the neo-liberal reforms?” May CWPE 1030.
Razmi, Rappetti, and Skott (2009), "The Real Exchange Rate as an Instrument of
Development Policy". UMAS working paper series.
Rodrik, Dani (2008), “The Real Exchange Rate and Economic Growth“. Working Paper,
John F. Kennedy School of Government Harvard University.
Shaikh, Anwar (2012), “Three Balances and Twin Deficits: Godley versus Ruggles and
Ruggles” in Contributions in Stock-Flow Modeling, Essays in Honor to Wynne
Godley. Edited by Dimitri Papadimitriou and Gennaro Zezza.
Taylor, Lance (2004), “Reconstructing Macroeconomics: Structuralist Proposals and
Critiques to the Mainstream” Harvard University Press.
UNCTAD Trade and Development Report 1981-2011 (2012), “Three Decades of
Thinking Development” Part I. Geneva.
115
Appendix to Chapter I-A
Table I.1A
116
OCDE Data
Year RULCoecd cpioecd ppioecd (pmg*e)oecd IntRateoecd RULCADJoecd
1960 157.3 13.7 20.4 25.0 5.1 105.2
1961 158.3 14.1 20.9 25.3 5.0 106.5
1962 157.9 14.4 21.1 25.6 4.9 108.1
1963 155.2 14.8 21.3 25.9 4.9 107.7
1964 152.7 15.4 22.0 26.5 5.3 106.7
1965 150.9 16.1 22.7 27.0 5.7 106.9
1966 150.3 16.7 23.3 27.7 6.3 107.9
1967 150.4 17.2 23.5 28.3 6.1 110.0
1968 147.3 18.0 24.1 28.4 6.5 110.1
1969 145.8 18.9 25.2 29.2 7.2 109.4
1970 150.2 20.0 26.7 30.8 7.7 112.4
1971 150.5 21.2 27.8 32.9 6.8 114.8
1972 149.5 22.4 29.0 36.3 6.6 115.5
1973 147.2 24.4 32.2 41.7 7.8 111.5
1974 150.7 27.5 38.3 46.6 9.4 108.3
1975 155.2 30.6 40.9 52.5 8.3 116.2
1976 150.2 33.7 44.3 52.9 8.4 114.3
1977 147.9 36.7 47.6 57.0 7.8 114.2
1978 147.1 39.3 50.2 66.1 7.8 115.3
1979 144.8 42.2 54.4 73.5 9.4 112.3
1980 144.4 46.9 61.0 79.7 11.2 111.0
1981 141.7 51.7 67.5 74.4 12.9 108.7
1982 140.3 55.8 72.0 71.8 11.9 108.7
1983 135.1 59.1 75.2 70.3 10.1 106.2
1984 132.1 61.9 78.8 68.8 10.2 103.8
1985 129.3 64.4 80.7 68.0 8.9 103.2
1986 130.3 66.0 79.3 84.7 7.2 108.5
1987 129.1 67.6 79.5 96.5 7.2 109.8
1988 126.5 69.5 81.7 103.2 7.4 107.7
1989 124.2 72.5 85.1 102.1 8.5 105.7
1990 123.2 75.9 87.4 112.8 9.7 106.9
1991 123.2 79.2 88.8 114.2 8.7 109.9
1992 122.2 81.8 89.3 119.2 7.7 112.0
1993 120.9 84.1 90.0 113.8 5.9 113.0
1994 116.1 86.0 91.3 116.3 6.4 109.3
1995 113.9 87.9 94.4 127.3 6.1 106.1
1996 112.6 89.6 95.0 123.7 5.3 106.2
1997 109.2 91.4 95.9 114.1 4.8 104.1
1998 108.3 92.8 95.5 111.3 4.0 105.2
1999 105.7 94.1 95.8 109.4 4.1 103.8
2000 101.7 96.2 100.0 101.4 4.8 97.8
2001 101.8 98.3 100.5 97.2 3.8 99.6
2002 100.0 100.0 100.0 100.0 3.4 100.0
2003 97.6 102.0 100.7 111.3 2.7 98.8
2004 92.9 104.0 103.5 119.7 3.0 93.3
2005 89.9 106.2 107.3 121.1 2.9 89.0
2006 86.9 108.5 110.8 122.2 3.6 85.1
2007 84.3 110.9 113.9 131.0 3.9 82.0
2008 85.5 114.4 120.3 136.9 3.2 81.3
2009 91.5 114.8 115.6 132.6 2.3 90.8
2010 83.4 116.6 119.7 131.4 2.0 81.3
Source: See table 1A and Chapter 1, section V.
117
Table I.1B
Source: Anwar Shaikh (2011).
118
Table I.1C
This appendix I.1-C describes the methodology to construct real effective exchange rates, relative real adjusted unit
labor costs, and real interest rate differentials for two set of countries, all these series were calculated in annual index
form (2002=100) for the period 1960-2010. The first set of 16 countries includes: Australia, Belgium, Canada,
Denmark, Finland, France, Germany, Italy, Japan, South Korea, Netherlands, Norway, Spain, Sweden, UK, and the
US. The second set of 4 countries includes: Argentina, El Salvador, Mexico, and Taiwan. We focus first on the
methodology for the first set of countries and then describe the methodology for the second set of countries.
Our real effective exchange rates (rxr) is a trade-weighted real effective exchange rate index which has as its
main components for each country the manufacturing price index (pmfg, national currency), the index of the nominal
exchange rate (e =Foreign Currency/Dollar) and a geometric traded weighted average of manufacturing prices
((𝒑𝒎𝒈 ∗ 𝒆)𝒐𝒆𝒄𝒅). The first both indexes were obtained from the U.S. Bureau of Labor Statistics.
(𝟏) 𝒓𝒙𝒓𝒄,𝒕 =𝒑𝒎𝒇𝒈𝒊,𝒕∗𝒆𝒊,𝒕∗𝟏𝟎𝟎
((𝒑𝒎𝒈∗𝒆)𝒐𝒆𝒄𝒅)𝒕
Where i stands for country (e.g., Australia,…, U.S.), t stands for time (e.g., 1960,…, 2010) and where 𝑝𝑚𝑓𝑔𝑖,𝑡 ∗
𝑒𝑖,𝑡 = 𝑝𝑚𝑓𝑔𝑖,𝑡 ∗𝑒𝑖,𝑡
100 .
The denominator of formula 1, which is a manufacturing price index for the 16 OECD countries, is a geometric
traded weighted average which uses total flows of trade (w=X+M in millions of US Dollars). Each year (t) takes into
account just 16 countries (i). The trade flows data were obtained from the IMF (IFS). The methodology for the
estimation of the denominator is described in the following three formulas:
(𝑝𝑚𝑓𝑔 ∗ 𝑒)𝑜𝑒𝑐𝑑𝑡 = ∏ 𝑥𝑖,𝑡
16
𝑖=1
𝑥𝑐,𝑡 = 𝑝𝑚𝑓𝑔𝑖,𝑡 ∗ 𝑒𝑖,𝑡
(𝑇𝑟𝑎𝑑𝑒 𝑊𝑒𝑖𝑔ℎ𝑡𝑠)𝑖,𝑡 where 𝑇𝑟𝑎𝑑𝑒 𝑊𝑒𝑖𝑔ℎ𝑡𝑠𝑖,𝑡 = 𝑤𝑖,𝑡/ ∑ 𝑤𝑗,𝑡16𝑗=1
Trade weights reflect country’s share in the total trade flows (M+X) of the 16 OECD countries. The real adjusted unit
labor cost ratio (rulcadjratio) was calculated on the basis of formula 2 below. The manufacturing unit labor cost
indexes from all the countries were obtained from the U.S. Bureau of Labor Statistics (national currency basis). The
indexes to calculate the adjustment component, CPI and PPI, were obtained from the U.S. Bureau of Labor Statistics
and OECD, respectively. PPI stands for manufacturing producer price index. Although CPI stands for consumer price
index, we calculated this index as a ratio of unit labor cost and the real unit labor cost, both in local currency.
(𝟐) 𝒓𝒖𝒍𝒄𝒂𝒅𝒋𝒓𝒂𝒕𝒊𝒐𝒊,𝒕 =𝑹𝑼𝑳𝑪𝒂𝒅𝒋𝒊,𝒕 ∗ 𝟏𝟎𝟎
𝑹𝑼𝑳𝑪𝒂𝒅𝒋𝑶𝑬𝑪𝑫𝒕
Where 𝑅𝑈𝐿𝐶𝑎𝑑𝑗𝑖,𝑡 = (𝑅𝑈𝐿𝐶𝑖,𝑡 ∗ (𝐶𝑃𝐼𝑖,𝑡
𝑃𝑃𝐼𝑖,𝑡)) and the real unit labor cost is 𝑅𝑈𝐿𝐶𝑖,𝑡 = (
𝑈𝐿𝐶𝑖,𝑡
𝐶𝑃𝐼𝑖,𝑡) ∗ 100
The denominator in (2), which is a manufacturing price index for the 16 OECD countries, is a geometric traded
weighted average that was calculated as a compound geometric traded weighted average as follows:
𝑅𝑈𝐿𝐶𝑎𝑑𝑗𝑂𝐸𝐶𝐷𝑡 = (𝑅𝑈𝐿𝐶𝑜𝑒𝑐𝑑𝑡 ∗ (𝐶𝑃𝐼𝑜𝑒𝑐𝑑𝑡
𝑃𝑃𝐼𝑜𝑒𝑐𝑑𝑡))
The geometric traded weighted average for the real unit labor for each year (t) was obtained as follows:
𝑅𝑈𝐿𝐶𝑜𝑒𝑐𝑑𝑡 = ∏ 𝑦𝑖,𝑡16𝑖=1 where 𝑦𝑖,𝑡 = 𝑅𝑈𝐿𝐶𝑖,𝑡
(𝑇𝑟𝑎𝑑𝑒 𝑊𝑒𝑖𝑔ℎ𝑡)𝑖,𝑡
The geometric traded weighted average for the CPI for each year (t) was obtained as follows:
119
𝐶𝑃𝐼𝑜𝑒𝑐𝑑𝑡 = ∏ 𝑧𝑐,𝑡16𝑖=1 where 𝑧𝑖,𝑡 = 𝐶𝑃𝐼𝑖,𝑡
(𝑇𝑟𝑎𝑑𝑒 𝑊𝑒𝑖𝑔ℎ𝑡)𝑖,𝑡
The geometric traded weighted average for the PPI for each year (t) was obtained as follows:
𝑃𝑃𝐼𝑜𝑒𝑐𝑑𝑡 = ∏ 𝑥𝑦𝑖,𝑡16𝑖=1 where 𝑥𝑦𝑖,𝑡 = 𝑃𝑃𝐼𝑖,𝑡
(𝑇𝑟𝑎𝑑𝑒 𝑊𝑒𝑖𝑔ℎ𝑡)𝑖,𝑡
The data to calculate real interest rate differentials was obtained from the International Monetary Fund (IMF,
IFS). We mainly used Treasury bill rates with a maturity of 3 months. In order to calculate the real interest rate
differential for each country, we first calculated the nominal interest rate differentials for each country under the basis
of the following formula:
(𝟑) 𝒊𝒏𝒕𝒓𝒂𝒕𝒆𝒅𝒊𝒇𝒇𝒕 = 𝑰𝒏𝒕𝑹𝒂𝒕𝒆𝒔𝒕 − 𝑰𝒏𝒕𝑹𝒂𝒕𝒆𝑶𝒆𝒄𝒅𝒕
Where 𝑖𝑛𝑡𝑟𝑎𝑡𝑒𝑑𝑖𝑓𝑓 stands for nominal interest rate differential, 𝐼𝑛𝑡𝑅𝑎𝑡𝑒𝑠 stands for nominal interest rate, and
𝐼𝑛𝑡𝑅𝑎𝑡𝑒𝑂𝑒𝑐𝑑 stands for a geometric traded-weighted average which was calculated on the basis of the following two
equations:
𝐼𝑛𝑡𝑅𝑎𝑡𝑒𝑂𝑒𝑐𝑑𝑡 = ∏ 𝑥𝑧𝑖,𝑡16𝑖=1 where 𝑥𝑧𝑖,𝑡 = 𝐼𝑛𝑡𝑅𝑎𝑡𝑒𝑠𝑖,𝑡
(𝑇𝑟𝑎𝑑𝑒 𝑊𝑒𝑖𝑔ℎ𝑡)𝑖,𝑡
Thus, the real interest rate differential for each country was calculated as a difference between intratediff and the
growth rate of a geometric traded-weighted average of the producer price index from the 16 OECD countries
(gppiratio):
(𝟒) 𝒓𝒆𝒂𝒍𝒊𝒏𝒕𝒓𝒂𝒕𝒆𝒅𝒊𝒇𝒇𝒕 = 𝒊𝒏𝒕𝒓𝒂𝒕𝒆𝒅𝒊𝒇𝒇𝑡 − 𝒈𝒑𝒑𝒊𝒓𝒂𝒕𝒊𝒐𝑡
Where 𝑔𝑝𝑝𝑖𝑟𝑎𝑡𝑖𝑜𝑡 = (100 ∗((𝑝𝑝𝑖 𝑝𝑝𝑖𝑂𝐸𝐶𝐷⁄ )𝑡−(𝑝𝑝𝑖 𝑝𝑝𝑖𝑂𝐸𝐶𝐷⁄ )𝑡−1)
(𝑝𝑝𝑖 𝑝𝑝𝑖𝑂𝐸𝐶𝐷⁄ )𝑡−1)
and 𝑝𝑝𝑖 𝑝𝑝𝑖𝑂𝐸𝐶𝐷⁄𝑡
= (𝑃𝑃𝐼𝑡 ∗ (100
𝑝𝑝𝑖𝑂𝐸𝐶𝐷𝑡)). The geometric traded-weighted average of the
producer price index was calculated on the basis of the following two calculations:
𝑝𝑝𝑖𝑂𝐸𝐶𝐷𝑡 = ∏ 𝑦𝑧𝑖,𝑡16𝑖=1 and 𝑦𝑧𝑖,𝑡 = 𝑃𝑃𝐼𝑖,𝑡
(𝑇𝑟𝑎𝑑𝑒 𝑊𝑒𝑖𝑔ℎ𝑡)𝑖,𝑡.
Table I.1D: Correlation analysis: real ULC ratio and trade balance
Coefficient Coefficient Coefficient
Argentina -0.497 France -0.197 Norway 0.911
Australia -0.195 Germany 0.288 Spain 0.635
Belgium 0.441 Italy -0.130 Sweden -0.805
Canada -0.367 Japan 0.731 Taiwan -0.196
Denmark 0.591 Korea -0.468 UK -0.223
El Salvador 0.210 Mexico -0.029 US 0.880
Finland -0.547 Netherlands -0.414
120
(1) Argentina
e 1/e 1/e rxr Interest Rate
Year (Peso/Dollar) (Dollar/Peso) 1970=100 1970=100 rxr RULCratio Nominal Int. Rate CPI 1970=100 Inflation Rate Real Int. Rate US Int. Rate CPI 1970=100 Inflation Rate Real Int. Rate Differential
1960 8.7E-12 1.1E+11 435 83 78 128 NA 15 - NA 4.0 76 - - NA
1961 8.7E-12 1.1E+11 435 94 87 130 NA 17 13 NA 3.5 77 1.0 2.6 NA
1962 8.7E-12 1.1E+11 435 119 111 133 NA 21 28 NA 3.5 78 1.2 2.3 NA
1963 1.4E-11 7.2E+10 273 91 85 132 NA 26 24 NA 3.7 79 1.2 2.5 NA
1964 1.4E-11 7.1E+10 270 109 101 133 NA 32 22 NA 4.0 80 1.6 2.4 NA
1965 1.7E-11 5.9E+10 224 114 106 138 NA 41 29 NA 4.2 81 1.2 3.0 NA
1966 2.1E-11 4.8E+10 181 118 110 148 NA 55 32 NA 5.2 84 3.5 1.7 NA
1967 3.3E-11 3.0E+10 114 93 86 147 NA 70 29 NA 5.0 86 3.1 1.9 NA
1968 3.5E-11 2.9E+10 108 99 92 143 NA 82 16 NA 5.7 90 4.0 1.7 NA
1969 3.5E-11 2.9E+10 108 101 94 137 NA 88 8 NA 7.0 95 5.3 1.7 NA
1970 3.8E-11 2.6E+10 100 100 93 138 NA 100 14 NA 7.3 100 5.8 1.5 NA
1971 4.5E-11 2.2E+10 84 108 101 148 NA 135 35 NA 5.7 104 4.4 1.2 NA
1972 5.0E-11 2.0E+10 76 150 140 137 NA 213 58 NA 5.7 108 3.3 2.4 NA
1973 5.0E-11 2.0E+10 76 229 213 150 NA 344 61 NA 7.0 114 5.8 1.2 NA
1974 5.0E-11 2.0E+10 76 254 237 155 NA 425 23 NA 7.8 127 11.1 -3.2 NA
1975 3.7E-10 2.7E+09 10 90 84 160 NA 1,203 183 NA 7.5 139 9.4 -1.9 NA
1976 1.4E-09 7.1E+08 3 121 112 103 NA 6,541 444 NA 6.8 147 5.8 1.0 NA
1977 4.1E-09 2.5E+08 1 107 100 100 NA 18,055 176 NA 6.7 156 6.6 0.1 NA
1978 8.0E-09 1.3E+08 0.476 141 131 107 NA 49,743 176 NA 8.3 168 7.5 0.7 NA
1979 1.3E-08 7.6E+07 0.288 202 188 115 NA 129,086 160 NA 9.7 184 9.6 0.1 NA
1980 1.8E-08 5.4E+07 0.206 275 256 135 87 259160 101 - 11.5 195 5.7 5.9 NA
1981 4.4E-08 2.3E+07 0.086 212 198 128 185 529921 104 81 14.4 215 10.3 4.1 76.6
1982 2.6E-07 3.9E+06 0.015 90 84 100 202 1403103 165 37 12.9 228 6.2 6.8 30.4
1983 1.1E-06 9.5E+05 0.004 95 89 117 739 6227133 344 395 10.4 235 3.2 7.2 387.9
1984 6.8E-06 1.5E+05 0.001 103 96 133 1182 45253676 627 556 11.9 245 4.3 7.6 548.0
1985 6.0E-05 1.7E+04 0.000 87 81 132 1161 349438691 672 489 9.6 254 3.6 6.1 482.9
1986 9.4E-05 1.1E+04 0.000 103 96 136 135 664274647 90 45 7.1 259 1.9 5.2 39.5
1987 0.000 4.7E+03 0.000 101 94 135 253 1536644785 131 121 7.7 269 3.7 3.9 117.5
1988 0.001 1.1E+03 0.000 106 98 114 524 6806654163 343 181 8.3 279 4.0 4.2 176.5
1989 0.042 2.4E+01 0.000 66 62 104 1387179 216438337547 3080 1384099 8.6 293 4.8 3.7 1384095
1990 0.488 2.1E+00 0.000 132 122 132 9695422 5224761477616 2314 9693108 8.3 309 5.4 2.9 9693105
1991 0.954 1.0E+00 0.000 175 163 151 71 14194154502705 172 -100 6.8 322 4.2 2.6 -102.9
1992 0.991 1.0E+00 0.000 205 190 152 15 17728499007856 25 -10 5.3 332 3.0 2.3 -12.1
1993 0.999 1.0E+00 0.000 218 203 167 6 19609767865395 11 -4 4.4 341 3.0 1.5 -5.8
1994 0.999 1.0E+00 0.000 221 206 137 8 20428932561364 4 3 6.3 350 2.6 3.7 -0.2
1995 1.00 1.0E+00 0.000 222 207 141 9 21118642800837 3 6 6.3 360 2.8 3.5 2.6
1996 1.00 1.0E+00 0.000 216 201 135 6 21151499089402 0 6 6.0 371 2.9 3.1 3.0
1997 1.00 1.0E+00 0.000 213 198 128 7 21263047378292 1 6 6.1 379 2.3 3.8 2.3
1998 1.00 1.0E+00 0.000 211 197 147 7 21458724075131 1 6 5.1 385 1.6 3.6 2.3
1999 1.00 1.0E+00 0.000 204 190 143 7 21208343526598 -1 8 5.5 394 2.2 3.3 4.9
2000 1.00 1.0E+00 0.000 196 182 141 8 21009846693305 -1 9 6.2 407 3.4 2.8 6.2
2001 1.00 1.0E+00 0.000 188 175 147 25 20785730789095 -1 26 4.1 418 2.8 1.3 24.7
2002 3.06 3.3E-01 0.000 76 71 95 41 26162677702574 26 15 3.1 425 1.6 1.5 14.0
2003 2.90 3.4E-01 0.000 89 83 92 4 29679693658431 13 -10 2.1 435 2.3 -0.2 -9.5
2004 2.92 3.4E-01 0.000 90 84 98 2 30990275625866 4 -2 2.8 446 2.7 0.1 -2.6
2005 2.90 3.4E-01 0.000 96 89 103 4 33977547636522 10 -6 3.9 461 3.4 0.5 -6.1
2006 3.05 3.3E-01 0.000 98 91 108 7 37681440104379 11 -4 4.8 476 3.2 1.5 -5.2
2007 3.10 3.2E-01 0.000 103 95 112 9 41009201119900 9 0 4.3 490 2.9 1.5 -1.6
Source: See Table 1A.
Argentina-US 1977=100 Argentina Data US Data
121
(2) Australia
Trade
Year e pmfg*e rxr CPI PPI RULC RULCAdj RULCAdjratio Int.Rate intratediff gppiratio realintratediff rxr1 rulcadjratio1 realintratediff1 Weight
1960 206 22 89 10.0 10.4 104 100 95.3 4.4 -0.7 88 95 0.029
1961 206 23 89 10.2 10.6 106 102 95.8 5.0 0.0 -0.04 0.0 89 96 -0.1 0.028
1962 206 23 88 10.2 10.6 108 104 96.4 4.3 -0.6 -0.85 0.2 88 96 0.2 0.027
1963 206 23 88 10.3 10.6 106 102 94.7 3.8 -1.1 -0.63 -0.4 87 95 -0.3 0.028
1964 206 23 88 10.5 10.9 105 101 94.6 4.1 -1.2 -1.07 -0.1 87 94 0.0 0.029
1965 206 24 89 10.9 11.3 108 104 97.1 4.8 -0.9 0.95 -1.8 89 97 -1.9 0.028
1966 206 25 90 11.3 11.7 111 107 99.3 4.9 -1.4 0.40 -1.8 90 99 -1.9 0.025
1967 206 26 91 11.6 12.0 109 105 95.6 4.6 -1.6 2.45 -4.0 91 95 -4.0 0.026
1968 206 26 93 11.9 12.4 111 107 96.8 4.8 -1.6 0.14 -1.8 93 97 -1.8 0.024
1969 206 30 103 12.3 14.1 109 95 87.2 5.3 -2.0 8.80 -10.8 103 87 -11.0 0.024
1970 206 31 100 12.7 14.7 109 94 83.9 6.3 -1.5 -1.31 -0.2 100 84 -0.1 0.023
1971 209 33 99 13.5 15.3 114 101 87.8 6.1 -0.6 -0.36 -0.3 99 88 -0.3 0.022
1972 219 36 100 14.3 16.1 116 103 89.4 4.9 -1.7 0.85 -2.5 100 89 -2.6 0.020
1973 261 48 116 15.6 17.5 115 103 92.5 6.3 -1.5 -1.84 0.3 116 92 0.5 0.022
1974 265 57 123 18.0 20.3 117 104 95.8 9.3 -0.1 -2.60 2.5 123 96 2.5 0.021
1975 241 59 112 20.8 23.1 128 114 98.5 8.5 0.1 6.89 -6.8 112 98 -7.0 0.020
1976 225 61 115 23.5 25.8 126 115 100.3 8.7 0.3 2.94 -2.6 115 100 -2.7 0.020
1977 204 60 105 26.4 28.5 123 114 100.2 9.7 2.0 2.55 -0.6 105 100 -0.8 0.019
1978 211 68 104 28.5 30.8 124 114 99.3 8.8 1.0 2.64 -1.7 104 99 -1.8 0.018
1979 206 77 105 31.1 35.3 117 103 92.0 9.6 0.2 5.82 -5.6 105 92 -5.8 0.017
1980 210 88 110 34.3 40.3 115 98 88.2 11.5 0.3 1.73 -1.5 110 88 -1.5 0.018
1981 211 95 128 37.6 43.7 117 101 92.8 13.8 0.8 -1.80 2.6 129 93 2.7 0.019
1982 187 92 129 41.8 47.7 118 103 95.2 15.2 3.2 2.29 1.0 129 95 0.9 0.021
1983 166 87 124 46.0 51.5 122 109 102.4 12.8 2.8 3.24 -0.5 125 102 -0.7 0.018
1984 162 90 130 47.8 54.3 114 101 96.8 12.2 2.1 0.62 1.4 131 97 1.5 0.020
1985 129 76 112 51.0 57.9 114 100 97.0 14.0 5.2 4.13 1.0 112 97 0.8 0.019
1986 123 77 91 55.7 61.1 113 103 95.1 14.0 6.7 7.45 -0.7 91 95 -1.0 0.017
1987 129 87 90 60.4 65.6 112 103 94.3 13.2 6.0 6.99 -1.0 90 94 -1.2 0.016
1988 144 104 101 64.8 70.3 108 100 92.6 12.2 4.7 4.40 0.3 101 92 0.2 0.018
1989 146 112 110 69.7 75.1 107 99 93.5 15.1 6.6 2.47 4.2 110 93 4.1 0.020
1990 144 115 102 74.4 79.6 111 104 96.9 13.5 3.8 3.23 0.6 102 97 0.4 0.017
1991 143 115 100 77.0 80.7 111 105 96.0 9.9 1.2 -0.16 1.4 100 96 1.4 0.017
1992 135 114 96 77.6 82.1 112 106 94.6 7.6 -0.1 1.09 -1.2 96 95 -1.2 0.017
1993 125 110 97 78.9 83.6 110 104 92.2 6.2 0.3 1.01 -0.7 97 92 -0.8 0.018
1994 135 122 105 80.5 84.3 108 104 94.8 8.2 1.8 -0.61 2.4 105 95 2.4 0.018
1995 136 126 99 84.3 87.4 109 105 99.2 8.5 2.4 0.32 2.1 99 99 2.1 0.017
1996 144 135 109 86.6 88.0 109 107 101.0 7.6 2.3 0.12 2.2 109 101 2.2 0.018
1997 137 131 115 86.8 89.3 109 106 101.7 6.1 1.3 0.36 0.9 115 102 0.9 0.018
1998 116 112 101 87.6 89.8 108 106 100.4 5.0 1.0 1.08 -0.1 101 100 -0.1 0.017
1999 119 115 105 89.0 90.4 107 106 101.6 5.5 1.4 0.35 1.0 105 102 1.0 0.017
2000 107 104 102 93.0 96.8 104 100 102.2 6.2 1.4 2.58 -1.2 102 102 -1.2 0.017
2001 95 92 95 97.3 99.8 100 98 98.1 5.0 1.1 2.56 -1.4 95 98 -1.5 0.016
2002 100 100 100 100.0 100.0 100 100 100.0 5.3 1.9 0.72 1.1 100 100 1.1 0.017
2003 120 126 113 102.8 100.6 98 101 101.8 4.9 2.2 -0.16 2.3 114 102 2.3 0.018
2004 136 152 127 105.2 104.4 100 101 108.3 5.3 2.3 1.03 1.3 127 108 1.2 0.018
2005 140 161 133 108.1 110.7 103 100 112.6 5.3 2.3 2.34 0.0 134 113 -0.1 0.019
2006 139 164 134 112.0 119.5 103 96 113.0 5.7 2.1 4.46 -2.4 135 113 -2.5 0.019
2007 154 186 142 114.6 122.3 103 96 117.4 6.3 2.4 -0.48 2.9 143 118 2.9 0.020
2008 157 195 142 119.7 132.4 103 93 114.4 5.7 2.5 2.57 -0.1 144 115 -0.2 0.023
2009 146 189 142 121.8 125.2 103 100 110.5 4.2 1.9 -1.61 3.5 144 111 3.5 0.025
2010 169 223 169 125.5 127.6 100 99 121.3 4.9 2.9 -1.63 4.5 172 122 4.5 0.027
Source: See Table 1A and 1B.
Excluding Country Being ExaminedInterest RatesReal Exchange Rate Real Adjusted Unit Labor Cost Ratio
122
(3) Belgium
Trade
Year e pmfg*e rxr CPI PPI RULC RULCAdj RULCAdjratio Int.Rate intratediff gppiratio realintratediff rxr1 rulcadjratio1 realintratediff1 Weight
1960 86 30 121 18.3 34.0 165 89 84.3 5.7 0.6 122 84 0.054
1961 86 30 119 18.3 34.0 172 93 87.3 6.0 1.0 -2.45 3.5 120 87 3.5 0.056
1962 86 30 118 18.9 34.3 171 94 86.9 5.3 0.4 0.15 0.2 119 86 0.1 0.058
1963 86 31 119 19.3 35.1 178 98 90.7 5.2 0.3 1.25 -1.0 121 90 -1.1 0.059
1964 86 32 121 20.1 36.4 181 100 93.8 5.9 0.6 0.30 0.3 123 93 0.2 0.059
1965 86 33 123 20.7 36.4 186 106 98.8 5.9 0.1 -2.82 2.9 125 99 3.2 0.058
1966 86 34 123 21.7 36.8 184 108 100.3 6.1 -0.2 -1.86 1.7 124 100 1.8 0.059
1967 86 35 124 22.3 36.8 184 112 101.6 6.2 0.1 -0.74 0.8 125 102 0.9 0.058
1968 86 35 125 22.8 37.2 177 109 98.7 6.4 0.0 -1.31 1.3 127 99 1.4 0.057
1969 85 37 126 23.9 38.5 170 105 96.3 7.0 -0.2 -1.02 0.8 127 96 0.8 0.060
1970 86 38 122 24.6 40.8 170 102 91.1 7.4 -0.4 0.02 -0.4 124 91 -0.3 0.059
1971 88 38 116 25.7 41.6 176 109 95.1 7.0 0.3 -2.26 2.5 117 95 2.7 0.058
1972 97 44 121 27.2 43.3 177 111 96.2 6.8 0.2 -0.04 0.3 122 96 0.2 0.060
1973 110 52 124 29.1 46.5 177 110 99.1 7.3 -0.5 -3.31 2.8 126 99 2.9 0.062
1974 110 59 126 32.6 55.8 184 108 99.3 9.2 -0.2 1.01 -1.2 128 99 -1.1 0.059
1975 116 66 125 37.0 58.2 189 120 103.2 7.5 -0.9 -2.45 1.6 126 103 1.8 0.058
1976 111 65 123 40.2 61.0 182 120 105.3 9.2 0.9 -3.20 4.1 125 106 4.2 0.058
1977 119 74 129 43.2 62.6 175 121 105.7 7.7 -0.1 -4.54 4.4 131 106 4.7 0.059
1978 136 87 131 45.1 63.0 169 121 105.1 7.6 -0.2 -4.57 4.3 133 105 4.6 0.060
1979 146 95 129 47.1 65.4 164 118 104.9 10.0 0.5 -4.20 4.7 132 105 4.9 0.060
1980 146 95 119 50.2 62.7 160 128 115.8 12.9 1.7 -14.55 16.2 120 117 17.0 0.056
1981 115 73 98 54.0 71.6 152 115 105.4 14.2 1.2 3.23 -2.0 98 106 -2.2 0.051
1982 93 65 90 58.9 81.2 139 101 92.6 13.7 1.8 6.36 -4.6 90 92 -5.0 0.050
1983 84 59 83 63.4 86.6 128 94 88.5 11.2 1.1 2.11 -1.0 83 88 -1.2 0.050
1984 74 54 78 67.3 93.2 125 90 87.1 11.9 1.7 2.71 -1.0 77 87 -1.1 0.046
1985 72 56 83 70.6 95.6 123 91 88.0 10.2 1.3 0.10 1.2 82 87 1.2 0.045
1986 96 78 92 71.6 84.6 123 104 96.2 8.4 1.1 -9.89 11.0 92 96 11.4 0.050
1987 114 93 97 72.5 80.4 124 112 101.9 7.6 0.4 -5.19 5.6 97 102 5.8 0.052
1988 116 99 96 73.5 81.8 120 108 100.5 7.3 -0.1 -1.07 1.0 96 100 1.0 0.052
1989 108 94 92 75.8 87.1 117 102 96.6 8.5 0.0 2.27 -2.3 92 96 -2.4 0.053
1990 128 114 101 78.4 86.3 119 108 101.3 9.8 0.1 -3.62 3.8 101 101 4.0 0.056
1991 125 110 96 80.9 85.4 122 115 105.0 9.3 0.5 -2.57 3.1 96 105 3.2 0.055
1992 133 120 101 82.9 83.8 121 120 107.0 9.0 1.3 -2.33 3.6 101 107 3.7 0.057
1993 123 114 100 85.1 81.7 121 126 111.2 7.9 1.9 -3.38 5.3 100 112 5.4 0.049
1994 128 120 104 87.1 83.0 113 118 108.3 6.7 0.3 0.19 0.1 104 109 0.1 0.049
1995 145 141 111 88.5 85.7 109 113 106.4 6.1 0.0 -0.08 0.1 111 107 0.1 0.051
1996 138 134 108 90.2 87.6 108 111 104.4 4.8 -0.4 1.52 -1.9 109 105 -2.0 0.049
1997 119 117 102 91.8 90.8 103 104 100.3 4.6 -0.2 2.64 -2.8 102 100 -3.0 0.047
1998 118 116 104 92.6 89.1 103 107 101.4 4.1 0.1 -1.45 1.5 104 101 1.6 0.049
1999 113 110 101 93.6 89.1 104 109 104.8 3.7 -0.4 -0.23 -0.2 101 105 -0.1 0.047
2000 98 97 95 96.1 98.5 99 96 98.5 4.8 0.0 5.88 -5.9 95 98 -6.2 0.045
2001 95 93 95 98.4 100.3 100 98 98.8 4.6 0.8 1.32 -0.5 95 99 -0.6 0.048
2002 100 100 100 100.0 100.0 100 100 100.0 4.1 0.6 0.17 0.5 100 100 0.4 0.052
2003 120 119 107 101.6 99.6 99 101 102.3 3.2 0.5 -1.14 1.6 108 102 1.7 0.054
2004 132 133 111 103.7 103.8 95 95 101.5 3.1 0.1 1.39 -1.3 112 102 -1.4 0.055
2005 132 136 112 106.6 110.6 92 89 100.1 2.7 -0.2 2.89 -3.1 113 100 -3.3 0.055
2006 133 140 115 108.6 117.3 93 86 101.3 3.3 -0.3 2.62 -2.9 115 101 -3.1 0.053
2007 145 152 116 110.4 121.1 92 84 102.6 4.1 0.2 0.48 -0.3 117 103 -0.4 0.056
2008 156 164 120 115.4 128.0 90 82 100.3 4.0 0.8 0.04 0.8 121 100 0.7 0.055
2009 147 156 117 115.4 122.7 101 95 104.1 2.2 -0.1 -0.27 0.2 119 104 0.2 0.056
2010 140 148 113 117.9 131.9 95 85 104.2 1.9 -0.1 3.80 -3.9 114 104 -4.1 0.053
Source: See Table 1A and 1B.
Real Exchange Rate Real Adjusted Unit Labor Cost Ratio Interest Rates Excluding Country Being Examined
123
(4) Canada
Trade
Year e pmfg*e rxr CPI PPI RULC RULCAdj RULCAdjratio Int.Rate intratediff gppiratio realintratediff rxr1 rulcadjratio1 realintratediff1 Weight
1960 162 37 148 15.3 17.9 185 157 149.6 4.5 -0.6 153 154 0.074
1961 155 35 137 15.5 18.0 178 153 143.3 4.4 -0.6 -1.69 1.1 140 147 1.2 0.072
1962 147 32 126 15.7 18.2 168 144 133.4 4.6 -0.3 0.29 -0.6 128 136 -0.7 0.069
1963 146 32 125 15.9 18.4 166 143 132.9 4.5 -0.4 0.06 -0.5 127 136 -0.5 0.067
1964 146 33 124 16.1 18.7 164 141 132.1 4.7 -0.6 -2.11 1.5 126 135 1.8 0.069
1965 146 33 124 16.5 18.9 163 142 132.9 4.9 -0.8 -1.74 0.9 126 136 1.1 0.070
1966 146 34 122 17.3 19.4 164 146 135.4 5.5 -0.8 -0.16 -0.6 124 139 -0.4 0.074
1967 146 35 123 17.8 19.9 167 149 135.7 5.6 -0.5 1.91 -2.4 125 139 -2.5 0.076
1968 146 35 125 18.5 20.2 162 148 134.4 6.7 0.2 -0.99 1.2 127 138 1.3 0.079
1969 146 37 126 19.3 20.9 159 147 134.5 7.7 0.4 -0.85 1.3 128 138 1.3 0.076
1970 150 38 124 20.0 21.5 165 154 136.9 7.1 -0.6 -3.43 2.8 126 140 3.0 0.072
1971 156 41 125 20.5 21.9 163 153 133.1 5.6 -1.2 -2.11 0.9 128 136 1.2 0.073
1972 159 43 120 21.6 22.9 158 149 128.9 6.3 -0.3 0.53 -0.9 121 132 -0.9 0.072
1973 157 46 111 23.2 25.4 152 139 124.7 7.0 -0.8 -0.13 -0.7 111 127 -0.8 0.065
1974 161 54 115 25.8 30.3 154 132 121.4 8.1 -1.3 0.23 -1.5 116 123 -1.5 0.063
1975 154 58 111 28.5 33.7 165 140 120.1 7.7 -0.6 4.17 -4.8 112 122 -5.1 0.063
1976 159 62 118 30.6 35.4 163 141 123.2 8.4 0.0 -2.83 2.8 119 125 3.0 0.063
1977 148 61 108 33.2 38.1 158 137 120.4 7.9 0.1 0.04 0.1 108 122 0.0 0.059
1978 138 60 91 36.1 41.7 151 131 113.3 9.0 1.2 3.65 -2.5 91 114 -2.9 0.056
1979 134 66 90 39.4 47.8 150 124 110.2 10.4 1.0 5.75 -4.8 89 111 -5.2 0.054
1980 134 72 90 43.4 54.1 152 122 109.6 12.4 1.1 0.96 0.2 90 110 0.0 0.052
1981 131 75 101 48.9 59.6 147 121 111.0 15.7 2.7 -0.32 3.1 101 112 3.0 0.058
1982 127 77 107 54.3 63.6 152 130 119.1 14.0 2.1 0.10 2.0 108 120 1.8 0.056
1983 127 84 119 57.4 65.8 145 126 119.1 10.6 0.5 -1.04 1.6 120 120 1.7 0.062
1984 121 82 119 59.9 68.8 133 116 111.4 11.9 1.7 -0.16 1.9 120 112 1.9 0.068
1985 115 81 119 62.3 70.7 132 116 112.5 10.4 1.5 0.20 1.3 120 113 1.2 0.067
1986 113 84 99 64.9 71.3 132 121 111.2 9.2 2.0 2.70 -0.7 99 112 -1.1 0.061
1987 118 92 96 67.9 73.3 129 119 108.8 9.4 2.2 2.52 -0.3 95 109 -0.6 0.056
1988 128 104 101 70.4 76.5 128 118 109.2 9.8 2.3 1.59 0.7 101 110 0.6 0.059
1989 133 109 107 74.0 78.0 125 119 112.3 10.2 1.7 -2.08 3.8 108 113 3.8 0.058
1990 135 110 98 78.1 78.2 124 123 115.5 11.2 1.5 -2.40 3.9 98 116 3.9 0.052
1991 137 114 99 82.6 77.4 124 132 120.4 9.2 0.4 -2.58 3.0 99 122 3.1 0.051
1992 130 106 89 83.8 77.8 122 132 117.8 7.4 -0.3 -0.01 -0.2 88 119 -0.2 0.051
1993 122 101 89 85.3 80.6 117 123 109.1 6.5 0.5 2.70 -2.2 88 110 -2.3 0.058
1994 115 102 88 85.4 85.5 113 113 103.5 7.8 1.4 4.57 -3.2 87 104 -3.5 0.057
1995 114 110 86 87.3 91.9 112 107 100.4 7.6 1.6 4.03 -2.4 85 100 -2.8 0.054
1996 115 110 89 88.7 92.2 113 108 102.0 6.2 1.0 -0.30 1.3 88 102 1.2 0.055
1997 113 107 93 90.0 93.0 108 105 100.6 5.3 0.6 -0.23 0.8 93 101 0.8 0.059
1998 106 100 90 91.4 93.3 107 105 99.7 5.2 1.1 0.79 0.3 89 100 0.2 0.060
1999 106 105 96 93.0 94.9 103 101 97.2 5.5 1.4 1.51 -0.1 95 97 -0.3 0.062
2000 106 105 103 95.5 99.0 98 94 96.6 6.0 1.2 -0.15 1.3 104 96 1.3 0.065
2001 101 101 104 97.8 100.0 101 98 98.8 4.9 1.0 0.53 0.5 104 99 0.4 0.063
2002 100 100 100 100.0 100.0 100 100 100.0 4.4 1.0 0.50 0.5 100 100 0.4 0.060
2003 112 112 100 102.9 98.8 101 105 106.3 3.9 1.2 -1.94 3.1 100 107 3.2 0.057
2004 121 122 102 104.8 102.0 102 104 111.9 3.7 0.7 0.44 0.3 102 113 0.2 0.054
2005 130 128 106 107.1 103.6 101 104 116.8 3.5 0.6 -1.92 2.5 106 118 2.6 0.057
2006 139 137 112 109.2 106.0 101 104 122.2 4.1 0.5 -0.96 1.5 113 124 1.5 0.055
2007 146 147 112 111.7 107.6 101 105 128.1 4.2 0.3 -1.31 1.6 113 130 1.7 0.053
2008 147 155 113 114.4 112.3 102 103 127.2 3.0 -0.3 -1.13 0.9 114 129 0.9 0.052
2009 138 140 105 115.1 108.4 104 110 121.2 2.2 -0.2 0.40 -0.6 106 122 -0.6 0.050
2010 153 155 118 117.3 109.4 96 103 126.4 2.2 0.2 -2.50 2.7 119 128 2.9 0.051
Source: See Table 1A and 1B.
Real Exchange Rate Real Adjusted Unit Labor Cost Ratio Interest Rates Excluding Country Being Examined
124
(5) Denmark
Trade
Year e pmfg*e rxr CPI PPI RULC RULCAdj RULCAdjratio Int.Rate intratediff gppiratio realintratediff rxr1 rulcadjratio1 realintratediff1 Weight
1960 114 16 63 9.8 16.7 148 87 82.3 6.0 0.9 63 82 0.021
1961 114 17 65 10.0 16.9 153 91 85.0 6.6 1.6 -0.82 2.4 65 85 2.4 0.020
1962 114 17 68 10.7 17.4 149 92 84.9 6.6 1.7 1.94 -0.3 67 85 -0.3 0.021
1963 114 18 68 11.2 17.9 150 94 87.4 6.4 1.6 1.64 -0.1 68 87 -0.2 0.020
1964 114 18 68 11.7 18.2 144 92 86.4 7.1 1.8 -1.21 3.0 67 86 3.0 0.021
1965 114 19 69 12.6 19.0 141 94 87.7 8.6 2.9 1.02 1.9 68 87 1.8 0.021
1966 114 19 70 13.4 19.4 143 99 92.0 8.7 2.4 -0.60 3.0 69 92 3.0 0.020
1967 113 20 70 14.2 19.6 138 100 90.6 9.1 3.0 0.58 2.4 70 90 2.3 0.020
1968 105 19 66 15.4 20.2 125 95 86.4 8.7 2.2 0.19 2.0 66 86 2.0 0.018
1969 105 19 67 15.9 20.9 128 98 89.3 9.7 2.4 -0.80 3.2 66 89 3.2 0.018
1970 105 20 66 16.9 22.5 129 96 85.8 10.9 3.2 1.49 1.7 65 86 1.6 0.018
1971 107 22 66 18.1 23.3 129 100 87.5 11.0 4.2 -0.82 5.0 66 87 4.9 0.017
1972 113 24 67 19.3 24.6 127 99 86.1 10.6 4.0 1.58 2.4 66 86 2.3 0.017
1973 131 32 76 21.0 28.3 133 99 88.8 11.8 4.0 3.33 0.7 76 89 0.6 0.018
1974 130 35 75 24.2 34.4 137 96 88.9 15.1 5.7 2.31 3.4 75 89 3.2 0.016
1975 138 43 81 26.5 36.3 135 99 84.8 13.0 4.6 -1.07 5.7 81 85 5.7 0.017
1976 131 44 83 28.9 39.2 136 100 87.9 15.0 6.7 -0.28 7.0 82 88 6.9 0.017
1977 131 48 84 32.1 42.1 132 100 87.7 15.7 8.0 -0.14 8.1 83 88 8.0 0.016
1978 143 56 85 35.4 44.0 130 104 90.6 16.5 8.7 -1.03 9.7 85 90 9.6 0.016
1979 150 61 83 38.7 48.3 125 101 89.6 16.9 7.5 1.36 6.1 83 89 6.0 0.016
1980 140 63 78 43.5 56.7 117 90 80.6 19.0 7.8 4.58 3.2 78 80 3.0 0.014
1981 111 53 72 48.7 65.6 113 84 77.2 19.6 6.7 4.66 2.0 71 77 1.9 0.014
1982 95 50 70 53.5 72.6 110 81 74.6 21.4 9.4 3.76 5.7 70 74 5.5 0.014
1983 86 48 68 57.3 76.2 104 78 73.8 15.3 5.3 0.46 4.8 68 73 4.7 0.014
1984 76 46 67 60.8 81.9 105 78 75.2 14.6 4.4 2.66 1.8 67 75 1.7 0.013
1985 74 47 70 63.7 79.1 106 85 82.8 12.2 3.3 -5.82 9.2 69 83 9.2 0.014
1986 97 65 77 66.0 76.6 109 94 86.7 10.5 3.3 -1.41 4.7 76 86 4.7 0.015
1987 115 82 84 68.7 77.2 115 103 93.6 12.5 5.3 0.58 4.7 84 93 4.7 0.015
1988 117 85 82 71.8 80.3 114 102 94.3 11.3 3.8 1.27 2.6 82 94 2.5 0.014
1989 108 80 78 75.2 84.8 109 96 91.0 10.2 1.7 1.32 0.3 78 91 0.3 0.014
1990 127 101 90 77.1 86.5 114 101 94.9 11.0 1.3 -0.74 2.0 90 95 2.0 0.015
1991 123 102 89 79.1 87.4 114 103 93.9 10.1 1.4 -0.47 1.8 89 94 1.8 0.015
1992 131 113 95 80.7 87.1 114 106 94.7 10.1 2.4 -0.90 3.3 95 95 3.3 0.015
1993 122 106 94 81.7 85.9 115 109 96.8 8.2 2.2 -2.17 4.4 93 97 4.4 0.014
1994 124 106 91 83.4 86.4 103 99 91.0 8.3 2.0 -0.91 2.9 91 91 2.9 0.014
1995 141 121 95 85.2 89.1 103 98 92.5 9.0 2.9 -0.21 3.1 95 92 3.1 0.015
1996 136 123 99 86.8 90.6 108 104 97.7 7.8 2.6 1.11 1.5 99 98 1.4 0.014
1997 119 108 94 88.8 92.1 101 97 93.6 7.1 2.4 0.58 1.8 94 94 1.8 0.013
1998 118 107 96 90.5 91.7 102 101 96.0 6.0 2.0 0.00 2.0 96 96 2.0 0.013
1999 113 104 95 92.7 92.0 101 102 97.8 6.1 2.0 0.08 1.9 95 98 1.9 0.013
2000 97 92 90 95.5 96.3 97 96 98.1 7.1 2.2 0.22 2.0 90 98 2.0 0.012
2001 95 90 93 97.7 99.1 99 97 97.9 6.4 2.6 2.40 0.2 93 98 0.1 0.012
2002 100 100 100 100.0 100.0 100 100 100.0 6.1 2.7 1.45 1.2 100 100 1.2 0.013
2003 120 120 108 102.2 100.0 101 103 104.2 5.2 2.4 -0.71 3.2 108 104 3.1 0.013
2004 132 130 109 103.3 101.0 98 100 107.3 5.0 2.0 -1.71 3.8 109 107 3.8 0.013
2005 132 132 109 105.1 104.2 99 100 112.6 4.6 1.6 -0.51 2.2 109 113 2.1 0.013
2006 133 135 110 107.1 107.5 96 96 112.3 4.6 1.1 -0.09 1.1 110 112 1.1 0.013
2007 145 148 113 109.0 111.1 98 96 116.8 5.1 1.2 0.57 0.7 113 117 0.7 0.013
2008 155 160 117 112.7 115.9 101 99 121.4 5.6 2.4 -1.24 3.6 117 122 3.6 0.013
2009 147 152 115 114.2 114.2 103 103 113.3 5.1 2.8 2.46 0.3 115 113 0.3 0.013
2010 140 140 106 116.8 118.3 95 94 115.5 4.2 2.2 0.10 2.1 106 116 2.1 0.012
Source: See Table 1A and 1B.
Real Exchange Rate Real Adjusted Unit Labor Cost Ratio Interest Rates Excluding Country Being Examined
125
(6) El salvador
e 1/e 1/e ES CPI US CPI rxr rxr ES RULC ES RULC US RULC Real ULC Ratio
Year (Colón/Dollar)(Dollar/Colón) 1970=100 1970=100 1970=100 1970=100 2002=100 1990=100 2002=100 2002=100 2002=100
1960 2.5 0.40 100 94 76 123 54 57 64 244 26
1961 2.5 0.40 100 91 77 118 52 61 68 243 28
1962 2.5 0.40 100 91 78 117 52 66 73 237 31
1963 2.5 0.40 100 93 79 118 52 65 72 225 32
1964 2.5 0.40 100 94 80 118 52 65 72 220 33
1965 2.5 0.40 100 95 81 117 51 69 77 213 36
1966 2.5 0.40 100 94 84 112 49 71 79 211 37
1967 2.5 0.40 100 95 86 111 49 73 81 216 38
1968 2.5 0.40 100 97 90 109 48 76 84 215 39
1969 2.5 0.40 100 97 94 103 45 83 92 216 43
1970 2.5 0.40 100 100 100 100 44 93 104 218 47
1971 2.5 0.40 100 100 104 96 42 106 118 208 57
1972 2.5 0.40 100 102 108 95 42 117 130 203 64
1973 2.5 0.40 100 109 114 95 42 125 139 197 71
1974 2.5 0.40 100 127 127 100 44 130 144 202 71
1975 2.5 0.40 100 151 139 109 48 112 125 200 62
1976 2.5 0.40 100 162 147 110 49 117 130 194 67
1977 2.5 0.40 100 181 156 116 51 133 147 191 77
1978 2.5 0.40 100 205 168 122 54 129 143 190 75
1979 2.5 0.40 100 235 187 126 55 145 161 188 86
1980 2.5 0.40 100 276 212 130 57 173 192 191 101
1981 2.5 0.40 100 316 234 135 59 171 190 181 105
1982 2.5 0.40 100 353 249 142 63 166 184 185 99
1983 2.5 0.40 100 400 257 156 69 155 173 171 101
1984 2.5 0.40 100 447 268 167 73 144 160 165 96
1985 2.5 0.40 100 546 277 197 87 126 140 162 86
1986 4.9 0.21 52 721 282 132 58 119 132 164 81
1987 5.0 0.20 50 900 293 154 68 130 145 153 94
1988 5.0 0.20 50 1078 305 177 78 116 129 147 88
1989 5.0 0.20 50 1268 319 199 87 107 119 145 82
1990 6.8 0.15 37 1572 336 171 75 100 111 143 78
1991 8.0 0.12 31 1798 351 160 70 97 108 142 76
1992 8.4 0.12 30 2000 361 166 73 99 110 140 78
1993 8.7 0.11 29 2370 372 183 81 108 120 136 89
1994 8.7 0.11 29 2621 382 197 87 108 120 131 92
1995 8.8 0.11 29 2884 392 210 92 108 120 125 95
1996 8.8 0.11 29 3166 404 224 98 101 112 120 93
1997 8.8 0.11 29 3309 413 229 101 97 108 116 93
1998 8.8 0.11 29 3393 420 231 102 97 108 115 93
1999 8.8 0.11 29 3410 429 227 100 105 116 111 105
2000 8.8 0.11 29 3488 443 225 99 103 115 107 107
2001 8.8 0.11 29 3619 456 227 100 95 106 106 100
2002 8.8 0.11 29 3686 463 227 100 90 100 100 100
2003 8.8 0.11 29 3764 474 227 100 88 98 97 100
2004 8.8 0.11 29 3932 486 231 102 83 93 88 105
2005 8.8 0.11 29 4116 503 234 103 78 87 84 104
2006 8.8 0.11 29 4283 519 236 104 75 84 80 104
2007 8.8 0.11 29 4479 534 240 105 75 84 76 109
2008 8.8 0.11 29 4779 554 246 108 72 80 78 102
2009 8.8 0.11 29 4829 553 250 110 71 78 79 100
2010 8.8 0.11 29 4886 561 249 109 72 80 70 113
Source: See Table 1A.
126
(7) Finland
Trade
Year e pmfg*e rxr CPI PPI RULC RULCAdj RULCAdjratio Int.Rate intratediff gppiratio realintratediff rxr1 rulcadjratio1 realintratediff1 Weight
1960 197 26 106 9.0 13.1 168 115 109.4 6.8 1.6 106 110 0.013
1961 197 28 109 9.4 13.2 166 118 111.0 6.8 1.7 -1.47 3.2 109 111 3.2 0.013
1962 197 28 108 9.6 13.4 174 126 116.1 7.0 2.1 0.72 1.4 108 116 1.3 0.013
1963 197 29 110 10.2 13.8 168 124 114.9 7.0 2.1 2.18 -0.1 111 115 -0.2 0.012
1964 197 30 113 11.3 14.9 163 123 115.1 7.0 1.7 4.45 -2.8 113 115 -2.8 0.013
1965 197 30 113 11.9 15.5 163 124 116.2 7.0 1.3 1.25 0.0 113 116 0.0 0.013
1966 197 31 112 12.1 15.8 165 126 116.9 7.0 0.7 -0.84 1.5 112 117 1.5 0.012
1967 182 30 106 12.8 16.3 164 129 117.2 7.0 0.9 2.32 -1.4 106 117 -1.5 0.011
1968 150 28 97 14.0 18.1 158 122 110.7 7.0 0.5 8.12 -7.6 97 111 -7.7 0.010
1969 150 31 105 14.5 18.7 154 119 108.7 7.0 -0.2 -0.98 0.8 105 109 0.8 0.011
1970 150 32 103 14.7 19.5 165 124 110.5 7.0 -0.7 -1.65 0.9 103 111 1.0 0.011
1971 150 33 102 15.7 20.5 176 135 117.2 8.5 1.7 0.88 0.8 102 117 0.8 0.011
1972 152 36 100 16.7 22.3 179 134 115.8 7.8 1.2 4.02 -2.9 100 116 -2.9 0.011
1973 165 46 110 18.4 26.2 185 130 117.0 9.3 1.5 5.88 -4.4 110 117 -4.5 0.010
1974 167 61 131 21.6 32.5 192 127 117.6 9.3 -0.2 4.57 -4.7 132 118 -4.8 0.011
1975 171 70 134 25.4 36.9 207 142 122.6 9.3 0.9 6.33 -5.4 135 123 -5.5 0.012
1976 163 75 142 29.1 40.0 210 152 133.3 9.3 0.9 -0.10 1.0 143 134 1.0 0.011
1977 156 78 137 32.5 44.0 201 148 129.8 8.3 0.5 2.38 -1.9 137 130 -2.0 0.011
1978 153 83 126 35.1 46.5 188 142 122.9 7.3 -0.6 0.15 -0.7 126 123 -0.7 0.010
1979 162 96 130 37.7 51.0 182 134 119.6 8.5 -0.9 1.20 -2.1 131 120 -2.1 0.011
1980 169 109 136 42.0 59.3 179 127 114.2 9.3 -2.0 3.77 -5.8 137 114 -5.8 0.012
1981 146 100 134 46.8 66.9 180 126 115.8 9.3 -3.7 2.08 -5.8 134 116 -5.7 0.011
1982 131 96 134 51.4 71.7 175 125 115.4 8.5 -3.4 0.48 -3.9 135 116 -3.9 0.011
1983 113 89 127 55.6 75.7 170 125 117.4 9.5 -0.6 1.02 -1.6 127 118 -1.6 0.011
1984 105 89 129 59.7 79.8 166 124 119.8 9.5 -0.7 0.60 -1.3 129 120 -1.3 0.010
1985 102 87 129 62.7 83.4 167 125 121.4 9.0 0.1 1.92 -1.8 129 122 -1.8 0.011
1986 124 107 127 64.5 79.1 166 136 125.3 7.0 -0.2 -3.47 3.3 127 126 3.3 0.011
1987 143 129 133 67.1 79.8 160 135 122.9 7.0 -0.2 0.64 -0.9 134 123 -0.9 0.012
1988 150 144 139 70.4 83.0 157 134 124.1 10.6 3.1 1.27 1.9 140 124 1.8 0.011
1989 147 148 145 75.2 87.1 155 134 126.7 12.1 3.6 0.77 2.8 146 127 2.8 0.011
1990 165 168 149 79.7 90.1 159 140 131.3 13.3 3.6 0.60 3.0 149 132 3.0 0.011
1991 156 151 133 83.2 90.3 164 151 137.1 11.7 3.0 -1.25 4.2 133 137 4.2 0.009
1992 140 139 117 85.6 91.3 147 138 123.2 12.0 4.3 0.55 3.7 117 123 3.7 0.009
1993 110 115 101 87.6 94.0 134 125 110.5 8.8 2.9 2.07 0.8 101 111 0.8 0.008
1994 120 124 107 88.5 95.3 128 119 108.5 9.0 2.6 -0.07 2.7 107 109 2.7 0.009
1995 144 160 125 89.2 104.9 132 113 106.0 8.8 2.7 6.52 -3.8 126 106 -3.9 0.010
1996 137 146 118 89.7 104.1 132 114 107.3 7.1 1.8 -1.39 3.2 118 107 3.2 0.010
1997 121 129 113 90.9 103.9 126 111 106.2 6.0 1.2 -1.20 2.4 113 106 2.4 0.010
1998 118 130 117 92.0 102.5 123 110 104.7 4.8 0.8 -0.89 1.6 117 105 1.6 0.011
1999 113 119 109 93.2 100.0 117 109 105.0 4.7 0.6 -2.68 3.3 109 105 3.3 0.010
2000 98 102 101 96.0 107.1 106 95 97.1 5.5 0.7 2.54 -1.9 101 97 -1.9 0.010
2001 95 99 102 98.4 103.9 106 101 101.1 5.0 1.2 -3.47 4.7 102 101 4.7 0.010
2002 100 100 100 100.0 100.0 100 100 100.0 5.0 1.5 -3.25 4.8 100 100 4.8 0.010
2003 120 114 102 100.9 97.1 96 100 100.9 4.1 1.4 -3.57 5.0 102 101 5.0 0.010
2004 132 122 102 101.1 97.3 93 97 103.8 4.1 1.1 -2.52 3.7 102 104 3.7 0.010
2005 132 119 98 101.7 99.5 92 94 106.1 3.4 0.4 -1.32 1.7 98 106 1.7 0.010
2006 133 116 95 103.4 102.7 84 85 99.5 3.8 0.2 -0.09 0.3 94 99 0.3 0.011
2007 145 125 96 105.9 105.4 77 78 94.6 4.3 0.4 -0.19 0.6 96 95 0.6 0.011
2008 156 128 94 110.1 109.2 80 81 99.1 4.3 1.1 -1.92 3.0 94 99 3.0 0.011
2009 147 113 85 110.3 100.9 98 107 117.7 3.7 1.4 -3.83 5.3 85 118 5.3 0.010
2010 140 112 85 111.5 105.6 87 92 113.7 3.0 1.0 1.05 0.0 85 114 0.0 0.009
Source: See Table 1A and 1B.
Real Exchange Rate Real Adjusted Unit Labor Cost Ratio Interest Rates Excluding Country Being Examined
127
(8) France
Trade
Year e pmfg*e rxr CPI PPI RULC RULCAdj RULCAdjratio Int.Rate intratediff gppiratio realintratediff rxr1 rulcadjratio1 realintratediff1 Weight
1960 142 28 112 12.0 20.0 145 87 82.9 6.2 1.1 113 81 0.082
1961 141 29 114 12.4 20.6 146 88 82.9 6.1 1.1 0.56 0.5 115 81 0.4 0.082
1962 142 29 114 13.1 20.7 144 91 84.6 6.0 1.1 -0.02 1.1 115 83 1.0 0.082
1963 142 29 114 13.9 21.3 138 90 83.8 5.9 1.1 1.51 -0.5 115 82 -0.7 0.086
1964 142 31 116 14.2 22.0 142 92 86.0 6.1 0.8 0.18 0.6 118 85 0.5 0.086
1965 142 31 115 14.4 22.2 141 91 85.3 6.3 0.6 -2.00 2.6 117 84 2.7 0.084
1966 141 31 111 14.8 22.9 137 88 81.9 6.5 0.2 0.02 0.1 112 80 0.1 0.084
1967 141 31 111 15.3 22.6 136 93 84.2 6.8 0.7 -1.95 2.6 112 83 2.7 0.084
1968 140 31 111 15.9 22.2 135 96 87.5 7.0 0.6 -4.12 4.7 112 86 5.0 0.083
1969 134 31 106 16.6 24.6 135 91 83.3 7.7 0.5 6.13 -5.6 107 82 -6.2 0.087
1970 126 30 98 17.4 26.5 134 88 78.7 8.5 0.8 1.35 -0.6 97 77 -0.8 0.087
1971 126 32 98 18.3 27.0 135 91 79.7 7.0 0.3 -1.95 2.2 98 78 2.4 0.088
1972 138 38 104 19.4 28.2 141 97 83.7 6.5 -0.1 0.27 -0.4 104 82 -0.4 0.094
1973 156 47 113 21.0 32.4 142 92 82.4 8.7 0.9 3.42 -2.5 114 81 -2.9 0.095
1974 144 49 104 23.7 41.9 137 77 71.5 11.8 2.4 8.67 -6.3 105 69 -7.4 0.091
1975 162 62 118 26.5 39.5 147 98 84.7 8.7 0.4 -11.81 12.2 120 83 13.3 0.095
1976 145 60 114 29.3 41.1 141 101 88.0 9.0 0.7 -3.97 4.7 116 87 5.0 0.095
1977 141 63 110 32.1 42.4 139 105 92.3 9.6 1.8 -4.08 5.9 112 92 6.1 0.094
1978 154 76 115 35.0 46.7 141 105 91.4 8.9 1.0 4.66 -3.6 116 91 -4.2 0.095
1979 163 87 118 38.7 53.0 138 101 90.0 9.7 0.2 4.50 -4.3 121 89 -4.8 0.097
1980 164 95 119 43.9 57.6 138 105 94.6 12.7 1.4 -3.02 4.4 121 94 4.6 0.099
1981 128 80 107 49.8 63.9 135 105 96.7 15.6 2.6 0.42 2.2 108 96 1.9 0.092
1982 106 72 101 55.7 71.0 136 106 97.9 15.2 3.3 4.10 -0.8 101 98 -1.5 0.092
1983 91 68 97 61.1 78.9 132 102 96.4 13.1 3.0 6.34 -3.3 96 96 -4.2 0.088
1984 79 63 92 65.6 89.3 131 97 93.0 12.2 2.0 8.08 -6.1 91 92 -6.9 0.081
1985 77 66 97 69.5 92.9 129 96 93.4 10.6 1.7 1.52 0.2 96 93 -0.1 0.082
1986 100 92 109 71.3 90.3 130 103 94.8 8.2 1.0 -1.05 2.0 110 94 1.9 0.088
1987 115 109 113 73.5 90.9 130 105 95.8 8.8 1.6 0.38 1.3 115 95 1.1 0.090
1988 116 113 110 75.5 95.6 126 100 92.5 8.5 1.1 2.40 -1.4 111 92 -1.7 0.089
1989 109 107 105 78.1 100.8 124 96 90.6 9.1 0.6 1.19 -0.6 106 90 -0.8 0.089
1990 127 129 114 80.8 99.6 123 99 93.0 10.1 0.4 -3.82 4.2 116 92 4.5 0.094
1991 123 126 111 83.5 98.3 122 104 94.6 9.4 0.7 -2.82 3.5 112 94 3.7 0.092
1992 131 137 115 85.6 96.7 122 108 96.4 9.5 1.8 -2.14 4.0 117 96 4.0 0.092
1993 122 128 112 87.5 94.0 121 112 99.4 7.6 1.7 -3.67 5.3 113 99 5.6 0.084
1994 125 128 110 89.2 95.0 115 108 99.1 6.5 0.1 -0.32 0.5 111 99 0.5 0.084
1995 139 144 113 90.7 100.8 113 101 95.4 7.1 1.0 2.68 -1.7 114 95 -2.0 0.086
1996 136 139 113 92.6 98.1 112 105 99.2 5.1 -0.2 -3.33 3.2 114 99 3.5 0.083
1997 119 122 107 93.7 97.5 109 105 100.6 4.5 -0.3 -1.62 1.3 107 101 1.5 0.080
1998 118 121 108 94.1 96.6 104 102 96.5 4.1 0.0 -0.47 0.5 109 96 0.5 0.084
1999 113 114 104 94.7 94.9 103 102 98.6 3.7 -0.4 -2.06 1.6 104 98 1.8 0.081
2000 98 99 97 96.5 99.0 100 98 99.8 4.8 0.0 -0.03 0.0 97 100 0.0 0.076
2001 95 95 98 97.9 100.2 100 98 98.2 4.6 0.8 0.68 0.1 98 98 0.0 0.078
2002 100 100 100 100.0 100.0 100 100 100.0 4.1 0.6 0.30 0.3 100 100 0.3 0.079
2003 120 116 104 102.1 100.9 97 98 99.5 3.2 0.5 0.18 0.3 104 100 0.3 0.081
2004 132 126 105 104.4 103.0 95 96 102.8 3.1 0.1 -0.72 0.8 106 103 0.9 0.080
2005 132 124 103 106.4 106.1 92 92 103.6 2.7 -0.2 -0.52 0.3 103 104 0.4 0.078
2006 133 126 103 108.0 109.2 91 90 105.2 3.3 -0.2 -0.40 0.2 103 106 0.2 0.077
2007 145 139 106 109.8 111.7 88 87 106.0 4.1 0.2 -0.47 0.7 107 106 0.7 0.078
2008 156 151 111 112.9 117.1 91 88 108.6 3.9 0.7 -0.80 1.5 112 109 1.5 0.078
2009 147 142 107 113.2 110.5 95 98 107.6 2.1 -0.2 -1.80 1.6 108 108 1.8 0.080
2010 140 135 103 115.2 112.9 90 92 113.0 1.7 -0.2 -1.28 1.1 103 114 1.1 0.074
Source: See Table 1A and 1B.
Real Exchange Rate Real Adjusted Unit Labor Cost Ratio Interest Rates Excluding Country Being Examined
128
(9) Germany
Trade
Year e pmfg*e rxr CPI PPI RULC RULCAdj RULCAdjratio Int.Rate intratediff gppiratio realintratediff rxr1 rulcadjratio1 realintratediff1 Weight
1960 50 16 64 27.7 39.0 95 68 64.3 6.4 1.3 60 60 0.135
1961 52 17 68 28.8 39.5 98 71 66.9 5.9 0.9 -0.79 1.7 64 63 1.2 0.140
1962 52 18 69 29.4 40.0 101 75 69.0 5.9 1.0 0.19 0.8 65 65 0.4 0.142
1963 52 18 70 30.2 40.1 101 76 70.5 6.1 1.2 -0.66 1.8 66 67 2.0 0.140
1964 52 19 70 30.7 40.6 99 75 70.1 6.2 0.9 -1.99 2.9 66 66 3.2 0.138
1965 52 19 71 31.9 41.5 99 76 70.9 7.1 1.4 -0.63 2.0 67 67 1.3 0.145
1966 52 20 71 33.1 42.3 99 78 72.0 8.1 1.8 -1.07 2.9 67 68 3.1 0.141
1967 52 20 71 33.7 42.0 97 78 70.8 7.0 0.9 -1.52 2.4 67 67 2.9 0.137
1968 52 20 71 34.3 41.5 95 78 70.9 6.5 0.0 -3.48 3.5 67 67 3.9 0.140
1969 53 21 72 34.9 42.4 95 78 71.5 6.8 -0.4 -2.36 1.9 68 68 2.1 0.144
1970 57 24 79 35.9 44.5 105 85 75.4 8.3 0.5 -1.12 1.7 76 72 1.4 0.149
1971 60 27 82 38.0 46.3 107 88 76.4 8.0 1.2 0.00 1.2 80 73 0.8 0.153
1972 65 31 84 40.0 47.6 106 89 77.1 7.9 1.3 -1.42 2.7 82 74 2.8 0.153
1973 78 39 93 42.8 50.1 105 90 80.5 9.3 1.5 -5.17 6.7 92 77 7.1 0.157
1974 80 43 92 45.7 57.6 108 86 79.2 10.4 1.0 -3.29 4.2 91 76 5.2 0.147
1975 84 47 90 48.5 60.2 108 87 74.8 6.9 -1.4 -2.21 0.8 88 71 1.4 0.147
1976 82 47 89 50.5 62.4 104 84 73.6 6.5 -1.9 -4.21 2.3 87 70 3.3 0.149
1977 89 53 93 52.4 64.2 106 87 76.0 5.3 -2.5 -4.42 1.9 92 72 3.0 0.152
1978 103 64 96 54.0 64.9 107 89 77.3 4.8 -3.1 -4.01 0.9 96 74 2.1 0.156
1979 113 71 97 56.0 68.0 106 88 78.0 6.4 -3.0 -3.40 0.4 96 74 1.6 0.155
1980 114 75 94 59.2 73.1 111 90 81.0 8.2 -3.1 -4.17 1.1 93 78 2.5 0.151
1981 91 62 83 62.7 78.8 109 87 80.0 10.2 -2.7 -2.48 -0.2 81 77 0.9 0.138
1982 85 61 85 66.1 83.4 110 87 80.0 8.6 -3.3 -0.69 -2.6 83 77 -2.0 0.143
1983 81 60 85 68.1 84.7 106 85 80.5 6.8 -3.3 -2.87 -0.4 83 78 0.7 0.141
1984 73 54 79 69.9 87.1 105 84 81.0 6.8 -3.4 -1.79 -1.6 76 78 -0.6 0.131
1985 70 54 80 71.3 89.3 104 83 80.8 6.0 -2.9 -0.01 -2.9 77 78 -2.4 0.134
1986 95 77 91 71.4 87.0 108 88 81.4 4.9 -2.3 -0.74 -1.6 90 79 -1.3 0.150
1987 115 96 99 71.4 84.9 114 96 87.7 4.7 -2.5 -2.69 0.2 99 86 1.1 0.154
1988 118 99 96 72.2 85.9 113 95 88.6 5.1 -2.4 -1.55 -0.8 96 87 -0.1 0.147
1989 110 94 92 74.3 88.6 112 94 88.8 6.6 -1.9 -0.94 -1.0 91 87 -0.5 0.145
1990 128 112 100 76.3 90.2 112 95 88.7 8.4 -1.3 -0.99 -0.3 100 87 0.1 0.157
1991 125 113 99 79.2 92.3 112 96 87.1 8.4 -0.4 0.81 -1.2 98 85 -1.3 0.162
1992 133 124 104 83.2 93.6 115 102 91.1 8.1 0.4 0.89 -0.5 104 89 -0.7 0.160
1993 125 119 104 86.8 93.8 115 107 94.4 6.4 0.4 -0.71 1.1 105 93 1.3 0.148
1994 128 120 103 89.3 94.3 108 103 93.9 6.0 -0.4 -0.80 0.4 104 93 0.6 0.146
1995 145 140 110 90.7 95.0 110 105 99.1 5.6 -0.4 -2.53 2.1 112 99 2.6 0.149
1996 138 136 110 92.1 95.1 111 108 101.5 4.8 -0.5 -0.53 0.0 111 102 0.2 0.143
1997 119 116 102 93.8 96.1 105 103 98.7 4.5 -0.3 -0.05 -0.2 102 98 -0.2 0.136
1998 118 118 106 94.8 96.0 105 104 99.0 4.0 0.0 0.35 -0.4 107 99 -0.4 0.144
1999 113 112 102 95.3 95.7 105 105 100.8 3.7 -0.4 -0.59 0.2 103 101 0.3 0.138
2000 98 96 94 96.7 98.8 101 99 101.3 4.8 0.0 -1.13 1.1 94 101 1.3 0.130
2001 95 93 96 98.6 99.8 100 98 98.8 4.2 0.4 0.53 -0.1 95 99 -0.3 0.137
2002 100 100 100 100.0 100.0 100 100 100.0 3.9 0.4 0.71 -0.3 100 100 -0.4 0.139
2003 120 120 108 101.1 100.2 97 98 99.0 3.0 0.3 -0.50 0.8 109 99 0.9 0.148
2004 132 132 110 102.7 101.4 92 93 99.9 3.0 0.1 -1.52 1.6 112 100 1.8 0.150
2005 132 131 108 104.3 103.2 88 88 99.4 2.7 -0.2 -1.82 1.6 110 99 1.9 0.147
2006 133 133 109 106.0 105.2 81 82 96.4 3.4 -0.1 -1.35 1.2 110 96 1.5 0.151
2007 145 148 113 108.4 107.0 77 78 95.6 4.0 0.1 -1.01 1.1 116 95 1.3 0.157
2008 156 159 116 111.1 109.7 81 82 100.5 4.0 0.8 -2.95 3.7 120 101 4.1 0.156
2009 147 149 112 111.6 106.7 98 102 112.4 3.2 0.9 1.21 -0.3 115 115 -0.6 0.158
2010 140 145 110 112.8 109.7 88 91 111.8 2.7 0.8 -0.72 1.5 112 114 1.4 0.152
Source: See Table 1A and 1B.
Real Exchange Rate Real Adjusted Unit Labor Cost Ratio Interest Rates Excluding Country Being Examined
129
(10) Italy
Trade
Year e pmfg*e rxr CPI PPI RULC RULCAdj RULCAdjratio Int.Rate intratediff gppiratio realintratediff rxr1 rulcadjratio1 realintratediff1 Weight
1960 330 25 101 4.7 6.6 151 106 101.0 5.0 -0.1 101 101 0.052
1961 330 26 101 5.1 6.7 142 108 101.7 5.2 0.2 -1.73 1.9 101 102 2.4 0.056
1962 330 26 102 5.2 6.8 148 112 103.8 5.8 0.9 1.80 -0.9 102 104 -0.6 0.059
1963 329 28 109 5.7 7.2 156 123 114.4 6.1 1.2 3.88 -2.7 110 115 -2.4 0.064
1964 328 30 112 6.0 7.4 157 126 117.8 7.4 2.1 0.14 2.0 113 119 1.2 0.059
1965 328 30 110 6.1 7.6 147 119 111.3 6.9 1.2 -1.20 2.4 110 112 2.5 0.060
1966 328 30 108 6.5 7.7 136 114 105.4 6.5 0.2 -0.44 0.7 109 106 0.9 0.061
1967 328 30 107 6.6 7.6 140 121 110.4 6.6 0.5 -2.33 2.8 107 111 3.4 0.065
1968 329 30 107 6.6 7.6 139 120 109.1 6.7 0.2 -2.51 2.8 107 110 2.7 0.063
1969 326 31 108 6.8 7.9 142 122 111.9 6.9 -0.4 -0.40 0.0 108 113 0.3 0.065
1970 327 34 111 7.2 8.5 155 131 116.2 9.0 1.3 0.87 0.4 112 117 0.3 0.066
1971 331 37 112 7.5 8.9 167 141 123.2 8.3 1.6 0.24 1.3 113 125 1.2 0.065
1972 351 40 109 7.9 9.2 161 139 120.4 7.5 0.9 -0.66 1.5 110 122 1.7 0.067
1973 352 45 107 8.8 10.8 155 127 114.0 7.4 -0.4 5.91 -6.3 107 115 -7.0 0.064
1974 315 49 106 10.5 15.1 156 108 99.6 9.9 0.4 17.93 -17.5 106 100 -18.6 0.066
1975 314 55 104 12.3 16.4 166 125 107.4 11.5 3.2 1.64 1.6 105 108 1.2 0.066
1976 247 50 95 14.3 20.2 155 110 96.1 13.1 4.7 13.59 -8.9 95 96 -10.3 0.064
1977 232 54 95 16.8 24.3 152 105 91.7 14.6 6.8 11.79 -5.0 95 91 -6.1 0.065
1978 241 62 93 18.8 25.7 148 108 93.8 13.7 5.9 0.45 5.4 93 93 5.2 0.066
1979 247 71 97 21.6 29.7 142 104 92.4 14.1 4.6 6.44 -1.8 96 92 -2.3 0.070
1980 239 80 101 26.1 35.6 131 96 86.3 15.3 4.0 7.03 -3.0 101 85 -3.8 0.071
1981 180 71 95 30.8 41.9 132 97 89.1 19.4 6.4 6.39 0.0 95 88 -0.9 0.070
1982 151 69 96 35.9 47.4 131 100 91.6 20.2 8.3 6.06 2.2 95 91 1.3 0.069
1983 135 68 96 41.2 52.6 129 101 95.3 18.3 8.2 6.21 2.0 96 95 1.0 0.067
1984 117 65 94 45.6 58.0 124 98 94.0 15.6 5.4 5.32 0.1 93 94 -0.7 0.064
1985 107 64 94 49.8 62.5 122 97 94.1 13.7 4.9 5.07 -0.2 93 94 -0.8 0.064
1986 137 86 102 52.8 62.6 120 101 93.4 11.5 4.3 1.99 2.3 102 93 2.0 0.068
1987 158 103 107 55.3 64.5 119 102 92.9 10.6 3.4 2.75 0.7 107 92 0.3 0.071
1988 157 105 102 58.0 66.7 117 101 94.1 10.9 3.5 0.78 2.7 102 94 2.4 0.068
1989 149 106 103 61.7 70.7 117 102 96.3 12.8 4.3 1.60 2.7 104 96 2.4 0.070
1990 171 125 110 65.6 73.6 118 105 98.7 13.5 3.9 1.38 2.5 111 99 2.2 0.073
1991 165 125 110 69.8 76.1 121 111 100.8 13.3 4.6 1.76 2.8 111 101 2.4 0.072
1992 166 129 109 73.5 77.5 117 111 99.1 13.3 5.6 1.34 4.2 109 99 3.8 0.071
1993 130 106 93 76.8 80.4 116 111 97.9 11.2 5.3 2.87 2.4 93 98 1.8 0.064
1994 127 104 89 80.0 83.4 107 103 94.3 10.5 4.1 2.27 1.9 89 94 1.5 0.065
1995 126 108 85 84.1 89.9 103 96 90.5 12.2 6.1 4.37 1.8 84 90 1.2 0.066
1996 133 119 96 87.5 91.6 104 99 93.6 9.4 4.1 1.24 2.9 96 93 2.6 0.067
1997 120 110 96 89.2 92.8 105 101 97.2 6.9 2.1 0.27 1.8 96 97 1.7 0.064
1998 118 111 99 91.0 92.9 103 101 96.0 4.9 0.8 0.56 0.3 99 96 0.2 0.066
1999 113 106 97 92.6 92.7 103 103 98.9 4.7 0.6 -0.51 1.1 97 99 1.1 0.062
2000 98 93 91 95.0 98.4 98 95 97.1 5.6 0.8 1.67 -0.9 91 97 -1.1 0.059
2001 95 93 95 97.5 99.6 99 97 97.3 5.2 1.3 0.67 0.7 95 97 0.6 0.062
2002 100 100 100 100.0 100.0 100 100 100.0 5.0 1.6 0.93 0.7 100 100 0.5 0.063
2003 120 122 109 102.7 100.7 103 105 106.4 4.2 1.5 0.03 1.5 110 107 1.4 0.065
2004 132 137 114 105.0 103.3 102 104 111.3 4.3 1.3 -0.26 1.6 115 112 1.5 0.065
2005 132 138 114 106.9 106.3 101 101 113.8 3.6 0.6 -0.71 1.3 115 115 1.3 0.063
2006 133 140 115 109.3 110.0 98 97 114.3 4.0 0.5 0.20 0.3 116 115 0.2 0.063
2007 145 159 122 111.3 113.4 97 96 116.6 4.5 0.6 0.27 0.3 123 118 0.3 0.066
2008 156 176 129 115.0 118.4 100 98 120.0 4.7 1.5 -1.15 2.6 131 122 2.6 0.065
2009 147 173 130 115.8 112.8 110 113 124.3 4.3 2.0 -0.90 2.9 133 126 2.9 0.064
2010 140 164 124 117.7 116.5 101 102 126.1 4.0 2.1 -0.24 2.3 126 128 2.2 0.061
Source: See Table 1A and 1B.
Real Exchange Rate Real Adjusted Unit Labor Cost Ratio Interest Rates Excluding Country Being Examined
130
(11) Japan
Trade
Year e pmfg*e rxr CPI PPI RULC RULCAdj RULCAdjratio Int.Rate intratediff gppiratio realintratediff rxr1 rulcadjratio1 realintratediff1 Weight
1960 35 16 65 18.5 51.8 193 69 65.6 6.1 1.0 64 64 0.053
1961 35 17 67 19.4 52.4 189 70 65.7 6.1 1.1 -1.10 2.2 65 64 1.6 0.059
1962 35 17 67 20.6 51.6 192 77 71.1 6.1 1.2 -2.48 3.6 65 70 3.8 0.058
1963 35 18 70 22.4 52.4 183 78 72.5 5.7 0.8 0.45 0.4 68 71 0.0 0.062
1964 35 18 67 23.2 52.4 174 77 72.4 5.7 0.4 -3.11 3.5 65 71 3.3 0.065
1965 35 18 67 24.7 53.0 176 82 77.0 5.7 0.0 -1.78 1.8 65 76 1.6 0.068
1966 35 19 67 25.9 54.3 169 81 74.9 6.3 0.0 -0.32 0.3 65 73 0.0 0.071
1967 35 19 68 26.9 55.7 162 78 71.1 6.3 0.2 1.75 -1.5 66 69 -2.3 0.077
1968 35 20 69 28.3 56.2 160 81 73.3 6.4 -0.1 -1.63 1.6 67 71 1.4 0.080
1969 35 21 72 30.1 57.3 156 82 74.9 6.5 -0.7 -2.48 1.8 70 73 1.8 0.083
1970 35 22 72 32.3 59.2 154 84 74.7 6.5 -1.2 -2.57 1.3 69 73 1.2 0.089
1971 36 23 71 34.4 58.7 158 92 80.4 6.2 -0.6 -4.73 4.2 69 79 4.4 0.091
1972 41 28 76 35.9 59.7 158 95 82.5 5.4 -1.2 -2.48 1.3 74 81 1.5 0.093
1973 46 34 81 40.2 69.1 159 92 82.8 6.5 -1.3 4.25 -5.5 80 81 -6.2 0.096
1974 43 37 79 49.6 88.1 164 93 85.4 8.0 -1.4 7.23 -8.6 77 84 -10.6 0.108
1975 42 38 72 55.4 90.5 167 102 87.8 7.4 -0.9 -3.76 2.9 70 86 4.1 0.101
1976 42 40 76 60.6 95.5 155 98 85.9 7.2 -1.2 -2.62 1.5 73 84 1.7 0.104
1977 47 46 82 65.6 98.7 152 101 88.3 5.7 -2.0 -3.91 1.9 80 87 2.4 0.106
1978 60 63 96 68.5 98.2 149 104 90.0 4.7 -3.1 -5.58 2.5 95 89 3.6 0.105
1979 57 60 81 70.9 103.1 141 97 86.6 6.7 -2.8 -3.18 0.4 79 85 1.5 0.099
1980 55 60 75 76.4 118.5 137 89 79.8 7.6 -3.7 2.48 -6.1 73 78 -6.6 0.107
1981 57 64 85 80.0 120.1 136 91 83.4 7.0 -5.9 -8.26 2.4 84 81 3.4 0.120
1982 50 57 80 82.4 120.7 134 91 84.1 6.7 -5.2 -5.84 0.7 78 82 2.4 0.116
1983 53 61 87 83.9 119.9 135 94 88.8 6.2 -3.9 -4.89 1.0 85 87 2.0 0.120
1984 53 63 92 85.8 120.0 133 95 91.6 5.9 -4.3 -4.43 0.1 91 90 1.2 0.123
1985 53 61 90 87.6 119.1 123 91 87.9 4.6 -4.2 -3.21 -1.0 88 86 0.2 0.121
1986 74 88 104 88.0 113.5 127 99 91.1 3.9 -3.3 -2.97 -0.3 104 90 0.7 0.117
1987 87 101 105 88.2 110.0 126 101 91.9 3.3 -3.9 -3.35 -0.6 106 91 0.7 0.112
1988 98 113 110 88.7 109.4 122 99 92.1 3.3 -4.1 -3.20 -0.9 111 91 0.0 0.116
1989 91 106 104 90.9 111.4 120 98 92.6 4.4 -4.1 -2.22 -1.9 104 92 -1.0 0.115
1990 86 101 90 93.6 113.1 116 96 89.9 6.5 -3.2 -1.20 -2.0 88 89 -1.2 0.109
1991 93 110 96 96.7 114.3 114 96 87.6 5.7 -3.0 -0.50 -2.5 96 86 -2.2 0.113
1992 99 118 99 98.4 113.3 115 100 89.0 4.0 -3.6 -1.45 -2.2 98 88 -1.4 0.111
1993 113 132 116 99.5 111.5 114 102 90.3 2.7 -3.3 -2.40 -0.9 118 89 -0.3 0.122
1994 123 140 120 100.2 109.7 114 104 95.0 2.7 -3.7 -3.01 -0.7 123 94 0.4 0.120
1995 133 148 117 100.1 108.7 110 102 95.7 1.5 -4.6 -4.06 -0.5 119 95 1.0 0.117
1996 115 126 102 100.1 106.9 106 100 93.8 1.3 -4.0 -2.28 -1.7 102 93 -0.5 0.111
1997 103 112 98 102.1 107.6 104 99 94.9 1.0 -3.7 -0.36 -3.4 98 94 -2.6 0.108
1998 96 105 94 102.7 106.0 105 102 96.8 0.6 -3.4 -1.06 -2.3 93 96 -1.4 0.095
1999 110 117 107 102.3 104.5 103 101 96.8 1.1 -3.0 -1.73 -1.3 108 96 -0.7 0.100
2000 116 119 117 101.6 104.5 98 95 97.0 1.2 -3.6 -4.18 0.6 120 97 1.7 0.106
2001 103 104 107 100.9 102.1 102 100 100.9 0.7 -3.2 -2.78 -0.4 108 101 0.6 0.097
2002 100 100 100 100.0 100.0 100 100 100.0 0.6 -2.8 -1.59 -1.2 100 100 -0.5 0.095
2003 108 104 93 99.7 99.2 93 94 94.9 0.5 -2.2 -1.53 -0.7 93 94 -0.1 0.094
2004 116 108 90 99.8 100.4 87 86 92.5 0.8 -2.2 -1.49 -0.7 89 92 -0.2 0.094
2005 114 103 85 99.4 102.1 81 79 88.2 0.7 -2.3 -1.89 -0.4 84 87 0.2 0.093
2006 108 93 76 99.7 104.3 77 74 86.8 1.1 -2.5 -1.08 -1.4 74 86 -1.0 0.091
2007 106 90 69 99.8 106.1 73 68 83.3 1.1 -2.8 -1.04 -1.7 66 82 -1.3 0.088
2008 121 97 71 101.1 111.0 71 65 79.9 0.9 -2.3 -0.99 -1.3 68 78 -0.8 0.091
2009 134 110 83 99.7 105.1 77 73 80.8 0.7 -1.6 -1.42 -0.2 81 79 0.2 0.087
2010 143 111 84 99.0 105.1 67 63 77.8 0.6 -1.3 -3.43 2.1 83 76 2.7 0.096
Source: See Table 1A and 1B.
Real Exchange Rate Real Adjusted Unit Labor Cost Ratio Interest Rates Excluding Country Being Examined
131
(12) Korea
Trade
Year e pmfg*e rxr CPI PPI RULC RULCAdj RULCAdjratio Int.Rate intratediff gppiratio realintratediff rxr1 rulcadjratio1 realintratediff1 Weight
1960 1924 51 204 6.9 2.6 113 295 280.6 10.2 5.1 204 281 0.002
1961 962 29 116 6.9 3.0 113 257 241.1 10.2 5.2 12.37 -7.2 116 242 -7.3 0.002
1962 962 33 129 6.9 3.4 113 228 210.6 10.2 5.3 11.80 -6.5 129 211 -6.4 0.003
1963 962 41 157 6.9 4.2 113 185 171.7 10.2 5.3 21.75 -16.4 157 172 -16.4 0.003
1964 585 35 133 6.9 6.0 113 130 121.5 10.5 5.2 38.02 -32.8 133 122 -33.2 0.002
1965 469 30 113 6.9 6.5 113 120 112.4 28.0 22.3 4.92 17.3 113 112 17.3 0.003
1966 461 34 122 6.9 7.4 113 106 98.3 28.0 21.7 10.11 11.6 123 98 11.7 0.004
1967 462 35 123 6.9 7.5 113 104 94.4 28.0 21.9 1.27 20.6 123 94 20.7 0.005
1968 452 37 129 6.9 8.1 113 96 87.6 23.0 16.5 5.08 11.5 129 87 11.5 0.006
1969 434 38 130 6.9 8.8 113 89 81.4 22.0 14.8 3.70 11.1 130 81 11.1 0.007
1970 403 39 126 6.9 9.7 113 80 71.4 19.0 11.3 4.57 6.7 126 71 6.6 0.007
1971 360 37 112 7.8 10.5 106 79 68.6 16.0 9.2 4.05 5.2 112 68 5.2 0.007
1972 318 39 108 8.7 12.0 107 77 67.0 11.0 4.4 9.51 -5.1 108 67 -5.2 0.007
1973 314 44 105 9.0 12.9 119 84 75.3 11.0 3.2 -3.62 6.8 105 75 7.0 0.010
1974 309 50 107 11.2 18.3 117 72 66.7 21.0 11.5 19.45 -7.9 107 66 -8.1 0.010
1975 258 48 92 14.0 23.1 106 64 55.5 21.1 12.8 18.40 -5.6 92 55 -5.9 0.011
1976 258 58 110 16.2 25.9 115 72 63.2 21.6 13.3 3.50 9.8 110 63 9.7 0.013
1977 258 64 112 17.9 28.3 123 77 67.9 21.5 13.7 1.42 12.3 112 67 12.2 0.014
1978 258 70 106 20.4 31.6 126 81 70.4 21.6 13.8 5.94 7.8 106 70 7.7 0.016
1979 258 84 114 24.1 37.5 130 84 74.9 25.2 15.7 9.45 6.3 114 75 6.0 0.017
1980 206 83 105 31.1 52.1 130 78 69.9 28.8 17.5 23.93 -6.4 105 70 -7.0 0.016
1981 184 86 116 37.7 62.7 123 74 67.8 23.6 10.7 8.94 1.7 116 67 1.5 0.019
1982 171 86 120 40.4 65.6 121 75 68.5 17.4 5.5 -1.87 7.4 121 68 7.3 0.020
1983 161 86 122 41.8 65.7 124 79 74.4 13.1 3.0 -4.15 7.1 122 74 7.2 0.022
1984 155 85 123 42.8 66.2 124 80 77.4 14.3 4.1 -3.81 7.9 124 77 8.0 0.024
1985 145 82 121 43.8 66.8 121 79 77.0 13.6 4.7 -1.57 6.3 122 77 6.2 0.024
1986 141 82 97 45.0 65.8 118 80 74.1 11.6 4.3 0.30 4.1 97 74 3.9 0.023
1987 151 91 94 46.4 66.1 118 83 75.4 12.4 5.2 0.21 5.0 94 75 5.0 0.026
1988 170 110 107 49.7 67.9 128 94 86.9 13.0 5.6 0.00 5.6 107 87 5.5 0.029
1989 185 124 122 52.5 68.9 129 98 92.9 14.7 6.2 -2.63 8.9 122 93 8.8 0.029
1990 176 121 107 57.0 71.8 127 101 94.3 15.0 5.3 1.41 3.9 107 94 3.7 0.028
1991 170 130 114 62.3 75.2 129 107 97.5 16.5 7.7 3.11 4.6 114 97 4.4 0.031
1992 159 128 107 66.2 76.8 131 113 100.6 15.1 7.4 1.60 5.8 108 101 5.6 0.031
1993 155 135 119 69.3 78.0 134 119 105.4 12.1 6.1 0.64 5.5 120 106 5.4 0.034
1994 155 144 124 73.8 80.1 134 124 113.1 12.3 5.9 1.27 4.6 125 114 4.5 0.035
1995 162 162 127 77.1 83.9 142 130 122.7 12.4 6.3 1.32 5.0 128 124 4.8 0.039
1996 155 157 127 80.8 86.6 142 133 125.2 10.9 5.6 2.58 3.1 128 126 2.8 0.041
1997 132 136 119 84.5 89.9 131 123 118.3 11.7 6.9 2.83 4.1 120 119 3.8 0.040
1998 89 103 93 90.7 100.9 119 107 101.6 12.8 8.8 12.70 -3.9 93 102 -4.6 0.032
1999 105 110 100 91.5 98.8 105 97 93.8 8.7 4.6 -2.34 6.9 100 94 6.9 0.036
2000 111 112 111 93.6 100.8 100 93 95.2 8.5 3.7 -2.30 6.0 111 95 6.0 0.041
2001 97 97 100 97.4 100.3 101 98 98.9 6.7 2.8 -0.95 3.8 100 99 3.7 0.038
2002 100 100 100 100.0 100.0 100 100 100.0 6.5 3.0 0.17 2.9 100 100 2.8 0.040
2003 105 105 94 103.5 102.2 95 97 97.8 4.9 2.2 1.49 0.7 94 98 0.6 0.041
2004 109 116 97 107.3 108.4 96 95 101.5 4.5 1.5 3.18 -1.7 97 102 -1.9 0.044
2005 122 127 105 110.3 110.7 97 97 108.5 4.7 1.7 -1.42 3.1 105 109 3.1 0.046
2006 131 130 107 112.7 111.7 93 94 110.6 5.1 1.5 -2.31 3.8 107 111 3.9 0.047
2007 135 135 103 115.5 113.3 91 92 112.6 5.4 1.5 -1.38 2.9 103 113 2.9 0.048
2008 114 119 87 120.9 123.0 87 85 104.7 5.8 2.6 2.82 -0.3 86 105 -0.5 0.051
2009 98 108 82 124.4 122.8 88 89 97.8 5.1 2.8 3.86 -1.1 81 98 -1.4 0.053
2010 108 125 95 127.9 122.8 85 88 108.6 4.6 2.6 -3.43 6.0 95 109 6.1 0.059
Source: See Table 1A and 1B.
Real Exchange Rate Real Adjusted Unit Labor Cost Ratio Interest Rates Excluding Country Being Examined
132
(13) Mexico
e 1/e 1/e MEX CPI US CPI rxr rxr Mex US Real ULC Ratio Real ULC Ratio
Year (Peso/Dollar) (Dollar/Peso) 1970=100 1970=100 1970=100 1970=100 1988=100 Real ULC Real ULC Mex/US 1988=100
1970 0.01 80 100 100 100 100 121 631 17285 0.037 50
1971 0.01 80 100 105 104 101 122 653 15692 0.042 57
1972 0.01 80 100 111 108 103 124 671 15786 0.042 58
1973 0.01 80 100 124 114 108 131 677 16429 0.041 56
1974 0.01 80 100 154 127 121 146 711 16549 0.043 59
1975 0.01 80 100 177 139 127 154 731 14171 0.052 71
1976 0.02 65 81.2 204 147 113 137 793 14295 0.055 76
1977 0.02 44 55.4 264 156 94 113 759 14426 0.053 72
1978 0.02 44 55.1 310 168 102 123 744 14918 0.050 68
1979 0.02 44 54.9 366 187 107 130 781 14912 0.052 72
1980 0.02 44 54.5 463 212 119 144 798 14398 0.055 76
1981 0.02 41 51.1 592 234 129 156 845 13630 0.062 85
1982 0.05 18 23.0 941 249 87 105 808 12511 0.065 89
1983 0.12 8 10.4 1900 257 77 93 586 11570 0.051 69
1984 0.17 6 7.5 3143 268 87 106 546 11896 0.046 63
1985 0.26 4 4.9 4958 277 87 105 550 11858 0.046 64
1986 0.61 2 2.1 9233 282 67 81 515 11348 0.045 62
1987 1.37 1 0.91 21404 293 67 81 506 12190 0.041 57
1988 2.27 0.44 0.55 45840 305 83 100 867 11884 0.073 100
1989 2.46 0.41 0.51 55012 320 87 106 950 11450 0.083 114
1990 2.81 0.36 0.44 69673 337 92 111 946 10932 0.087 119
1991 3.02 0.33 0.41 85463 351 101 122 964 10330 0.093 128
1992 3.09 0.32 0.40 98716 362 110 133 1017 9973 0.102 140
1993 3.12 0.32 0.40 108342 372 117 141 993 9657 0.103 141
1994 3.38 0.30 0.37 115889 382 112 136 948 9564 0.099 136
1995 6.42 0.16 0.19 156450 393 78 94 777 9328 0.083 114
1996 7.60 0.13 0.16 210235 404 86 103 733 8977 0.082 112
1997 7.92 0.13 0.16 253597 414 97 117 784 8620 0.091 125
1998 9.14 0.11 0.14 293991 420 96 116 841 8455 0.100 136
1999 9.56 0.10 0.13 342751 429 104 126 879 8189 0.107 147
2000 9.46 0.11 0.13 375284 444 112 135 963 8092 0.119 163
2001 9.34 0.11 0.13 399181 456 117 141 956 7574 0.126 173
2002 9.66 0.10 0.13 419262 464 117 142 864 6639 0.130 179
2003 10.79 0.09 0.12 438326 474 107 129 941 6156 0.153 210
2004 11.29 0.09 0.11 458876 487 104 126 920 5522 0.167 229
2005 10.90 0.09 0.11 477177 503 109 131 914 5187 0.176 242
2006 10.90 0.09 0.11 494496 520 109 132 894 4923 0.182 249
2007 10.93 0.09 0.11 514111 534 110 133 852 4629 0.184 252
2008 11.13 0.09 0.11 540460 555 109 132 790 4578 0.173 237
2009 13.51 0.07 0.09 569090 553 95 115 703 3759 0.187 256
2010 12.64 0.08 0.10 592745 562 104 126 654 3479 0.188 258
2011 12.42 0.08 0.10 612942 580 106 129 637 3494 0.182 250
Source: See Table 1A.
133
(14) Netherlands
Trade
Year e pmfg*e rxr CPI PPI RULC RULCAdj RULCAdjratio Int.Rate intratediff gppiratio realintratediff rxr1 rulcadjratio1 realintratediff1 Weight
1960 62 22 88 18.3 31.6 171 99 94.4 4.2 -0.9 88 94 0.062
1961 64 23 92 18.9 31.6 180 107 100.8 3.9 -1.1 -2.35 1.3 91 101 1.3 0.065
1962 65 24 92 19.3 31.7 181 110 101.9 4.2 -0.7 -0.44 -0.3 92 102 -0.2 0.065
1963 65 24 94 19.8 32.4 187 114 105.8 4.2 -0.7 1.16 -1.8 93 106 -1.9 0.065
1964 65 26 98 21.0 34.4 187 114 106.7 4.9 -0.4 2.81 -3.2 98 107 -3.5 0.068
1965 65 27 101 22.1 35.4 187 116 108.8 5.5 -0.2 -0.07 -0.2 101 109 -0.1 0.067
1966 64 28 101 23.5 37.2 184 116 107.8 6.6 0.3 2.07 -1.8 101 108 -1.8 0.064
1967 65 29 102 24.0 37.6 186 119 107.9 6.2 0.1 0.21 -0.1 102 108 -0.1 0.064
1968 64 29 101 25.1 38.2 177 116 105.0 6.5 0.0 -0.74 0.8 101 105 0.8 0.063
1969 64 29 98 26.9 37.3 170 123 112.0 7.5 0.3 -6.65 6.9 98 113 7.3 0.065
1970 65 30 98 28.1 39.1 175 126 112.0 8.2 0.5 -1.33 1.8 97 113 1.8 0.068
1971 67 33 102 30.1 40.8 178 131 114.2 7.4 0.6 0.27 0.3 102 115 0.1 0.070
1972 73 38 106 32.5 42.7 173 131 113.7 6.9 0.3 0.61 -0.3 106 115 -0.4 0.070
1973 84 47 112 35.1 45.4 175 135 121.1 7.9 0.1 -4.34 4.5 113 123 4.7 0.072
1974 87 60 128 38.5 49.5 176 137 126.4 9.8 0.4 -8.38 8.8 131 129 9.4 0.070
1975 92 60 114 42.4 52.4 174 141 121.2 8.8 0.5 -0.78 1.2 115 123 1.2 0.072
1976 88 60 113 46.1 56.1 165 135 118.5 9.0 0.6 -1.28 1.9 114 120 1.9 0.073
1977 95 66 115 49.2 59.2 160 133 116.4 8.1 0.3 -1.90 2.2 117 118 2.4 0.071
1978 108 77 117 51.2 60.0 157 134 116.5 7.7 -0.1 -3.73 3.6 118 118 3.9 0.070
1979 116 87 118 53.4 61.6 156 135 120.6 8.8 -0.7 -5.42 4.8 119 122 5.2 0.071
1980 117 90 114 56.9 66.1 151 130 117.4 10.2 -1.0 -4.27 3.2 115 119 3.6 0.069
1981 93 74 100 60.6 71.9 145 123 112.8 11.6 -1.4 -1.62 0.2 100 114 0.5 0.063
1982 87 72 100 64.3 76.0 144 122 111.8 10.1 -1.8 -0.89 -0.9 100 113 -0.8 0.064
1983 82 67 95 66.2 77.0 138 119 112.0 8.6 -1.5 -3.02 1.6 95 113 1.9 0.062
1984 73 63 92 68.3 80.4 130 110 106.4 8.3 -1.9 -0.26 -1.6 91 107 -1.5 0.058
1985 70 60 88 69.9 81.7 129 111 107.3 7.3 -1.5 -0.90 -0.6 87 108 -0.5 0.059
1986 95 83 98 70.0 79.5 129 114 105.1 6.3 -0.9 -0.96 0.1 98 105 0.2 0.054
1987 115 100 104 69.7 78.5 133 118 107.4 6.4 -0.8 -1.47 0.7 104 108 0.8 0.054
1988 118 106 102 70.1 78.9 129 115 106.7 6.4 -1.0 -2.16 1.1 103 107 1.3 0.052
1989 110 100 98 70.9 81.7 124 108 101.7 7.2 -1.3 -0.61 -0.7 97 102 -0.6 0.050
1990 128 117 104 72.6 86.9 125 105 98.0 8.9 -0.8 3.48 -4.2 104 98 -4.4 0.054
1991 125 114 100 75.5 87.2 126 109 99.3 8.7 0.0 -1.16 1.2 100 99 1.2 0.053
1992 133 122 103 78.3 86.2 126 114 102.3 8.1 0.4 -1.69 2.1 103 102 2.2 0.053
1993 125 118 104 79.8 84.5 125 118 104.7 6.4 0.4 -2.78 3.2 104 105 3.3 0.053
1994 128 120 103 82.1 85.2 117 112 102.7 6.9 0.5 -0.61 1.1 103 103 1.1 0.053
1995 145 139 109 83.5 87.4 112 107 101.2 6.9 0.8 -0.71 1.6 110 101 1.6 0.056
1996 138 131 106 85.2 88.6 110 105 99.3 6.2 0.9 0.76 0.1 106 99 0.0 0.055
1997 119 115 101 87.1 91.3 109 104 100.2 5.6 0.8 1.97 -1.2 101 100 -1.3 0.053
1998 118 115 103 88.9 89.2 109 109 103.2 4.6 0.6 -1.87 2.5 103 103 2.6 0.055
1999 113 109 100 90.8 89.3 106 108 103.9 4.6 0.5 -0.14 0.7 100 104 0.6 0.053
2000 98 96 94 92.9 100.4 101 93 95.5 5.4 0.6 7.71 -7.1 94 95 -7.6 0.051
2001 95 94 97 96.8 101.3 101 96 96.6 5.0 1.1 0.37 0.7 97 96 0.7 0.053
2002 100 100 100 100.0 100.0 100 100 100.0 4.9 1.4 -0.82 2.3 100 100 2.3 0.052
2003 120 123 111 102.1 100.7 99 101 102.1 4.1 1.4 -0.04 1.5 111 102 1.4 0.055
2004 132 136 114 103.4 104.5 96 95 101.6 4.1 1.1 0.94 0.2 115 102 0.1 0.056
2005 132 139 115 105.1 111.4 91 86 96.8 3.4 0.4 2.89 -2.5 116 97 -2.6 0.055
2006 133 139 114 106.3 117.1 89 81 95.2 3.8 0.2 1.85 -1.6 115 95 -1.8 0.056
2007 145 154 117 108.1 122.9 86 76 92.1 4.3 0.4 2.08 -1.7 119 92 -1.8 0.059
2008 156 171 125 110.7 132.1 89 74 91.4 4.2 1.0 1.71 -0.7 127 91 -0.9 0.061
2009 147 156 118 112.0 116.7 96 92 100.9 3.7 1.4 -8.06 9.4 119 101 9.9 0.063
2010 140 151 115 113.5 127.3 88 78 96.3 3.0 1.0 5.33 -4.3 116 96 -4.7 0.061
Source: See Table 1A and 1B.
Real Exchange Rate Real Adjusted Unit Labor Cost Ratio Interest Rates Excluding Country Being Examined
134
(15) Norway
Trade
Year e pmfg*e rxr CPI PPI RULC RULCAdj RULCAdjratio Int.Rate intratediff gppiratio realintratediff rxr1 rulcadjratio1 realintratediff1 Weight
1960 112 10 40 10.5 15.1 89 62 58.9 4.6 -0.5 39 58 0.015
1961 112 10 41 10.8 15.7 91 63 59.1 4.6 -0.4 1.59 -2.0 40 59 -2.0 0.015
1962 112 11 43 11.3 16.7 96 65 59.9 5.0 0.0 5.80 -5.8 43 59 -5.9 0.015
1963 112 11 43 11.7 17.0 94 65 60.1 5.0 0.1 0.42 -0.4 43 60 -0.4 0.015
1964 112 12 44 12.1 17.8 92 63 58.6 4.9 -0.4 1.33 -1.7 44 58 -1.7 0.015
1965 112 12 45 12.9 18.4 90 63 58.9 5.0 -0.7 0.51 -1.3 45 58 -1.2 0.015
1966 112 12 44 13.3 18.6 92 65 60.4 5.0 -1.3 -1.54 0.2 44 60 0.3 0.015
1967 112 13 45 13.8 19.3 95 68 61.7 5.0 -1.1 2.79 -3.9 45 61 -3.9 0.016
1968 112 13 47 14.2 20.3 96 67 60.8 4.9 -1.5 2.76 -4.3 47 60 -4.3 0.014
1969 112 14 48 14.7 21.4 93 64 58.6 5.1 -2.1 0.75 -2.8 48 58 -2.8 0.014
1970 112 16 51 16.2 23.7 92 63 55.6 6.3 -1.4 4.19 -5.6 50 55 -5.7 0.014
1971 113 17 52 17.3 24.9 95 66 57.3 6.4 -0.4 1.07 -1.4 51 57 -1.5 0.014
1972 121 19 54 18.5 26.7 94 65 56.6 6.3 -0.3 2.84 -3.2 53 56 -3.2 0.013
1973 139 24 59 20.0 28.8 93 64 57.7 6.2 -1.6 -2.93 1.3 58 57 1.4 0.014
1974 145 29 63 21.8 32.3 96 65 59.9 7.1 -2.3 -5.50 3.2 62 59 3.3 0.014
1975 153 35 67 24.3 37.6 104 67 57.7 7.3 -1.0 8.81 -9.9 67 57 -10.0 0.015
1976 146 36 68 26.5 40.0 107 71 62.3 7.3 -1.1 -1.76 0.7 68 62 0.7 0.015
1977 150 40 71 29.1 39.7 108 79 69.2 7.4 -0.4 -7.67 7.3 70 69 7.4 0.015
1978 152 43 65 31.3 41.1 107 82 70.8 8.5 0.6 -1.82 2.4 64 70 2.4 0.013
1979 158 49 67 32.9 43.8 100 75 66.4 8.6 -0.9 -1.59 0.7 67 66 0.8 0.013
1980 162 52 65 36.4 49.1 100 74 66.5 10.3 -1.0 -0.08 -0.9 65 66 -0.9 0.014
1981 139 48 65 41.4 54.0 98 75 69.1 12.3 -0.6 -0.60 0.0 64 69 0.0 0.014
1982 124 46 64 46.2 57.4 93 75 69.0 13.2 1.3 -0.26 1.5 63 69 1.5 0.014
1983 109 45 64 50.0 60.8 92 75 71.0 12.9 2.8 1.25 1.5 63 71 1.5 0.014
1984 98 43 63 53.2 64.8 88 72 69.8 12.2 2.0 1.81 0.2 63 69 0.1 0.013
1985 93 43 63 56.3 67.5 88 74 71.3 12.6 3.7 1.72 2.0 62 71 1.9 0.014
1986 108 53 62 60.2 68.1 91 81 74.2 13.5 6.3 2.67 3.6 62 74 3.5 0.013
1987 118 62 65 65.4 71.9 91 83 75.2 13.6 6.3 5.32 1.0 64 75 0.9 0.013
1988 122 72 70 69.8 76.2 91 83 77.3 13.0 5.5 3.09 2.5 70 77 2.4 0.012
1989 116 71 69 73.0 80.2 89 81 76.4 10.8 2.3 1.04 1.3 69 76 1.3 0.012
1990 128 78 69 76.0 82.2 88 82 76.3 10.7 1.0 -0.28 1.3 69 76 1.3 0.013
1991 123 78 69 78.6 84.1 90 84 76.3 9.9 1.1 0.81 0.3 68 76 0.3 0.012
1992 129 84 70 80.4 84.3 89 85 75.6 9.8 2.1 -0.27 2.4 70 75 2.3 0.012
1993 112 77 68 82.2 84.1 88 86 76.4 6.5 0.6 -1.13 1.7 68 76 1.7 0.011
1994 113 81 69 83.5 85.8 90 87 79.7 7.1 0.7 0.63 0.1 69 80 0.1 0.011
1995 126 100 78 85.5 87.5 93 91 85.6 6.8 0.8 -1.43 2.2 78 85 2.2 0.011
1996 124 96 78 86.6 88.7 92 90 84.7 5.9 0.7 0.80 -0.1 77 85 -0.1 0.012
1997 113 91 79 88.9 89.4 93 92 88.8 5.1 0.4 -0.24 0.6 79 89 0.6 0.012
1998 106 92 82 90.9 90.2 99 100 94.7 5.4 1.3 1.36 0.0 82 95 -0.1 0.011
1999 102 92 84 92.9 93.3 98 98 94.2 5.4 1.3 3.17 -1.9 84 94 -1.9 0.011
2000 91 86 85 95.8 102.8 97 91 92.8 6.4 1.6 5.45 -3.9 84 93 -4.0 0.012
2001 89 88 91 98.7 103.1 98 94 94.1 6.3 2.5 -0.18 2.6 91 94 2.6 0.012
2002 100 100 100 100.0 100.0 100 100 100.0 6.3 2.9 -2.53 5.4 100 100 5.4 0.012
2003 113 113 101 102.5 101.7 93 94 95.1 4.5 1.8 1.01 0.8 101 95 0.8 0.012
2004 119 120 100 102.9 108.2 91 86 92.7 3.6 0.6 3.44 -2.8 100 93 -2.9 0.012
2005 124 128 106 104.5 115.1 92 83 93.7 3.3 0.3 2.67 -2.3 106 94 -2.4 0.013
2006 125 141 116 107.0 124.1 99 85 100.0 3.9 0.4 4.37 -4.0 116 100 -4.1 0.014
2007 136 160 122 107.7 131.3 102 84 101.8 4.8 0.9 2.95 -2.1 122 102 -2.1 0.014
2008 142 163 119 111.8 139.6 100 80 99.0 4.3 1.1 0.65 0.5 119 99 0.4 0.015
2009 127 154 116 114.2 139.7 101 83 91.3 3.3 1.0 4.13 -3.1 116 91 -3.2 0.014
2010 132 158 120 117.0 148.9 97 76 93.9 2.8 0.8 2.94 -2.1 121 94 -2.2 0.014
Source: See Table 1A and 1B.
Real Exchange Rate Real Adjusted Unit Labor Cost Ratio Interest Rates Excluding Country Being Examined
135
(16) Spain
Trade
Year e pmfg*e rxr CPI PPI RULC RULCAdj RULCAdjratio Int.Rate intratediff gppiratio realintratediff rxr1 rulcadjratio1 realintratediff1 Weight
1960 293 24 94 3.5 9.3 115 43 41.1 4.6 -0.5 94 41 0.009
1961 293 24 93 3.5 9.3 114 43 40.6 4.6 -0.4 -2.25 1.8 93 40 2.0 0.011
1962 293 24 92 3.7 9.3 108 43 40.0 4.6 -0.3 -0.85 0.5 92 40 0.7 0.013
1963 293 24 91 4.1 9.3 99 43 40.2 4.6 -0.3 -1.07 0.8 91 40 0.9 0.014
1964 293 24 89 4.4 9.3 93 43 40.5 4.6 -0.7 -3.20 2.5 89 40 2.6 0.014
1965 293 24 89 4.9 10.2 93 45 41.6 4.6 -1.1 7.02 -8.2 89 41 -8.1 0.016
1966 293 25 92 5.2 10.5 95 47 43.9 4.6 -1.7 -0.31 -1.4 92 43 -1.2 0.018
1967 285 26 92 5.6 10.6 101 53 48.6 5.1 -1.0 -0.20 -0.8 91 48 -0.9 0.017
1968 251 24 83 5.9 10.8 97 53 47.9 5.1 -1.4 -0.31 -1.0 83 47 -1.1 0.016
1969 251 25 85 6.0 11.1 95 51 46.8 5.5 -1.7 -1.85 0.1 84 46 0.3 0.016
1970 251 26 84 6.3 11.3 96 54 48.0 6.5 -1.2 -4.22 3.0 83 47 3.1 0.017
1971 253 27 83 6.8 11.9 100 57 49.8 5 -1.8 1.49 -3.3 83 49 -3.3 0.017
1972 274 31 86 7.4 12.7 97 57 49.0 5 -1.6 2.60 -4.2 86 48 -4.0 0.019
1973 302 38 90 8.2 14.0 97 57 51.3 6 -1.8 -0.69 -1.1 90 51 -1.0 0.019
1974 305 45 96 9.5 16.5 102 59 54.5 7 -2.4 -0.79 -1.6 96 54 -1.4 0.021
1975 307 52 99 11.2 18.1 115 71 61.1 7 -1.3 2.10 -3.4 99 60 -3.4 0.021
1976 263 52 98 13.1 20.5 126 81 70.7 7 -1.4 4.61 -6.0 98 70 -6.1 0.021
1977 232 54 95 16.3 24.6 132 88 76.7 8.03 0.2 11.80 -11.6 95 76 -11.9 0.019
1978 230 63 95 19.6 28.6 136 93 80.4 9.019 1.2 10.44 -9.3 95 80 -9.5 0.019
1979 262 82 111 22.6 32.8 143 99 88.0 14.5041 5.1 5.60 -0.5 112 88 -0.6 0.020
1980 246 87 110 26.1 38.5 138 94 84.3 15.8304 4.6 4.69 -0.1 110 84 -0.2 0.022
1981 191 76 102 30.0 44.5 134 90 83.0 15.8058 2.9 4.67 -1.8 102 83 -2.0 0.021
1982 160 70 97 34.3 50.0 129 88 81.2 15.8437 3.9 5.38 -1.5 97 81 -1.6 0.022
1983 123 59 84 38.5 57.1 123 83 78.1 18.3546 8.3 9.11 -0.8 84 78 -1.2 0.021
1984 110 59 86 42.8 64.0 118 79 75.9 14.9767 4.8 7.13 -2.3 86 75 -2.6 0.021
1985 104 60 89 46.6 69.1 114 76 74.1 12.1358 3.3 5.32 -2.0 88 74 -2.2 0.021
1986 126 79 93 50.7 69.8 115 84 77.0 9.99417 2.8 2.75 0.0 93 77 -0.1 0.022
1987 143 93 96 53.3 70.3 114 86 78.7 12.0954 4.9 0.58 4.3 96 78 4.2 0.025
1988 151 102 99 55.9 72.4 112 86 80.2 11.265 3.8 0.25 3.6 99 80 3.5 0.026
1989 149 106 104 59.7 75.5 112 88 83.4 13.6338 5.1 -0.01 5.1 104 83 5.1 0.027
1990 173 128 113 63.7 77.1 117 96 90.2 14.4233 4.7 -0.57 5.3 113 90 5.3 0.030
1991 169 129 113 67.5 78.3 118 102 92.5 12.4063 3.7 -0.02 3.7 113 92 3.6 0.031
1992 172 134 112 71.5 79.3 121 109 97.6 12.065 4.4 0.76 3.6 113 98 3.5 0.032
1993 138 111 97 74.7 81.3 125 115 101.4 10.3717 4.4 1.64 2.8 97 101 2.6 0.028
1994 132 108 93 78.3 84.8 120 110 101.0 9.05167 2.7 2.81 -0.2 93 101 -0.3 0.030
1995 141 123 97 81.9 90.2 115 105 98.7 10.5308 4.5 2.92 1.6 97 99 1.4 0.031
1996 139 126 102 84.8 91.7 115 107 100.5 7.985 2.7 1.03 1.7 102 101 1.6 0.033
1997 120 111 97 86.5 92.6 115 107 102.9 5.71292 0.9 0.01 0.9 97 103 0.9 0.032
1998 118 109 98 88.1 92.0 112 107 101.5 4.31375 0.3 -0.22 0.5 98 102 0.5 0.034
1999 113 105 96 90.1 92.7 107 104 100.2 3.86875 -0.2 0.42 -0.7 96 100 -0.7 0.035
2000 98 93 92 93.2 98.2 104 99 100.9 5.07 0.3 1.45 -1.2 92 101 -1.3 0.033
2001 95 92 95 96.6 99.8 102 98 98.7 4.51667 0.7 1.14 -0.5 95 99 -0.5 0.035
2002 100 100 100 100.0 100.0 100 100 100.0 4.14917 0.7 0.72 0.0 100 100 -0.1 0.036
2003 120 122 109 103.1 100.7 100 102 103.3 3.16917 0.5 -0.06 0.5 110 103 0.5 0.040
2004 132 138 115 106.2 104.0 98 100 107.2 3.13458 0.2 0.54 -0.4 116 107 -0.4 0.041
2005 132 143 118 109.8 108.7 98 99 110.8 2.79042 -0.1 0.84 -1.0 119 111 -1.0 0.040
2006 133 150 123 113.6 113.8 96 96 113.0 3.52167 -0.1 1.37 -1.4 124 114 -1.5 0.040
2007 145 170 130 116.8 117.3 98 97 118.6 4.18875 0.3 0.24 0.1 131 119 0.0 0.042
2008 156 191 140 121.6 123.4 100 98 121.1 4.03958 0.8 -0.41 1.2 142 122 1.2 0.041
2009 147 181 136 121.3 117.4 101 104 114.6 2.4875 0.2 -0.99 1.2 138 115 1.2 0.040
2010 140.3 172 131 123.4 122.5 94 95 116.5 2.97208 1.0 0.78 0.2 132 117 0.2 0.037
Source: See Table 1A and 1B.
Real Exchange Rate Real Adjusted Unit Labor Cost Ratio Interest Rates Excluding Country Being Examined
136
(17) Sweden
Trade
Year e pmfg*e rxr CPI PPI RULC RULCAdj RULCAdjratio Int.Rate intratediff gppiratio realintratediff rxr1 rulcadjratio1 realintratediff1 Weight
1960 188 36 146 9.9 11.7 199 168 160.2 5.2 0.1 148 163 0.034
1961 188 37 146 10.2 11.9 203 173 162.9 5.3 0.3 -0.77 1.1 148 166 1.1 0.034
1962 189 38 147 10.3 12.1 207 178 164.3 5.0 0.1 0.74 -0.7 149 167 -0.7 0.033
1963 187 38 147 10.7 12.5 209 179 166.6 4.4 -0.5 2.21 -2.7 149 170 -2.8 0.033
1964 189 39 149 11.1 12.8 202 176 165.0 5.2 -0.1 -0.70 0.6 151 168 0.7 0.034
1965 189 40 149 11.8 13.1 197 177 165.7 6.1 0.3 -0.58 0.9 151 169 1.0 0.034
1966 188 41 148 12.6 13.3 193 183 169.2 6.4 0.1 -0.75 0.8 150 172 0.8 0.033
1967 188 42 149 12.9 13.7 194 182 165.0 5.4 -0.7 2.10 -2.8 151 168 -2.9 0.032
1968 188 42 149 13.3 13.7 186 181 164.0 6.1 -0.4 -2.43 2.1 150 167 2.1 0.031
1969 188 43 147 13.6 14.3 183 174 159.2 7.0 -0.2 -0.46 0.3 149 162 0.3 0.031
1970 188 44 143 14.6 15.2 184 176 156.3 7.9 0.2 0.58 -0.4 145 159 -0.4 0.032
1971 191 47 142 15.8 15.6 183 185 161.5 6.5 -0.2 -1.73 1.5 144 164 1.5 0.030
1972 204 53 145 16.7 16.3 184 189 163.4 5.5 -1.1 0.35 -1.4 147 166 -1.4 0.030
1973 223 62 148 18.4 18.2 176 178 159.2 5.1 -2.7 0.26 -3.0 149 162 -2.9 0.030
1974 219 73 157 20.0 22.7 182 161 148.3 7.3 -2.1 4.93 -7.0 159 150 -7.1 0.030
1975 235 89 169 21.9 24.1 200 182 156.8 8.0 -0.3 -0.26 0.0 171 159 0.0 0.031
1976 223 89 168 24.3 26.3 209 193 168.9 7.9 -0.5 0.57 -1.0 171 172 -1.1 0.030
1977 218 91 160 27.3 28.8 207 196 171.8 9.6 1.8 1.84 0.0 162 174 -0.3 0.027
1978 215 98 149 29.4 31.1 208 196 170.0 8.4 0.5 2.46 -1.9 150 172 -2.1 0.025
1979 227 112 152 31.3 34.8 195 175 155.9 8.6 -0.8 3.14 -4.0 154 158 -4.0 0.026
1980 230 124 156 35.8 39.7 186 168 151.3 11.7 0.4 1.71 -1.3 157 153 -1.4 0.025
1981 192 112 150 39.9 44.1 185 168 154.2 13.0 0.1 0.65 -0.6 151 156 -0.7 0.023
1982 155 100 139 43.5 50.6 176 151 139.3 13.1 1.2 7.38 -6.2 140 140 -6.4 0.023
1983 127 88 125 48.7 56.1 162 141 132.3 12.3 2.2 6.24 -4.0 126 133 -4.2 0.023
1984 118 88 127 52.4 61.1 155 133 128.1 12.1 1.9 3.91 -2.0 128 129 -2.2 0.022
1985 113 90 132 56.4 64.5 158 138 133.5 13.6 4.8 3.00 1.8 133 134 1.6 0.023
1986 136 120 141 58.8 65.8 161 144 132.4 10.0 2.8 3.92 -1.1 142 133 -1.2 0.024
1987 153 141 146 61.4 68.0 161 146 132.7 10.5 3.3 3.04 0.3 147 134 0.1 0.025
1988 158 152 148 64.0 72.0 163 145 134.3 10.7 3.3 3.09 0.2 149 135 0.0 0.024
1989 151 155 152 67.7 77.8 165 144 136.2 11.3 2.8 3.67 -0.8 154 137 -1.0 0.024
1990 164 176 156 75.5 81.2 162 151 141.0 13.4 3.7 1.61 2.1 158 142 2.0 0.023
1991 161 177 155 82.3 82.2 160 160 145.8 11.1 2.4 -0.28 2.7 157 147 2.7 0.021
1992 167 183 154 83.7 80.9 157 162 145.1 11.4 3.7 -2.17 5.9 155 146 5.9 0.021
1993 125 143 125 85.5 85.4 139 140 123.6 8.5 2.5 4.70 -2.2 126 124 -2.3 0.019
1994 126 145 125 87.4 89.3 127 124 113.5 8.6 2.2 3.02 -0.9 125 114 -0.9 0.020
1995 136 167 131 91.5 97.9 120 112 105.9 9.5 3.4 6.15 -2.7 132 106 -2.9 0.022
1996 145 169 137 92.5 95.7 124 120 113.1 6.9 1.7 -2.94 4.6 138 113 4.6 0.022
1997 127 144 127 93.1 96.5 119 115 110.3 5.4 0.6 -0.16 0.8 127 111 0.7 0.021
1998 122 135 122 94.0 95.9 115 113 107.4 4.6 0.6 -0.19 0.7 122 108 0.7 0.022
1999 118 125 114 97.1 95.4 105 107 103.3 4.0 -0.1 -0.80 0.7 114 103 0.8 0.021
2000 106 111 110 95.4 99.4 104 100 101.9 4.7 -0.2 -0.19 0.0 110 102 0.0 0.020
2001 94 98 101 97.6 100.7 109 105 105.9 4.6 0.7 0.83 -0.1 101 106 -0.1 0.018
2002 100 100 100 100.0 100.0 100 100 100.0 4.7 1.2 -0.22 1.5 100 100 1.4 0.019
2003 120 117 105 102.3 98.8 94 98 99.0 3.8 1.1 -1.94 3.1 105 99 3.1 0.020
2004 132 122 102 102.6 99.4 87 90 96.1 3.3 0.3 -2.11 2.4 102 96 2.5 0.021
2005 130 119 98 102.9 103.2 84 83 93.8 2.6 -0.4 0.22 -0.6 98 94 -0.6 0.020
2006 132 119 97 105.1 107.2 78 76 89.3 3.0 -0.6 0.59 -1.1 97 89 -1.1 0.020
2007 144 132 101 107.0 111.7 79 75 92.0 3.9 0.0 1.29 -1.3 101 92 -1.3 0.021
2008 148 132 97 113.2 115.0 81 80 98.3 3.9 0.7 -2.52 3.2 97 98 3.2 0.021
2009 127 117 88 112.9 116.5 95 92 100.9 1.8 -0.5 5.45 -5.9 88 101 -6.0 0.019
2010 135 121 92 112.6 116.9 78 75 92.7 1.7 -0.3 -3.08 2.8 92 93 2.9 0.020
Source: See Table 1A and 1B.
Real Exchange Rate Real Adjusted Unit Labor Cost Ratio Interest Rates Excluding Country Being Examined
137
(18) Taiwan
Year e (Pmfg*e) (Pmfg*e)oecd rxr RULC CPI PPI RULCadj RULCADJoecd RULCAdjratio Deviation
1976 91 72 53 136 132 39 70 74 114 64 71
1977 91 74 57 130 126 42 71 73 114 65 65
1978 93 81 66 122 119 44 74 71 115 62 60
1979 96 93 73 127 125 48 84 72 112 64 63
1980 96 101 80 127 120 58 102 68 110 62 65
1981 94 104 74 140 110 67 110 67 108 62 77
1982 88 99 72 138 114 69 110 71 108 66 72
1983 86 96 70 136 108 70 109 69 105 66 71
1984 87 97 69 141 113 70 109 72 103 70 71
1985 87 96 68 142 121 70 106 80 102 78 64
1986 91 97 85 115 125 70 103 85 108 79 36
1987 109 107 97 111 127 71 99 90 109 82 29
1988 121 116 103 112 132 71 98 96 107 90 22
1989 131 121 102 118 133 75 97 102 105 97 22
1990 128 122 113 108 140 78 97 112 107 105 3
1991 129 123 114 108 138 80 97 115 110 104 3
1992 137 124 119 104 138 84 94 124 112 111 -7
1993 131 124 114 109 139 87 96 126 113 111 -3
1994 131 124 116 107 135 90 98 125 109 114 -7
1995 130 133 127 104 132 93 105 117 106 110 -6
1996 126 130 124 105 127 96 104 118 106 111 -6
1997 120 127 114 111 125 97 104 117 104 112 -1
1998 103 115 111 103 121 99 104 115 105 109 -6
1999 107 109 109 100 117 99 99 116 104 111 -12
2000 111 112 101 110 111 100 101 109 98 112 -1
2001 102 104 97 107 112 100 100 112 99 113 -7
2002 100 100 100 100 100 100 100 100 100 100 0
2003 100 99 111 89 96 100 102 94 98 96 -7
2004 104 104 120 87 93 101 110 86 92 93 -7
2005 108 105 121 87 89 104 110 84 88 96 -9
2006 106 106 122 87 87 104 117 78 84 93 -6
2007 105 109 131 83 79 106 124 68 81 84 -1
2008 110 111 137 81 77 110 131 65 78 83 -2
2009 105 99 130 76 72 109 119 66 88 75 1
Source: See Table 1A.
138
(19) United Kingdom
Trade
Year e pmfg*e rxr CPI PPI RULC RULCAdj RULCAdjratio Int.Rate intratediff gppiratio realintratediff rxr1 rulcadjratio1 realintratediff1 Weight
1960 187 17 69 7.2 11.1 126 82 77.7 5.6 0.5 65 74 0.148
1961 187 18 70 7.3 11.3 133 85 79.9 6.0 1.0 0.35 0.6 66 77 -0.2 0.141
1962 187 18 71 7.5 11.6 131 86 79.2 5.3 0.4 1.33 -0.9 67 76 -1.7 0.135
1963 186 18 70 7.6 11.7 129 84 78.0 4.8 -0.1 0.02 -0.1 67 75 -0.3 0.133
1964 186 19 71 8.0 12.1 123 82 76.5 5.5 0.2 -0.39 0.6 67 74 0.4 0.129
1965 186 19 72 8.5 12.5 123 84 78.2 6.6 0.8 0.87 0.0 69 76 -0.8 0.122
1966 186 20 72 8.7 12.9 126 85 79.0 6.8 0.5 -0.25 0.7 69 77 0.2 0.116
1967 183 20 70 8.8 13.0 123 83 75.7 6.7 0.6 0.30 0.3 67 73 -0.1 0.113
1968 159 18 63 9.3 13.5 116 80 72.4 7.6 1.1 1.29 -0.2 60 70 -0.9 0.106
1969 159 19 64 9.8 14.0 117 82 74.8 8.8 1.6 -0.53 2.1 60 72 1.6 0.101
1970 160 20 65 10.5 15.0 124 86 76.8 7.9 0.2 0.89 -0.7 62 75 -1.2 0.096
1971 163 23 69 11.4 16.4 123 86 74.9 6.7 -0.1 4.69 -4.8 66 73 -5.3 0.096
1972 166 25 68 12.3 17.2 126 89 77.5 7.5 1.0 0.95 0.0 65 76 -0.5 0.091
1973 163 25 60 13.3 18.5 123 88 79.3 10.4 2.6 -3.16 5.8 57 78 5.7 0.087
1974 156 27 57 15.6 18.8 124 103 94.9 12.5 3.1 -14.54 17.6 54 94 18.5 0.085
1975 148 33 64 19.2 23.1 137 114 97.8 11.5 3.1 15.03 -11.9 61 98 -13.6 0.086
1976 120 31 58 22.3 26.9 131 108 94.9 12.1 3.7 7.58 -3.9 56 94 -5.3 0.079
1977 116 33 59 25.9 32.0 126 102 89.5 10.1 2.3 10.59 -8.3 56 89 -9.3 0.082
1978 128 42 63 28.3 34.9 133 108 93.4 11.3 3.5 3.42 0.1 61 93 -0.4 0.085
1979 141 53 73 32.1 39.1 138 114 101.2 12.6 3.2 3.11 0.1 70 101 -0.4 0.087
1980 155 73 92 37.8 45.3 143 120 107.8 13.8 2.6 3.35 -0.7 91 109 -1.2 0.089
1981 135 73 99 42.2 50.1 142 119 109.8 14.7 1.7 0.20 1.5 98 111 1.2 0.084
1982 116 67 94 46.1 54.4 135 115 105.4 12.8 0.9 1.73 -0.9 93 106 -1.1 0.085
1983 101 61 88 48.1 57.9 129 107 100.9 11.2 1.1 1.91 -0.8 86 101 -1.1 0.084
1984 89 55 80 50.7 61.4 126 104 100.0 11.3 1.1 1.16 -0.1 78 100 -0.4 0.080
1985 86 56 82 53.7 65.2 126 103 100.3 11.1 2.3 3.63 -1.4 80 100 -1.8 0.083
1986 98 67 79 55.5 66.1 128 107 98.7 10.0 2.8 3.16 -0.4 77 99 -0.9 0.081
1987 109 78 81 57.8 68.3 130 110 100.0 9.4 2.1 3.18 -1.0 80 100 -1.4 0.084
1988 119 89 86 60.7 70.8 125 107 99.6 9.7 2.2 0.91 1.3 85 100 1.1 0.086
1989 109 87 85 65.4 74.2 120 106 99.9 10.7 2.2 0.55 1.7 84 100 1.4 0.084
1990 119 99 88 71.6 78.9 119 108 100.8 12.1 2.4 3.43 -1.0 87 101 -1.5 0.085
1991 118 101 88 75.8 83.1 121 110 100.0 10.2 1.5 3.77 -2.3 87 100 -2.8 0.081
1992 118 102 86 78.6 85.7 109 100 89.7 8.9 1.2 2.56 -1.3 85 89 -1.6 0.080
1993 100 89 79 79.9 89.1 108 97 85.8 6.6 0.7 3.06 -2.3 77 85 -2.7 0.078
1994 102 94 81 81.7 91.3 105 94 86.2 7.8 1.4 1.07 0.4 80 85 0.2 0.077
1995 105 102 80 84.6 95.0 105 94 88.2 7.9 1.9 0.65 1.2 79 87 1.0 0.076
1996 104 104 84 86.8 97.5 103 92 86.4 7.3 2.0 1.97 0.1 83 85 -0.3 0.080
1997 109 112 99 89.4 98.4 102 93 89.2 7.0 2.2 -0.06 2.3 98 88 2.1 0.084
1998 110 114 103 92.5 98.5 105 98 93.4 5.8 1.7 0.50 1.2 103 93 1.1 0.082
1999 108 110 100 93.8 99.0 104 99 95.3 5.4 1.3 0.30 1.0 100 95 0.9 0.080
2000 101 100 98 96.6 100.4 100 96 98.2 5.8 1.0 -2.85 3.8 98 98 4.0 0.077
2001 96 96 98 98.3 100.1 99 97 97.6 5.0 1.2 -0.84 2.0 98 97 2.0 0.078
2002 100 100 100 100.0 100.0 100 100 100.0 4.8 1.4 0.39 1.0 100 100 0.9 0.079
2003 109 108 97 102.9 100.6 97 99 100.5 4.2 1.5 -0.07 1.6 97 101 1.5 0.076
2004 122 119 100 106.0 101.6 93 97 103.5 4.8 1.9 -1.75 3.6 100 104 3.6 0.075
2005 121 121 100 109.0 103.6 91 95 107.3 4.4 1.4 -1.61 3.0 100 108 3.1 0.075
2006 123 123 101 112.5 105.7 89 95 111.7 4.5 1.0 -1.22 2.2 101 113 2.2 0.077
2007 133 136 104 117.3 108.2 85 93 113.0 5.2 1.3 -0.50 1.8 104 114 1.7 0.070
2008 123 126 92 121.9 115.5 84 88 108.7 4.3 1.1 1.07 0.0 91 109 -0.1 0.066
2009 104 111 84 121.3 120.4 88 89 97.5 2.7 0.4 8.51 -8.1 82 97 -8.7 0.065
2010 103 109 83 126.9 125.7 85 85 105.1 2.4 0.4 0.81 -0.4 82 105 -0.5 0.064
Source: See Table 1A and 1B.
Real Exchange Rate Real Adjusted Unit Labor Cost Ratio Interest Rates Excluding Country Being Examined
139
(20) United States
Trade
Year e pmfg*e rxr CPI PPI RULC RULCAdj RULCAdjratio Int.Rate intratediff gppiratio realintratediff rxr1 rulcadjratio1 realintratediff1 Weight
1960 100 41 165 17.9 24.0 244 182 172.8 4.0 -1.1 189 201 0.217
1961 100 41 163 18.1 24.0 243 182 171.1 3.5 -1.5 -2.25 0.8 185 197 2.0 0.207
1962 100 41 162 18.3 24.0 237 180 166.6 3.5 -1.4 -0.85 -0.6 184 190 -0.1 0.207
1963 100 40 156 18.5 24.0 225 173 160.6 3.7 -1.2 -1.07 -0.2 175 181 0.5 0.200
1964 100 41 153 18.8 24.0 220 172 160.9 4.0 -1.3 -3.20 1.9 170 181 3.0 0.199
1965 100 41 151 19.0 24.5 213 166 155.0 4.2 -1.5 -1.00 -0.5 168 173 0.1 0.197
1966 100 42 150 19.7 25.3 211 165 152.5 5.2 -1.1 0.31 -1.4 167 170 -1.3 0.203
1967 100 43 152 20.3 25.6 216 171 155.7 5.0 -1.1 0.59 -1.7 168 174 -1.6 0.202
1968 100 44 156 21.1 26.3 215 173 157.1 5.7 -0.8 0.06 -0.8 175 177 -0.8 0.208
1969 100 45 155 22.2 27.3 216 176 161.0 7.0 -0.2 -0.59 0.4 172 181 0.7 0.197
1970 100 47 153 23.5 28.2 218 182 162.1 7.3 -0.4 -2.64 2.2 170 182 3.0 0.193
1971 100 48 148 24.6 29.1 208 175 152.9 5.7 -1.1 -0.90 -0.2 161 168 0.3 0.186
1972 100 49 135 25.4 30.0 203 172 148.5 5.7 -0.9 -1.04 0.2 145 162 0.6 0.184
1973 100 49 118 26.8 32.7 197 161 144.7 7.0 -0.8 -1.79 1.0 122 157 1.5 0.179
1974 100 54 116 29.8 37.8 202 159 146.7 7.8 -1.6 -2.86 1.3 120 160 2.2 0.186
1975 100 62 117 32.6 41.9 200 155 133.8 7.5 -0.8 3.65 -4.5 121 143 -5.2 0.185
1976 100 64 121 34.5 43.7 194 153 134.0 6.8 -1.6 -3.68 2.1 126 143 3.2 0.188
1977 100 67 118 36.8 46.5 191 151 132.2 6.7 -1.1 -1.02 -0.1 122 141 0.4 0.190
1978 100 71 108 39.5 50.1 190 150 130.2 8.3 0.4 2.25 -1.8 110 138 -2.4 0.190
1979 100 77 104 43.3 55.8 188 146 129.9 9.7 0.3 2.60 -2.3 105 138 -3.0 0.186
1980 100 83 104 48.1 63.2 191 145 130.6 11.5 0.3 1.07 -0.8 105 139 -1.1 0.186
1981 100 88 118 52.8 69.1 181 138 127.3 14.4 1.5 -1.10 2.6 123 135 2.5 0.203
1982 100 92 129 55.8 71.9 185 144 132.4 12.9 1.0 -2.43 3.4 137 142 3.8 0.198
1983 100 93 132 58.3 73.0 171 136 128.5 10.4 0.4 -2.81 3.2 142 137 3.8 0.203
1984 100 94 137 60.6 74.6 165 134 129.4 11.9 1.7 -2.45 4.2 150 139 4.6 0.223
1985 100 94 138 62.7 75.2 162 135 131.1 9.6 0.8 -1.70 2.5 151 141 2.7 0.218
1986 100 97 115 63.8 73.7 164 142 130.9 7.1 -0.2 -0.18 0.0 119 140 0.0 0.205
1987 100 96 99 66.0 75.5 153 134 122.0 7.7 0.5 2.21 -1.7 99 128 -2.5 0.194
1988 100 98 95 68.4 78.1 147 129 119.8 8.3 0.8 0.67 0.2 94 125 -0.2 0.196
1989 100 102 100 71.4 82.1 145 126 119.1 8.6 0.1 0.82 -0.8 100 124 -1.0 0.199
1990 100 106 94 74.9 85.7 143 125 116.8 8.3 -1.4 1.63 -3.1 92 121 -3.2 0.185
1991 100 108 95 77.6 86.7 142 127 115.8 6.8 -1.9 -0.37 -1.5 94 120 -1.1 0.186
1992 100 109 91 79.6 87.8 140 127 113.1 5.3 -2.4 0.75 -3.1 90 116 -2.8 0.190
1993 100 109 96 81.6 89.2 136 124 110.0 4.4 -1.5 0.67 -2.2 95 113 -1.9 0.212
1994 100 109 94 83.3 90.3 131 121 110.4 6.3 -0.1 -0.14 0.0 92 113 0.1 0.211
1995 100 110 86 85.3 93.0 125 115 108.5 6.3 0.2 -0.34 0.5 83 111 0.6 0.200
1996 100 109 88 87.6 95.1 120 111 104.2 6.0 0.7 1.66 -0.9 85 105 -1.5 0.207
1997 100 108 95 89.6 95.4 116 109 104.4 6.1 1.3 -0.78 2.1 94 106 2.1 0.222
1998 100 107 96 90.7 94.4 115 111 105.3 5.1 1.1 -0.61 1.7 95 107 1.7 0.226
1999 100 105 96 92.7 96.0 111 107 103.1 5.5 1.4 1.53 -0.2 95 104 -0.9 0.234
2000 100 102 101 95.7 99.9 107 103 105.2 6.2 1.4 -0.40 1.8 101 107 1.6 0.248
2001 100 102 105 98.4 100.7 106 104 104.2 4.1 0.2 0.28 0.0 106 106 -0.2 0.242
2002 100 100 100 100.0 100.0 100 100 100.0 3.1 -0.3 -0.17 -0.2 100 100 0.0 0.234
2003 100 99 89 102.3 102.6 97 97 98.2 2.1 -0.6 1.87 -2.5 86 98 -2.9 0.217
2004 100 98 82 105.0 107.0 88 86 92.4 2.8 -0.2 1.45 -1.6 78 90 -2.0 0.211
2005 100 101 83 108.5 112.9 84 81 90.7 3.9 1.0 1.80 -0.8 79 88 -1.5 0.215
2006 100 102 83 112.0 117.4 80 77 90.0 4.8 1.2 0.69 0.5 79 87 0.1 0.213
2007 100 101 77 115.3 121.9 76 72 88.1 4.3 0.5 1.01 -0.6 72 85 -0.8 0.204
2008 100 102 75 119.7 131.6 78 71 87.6 2.2 -1.0 2.22 -3.2 69 85 -3.5 0.200
2009 100 106 80 119.2 125.1 79 75 82.5 1.4 -0.9 -1.13 0.2 76 79 0.7 0.202
2010 100 105 80 121.2 131.3 70 65 79.8 1.1 -0.9 1.37 -2.2 76 75 -2.5 0.210
Source: See Table 1A and 1B.
Real Exchange Rate Real Adjusted Unit Labor Cost Ratio Interest Rates Excluding Country Being Examined
140
Appendix to Chapter I-B
(1) Argentina
Dependent variable is dLRXR
42 observations used for estimation from 1966 to 2007 42 observations used for estimation from 1966 to 2007
Regressor Coefficient Standard Error T-Ratio[Prob]
dLRXR1 .40535 .21387 1.8953[.067]
dLRXR2 .081804 .20155 .40587[.688]
dLRXR3 .52997 .20825 2.5449[.016]
dLRXR4 .29139 .18759 1.5534[.130]
dLRXR5 .20919 .17728 1.1800[.247]
dLRULC .58344 .27632 2.1115[.043]
dLRULC1 -.85260 .27249 -3.1289[.004]
dLRULC2 -.58834 .29136 -2.0193[.052]
dLRULC3 -.66282 .27904 -2.3753[.024]
dLRULC4 -.49782 .27699 -1.7973[.082]
ecm(-1) -.75821 .23545 -3.2202[.003]
List of additional temporary variables created:
dLRXR = LRXR-LRXR(-1)
dLRXR1 = LRXR(-1)-LRXR(-2)
dLRXR2 = LRXR(-2)-LRXR(-3)
dLRXR3 = LRXR(-3)-LRXR(-4)
dLRXR4 = LRXR(-4)-LRXR(-5)
dLRXR5 = LRXR(-5)-LRXR(-6) 5.3322 3.3146 4.3207 2.5502 3.3986
dLRULC = LRULC-LRULC(-1)
dLRULC1 = LRULC(-1)-LRULC(-2)
dLRULC2 = LRULC(-2)-LRULC(-3) 10.6644 6.6292 8.6414 5.1003 6.7972
dLRULC3 = LRULC(-3)-LRULC(-4)
dLRULC4 = LRULC(-4)-LRULC(-5)
ecm = LRXR -1.0777*LRULC
R-Squared .49758 R-Bar-Squared .31336
S.E. of Regression .18518 F-Stat. F(10,31) 2.9711[.010]
Mean of Dependent Variable .0049869 S.D. of Dependent Variable .22347
Residual Sum of Squares 1.0287 Equation Log-likelihood 18.3014
Akaike Info. Criterion 6.3014 Schwarz Bayesian Criterion -4.1246
DW-statistic 1.9719
Estimated Long Run Coefficients using the ARDL Approach
ARDL(6,5) selected based on R-BAR Squared Criterion
**************************************************************
Dependent variable is LRXR
D:Heteroscedasticity*CHSQ(1) = .69765[.404]*F(1,40) = .67565[.416]
* * * *
**************************************************************
Regressor Coefficient Standard Error T-Ratio[Prob]
LRULC 1.0777 .0089654 120.2063[.000]
* Test Statistics * LM Version * F Version *
**************************************************************
**************************************************************
Diagnostic Tests
* * * *
B:Functional Form *CHSQ(1) = 2.2282[.136]*F(1,29) = 1.6247[.213]
* * * *
* C:Normality *CHSQ(2) = 9.2334[.010]* Not applicable *
F-statistic 95% Lower Bound 95% Upper Bound 90% Lower Bound 90% Upper Bound
A:Serial Correlation*CHSQ(1) = .016740[.897]*F(1,29) = .011563[.915]
****************************************************************
****************************************************************
****************************************************************
Error Correction Representation for the Selected ARDL Model
ARDL(6,5) selected based on R-BAR Squared Criterion
****************************************************************
* * * *
W-statistic 95% Lower Bound 95% Upper Bound 90% Lower Bound 90% Upper Bound
************************************************************************
Testing for existence of a level relationship among the variables in the ARDL model
************************************************************************
141
(2) Australia
Dependent variable is dLRXR Dependent variable is LRXR
44 observations used for estimation from 1967 to 2010 44 observations used for estimation from 1967 to 2010
Regressor Coefficient Standard Error T-Ratio[Prob] Regressor Coefficient Standard Error T-Ratio[Prob]
dLRXR1 .26570 .14910 1.7821[.084] LRULC 2.1167 .39574 5.3487[.000]
dLRULC .62238 .35315 1.7623[.087] RIDIFF -.043949 .017472 -2.5155[.017]
dLRULC1 -.10651 .37538 -.28374[.778] INPT -5.0142 1.8156 -2.7617[.010]
dLRULC2 -.70736 .39144 -1.8071[.080]
dRIDIFF .0083861 .0047129 1.7794[.084]
dRIDIFF1 .026570 .0095796 2.7736[.009]
dRIDIFF2 .025890 .0084936 3.0481[.005]
dRIDIFF3 .024225 .0066552 3.6400[.001]
dRIDIFF4 .0087199 .0050106 1.7403[.091]
ecm(-1) -.51808 .12623 -4.1043[.000]
List of additional temporary variables created:
dLRXR = LRXR-LRXR(-1)
dLRXR1 = LRXR(-1)-LRXR(-2)
dLRULC = LRULC-LRULC(-1)
dLRULC1 = LRULC(-1)-LRULC(-2)
dLRULC2 = LRULC(-2)-LRULC(-3)
dRIDIFF = RIDIFF-RIDIFF(-1) Testing for existence of a level relationship among the variables in the ARDL model
dRIDIFF1 = RIDIFF(-1)-RIDIFF(-2)
dRIDIFF2 = RIDIFF(-2)-RIDIFF(-3) F-statistic 95% Lower Bound 95% Upper Bound 90% Lower Bound 90% Upper Bound
dRIDIFF3 = RIDIFF(-3)-RIDIFF(-4) 6.6668 4.0902 5.2618 3.3274 4.3578
dRIDIFF4 = RIDIFF(-4)-RIDIFF(-5)
ecm = LRXR -2.1167*LRULC + .043949*RIDIFF + 5.0142*INPT W-statistic 95% Lower Bound 95% Upper Bound 90% Lower Bound 90% Upper Bound
20.0004 12.2706 15.7855 9.9822 13.0733
R-Squared .55987 R-Bar-Squared .38949
S.E. of Regression .064478 F-Stat. F(10,33) 3.9433[.001]
Mean of Dependent Variable .014770 S.D. of Dependent Variable .082521
Residual Sum of Squares .12888 Equation Log-likelihood 65.8943
Akaike Info. Criterion 52.8943 Schwarz Bayesian Criterion 41.2971
DW-statistic 2.2916
**************************************************************
****************************************************************
*************************************************************************
B:Functional Form *CHSQ(1) = 5.9573[.015]*F(1,30) = 4.6978[.038]
* * * *
C:Normality *CHSQ(2) = .49038[.783]* Not applicable
* * * *
D:Heteroscedasticity*CHSQ(1) = 3.3458[.067]*F(1,42) = 3.4566[.070]
*************************************************************************
**************************************************************
**************************************************************
* Test Statistics * LM Version * F Version *
* * * *
* * * *
A:Serial Correlation*CHSQ(1) = 5.5841[.018]*F(1,30) = 4.3608[.045]
****************************************************************
Error Correction Representation for the Selected ARDL Model
ARDL(2,3,5) selected based on R-BAR Squared Criterion
Estimated Long Run Coefficients using the ARDL Approach
ARDL(2,3,5) selected based on R-BAR Squared Criterion
**************************************************************
****************************************************************
****************************************************************
Diagnostic Tests
(3) Belgium
Dependent variable is dLRXR Dependent variable is LRXR
31 observations used for estimation from 1980 to 2010 31 observations used for estimation from 1980 to 2010
Regressor Coefficient Standard Error T-Ratio[Prob] Regressor Coefficient Standard Error T-Ratio[Prob]
dLRULC .48023 .079839 6.0150[.000] LRULC .91101 .016956 53.7274[.000]
dRIDIFF .0058872 .0016579 3.5511[.001] RIDIFF .010786 .0054136 1.9925[.057]
dRIDIFF1 -.0024298 .0016571 -1.4663[.155] T .011174 .0020386 5.4813[.000]
dT .0058903 .0010178 5.7871[.000]
ecm(-1) -.52715 .082453 -6.3933[.000]
List of additional temporary variables created:
dLRXR = LRXR-LRXR(-1)
dLRULC = LRULC-LRULC(-1)
dRIDIFF = RIDIFF-RIDIFF(-1)
dRIDIFF1 = RIDIFF(-1)-RIDIFF(-2)
dT = T-T(-1)
ecm = LRXR -.91101*LRULC -.010786*RIDIFF -.011174*T
R-Squared .76013 R-Bar-Squared .71215
S.E. of Regression .035385 F-Stat. F(4,26) 19.8056[.000]
Mean of Dependent Variable -.0047730 S.D. of Dependent Variable .065954
Residual Sum of Squares .031303 Equation Log-likelihood 62.9325
Akaike Info. Criterion 56.9325 Schwarz Bayesian Criterion 52.6305
DW-statistic 1.4089 F-statistic 95% Lower Bound 95% Upper Bound 90% Lower Bound 90% Upper Bound
17.7332 4.2593 5.4149 3.4573 4.4664
W-statistic 95% Lower Bound 95% Upper Bound 90% Lower Bound 90% Upper Bound
53.1995 12.7779 16.2446 10.3720 13.3991
D:Heteroscedasticity*CHSQ(1) = .36322[.547]*F(1,29) = .34382[.562]
Testing for existence of a level relationship among the variables in the ARDL model
**************************************************************************
**************************************************************************
A:Serial Correlation*CHSQ(1) = 2.6452[.104]*F(1,24) = 2.2389[.148]
* * * *
B:Functional Form *CHSQ(1) = 4.4148[.036]*F(1,24) = 3.9855[.057]
* * * *
C:Normality *CHSQ(2) = .97586[.614]* Not applicable
* * * *
**************************************************************
**************************************************************
**************************************************************
Diagnostic Tests
* Test Statistics * LM Version * F Version *
* * * *
****************************************************************
****************************************************************
****************************************************************
****************************************************************
Error Correction Representation for the Selected ARDL Model
ARDL(1,0,2) selected based on Akaike Information Criterion
Estimated Long Run Coefficients using the ARDL Approach
ARDL(1,0,2) selected based on Akaike Information Criterion
**************************************************************
142
(4) Canada
Dependent variable is dLRXR Dependent variable is LRXR
27 observations used for estimation from 1984 to 2010 27 observations used for estimation from 1984 to 2010
Regressor Coefficient Standard Error T-Ratio[Prob] Regressor Coefficient Standard Error T-Ratio[Prob]
dLRXR1 -.44182 .18345 -2.4084[.029] LRULC 1.0060 .013063 77.0078[.000]
dLRULC .21813 .38971 .55973[.584] RIDIFF -.22945 .10203 -2.2489[.042]
dLRULC1 .57853 .39567 1.4622[.164]
dLRULC2 .38615 .35591 1.0850[.295]
dLRULC3 .77113 .36684 2.1021[.053]
dRIDIFF .017058 .0072385 2.3565[.032]
dRIDIFF1 .067051 .012973 5.1685[.000]
dRIDIFF2 .035111 .0095352 3.6823[.002]
dRIDIFF3 .024035 .0062534 3.8435[.002]
dRIDIFF4 .028331 .0061105 4.6365[.000]
dRIDIFF5 .019216 .0062076 3.0956[.007]
ecm(-1) -.30239 .11137 -2.7153[.016]
List of additional temporary variables created:
dLRXR = LRXR-LRXR(-1)
dLRXR1 = LRXR(-1)-LRXR(-2)
dLRULC = LRULC-LRULC(-1)
dLRULC1 = LRULC(-1)-LRULC(-2)
dLRULC2 = LRULC(-2)-LRULC(-3) F-statistic 95% Lower Bound 95% Upper Bound 90% Lower Bound 90% Upper Bound
dLRULC3 = LRULC(-3)-LRULC(-4) 10.2300 3.1045 4.3195 2.3996 3.4815
dRIDIFF = RIDIFF-RIDIFF(-1)
dRIDIFF1 = RIDIFF(-1)-RIDIFF(-2) W-statistic 95% Lower Bound 95% Upper Bound 90% Lower Bound 90% Upper Bound
dRIDIFF2 = RIDIFF(-2)-RIDIFF(-3) 30.6901 9.3134 12.9586 7.1989 10.4444
dRIDIFF3 = RIDIFF(-3)-RIDIFF(-4)
dRIDIFF4 = RIDIFF(-4)-RIDIFF(-5)
dRIDIFF5 = RIDIFF(-5)-RIDIFF(-6)
ecm = LRXR -1.0060*LRULC + .22945*RIDIFF
R-Squared .86037 R-Bar-Squared .72075
S.E. of Regression .034879 F-Stat. F(11,15) 7.2823[.000]
Mean of Dependent Variable -.4651E-3 S.D. of Dependent Variable .066003
Residual Sum of Squares .015815 Equation Log-likelihood 62.1641
Akaike Info. Criterion 48.1641 Schwarz Bayesian Criterion 39.0932
DW-statistic 2.4432
**************************************************************************
****************************************************************
* Test Statistics * LM Version * F Version *
* * * *
**************************************************************************
Diagnostic Tests
**************************************************************
**************************************************************
C:Normality *CHSQ(2) = .50617[.776]* Not applicable
* * * *
D:Heteroscedasticity*CHSQ(1) = .8995E-3[.976]*F(1,25) = .8329E-3[.977]
Testing for existence of a level relationship among the variables in the ARDL model
****************************************************************
****************************************************************
****************************************************************
A:Serial Correlation*CHSQ(1) = 3.1367[.077]*F(1,12) = 1.5773[.233]
* * * *
B:Functional Form *CHSQ(1) = 1.1871[.276]*F(1,12) = .55185[.472]
* * * *
**************************************************************
**************************************************************
Error Correction Representation for the Selected ARDL Model
ARDL(2,4,6) selected based on R-BAR Squared Criterion
Estimated Long Run Coefficients using the ARDL Approach
ARDL(2,4,6) selected based on R-BAR Squared Criterion
(5) Denmark
Dependent variable is dLRXR Dependent variable is LRXR
38 observations used for estimation from 1973 to 2010 38 observations used for estimation from 1973 to 2010
Regressor Coefficient Standard Error T-Ratio[Prob] Regressor Coefficient Standard Error T-Ratio[Prob]
dLRULC .50602 .073873 6.8498[.000] LRULC 1.0107 .0067362 150.0445[.000]
dRIDIFF .8609E-3 .0028575 .30128[.765] RIDIFF -.026734 .0081808 -3.2679[.003]
dRIDIFF1 .015953 .0039567 4.0318[.000]
dRIDIFF2 .010783 .0035189 3.0644[.005]
dRIDIFF3 .0056671 .0033878 1.6728[.105]
dRIDIFF4 .0058237 .0031650 1.8401[.076]
dRIDIFF5 .0093513 .0028255 3.3096[.002]
ecm(-1) -.50065 .074386 -6.7304[.000]
List of additional temporary variables created:
dLRXR = LRXR-LRXR(-1)
dLRULC = LRULC-LRULC(-1)
dRIDIFF = RIDIFF-RIDIFF(-1)
dRIDIFF1 = RIDIFF(-1)-RIDIFF(-2)
dRIDIFF2 = RIDIFF(-2)-RIDIFF(-3)
dRIDIFF3 = RIDIFF(-3)-RIDIFF(-4)
dRIDIFF4 = RIDIFF(-4)-RIDIFF(-5)
dRIDIFF5 = RIDIFF(-5)-RIDIFF(-6)
ecm = LRXR -1.0107*LRULC + .026734*RIDIFF F-statistic 95% Lower Bound 95% Upper Bound 90% Lower Bound 90% Upper Bound
9.3503 2.9586 4.1766 2.3163 3.3767
R-Squared .69944 R-Bar-Squared .61652
S.E. of Regression .034073 F-Stat. F(7,30) 9.6408[.000] W-statistic 95% Lower Bound 95% Upper Bound 90% Lower Bound 90% Upper Bound
Mean of Dependent Variable .012408 S.D. of Dependent Variable .055023 28.0508 8.8759 12.5298 6.9488 10.1300
Residual Sum of Squares .033669 Equation Log-likelihood 79.6268
Akaike Info. Criterion 70.6268 Schwarz Bayesian Criterion 63.2577
DW-statistic 2.0401
D:Heteroscedasticity*CHSQ(1) = .35968[.549]*F(1,36) = .34400[.561]
**************************************************************************
**************************************************************************
Testing for existence of a level relationship among the variables in the ARDL model
A:Serial Correlation*CHSQ(1) = .25109[.616]*F(1,28) = .18625[.669]
* * * *
B:Functional Form *CHSQ(1) = .36302[.547]*F(1,28) = .27007[.607]
* * * *
C:Normality *CHSQ(2) = 2.3731[.305]* Not applicable
* * * *
****************************************************************
****************************************************************
****************************************************************
**************************************************************
* Test Statistics * LM Version * F Version *
* * * *
**************************************************************
**************************************************************
Diagnostic Tests
Error Correction Representation for the Selected ARDL Model
ARDL(1,0,6) selected based on R-BAR Squared Criterion
**************************************************************** **************************************************************
Estimated Long Run Coefficients using the ARDL Approach
ARDL(1,0,6) selected based on R-BAR Squared Criterion
143
(6) El Salvador
Dependent variable is dLRXR Dependent variable is LRXR
50 observations used for estimation from 1961 to 2010 50 observations used for estimation from 1961 to 2010
Regressor Coefficient Standard Error T-Ratio[Prob] Regressor Coefficient Standard Error T-Ratio[Prob]
dLRULC -.17388 .093896 -1.8518[.070] LRULC 1.0630 .032679 32.5295[.000]
dD86 -.43729 .047818 -9.1449[.000] D86 -3.9009 1.3806 -2.8254[.007]
ecm(-1) -.077345 .022510 -3.4361[.001]
List of additional temporary variables created:
dLRXR = LRXR-LRXR(-1)
dLRULC = LRULC-LRULC(-1)
dD86 = D86-D86(-1)
ecm = LRXR -1.0630*LRULC + 3.9009*D86
R-Squared .71299 R-Bar-Squared .68748
S.E. of Regression .046335 F-Stat. F(2,47) 55.8947[.000]
Mean of Dependent Variable .014193 S.D. of Dependent Variable .082883
Residual Sum of Squares .096610 Equation Log-likelihood 85.2804
Akaike Info. Criterion 80.2804 Schwarz Bayesian Criterion 75.5003
DW-statistic 1.4150
F-statistic 95% Lower Bound 95% Upper Bound 90% Lower Bound 90% Upper Bound
13.5465 2.8709 4.0302 2.2916 3.3084
W-statistic 95% Lower Bound 95% Upper Bound 90% Lower Bound 90% Upper Bound
40.6394 8.6126 12.0905 6.8748 9.9252
****************************************************************
****************************************************************
D:Heteroscedasticity*CHSQ(1) = .76582[.382]*F(1,48) = .74662[.392]
Testing for existence of a level relationship among the variables in the ARDL model
**************************************************************************
**************************************************************************
A:Serial Correlation*CHSQ(1) = 4.9770[.026]*F(1,44) = 4.8639[.033]
* * * *
B:Functional Form *CHSQ(1) = .064081[.800]*F(1,44) = .056463[.813]
* * * *
C:Normality *CHSQ(2) = 47.5221[.000]* Not applicable
* * * *
**************************************************************** **************************************************************
ARDL(1,1,1) selected based on R-BAR Squared Criterion
**************************************************************
**************************************************************
Diagnostic Tests
* Test Statistics * LM Version * F Version *
* * * *
****************************************************************
Error Correction Representation for the Selected ARDL Model
ARDL(1,1,1) selected based on R-BAR Squared Criterion
**************************************************************
Estimated Long Run Coefficients using the ARDL Approach
(7) Finland
Dependent variable is dLRXR Dependent variable is LRXR
40 observations used for estimation from 1971 to 2010 40 observations used for estimation from 1971 to 2010
Regressor Coefficient Standard Error T-Ratio[Prob] Regressor Coefficient Standard Error T-Ratio[Prob]
dLRXR1 .56262 .23404 2.4039[.024] LRULC 1.0036 .0041384 242.5106[.000]
dLRXR2 -.020238 .22172 -.091280[.928] RIDIFF -.027272 .011870 -2.2976[.031]
dLRXR3 .32550 .18900 1.7223[.097]
dLRXR4 -.16726 .16151 -1.0356[.310]
dLRULC .022051 .28992 .076059[.940]
dLRULC1 -.35466 .32122 -1.1041[.280]
dLRULC2 -.86432 .42513 -2.0331[.052]
dRIDIFF .0015177 .0055102 .27543[.785]
dRIDIFF1 .012699 .0071897 1.7663[.089]
dRIDIFF2 .011612 .0063274 1.8353[.078]
dRIDIFF3 .0084019 .0049729 1.6895[.103]
dRIDIFF4 .012626 .0047002 2.6863[.012]
dRIDIFF5 .0060269 .0037628 1.6017[.121]
ecm(-1) -.47434 .17102 -2.7736[.010]
List of additional temporary variables created:
dLRXR = LRXR-LRXR(-1)
dLRXR1 = LRXR(-1)-LRXR(-2)
dLRXR2 = LRXR(-2)-LRXR(-3)
dLRXR3 = LRXR(-3)-LRXR(-4) F-statistic 95% Lower Bound 95% Upper Bound 90% Lower Bound 90% Upper Bound
dLRXR4 = LRXR(-4)-LRXR(-5) 4.3885 2.9485 4.1225 2.3296 3.3523
dLRULC = LRULC-LRULC(-1)
dLRULC1 = LRULC(-1)-LRULC(-2) W-statistic 95% Lower Bound 95% Upper Bound 90% Lower Bound 90% Upper Bound
dLRULC2 = LRULC(-2)-LRULC(-3) 13.1654 8.8454 12.3675 6.9888 10.0570
dRIDIFF = RIDIFF-RIDIFF(-1)
dRIDIFF1 = RIDIFF(-1)-RIDIFF(-2)
dRIDIFF2 = RIDIFF(-2)-RIDIFF(-3)
dRIDIFF3 = RIDIFF(-3)-RIDIFF(-4)
dRIDIFF4 = RIDIFF(-4)-RIDIFF(-5)
dRIDIFF5 = RIDIFF(-5)-RIDIFF(-6)
ecm = LRXR -1.0036*LRULC + .027272*RIDIFF
R-Squared .58530 R-Bar-Squared .32612
S.E. of Regression .056155 F-Stat. F(13,26) 2.6057[.018]
Mean of Dependent Variable -.0048851 S.D. of Dependent Variable .068406
Residual Sum of Squares .075681 Equation Log-likelihood 68.6446
Akaike Info. Criterion 52.6446 Schwarz Bayesian Criterion 39.1336
DW-statistic 2.0779
****************************************************************
****************************************************************
* * * *
D:Heteroscedasticity*CHSQ(1) = 1.9273[.165]*F(1,38) = 1.9236[.174]
Testing for existence of a level relationship among the variables in the ARDL model
**************************************************************************
**************************************************************************
****************************************************************
****************************************************************
* * * *
A:Serial Correlation*CHSQ(1) = .40198[.526]*F(1,23) = .23349[.634]
* * * *
B:Functional Form *CHSQ(1) = .34464[.557]*F(1,23) = .19989[.659]
* * * *
C:Normality *CHSQ(2) = .57980[.748]* Not applicable
**************************************************************
**************************************************************
Diagnostic Tests
* Test Statistics * LM Version * F Version *
**************************************************************
**************************************************************
Error Correction Representation for the Selected ARDL Model
ARDL(5,3,6) selected based on R-BAR Squared Criterion
Estimated Long Run Coefficients using the ARDL Approach
ARDL(5,3,6) selected based on R-BAR Squared Criterion
144
(8) France
Dependent variable is dLRXR Dependent variable is LRXR
40 observations used for estimation from 1971 to 2010 40 observations used for estimation from 1971 to 2010
Regressor Coefficient Standard Error T-Ratio[Prob] Regressor Coefficient Standard Error T-Ratio[Prob]
dLRXR1 .55208 .17242 3.2019[.004] LRULC 1.0974 .0085758 127.9638[.000]
dLRULC .21613 .32117 .67294[.507] RIDIFF .027787 .013447 2.0664[.050]
dLRULC1 -.17182 .41566 -.41337[.683] T -.0096015 .0011410 -8.4147[.000]
dLRULC2 -.10741 .38142 -.28160[.781]
dLRULC3 -.032503 .34484 -.094256[.926]
dLRULC4 -.69415 .35924 -1.9323[.065]
dLRULC5 -1.1608 .34857 -3.3301[.003]
dRIDIFF .0064686 .0051042 1.2673[.217]
dRIDIFF1 -.011339 .0070582 -1.6065[.121]
dRIDIFF2 -.0071440 .0058709 -1.2168[.235]
dRIDIFF3 -.0088214 .0048473 -1.8199[.081]
dRIDIFF4 -.0064332 .0037576 -1.7121[.099]
dRIDIFF5 .0072018 .0023719 3.0363[.006]
dT -.0051254 .0015472 -3.3128[.003]
ecm(-1) -.53382 .15465 -3.4518[.002]
List of additional temporary variables created:
dLRXR = LRXR-LRXR(-1)
dLRXR1 = LRXR(-1)-LRXR(-2)
dLRULC = LRULC-LRULC(-1) F-statistic 95% Lower Bound 95% Upper Bound 90% Lower Bound 90% Upper Bound
dLRULC1 = LRULC(-1)-LRULC(-2) 4.8798 4.1522 5.3039 3.4142 4.4414
dLRULC2 = LRULC(-2)-LRULC(-3)
dLRULC3 = LRULC(-3)-LRULC(-4) W-statistic 95% Lower Bound 95% Upper Bound 90% Lower Bound 90% Upper Bound
dLRULC4 = LRULC(-4)-LRULC(-5) 14.6393 12.4565 15.9118 10.2425 13.3243
dLRULC5 = LRULC(-5)-LRULC(-6)
dRIDIFF = RIDIFF-RIDIFF(-1)
dRIDIFF1 = RIDIFF(-1)-RIDIFF(-2)
dRIDIFF2 = RIDIFF(-2)-RIDIFF(-3)
dRIDIFF3 = RIDIFF(-3)-RIDIFF(-4)
dRIDIFF4 = RIDIFF(-4)-RIDIFF(-5)
dRIDIFF5 = RIDIFF(-5)-RIDIFF(-6)
dT = T-T(-1)
ecm = LRXR -1.0974*LRULC -.027787*RIDIFF + .0096015*T
R-Squared .70612 R-Bar-Squared .50169
S.E. of Regression .040534 F-Stat. F(14,25) 3.9474[.001]
Mean of Dependent Variable .0014960 S.D. of Dependent Variable .057421
Residual Sum of Squares .037790 Equation Log-likelihood 82.5344
Akaike Info. Criterion 65.5344 Schwarz Bayesian Criterion 51.1789
DW-statistic 2.5046
* * * *
D:Heteroscedasticity*CHSQ(1) = .14012[.708]*F(1,38) = .13358[.717]
Testing for existence of a level relationship among the variables in the ARDL model
**************************************************************************
**************************************************************************
* * * *
A:Serial Correlation*CHSQ(1) = 4.4563[.035]*F(1,22) = 2.7583[.111]
* * * *
B:Functional Form *CHSQ(1) = 2.9319[.087]*F(1,22) = 1.7401[.201]
* * * *
C:Normality *CHSQ(2) = .30689[.858]* Not applicable
**************************************************************
**************************************************************
ARDL(2,6,6) selected based on R-BAR Squared Criterion
Diagnostic Tests
**************************************************************
**************************************************************
* Test Statistics * LM Version * F Version *
ARDL(2,6,6) selected based on R-BAR Squared Criterion
****************************************************************
****************************************************************
****************************************************************
****************************************************************
Error Correction Representation for the Selected ARDL Model Estimated Long Run Coefficients using the ARDL Approach
(9) Germany
Dependent variable is dLRXR Dependent variable is LRXR
46 observations used for estimation from 1965 to 2010 46 observations used for estimation from 1965 to 2010
Regressor Coefficient Standard Error T-Ratio[Prob] Regressor Coefficient Standard Error T-Ratio[Prob]
dLRXR1 .40819 .15332 2.6624[.012] LRULC 1.0318 .0079782 129.3300[.000]
dLRXR2 -.058229 .16770 -.34722[.730]
dLRXR3 .11046 .16062 .68766[.496]
dLRXR4 .27257 .15956 1.7083[.096]
dLRULC .48454 .23040 2.1031[.043]
dLRULC1 -.68541 .23515 -2.9148[.006]
dLRULC2 -.20137 .32061 -.62807[.534]
dLRULC3 .012579 .31336 .040142[.968]
dLRULC4 -.89557 .30219 -2.9636[.005]
ecm(-1) -.31473 .10687 -2.9450[.006]
List of additional temporary variables created:
dLRXR = LRXR-LRXR(-1)
dLRXR1 = LRXR(-1)-LRXR(-2)
dLRXR2 = LRXR(-2)-LRXR(-3)
dLRXR3 = LRXR(-3)-LRXR(-4)
dLRXR4 = LRXR(-4)-LRXR(-5)
dLRULC = LRULC-LRULC(-1) F-statistic 95% Lower Bound 95% Upper Bound 90% Lower Bound 90% Upper Bound
dLRULC1 = LRULC(-1)-LRULC(-2) 6.1673 3.2641 4.2781 2.4777 3.3540
dLRULC2 = LRULC(-2)-LRULC(-3)
dLRULC3 = LRULC(-3)-LRULC(-4) W-statistic 95% Lower Bound 95% Upper Bound 90% Lower Bound 90% Upper Bound
dLRULC4 = LRULC(-4)-LRULC(-5) 12.3346 6.5282 8.5561 4.9555 6.7080
ecm = LRXR -1.0318*LRULC
R-Squared .50431 R-Bar-Squared .36269
S.E. of Regression .045984 F-Stat. F(9,36) 3.9566[.001]
Mean of Dependent Variable .011365 S.D. of Dependent Variable .057602
Residual Sum of Squares .074010 Equation Log-likelihood 82.6693
Akaike Info. Criterion 71.6693 Schwarz Bayesian Criterion 61.6118
DW-statistic 1.9304
Error Correction Representation for the Selected ARDL Model Estimated Long Run Coefficients using the ARDL Approach
ARDL(5,5) selected based on R-BAR Squared Criterion ARDL(5,5) selected based on R-BAR Squared Criterion
**************************************************************** *************************************************************
**************************************************************** *************************************************************
Diagnostic Tests
*************************************************************
* Test Statistics * LM Version * F Version *
*************************************************************
* * * *
A:Serial Correlation*CHSQ(1) = .19922[.655]*F(1,34) = .14789[.703]
* * * *
B:Functional Form *CHSQ(1) = 1.5247[.217]*F(1,34) = 1.1656[.288]
**************************************************************** * * * *
C:Normality *CHSQ(2) = .74424[.689]* Not applicable
* * * *
D:Heteroscedasticity*CHSQ(1) = .11744[.732]*F(1,44) = .11262[.739]
Testing for existence of a level relationship among the variables in the ARDL model
*************************************************************************
*************************************************************************
****************************************************************
145
(10) Italy
Dependent variable is dLRXR Dependent variable is LRXR
30 observations used for estimation from 1981 to 2010 30 observations used for estimation from 1981 to 2010
Regressor Coefficient Standard Error T-Ratio[Prob] Regressor Coefficient Standard Error T-Ratio[Prob]
dLRXR1 .53608 .19226 2.7882[.011] LRULC .75821 .12648 5.9946[.000]
dLRXR2 .27553 .16971 1.6235[.119] RIDIFF .018644 .0086000 2.1678[.042]
dLRXR3 -.083448 .16019 -.52094[.608] INPT 1.1051 .57636 1.9173[.069]
dLRXR4 .38807 .17948 2.1622[.042]
dLRULC .65781 .14603 4.5047[.000]
dRIDIFF .029178 .0085707 3.4044[.003]
ecm(-1) -.86759 .19711 -4.4016[.000]
List of additional temporary variables created:
dLRXR = LRXR-LRXR(-1)
dLRXR1 = LRXR(-1)-LRXR(-2)
dLRXR2 = LRXR(-2)-LRXR(-3)
dLRXR3 = LRXR(-3)-LRXR(-4)
dLRXR4 = LRXR(-4)-LRXR(-5)
dLRULC = LRULC-LRULC(-1)
dRIDIFF = RIDIFF-RIDIFF(-1)
ecm = LRXR -.75821*LRULC -.018644*RIDIFF -1.1051*INPT
R-Squared .65323 R-Bar-Squared .52113
S.E. of Regression .041452 F-Stat. F(7,22) 5.6513[.001] F-statistic 95% Lower Bound 95% Upper Bound 90% Lower Bound 90% Upper Bound
Mean of Dependent Variable .0074910 S.D. of Dependent Variable .059901 5.4961 4.2795 5.4490 3.4505 4.4929
Residual Sum of Squares .036083 Equation Log-likelihood 58.2786
Akaike Info. Criterion 49.2786 Schwarz Bayesian Criterion 42.9732 W-statistic 95% Lower Bound 95% Upper Bound 90% Lower Bound 90% Upper Bound
DW-statistic 2.0061 16.4884 12.8386 16.3471 10.3515 13.4787
**************************************************************** *************************************************************
* * * *
Estimated Long Run Coefficients using the ARDL Approach
ARDL(5,0,1) selected based on R-BAR Squared Criterion
**************************************************************** *************************************************************
**************************************************************** *************************************************************
* * * *
C:Normality *CHSQ(2) = 1.1390[.566]* Not applicable
* * * *
Diagnostic Tests
*************************************************************
* Test Statistics * LM Version * F Version *
D:Heteroscedasticity*CHSQ(1) = .65600[.418]*F(1,28) = .62596[.435]
**************************************************************** Testing for existence of a level relationship among the variables in the ARDL model
*************************************************************************
*************************************************************************
Error Correction Representation for the Selected ARDL Model
ARDL(5,0,1) selected based on R-BAR Squared Criterion
A:Serial Correlation*CHSQ(1) = .0078104[.930]*F(1,20) = .0052083[.943]
* * * *
B:Functional Form *CHSQ(1) = .54331[.461]*F(1,20) = .36889[.550]
(11) Japan
Dependent variable is dLRXR Dependent variable is LRXR
34 observations used for estimation from 1977 to 2010 34 observations used for estimation from 1977 to 2010
Regressor Coefficient Standard Error T-Ratio[Prob] Regressor Coefficient Standard Error T-Ratio[Prob]
dLRXR1 .48341 .20986 2.3035[.030] LRULC 1.0127 .0033600 301.3871[.000]
dLRXR2 .48139 .19104 2.5198[.018] RIDIFF -.067310 .015429 -4.3626[.000]
dLRULC 1.7631 .42071 4.1908[.000]
dLRULC1 .11893 .48615 .24463[.809]
dLRULC2 .93748 .47710 1.9649[.060]
dRIDIFF -.0058566 .010096 -.58010[.567]
dRIDIFF1 .016718 .0086369 1.9357[.064]
ecm(-1) -.83590 .17764 -4.7056[.000]
List of additional temporary variables created:
dLRXR = LRXR-LRXR(-1)
dLRXR1 = LRXR(-1)-LRXR(-2)
dLRXR2 = LRXR(-2)-LRXR(-3)
dLRULC = LRULC-LRULC(-1)
dLRULC1 = LRULC(-1)-LRULC(-2)
dLRULC2 = LRULC(-2)-LRULC(-3)
dRIDIFF = RIDIFF-RIDIFF(-1)
dRIDIFF1 = RIDIFF(-1)-RIDIFF(-2)
ecm = LRXR -1.0127*LRULC + .067310*RIDIFF F-statistic 95% Lower Bound 95% Upper Bound 90% Lower Bound 90% Upper Bound
7.4704 3.0194 4.2314 2.3560 3.4172
R-Squared .66893 R-Bar-Squared .54478
S.E. of Regression .071397 F-Stat. F(7,26) 6.9274[.000] W-statistic 95% Lower Bound 95% Upper Bound 90% Lower Bound 90% Upper Bound
Mean of Dependent Variable .0036292 S.D. of Dependent Variable .10582 22.4113 9.0581 12.6942 7.0680 10.2515
Residual Sum of Squares .12234 Equation Log-likelihood 47.4203
Akaike Info. Criterion 37.4203 Schwarz Bayesian Criterion 29.7885
DW-statistic 2.2650
Error Correction Representation for the Selected ARDL Model
ARDL(3,3,2) selected based on Akaike Information Criterion
****************************************************************
****************************************************************
****************************************************************
****************************************************************
Estimated Long Run Coefficients using the ARDL Approach
ARDL(3,3,2) selected based on Akaike Information Criterion
*************************************************************
*************************************************************
Diagnostic Tests
*************************************************************
*************************************************************
* Test Statistics * LM Version * F Version *
* * * *
A:Serial Correlation*CHSQ(1) = 1.9929[.158]*F(1,23) = 1.4321[.244]
* * * *
B:Functional Form *CHSQ(1) = .046899[.829]*F(1,23) = .031769[.860]
* * * *
C:Normality *CHSQ(2) = .68256[.711]* Not applicable
* * * *
D:Heteroscedasticity*CHSQ(1) = .0090781[.924]*F(1,32) = .0085464[.927]
Testing for existence of a level relationship among the variables in the ARDL model
*************************************************************************
*************************************************************************
146
(12) Korea
Dependent variable is LRXR
46 observations used for estimation from 1965 to 2010
Regressor Coefficient Standard Error T-Ratio[Prob]
LRULC .26625 .12045 2.2104[.034]
RIDIFF .010714 .0050804 2.1089[.042]
INPT 3.6918 .48425 7.6237[.000]
T -.0080882 .0022253 -3.6346[.001]
F-statistic 95% Lower Bound 95% Upper Bound 90% Lower Bound 90% Upper Bound
7.3644 5.2980 6.3544 4.4217 5.4057
W-statistic 95% Lower Bound 95% Upper Bound 90% Lower Bound 90% Upper Bound
22.0931 15.8939 19.0631 13.2650 16.2170
Error Correction Representation for the Selected ARDL Model
ARDL(1,4,1) selected based on R-BAR Squared Criterion
************************************************************************
Dependent variable is dLRXR
46 observations used for estimation from 1965 to 2010
Estimated Long Run Coefficients using the ARDL Approach
ARDL(1,4,1) selected based on R-BAR Squared Criterion
*************************************************************
************************************************************************
Regressor Coefficient Standard Error T-Ratio[Prob]
dLRULC .51550 .22220 2.3200[.026]
dLRULC1 .21920 .20272 1.0813[.286]
dLRULC2 -.10162 .15085 -.67363[.505]
dLRULC3 .26636 .13940 1.9108[.064]
dRIDIFF .0046282 .0030775 1.5039[.141]
dINPT 2.3238 .58458 3.9751[.000]
dT -.0050911 .0018248 -2.7900[.008]
ecm(-1) -.62945 .13995 -4.4977[.000]
************************************************************************
List of additional temporary variables created:
dLRXR = LRXR-LRXR(-1) dLRULC = LRULC-LRULC(-1) dLRULC1 = LRULC(-1)-LRULC(-2)
dLRULC2 = LRULC(-2)-LRULC(-3) dLRULC3 = LRULC(-3)-LRULC(-4)
dRIDIFF = RIDIFF-RIDIFF(-1) dINPT = INPT-INPT(-1) dT = T-T(-1)
ecm = LRXR -.26625*LRULC -.010714*RIDIFF -3.6918*INPT + .0080882*T
************************************************************************
R-Squared .62619 R-Bar-Squared .53274
S.E. of Regression .067136 F-stat. F( 7, 38) 8.6152[.000]
Mean of Dependent Variable -.0073507 S.D. of Dependent Variable .098214
Residual Sum of Squares .16226 Equation Log-likelihood 64.6143
Akaike Info. Criterion 54.6143 Schwarz Bayesian Criterion 45.4711
DW-statistic 1.6793
*************************************************************************
*************************************************************
Diagnostic Tests
***************************************************************
***************************************************************
* Test Statistics * LM Version * F Version
* * * *
A:Serial Correlation*CHSQ( 1)= 3.2497[.071]*F( 1, 35)= 2.6606[.112]
* * * *
B:Functional Form *CHSQ( 1)= .037830[.846]*F( 1, 35)= .028807[.866]
* * * *
C:Normality *CHSQ( 2)= .48002[.787]* Not applicable
* * *
D:Heteroscedasticity*CHSQ( 1)= .90676[.341]*F( 1, 44)= .88477[.352]
Testing for existence of a level relationship among the variables in the ARDL model
*************************************************************************
(13) Mexico
Dependent variable is dLRXR Dependent variable is LRXR
30 observations used for estimation from 1982 to 2011 30 observations used for estimation from 1982 to 2011
Regressor Coefficient Standard Error T-Ratio[Prob] Regressor Coefficient Standard Error T-Ratio[Prob]
dLRULC .12161 .041650 2.9197[.008] LRULC .17969 .052885 3.3977[.002]
dDIFF .0021185 .0011547 1.8346[.079] DIFF .0031304 .0016530 1.8938[.070]
dINPT 2.6413 .47614 5.5474[.000] INPT 3.9029 .26540 14.7060[.000]
dD86 -.23659 .081241 -2.9122[.008] D86 -.34959 .13820 -2.5296[.018]
dD95 -.29034 .081283 -3.5720[.002] D95 -.42902 .15297 -2.8045[.010]
ecm(-1) -.67676 .11462 -5.9043[.000]
List of additional temporary variables created: **************************************************************
dLRXR = LRXR-LRXR(-1)
dLRULC = LRULC-LRULC(-1) **************************************************************
dDIFF = DIFF-DIFF(-1)
dINPT = INPT-INPT(-1)
dD86 = D86-D86(-1)
dD95 = D95-D95(-1)
ecm = LRXR -.17969*LRULC -.0031304*DIFF -3.9029*INPT + .34959*D86 + .
42902*D95
R-Squared .75097 R-Bar-Squared .69909
S.E. of Regression .074843 F-stat. F( 5, 24) 14.4749[.000]
Mean of Dependent Variable -.0063348 S.D. of Dependent Variable .13644
Residual Sum of Squares .13444 Equation Log-likelihood 38.5499
Akaike Info. Criterion 32.5499 Schwarz Bayesian Criterion 28.3463 F-statistic 95% Lower Bound 95% Upper Bound 90% Lower Bound 90% Upper Bound
DW-statistic 1.5031 7.3264 4.2418 5.4228 3.3965 4.4090
W-statistic 95% Lower Bound 95% Upper Bound 90% Lower Bound 90% Upper Bound
21.9792 12.7255 16.2683 10.1894 13.2271
Error Correction Representation for the Selected ARDL Model
ARDL(1,0,0) selected based on Akaike Information Criterion
****************************************************************
****************************************************************
*************************************************************
*************************************************************
Estimated Long Run Coefficients using the ARDL Approach
ARDL(1,0,0) selected based on Akaike Information Criterion
****************************************************************
****************************************************************
* * * *
* * * *
* * * *
* * * *
Diagnostic Tests
* Test Statistics * LM Version * F Version
A:Serial Correlation*CHSQ( 1)= 1.9109[.167]*F( 1, 23)= 1.5647[.224]
B:Functional Form *CHSQ( 1)= .76298[.382]*F( 1, 23)= .60022[.446]
C:Normality *CHSQ( 2)= .71065[.701]* Not applicable
D:Heteroscedasticity*CHSQ( 1)= .13467[.714]*F( 1, 28)= .12626[.725]
Testing for existence of a level relationship among the variables in the ARDL model
*************************************************************************
*************************************************************************
147
(14) Netherlands
Dependent variable is dLRXR Dependent variable is LRXR
21 observations used for estimation from 1990 to 2010 21 observations used for estimation from 1990 to 2010
Regressor Coefficient Standard Error T-Ratio[Prob] Regressor Coefficient Standard Error T-Ratio[Prob]
dLRXR1 .34769 .21517 1.6159[.145] LRULC .91858 .014446 63.5856[.000]
dLRXR2 .42197 .23583 1.7894[.111] RIDIFF .038162 .013898 2.7459[.033]
dLRXR3 .77366 .22184 3.4874[.008] T .0095382 .0016177 5.8960[.001]
dLRULC -2.4747 .76576 -3.2317[.012]
dLRULC1 -2.5345 .68870 -3.6801[.006]
dLRULC2 -1.0432 .66354 -1.5721[.155]
dLRULC3 -2.2838 .68239 -3.3468[.010]
dRIDIFF .027913 .0084471 3.3044[.011]
dRIDIFF1 .0085820 .0098869 .86802[.411]
dRIDIFF2 .0050687 .0071135 .71254[.496]
dRIDIFF3 .0092427 .0041720 2.2154[.058]
dT .0091996 .0028638 3.2124[.012]
ecm(-1) -.96451 .21907 -4.4026[.002]
List of additional temporary variables created:
dLRXR = LRXR-LRXR(-1)
dLRXR1 = LRXR(-1)-LRXR(-2)
dLRXR2 = LRXR(-2)-LRXR(-3)
dLRXR3 = LRXR(-3)-LRXR(-4)
dLRULC = LRULC-LRULC(-1) F-statistic 95% Lower Bound 95% Upper Bound 90% Lower Bound 90% Upper Bound
dLRULC1 = LRULC(-1)-LRULC(-2) 6.7875 4.5911 5.9289 3.6824 4.7904
dLRULC2 = LRULC(-2)-LRULC(-3)
dLRULC3 = LRULC(-3)-LRULC(-4) W-statistic 95% Lower Bound 95% Upper Bound 90% Lower Bound 90% Upper Bound
dRIDIFF = RIDIFF-RIDIFF(-1) 20.3624 13.7734 17.7866 11.0473 14.3713
dRIDIFF1 = RIDIFF(-1)-RIDIFF(-2)
dRIDIFF2 = RIDIFF(-2)-RIDIFF(-3)
dRIDIFF3 = RIDIFF(-3)-RIDIFF(-4)
dT = T-T(-1)
ecm = LRXR -.91858*LRULC -.038162*RIDIFF -.0095382*T
R-Squared .85337 R-Bar-Squared .51123
S.E. of Regression .031858 F-Stat. F(12,8) 2.9100[.069]
Mean of Dependent Variable .0082732 S.D. of Dependent Variable .045569
Residual Sum of Squares .0060896 Equation Log-likelihood 55.7321
Akaike Info. Criterion 40.7321 Schwarz Bayesian Criterion 32.8982
DW-statistic 2.1301
Error Correction Representation for the Selected ARDL Model
****************************************************************
****************************************************************
****************************************************************
****************************************************************
ARDL(4,4,4) selected based on R-BAR Squared Criterion
Estimated Long Run Coefficients using the ARDL Approach
ARDL(4,4,4) selected based on R-BAR Squared Criterion
*************************************************************
*************************************************************
Diagnostic Tests
*************************************************************
*************************************************************
* * * *
* * * *
* * * *
* * * *
* Test Statistics * LM Version * F Version *
A:Serial Correlation*CHSQ(1) = .50818[.476]*F(1,5) = .12400[.739]
B:Functional Form *CHSQ(1) = .53173[.466]*F(1,5) = .12989[.733]
C:Normality *CHSQ(2) = .64947[.723]* Not applicable
D:Heteroscedasticity*CHSQ(1) = .0082093[.928]*F(1,19) = .0074304[.932]
Testing for existence of a level relationship among the variables in the ARDL model
*************************************************************************
*************************************************************************
(15) Norway
Dependent variable is dLRXR Dependent variable is LRXR
43 observations used for estimation from 1968 to 2010 43 observations used for estimation from 1968 to 2010
Regressor Coefficient Standard Error T-Ratio[Prob] Regressor Coefficient Standard Error T-Ratio[Prob]
dLRULC .57224 .20933 2.7337[.010] LRULC .95526 .026719 35.7526[.000]
dLRULC1 .22521 .19421 1.1596[.254] RIDIFF -.094224 .037123 -2.5382[.016]
dLRULC2 -.27905 .20521 -1.3598[.183] T .0086388 .0035585 2.4276[.021]
dRIDIFF -.0041701 .0025335 -1.6460[.109]
dRIDIFF1 .0074972 .0034887 2.1490[.039]
dRIDIFF2 .0070926 .0028787 2.4639[.019]
dRIDIFF3 .0069666 .0023156 3.0086[.005]
dT .0013155 .6410E-3 2.0522[.048]
ecm(-1) -.15228 .060662 -2.5103[.017]
List of additional temporary variables created:
dLRXR = LRXR-LRXR(-1)
dLRULC = LRULC-LRULC(-1)
dLRULC1 = LRULC(-1)-LRULC(-2)
dLRULC2 = LRULC(-2)-LRULC(-3)
dRIDIFF = RIDIFF-RIDIFF(-1)
dRIDIFF1 = RIDIFF(-1)-RIDIFF(-2)
dRIDIFF2 = RIDIFF(-2)-RIDIFF(-3)
dRIDIFF3 = RIDIFF(-3)-RIDIFF(-4)
dT = T-T(-1) F-statistic 95% Lower Bound 95% Upper Bound 90% Lower Bound 90% Upper Bound
ecm = LRXR -.95526*LRULC + .094224*RIDIFF -.0086388*T 5.1186 4.0981 5.3072 3.3817 4.4189
R-Squared .43834 R-Bar-Squared .26283 W-statistic 95% Lower Bound 95% Upper Bound 90% Lower Bound 90% Upper Bound
S.E. of Regression .036659 F-Stat. F(8,34) 3.1218[.009] 15.3558 12.2944 15.9217 10.1450 13.2568
Mean of Dependent Variable .023186 S.D. of Dependent Variable .042697
Residual Sum of Squares .043004 Equation Log-likelihood 87.5003
Akaike Info. Criterion 76.5003 Schwarz Bayesian Criterion 66.8137
DW-statistic 2.1293
Error Correction Representation for the Selected ARDL Model
****************************************************************
ARDL(1,3,4) selected based on R-BAR Squared Criterion
****************************************************************
****************************************************************
****************************************************************
ARDL(1,3,4) selected based on R-BAR Squared Criterion
Estimated Long Run Coefficients using the ARDL Approach
*************************************************************
*************************************************************
Diagnostic Tests
*************************************************************
*************************************************************
* * * *
* * * *
* * * *
* * * *
* Test Statistics * LM Version * F Version *
A:Serial Correlation*CHSQ(1) = .22725[.634]*F(1,31) = .16470[.688]
B:Functional Form *CHSQ(1) = .020956[.885]*F(1,31) = .015115[.903]
C:Normality *CHSQ(2) = 1.7830[.410]* Not applicable
D:Heteroscedasticity*CHSQ(1) = 1.1528[.283]*F(1,41) = 1.1295[.294]
Testing for existence of a level relationship among the variables in the ARDL model
*************************************************************************
*************************************************************************
148
(16) Spain
Dependent variable is dLRXR Dependent variable is LRXR
29 observations used for estimation from 1982 to 2010 29 observations used for estimation from 1982 to 2010
Regressor Coefficient Standard Error T-Ratio[Prob] Regressor Coefficient Standard Error T-Ratio[Prob]
dLRXR1 .048945 .17278 .28327[.780] LRULC 1.0107 .0061572 164.1513[.000]
dLRXR2 -.35103 .20355 -1.7245[.102] RIDIFF -.071146 .033291 -2.1371[.048]
dLRXR3 -.64338 .21039 -3.0580[.007]
dLRULC 1.0877 .37569 2.8951[.010]
dLRULC1 .69388 .39272 1.7668[.094]
dLRULC2 1.3162 .41509 3.1709[.005]
dRIDIFF .0077016 .0074099 1.0394[.312]
dRIDIFF1 .022884 .0080519 2.8421[.011]
dRIDIFF2 .029094 .0076673 3.7945[.001]
dRIDIFF3 .021294 .0061068 3.4870[.003]
ecm(-1) -.32093 .099445 -3.2272[.005]
List of additional temporary variables created:
dLRXR = LRXR-LRXR(-1)
dLRXR1 = LRXR(-1)-LRXR(-2)
dLRXR2 = LRXR(-2)-LRXR(-3)
dLRXR3 = LRXR(-3)-LRXR(-4)
dLRULC = LRULC-LRULC(-1)
dLRULC1 = LRULC(-1)-LRULC(-2) F-statistic 95% Lower Bound 95% Upper Bound 90% Lower Bound 90% Upper Bound
dLRULC2 = LRULC(-2)-LRULC(-3) 7.3369 3.0491 4.2698 2.3849 3.4580
dRIDIFF = RIDIFF-RIDIFF(-1)
dRIDIFF1 = RIDIFF(-1)-RIDIFF(-2) W-statistic 95% Lower Bound 95% Upper Bound 90% Lower Bound 90% Upper Bound
dRIDIFF2 = RIDIFF(-2)-RIDIFF(-3) 22.0107 9.1474 12.8093 7.1546 10.3739
dRIDIFF3 = RIDIFF(-3)-RIDIFF(-4)
ecm = LRXR -1.0107*LRULC + .071146*RIDIFF
R-Squared .73910 R-Bar-Squared .54343
S.E. of Regression .040731 F-Stat. F(10,18) 4.5327[.003]
Mean of Dependent Variable .0090444 S.D. of Dependent Variable .060280
Residual Sum of Squares .026545 Equation Log-likelihood 60.2959
Akaike Info. Criterion 47.2959 Schwarz Bayesian Criterion 38.4085
DW-statistic 2.2928
*************************************************************
* Test Statistics * LM Version * F Version *
Error Correction Representation for the Selected ARDL Model
****************************************************************
****************************************************************
****************************************************************
****************************************************************
ARDL(4,3,4) selected based on R-BAR Squared Criterion
Estimated Long Run Coefficients using the ARDL Approach
ARDL(4,3,4) selected based on R-BAR Squared Criterion
*************************************************************
*************************************************************
Diagnostic Tests
*************************************************************
* * * *
* * * *
* * * *
* * * *
A:Serial Correlation*CHSQ(1) = 1.7042[.192]*F(1,15) = .93649[.349]
B:Functional Form *CHSQ(1) = .38639[.534]*F(1,15) = .20256[.659]
C:Normality *CHSQ(2) = .18738[.911]* Not applicable
D:Heteroscedasticity*CHSQ(1) = 5.4241[.020]*F(1,27) = 6.2119[.019]
Testing for existence of a level relationship among the variables in the ARDL model
*************************************************************************
*************************************************************************
(17) Sweden
Dependent variable is dLRXR Dependent variable is LRXR
48 observations used for estimation from 1963 to 2010 48 observations used for estimation from 1963 to 2010
Regressor Coefficient Standard Error T-Ratio[Prob] Regressor Coefficient Standard Error T-Ratio[Prob]
dLRXR1 .28887 .16350 1.7668[.084] LRULC .99276 .0084742 117.1503[.000]
dLRXR2 -.16652 .15848 -1.0507[.299]
dLRULC .36927 .16959 2.1775[.035]
dLRULC1 -.35034 .17937 -1.9532[.057]
ecm(-1) -.18792 .076849 -2.4453[.019]
List of additional temporary variables created:
dLRXR = LRXR-LRXR(-1)
dLRXR1 = LRXR(-1)-LRXR(-2)
dLRXR2 = LRXR(-2)-LRXR(-3)
dLRULC = LRULC-LRULC(-1)
dLRULC1 = LRULC(-1)-LRULC(-2)
ecm = LRXR -.99276*LRULC
R-Squared .30803 R-Bar-Squared .22566
S.E. of Regression .047049 F-Stat. F(4,43) 4.6742[.003]
Mean of Dependent Variable -.010005 S.D. of Dependent Variabl .053467
Residual Sum of Squares .092972 Equation Log-likelihood 81.8106 F-statistic 95% Lower Bound 95% Upper Bound 90% Lower Bound 90% Upper Bound
Akaike Info. Criterion 75.8106 Schwarz Bayesian Criterion 70.1970 3.6058 3.2643 4.2434 2.4721 3.3695
DW-statistic 1.9417
W-statistic 95% Lower Bound 95% Upper Bound 90% Lower Bound 90% Upper Bound
7.2116 6.5287 8.4868 4.9443 6.7390
Error Correction Representation for the Selected ARDL Model
****************************************************************
****************************************************************
****************************************************************
****************************************************************
ARDL(3,2) selected based on R-BAR Squared Criterion
Estimated Long Run Coefficients using the ARDL Approach
ARDL(3,2) selected based on R-BAR Squared Criterion
*************************************************************
*************************************************************
Diagnostic Tests
*************************************************************
* Test Statistics * LM Version * F Version *
*************************************************************
* * * *
* * * *
* * * *
* * * *
A:Serial Correlation*CHSQ(1) = .10196[.749]*F(1,41) = .087275[.769]
B:Functional Form *CHSQ(1) = 1.0759[.300]*F(1,41) = .94009[.338]
C:Normality *CHSQ(2) = 3.4785[.176]* Not applicable
D:Heteroscedasticity*CHSQ(1) = .80204[.370]*F(1,46) = .78169[.381]
Testing for existence of a level relationship among the variables in the ARDL model
*************************************************************************
*************************************************************************
149
(18) Taiwan
Dependent variable is LRXR
24 observations used for estimation from 1986 to 2009
Regressor Coefficient Standard Error T-Ratio[Prob] Regressor Coefficient Standard Error T-Ratio[Prob]
dLRXR1 -.32527 .14379 -2.2621[.035] LRULC .98642 .0061526 160.3263[.000]
dLRULC .64886 .18205 3.5642[.002]
dLRULC1 .49253 .19238 2.5602[.019]
ecm(-1) -.36361 .063017 -5.7699[.000]
F-statistic 95% Lower Bound 95% Upper Bound 90% Lower Bound 90% Upper Bound
21.0174 3.5606 4.6285 2.6681 3.5459
W-statistic 95% Lower Bound 95% Upper Bound 90% Lower Bound 90% Upper Bound
42.0348 7.1211 9.2571 5.3363 7.0919
Error Correction Representation for the Selected ARDL Model
ARDL(2,2) selected based on Akaike Information Criterion
Estimated Long Run Coefficients using the ARDL Approach
*****************************************************************
Dependent variable is dLRXR
24 observations used for estimation from 1986 to 2009
*****************************************************************
*****************************************************************
List of additional temporary variables created:
dLRXR = LRXR-LRXR(-1)
dLRXR1 = LRXR(-1)-LRXR(-2)
dLRULC = LRULC-LRULC(-1)
dLRULC1 = LRULC(-1)-LRULC(-2)
ecm = LRXR -.98642*LRULC
*****************************************************************
R-Squared .65274 R-Bar-Squared .57964
S.E. of Regression .041287 F-stat. F( 3, 20) 11.9049[.000]
Mean of Dependent Variable -.026046 S.D. of Dependent Variable .063680
Residual Sum of Squares .032388 Equation Log-likelihood 45.2417
Akaike Info. Criterion 40.2417 Schwarz Bayesian Criterion 37.2965
DW-statistic 1.7295
*************************************************************
*************************************************************
ARDL(2,2) selected based on Akaike Information Criterion
Diagnostic Tests
*************************************************************
*************************************************************
* Test Statistics * LM Version * F Version *
* * * *
* * * *
* * * *
* * * *
A:Serial Correlation*CHSQ( 1)= .76738[.381]*F( 1, 18)= .59454[.451]
B:Functional Form *CHSQ( 1)= 2.2999[.129]*F( 1, 18)= 1.9078[.184]
C:Normality *CHSQ( 2)= .75513[.686]* Not applicable
D:Heteroscedasticity*CHSQ( 1)= 2.6700[.102]*F( 1, 22)= 2.7539[.111]
Testing for existence of a level relationship among the variables in the ARDL model
*************************************************************************
*************************************************************************
(19) United Kingdom
Dependent variable is dLRXR Dependent variable is LRXR
36 observations used for estimation from 1975 to 2010 36 observations used for estimation from 1975 to 2010
Regressor Coefficient Standard Error T-Ratio[Prob] Regressor Coefficient Standard Error T-Ratio[Prob]
dLRULC .39797 .061974 6.4216[.000] LRULC .97243 .0048563 200.2407[.000]
dRIDIFF .011914 .0030069 3.9623[.000] RIDIFF .043989 .0080380 5.4726[.000]
dD97 .17009 .052367 3.2481[.003] D97 .64510 .19793 3.2593[.003]
ecm(-1) -.40925 .063965 -6.3981[.000]
List of additional temporary variables created:
dLRXR = LRXR-LRXR(-1)
dLRULC = LRULC-LRULC(-1)
dRIDIFF = RIDIFF-RIDIFF(-1)
dD97 = D97-D97(-1)
ecm = LRXR -.97243*LRULC -.043989*RIDIFF -.64510*D97
R-Squared .64360 R-Bar-Squared .58420
S.E. of Regression .051121 F-Stat. F(3,32) 18.0580[.000]
Mean of Dependent Variable .011382 S.D. of Dependent Variable .079279
Residual Sum of Squares .078402 Equation Log-likelihood 59.2479
Akaike Info. Criterion 53.2479 Schwarz Bayesian Criterion 48.4974
DW-statistic 1.7934
F-statistic 95% Lower Bound 95% Upper Bound 90% Lower Bound 90% Upper Bound
9.7724 2.6865 4.0691 2.1559 3.3558
W-statistic 95% Lower Bound 95% Upper Bound 90% Lower Bound 90% Upper Bound
39.0897 10.7461 16.2765 8.6236 13.4232
Error Correction Representation for the Selected ARDL Model
*****************************************************************
ARDL(1,0,1,1) selected based on Akaike Information Criterion
Estimated Long Run Coefficients using the ARDL Approach
ARDL(1,0,1,1) selected based on Akaike Information Criterion
*************************************************************
*************************************************************
Diagnostic Tests
A:Serial Correlation*CHSQ(1) = .43032[.512]*F(1,29) = .35084[.558]
B:Functional Form *CHSQ(1) = 1.9352[.164]*F(1,29) = 1.6475[.209]
C:Normality *CHSQ(2) = 5.7818[.056]* Not applicable
*****************************************************************
*****************************************************************
*****************************************************************
*************************************************************
* Test Statistics * LM Version * F Version *
*************************************************************
* * * *
* * * *
* * * *
* * * *
D:Heteroscedasticity*CHSQ(1) = .12413[.725]*F(1,34) = .11764[.734]
Testing for existence of a level relationship among the variables in the ARDL model
*************************************************************************
*************************************************************************
150
(20) United States
Dependent variable is LRXR
48 observations used for estimation from 1963 to 2010
Regressor Coefficient Standard Error T-Ratio[Prob] Regressor Coefficient Standard Error T-Ratio[Prob]
dLRXR1 .43041 .12169 3.5369[.001] LRULC .97289 .0046110 210.9938[.000]
dLRULC .40035 .086415 4.6329[.000] RIDIFF .027164 .011186 2.4284[.019]
dRIDIFF .011178 .0042100 2.6551[.011]
ecm(-1) -.41151 .088215 -4.6648[.000]
F-statistic 95% Lower Bound 95% Upper Bound 90% Lower Bound 90% Upper Bound
4.9591 2.8578 4.0571 2.2848 3.3341
W-statistic 95% Lower Bound 95% Upper Bound 90% Lower Bound 90% Upper Bound
14.8774 8.5735 12.1712 6.8544 10.0024
*************************************************************
*************************************************************
*****************************************************************
List of additional temporary variables created:
Estimated Long Run Coefficients using the ARDL Approach
ARDL(2,0,0) selected based on Akaike Information Criterion
Mean of Dependent Variable -.018536 S.D. of Dependent Variable .077844
Residual Sum of Squares .15268 Equation Log-likelihood 69.9061
Akaike Info. Criterion 65.9061 Schwarz Bayesian Criterion 62.1637
DW-statistic 1.7381
dLRXR = LRXR-LRXR(-1)
dLRXR1 = LRXR(-1)-LRXR(-2)
dLRULC = LRULC-LRULC(-1)
dRIDIFF = RIDIFF-RIDIFF(-1)
ecm = LRXR -.97289*LRULC -.027164*RIDIFF
*****************************************************************
R-Squared .46393 R-Bar-Squared .42737
S.E. of Regression .058906 F-stat. F( 3, 44) 12.6927[.000]
Error Correction Representation for the Selected ARDL Model
ARDL(2,0,0) selected based on Akaike Information Criterion
*****************************************************************
Dependent variable is dLRXR
48 observations used for estimation from 1963 to 2010
*****************************************************************
Diagnostic Tests
*************************************************************
*************************************************************
* Test Statistics * LM Version * F Version *
* * * *
* * * *
Testing for existence of a level relationship among the variables in the ARDL model
*************************************************************************
*************************************************************************
* * * *
* * * *
A:Serial Correlation*CHSQ( 1)= 1.8809[.170]*F( 1, 43)= 1.7537[.192]
B:Functional Form *CHSQ( 1)= .028815[.865]*F( 1, 43)= .025829[.873]
C:Normality *CHSQ( 2)= 3.5199[.172]* Not applicable
D:Heteroscedasticity*CHSQ( 1)= .9834E-3[.975]*F( 1, 46)= .9424E-3[.976]
Appendix to Chapter II-A
151
Table II.1
Mex Manufacturing GDP Mex Manufacturing GDP Nominal Wages MEX Real Wages
Years Current Prices Deflactor Constant Prices Thousand of Pesos CPI 1988=100 Thousand of Pesos
Millions of Pesos 1988=100 Millions of Pesos 1970-2011 1988=100
1970 105 0.2 49194 39270 0.22 18001466
1971 118 0.2 51116 43324 0.23 18829829
1972 135 0.2 56098 49634 0.24 20556074
1973 164 0.3 62001 58984 0.27 21795372
1974 216 0.3 65929 78645 0.33 23476466
1975 257 0.4 69225 97378 0.39 25289112
1976 316 0.4 72732 125777 0.45 28201744
1977 441 0.6 75312 160382 0.58 27862755
1978 551 0.7 82682 194927 0.68 28831075
1979 715 0.8 91476 249083 0.80 31170764
1980 989 1.0 98429 324783 1.01 32167345
1981 1326 1.3 104769 447234 1.29 34623548
1982 2033 2.0 101888 674799 2.05 32873852
1983 3772 4.0 93918 980796 4.14 23668594
1984 6618 6.7 98629 1555173 6.86 22683436
1985 11069 10.6 104625 2540625 10.82 23491222
1986 19446 19.6 99119 4276803 20.14 21233761
1987 49551 48.5 102136 9929307 46.69 21264827
1988 91240 100.0 91240 26052302 100 26052302
1989 110229 112.1 98302 35366285 120 29469969
1990 140608 133.9 105045 46116322 152 30341285
1991 178729 164.6 108597 59038517 186 31666744
1992 208365 184.1 113150 73313769 215 34044098
1993 219934 195.7 112386 79694402 236 33718931
1994 245012 209.4 117011 86580024 253 34246697
1995 350156 314.9 111205 96209084 341 28189389
1996 494520 401.4 123208 126268832 459 27531910
1997 615478 454.4 135448 164832219 553 29794952
1998 749293 515.1 145474 208037397 641 32437859
1999 884331 583.4 151591 254612869 748 34052276
2000 1013598 625.6 162011 311239497 819 38017176
2001 1031218 661.6 155857 332952836 871 38234723
2002 1068603 690.2 154822 336543221 915 36795921
2003 1327543 734.8 180659 354548000 956 37078548
2004 1514525 809.5 187087 375052000 1001 37466280
2005 1559631 810.5 192419 397531000 1041 38188853
2006 1832579 911.5 201056 419685000 1079 38905035
2007 1905965 939.7 202819 429846000 1122 38326606
2008 2027255 1009.4 200840 433761000 1179 36790191
2009 1928312 1047.7 184044 409693000 1241 33000659
2010 2199445 1101.2 199724 424723000 1293 32846003
2011 2390254 1143.6 209005 445701000 1337 33332572
Source: Own elaboration based on Instituto Nacional de Estadistica Geografia e Informatica (BIE INEGI, Mexico). Cuentas de
Produccion from 1970-1984; 1980-1993; 1993-2004; 2003-2012.
GDP: deflacted by using the implicit prices of the Manufacturing, 1988=100.
Wages: deflacted by the Consumer Price Index, 1988=100.
152
Table II.2
US Manufacturing GDP US US Manufacturing GDP US Nominal Wages US Real Wages
Years Current Prices Deflactor Constant Prices 1970-2011 CPI 1988=100 Billion of Dollars
Billion of Dollars 1988=100 Billion of Dollars Billion of Dollars 1988=100
1970 236 45.9 513 158.4 32.8 483
1971 249 47.4 526 160.5 34.2 469
1972 274 48.2 568 175.6 35.3 497
1973 303 49.8 607 196.6 37.5 524
1974 318 54.4 585 211.8 41.7 508
1975 337 59.1 571 211.6 45.5 465
1976 387 61.9 625 238 48.1 495
1977 439 64.9 676 266.7 51.2 521
1978 490 68.7 713 300.1 55.1 545
1979 544 74.2 733 335.2 61.4 546
1980 557 81.5 683 356.2 69.7 511
1981 617 87.3 706 387.6 76.8 504
1982 603 90.2 669 385.7 81.6 473
1983 653 91.9 711 400.7 84.2 476
1984 724 93.5 775 445.4 87.8 507
1985 740 94.7 782 468.5 91.0 515
1986 766 95.0 806 480.7 92.6 519
1987 823 97.9 841 558.2 96.0 581
1988 900 100.0 900 595.8 100.0 596
1989 950 102.9 923 616.8 104.8 588
1990 969 105.6 917 628.5 110.5 569
1991 977 107.7 907 636.2 115.1 553
1992 1017 108.7 935 663 118.6 559
1993 1059 109.8 965 684.9 122.1 561
1994 1127 111.3 1013 719.1 125.3 574
1995 1181 112.3 1052 737.3 128.8 572
1996 1209 112.5 1075 745.4 132.6 562
1997 1277 111.3 1147 774.3 135.7 571
1998 1327 109.1 1216 821.7 137.8 596
1999 1368 108.3 1263 854.5 140.8 607
2000 1416 108.9 1299 903.1 145.6 620
2001 1344 108.1 1243 874.1 149.7 584
2002 1356 106.6 1272 857.5 152.1 564
2003 1374 105.3 1306 879.3 155.5 565
2004 1483 105.7 1403 881.9 159.7 552
2005 1569 106.9 1468 900.5 165.1 545
2006 1648 107.5 1533 925.5 170.4 543
2007 1698 107.9 1574 938.6 175.3 536
2008 1629 109.7 1485 941.1 182.0 517
2009 1540 107.3 1435 848.4 181.3 468
2010 1631 108.2 1506 859.5 184.3 466
2011 1732 111.0 1560 904.9 190.1 476
Source: Own elaboration based on U.S. Department of Commerce, Bureau of Economic Analysis (BEA).
GDP: deflacted by using the implicit prices of durable and non-durable goods indices (average), 1988=100.
Wages: deflacted by the Consumer Price Index, 1988=100.
153
Table II.3A
Years Real GDP Real Wages Workers Unit Labor
Thousands Costs
1970 0.049 0.018 1726 631
1971 0.051 0.019 1772 653
1972 0.056 0.021 1831 671
1973 0.062 0.022 1925 677
1974 0.066 0.023 1996 711
1975 0.069 0.025 2002 731
1976 0.073 0.028 2046 793
1977 0.075 0.028 2051 759
1978 0.083 0.029 2133 744
1979 0.091 0.031 2291 781
1980 0.098 0.032 2441 798
1981 0.105 0.035 2557 845
1982 0.102 0.033 2505 808
1983 0.094 0.024 2326 586
1984 0.099 0.023 2374 546
1985 0.105 0.023 2451 550
1986 0.099 0.021 2404 515
1987 0.102 0.021 2430 506
1988 0.091 0.026 3035 867
1989 0.098 0.029 3168 950
1990 0.105 0.030 3275 946
1991 0.109 0.032 3307 964
1992 0.113 0.034 3380 1017
1993 0.112 0.034 3310 993
1994 0.117 0.034 3239 948
1995 0.111 0.028 3067 777
1996 0.123 0.028 3278 733
1997 0.135 0.030 3566 784
1998 0.145 0.032 3773 841
1999 0.152 0.034 3913 879
2000 0.162 0.038 4102 963
2001 0.156 0.038 3899 956
2002 0.155 0.037 3637 864
2003 0.153 0.036 3531 833
2004 0.159 0.0358 3506 790
Source: Own elaboration based on Instituto Nacional de
Estadistica Geografia e Informatica (BIE INEGI, Mexico). Cuentas de
Produccion from 1970-1984; 1980-1993; 1993-2004.
GDP: deflacted by us ing the impl ici t prices of the Manufacturing, 1988=100.
Wages: deflacted by the Consumer Price Index, 1988=100.
Manufacturing Unit Labor Costs in Mexico 1970-2004
Billions of Pesos
154
Table II.3B
Years Real GDP Real Wages Workers Unit Labor
Thousands Costs
1970 0.049 0.018 1726 631
1971 0.051 0.019 1772 653
1972 0.056 0.021 1831 671
1973 0.062 0.022 1925 677
1974 0.066 0.023 1996 711
1975 0.069 0.025 2002 731
1976 0.073 0.028 2046 793
1977 0.075 0.028 2051 759
1978 0.083 0.029 2133 744
1979 0.091 0.031 2291 781
1980 0.098 0.032 2441 798
1981 0.105 0.035 2557 845
1982 0.102 0.033 2505 808
1983 0.094 0.024 2326 586
1984 0.099 0.023 2374 546
1985 0.105 0.023 2451 550
1986 0.099 0.021 2404 515
1987 0.102 0.021 2430 506
1988 0.091 0.026 3035 867
1989 0.098 0.029 3168 950
1990 0.105 0.030 3275 946
1991 0.109 0.032 3307 964
1992 0.113 0.034 3380 1017
1993 0.112 0.034 3310 993
1994 0.117 0.034 3239 948
1995 0.111 0.028 3067 777
1996 0.123 0.028 3278 733
1997 0.135 0.030 3566 784
1998 0.145 0.032 3773 841
1999 0.152 0.034 3913 879
2000 0.162 0.038 4102 963
2001 0.156 0.038 3899 956
2002 0.155 0.037 3637 864
2003 0.181 0.037 4587 941
2004 0.187 0.037 4596 920
2005 0.192 0.038 4607 914
2006 0.201 0.039 4619 894
2007 0.203 0.038 4509 852
2008 0.201 0.037 4312 790
2009 0.184 0.033 3920 703
2010 0.200 0.033 3978 654
2011 0.209 0.033 3994 637
Source: Own elaboration based on Instituto Nacional de
Estadistica Geografia e Informatica (BIE INEGI, Mexico). Cuentas de
Produccion from 1970-1984; 1980-1993; 1993-2004; 2003-2012.
GDP: deflacted by using the implicit prices of the Manufacturing, 1988=100.
Wages: deflacted by the Consumer Price Index, 1988=100.
Manufacturing Unit Labor Costs in Mexico 1970-2011
Billions of Pesos
155
Table II.4
Years Real GDP Real Wages Workers Unit Labor
Thousands Costs
1970 513 483 18377 17285
1971 526 469 17597 15692
1972 568 497 18046 15786
1973 607 524 19051 16429
1974 585 508 19034 16549
1975 571 465 17376 14171
1976 625 495 18050 14295
1977 676 521 18721 14426
1978 713 545 19533 14918
1979 733 546 20010 14912
1980 683 511 19222 14398
1981 706 504 19090 13630
1982 669 473 17699 12511
1983 711 476 17273 11570
1984 775 507 18171 11896
1985 782 515 17995 11858
1986 806 519 17638 11348
1987 841 581 17635 12190
1988 900 596 17955 11884
1989 923 588 17969 11450
1990 917 569 17631 10932
1991 907 553 16951 10330
1992 935 559 16678 9973
1993 965 561 16617 9657
1994 1013 574 16871 9564
1995 1052 572 17143 9328
1996 1075 562 17164 8977
1997 1147 571 17326 8620
1998 1216 596 17246 8455
1999 1263 607 17051 8189
2000 1299 620 16948 8092
2001 1243 584 16121 7574
2002 1272 564 14976 6639
2003 1306 565 14216 6156
2004 1403 552 14024 5522
2005 1468 545 13954 5187
2006 1533 543 13897 4923
2007 1574 536 13609 4629
2008 1485 517 13142 4578
2009 1435 468 11528 3759
2010 1506 466 11238 3479
2011 1560 476 11456 3494
Source: Own elaboration based on U.S. Department of
Commerce, Bureau of Economic Analysis (BEA).
GDP: deflacted by using the implicit prices of durable and
non-durable goods indices (average), 1988=100.
Wages: deflacted by the Consumer Price Index, 1988=100.
Manufacturing Unit Labor Costs in the U.S. 1970-2011
Billions of Dollars
156
Table II.5
Table II.6A
Years Mexico U.S. U.S./Mex RULCI1 1977=100
1970 631 17285 27 144
1971 653 15692 24 126
1972 671 15786 24 124
1973 677 16429 24 128
1974 711 16549 23 122
1975 731 14171 19 102
1976 793 14295 18 95
1977 759 14426 19 100
1978 744 14918 20 106
1979 781 14912 19 100
1980 798 14398 18 95
1981 845 13630 16 85
1982 808 12511 15 81
1983 586 11570 20 104
1984 546 11896 22 115
1985 550 11858 22 113
1986 515 11348 22 116
1987 506 12190 24 127
RealUnit Labor Cost Ratio 1970-1987
Years Mexico U.S. U.S./Mex RULCI2 1988=100
1970 631 17285 27 200
1971 653 15692 24 175
1972 671 15786 24 172
1973 677 16429 24 177
1974 711 16549 23 170
1975 731 14171 19 141
1976 793 14295 18 131
1977 759 14426 19 139
1978 744 14918 20 146
1979 781 14912 19 139
1980 798 14398 18 132
1981 845 13630 16 118
1982 808 12511 15 113
1983 586 11570 20 144
1984 546 11896 22 159
1985 550 11858 22 157
1986 515 11348 22 161
1987 506 12190 24 176
1988 867 11884 14 100
1989 950 11450 12 88
1990 946 10932 12 84
1991 964 10330 11 78
1992 1017 9973 10 72
1993 993 9657 10 71
1994 948 9564 10 74
1995 777 9328 12 87
1996 733 8977 12 89
1997 784 8620 11 80
1998 841 8455 10 73
1999 879 8189 9 68
2000 963 8092 8 61
2001 956 7574 8 58
2002 864 6639 8 56
2003 833 6156 7 54
2004 790 5522 7 51
Relative Unit Labor Costs, 1970-2004
157
Table II.6B
Years Mexico U.S. U.S./Mex RULCI2 1988=100
1970 631 17285 27 200
1971 653 15692 24 175
1972 671 15786 24 172
1973 677 16429 24 177
1974 711 16549 23 170
1975 731 14171 19 141
1976 793 14295 18 131
1977 759 14426 19 139
1978 744 14918 20 146
1979 781 14912 19 139
1980 798 14398 18 132
1981 845 13630 16 118
1982 808 12511 15 113
1983 586 11570 20 144
1984 546 11896 22 159
1985 550 11858 22 157
1986 515 11348 22 161
1987 506 12190 24 176
1988 867 11884 14 100
1989 950 11450 12 88
1990 946 10932 12 84
1991 964 10330 11 78
1992 1017 9973 10 72
1993 993 9657 10 71
1994 948 9564 10 74
1995 777 9328 12 87
1996 733 8977 12 89
1997 784 8620 11 80
1998 841 8455 10 73
1999 879 8189 9 68
2000 963 8092 8 61
2001 956 7574 8 58
2002 864 6639 8 56
2003 941 6156 7 48
2004 920 5522 6 44
2005 914 5187 6 41
2006 894 4923 6 40
2007 852 4629 5 40
2008 790 4578 6 42
2009 703 3759 5 39
2010 654 3479 5 39
2011 637 3494 5 40
Relative Unit Labor Costs, 1970-2011
158
Table II.7
Table II.8
Table II.9
Years Real GDP Real Wages Workers Unit Labor
Thousands Costs
2003 1.3275 0.3545 4587 1225
2004 1.3748 0.3583 4596 1198
2005 1.4140 0.3652 4607 1190
2006 1.4774 0.3720 4619 1163
2007 1.4904 0.3665 4509 1109
2008 1.4758 0.3518 4312 1028
2009 1.3524 0.3156 3920 915
2010 1.4676 0.3141 3978 851
2011 1.5358 0.3187 3994 829
Source: Own elaboration based on Instituto Nacional de
Estadistica Geografia e Informatica (BIE INEGI, Mexico).
GDP: deflacted by using the implicit prices of the Manufacturing, 2003=100.
Wages: deflacted by the Consumer Price Index, 2003=100.
Manufacturing Unit Labor Costs in Mexico 2003-2011
Billions of Pesos
Years Real GDP Real Wages Workers Unit Labor
Thousands Costs*
2003 1374 879 14216 9096
2004 1477 859 14024 8159
2005 1545 848 13954 7664
2006 1614 845 13897 7274
2007 1657 833 13609 6840
2008 1563 804 13142 6764
2009 1510 728 11528 5553
2010 1586 725 11238 5140
2011 1643 740 11456 5163
Source: Own elaboration based on U.S. Department of
Commerce, Bureau of Economic Analys is (BEA).
GDP: deflacted by us ing the impl ici t prices of durable and
non-durable goods indices (average), 2003=100.
Wages : deflacted by the Consumer Price Index, 2003=100.
Manufacturing Unit Labor Costs in the U.S. 2003-2011
Billions of Dollars
Years Mexico U.S. U.S./Mex RULCI3 2003=100
2003 1225 9096 7 100
2004 1198 8159 7 92
2005 1190 7664 6 87
2006 1163 7274 6 84
2007 1109 6840 6 83
2008 1028 6764 7 89
2009 915 5553 6 82
2010 851 5140 6 81
2011 829 5163 6 84
Relative Unit Labor Costs 2003-2011
159
Table II.10
Years Real Exchange Real Exchange
P. per D. Index Mexico U.S. Index 1970=100 Index1 1977=100
1970 0.01 100 100 100 100 94
1971 0.01 100 105 104 99 93
1972 0.01 100 111 108 97 91
1973 0.01 100 124 114 92 86
1974 0.01 100 154 127 83 77
1975 0.01 100 177 139 79 74
1976 0.02 123 204 147 88 83
1977 0.02 181 264 156 107 100
1978 0.02 182 310 168 98 92
1979 0.02 182 366 187 93 87
1980 0.02 184 463 212 84 79
1981 0.02 196 592 234 77 73
1982 0.05 435 941 249 115 107
1983 0.12 960 1900 257 130 121
1984 0.17 1342 3143 268 114 107
1985 0.26 2052 4958 277 115 107
1986 0.61 4863 9233 282 149 139
1987 1.37 10955 21404 293 150 140
1988 2.27 18180 45840 305 121 113
1989 2.46 19694 55012 320 114 107
1990 2.81 22501 69673 337 109 102
1991 3.02 24143 85463 351 99 93
1992 3.09 24756 98716 362 91 85
1993 3.12 24922 108342 372 86 80
1994 3.38 27001 115889 382 89 83
1995 6.42 51352 156450 393 129 121
1996 7.60 60796 210235 404 117 109
1997 7.92 63348 253597 414 103 97
1998 9.14 73085 293991 420 104 98
1999 9.56 76484 342751 429 96 90
2000 9.46 75645 375284 444 89 84
2001 9.34 74740 399181 456 85 80
2002 9.66 77248 419262 464 85 80
2003 10.79 86312 438326 474 93 87
2004 11.29 90288 458876 487 96 90
2005 10.90 87183 477177 503 92 86
2006 10.90 87194 494496 520 92 86
2007 10.93 87426 514111 534 91 85
2008 11.13 89038 540460 555 91 86
2009 13.51 108108 569090 553 105 98
2010 12.64 101088 592745 562 96 90
2011 12.42 99387 612942 580 94 88
Source: Own elaboration based on INEGI (BIE), Banco de Mexico and BEA.
Fix Mexican nominal exchange rate.
Nominal and Real Exchange Rates, 1970-2011
Exchange Rate Consumer Price Index
160
Table II.11
Years NCI RNCF Oil Exports R. Oil Exports G US CPI US dollars Mex Mex Nominal GDP Mex Real GDP General government final
1988=100 GDP Deflator 1988=100 consumption expenditure (% of GDP)
1970 1375 4192 44627 110221 6371 33 40 35542 87781 7.26
1971 1372 4009 35691 82930 6939 34 43 39201 91084 7.62
1972 1563 4424 19966 43567 8508 35 46 45178 98579 8.63
1973 1892 5040 21316 41006 9761 38 52 55271 106328 9.18
1974 3493 8382 124497 194539 10284 42 64 71977 112471 9.14
1975 4668 10264 449065 606883 12270 45 74 88004 118932 10.32
1976 3337 6938 546068 761748 13670 48 72 89024 124185 11.01
1977 2498 4876 985924 1547054 13815 51 64 81826 128396 10.76
1978 3372 6118 1719175 2346007 15274 55 73 102517 139897 10.92
1979 4886 7962 3931729 4484742 16725 61 88 134540 153464 10.90
1980 11235 16129 12049697 10392920 16828 70 116 194357 167634 10.04
1981 17376 22614 16804771 12252625 19632 77 137 250083 182339 10.77
1982 2590 3175 18579709 19379037 18971 82 96 173721 181195 10.47
1983 -2742 -3256 16706012 19480591 15280 84 86 148867 173591 8.80
1984 -830 -945 17971698 18404117 16610 88 98 175632 179858 9.24
1985 -3223 -3544 16077375 16081680 17031 91 100 184473 184522 9.23
1986 1976 2133 7080817 9715085 16165 93 73 129440 177596 9.10
1987 1861 1938 11499973 14830968 15904 96 78 140264 180891 8.79
1988 -4358 -4358 9847685 9847685 15410 100 100 183144 183144 8.41
1989 6217 5931 11916705 10198821 15779 105 117 222977 190833 8.27
1990 10818 9791 15300860 11677914 16810 110 131 262710 200505 8.38
1991 22341 19405 12815705 8516706 18970 115 150 314454 208971 9.08
1992 25458 21466 8125577 4839330 21505 119 168 363609 216554 9.93
1993 29340 24020 7287886 3990629 29245 122 224 503963 225349 12.98
1994 11271 8997 7246852 3962947 29551 125 223 527319 236002 12.52
1995 11167 8669 8169195 6162714 29293 129 155 343793 222412 13.17
1996 4281 3228 11508872 7861263 29328 133 169 397404 235478 12.45
1997 18177 13398 11085691 6704334 30485 136 191 480555 251874 12.10
1998 18131 13159 6989647 4226793 30574 138 190 502010 263716 11.59
1999 14592 10361 9742451 5356777 30880 141 214 579460 270750 11.41
2000 21577 14823 16081846 7801346 32072 146 240 683648 285090 11.25
2001 25055 16737 12641908 5722736 32258 150 256 724704 283364 11.38
2002 21792 14330 14305836 6258058 32551 152 261 741559 283738 11.47
2003 17592 11311 18595638 7641235 32553 156 248 713284 287774 11.31
2004 10864 6804 23301112 9183296 32049 160 257 770268 300136 10.68
2005 15779 9558 31931584 11623865 33070 165 280 866346 309238 10.69
2006 6446 3782 38807909 13242745 34091 170 298 966736 324704 10.50
2007 24554 14009 42857614 13881990 35384 175 312 1043395 334927 10.56
2008 27000 14835 50644202 15711437 36957 182 324 1099073 339617 10.88
2009 13184 7270 31044613 11200877 38794 181 277 895313 323653 11.99
2010 25964 14086 41888883 13584185 39660 184 309 1051628 340193 11.66
2011 40699 21405 56966984 17216036 40970 190 331 1170086 353952 11.58
Source: Own elaboration based on Banco de Mexico and World Development Indicators.
NCI = Net Capital Inflows: Capital Account + Net Errors and Omissions.
NCI were deflated using the US consumer price index 1988=100.
Oil Exports were deflated using the Mexican GDP deflator in US dollars 1988=100.
General government final consumtion expenditure and Nomianl and Real GDP from Mexico were obtained from World Bank (WDI).
G=Real general government final consumption expenditure.
Millions of DollarsMillions of Dollars
161
Table II.12
Years RERI2 RULCI2 RNCF G
1970 83 200 4192 6371
1971 82 175 4009 6939
1972 80 172 4424 8508
1973 76 177 5040 9761
1974 68 170 8382 10284
1975 65 141 10264 12270
1976 73 131 6938 13670
1977 88 139 4876 13815
1978 81 146 6118 15274
1979 77 139 7962 16725
1980 70 132 16129 16828
1981 64 118 22614 19632
1982 95 113 3175 18971
1983 107 144 -3256 15280
1984 95 159 -945 16610
1985 95 157 -3544 17031
1986 123 161 2133 16165
1987 124 176 1938 15904
1988 100 100 -4358 15410
1989 95 88 5931 15779
1990 90 84 9791 16810
1991 82 78 19405 18970
1992 75 72 21466 21505
1993 71 71 24020 29245
1994 74 74 8997 29551
1995 107 87 8669 29293
1996 97 89 3228 29328
1997 85 80 13398 30485
1998 86 73 13159 30574
1999 79 68 10361 30880
2000 74 61 14823 32072
2001 71 58 16737 32258
2002 71 56 14330 32551
2003 77 48 11311 32553
2004 79 44 6804 32049
2005 76 41 9558 33070
2006 76 40 3782 34091
2007 75 40 14009 35384
2008 76 42 14835 36957
2009 87 39 7270 38794
2010 79 39 14086 39660
2011 78 40 21405 40970
Indexes Base 1988=100 Millions of Dollars
162
Appendix to Chapter II-B
(1) 1971 – 2011
*******************************************************
Dependent variable is dLRXR Dependent variable is LRXR
41 observations used for estimation from 1971 to 2011 41 observations used for estimation from 1971 to 2011
*******************************************************
Regressor Coefficient Standard Error T-Ratio[Prob] Regressor Coefficient Standard Error T-Ratio[Prob]
List of additional temporary variables created: ********************************************************************
dLRXR = LRXR-LRXR(-1) * Test Statistics * LM Version * F Version *
dLRULC = LRULC-LRULC(-1) ********************************************************************
dT = T-T(-1) * * * *
ecm = LRXR -.82347*LRULC -.033160*T * A:Serial Correlation*CHSQ( 1)= .96158[.327]*F( 1, 37)= .88861[.352]*
* * * *
R-Squared .36868 R-Bar-Squared .33545 * B:Functional Form *CHSQ( 1)= .59087[.442]*F( 1, 37)= .54102[.467]*
S.E. of Regression .10232 F-stat. F( 2, 38) 11.0954[.000] * * * *
Mean of Dependent Variable -.0015154 S.D. of Dependent Variable .12551 * C:Normality *CHSQ( 2)= 5.0281[.081]* Not applicable *
Residual Sum of Squares .39782 Equation Log-likelihood 36.8476 * * * *
Akaike Info. Criterion 33.8476 Schwarz Bayesian Criterion 31.2773 * D:Heteroscedasticity*CHSQ( 1)= .33176[.565]*F( 1, 39)= .31815[.576]*
DW-statistic 1.7177
Error Correction Representation for the Selected ARDL Model
ARDL(1,0) selected based on R-BAR Squared Criterion
*****************************************************************
dLRULC .51356 .10911 4.7068[.000]
dT .020680 .0045762 4.5191[.000]
ecm(-1) -.62365 .13247 -4.7080[.000]
Estimated Long Run Coefficients using the ARDL Approach
ARDL(1,0) selected based on R-BAR Squared Criterion
LRULC .82347 .010039 82.0295[.000]
T .033160 .0018092 18.3285[.000]
Diagnostic Tests
*****************************************************************
*****************************************************************
*****************************************************************
(1) F-Test
Dependent variable is DLRXR
List of the variables added to the regression:
LRXR(-1) LRULC(-1)
40 observations used for estimation from 1972 to 2011
Regressor Coefficient Standard Error T-Ratio[Prob]
Joint test of zero restrictions on the coefficients of additional variables:
Lagrange Multiplier Statistic CHSQ( 2)= 9.6617[.008]
Likelihood Ratio Statistic CHSQ( 2)= 11.0587[.004]
F Statistic F( 2, 35)= 5.5731[.008]
***********************************************************
***********************************************************
***********************************************************
T .022450 .0069122 3.2479[.003]
DLRXR(-1) .50848 .19383 2.6233[.013]
DLRULC(-1) -.39614 .18162 -2.1811[.036]
LRXR(-1) -.68267 .20465 -3.3359[.002]
LRULC(-1) .55462 .16672 3.3266[.002]
Variable Addition Test (OLS case)
163
(2) 1976 – 2011
******************************************************
Dependent variable is dLRXR Dependent variable is LRXR 36 observations used for estimation from 1976 to 2011 36 observations used for estimation from 1976 to 2011
******************************************************
Regressor Coefficient Standard Error T-Ratio[Prob] Regressor Coefficient Standard Error T-Ratio[Prob]
* Test Statistics * LM Version * F Version *
* * * *
List of additional temporary variables created: * A:Serial Correlation*CHSQ( 1)= .0026921[.959]*F( 1, 24)= .0017949[.967]*
dLRXR = LRXR-LRXR(-1) * * * *
dLRXR1 = LRXR(-1)-LRXR(-2) * B:Functional Form *CHSQ( 1)= .39209[.531]*F( 1, 24)= .26427[.612]*
dLRXR2 = LRXR(-2)-LRXR(-3) * * * *
dLRULC = LRULC-LRULC(-1) * C:Normality *CHSQ( 2)= 19.7140[.000]* Not applicable *
dNCF = NCF-NCF(-1) * * * *
dLG = LG-LG(-1) * D:Heteroscedasticity*CHSQ( 1)= .65076[.420]*F( 1, 34)= .62592[.434]*
dLG1 = LG(-1)-LG(-2)
dLG2 = LG(-2)-LG(-3)
dINPT = INPT-INPT(-1)
dT = T-T(-1)
ecm = LRXR -.55331*LRULC + .1533E-4*NCF + .36229*LG -4.8810*INPT -.036
097*T
R-Squared .80577 R-Bar-Squared .72808
S.E. of Regression .068927 F-stat. F( 9, 26) 11.5238[.000]
Mean of Dependent Variable .0050645 S.D. of Dependent Variable .13218
Residual Sum of Squares .11877 Equation Log-likelihood 51.7711
Akaike Info. Criterion 40.7711 Schwarz Bayesian Criterion 32.0618
DW-statistic 1.9665
Diagnostic Tests
*****************************************************************
*****************************************************************
LRULC .55331 .11449 4.8327[.000]
NCF -.1533E-4 .5450E-5 -2.8121[.009]
LG -.36229 .16677 -2.1724[.040]
INPT 4.8810 1.6020 3.0468[.005]
T .036097 .0078784 4.5818[.000]
******************************************************************
******************************************************************
dLRXR1 -.20912 .13394 -1.5613[.131]
dLRXR2 -.16081 .12525 -1.2839[.211]
dLRULC .41723 .11036 3.7807[.001]
dNCF -.1156E-4 .2950E-5 -3.9178[.001]
dLG -.20436 .25102 -.81412[.423]
dLG1 .038903 .20397 .19073[.850]
dLG2 .71421 .20009 3.5695[.001]
dINPT 3.6806 1.6750 2.1973[.037]
dT .027219 .0083708 3.2517[.003]
ecm(-1) -.75406 .15841 -4.7601[.000]
Error Correction Representation for the Selected ARDL Model
ARDL(3,0,0,3) selected based on R-BAR Squared Criterion
******************************************************************
******************************************************************
Estimated Long Run Coefficients using the ARDL Approach
ARDL(3,0,0,3) selected based on R-BAR Squared Criterion
(2) F-Test
Dependent variable is DLRXR
List of the variables added to the regression: LRXR(-1) LRULC(-1) NCF(-1) LG(-1) D82
36 observations used for estimation from 1976 to 2011
Regressor Coefficient Standard Error T-Ratio[Prob]
Joint test of zero restrictions on the coefficients of additional variables:
Lagrange Multiplier Statistic CHSQ( 5)= 18.6635[.002]
Likelihood Ratio Statistic CHSQ( 5)= 26.3054[.000]
F Statistic F( 5, 25)= 5.3827[.002]
*************************************************************
INPT 3.5168 2.1747 1.6171[.118]
T .024242 .012847 1.8870[.071]
DLRXR(-1) .46563 .20403 2.2821[.031]
DLRULC(-1) -.11040 .17097 -.64572[.524]
DNCF(-1) -.4359E-5 .4000E-5 -1.0899[.286]
DLG(-1) .39069 .29411 1.3284[.196]
LRXR(-1) -.89533 .26181 -3.4198[.002]
LRULC(-1) .42272 .19267 2.1940[.038]
NCF(-1) -.6059E-5 .4860E-5 -1.2466[.224]
LG(-1) -.19776 .21454 -.92177[.365]
D82 .35917 .11388 3.1539[.004]
*************************************************************
*************************************************************
Variable Addition Test (OLS case)
164
(3) 1983 – 2011
*******************************************************
Dependent variable is dLRXR Dependent variable is LRXR 29 observations used for estimation from 1983 to 2011 29 observations used for estimation from 1983 to 2011
*******************************************************
Regressor Coefficient Standard Error T-Ratio[Prob] Regressor Coefficient Standard Error T-Ratio[Prob]
* Test Statistics * LM Version * F Version *
* * * *
List of additional temporary variables created: * A:Serial Correlation*CHSQ( 1)= .0078085[.930]*F( 1, 16)= .0043093[.948]*
dLRXR = LRXR-LRXR(-1) * * * *
dLRXR1 = LRXR(-1)-LRXR(-2) * B:Functional Form *CHSQ( 1)= .0033284[.954]*F( 1, 16)= .0018365[.966]*
dLRXR2 = LRXR(-2)-LRXR(-3) * * * *
dLRULC = LRULC-LRULC(-1) * C:Normality *CHSQ( 2)= 1.0653[.587]* Not applicable *
dNCF = NCF-NCF(-1) * * * *
dLG = LG-LG(-1) * D:Heteroscedasticity*CHSQ( 1)= 1.5352[.215]*F( 1, 27)= 1.5092[.230]*
dLG1 = LG(-1)-LG(-2)
dLG2 = LG(-2)-LG(-3)
dINPT = INPT-INPT(-1)
dT = T-T(-1)
ecm = LRXR -.61681*LRULC + .1760E-4*NCF + .41320*LG -4.9696*INPT -.041
738*T
R-Squared .88990 R-Bar-Squared .81867
S.E. of Regression .049617 F-stat. F( 9, 19) 15.2680[.000]
Mean of Dependent Variable -.0067989 S.D. of Dependent Variable .11652
Residual Sum of Squares .041851 Equation Log-likelihood 53.6944
Akaike Info. Criterion 41.6944 Schwarz Bayesian Criterion 33.4907
DW-statistic 2.0232
LG -.41320 .17694 -2.3352[.032]
INPT 4.9696 1.5497 3.2068[.005]
T .041738 .011779 3.5433[.002]
****************************************************************
****************************************************************
Diagnostic Tests
******************************************************************
******************************************************************
Estimated Long Run Coefficients using the ARDL Approach
ARDL(3,0,1,3) selected based on Akaike Information Criterion
dLRXR1 -.13082 .11256 -1.1622[.260]
dLRXR2 -.16238 .10135 -1.6021[.126]
dLRULC .40623 .085893 4.7295[.000]
dNCF -.8052E-5 .2485E-5 -3.2402[.004]
dLG -.085866 .20115 -.42687[.674]
dLG1 .085821 .18082 .47461[.640]
dLG2 .83854 .15635 5.3631[.000]
dINPT 3.2730 1.4364 2.2786[.034]
dT .027488 .0073250 3.7527[.001]
ecm(-1) -.65860 .14174 -4.6466[.000]
LRULC .61681 .14906 4.1381[.001]
NCF -.1760E-4 .6009E-5 -2.9285[.009]
Error Correction Representation for the Selected ARDL Model
ARDL(3,0,1,3) selected based on Akaike Information Criterion
******************************************************************
******************************************************************
(3) F-Test
Dependent variable is DLRXR List of the variables added to the regression:
LRXR(-1) LRULC(-1) NCF(-1) LG(-1)
29 observations used for estimation from 1983 to 2011
Regressor Coefficient Standard Error T-Ratio[Prob]
Joint test of zero restrictions on the coefficients of additional variables:
Lagrange Multiplier Statistic CHSQ( 4)= 14.6307[.006]
Likelihood Ratio Statistic CHSQ( 4)= 20.3639[.000]
F Statistic F( 4, 19)= 4.8364[.007]
*************************************************************
*************************************************************
INPT 4.4636 2.3349 1.9117[.071]
T .011483 .013850 .82911[.417]
DLRXR(-1) .58137 .22106 2.6299[.016]
DLRULC(-1) -.13388 .16794 -.79717[.435]
DNCF(-1) .1498E-5 .4199E-5 .35680[.725]
DLG(-1) .29351 .31066 .94480[.357]
LRXR(-1) -1.2959 .31726 -4.0847[.001]
LRULC(-1) .38159 .19698 1.9372[.068]
NCF(-1) -.1130E-4 .5122E-5 -2.2072[.040]
LG(-1) -.056819 .22257 -.25529[.801]
*************************************************************
Variable Addition Test (OLS case)
165
(4) Exogenity Test 1
Dependent variable is LRXR List of the variables added to the regression:
RLRULC1
40 observations used for estimation from 1972 to 2011
Regressor Coefficient Standard Error T-Ratio[Prob]
Joint test of zero restrictions on the coefficients of additional variables:
Lagrange Multiplier Statistic CHSQ( 1)= .31633[.574]
Likelihood Ratio Statistic CHSQ( 1)= .31759[.573]
F Statistic F( 1, 35)= .27900[.601]
Variable Addition Test (OLS case)
***************************************************************
***************************************************************
***************************************************************
INPT 1.2551 .50370 2.4918[.018]
LRXR(-1) .69304 .11869 5.8389[.000]
LRULC .53189 .27858 1.9093[.064]
LRULC(-1) -.50559 .27670 -1.8272[.076]
RLRULC1 -.17459 .33053 -.52820[.601]
(5) Exogenity Test 2
Dependent variable is LRULC
List of the variables added to the regression: RLRXR1
40 observations used for estimation from 1972 to 2011
Regressor Coefficient Standard Error T-Ratio[Prob]
Joint test of zero restrictions on the coefficients of additional variables:
Lagrange Multiplier Statistic CHSQ( 1)= 7.1767[.007]
Likelihood Ratio Statistic CHSQ( 1)= 7.9096[.005]
F Statistic F( 1, 35)= 7.6526[.009]
Variable Addition Test (OLS case)
*************************************************************
*************************************************************
*************************************************************
INPT -1.9692 .78445 -2.5103[.017]
LRULC(-1) .95558 .034542 27.6642[.000]
LRXR 1.6037 .44912 3.5707[.001]
LRXR(-1) -1.1202 .31635 -3.5411[.001]
RLRXR1 -1.3097 .47345 -2.7663[.009]
166
Appendix to Chapter III
Figure 11: Currency Valuation and Economic Growth
-.8
-.4
.0
.4
.8
-30 -20 -10 0 10 20 30 40
GROWTH
Overv
alu
ation / U
nderv
alu
ation
96 Countries, 1960-1980
74 801
107 625
-1.0
-0.5
0.0
0.5
1.0
-20 -10 0 10 20
GROWTH
Overv
alu
ation / U
nderv
alu
ation
96 Countries, 1981-2010
180 1,217
253 1,187-.4
-.2
.0
.2
.4
-4 0 4 8 12 16
GROWTH
Overv
alu
atio
n / U
nd
erv
alu
ation
25 Developed Countries, 1960-1980
10
13
179
227-.4
-.3
-.2
-.1
.0
.1
.2
.3
.4
-5 0 5 10 15
GROWTH
Overv
alu
ation / U
nde
rvalu
ation
25 Developed Countries, 1981-2010
30 304
56 360
-.8
-.4
.0
.4
.8
-30 -20 -10 0 10 20 30 40
GROWTH
Overv
alu
atio
n / U
nderv
alu
ation
71 Developing Countries, 1960-1980
70 562
88 458-1.5
-1.0
-0.5
0.0
0.5
1.0
1.5
-20 -10 0 10 20
GROWTH
Overv
alu
ation / U
nderv
alu
ation
71 Developing Countries, 1981-2010
155 890
192 850-.6
-.4
-.2
.0
.2
.4
.6
-20 -10 0 10 20 30 40
GROWTH
Overv
alu
ation / U
nderv
alu
ation
35 African Countries, 1960-1980
50 208
42 229
-.8
-.4
.0
.4
.8
-30 -20 -10 0 10 20 30
GROWTH
Ove
rvalu
ation / U
nderv
alu
ation
35 African Countries, 1981-2010
73 433
108 419
-.6
-.4
-.2
.0
.2
.4
.6
-10 0 10 20
GROWTH
Overv
alu
ation / U
nde
rva
luation
18 Asian Countries, 1960-2010
16 156
19 140
-.6
-.4
-.2
.0
.2
.4
.6
-12 -8 -4 0 4 8 12 16
GROWTH
Overv
alu
ation / U
nderv
alu
ation
18 Asian Countries, 1981-2010
20 258
24 235
-.4
-.2
.0
.2
.4
-10 0 10 20
GROWTH
Overv
alu
ation / U
nderv
alu
ation
20 Latin American Countries, 1960-1980
13
22 148
186
-0.6
-0.4
-0.2
0.0
0.2
0.4
0.6
0.8
1.0
-10 0 10
GROWTH
Overv
alu
ation / U
nde
rvalu
ation
20 Latin American Countries, 1981-2010
50 243
62 222
167
Figure 12: Undervaluation and Effective Demand: Developed Countries
-8
-6
-4
-2
0
2
4
6
8
10
90 92 94 96 98 00 02 04 06 08 10 12
G - T X - M I - S
-20
-10
0
10
20
30
-4
-2
0
2
4
90 92 94 96 98 00 02 04 06 08 10 12
UNDERVAL GROWTH
-20
-10
0
10
20
30
60
61
62
63
64
65
90 92 94 96 98 00 02 04 06 08 10 12
UNDERVAL WS
Belgium1990 - 2012
Belgium1990-2012
Belgium1990-2012
% o
f G
DP
%
%
% o
f G
DP
% o
f GD
P%
-10.0
-7.5
-5.0
-2.5
0.0
2.5
5.0
7.5
10.0
90 92 94 96 98 00 02 04 06 08 10 12
G - T X - M I - S
% o
f G
DP
Canada1990 - 2012
-20
-10
0
10
20
30
-4
-2
0
2
4
6
90 92 94 96 98 00 02 04 06 08 10 12
UNDERVAL GROWTH
%
%
Canada1990 - 2012
-20
-10
0
10
20
30
56
58
60
62
64
90 92 94 96 98 00 02 04 06 08 10 12
UNDERVAL WS
% fo
GD
P
%
Canada1990 - 2012
-12
-8
-4
0
4
8
90 92 94 96 98 00 02 04 06 08 10 12
G - T X - M I - S
% o
f G
DP
France1990 - 2012
-20
-10
0
10
20
30
-4
-2
0
2
4
90 92 94 96 98 00 02 04 06 08 10 12
UNDERVAL GROWTH
France1990 - 2012
%
%
-20
-10
0
10
20
30
62
62
63
63
64
64
90 92 94 96 98 00 02 04 06 08 10 12
UNDERVAL WS
France1990 - 2012
%
% o
f GD
P
-30
-20
-10
0
10
20
90 92 94 96 98 00 02 04 06 08 10 12
G - T X - M I - S
Greece1990 - 2012
% o
f G
DP
-30
-20
-10
0
10
20
30
-8
-4
0
4
8
90 92 94 96 98 00 02 04 06 08 10 12
UNDERVAL GROWTH
Greece1990 - 2012
%
%
-30
-20
-10
0
10
20
30
51
52
53
54
55
90 92 94 96 98 00 02 04 06 08 10 12
UNDERVAL WS
Greece1990 - 2012
%
% o
f GD
P
-32
-28
-24
-20
-16
-12
-8
-4
0
4
90 92 94 96 98 00 02 04 06 08 10 12
G - T X - M I - S
UK1990 - 2012
% o
f G
DP
-20
-10
0
10
20
-6
-4
-2
0
2
4
6
90 92 94 96 98 00 02 04 06 08 10 12
UNDERVAL GROWTH
UK1990 - 2012
%
%
-20
-10
0
10
20
61
62
63
64
65
66
67
90 92 94 96 98 00 02 04 06 08 10 12
UNDERVAL WS
UK1990 - 2012
%
% o
f GD
P
-32
-28
-24
-20
-16
-12
-8
-4
0
4
90 92 94 96 98 00 02 04 06 08 10 12
G - T X - M I - S
% o
f G
DP
-4
-2
0
2
4
6
8
-4
-2
0
2
4
6
90 92 94 96 98 00 02 04 06 08 10 12
UNDERVAL GROWTH
%
%
-4
-2
0
2
4
6
8
61
62
63
64
65
66
67
90 92 94 96 98 00 02 04 06 08 10 12
UNDERVAL WS
%
% o
f GD
P
US1990 - 2012 US
1990 - 2012
US1990 - 2012
168
Figure 13: Undervaluation and Effective Demand: African Countries
-40
-30
-20
-10
0
10
20
30
40
90 92 94 96 98 00 02 04 06 08 10 12
G - T X - M I - S
Cameron1990 - 2012
% o
f G
DP
-30
-20
-10
0
10
20
30
-12
-8
-4
0
4
8
90 92 94 96 98 00 02 04 06 08 10 12
UNDERVAL GROWTH
Cameron1994 - 2001
%
%
-30
-20
-10
0
10
20
30
50
51
52
53
54
55
56
90 92 94 96 98 00 02 04 06 08 10 12
UNDERVAL WS
Cameron1990 - 2012
% o
f GD
P
%
-4
0
4
8
12
16
20
97 98 99 00 01 02 03 04 05 06 07 08 09 10 11 12
G - T X - M I - S
Cote d´Ivoire1997 - 2012
% o
f G
DP
-10
-5
0
5
10
15
-4
-2
0
2
4
6
97 98 99 00 01 02 03 04 05 06 07 08 09 10 11 12
UNDERVAL GROWTH
Cote d´Ivoire1997 - 2012
%
%
-10
-5
0
5
10
15
45
46
47
48
49
50
51
97 98 99 00 01 02 03 04 05 06 07 08 09 10 11 12
UNDERVAL WS
Cote d´Ivoire1997 - 2012
%
% o
f GD
P
-20
-15
-10
-5
0
5
10
90 92 94 96 98 00 02 04 06 08 10 12
G - T X - M I - S
Egypt1990 - 2012
% o
f G
DP
-20
0
20
40
60
80
0
2
4
6
8
90 92 94 96 98 00 02 04 06 08 10 12
UNDERVAL GROWTH
Egypt1990 - 2012
%
%
-20
0
20
40
60
80
33
34
35
36
37
38
39
90 92 94 96 98 00 02 04 06 08 10 12
UNDERVAL WS
Egypt1990 - 2012
%
% o
f GD
P
-10
0
10
20
30
40
90 92 94 96 98 00 02 04 06 08 10 12
G - T X - M I - S
% o
f G
DP
Nigeria1990 - 2012
-80
-40
0
40
80
120
-10
0
10
20
30
40
90 92 94 96 98 00 02 04 06 08 10 12
UNDERVAL GROWTH
Nigeria1990 - 2012
%
%
-80
-40
0
40
80
120
25
30
35
40
45
50
55
90 92 94 96 98 00 02 04 06 08 10 12
UNDERVAL WS
Nigeria1990 - 2012
%
% o
f GD
P
-32
-28
-24
-20
-16
-12
-8
-4
0
4
8
1990 1992 1994 1996 1998 2000 2002 2004 2006 2008 2010
G - T X - M I - S
Senegal1990 - 2010
% o
f G
DP
-30
-20
-10
0
10
20
-2
0
2
4
6
8
1990 1992 1994 1996 1998 2000 2002 2004 2006 2008 2010
UNDERVAL GROWTH
Senegal1990 - 2010%
%
-30
-20
-10
0
10
20
37
38
39
40
41
42
1990 1992 1994 1996 1998 2000 2002 2004 2006 2008 2010
UNDERVAL WS
Senegal1990 - 2010
%
% o
f GD
P
-6
-4
-2
0
2
4
6
90 92 94 96 98 00 02 04 06 08 10 12
G - T X - M I - S
South Africa1990 - 2012
% o
f G
DP
-30
-20
-10
0
10
20
-4
-2
0
2
4
6
90 92 94 96 98 00 02 04 06 08 10 12
UNDERVAL GROWTH
South Africa1990 - 2012
%
%
-30
-20
-10
0
10
20
50
52
54
56
58
60
90 92 94 96 98 00 02 04 06 08 10 12
UNDERVAL WS
South Africa1990 - 2012
%
% o
f GD
P
169
Figure 14: Undervaluation and Effective Demand: Asian Countries
-32
-28
-24
-20
-16
-12
-8
-4
0
4
8
90 92 94 96 98 00 02 04 06 08 10 12
G - T X - M I - S
Tanzania1990 - 2012
% o
f G
DP
-20
0
20
40
60
0
2
4
6
8
90 92 94 96 98 00 02 04 06 08 10 12
UNDERVAL GROWTH
Tanzania1990 - 2012%
%
-20
0
20
40
60
40
44
48
52
56
90 92 94 96 98 00 02 04 06 08 10 12
UNDERVAL WS
Tanzania1990 - 2012%
% o
f GD
P
-4
-2
0
2
4
6
8
10
90 92 94 96 98 00 02 04 06 08 10 12
G - T X - M I - S
China1990 - 2012
% o
f G
DP
0
10
20
30
40
50
2
4
6
8
10
12
14
16
90 92 94 96 98 00 02 04 06 08 10 12
UNDERVAL GROWTH
China1990 - 2012
%
%
0
10
20
30
40
50
40
44
48
52
56
90 92 94 96 98 00 02 04 06 08 10 12
UNDERVAL WS
China1990 - 2012
%
% o
f GD
P
-10
-5
0
5
10
15
20
90 92 94 96 98 00 02 04 06 08 10 12
G - T X- M I - S
Hong Kong1990 - 2012
% o
f G
DP
-40
-30
-20
-10
0
10
20
-8
-4
0
4
8
12
90 92 94 96 98 00 02 04 06 08 10 12
UNDERVAL GROWTH
Hong Kong199 - 2010%
%
-40
-30
-20
-10
0
10
20
46
48
50
52
54
56
90 92 94 96 98 00 02 04 06 08 10 12
UNDERVAL WS
Hong Kong1990 - 2012
%
% o
f GD
P
-12
-8
-4
0
4
8
90 92 94 96 98 00 02 04 06 08 10 12
G - T X - M I - S
India1990 - 2012
% o
f G
DP
-10
0
10
20
30
0
2
4
6
8
10
12
90 92 94 96 98 00 02 04 06 08 10 12
UNDERVAL GROWTH
India1990 - 2012
%
%
-10
0
10
20
30
45
50
55
60
65
70
90 92 94 96 98 00 02 04 06 08 10 12
UNDERVAL WS
India1990 - 2012
%
% o
f GD
P
-12
-8
-4
0
4
8
90 92 94 96 98 00 02 04 06 08 10 12
G - T X - M I - S
Indonesia1990 - 2012
% o
f G
DP
-40
0
40
80
120
-15
-10
-5
0
5
10
90 92 94 96 98 00 02 04 06 08 10 12
UNDERVAL GROWTH
Indonesia1990 - 2012
%
%
-40
0
40
80
120
44
45
46
47
48
90 92 94 96 98 00 02 04 06 08 10 12
UNDERVAL WS
Indonesia1990 - 2012
%
% o
f GD
P
-8
-4
0
4
8
12
90 92 94 96 98 00 02 04 06 08 10 12
G - T X - M I - S
Japan1990 - 2012
% o
f G
DP
-30
-20
-10
0
10
20
-6
-4
-2
0
2
4
6
90 92 94 96 98 00 02 04 06 08 10 12
UNDERVAL GROWTH
Japan1990 - 2012
%
%
-30
-20
-10
0
10
20
.50
.52
.54
.56
.58
90 92 94 96 98 00 02 04 06 08 10 12
UNDERVAL WS
Japan1990 - 2012
%
% o
f GD
P
170
Figure 15: Undervaluation and Effective Demand: Latin American Countries
-5.0
-2.5
0.0
2.5
5.0
7.5
10.0
12.5
15.0
90 92 94 96 98 00 02 04 06 08 10 12
G - T X - M I - S
Korea1990 - 2012
% o
f G
DP
-30
-20
-10
0
10
20
-8
-4
0
4
8
12
90 92 94 96 98 00 02 04 06 08 10 12
UNDERVAL GROWTH
Korea1990 - 2012
%
%
-30
-20
-10
0
10
20
52
54
56
58
60
90 92 94 96 98 00 02 04 06 08 10 12
UNDERVAL WS
Korea1990 - 2012
%
% o
f GD
P
-8
-4
0
4
8
12
16
20
24
28
96 97 98 99 00 01 02 03 04 05 06 07 08 09 10 11 12
G - T X - M I - S
Malysia1996 - 2012
% o
f G
DP
0
4
8
12
16
-8
-4
0
4
8
12
96 97 98 99 00 01 02 03 04 05 06 07 08 09 10 11 12
UNDERVAL GROWTH
Malysian1996 - 2012
%
%
0
4
8
12
16
50
51
52
53
54
55
56
96 97 98 99 00 01 02 03 04 05 06 07 08 09 10 11 12
UNDERVAL WS
Malysian1996 - 2012
%
% o
f GD
P
-20
-10
0
10
20
30
40
50
90 92 94 96 98 00 02 04 06 08 10 12
G - T X - M I - S
Singapore1990 - 2012
% o
f G
DP
-20
-15
-10
-5
0
5
10
-4
0
4
8
12
16
90 92 94 96 98 00 02 04 06 08 10 12
UNDERVAL GROWTH
Singapore1990 - 2012
%
%
-20
-15
-10
-5
0
5
10
40
42
44
46
48
50
90 92 94 96 98 00 02 04 06 08 10 12
UNDERVAL WS
Singapore1990 - 2012
%
% o
f GD
P
-16
-12
-8
-4
0
4
8
12
16
20
24
90 92 94 96 98 00 02 04 06 08 10 12
G - T X - M I - S
Thailand1990 - 2012
% o
f G
DP
-12
-8
-4
0
4
8
-12
-8
-4
0
4
8
12
90 92 94 96 98 00 02 04 06 08 10 12
UNDERVAL GROWTH
Thailand1990- 2012
%
%
-12
-8
-4
0
4
8
35
36
37
38
39
40
41
90 92 94 96 98 00 02 04 06 08 10 12
UNDERVAL WS
Thailand1990 - 2012
%
% o
f GD
P
-8
-4
0
4
8
12
16
90 92 94 96 98 00 02 04 06 08 10 12
G - T X - M I - S
Argentina1990 - 2012
% o
f G
DP
-40
0
40
80
120
-15
-10
-5
0
5
10
15
90 92 94 96 98 00 02 04 06 08 10 12
UNDERVAL GROWTH
Argentina1990 - 2012
%
%
-40
0
40
80
120
32
36
40
44
48
52
90 92 94 96 98 00 02 04 06 08 10 12
UNDERVAL WS
Argentina1990 - 2012
%
% o
f GD
P
171
-12
-8
-4
0
4
8
12
90 92 94 96 98 00 02 04 06 08 10 12
G - T X - M I - S
Brazil1990 - 2012
% o
f G
DP
-60
-40
-20
0
20
40
-8
-4
0
4
8
90 92 94 96 98 00 02 04 06 08 10 12
UNDERVAL GROWTH
Brazil1990 - 2012
%
%
-60
-40
-20
0
20
40
50
52
54
56
58
90 92 94 96 98 00 02 04 06 08 10 12
UNDERVAL WS
Brazil1990 - 2012
%
% o
f GD
P
-10
-5
0
5
10
15
20
25
90 92 94 96 98 00 02 04 06 08 10 12
G - T X - M I - S
Chile1990 - 2012
% o
f G
DP
-20
0
20
40
60
-4
0
4
8
12
16
90 92 94 96 98 00 02 04 06 08 10 12
UNDERVAL GROWTH
Chile1990 - 2012
%
%
-20
0
20
40
60
43
44
45
46
47
90 92 94 96 98 00 02 04 06 08 10 12
UNDERVAL WS
Chile1990 - 2012
%
% o
f GD
P
-8
-6
-4
-2
0
2
4
6
8
90 92 94 96 98 00 02 04 06 08 10 12
G - T X - M I - S
Colombia1990 - 2012
%
-20
-10
0
10
20
30
40
-6
-4
-2
0
2
4
6
8
90 92 94 96 98 00 02 04 06 08 10 12
UNDERVAL GROWTH
Colombia1990 - 2012
%
%
-20
-10
0
10
20
30
40
46
47
48
49
50
51
52
90 92 94 96 98 00 02 04 06 08 10 12
UNDERVAL WS
Colombia1990 - 2012
%
% o
f GD
P
-20
-16
-12
-8
-4
0
4
90 92 94 96 98 00 02 04 06 08 10 12
G - T X - M I - S
Guatemala1990 - 2012
% o
f G
DP
-20
0
20
40
60
0
2
4
6
8
90 92 94 96 98 00 02 04 06 08 10 12
UNDERVAL GROWTH
Guatemala1990 - 2012
%
%
-20
0
20
40
60
42
43
44
45
46
47
90 92 94 96 98 00 02 04 06 08 10 12
UNDERVAL WS
Guatemala1990 - 2012
%
% o
f GD
P
-6
-4
-2
0
2
4
6
90 92 94 96 98 00 02 04 06 08 10 12
G - T X - M I - S
Mexico1990 - 2012
% o
f G
DP
-20
-10
0
10
20
-8
-4
0
4
8
90 92 94 96 98 00 02 04 06 08 10 12
UNDERVAL GROWTH
Mexico1990 - 2012%
%
-20
-10
0
10
20
36
38
40
42
44
46
90 92 94 96 98 00 02 04 06 08 10 12
UNDERVAL WS
Mexico1990 - 2012
%
% o
f GD
P
-8
-4
0
4
8
12
16
90 92 94 96 98 00 02 04 06 08 10 12
G - T X - M I - S
Panama1990 - 2012
% o
f G
DP
-10
0
10
20
30
40
0
4
8
12
16
90 92 94 96 98 00 02 04 06 08 10 12
UNDERVAL GROWTH
Panama1990 - 2012
%
%
-10
0
10
20
30
40
38
40
42
44
46
90 92 94 96 98 00 02 04 06 08 10 12
UNDERVAL WS
Panama1990 - 2012
%
% o
f GD
P
172
Table III.1: Developed Countries
Australia (AUS) Finland (FIN) Israel (ISR) New Zealand (NZL) Sweden (SWE) Austria (AUT) France (FRA) Italy (ITA) Norway (NOR) Switzerland (CHE) Belgium (BEL) Greece (GRC) Japan (JPN) Portugal (PRT) Taiwan (TWN)
Canada (CAN) Hong Kong (HKG) Korea (KOR) Singapore (SGP) U. K. (GBR) Denmark (DNK) Ireland (IRL) Netherlands (NLD) Spain (Spain) USA (USA)
Table III.2: Developing Countries
Algeria Costa Rica India Nepal Syria Argentina Cote d`Ivoire Indonesia Nicaragua Tanzania
Bangladesh Dominican Rep. Jamaica Niger Thailand Benin Ecuador Jordan Nigeria Togo Bolivia Egypt Kenya Pakistan Trinidad &Tobago Botswana El Salvador Madagascar Panama Tunisia Brazil Ethiopia Malawi Papua New Guinea Turkey Burkina Faso Gabon Malaysia Paraguay Uganda Burundi Gambia, The Mali Peru Uruguay Cameroon Ghana Mauritania Philippines Venezuela Cent'l Africa R. Guatemala Mauritius Romania Zambia Chile Guinea Mexico Senegal China Guinea-Bissau Morocco Sierra Leone Colombia Haiti Mozambique South Africa Congo, R of Honduras Namibia Sri Lanka
-8
-4
0
4
8
12
1990 1993 1996 1999 2002 2005 2008 2011
G - T X - M I - S
Peru1990 - 2012
% o
f G
DP
-25
-20
-15
-10
-5
0
5
-10
-5
0
5
10
15
90 92 94 96 98 00 02 04 06 08 10 12
UNDERVAL GROWTH
Peru1990 - 2012
%
%
-25
-20
-15
-10
-5
0
5
.28
.32
.36
.40
.44
.48
90 92 94 96 98 00 02 04 06 08 10 12
UNDERVAL WS
Pero1990 - 2012
%
% o
f GD
P
-6
-4
-2
0
2
4
6
8
90 92 94 96 98 00 02 04 06 08 10 12
G - T X - M I - S
Uruguay1990 - 2012
% o
f G
DP
-40
-20
0
20
40
60-8
-4
0
4
8
12
90 92 94 96 98 00 02 04 06 08 10 12
UNDERVAL GROWTH%
%
Uruguay1990 - 2012
-40
-20
0
20
40
60
40
44
48
52
56
60
90 92 94 96 98 00 02 04 06 08 10 12
UNDERVAL WS
Uruguay1990 - 2012
%
% o
f GD
P
173
Table III.3: African Countries
Algeria Congo, R of Guinea Mauritius Sierra Leone Benin Cote d`Ivoire Guinea-Bissau Morocco South Africa Botswana Egypt Kenya Mozambique Tanzania Burkina Faso Ethiopia Madagascar Namibia Togo Burundi Gabon Malawi Niger Tunisia Cameroon Gambia, The Mali Nigeria Uganda Cent'l Africa R. Ghana Mauritania Senegal Zambia
Table III.4: Asian Countries
Bangladesh Indonesia Nepal Singapore Thailand China Japan Pakistan Sri Lanka Turkey Hong Kong Korea Papua New Guinea Syria India Malaysia Philippines Taiwan
Table III.5: Latin American Countries
Argentina Colombia El Salvador Jamaica Paraguay Bolivia Costa Rica Guatemala Mexico Peru Brazil Dominican Rep. Haiti Nicaragua Uruguay Chile Ecuador Honduras Panama Venezuela