international trade - makro.uni-jena.de · 6.4mobility of human capital: brain drain, brain gain...

International Trade

(Module MW21.5)

PD Dr. M. Pasche

Friedrich Schiller University Jena

Creative Commons 3.0 license – 2018 (with except for included graphics from other sources)

Work in progress! Bug Report to: [email protected]

S.1

Outline:

1. Introduction

2. Technological Differences

2.1 Absolute and Comparative Cost Advantages2.2 Equilibrium, Welfare and Income2.3 Extension to Specific Factors or Increasing Costs2.4 Empirical Relevance

3. Resource Endowment

3.1 Preliminaries from Production Theory3.2 The Neoclassical Baseline Model (Heckscher/Ohlin)3.3 Important Implications3.4 Empirical Relevance

4. Intermediate Products and Global Value Chains

4.1 Role and Measurement of GVC4.2 Comparative Advantages in Intermediate Goods and Services4.3 Some Implications of Globalization

S.2

5. Economies of Scale and Imperfect Competition

5.1 Preliminaries: Monopolistic Competition and Economies ofScale

5.2 Trade with Increasing Returns5.3 Empirical Relevance

6. International Factor Mobility and Multinational Firms

6.1 Factor Movements in a Ricardo-Heckscher-Ohlin framework6.2 Mobility of Capital6.3 Mobility of Labor6.4 Mobility of Human Capital: Brain Drain, Brain Gain6.5 Foreign Direct Investments and Multinational Firms

7. Tariffs, Subsidies, Quotas, and Trade Policy

7.1 Effects of Tariffs and Optimal Tariff Theory7.2 Quotas and Voluntary Export Restraints7.3 Export Subsidies and Strategic Trade Policy7.4 Norms and Regulation

S.3

8. Further Topics in Globalization and Trade

8.1 Trade and Unemployment8.2 Multinational Firms in Developing Countries8.3 “Fair” Trade, Labor Standards and Ethical Consumption8.4 Does Trade Lead to Inefficient Use of Nature?

Basic Literature:

I Krugman, P.R., Obstfeld, M., Melitz. M. (2014), Internationaleconomics: theory and policy. 10th ed., Boston Mass. et al.:Pearson.

* Feenstra, R.C. (2015), Advanced International Trade: Theoryand Evidence. 2nd ed., New York: Worth.

I Gandolfo, G. (2013), International trade theory and policy.2nd ed., Berlin: Springer.

References to more specific literature can be found in the slide collection.

S.4

Preliminary time schedule (winter 18/19):

Week Monday Friday

42 1 243 2,3 ex44 3 345 3 346 ex ex47 4 548 5 ex49 6 650 6 ex51 7 7

02 7 703 ex ex04 8 805 8 ex06 ex t.b.a.

ex = exercise classS.5

Type of module:

I It is compulsory in the specialization area“World Economy”.

I It is elective for the specialization“Quantitative Macroeconomics”.

I Additional (elective) for all other.

Contact hours:

I The course has 4 contact hours per week:approx. 2/3 lecture + 1/3 exercise class (see schedule)

Examination:

I Written exam at the end of the term (60 minutes).

S.6

No-screen-policy / digital detox:

I Nobody can be forced to attend the course in person. But ifyou attend (which is highly recommendable!) a key successfactor is attention.

I Smartphones, tablets, and notebooks are attention killers.Neuroscientists and many didacts warn against the (excessive)use in the classroom. In fact, there is no need for it. It is muchmore important to follow my explanations actively and withattention, to ask if something is unclear, to comment if youhave important additional insights or if you are of opposingopinion.

I Moreover, the use of electronic devices also detracts attentionof your classmates (negative externality).

I So I will follow a no-screen-policy, and I expect that youcommit yourself to this policy as well – to your own benefit.

S.7

1. Introduction

WTO Annual Report 2018:

S.8

1. Introduction

WTO Statistical Review 2018:

S.9

1. Introduction

World export volume and value: (1980 = 100)

1950

1960

1970

1980

1990

2000

2010

0

500

1000

Year

Tra

de

Value

Volume

Source: WTO Database

S.10

1. Introduction

Export values: (1980 = 100)

1950

1960

1970

1980

1990

2000

2010

0

500

1000

1500

2000

Year

Tra

de

Europe

North America

ASEAN

Africa

Source: WTO Database

S.11

1. Introduction

Source: WTO Statistical Review 2018

S.12

1. Introduction

Source: WTO Statistical Review 2018

S.13

1. Introduction

Growth rates of trade volume and GDP (world):

1985

1990

1995

2000

2005

2010

2015

−10

0

10

Year

Gro

wth

rate

Trade growth

GDP growth

Source: WTO Database, World Bank Databse

S.14

1. Introduction

The KOF index of globalization:

Potrafke, N. (2015), The Evidence on Globalization. The WorldEconomy 38(3), 509-552.

S.15

1. Introduction

Stylized facts:

I The trade volume has increased drastically.

I Trade increase pronounced for Asia, less pronounced forAfrica.

I Trade share of developing countries increased, but mainlybecause of trade within this group.

I Growth rates are higher than the GDP growth rates, implyingan increasing trade/GDP ratio.

⇒ Increasing international division of labor

⇒ Increasing trade within value chains, importance ofintermediate products

⇒ Increasing importance of developing and emerging regions,especially Asia

S.16

1. Introduction

What will be explained?

Real Theory:

Explaining trade, specialisation/trade patterns, real exchange rates,gains from trade and effects on real income and wealth by meansof technologies, factor endowments, analysis of (strategic) tradepolicy.(this course)

Monetray Theory:

Explaining nominal exchange rates, dynamics of balance ofpayments, macroeconomic effects of trade and capital movementsin nominal terms.(course “Monetary Macroeconomics”)

S.17

1. Introduction

I Trade requires that a (voluntary) exchange of goods ismutually beneficial.

I A sufficient condition for mutual profitability is that theimported good X is relative cheaper (in terms of a unit of theexported good) compared to the opportunity cost ofproducing the good X itself.

⇒ Differences in relative prices

Which sources induce differences in relative prices and makespecialization and exchange of goods profitabe?

S.18

1. Introduction

1. Technological/Productivity differences between countries

⇒ creates comparative cost advantages⇒ creates comparative price advantages⇒ creates incentive to trade

2. Different factor endowments of countries

a) (non-) availability of ressources (e.g. oil, special metals),temporal/permanent

b) different relative endowments (e.g. (human) capital in relationto labor force)

⇒ creates different factor cost relations⇒ creates comparative cost advantages⇒ creates comparative price advantages⇒ creates incentive to trade

S.19

1. Introduction

3. Increasing returns

⇒ Creates cost advantages due to specialization and massproduction, even though the price relation in autarky does notdiffer.

4. Preferences

a) Different preferences and symmetry otherwise createcomparative price advantage.

b) Preferences for product variety create an incentive to importdifferent product variants.

S.20

2. Technological Differences2.1 Absolute and Comparative Cost Advantages

Basic Model – Assumptions:

I Two countries i = 1, 2

I Two homogeneous goods y ij with j = 1, 2

I One homogeneous production factor (labor) L

I No transportation costs, no transaction costs etc.

I Linear production function:

y ij = qi

jLij

with labor productivity qij =

y ij

Lij

⇒ labor coefficient aij =

1

qij

=Li

j

y ij

S.21


I Country i is endowed with labor force Li .I Full employment:

Li1 + Li

2 = Li , i = 1, 2

I Perfect competition: price equals marginal cost

Cost of good j :

cj = w iLij = w iai

jyij

Marginal cost of good j :

mcj = w iaij = pi

j

⇒ pi =pi

1

pi2

=ai

1

ai2

where pi is the relative price of good 1.

The relative price pi denotes the opportunity costs of good 1.S.22


Transformation curve:

y i1 =

1

ai1

Li1 ⇒ Li

1 = ai1y

i1

y i2 =

1

ai2

Li2 =

1

ai2

(Li − Li1) =

1

ai2

(Li − ai1y

i1)

y i2 =

Li

ai2

− ai1

ai2

y i1

⇒ −dy i2

dy i1

=ai

1

ai2

=pi

1

pi2

= pi

The marginal rate of transformation equals the relative price ofgood 1.

S.23


L1 Y2

L2

Y1

L1 + L2 = L

transformation curve

S.24


Two countries with different technologies = different laborproductivities

Adam Smith (Wealth of Nations, 1776) considers the case thateach country has absolute cost advantages: a1

1 < a21 and a1

2 > a22.

Good 1 (apple) has a lower price in country 1, and good 2 (peach)has a lower price in country 2.

These absolute cost advantages imply relative advantages:

p1 =a1

1

a12

<a2

1

a22

= p2

S.25


Incentive for commodity exchange:

I For an apple from country 1 you get more peaches whenexporting it into country 2. Peaches become therefore lessscarce in country 1.

I For a peach from country 2 you get more apples whenexporting it to country 1. Apples becomes less scarce incountry 2.

Each country specializes in the production of the good where it hasa comparative cost advantages. On the world market the relativeprice will be p = p1/p2 with

p11

p12

<p1

p2<

p21

p22

We call p = p1/p2 the relative price on the world market or theterms of trade or the real exchange rate.

S.26


What happens if one country has absolute cost advantages in bothsectors?

David Ricardo (Principles of Political Economy and Taxation,1817) considers that comparative cost advantages are sufficient.Mutually beneficial division of labor (specialisation) is possibleeven if one country is less productive in both sectors:

p1 =p1

1

p12

=a1

1

a12

<a2

1

a22

=p2

1

p22

= p2

does not imply any specific inter-country relations a1i ≶ a2

i

Example: a11 = 5, a1

2 = 10, a21 = 6, a2

2 = 18 impliesp1 = 1/2 > 1/3 = p2.

S.27


I Obviously, absolute cost advantages imply comparativeadvantages, but not vice versa. Ricardo’s approach is thereforemore general. The trade theory which is based ontechnological or prductivity differences is therefore calledRicardian.

I Ricardo’s argument could be taken as an explanation for theexistence of international division of labor (positive theory).Originally it was also a normative argument to liberalize tradewith the aim to have welfare gains from specialization.

S.28

2. Technological Differences2.2 Equilibrium, Welfare and Income

Welfare effects from trade:

I Without loss of generality we assume p1 < p2.

I We assume that we have a welfare function W (y i1, y

i2) (for

both countries) with convex indifference curves in the(y i

1, yi2)-space. For the sake of simplicity we assume that a

welfare function is the utility function of the representativehousehold. The arguments are the consumed quantities.

S.29


Autarky case:

From

maxy i

1,yi2

W (y i1, y

i2) s.t. y i

2 =Li

ai2

− ai1

ai2

y i1

we have the tangential solution

∂W /∂y i2

∂W /∂y i1

=ai

1

ai2

=pi

1

pi2

= pi

y2

y1

W (y1, y2)

y∗2

y∗1

S.30


Free trade:

Country 1 specialises on good 1 and produces y∗i = L1/a11. From

selling good 1 with price p1 we have the budget constraint

p1y∗1 = p1y

11 + p2y

12

⇒ y12 =

p1

p2y∗1 −

p1

p2y1

1

=p1

p2

L1

a11

− p1

p2y1

1

Maximizing welfare under this constraint gives

∂W /∂y i2

∂W /∂y i1

=p1

p2≡ p

where p1/p2 > p11/p

12 compared to the autarky case. Starting from

y∗1 the budget constraint is steeper, implying that the tangentialpoint under free trade has a longer distance to the origin than incase of autarky (⇒ higher welfare). S.31


The physical exports of one country must equal the physicalimports of the other country:

y∗1 − y11 = y2

1 , y∗2 − y22 = y1

2 (1)

An equilibrium implies that the expenditures for import equal theearnings from export:

p1(y∗1 − y11 ) = p2y

12 , p2(y∗2 − y2

2 ) = p1y21 (2)

Combining (1) and (2) gives

p1

p2=

y12

y21

=y∗2 − y2

2

y∗1 − y11

≡ p

The relation of imported goods equals the relation of exportedgoods and this equals the price relation.

S.32


y2

y1

Country 1

W 1P1 = C 1

t.o.t.

P1∗

W 1∗

C 1∗

Im2

Ex1

y2

y1

Country 2

W 2

P2 = C 2

t.o.t.

P2∗

Ex2

Im1

W 2∗

C 2∗

(P = production point, C = consumption point,

with “*” is free trade, without “*” is autarky)

S.33


I The terms of trade p = p1/p2 determine how the additionalwealth is distributed across the countries.

I In extreme cases only one country benefits (note: voluntaryexchange!)

I “Improving” terms of trade means that you get more importgoods for one unit of the export good.

S.34


Relative supply and relative demand:Assume again p1 < p2.

I Considering a relative price p < p1 < p2 then both countrieswill specialize in good 2 and the relative supply y∗1 /y

∗2 is zero.

I Considering a relative price p1 < p2 < p then both countrieswill specialize in good 1 and the relative supply y∗1 /y

∗2 is

infinity.

I Considering a relative price p1 < p < p2 then each countryspecializes in the good where it has a comparative costadvantage and the relative supply y∗1 /y

∗2 has a positive finite

value.

I The relative demand y1/y2 = (y11 + y2

1 )/(y12 + y2

2 ) isnegatively sloped in the relative price p due to the substitutioneffect. We assume identical and homothetic preferences.

S.35


p = p1

p2

y 11 +y 2

1

y 12 +y 2

2

a11/a

12

a21/a

22 relative supply

L1/a11

L2/a22

p∗

relative demand

S.36


I Typically it is assumed that relative demand intersects relativesupply in the region [p1, p2], determining terms of trade p inthe interior of the interval.

I Depending on the terms of trade the wealth surplus isdistributed to both countries.

I Since both countries may be very different regarding theirsize, countries may not specialize completely (as in standardtextbook example).

S.37


General result from Ricardian model:

I Trade is possible and beneficial when there are comparativecost advantages.

I The structure of comparative advanatages determines thetrade pattern.

I The reason why countries with comparative disadvages inboth sectors (overall low productivity) also gain from trade isthat the wages are lower than in the other country. A lowerwage compensates the disadvantage of higher cost due tolower productivity.

S.38


I Assume a11 > a2

1, a12 > a2

2 but a11/a

12 < p < a2

1/a22.

Hence country 1 specializes in good 1 and country 2specializes in good 2.

I Take good 2 as a numeraire.I From price = marginal cost we have for the real wages w i :

country 1 w1 =p

a11

and w2 =1

a22

country 2

From p < a21/a

22 and a1

1 > a21 we have

w1 =p

a11

<1

a22

= w2

I A country which has an absolute disadvantage butcomparative advantage in good 1 will have a lower real wage.The trade pattern depends on comparative advantages, butthe absolute advantages determine the level of wages.

S.39


Some partialities/misconceptions

1. “A country must be able to produce a good more efficient = toless cost than the foreign country, it must have a competitiveadvantage, otherwise it cannot gain from trade.”

This refers to Smith’s theory of absolute cost advantages. Ricardo’s

theory proves this claim as false! Trade is also possible and beneficial if a

country has a lower productivity = higher costs in all sectors. A difference

of opportunity costs is a sufficient argument. The less productive country

has lower relative wages (and lower nominal wages expressed in the same

currency than the more productive country). This is translated to a lower

relative price for the good with the lower opportunity costs.

S.40


2. “Trade and international competition harms the workers becausethe price advantage depends on low wages. The country with lowwages is exploitet.”

If the comparative advantage does not come from a high productivity, it

comes from low wages. The workers would be harmed if the real wages

would suffer from trade. This is not the case – at least in this simple

model. For one working hour, more goods could be purchased compared

to the autarky case. The development of the absolute level of wages

depends on the development of productvities.

S.41


Example: USA and some fast developing countries

Note that the overall wage level is determined by the productivityof all sectors, not only of the exporting sector.

wage [$/h] 1975 wage [$/h] 2000(USA = 100) (USA = 100)

USA 100 100South Korea 5 41Taiwan 6 30Singapore 13 37

(Source: Bureau of Labor Statistics, see Krugman/Obstfeld)

S.42


Back to the “exploitation” argument:

I If you compare wages between a rich and a poor country thenyou have to take into account

I differences in purchasing power (foreign wages should bemultiplied with PPP adjusted exchange rates)

I different labor productivity levels in the specific sector

I If wage differentials could not be explained by thesedifferences then the theory might be incomplete. E.g.monopsonistic labor markets?

I Then one have to ask which wage differential would have beenrealized in case of autarky.

S.43

2. Technological Differences2.3 Extension to Specific Factors or Increasing Costs

Ricardo-Viner-Model:

I Ricardo: One sectoral mobile production factor (labor) withconstant returns

I Now: an additional sectoral immobile production factor (e.g.land, sector-specific capital, sector-specific human capital),here denoted as C

I Even if we assume a production function with constantreturns, e.g. yi = C 1−α

i Lαi , then yi is concave in Li

(decreasing marginal productivity = increasing costs)

I The result is a concave transformation curve.

S.44


L1 Y2

L2

Y1

L1 + L2 = L

Transformation Curvewith decreasing marginal returns

price relation

S.45


Allocation of labor in the economy:

w w

L (total)

∂Y1

∂L1· p1 ∂Y2

∂L2· p2

w∗ w∗

L1 L2

S.46


Change of goods prices:

w w∂Y1

∂L1· p1 ∂Y2

∂L2· p2

w∗ w∗

L1 L2

∂Y1

∂L1· p′1

w∗∗ w∗∗

L′1 L′2

S.47


Since we have partial specialisation, a smooth change in therelative prices induces a smooth change in the relative supply:

p

y 11 +y 2

1

y 12 +y 2

2

relative demand

relative supply

p∗

y 11 +y 2

1

y 12 +y 2

2

S.48


Technological change in one sector:

L1 Y2

L2

Y1

L1 + L2 = L

t.o.t.

S.49


w w∂Y1

∂L1· p1 ∂Y2

∂L2· p2

w∗ w∗

L1 L2

∂Y new2

∂L2· p2

w∗∗ w∗∗

L′1 L′2

Specialization to the good which utilizes the specific factor whose endowment

has been increased. ⇒ similar to Heckscher-Ohlin model (next chapter)

S.50

2. Technological Differences2.4 Empirical Relevance

Empirical studies show that differences in productivity (technology)play a much more important role than the relative endowments offactors (detailed discussion in Feenstra).

From the Ricardo model we have concluded: “The trade patterndepends on comparative advantages, but the absolute advantagesdetermine the level of wages.”

⇒ Differences in productivity should correlate with differences inwages.

S.51


(Source: Feenstra)

S.52


Consider many sectors in twocountries. The relation of exportsin one sector should correlatewith the relation ofproductivities: The higher theproductivity in country 1 (USA)relative to country 2 (GB), thehigher are the exports of country1 relative to country 2.

(Source: Krugman/Obstfeld)

S.53


I Balassa, B. (1963), An Empirical Demonstration of ClassicalComparative Cost Theory. Review of Economics and Statistics4, 231-238.

I Leromain, E., Orefice, G. (2014), New Revealed ComparativeAdvantage Index: Dataset and Empirical Distribution.International Economics 139, 48-70

S.54

3. Resource Endowment3.1 Preliminaries from Production Theory

I We now argue with 2 countries, 2 homogenous goods, and 2production factors: labor L and capital C .

I The production function is assumed to have “neoclassical”features: both factors are substitutable, decreasing marginalreturns, constant returns to scale.

I Both countries have the same production technology. We haveto examine some basic implications from production theory.

S.55


Producing one unit to minimal cost:

minLi ,Ci

wLi + rCi s.t. yi (Li ,Ci ) = 1

First order conditions lead to:

w

r=∂yi/∂Li

∂yi/∂Ci= −dCi

dLi

Marginal rate of substitution (MRS)equals factor price relation.

C

L

y(C , L) = 1

slope: −w/r

aC

aL

minimal cost combination

S.56


Let aLi , aCi be the solution of the cost minimization problem forone unit of output of good i ⇒ total inputs are:L = aL · y , C = aC · y .

The unit costs can therefore be written as

ci (w , r) = w · aLi (w , r) + r · aCi (w , r)

Perfect competition implies zero profits:

pi = ci (w , r) (3)

Note that the unit cost are different for i = 1, 2 as long as thefactor intensities aLi/aCi differ.

S.57


Edgeworth-Box and Contract Curve:

L1

sector 1C1

isoquant y1

L1

C1

C2sector 2

L2

C2

L2

L1

sector 1C1

C2sector 2

L2total capital endowment

tota

lla

bor

end

owm

ent

y1

y2

L1

C1

C2

L2

inefficient allocation

y ′1

efficient allocationL∗1

C∗1

C∗2

L∗2y2

contract curve

(capital intensive)

(labor intensive)

Combining both diagramssuch that chosen inputcombinations overlap

⇒Box with sizeC = C1 + C2 andL = L1 + L2

S.58


Contract Curve ⇒ efficient resource allocation and production

For each point on the contract curve we have identical MRS inboth sectors:

w

r=∂y1/∂L1

∂y1/∂C1=∂y2/∂L2

∂y2/∂C2

The slope of a line from the origin to the point on the contractcurve reflects the intensity of factors utilized in the production ofthe good (here: good 1 is capital-intensive, good 2 islabor-intensive).

If the contract curve is only above or only below the connection ofthe origins, then we have unambiguous factor intensities⇒ non-reversal of factor intensities.

S.59


Reversal of factor intensities:

L1

sector 1C1

C2sector 2

L2

good 1 is capital-intensivegood 2 is labor-intensive

good 1 is labor-intensivegood 2 is capital-intensive

S.60


Profit maximization and factor prices:

From profit maximization under perfect competition we have realreturns of the factors according to their marginal productivity:

∂yi

∂Li=

w

pi,

∂yi

∂Ci=

r

pi

Total differential of the production function

dyi =∂yi

∂LidLi +

∂yi

∂CidCi

Using the condition: marginal productivity = real return

dyi =w

pidLi +

r

pidCi

S.61


Relation of goods prices and marginal rate of transformation:

The marginal rate of transformation (slope of the transformationcurve) is given by

dy2

dy1=

wp2dL2 + r

p2dC2

wp1dL1 + r

p1dC1

=p1

p2

(wdL2 + rdC2

wdL1 + rdC1

)Full employment implies L = L1 + L2,C = C1 + C2 and hencedL1 = −dL2, dC1 = −dC2. Therefore we have

dy2

dy1= −p1

p2(4)

Marginal rate of transformation equals goods price relation!

This is equivalent to maximizing the GDP (see Feenstra):

maxy1,y2

= p1y1 + p2y2 s.t. y2 = y2(y1, L,C )

S.62


Unambiguous relation between goods prices and factorprices:

Lemma: If both goods areproduced and if there areno factor intensityreversals, then each pricevector (p1, p2) correspondswith a unique factor pricevector (w , r).

r

w

p1 = c1(w , r)

p2 = c2(w , r)

r∗

w∗

(p1, p2 determined on world market)

Zero profit conditions forthe labor-intensive and thecapital-intensive good

S.63


The “cone of diversification”:

Recall, that we have from cost minimization

ci (w , r) = waLi + raCi (5)

with aLi (w , r), aCi (w , r) as the optimal factor demand per unit.

Remark: The envelope theorem applies and we have

∂ci

∂w= aLi

Full employment requires

aL1 · y1︸︷︷︸L∗1

+ aL2 · y2︸︷︷︸L∗2

= L (6)

aC1 · y1︸︷︷︸C∗1

+ aC2 · y2︸︷︷︸C∗2

= C (7)

S.64


I Let (w , r) be given. This determines the per unit factorrequirement aLi (w , r), aCi (w , r).

I Due to linear homogeneity, the MRS is constant along anexpansion path. The vectors (aLi , aCi ) in the (L,C )-spacecharacterize such expansion paths where goods i = 1, 2 areproduced to minimal cost.

I Let (L,C ) be the positive factor endowment of the country. Iffactor prices and hence (aLi , aCi ) are given, then there is aunique solution of (6) and (7) which determines the outputsy1, y2. The outputs determine the length of the vectors(aLi , aCi ).

S.65


C

L

y1 = 1minimal cost

y1 = 2

y1 = 3

expansion path 1

expansion path 2(capital-intensive sector)

(labor-intensive sector)y2 = 4

y2 = 8

y2 = 6

C2

L2

C1

L1

C1 + c2 = C endowment vector

L = L1 + L2

“cone of diversification”

S.66


I If the factor endowment vector lies in the “cone ofdiversification”, then both goods are produced with positivequantities. Otherwise the country is fully specialized to onegood.

The other way round:

I The relative goods prices p1/p2 are determined by the demandside (e.g. by the world market). This determines the desiredoutput bundle (y1, y2) (isoquants in the figure). The commontangential point with the relative factor price line determines(w , r) and henceforth the (aLi , aCi ) vectors and the cone.

I For extreme goods prices, the economy produces at oneendpoint of the transformation curve (full specialization, nounambiguous relation between goods and factor prices).

S.67


Summary:

I We have 2 full employment and 2 zero profit conditions:

aL1(w , r) · y1(p1, p2) + aL2(w , r) · y2(p1, p2) = L

aC1(w , r) · y1(p1, p2) + aC2(w , r) · y2(p1, p2) = C

p1 = c1(w , r)

p2 = c2(w , r)

these are 4 equations with 4 variables p1, p2,w , r .

I From the demand side (or world market) p1, p2 are given.

I In case of partial specialization y1 > 0, y2 > 0 there is aunique solution for (w , r).

I Remark: Two countries with the same technology andhenceforth (aLi , aCi ) which face the same goods prices willhave the same factor prices! (to be shown later on)

S.68

3. Resource Endowment3.2 The Neoclassical Baseline Model (Heckscher/Ohlin)

Assumptions:

I 2 goods, 2 countries, 2 production factors C , L

I identical production function in both countries:

y1i (C 1

i , L1i ) = y2

i (C 2i , L

2i ), i = 1, 2

with∂y j

i

∂x> 0,

∂2y ji

∂x2< 0 x = C j

i , Lji

and constant returns to scale

I factors are mobile between sectors but immobile betweencountries

I unambiguous factor intensities of good i = 1, 2 (no reversals)

S.69


Assumptions: (continued)

I perfect competition, full employment

I no transportation cost, no transaction cost

I identical homothetic preferences

I different relative factor endowments (w.l.o.g.):

C 1

L1>

C 2

L2

S.70


I Suppose that good 1 is capital intensive (see next figure).

I For two countries with different relative factor endowments wehave two different transformation curves.

I Autarky: Employing the same scheme of welfare indifferencecurves (identical preferences!), we obtain different relativeprices p1

1/p12 6= p2

1/p22 .

⇒ Sufficient condition for trade.

S.71


Good y1 capital-intensive

Good y2 labor-intensivey1

2

y11

Country 1 (capital abundant)

relative price

P = C

y12

y11

Country 2 (labor abundant)

relative price

S.72


I The terms of trade will be in between the autarky pricerelation:

p1 < p < p2

⇒ Specialization: production and consumption points differ.

⇒ Export and import

⇒ Increased welfare!

Theorem (Heckscher 1919, Ohlin 1924):

Under the given conditions each country specializes on theproduction of the good (and export it) which uses its abundantfactor intensively.

[The relative capital abundant country specializes on the capitalintensive good (and exports this), while the relative labor abundantcountry specializes on the labor intensive good (and exports this).]

S.73


From autarky to trade:

Good y1 capital-intensive

Good y2 labor-intensivey1

2

y11

Country 1 (capital abundant)

WP = C

t.o.t.

P′

W ′

C ′

imp

ort

export

y22

y21

Country 2 (labor abundant)

W

P = C

t.o.t.

P′

W ′

C ′

impot

exp

ort

S.74

3. Resource Endowment3.3 Relations between commodity prices, factor prices and factor endowment

I For a globally efficient allocation, the production factorsshould be directed into the most productive opportunities,implying a global equalization of factor prices. With free tradewe have a tendency into this direction even without globalfactor mobility!

I In a labor abundant country, labor is relative cheap comparedto scarce capital.

I If this country specializes on the production of the laborintensive good according to the HO theorem, labor demandincreases and induces a rise of wages.

I Vice versa for capital in the other country.

S.75


I Result: Tendency to equalize factor prices.

I Mobitlity of goods (trade) compensates the immobility ofproduction factors.

I If a country is labor abundant and it specializes on the laborintensive good, then it’s export goods embody more laborthan the imported and the consumed bundle of goods. So theforeign country utilizes the domestic labor force(Heckscher-Ohlin-Vanek approach, discussed later on).

Factor Price Equalization Theorem (Samuelson 1949):

Under the assumption of the HO model the factor prices (w , r)equalizes across the countries if there is free trade.

S.76


There are some empirically relevant violations of the underlyingassumptions which may prevent from factor price equalization:

1. Full specialization – the goods price relation is not atangential point of the transformation curve, henceforth thereis no unambiguous relation between pj

1/pj2 and w j/r j .

2. Different technologies/labor productivities (like in theRicardian model).

3. Reversals in factor intensity of production – it is no longerclear which good is “capital intensive”, and hence thespecialization pattern is ambiguous and may reverse.

4. Goods prices do not equalize due totransaction/transportation cost and trade barriers.

S.77


Edgeworth-Box Analysis:

I According to the assumptions of the HO theorem we constructtwo Edgeworth-Boxes where the size of the boxes differaccording to the different (relative) endowments of factors.

I Asuume that point M and N are chosen.

I Since they are on a contract curve, we have

MPL11

MPC 11

=MPL1

2

MPC 12

in M (8)

MPL21

MPC 21

=MPL2

2

MPC 22

in N (9)

(marginal product of labor/capital)

S.78


lab

or

capitalsector 1

sector 2country 1

sector 2country 2

M

N

S.79


I Since the production functions are linear homogenous, themarginal products of both factors are constant along anexpansion path starting in the origin.

I Thus it follows

MPL11

MPC 11

(in M) =MPL2

1

MPC 21

(in N) (10)

and from (8) and (9) we have

MPL12

MPC 12

(in M) =MPL2

2

MPC 22

(in N) (11)

I We have factor prices according to their marginal product.Thus, it follows

w1

r1=

w2

r2

S.80


But do the points M and N represent a trade situation?

I If M and N represent a specialization pattern according to theHO theorem, then we must have the same goods pricerelation = the same marginal rate of transformation.

I Under the assumption of perfect competition we have fromthe profit maximization condition

p11 ·MPL1

1 = w1 = p12 ·MPL1

2 (12)

⇒ p11

p12

=MPL1

2

MPL11

(13)

and for country 2p2

1

p22

=MPL2

2

MPL21

(14)

S.81


I Along the expansion path we have constant marginalproductivities. Since M and N are on the same expansion pathwe have:

MPL11 = MPL2

1, MPL12 = MPL2

2

I Thus, it follows

p21

p22

=p1

1

p12

=p1

p2(t.o.t.)

I Rearranging (12) to the absolute real wage

w1

p11

= MPL11 = MPL2

1 =w2

p21

implies that absolute real wages are equalized in case offree trade.

S.82


Harrod-Johnson diagram:

I From the properties of the production function and the costminimization calculus we have the result that the optimalcapital intensity Ci/Li depends positively on the relative factorprice w/r .

I By assumption, for each relative factor price one good isunambiguously the capital intensive good while the other islabor intensive.

I Each point on the contract curve is characetrized by w/l , andit corresponds with a point on the transformation curve, andhenceforth with a relative goods price pj

1/pj2.

S.83


lab

or

capitalsector 1

sector 2

A

factor price relations

B

C

y2

y1

A

goods price relations

B

C

p2

p1relative price oflabor intensive good 2

wr

relative price of labor

A

B

C

S.84


The Harrod-Johnson diagram:

p2

p1

Ci

Li

wr

goods/factor price relations

labor intensivegood 2

capital intensivegood 1

factor price / factor intensity relations

(CL

)

(C1L1

)(p2p1

) (C2L2

)

(w/r)

(C1L1

)∗(p2p1

)∗ (C2L2

)∗

(w/r)∗

S.85


I For a given relative factor price w/r and relative goods pricepj

2/pj1 the Harrod-Johnson diagram shows the chosen

capital/labor ratios in both sectors.

I Since we have assumed full employment when deriving thecurves, it is ensured that the average capital/labor in bothsectors (sectoral ratio weighted with the outputs) is equal tothe relative factor endowment of the country.

I Given the relative factor endowment we can derive an intervalfor the relative factor price which ensures the production ofboth goods.

I Now we have trade and relative goods prices according to thet.o.t. Both countries specialize to the good which utilizes theabundant factor intensively. As the Harrod-Johnson diagramshows, this implies the convergence of relative factor prices.

S.86


Factor price equalization:(Both cuntries use the same technology!)

p2

p1

Ci

Li

wr

goods/factor price relations

labor intensivegood 2

capital intensivegood 1

factor price / factor intensity relations(C1

L1

) (C2

L2

)(country 2 is capital abundant)

p22

p21

p12

p11

p2p1

t.o.t.

S.87


Dixit-Norman diagram:

I We combine the Edgeworth-Box analysis with the idea ofdiversification cones in order to represent two countries andtwo sectors in one diagram.

I Height and width of the box are determined by the worldendowments of capital and labor.

I If goods prices are equalized, then factor prices are equalizedas well. Therefore, also the expansion paths are identical inboth countries.

S.88


The Dixit-Norman diagram:

C

country 1 L

C 1

L1

exp. path 1

exp. path 2

y12 = 4

y11 = 3

C

country 1L

P

FPE set

yD2 = 10

yD1 = 5

C

country 2L

C 2

L2

exp. path 1

exp. path 2

y22 = 6

y21 = 2

C

country 2L

C 2

L2

exp. path 1

exp. path 2

y22 = 6

y21 = 2

C

country 2L

C 2

L2

exp. path 1

exp. path 2

y22 = 6

y21 = 2

S.89


I With given (w , r) also (aLi , aCi ) is given which is identical in bothcountries. Therefore we have two diversification cones which build aparallelogram.

I Since (w , r) and (p1, p2) are unambiguously correlated we have theworld demand for both goods: dw

1 , dw2 . These are represented by the

points Dwi = (aLi , aCi )d

wi on the vertexes of the parallelogram.

I Point B lies in both diversification cones. Full employment impliesthat the factors utilized in the production of both goods must sumup to the country’s factor endowment. Thus, country 1 producesy1

1 , y12 , and country 2 produces y2

1 , y22 which sum up to the world

demand Dw1 ,D

w2 .

I Result: Each resource allocation B in the parallelogram representsfree trade with factor price equalization. Therefore, theparallelogram is called Factor Price Equalization (FPE) set.

S.90


I Assume that factors are mobile across countries. As aconsequence there will be factor price equalization, and theresulting endowment vector is in the displayed FPE vertex.

I Now assume that factors are immobile across countries andthe endowment vector is exogenously given. In case of freetrade, (p1, p2) and hence the world demand is given. Thespecialization and corresponding outputs can be seen in thegraphic.

I As factor prices will equalize according to Samuelson (1949)theorem, the endowment vector must be within the FPE set.

S.91


Change of goods prices (terms of trade):

I What happens with the factor prices when we come from autarky tofree trade or if the t.o.t. change for other reasons?

I Recall, that the unit cost function (5) is given byci (w , r) = aLiw + aCi r , and from perfect competition we have

pi = ci (w , r) = aLiw + aCi r

Total differentiation leads to

dpi = aLidw + aCidr

Dividing both sides by pi = ci

dpi

pi=

waLi

ci

dw

w+

raCi

ci

dr

r

⇒ pi = θLi w + θCi r

with θLi + θCi = 1. Price changes (in %) are linearly related tochanges of the factor prices (in %), depending on their unit costshares. S.92


I We have the linear system(p1

p2

)=

(θL1 θC1

θL2 θC2

)(wr

)I Assume that y1 is labor intensive: θL1 − θL2 > 0 while good 2

is capital intensive: θC2 − θC1 > 0.

I Assume a change of the terms of trade where the laborintensive good becomes relative more expensive: p1 − p2 > 0.

I Solving the linear system above gives (see Feenstra)

w =(θC2 − θC1)p1 + θC1(p1 − p2)

(θC2 − θC1)> p1

r =(θL1 − θL2)p2 − θL2(p1 − p2)

(θL1 − θL2)< p2

S.93


As a result, we havew > p1 > p2 > r

Furthermore, this is consistent with the fact that due topi = θLi w + (1− θLi )r the price change must be in between theprice changes of the factors.

Stolper-Samuelson theorem (1941):

An increase of the relative price of a good will increase the realreturn to the factor (real factor price) which is intensivelay used forthis good, and it will reduce the real return of the other factor.

[Here: w/p1 increases while r/p1 decreases.]

S.94


Illustration of the Stolper-Samuelson theorem:

r

w

p2 = c2(w , r)

p1 = c1(w , r)

p′1 = c1(w , r)

p1

p2↑ ⇒ w

r ↑

S.95


Change of endowment:

I Assume a change in the factor endowment when the productprices are fixed and exogenously given.

I Total differentiation of the full employment condition gives

aL1dy1 + aL2dy2 = dL (15)

aC1dy1 + aC2dy2 = dC (16)

Dividing by L resp. C and re-arranging gives

y1aL1

L

dy1

y1+

y2aL2

L

dy2

y2=

dL

L(17)

y1aC1

C

dy1

y1+

y2aC2

C

dy2

y2=

dC

C(18)

S.96


or in growth rates:

λL1 · y1 + λL2 · y2 = L (19)

λC1 · y1 + λC2 · y2 = C (20)

where λji is the fraction of the factor j which is employed in sectori . Obviously, we have λj1 + λj2 = 1.

S.97


I We have the linear system(λL1 λL2

λC1 λC2

)(y1

y2

)=

(L

C

)I Again, assume that sector 1 is labor-intensive and sector 2

capital-intensive, implying λL1 > λC1 and λC2 > λL2.

I Assume a change in the endowment such that L > 0, C = 0.

I From solving the linear system we have (see Feenstra):

y1 =λC2

(λC2 − λL2)L > L > 0

and

y2 =−λC1

(λC2 − λL2)L < 0

S.98


Summarizing the analysis we have in case of an increase in laborforce:

y1 > L > 0 > y2

Rybczynski (1955) theorem:

An increase in the endowment of one factor will increase theoutput of the sector which uses this factor intensively, and decreasethe output of the other sector.

S.99


Illustration of the Rybczynski theorem:

y2

y1

t.o.t.

S.100


Discussion:

I “Dutch Disease”: The discovery of oil at the Netherland’scoast led to a boom of their oil industry. This led to acontraction of other industries since resources are redirectedto the oil industry.

I The dependency of developing countries on a specific naturalresource which is abundant in this country, may prevent theeconomy from developing other industries. New discoveries ofthat resource may discourage the country from economicdevelopment.

I In OECD countries we have stagnation in the growth of laborforce, but an accumulation of capital (physical and human).Labor intensive industries, i.e. such with use unskilled laborshrink in absolute terms.

S.101

3. Resource Endowment3.4 The Empirical Picture

The Leontief (1953) Paradoxon:

Leontief analysed the exports and imports of the US industry (datafrom 1947). He calculated the amount of capital and labor tocreate $1 million worth of exports or imports. The US could beregarded as a capital abundant country.

US 1947 Exports Imports

Capital ($ million) 2.5 3.1Labor (person-years) 182 170capital/labor ($/person-year) 13700 18200

The result seem not to confirm the prediction of the HO approach.

S.102


Possible explanations for the result:

1. Data are untypical because of World War II; biased dataset

2. Protectionism distorts the specialization pattern.

3. The technologies of the US and foreign industries differ (as inthe Ricardo case)

4. Other factors than labor and capital are neglected (e.g.human capital, different labor skills!)

5. The test is misleading (Leamer 1980): One should measurehow much labor is embodied in exports compared to imports,not the technologically determined labor intensity of exportand import goods. Then you should compare the laborintensity of what is produced (which is partially exported)with the labor intensity of what is consumed (which ispartially imported).

S.103


The Heckscher-Ohlin-Vanek approach:

I The approach is formulated for n goods, m factors and lcountries. We will concentrate on the 2-2-2-case.

I From cost minimization we have the optimal factor demandfor one unit given by the matrix

A =

(aL1 aL2

aC1 aC2

)I Let Y j = (y j

1, yj2)T be the vector of outputs of country j .

I Let D j = (x j1, x2j )T be the vecor of demand of country j .

I Then T j = Y j − D j is the vector of net exports of country j .I The factor content of traded goods is

F j =

(F j

L

F jC

)= AT j

The goal is to derive a relation between factor endowment of country j and thefactor content of the traded goods.

S.104


I Total factor endowment (with full employment assumption) isgiven by

V j = AY j

I Let Dw be the vector of world demand and V w = AY w theworld endowment of factors.

I Since preferences are identical and homothetic, and since thegoods prices are equalized due to trade, we have the sameproportions in the consumed bundle of goods in all countries.Hence, the demand of country j is a fixed share of the worlddemand:

D j = s jDw

S.105


I World consumption must equal world production: Dw = Y w .

I Multiplying with A gives the factor content of consumption orproduction. Thus we have for country j (see Feenstra):

AD j = s jADw = s jAY w = s jV w

⇒ F j = AT j = V j − s jV w (21)

or in terms of an individual factor:

F ji = V j

I − s jV wi

S.106


Theorem (Leamer 1980):

If capital is abundant relative to labor in country j , then (21)implies that the capital/labor relation embodied in production ofcountry j exceeds the capital/labor relation in country j ’sconsumption:

C j

Lj>

C j − F jC

Lj − F jL

S.107


Proof:

F jC = C j − s jCw ⇒ Cw =

1

s j(C j − F j

C )

F jL = Lj − s jLw ⇒ Lw =

1

s j(Lj − F j

L)

⇒ C j

Cw=

s jC j

C j − F jC

andLj

Lw=

s jLj

Lj − F jL

⇒ C j

Cw>

Lj

Lw⇐⇒ C j

Lj>

Cw

Lw

(= country j is capital abundant)

⇒ s jC j

C j − F jC

>s jLj

Lj − F jL

⇒ C j

Lj>

C j − F jC

Lj − F jL

The capital intensity in production exceeds the relative capital content ofconsumed goods.

S.108


Differenecs of relative factor content in produced and consumedbundle of goods:

country 1

cap

ital

country 2

labor

FPE set

P

C

FL

FC

FL = labor embodied in traded goods

FC = capital embodied in traded goods

S.109


With Leontief’s dataset the test is reproduced on the basis of theHOV approach. The table shows the amount of capital/laborembodied in production and consumption.

US 1947 Production Consumption

Capital ($ million) 327 305Labor (person-years) 47 million 45 millioncapital/labor ($/person-years) 6949 6737

The result is in line with the HOV approach.

S.110


I Later tests (e.g. Baldwin (1971)), however, show that also theHOV model performs poor in explaining the empirical data.

I Consequence: Trade patterns cannot be explained only byfactor endowment differences!

I An extensive discussion of the empirical picture of the HOVtheorem can be found in Feenstra.

S.111

3. Resource Endowment3.5 Combining Ressource Endowment and Technological Differences

Approach of Trefler (1993):

I Different technologies imply different optimal factor demand,i.e. differences in the matrix A = Aj .

I The productivities of a factor could be expressed in relation tothe productivity of a benchmark country (e.g. USA) whichproductivity is normalized to one. A higher productivityappears like an augmentation of the factor. Hence, theendowment should be measured in “efficiency units”.

I Let πji be the productivity of factor i in country j relative to

the USA.

I In terms of efficiency units we calculate the effectiveendowments as πj

iVji .

S.112


The HOV theorem can then be re-written as

F ji = πj

iVji − s j

∑k

πki V

ki

(where the world consists of several countries with summationindex k)

I Problem: For almost all datasets you can find πji such that the

system of equation holds true with equality, so that thetheorem is not testable.

⇒ One have to impose some empirically based restrictions on πji .

I For empirical tests (test procedures and results) see Feenstra.

S.113


Summing up:

I The trade (specialization) pattern is determined byI differences in technologies/productivities andI differences in factor endowments

I Trade leads to an equalization of relative goods prices.I The equalization of factor prices is limited:

I Tendency to equalization (HOV approach) only betweentechnologically similar countries

I Differences in real wages could partially explained byproductivity differences and other violations of the HOVapproach.

I Taking human capital (skilled labor) into account, theRybczynski approach argues that due to accumulation ofhuman capital by skilled workers (education) unskilled workerwill lose.

S.114


Problems:

I Technology based trade as well as endowment based tradepredict specialization and exchange of different goods: exporty1 from sector 1 and import y2 from sector 2. To a largeextent international trade is intra-industrial! This is notexplained yet.

I Trade happens to a large extent between countries withsimilar technologies and similar factor endowments, e.g.OECD countries. This is not explained yet.

S.115

4. Intermediate Products and Global Value Chains4.1 Role and Measurement of GVC

Value chains / production networks:

I Segmentation of the production process into different parts

I Intermediate goods or “tasks” (could be physical or services)

I Each task produces a “Value Added” and creates income ofproduction factors.

I With small transportation costss and low trade barriers, thesetasks could be allocated to different countries:

I Off-shoring: domestic firm build up production plant forintermnediate good/task in a foreign country (becoming aMNF).

I Out-sourcing: Intermediate good/task is not produced withinfirm but bought from another foreign firm.

⇒ Global Value Chains (GVC)

S.116


Country 1 Country 2 Country 3

Inputs

Final good X

export

export

Breaking up this box ...

S.117


Country 1 Country 2 Country 3

Inputs

Final good X

export

export

Inputs

Task 4

Inputs

Task 3

Task 1

Inputs

Task 2

Inputs

S.118


Some literature:

I Gereffi, G., M. Korzeniewicz (1994), Commodity Chains andGlobal Capitalism. Praeger Publishers, New York.

I Grossman, G., E. Rossi-Hansberg (2008), Trading Tasks: aSimple Model of Offshoring. American Economic Review 98,978-1997.

I Lanz R., Miroudot S., H.K. Nordas (2011), Trade in Tasks.OECD Trade Policy Working Paper 117, OECD Publishing,Paris.

S.119


I A country might not have a comparative advantage for theentire good but for some “tasks”⇒ ability to attract tasks which generate Value Added.

I Especially poor countries which have advantages only indelivering resources and raw agricultural products with lowvalue added might engage in parts of GVC with higher valueadded. This might require knowledge transfer and compliancewith norms.

I Often requires Foreign Direct Investment (capital flows)

I Building “hubs”: regions as platforms for performing a coupleof tasks (instead of producing final goods). Main hubs: China,USA, Germany (see figure below).

⇒ More efficient use of global resources.

S.120


I For the economic trade analysis it is not so relevant whetherwe speak about “goods” or “tasks” as long as the driver ofspecialization patterns is still the comparative cost advantage!

I Specific issues of global production networks and theirimplications will not be discussed (intensively) in this course(“Trade and Development” by Andreas Freytag)

S.121


GVC of producing a Wimbledon tennis ball:

Source: www.themanufacturer.com S.122


GVC of producing a blue jeans:

Source: Diercke Weltatlas Digital S.123


Trade in component: three main hubs:

Source: World Bank (2017) S.124


Measurement of the role of GVC:

I Requires World-Input-Output-Database (WIOD), availablefrom WTO.

I Output of sectors which is exported and is used as an input ofanother sector (instead of “final demand”).

I Production length index: average number of production stages⇒ simple GVC = one border-crossing; complex GVC =several border-crossings during production process.

I Participation index: intensity of country sector’s engagenent inGVCs.

I Import intensity of final demand (counts the imports of partsembodied in the final good which might be produceddomestically).

World Bank (2017), Global Value Chain Report 2017: Measuring and Analyzing

the Impact of GVCs on Economic Development.S.125


Source: World Bank (2017)

S.126

4. Intermediate Products and Global Value Chains4.2 Comparative Advantages in Intermediate Goods

Trade with intermediate products or “tasks”

I We turn back to the case of homogenous goods, constantreturns to scale, and perfect competition.

I International division of labor also means that the productionchain from input of raw materials and factors to the finaloutput can be decomposed. Parts of the production processcan be “outsourced” to other countries where it is chaper toproduce an intermediate good.

I To a large extent, e.g. “German” products are manufacturedin other countries.

I As long as large parts of the value adding parts of productionare outsourced to other countries, but the final good sales areaccounted for the German trade balance, the export strengthof Germany might give a false impression (“bazar economy” –H.W. Sinn).

S.127


A simple model with intermediate goods:

I Inputs: Unskilled labor Li , skilled labor Hi , capital Ci with thefactor prices w , q, r .

I Two intermediate goods i = 1, 2 with a linear homogenousproduction function yi = yi (Li ,Hi ,Ci ).

I Let xi be the net exports of an intermediate good i = 1, 2.

I Final good sector with a linear homogenous productionfunction yn = yn(y1 − x1, y2 − x2), that is, it “bundlestogether” the intermediate goods without any additionalfactor input.

I Full employment L1 + L2 = L,H1 + H2 = H,C1 + C2 = C .

S.128


I Intermediate good firms maximize profits with a given relativeprice, and they produce to minimal costs, given the factorprices.

I We have trade only in y1, y2. Assume:I Good y1 uses intensively unskilled laborI Good y2 uses intensively skilled labor.

I HOV approach: we have production y1, y2 and net exportx1, x2, and the output of the final good (since only this goodis consumed, it is a proxy for welfare) is yn(y1 − x1, y2 − x2)(see graphic)

S.129


y2

y1

yn

y ′n

p1/p2 ↓ ⇒ yn ↑

trade volume ncreases

S.130


I Cost minimization and zero profit condition gives

pi = ci (w , q, r)

and expressed in growth rates (see above)

pi = θiLw + θiH q + θiC r (22)

where θiL + θiH + θiC = 1.

S.131


I For the two intermediate good sector we have two equationsbut three unknowns w , q, r when goods price changes aregiven.

I Therefore, we assume that in both sectors we have equalcapital cost shares θ1C = θ2C .

I It follows that θ1L + θ1H = θ2L + θ2H and henceforthθ1L − θ2L = −(θ1H − θ2H).

I The difference of (22) for both sectors can then be written as

p1 − p2 = (θ1L − θ2L)︸︷︷︸+

(w − q)

S.132


I In the graphic above, the country outsources/offshoresunskilled-labor intensive intermediate goods (as it is the casein highly developed countries) and imports y1. On the otherhand, it exports intermediate goods which use intensivelyskilled work.

I The analysis shows that a decrease of the relative price of theunskilled-labor intensive good (p1 − p2 < 0) leads to adecrease of the relative wage of unskilled labor:w − q = (p1 − p2)/(θ1L − θ2L) < 0 (since (θ1L − θ2L) > 0by definition that sector 1 is unskilled-labor intensive).

I This is according to the Stolper-Samuelson theorem.

I Increasing division of labor by looking for cost reducingoutsourcing opportunities induce a larger spread of wagesbetween skilled and unskilled workers.

S.133


I For the final goods price we have pn = θn1p1 + θn2p2, so thatin our case we have

p1 < pn < p2

I The logic of those intermediate good models is that firms nowcompare the prices of imported intermediate goods with thedomestic production costs of this good. Firms are looking inwhich countries certain parts of the production chain shouldbe located in order to minimize costs.

I It might be the case that increasing division of labor fostersthe world trade volume growth much more than the growth ofthe final good (see the empirical picture in section 1!).

I Remark: Transportation costs, transaction costs (differentlanguage, culture), quality standards problem.

S.134

4. Intermediate Products and Global Value Chains4.3 Some Implications of Globalization

Deepening the specialization patterns:

I Decline of transportation and communication cost as well as generalcosts of trade ⇒ increased fragmentation of production process(GVC) = deepening the specialization patterns ⇒ more efficient useof global production factors.

I Developing countries could attract tasks in GVC. They might nothave comparative advantages for the entire product but for specifictasks ⇒ attracting tasks with a higher Value-Added rather thanjust exporting resources and raw agricultural products.

I Requires transfer of knowledge and building up skills ⇒ importantfor long-run development ⇒ building Regional Value Chains (RVC)as a “learning platform”.

I Attracting tasks from GVC pulls other regional sectors which deliverother input factors and services.

S.135


Problems:

I Globalization by building up GVCs might enhance vulnerabilityto external shocks as GVCs are a transmission mechanism forshocks. On the other hand, larger integration might lead todiversification.

I The representative agent always has a utility advantage. Buthouseholds are heterogeneous (different endowment of laborand capital, different skills/human capital endowment)⇒ according to Stolper-Samuelson theorem (see above) somehouseholds can lose.

I Moreover, structural change requires reallocation of resources,leading to frictions and temporary unemployment.

S.136


Remarks on the political economy of globalization:

I Market regulations and institutions which shape the marketoutcome according to the regional or national preferences (e.g.environmental or consumer protection) are threatened byglobalization. Regional/national specifities are seen as a barrier totrade and might be harmonized or relaxed. This underminesdemoctratic voting about specific market rules.

* Rodrik, D. (2011), The Globalization Paradox: Democracy andthe Future of the World Economy. New York and London:W.W. Norton.

* Rodrik, D. (2017), Straight Talk on Trade: Ideas for a SaneWorld Economy. Princeton University Press.

I Trade only with mutual benefits. But what about corruption andrent-seeking? Local authorities are selling resource extractionlicences or land at too low prices against side-payments⇒ local population is losing; foreign investors could“grab”.

S.137

5. Economies of Scale and Imperfect Competition5.1 Preliminaries: Monopolistic Competition and Economies of Scale

Economies of scale in the firm level:

In general, the property

y(λL, λC ) > λy(L,C )

defines increasing returns to scale.

We confine our analysis to homogenous functions. A functiony = y(L,C ) is homogenous of degree r if

λry = y(λL, λC )

It is linear homogenous for r = 1, implying constant returns toscale. If r > 1 then we have increasing returns to scale.

S.138


I Increasing returns imply that marginal input demand for oneadditional unit of output is decreasing.

I Therefore also the marginal cost (with given factor prices) aredecreasing.

I Therefore the marginal cost will always be below thedecreasing average cost function.

This implys that the rule price =marginal cost cannot hold truesince the firm would facenegative profits. This makesperfect competition impossible.

output

ACMC

price

losses

S.139


I We have the case of a natural monopoly where it istechnical efficient that one firm produces the complete output(demand) in order to utilize the economies of scale.

I The natural monopolist will maximize profits, thus setting theprice according to Cournot rule: marginal return equalsmarginal cost. The difference between the monopoly price andmarginal cost is determined by the price elasticity of demand.

I Regulatory problem: monopoly is technical efficient butreduces static welfare due to rent extracting.

S.140


output

demand

MR MC

yM

pM

AC

profits

S.141


I Remark: If we have one fixed input factor which induces fixedcost and the variable factor with constant returns, then theaverage cost function is decreasing, and the marginal costfunction is always below average cost. Thus, the result issimilar to increasing returns. E.g.: c(y) = cfix + ay .

I General problem: Assuming increasing returns to scale on thefirm level contradicts the assumption of perfect competition inthe trade models (see above).

I Taken seriously, the Ricardo-Viner model suffers from thisinconsistency that we have fixed costs plus perfectcompetition (= negative profits).

I We have to relax the market structure assumption:monopolistic competition, monopoly.

I But there is another possibility : increasing returns on thersector level.

S.142


Increasing returns on the sector level:

I The single firm i faces constant returns to scale. The ACfunction is non-decreasing in the firm-controlled output (e.g.constant in case of Cobb-Douglas).

I But the production of other firms j in the sector has apositive externality to the productivity of firm i (e.g.agglomeration benefits). This can be formalized by making thetotal factor productivity dependent on the aggregated output:

yi = T (y)yi (L,C ), y =∑j 6=i

yj , dT/dy > 0

λyi = T (y)yi (λL, λC )

I The aggregated production function has increasing returns toscale while the single firm operates with constant returns. Theprofit maximization behavior is then compatible with theperfect competition assumption.

S.143


I In case of increasing returns to scale, the transformation curve isconvex to the origin.

I In order to determine the production point in autarky it depends onthe curvature of the welfare indifference curve. Since we usuallyassume impefect substitutability, an indifference curve will neverintersect the axis. Therefore we can can conclude that there will bean interior solution in autarky.

y2

y1

W

y2

y1

W

y2

y1

W

S.144


(Imperfect) market structures:

Heterogeneous oligopoly as a “generic” market structure.

I Degree of heterogeneity is positive, limited number of firms:heterogeneous oligopoly.

I Degree of heterogeneity is zero, non-increasing returns toscale, limited number of firms: homogenous oligopoly.

I Degree of heterogeneity is zero, non-increasing returns toscale, many firms: perfect competition.

I Degree of heterogeneity is zero, increasing returns to scale:(natural) monopoly.

I Arbitrary degree of heterogeneitiy, arbitrary scale effects, onefirm: monopoly.

I Degree of heterogeneity is positive, increasing returns to scale,many firms: monopolistic competition (Chamberlin).

S.145


Monopolistic competition:(a technical description is provided in the context of th Krugman model)

I Heterogeneous = differentiated products, each produced byone firm. Each firm is a monopolist in the market for hisspecific product variant (because of scale effects).

I The monopolist faces a residual inverse demand functionpi (xi , x−i ) with x−i as the vector of output of all othercompeting firms. It is assumed that the product variants areseen as substitutional goods in the household’s optimizationproblem.

I In contrast to an oligopoly we assume that the market powerof the firm is low, i.e. there is no strategic impact of xi onx−i . The firm has to take x−i as given. It behaves accordingto Cournot’s monopoly solution.

S.146


I In the short run the firm earns profits. This attracts morefirms to enter the market with new product variants. Thetotal demand is then allocated to more firms, implying thatthe residual demand curve of a single firm shifts to the left.

I This process stops when the demand curve is tangential to theaverage cost function. Then we have zero profits and marketentry stops (Chamberlin solution).

I For simplicity we assume symmetry of all firms. Then it ispossible to determine algebraically the number of firms whichearn (nearly) zero profit.

I Following graphic: We have assumed linear costs with fixedcosts (decreasing AC), and a linear demand function.

S.147


output

MC

AC

DMR

DMR

D′MR’

pM

yM

Chamberlin solution

S.148


Oligopoly:

I Strategic interdependency of firms: firms set their strategicvariable (price or quantity) as a best response to the expectedbehavior of the other firms. In a Nash equilibrium expectedand realized variables are identical (no systematic error) andthe variables are determined as a mutual best response.

I This is relevant for large firms or in case of trade where eachcountry has a monopoly incumbant in the market.

I Ricarian framework with oligopolistic competition:

Ruffin, R.J. (2003), Oligopoly and trade: what, how much, and for

whom? Journal of International Economics 60, 315-335.

I We will discuss the oligopoly case in the context of strategictrade policy.

S.149

5. Economies of Scale and Imperfect Competition5.2 Trade with Increasing Returns

Implications of increasing returns for trade(case of homogenous goods):

I In case of free trade, we have usually complete specializationin order to utilize the scale effects.

I The country size plays a role since large countries can utilizethe scale effects more than a small country. Even in case ofidentical production technology, there is asymmetry due todifferent country size.

I The Heckscher-Ohlin theorem as well as Samuelson-Stolpertheorem do not hold true anymore! The price relation is not atangent to the transformation curve. There is not necessarilyan unambigous relation between goods and factor prices.Depending on the type of the economies of scale the mostcompetitive case implies price = average costs which are notconstant.

S.150


We have an incentive to specialize and to trade even if bothcountries are completely identical!

y2

y1

t.o.t.

y2

y1

t.o.t.

S.151


We may have positive welfare effects from trade even if eachcountry specializes to the good where it has a comparativedisadvantage! (scale effects overcompensates this)

y2

y1

W

y2

y1

W

S.152


It is possible that a country specializes to the good where it hascost disadvantages, but due to the economies of scale (here:monopolistic competition) there is no incentive to reverse thespecialization pattern, since the country would start producing thegood with larger average costs.

y1

demand

AC 1

(monopolistic competition)

AC 2

S.153


Market-power and trade:

How trade affects the imperfect competitive sector:

I If there is imperfect competition in sector 1 but constantreturns to scale (concave transformation curve), thenp1 > MC1 while p2 = MC2.

I The autarky price relation is no longer a tangent at thetransformation curve (see point A in the following figure)!The tangential property was derived by using the perfectcompetition assumption.

S.154


y2

y1

p in autarky

A

W0

B

W1

t.o.t.

C

D

W2

S.155


I If there is free trade, the world market for good 1 will be morecompetitive than in autarky case. The market power of thefirms in sector 1 vanishes due to international competition.Domestic firms in sector 1 are forced to produce with a givencompetitive price. This leads to a specialization in point C.

I Welfare increases from W0 to W2. This can be decomposedinto twi effects:

I From point A to B (and hence W0 to W1) we have the effct ofincreased competitiveness. The market power has vanished dueto trade. Thus, if we would have had competition in theautarky case, then B would have been realized.

I From B to C (and hence W1 to W2) is the effect fromspecialization and trade.

S.156


I Result: Monopolist will lose its market power when we havefree trade.

I Does the result hold true in case of increasing returns to scaleon the firm level? Note, that in this case perfect competitionis impossible.

Approach of Markusen-Melvin:

I Increasing returns to scale in sector 1. In autarky we have amonopolist in sector 1 which charges a monopoly price.

I Assume that the country has a comparative price disadvantagefor good 1. The monopolist sees the power eroding effect offree trade and will adapt his prices to a competitive level.

I Assume that with this competitive behavior the country willspecialize to good 1 (see graphic).

S.157


y2

y1

Country 1 (dominant)

A

W 1A

W 1B

B

C

W 1C

y2

y1

Country 2

B

W 2B

C

W 2C

S.158


I Now both countries are specialized. If the monopolist incountry 1 would decide now to charge the monopoly priceagain, the welfare effect would be even higher for this country,but it may harm the other country which is specialized togood 2.

I If the other country would try to escape from this situation byproducing good 1 again, it would start producing withextremly high average costs! Country 1 is able to exploit thescale effects and to extract all the rents in a way that preventscountry 2 from engaging in the production of good 1 again.

I Typically, country 1 is a large country which has more powerto determine the pattern of specialization.

S.159

5. Economies of Scale and Imperfect Competition5.3 Trade with Differentiated Products

I Inter-industrial trade has been explained by comparativeprice advantages. A comparative price advantage can onlyoccur when comparing identical (homogenous) goods.

I The incentive for intra-industrial trade can not becomparative price advantages since there are different productvariants in both countries!

I Why is there only one firm which produces a product variant?We assume increasing returns to scale, or, alternatively, a costfunction with fixed costs and constant marginal costs(decreasing AC):

C (y) = F + ay

AC (y) = F/y + a

I We have n firms with the same cost function, each producinga differentiated good which are close substitutes. We have amarket with monopolistic competition.

S.160


The Krugman model:

1. variant: heuristic linear demand function

I Assume that the residual demand is identical for each firmand given by

yd = S [1/n − b(p − p)]

where S are the total purchases, p the price of the firm and pthe average price in the market. b is a parameter, indicatinghow sensitive the demand responds to price deviations fromthe average. If all firms charge the same price, the totaldemand is devided into equal shares S/n.

S.161


I Maximizing the profits

maxpπ = pyd−C (yd ) = pS [1/n−b(p−p)]−F−aS [1/n−b(p−p)]

and considering that in equilibrium all firms will charge thesame price so that p = p for all firms, gives the monopolyprice

p =1

bn+ a

and the monopoly quantity

y =S

n

S.162


I The average costs are therefore

AC =Fn

S+ a

The more firms are in the market the less could the single firmexploit the decreasing cost effect, i.e. the average costincrease with n (and vice versa for market size S).

I In the long run the Chamberlin solution implies that priceequals average cost (zero profit):

1

bn+ a =

Fn

S+ a (23)

⇒ n =√

S/bF

which determines the equilibrium number of firms andhenceforth the number of product variants.

S.163


Graphical representation of the solution:

p,AC

n

p

AC

a +√

F/Sb

√S/bF

S.164


I With this equilibrium value of n the price and output of eachfirm is given by

p = a +√F/Sb

y =√SbF

I This implies that with the market size S the number of firmsas well as the firm size will increase while the price willdecrease.

S.165


Effects of trade:

Firms are producing for the world market demand. This means thatthe total market size S will increase to S ′:

p,AC

n

p

AC

a +√

F/Sb

√S/bF

AC ′

a +√

F/S ′b

√S ′/bF

S.166


The diagram shows the sitauion of the integrated world market.Compared to autarky, we have

I more product variants,

I produced to lower aveage costs.

I and sold to a lower price.

Why is there an incentive to trade?

I Consumer’s preferences for more product varieties!

S.167


Numerical example (taken from Krugman/Obstfeld):

I Let b = 1/30, 000,F = 750, 000, 000, a = 5000 in bothcountries.

I In the domestic market, the market size is S1 = 900, 000, andin the foreign market S2 = 1, 600, 000.

domestic market foreign market integratedin autarky in autarky world market

market size 900,000 1,600,000 2,500 ,000firms/variants 6 8 10sales per firm 150,000 200,000 250,000AC 10,000 8,750 8,000price 10,000 8,750 8,000

S.168


2. variant: preferences for variety

I Utility function depends not only on the consumed quantitybut also on the number of variants (Lancaster), e.g.:

U =n∑

i=1

y(σ−1)/σi , σ > 1

and in case of symmetry

= ny(σ−1)/σi

I It is possible to maximize utility with a given budgetconstraint BC =

∑ni=1 piyi = npyi (in case of symmetry).

I Deriving the demand function, determining the profitmaximizing price, determining the number of firms inequilibrium ⇒ see below.

S.169


The additional information is that in an integrated world marketwe obtain a higher welfare indifference curve:

n

yi

BC = npyi

Uaut = ny(σ−1)/σi

n

yi

Utrade = n′(y ′i )(σ−1)/σ

n′

y ′i

S.170


I From monopoly pricing we know that the price is a markupover marginal cost, where the markup is determined by theprice elasticity of demand:

R = p(yd ) · yd

MR =dp

dyd

yd

pp + p

= p(1− 1/η)

with η as the price elasticity of demandη = −(dyd/dp)(p/yd ) = σ (in case of the utility functiongiven above).

I Assume, that labor is the unique production factor, and thefirm’s labor demand is given by

Li = α + βyi

where MC is given by βw and AC is given by wα/yi + wβ.S.171


I From profit maximizitaion condition MR = MC and zeroprofit condition p = AC (Chamberlin solution) we have thelinear system

pi

w= β

(η

η − 1

)pi

w=

(α

yi

)+ β

This gives a unique solution for

yi = α(η − 1)

β

pi =wβη

(η − 1)

S.172


I To obtain the number of firms in equilibrium we have to usethe full employment condition:

L =n∑

i=1

(α + βyi ) = n(α + βyi )

⇒ n =L

α + βyi=

L

αη

where yi = α(η − 1)/β is determined by the linear systemabove.

I In this case, the number of variants and the country size (L)are proportional, implying that doubling the market (twoequally sized countries) means doubling the number ofvariants, in contrast to the case of linear demand functionsabove.

S.173


Summary:

I The argument for intra-industrial trade cannot be acomparative price advantage because different types of goodsare traded which can produced only by one firm (scale effects).

I Opening the market: Domestic customers will also purchaseforeign product variants. Domestic firms will also produce forforeign demand.

I By enlarging the market each firm can produce to loweraverage cost which induces profits which induces marketentry. Finally, lower AC lead to lower prices.

I The sum of product variants of country 1 and 2 will decreasebut the number of product variants available to the consumerswill increase in each country.

I There are also welfare gains from trade.

S.174

5. Economies of Scale and Imperfect Competition5.4 Empirical Relevance

How could intra-industrial trade be measured?

I We have to disaggregate into i = 1, ..., n industries (sectors).I Inter-industrial trade means exporting e.g. y1 and importing

y2, that is Ex1 > 0, Im1 = 0,Ex2 = 0, Im2 > 0I Intra-industrial trade means exporting and importing e.g. y1,

that is Ex1 > 0, Im1 > 0.I The Grubel-Lloyd index measures the “degree” of

intra-industrial trade in sector i :

Ii = 1− |Exi − Imi |Exi + Imi

Grubel, H.G., Lloyd, P.J. (1975), Intra-industry trade: the theory and

measurement of international trade in differentiated products. The

Economic Journal 85 (1975).

I In case of pure inter-industrial trade we would have eitherExi = 0 or Imi = 0 so that the index is Ii = 0. The maximumof intra-industrial trade is defined as Exi = Imi . In this casethe numerator is zero and we have Ii = 1. S.175


US 1993, selected sectors

sector GL index

anorganic chemical products 0.99energy producing equipment 0.97electric equipment 0.96organic chemical products 0.91medical products and pharma 0.86technical office equipment 0.81telecommunication products 0.69automotive 0.65iron and steel 0.43textiles, clothes 0.27shoes 0

(Source: Krugman/Obstfeld)

S.176


I SITC scheme (standard international trade classification).See http://unstats.un.org

I Classification by digits – the more digits the more is a sectorsubdivided.

I Example:I Section: 7 - Machinery and transport equipmentI Division: 73 - Metalworking machineryI Group: 733 - Machine tools for working metal, sintered metal

carbides or cermets, without removing materialI Subgroup: 733.1 - Machine tools (including presses) for

working metal by forging, hammering or die-stamping; machinetools (including presses) for working metal by bending, folding,straightening, flattening, shearing, punching or notching;presses for working metal or metal carbides

I Basic heading: 733.12 - Bending, folding, straightening orflattening machines (including presses), numerically controlled

S.177


I If the index is applied only to the first digit SITC, then theexchange e.g. of leather and wood products appear as“intra-industrial” (both belong to SITC 6), it overestimatesthe extent of intra-industrial trade. Therefore, only moredetailed classifications are reasonable for applying theGrubel-Lloyd index.

I The index can be aggregated over all sectors where thesector-specific indices are weigthed with the trade volume:

I =n∑

i=1

gi Ii

with the weights

gi =(Exi + Imi )∑nj=1(Exj + Imj )

S.178


The case of China

Trade-weighted average Grubel-Lloyd index, different levels ofaggregation

3-digit 2-digit 1-digityear 237 sectors 67 sectors 10 sectors

1980 0.20 0.30 0.631985 0.20 0.29 0.441990 0.36 0.45 0.601995 0.38 0.48 0.672000 0.39 0.48 0.572005 0.42 0.49 0.58

Source: van Marrewijk, C. (2008), Intra-industry trade, in: K. Reinert and R.

Rajan (eds.): Princeton encyclopedia of the world economy, Princeton

University Press, Princeton, N.J.

S.179


Bayoumi, T. (2011), Changing Patterns of Global Trade. IMF Paper.

S.180


I There are more indices with different properties, e.g. adynamic variant of the GL index:

DIi = 1−|(Exit − Exi ,t−1)− (Imi ,t − Imi ,t−1)||Exit − Exi ,t−1|+ |Imi ,t − Imi ,t−1|

= 1− |∆Ex −∆Im||∆Ex |+ |∆Im|

I This allows for an analysis of changes in intra-industrial tradeflows (e.g. as a response to exogenous effects) which mighthave been unrevealed when using the static GL concept.

Brulhart, M. (1994), Marginal Intra-Industry Trade: Measurement and

Relevance for the Pattern of Industrial Adjustment. Weltwirtschaftliches

Archiv 130, 600-613.

S.181


Summary:

I Horizontal intra-industrial trade (differentiated productvariants) can be explained by preferences for variety.

I Intra-industrial trade could be measured e.g. with theGrubel-Lloyd index.

I Intra-industrial trade has gained more and more importance,and now dominates the trade.

I For vertical trade of intermediate goods, standard models canbe used for an explanation.

S.182

6. International Factor Mobility and Multinational Firms6.1 Factor Movements in a Ricardo-Heckscher-Ohlin framework

Source: Markusen, J.R. (1983), Factor Movements and Commodity Trade as

Complements. Journal of International Economics 14, 341-356.

I Ricardo: Trade due to productivity differences; no FPE, factor pricedifferences reflect productivity differences.

I HOV: Trade due to different factor endowments; differentendowments create factor price differences; this creates goods pricedifferences and an incentive to trade; with (partial) specializationand free trade we have FPE.

Basic idea of the Markusen model:

I Starting with identical relative factor endowment.I Different productivities create factor price differentials. These

differentials create factor movements.I The different relative factor endowments are then a result of

productivity differences.I Result: FPE (trivial), factor movements enlarge specialization

and trade!S.183


Model setup:

I The standard 2-2-2 scenario.

I Identical initial factor endowments.

I Good 2 is the labor-intensive good (no reversals).

I Identical homothetic preferences.

I Production functions with constant returns to scale forcountry j ∈ {h, f }:

Y j1 = G (Lj

1,Cj1)

Y j2 = αjF (Lj

2,Cj2)

with αh > αf , i.e. the home country has a higher productivityin sector 2. (Since Y2 is labor-intensive, this has the same effect as

labor augmentation in country 2.)

S.184


y1

y2foreign home

F

H

homotheticpreferences

sector 2

cap

ita

l

foreign=home contract curve

sector 1

labor

F

H

S.185


1. Even if country h as a higher productivity in sector 2, therelation of marginal returns to labor and capital are identicalto those of country f ⇒ identical contract curve in theEdgeworth box.

2. If both countries would choose the same point F on thecontract curve (implying identical relative factor prices), thencountry h would produce more of Y2 but the same amount ofY1 compared to country f , implying a different marginal rateof transformation.

If we assume trade and therefore an identical price relation,then the production point must be“downhill” in point H.

3. For homothetic preferences (same price relation = samestructure of the consumed goods bundle), country h exportsgood 2, while country f exports good 1.

S.186


⇒ But then the factor prices can not be equal(see points F and H on the contract curve):

I Let p = p2/p1 be the world market price of good 2(expressed in units of good 1).

I Competitive equilibrium requires that factor price equals thevalue of its marginal product:

w j = pαj ∂F

∂L=∂G

∂L, r j = pαj ∂F

∂C=∂G

∂C

I Note that C1/L1 in point H is larger than C1/L1 in point F .

⇒ Thus, we concludewh

rh>

w f

r f

and according to the Ricardo result, real wages in country hare higher than in f (and vice versa for capital).

S.187


This creates an incentive of the factors to move (if one assumethat differentials in factor prices are the only reason to move):

I Labor moves from country f to h, capital moves from h to funtil factor prices are equalized.

I Each country receives the factor which is intensively utilizedin the export sector.

I This adds an HOV base for trade and reinforces the directionof specialization and trade.

S.188


Home country h:(Rybczynski effects)

Labor inflow (c.p.):y1

y2

Capital outflow (c.p.):y1

y2

S.189


Recall the Rybczynski theorem:

I An increase of the endowment of a factor induces an increaseof production of the good which intensively utilizes this factorwhile the other good’s production is reduced (with constantterms of trade).

I Both, the inflow labor, and the ourflow of capital work in thesame direction: Country h enforces the production of good 2,country f enforces the production of good 1.

I Since the goods prices and therefore consumption points areunchanged, this implies an enlargement of trade.

S.190

6. International Factor Mobility and Multinational Firms6.2 Mobility of Capital

Capital flows as an intertemporal decision problem:

I Capital is accumulated savings. Since in all models the goodsY1,Y2, ... have been interpreted as consumption goods, thecomplete world production is consumed. For capitalaccumulation we need savings (waive for consumption) andproduction of investment goods.

I Requires an intertemporal decision: waive for consumptiontoday enables larger output and consumption tomorrow.

I In open economies with trade, the domestic savings need notto be equal the domestic investments into the capital stock:

GDP = C + I + (Ex − Im) ⇐⇒ S = I + (Ex − Im)

S − I = (Ex − Im)

I Savings and henceforth capital is transferred to the foreign,this is accounted in the capital balance.

S.191


What does this practically mean?

(Case of net capital exports, savings flow from h to f )

I Domestic agents have claims against foreign agents:I Domestic agents simply accumulate foreign currency = claim

against foreign central bankI Alternatively, they invest it in foreign bond = interest bearing

claimI Alternatively, they hold foreign equity capital = claims on

returns / profitsI Alternatively, they use revenues from trade to pay back debt or

to reduce foreign claims against domestic agents.I Special form: Foreign Direct Investment

I Cross-border ownership of the (foreign) capital stock =extension of domestic capital stock on foreign territory

I Ranges from building new production capacities with positiveknowledge spillover effects to the host country to simpleresource and land grabbing.

I It should be acknowledged that surplus from FDI is typicallyrepatriated to the domestic country. S.192


What drives capital flows between countries?

I real return r , r∗

I capital flow controls

I taxes, differences in taxation systems

I uncertainty (e.g. exchange rates ⇒ “MonetaryMacroeconomics”)

I stability of the (political) environment

Different types of capital flows:

I Long run: foreign direct investment – claims on the foreign’scapital stock and its management

I Short run: portfolio investments – adjusts immediately ifprices, interest rates, exchnge rates etc. change

S.193


Microeconomic foundation of intertemporal consumption andcapital flow:

I 2 periods, only one good that can either be consumed orinvested

I perfect capital market with one interest rate for lending andborrowing

I Intertemporal budget constraint with lending/borrowing:

Ct+1 = Yt+1 + (1 + r)(Yt − Ct)

⇒ Ct +1

1 + rCt+1 = Yt +

1

1 + rYt+1

S.194


I Consumers maximize the net present utility

U = U(Ct) +1

1 + ρU(Ct+1)

with ρ as the intertemporal discount factor (reflectingpreference for the presence or “impatience”).

They have to maximize utility conditional to the intertemporalbudget constraint (DD in the following figure).

S.195


Yt+1,Ct+1

Yt ,Ct

45◦

PYt+1

Yt

D

D

determined by r

CCt+1

Ct

A

(a)

(a) Borrowing in t (Import)

(b)

(b) Repaying in t + 1 (Export)

S.196


I In t the households consume more than it is produced =positive net imports = negative trade balance AC . Theconsumption is partially financed by foreign savings(increased debt = positive capital balance)

I In t + 1 the debt is repaid including interest rate: AP, implyinga positive trade balance (and negative capital balance).

I Both balances in t and t + 1 reflect optimal behavior, notdisequilibrium.

S.197


I Now consider that Y can not only be consumed but also beinvested.

I Investment leads to capital accumulation. Thus theproduction frontier shifts and we have

Yt+1 = Yt + F (It)

with F (·) as the production effect of investments andF ′ > 0,F ′′ < 0

I Note that we have here only one good Y , so that capitalaccumulation can not have effects according to theRybczynski theorem.

I Saving and investment therefore lead to an intertemporalproduction frontier (RP in the following graphic).

S.198


Yt+1,Ct+1

Yt ,Ct

45◦

P

Yt

D

D

R

D’

D’

QYt+1

It

shift of budget constraint

S.199


I The calculus of an investor is whether to invest into thephysical project It or to invest the funds at the world capitalmarket at the rate r . The returns from investment F (It) aretherefore discounted with 1/(1 + r).

I Profit maximization gives

maxItπ =

1

1 + rF (It)− It

FOC ⇒ dF

dIt= 1 + r

which is the tangential point Q at the intertemporalproduction frontier RB.

I Due to the increased production in t + 1 the intertemporalbudget constraint shifts to D ′D ′.

I The same calculus for intertemporal consumption is applied tothe new constraint. The achievable utility increases.

S.200


Yt+1,Ct+1

Yt ,Ct

45◦

Yt

D

D

R

D’

D’

QYt+1

It

CCt+1

Ct

Imt

Ext+1

A B

S.201


The complete result of intertemporal consumtpionand investment decisions:

I Recall that S − I = (Ex − Im). In t we have a negative tradebalance (positive capital balance) AC , containing excessconsumption BC and investment AB.

I In t + 1 the foreign debt is repaid inclduing interest paymentsAQ which implies a positive trade balance.

⇒ Borrowing and repaying capital is indirectly borrowing andregiving goods.

S.202


Consider two countries with different investment opportunities anddifferent time preferences:

I An optimal positive trade balance in t in country 1 implies anoptimal negative trade balance in t in country 2, and viceversa in t + 1.

I Given the time preferences and different returns oninvestments (different intertemporal production frontiers), theoptimal plans for consumption and investment (andhenceforth exports and imports) must be consistent.

I The world market interest rate r adjusts so that intertemporalplans of both countries are coordinated.

S.203


Yt+1, Ct+1

Yt , Ct

45◦

Yt

Country 1

−(1 + r)

Yt+1

ItCtExt

Imt+1

Yt+1, Ct+1

Yt , Ct

45◦

Yt

Country 2

−(1 + r)

Yt+1

It

Ct

Ext+1

Imt

Left side: low marginal productivity of It , patient consumersRight side: large marginal productivity of It , less patient consumers

S.204


Im1t+1

Ex2t+1

Im2t , Ex1

t

offer curve

capital exporting country 1

offer curve

capital importing country 2

(1+r)

S.205


Possible effects of capital flows from the developed to lessdeveloped countries:

I A country with a higher return on investments has acomparative advantage for attracting investments.

I The capital receiving countries build up a capital stock muchfaster.

I This has also a positive effect on the labor productivity: wageswill rise.

I In the country with capital outflows there might be reverseeffects: unskilled labor will (relativley) lose.

I Not explained in the model: inflowing financial sources can bealso used for consumption.

S.206

6. International Factor Mobility and Multinational Firms6.3 Mobility of Labor

I Labor can be more or less skilled, depending on the level ofeducation. Many models distinguish raw labor force (unskilledlabor) L and human capital HC .

I Wages for high-skilled labor will exceed wages for unskilledlabor, the wage difference depends on the differential of(expected) marginal return to labor and can be interpreted asthe rent for former investment in human capital.

I Wage differences across countries depend on technologicaldifferences, the level of education, and the relativeendowments of factors.

I The decision to migrate and the decision to accumulateHC (invest into education) become interdependent.

S.207


Debate about migration:

I 1990ies/2000: Germany faced an outflow of high-skilled labor(engineers, physicians, top managers, IT specialists etc.)which exceeded the inflow (loss of human capital). Discussingmeasures to reverse this trend.

I In the 1950ies and 1960ies also unskilled or less skilled laborwas scarce! Labor force from Italy and Spain, later fromTurkey have been encouraged to migrate to Germany.

I In both cases the lack of labor force or human capital inducesan underutilization of production capacities and may lead to aloss of growth rate.

I 2015: Strong inflow of refugees – integration in labor marketmight become problematic.

S.208


Motives of migration (e.g.):

I Higher wages / better working conditions in foreign country

I Better supply with goods and access to infrastructure

I Better social security

I More economic and political freedom

Barriers of mobility (e.g.):

I Different language and culture

I Losing existing social contacts (family, friends etc.)

I Transaction and transportation costs

I Uncertainty

S.209


Assume that wage differentials ≈ differences in labor productivityare the driving force:

I Real wages will be equalized. Migrants as well as the worker inthe sending country receive higher wages, while the worker inthe receiving country face a wage loss.

I Global factor allocation is improved in terms of efficiency,therefore the output will increase (see triangle ABC in thegraphic).

S.210


w w∂Y1

∂L1· p1 ∂Y2

∂L2· p2

L1 L2

w2

w1

B

C

w∗ w∗

migration

A

S.211


I In times of small and moderate migration, these effects mightbe small and statistically insignificant.

I In time of mass migration (e.g. late 19th century) theseeffects have been significant (source: cf. Krugman/Obstfeld)

real wage 1870 increase or real wage(US = 100) in % 1870-1913

receiving countriesArgentinia 53 51Australia 110 1Canada 86 121USA 100 47sending countriesIreland 43 84Italy 23 112Norway 24 193Sweden 24 250

S.212

6. International Factor Mobility and Multinational Firms6.4 Mobility of Human Capital: Brain Drain, Brain Gain

I “Brain Drain” = Migration of HC from the perspective of thesending country

I “Brain Gain” = Migration of HC from the receiving country,but also HC losses might have stimulating effects

I Typical situation: Return on HC is higher in the developedand industrialized countries compared to developing countries.Thus, there is an outflow of HC from developing countries.The share of skilled workers and hence average wagesdecrease.

⇒ Thus rich countries were seen to become richer, the poorcountries become poorer at the first sight. Should there be ataxation-compensation scheme? (Bhagwati)

S.213


I It is more likely that high-skilled workers migrate rather thanless-skilled (e.g. larger wage differentials, lower barriers tomigrate).

I Slight increase in migration.

Iranzo, S., Peri, G. (2009), Migration and trade: Theory with an application to the Eastern-Western European

integration. Journal of International Economics 79, 1-19

S.214


Problems:

I The sending country has invested into HC but it utilizes neither itsproductivity effect nor its positive externalities. Typically, thegoverment spend money for education, financed by taxes.

I According to the Rybczynski theorem, the HC-intensive sector inthe sending country shrinks while the unskilled-labor intensive sectorgrows. This reverses the developing process.

I Loss of income and therefore tax base.

I Since skilled workers are more productive than unskilled workers, theoutput per capita will also decrease in the sending country (c.p.).

I In a dynamic perspective, growth models like Uzawa/Lucas showthat loss of HC is more harmful than a loss of physical capital: theconvergence process slows down.

S.215


But there are also positive effects:

I Migration of HC leads also to FPE (in tendency). In the sendingcountry, the investment into new HC is stimulated since its marginalproductivity increases (vice versa for the receiving country). It ispossible that the encouraging effect on skill-creation outweights theloss effect of emigration.

I Furthermore, the decision to become more educated now not onlydepends on the domestic marginal product but also on the foreignreturn on HC, i.e. workers invest into HC with the perspective toeventually leave the country. Thus, international wage differentialsnot only induce migration but also alter the incentive scheme forHC accumulation.

I High-skilled workers in the foreign country will partially transferincome to the sending country.

I Return to home country after being additionally skilled.

I HC flows as a vehicle of knowledge diffusion.

S.216


I The empirical literature is inconclusive in most points.

I Correlation between emigration rate and the effect on theGDP growth rate: ambigous effect

Source: Rapoport (2004)

S.217


Some introductory literature:

I Rapoport, E. (2004), Who is Afraid of Brain Drain? HumanCapital Flight and Growth in Developing Countries. BrusselsEconomic Review 47(1), 89-101.

I Commander, S., Kangasniemi, M., Winters, L.A. (2004), TheBrain Drain: a Review of Theory and Facts. BrusselsEconomic Review 47(1), 29-44.

I Schaeffer, P. (2005), Human Capital, Migration Strategy, andBrain Drain. Journal of International Trade and EconomicDevelopment 14(3), 319-335.

S.218

6. International Factor Mobility and Multinational Firms6.5 Foreign Direct Investments and Multinational Firms

(a) Foreign Direct Investment

I Foreign Direct Investement: Investing capital or buying assetsin a foreign country in order to exercise some managerialcontrol in the long run. Commonly used (heuristic) crietrion:aquiring at least 10% of the capital of a firm.

I FDI are done by multinational firms (production firms as wellas financial intermediates). In most models it is assumed thata domestic producing firm invests into foreign productioncapacities.

I Decision whether to produce in the domestic country andexporting the goods, or to invest in the foreign country, andproducing and selling there (third possibility: Joint Ventureand contractual agreement with foreign firm). This combinesan ownership decision and a localization decision.

S.219



I FDI can be a substitute for exports (final goods), but it canalso stimulate trade e.g. in case of FDI for outsourcing theproduction of intermediate goods, or producing abroad (withcost advantages) and importing the final goods.

I If a firm outsources parts of the production chain, thenintra-firm trade occurs.

I FDI are seen as the “primary motor of globalization” (RenatoRuggiero, former WTO director)

S.220



I Firms have intangible assets like knowledge (e.g. patents),technological skills. Producing abroad is then technology orknowledge transfer, affecting the production function in theother country.

I The investing firm has specific knowledge and transfers it tothe country with cost advantages in production by FDI. Thereceiving country faces a positive spillover effect (example:Siemens in China)

I Or the domestic firm invests into foreign firms in order toshare their superior knowledge and utilizing it in the homecountry.

S.221



UNCTAD World Investment Report 2014:

I Total FDI in 2013: 1.452 trillion USDI FDI outflow 2013:

I Developed countries: 899 billion USDI Developing and emerging countries: 553 billion USD

I FDI inflow 2013:I Developed countries: 566 billion USDI Developing and emerging countries: 886 billion USD

S.222



(Source: UNCTAD World Investment Report 2014)

S.223


(b) Multinational Firms

I Most trade theories assume that a single firm, residing in onecountry, produces a certain good which can then be exported.

I Empirically seen, many firms produce a good in differentcountries, or they segment the production chain and producedifferent parts of the good in different countries. Theheadquarter of the firm resides in one country.

I Firms which are producing in more than one country and/orwhere production and headquarter services operate indifferent countries, are called Multinational Firm (MNF).

S.224



I Horizontal MNF: The complete production of good y isdone simultanously in different countries. Note that thiserodes the incentive to export.

I Vertical MNF: Different parts of the production process ofthe good y are done in different countries

I trade with intermediate productsI intra-firm trade with intermediate products

MNF activities could stimulate trade (dominating effect) bzt couldalso dampen trade in case of horizontal MNF.

S.225



What are motives to become a MNF?

I Firms can have a competitive advantage in different locations⇒ approach by M. Porter

I Different dimensions which are important for the locationdecision (according to J.H. Dunning 1980):

I Ownership advantagesI Location related advantagesI Internalisation advantages

S.226



Markusen, J.R. (2002), Multinational Firms and the Theory ofInternational Trade. Colorado.

I Partial equilibrium approach: The consequences of firm’sdecision on total economy are neglected.

I Options of a firm:I Produce only national (don’t become MNF), and export into

the foreign country.I Produce in foreign country while headquarter services stay in

home country, and export into home country (vertical MNF)I Produce in both countries, no exports/imports (horizontal

MNF). In this case, the firm is able to charge different prices.

I The firm is involved into Cournot competition (notexplicitely modelled here) where in both countries we have thesame linear demand scheme.

S.227



profit domestic country profit foreign country

national MCd , Ld ,G MCd ,T , Lf

vertical MCf ,T , Ld MCf , Lf ,Ghorizontal MCd , Ld ,G MCf , Lf ,G

MCi = marginal cost in i ∈ {d , f },Li = market size/demand,G = fixed costs for setting up a production plant,T = transportation cost.

The firm calculates the possible total profits (foreign + domestic)and decides about the internationalization policy (national,vertical, horizontal).

S.228



A baseline parametrization of the model: Let MCd = MCf andLd + Lf = L so that we can calculate the domestic market shareLd/L. In case of a horizontal MNF, there are the same profits perunit in both countries so that different domestic market shareshave no impact on total profits.

0

profits

1Ld

L0.5

horizontal MNF

national firm

vertical MNF

S.229



Increasing world demand L: All curves shifts upwards but profitsfor horizontal MNF more than the others: In case of trade betweentwo large countries of similar size it is more likely to havehorizontal MNF.

0

profits

1Ld

L0.5

horizontal MNF

national firm

vertical MNF

horizontal MNF

national firm

vertical MNF

horizontal MNF

national firm

vertical MNF

S.230



Reducing transportation costs: Does not affect horizontal MNF,but the other curves shift upwards. It becomes more likely toobserve national production or vertical MNF, depending on thecountry size differences.

0

profits

1Ld

L0.5

horizontal MNF

national firm

vertical MNF

national firm

vertical MNF

national firm

vertical MNF

S.231



Different marginal costs in foreign country (MCd > MCf ):Vertical MNF become more likely.

0

profits

1Ld

L0.5

horizontal MNF

national firm

vertical MNF

horizontal MNF

national firm

vertical MNF

horizontal MNF

national firm

vertical MNF

S.232



I The aim of the model is to explain which determinants willlead to which internationalization strategy (national,horizontal, vertical).

I Since it does not relate the output to different productionfactors via a production function, it cannot explain whichparts of the production process (value creation chain) islocated in which country – see chapter 4.

I It is a partial approach, not a general equilibrium approach.

I Further consequences from internationalization: knowledgetransfer, technology transfer, building up human capital (notdiscussed here).

S.233

7. Tariffs, Subsidies, Quotas, and Trade Policy7.1 Effects of Tariffs and Optimal Tariff Theory

I We have analyzed trade patterns and the effects of trade ondistribution and welfare by assuming that prices and quantitiesare determined by markets (typically: perfect competition).

I We now investigate the effect of distortions by policymeasures like

I (import) tariffs:specific tariff p + t, ad valorem tariff (1 + t)p

I (export) subsidiesI import quotasI “voluntary” export restraintsI other measures, i.e. standards (technological specifications)

The case of standards is highly relevant and is the subject ofall free trade agreements. Problem: standards are often notintroduced as a meausre of trade policy but for other (good)reasons.

S.234


I Partial versus total welfare analysis:I partial: effects on consumer and producer surplus and tarfiff

revenuesI total: all effects on factor allocation and relative prices ⇒

transformation curve analysis

I Small versus large country:I small country: terms of trade are given and unaffected by the

tariffI large country: tariff changes demand, demand changes terms

of trade.

I What happens with tariff revenues?I not considered in welfare analysis (only private welfare)I considering governmental spending of revenues (private plus

gov. welfare)I revenues are given to the consumers

S.235


Tariffs:

I Protect the domestic industry from foreign competitors; thereforeprone to political pressure.

I What might be “good” for some domestic sectors may be harmfulfor domestic consumers and may distort factor allocation. Underspecific conditions domestic welfare may increase – at the expenseof the welfare of other countries (“beggar-by-neighbour-policy”)

I Answer to tariffs (and other trade barriers) may be retorsion tariffs⇒ mutually “optimal” trade barriers lead to a pareto-inferiorequilibrium ⇒ necessity of negotiations (GATT, WTO)

I WTO allows tariffs (i) in cases of “unusual” import competition(“escape clause”), (ii) as antidumping duties.

I Success in negotiations (see graphic), but tendency to substitutetariffs for other forms of trade barriers.

I Detailed discussion of trade policy in MW52.2 “IPE I: Trade andDevelopment” (Prof. Freytag)

S.236


Average tariff rates (1) (MFN = “Most Favored Nation”)

S.237


Average tariff rates (2)

S.238


Average tariff rates (3)

S.239


Welfare analysis:

I We have consumption of a numeraire good c0, and anadditive utility function u = c0 + U(c) with c as the otherconsumption goods. Let p be the price vector and I the givenincome, then the consumer’s program

maxc0,c

u = c0 + U(c) s.t. c0 + p′c = I

leads to the demand function c = d(p) which implies theoptimal consumption of the numeraire good: c0 = I − p′d(p).

I The welfare can then be written as

W (p, I ) = I − p′d(p) + u(d(p))

(see Feenstra, but we neglect the summation over differenthouseholds by assuming one representative household)

I Due to the envelopen theorem we have ∂W /∂p = −d(p).

S.240


I Simplification: We treat c as a single good and p as a scalar.Both goods are traded: While the numeraire good has a fixedworld market price of 1, the other good has a price ofp = p∗ + t with p∗ as the world market price and t as thetariff (later we also consider ad valorem tariffs p = p∗(1 + t)).

I While c is the consumption of the non-numeraire good, it is ythe production and m are the imports, hence c = y + m orm = d(p)− y .

I We assume perfect competition in the numeraire good sector(zero profit condition).

S.241


I What is the income I? Income is generated (i) by (given!)factor income L, (ii) by the tariff revenues t ·m, and (iii) bythe profits of the non-numeraire good sector (if there are anyprofits) π = py − C (y), so that we can write

W (t) = L + mt + [py − C (y)]− pd(p) + u(d(p))

I Tariffs have an impact on the price p = p(t), and hence onthe output y = y(p(t)) of the import-competing industry, andon the imports m = m(p(t)). Recall the envelopen theorem,and m = d(p)− y and p(t) = p∗(t) + t, then we have

dW

dt= −d(p)

dp

dt+ m +

(tdm

dp+ y

)dp

dt+ [p − C ′(y)]

dy

dt

= tdm

dp

dp

dt−m

dp∗

dt+ [p − C ′(y)]

dy

dt(24)

S.242


(a) Perfect competition, small country

I Small country means that tariffs have no impact on the termsof trade p∗. Therefore we have p = p∗ + t with dp∗/dt = 0and dp/dt = 1.

I Perfect competition imply p = C ′(y) (zero profits).

I The effect on the welfare (24) is then

dW

dt= t

dm

dp

I The welfare maximizing tariff is obviously t = 0 implying thatany tariff will harm the domestic welfare!

S.243


Graphical analysis – partial equilibrium:

p

y , c

D

SCase of small country:

p∗

y0 c0

p∗ + t

y1 c1

a

producer surplus: +a

c

tariff revenues: +c

b d

consumer surplus: -(a+b+c+d)

p

m

MX

m0

X + t

m1

c b+d

S.244


I Left side: domestic market, right side: (residual) importdemand, changes in demand do not affect p∗ (totally elasticsupply function).

I The tariffs increases the domestic price. Demand is reduced toc1, supply is increased to y1

I Disentangling welfare effects:I Consumer surplus loss: −(a + b + c + d)I Producer surplus gain: +aI Tariff revenues: +cI Total effect (sum): −(b + d) < 0

I The deadweight loss −(b + d) is the triangle under the importfunction (right side)

S.245


Graphical analysis – total equilibrium:

I The tariff does not only have an impact on the market for theimported good.

I An increase in the output of the import-competing domesticindustry (the primary aim of the policy measure) implies adecrease of hte other sector’s output (in case of fullemployment) and hence a reallocation of resources.

I Welfare effects can be properly analysed in a total equilibriumframework, i.e. by means of transformation and welfareindifference curves.

I Producers and consumers make their decisions according tothe domestic price (relation) p = p∗ + t (steeper price lines inthe graphic). The consumption point, however, is determinedby the fixed terms of trade = opportunities to exchange theproduced bundle of goods on the world market.

S.246


y2

y1

t.o.t.

P0

C0

W0

dom. p

P1

Small country:

tariff changes dom. p

but not the t.o.t.

How to determine the new consumption?

homothetic preferences

C1

W1

S.247


Not discussed here:

I In case of differentiated goods, tariffs may have an impact onthe traded variety of goods! Since variety may enter the utilityfunction, this has also welfare effects.

I Due to the reallocation of resources, tariffs have an impact onincome distribution.

S.248


(b) Perfect competition, large country

I A large country affects the terms of trade, i.e. changes in thedemand for the import good has price effects p∗, hence:p = p∗(t) + t with dp∗/dt < 0.

I With perfect competition (p = C ′(y)), the welfare effect (24) is now

dW

dt= t

dm

dp

dp

dt−m

dp∗

dt

I Starting with free trade, i.e. t = 0 the initial welfare effect is thenpositive:

dW

dt

∣∣∣t=0

= −mdp∗

dt> 0

The decreased world demand dampens the world price p∗ which isan improvement of the terms of trade for the home country.

I From the first order condition dW /dt = 0 we obtain:

t∗

p∗=

(dp∗

dt

m

p∗

)/

(dm

dp

dp

dt

)> 0

S.249


Graphical analysis – partial equilibrium:

p

y , c

D

SCase of large country:

p0

y0 c0

p∗

p∗ + t

y1 c1

tt a

producer surplus: +a

ce

tariff revenues: +c+e

b d

consumer surplus: -(a+b+c+d)

y1 c1

p

m

M

X

m0

X + t

m1

X + t

c

eb+d

X + t

S.250


I The terms of trade in the free trade case is p∗0 = p0. Thedecreased demand leads to a lower world market price p∗, sothe domestic price increases less than t. The world supplycurve is therefore not perfectly elastic.

I Disentangling the welfare effects:I Consumer surplus loss: −(a + b + c + d)I Producer surplus gain: +aI Tariff revenue: +(c + e)I Total effect (sum): e − (b + d)

I The total effect is ambigous and may be positive, dependingon the terms of trade effect e (⇒ elasticity of the worldsupply function).

S.251


Graphical analysis – total equilibrium:

I We need the concept of the offer curve.

I This curve relates the planned exchange of export goods withimport goods for different terms of trade, as given by thedistance of production and consumption points.

I We first start with linear transformation curves where we havefull specialization = given production point.

S.252


y2

y1

t.o.t.

P

C1

C2

C3

offer curve

S.253


y2

y1P1

OC 1

y2

y1

P1

OC 1

y1

y2 P2

OC 2

terms of trade

S.254


We now turn to concave (“neoclassical”) transformation curves.

I Consider the graphic on the next slide:

I For each terms of trade relation we determine thecorresponding production and consumption decision of firmsand households – and thus the export and import plans!

I These export / import plans (together with the correspondingt.o.t.) are displayed in a separate diagram.

I Combining all these point gives the offer curve (set of allexport/import quantities for allpossible t.o.t.). The origin isthe autarky situation.

S.255


y1 y1

y2

p

P = C

O

export / import plans

p′

P′

C ′

Ex′1

Im′2

O′

p′

Ex1

Im2

p′′

C ′′

P′′ Ex′′1

Im′′2

Ex′′1

Im′′2

O′′

p′′

OC

S.256


I Consider another country with corresponding plans toexchange goods: The intersection point of offer curvesdetermines the terms of trade relation where the countryspecific plans to exchange goods can be realized (equilibrium).

I Note, that relative price changes may have differentsubstitution effects. Depending on the price elasticity, offercurves can be “sloped backwards”. In this case, multipleequilibrium t.o.t. may occur. Some of them may not be stablesince marginal deviation leads to excess demand of exportgood and excess supply of the import good so that the t.o.t.relation moves to a stable relation (see graphic).

S.257


y2

y1

OC 1

OC 2

t.o.t.

y2

y1

OC 1

OC 2

t.o.t.

y2

y1

OC 1

OC 2t.o.t.

t.o.t.

(unstable)

S.258


Effect of a tariff on the terms of trade:

I A tariff makes the imported good more expensive. We assumethat the tariff revenues are given to the consumers(compensation for the higher import good prices).

I The domestic demand for imported goods decrease.Consumers are less willing to give away export goods in orderto obtain import goods. Thus, the offer curve rotates (aroundthe origin) to the left side (see graphic).

I This leads to an improvement of the terms of trade (steepert.o.t. relation): You have to renounce less export goods toobtain a certain amount of import goods. Also the total tradevolume decreases (see graphic).

S.259


Im1

Ex1

OC 1

OC 2(foreign)

t.o.t.OC 1

t

Tariffs shifts OC:

t.o.t.

- “improvement” of t.o.t.

- trade volume reduction

S.260


Optimal tariff theory:

I In case of perfect competition and a small country the optimaltariff is zero. In case of a large country we have seen thatthere is a positive tariff which maximizes domestic welfare.

I We now determine this optimal tariff graphically in a generalequilibrium setting (offer curves).

I As we have seen by the graphical construction, on each pointof an offer curve we have a different welfare level.

I By a similar construction principle we can obtain tradeindifference curves in the export-import-diagram. By shiftingthe production panel around a given welfare indifference curve(see graphic, left side) so that we obtain a continuum oftangential points, the origin of the shifted production paneldescribes the trade indifference curve: The relation ofexported to imported goods which give the consumers thesame welfare level.

S.261


y1 y1

y2

export / import

p

p

Ex1

Im2

Ex1

Im2

C

P

W

trade indifference curve

S.262


I When a certian terms of trade relation is given, the tangentialpoint where production and consumption point coincide,determines the same point on the offer curve and thecorresponding trade indifference curve (see previous graphic).

I For a scheme of different welfare indifference curve we obtaina corresponding scheme of trade indifference curve (see nextgraphic).

S.263


y1 y1

y2

export / import

p

p

W

trade indifference curves

p′

t.o.t.improvement

p′

W ′

S.264


Now it is easy to determine the optimal tariff by determining theoptimal equilibrium terms of trade:

I Since the offer curve of the other country is given, wedetermine the tangential point of the other country’s offercurve with the highest possible domestic trade indifferencecurve.

I This must then be the intersection point of the domestic offercurve (see graphic).

I Thus, implicitely we have determined the corresponding tariff.

S.265


y2

y1

OC 2 (foreign)

OC 1t.o.t. (free trade)

W 1

W 1t

Maximize W 1 subject to

foreign offer curve OC 2

OC 1OC 1t

⇒ optimal tariff

⇒ optimal OC 1t

t.o.t. (tariff)

S.266


Retorsion tariff:

I In the free trade case both trade indifference curves must betangential since this situation is Pareto efficient.

I The optimal tariff is a “beggar-my neighbour” policy becauseit worsens the terms of trade relation of the other country andharms its welfare! Its trade indifference curve is on a lowerlevel than in the free trade case.

I The other country takes the new offer curve of the homecountry as given and chooses the tangential point with itsown highest trade indifference curve (see graphic). This is the“best response” to the tariff of the home country. The tradevolume again decreases.

S.267


y2

y1

OC 2 (foreign)

OC 1t.o.t. (free trade)

W 1

W 2

W 1t

OC 1OC 1t

t.o.t. (tariff)

W 2r

OC 2 (foreign)

OC 2r

Optimal retorsion tariff

Reduction of trade volume

t.o.t. (tariff + retorsion)

W 1t

t.o.t. (tariff)

S.268


I This may induce a “tariff race” which ends up in anequilibrium situation with mutually best responses, where theresulting offer curve of country i = 1, 2 is tangential to thetrade indifference curve of country j = 1, 2, j 6= i .

I It depends on the curvature of the offer curves whether thisequilibrium point harms both countries or only one country,compared to the free trade situation (where in the graphicmust the equilibrium be if one country should benefit?)

I In case of one large and one (or many) small countries theeffects may be asymmetric so that the large country maybenefit even in case of retorsion tariffs.

S.269


(c) Foreign Monopoly

I We turn back to the general welfare effect of tariffs asdenoted in eq. (24).

I Assume that the import good is produced by a foreignmonopolist, and we have no import-competing domestic firm.This is a simplifying assumption since the aim of tariffs is toprotect the domestic industry.

I From the analysis of a large country we know that the welfareincreasing effect of a tariffs comes from the fact that itchanges the terms of trade dp∗/dt < 0 so that the increase ofthe domestic price is lower than t, or dp/dt < 1 (“partialpass-through” of tariff). We will see that this result can alsobe established in case of a foreign monopoly.

S.270


The domestic imports m equal the foreign exports x . We use theinverse domestic demand function p(x). The calculus of the foreignmonopolist, facing a specific tariff, is given by

maxxπ∗(x) = x(p(x)− t)− C ∗(x)

⇒ π∗′(x) = p(x) + xp′(x)− (C ∗

′(x) + t) = 0 (FOC)

with p(x) + xp′(x) as marginal revenues and (C ∗′(x) + t) as

marginal costs including tariff. Total differentiation gives

π∗′′

(x)dx − dt = 0 ⇒ dx

dt=

1

π∗′′(x)< 0

S.271


Now consider the effect on the domestic price:

dp

dt= p′(x)

dx

dt=

p′(x)

π∗′′(x)> 0

The price increasing effect is not surprising. The question iswhether dp/dt < 1. This requires

p′(x) > π∗′′

(x) = 2p′(x) + xp′′(x)− C ∗′′

(x)

With assuming constant marginal costs (C ∗′′

(x) = 0) this can betransformed to

p′(x) + xp′′(x) < 0

Then we have a partial pass-through and hence positive welfareeffects of a positive tariff if and only if the import demand functionis not “too convex”, e.g. linear.

S.272


I In Feenstra it is shown that a similar result holds true in caseof an ad valorem tariff p = p∗(1 + t). In this case thenecessary and sufficient condition is that the elasticity ofdemand increases as the consumption of the import goodfalls. This holds true e.g. for linear demand functions.

Theorem (Brander and Spencer 1984):

When the home country imports from a foreign monopolist withconstant marginal costs, then

I a small positive specific tariff increases home welfare ifmarginal revenue is steeper than the demand function,

I a small positive ad valorem tariff increases home welfare if theelasticity of demand increases as consumption of the importedgood falls.

S.273


(d) Tariffs in Cournot duopoly

I Consider the case that there is one domestic and one foreignfirm producing the (homogenous) import good.

I Domestic quantity is y , foreign quantity is x , world marketsupply is z = y + x .

I The world market demand is z(p) or as the inverse demandp(z) with p′(z) < 0.

S.274


I Foreign and domestic firm’s calculus is

maxxπ∗ = x [p(z)− t]− C ∗(x)

maxyπ = yp(z)− C (y)

with the FOC

π∗x = p(z) + xp′(z)− [C ∗′(x) + t] = 0 (25)

πy = p(z) + yp′(z)− C ′(y) = 0 (26)

I The optimal quantities x and y will therefore be mutuallydependent (“best response functions”):

x = r∗(y , t), y = r(x) (27)

S.275


I For “typical” demand and cost functions the best responsefunctions (reaction functions) are downward sloping and willhave an intersection point which is stable (standard Cournotduopoly solution, see next graphic).

I The intersection point is a Nash equilibrium which is thesolution of the system (27).

I From the foreign firm’s second order conditionπ∗xx = 2p′ + xp′′ − C ∗

′′< 0 and the total derivative of the

FOC: π∗xxdx − dt = 0 we obtain dx/dt = 1/π∗xx < 0 whichimplies that a tariff will shift the foreign firm’s reaction curvedownwards (see next graphic).

S.276


x

y

Rhome reaction curve

R

R∗

R∗

foreign reaction curve

A

foreign iso-profit curve

domestic iso-profit curve

R∗t

R∗t

B

S.277


What happens with price and quantity in case of the tariff?

I We add the two FOC (25) and (26) and rearrange:

2p(z) + zp′(z) = C ′(y) + C ∗′(x) + t

I Total differentiation and assuming constant marginal costs(C ′′ = C ∗

′′= 0) gives us

dz

dt=

1

3p′(z) + zp′′(z)

dp

dt=

p′(z)

3p′(z) + zp′′(z)

S.278


I It is unambigously clear that the domestic output will increasesince the reaction functions are downwards sloped (seeprevious graphic). This implies that output decreases if

dz/dt < 0 ⇐⇒ 3p′(z) + zp′′(z) < 0

I The price will increase less than the tariff (dp/dt < 1) if

p′(z) > 3p′(z) + zp′′(z) ⇐⇒ 2p′(z) + zp′′(z) < 0 (28)

I This is a weaker condition compared to the foreign monopolycase. The marginal revenue curve must slope down which isfor example the case for linear demand functions. Mostdemand functions satisfy this condition.

S.279


What is the impact on welfare?

I Recall (24) with m = x in our case:

dW

dt= t

dx

dt− x

dp∗

dt+ [p − C ′(y)]

dy

dt

Starting with the free trade case t = 0, the first term becomeszero, the second term is positive if (28) holds true. The thirdterms is also positive since the reaction curves are downwardssloping.

I As a result, a positive (small) tariff will increase domesticwelfare in case of “typical” demand functions.

S.280


I The analysis shows that tariffs leads to a “profit shifting” inmarkets with imperfect competition.

I Thus, trade policy is tempted to use this instrumentstrategically.

I Recall, that the same considerations also hold true for theother country which might induce a Pareto inefficient tariffequilibrium.

I Furthermore, it can be argued that (increased) profits mayinduce market entries so that profits erode. It can be shownthat in this case we need much stronger conditions to havepositive welfare effects of tariffs.

S.281


(e) Bertrand duopoly

I Consider heterogenous goods where firms are engaged in pricecompetition. The price of the imported good is p = p∗(1 + t)(ad valorem tariff), the price of the domestic good is q.

I We have therefore two different demand functions for thedomestic good y = d(p, q) and the foreign good x = d∗(p, q).

S.282


I The firm’s calculus is

maxpπ∗ =

pd∗(p, q)

1 + t− C ∗(d∗(p, q))

maxqπ = qd(p, q)− C (d(p, q))

with the FOC

p

(1− 1

η∗

)= (1 + t)C ∗

′(d∗(p, q)) (29)

q

(1− 1

η

)= C ′(d(p, q)) (30)

where η∗, η are the (positive) price elasticities of demand.We assume that second order conditions and the stabilitycondition of the solution are fulfilled.

S.283


I Solving (29) and (30) gives the best response functions

p = r∗(q, t), q = r(p) (31)

I Under “typical” demand conditions the reaction curves areupwards sloped (for details see Feenstra): a competitor’s priceincrease shifts the demand to the other firm, and its bestresponse to this is to increase the price as well, but theresponse is dampened, i.e. r∗q (q) < 1 and rp(p) < 1.

I A condition which guarantees this feature: the elasticitiesincrease as the relative goods price increase:η∗′(p/q) > 0, η′(q/p) > 0 (for details see Feenstra).

I From the second order conditions (same argumentation asbefore) we see that a tariff will shift the foreign firm’s reactionfunction to the right (see graphic).

S.284


q

p

R


R∗

R∗


Afor. iso profit

dom. iso profit

R∗t

R∗t

B

S.285


What are the effects on prices, quantities, and welfare?

I We have a positive terms of trade effect if there is a limitedpass-through of the tariff, i.e.

dp

dt

1 + t

p< 1 ⇒ dp∗

dt< 0

I It can be shown that the assumption we have made toguarantee upwards sloped reaction functions,η∗′(p/q) > 0, η′(q/p) > 0, is also sufficient to guarantee a

positive terms of trade effect!

I A tariff shifts the reaction function of the foreign firm to theright. Since the slope of each reaction function is less than 1,the new relative price p/q will increase.

S.286


I The welfare function and its differentiation in case of an advalorem tariff is

W = L + tpx + [qy − C (y)]− px − qy + U(x , y)

dW

dt= tp∗

dx

dt− x

dp∗

dt+ [q − C ′(y)]

dy

dtI Starting with a free trade situation (t = 0), the first term

becomes zero, the second term is positive under theassumptions made before (positive terms of trade effect). Thethird term, however, has an ambigous sign!

I The demand = output decreases due to the increaseddomestic price, but it increases due to the increased foreignprice (countervailing effect). If the domestic price does notincrease too much, then the domestic output will increase =the third term is also positive. Otherwise negative welfareeffects are possible.

S.287


Summarizing the results:

I Small countries with perfect competition will unambigously beharmed by tariffs.

I Large countries have an impact on the terms of trade. Theymay benefit from positive tariffs – to the expense of the othercountry. Retorsion tariffs and an inefficient tariff equilibriummight be the result.

I Tariffs may discipline a foreign monopoly. Similar effects thanin case of a large country occur.

I In case of duopoly markets the tariff effect depends on thetype of competition: In Cournot competition we have similareffects as in case of a foreign monopoly, and profits are shiftedfrom the foreign to the domestic firm. In case of Bertrandcompetition, welfare effects are ambigous.

S.288

7. Tariffs, Subsidies, Quotas, and Trade Policy7.2 Quotas and Voluntary Export Restraints

I Import quotas are constraints to the imported quantity.

I To import the good a firm needs a license.

(a) Import tariffs and quotas are equivalent in case ofperfect competition

I A tariff causes a reduction of import quantity and an increasein domestic firm’s output. An import quota will cause directlythe same reduction of import quantity. Since the consumerwill then demand more domestic products, domestic outputand price will rise.

I The partial equilibrium analysis is identical with the tariffcase: The import reduction of a small country has no effect onthe world market price, while a large country’s import quotahas a price reducing effect on the world market.

S.289


Small country:

p

y , c

D

S

p∗

y0 c0

p∗q

y1 c1

a cb d

p

m

MX

m0

quota m

m1

Xq

c b+d

S.290


I Disentangling welfare effects:I Consumer surplus loss: −(a + b + c + d)I Firm surplus gain: +aI What is about c?

I Area c are not tariff revenues anymore! Note, that animporter can buy goods on the world market at price p∗ andsell it to the domestic consumers to the price p∗q. Hence, areac represents the profits or rents of an importer. This maycreate different incentives and may lead to different outcomes:

(⇒ next slide)

S.291


1. Import licenses are given to domestic firms. Then c are the importerprofits as a part of the income. Total welfare (sum): −(b + d) as inthe case of tariffs.

2. There are rent-seeking activities of firms in order to obtain aprofitable import license. Consider that these activities are inefficientand resource consuming. In the worst case the loss due to rentseeking is as high as the additional rents. Hence, c does not accountfor welfare, and the total welfare is then −(b + c + d).

3. The import licenses are auctioned. If we assume that the highestpossible auction price is paid, then c repreents the auction revenuesfor the government, similar to the tariff revenues. The total welfareis again −(b + d).

4. The authority to control the quotas is given to foreign exporters.They are then able to extract the rents of the importers so that cflows to foreign firms. The resulting welfare is again −(b + c + d).

S.292


Large country:

p

y , c

D

S

p0

y0 c0

p∗

p∗q

y1 c1

tt a ce

b d

y1 c1

p

m

M

X

m0

Xq

quota m

m1

Xq

c

eb+d

Xq

S.293


(b) Import tariffs and quotas are not equivalent in case ofimperfect competition:

I Consider the case of a domestic monopoly (next slide).(a) In the free trade case, the monopolist faces a given p∗ and a

perfectly elastic world demand, i.e. demand and marginalrevenues are horizontal. In Cournot maximum he produces y0

which is the same solution as in case of perfect competition.(b) In case of a tariff, the price is p∗ + t. Output moves to y1,

consumption moves to c1, the imports fall from m0 to m1.(c) Now consider an import quota of m1. At price p∗ we have

imports m1. The residual demand for the monopolist is thenD −m1 with a corresponding marginal revenue curve MR.The monopolist is then able to charge the monopoly price forthe residual demand, since the consumers cannot substitutethe expensive good with imports. The output falls to y2, theprice increases to p2.

S.294


(a) Free trade

p

y , c

domestic D MC

p∗ world D

y0 c0

m0

(b) Tariff

p

y , c

domestic D MC

y0 c0

p∗ + t

y1 c1

m1

m0

(c) Quota

p

y , c

D MC

y0 c0

m1

D’

mMR

yM

p∗M

S.295


Cournot duopoly: same effect on quantity and prices

x

y


R

R∗

R∗


A

x quota

R∗t

R∗t

B

S.296


Bertrand duopoly:

I The effects of an import quota (or a “voluntary” exportrestraint, VER) are very different from tariff effects.

I Since x = x(p, q) depends on both prices, there is acontinuum of (p, q) with x(p, q) = x . which reflects theimport quota (see next graphic).

I Consider x on the level of the free trade (!) Bertrand solution.

I The existence of such an additional constraint when settingprices strategically changes the calculus of the firms. Note,that this constraint is common knowledge.

S.297


q

p

R


R∗

R∗


dom. iso profit

A

x(p, q)

B

A

S.298


I The function x(p, q) which implicitely defines the importquota (or VER) must lie between the reaction functions (seegraphic).

I Starting from a point on x on the upper right side of theBertrand solution, it would be profitable for the foreign firmto reduce its price according to its reaction function R∗.Unfortunately, this would shift the demand to the foreign firm,and the constraint x would be violated.

I According to this additional constraint, the best response isnow given by x(p, q) and the best response function is kinked(strong line in the graphic).

I The domestic firm can exploit this strategically by choosing aprice pS so that the best response qS with x(pS , qS ) = x givesthe highest possible iso-profit curve. This is comparable to theStackelberg solution in a sequential Bertrand game (“firstmover advantage”).

S.299


I The result is that we have no impact on the imported quantity(note that we have neglected effects on the total purchases).But there are price effects, and we have a welfareredistribution from consumers to firms.

I Note, that also the foreign firm benefits since it realizes ahigher iso profit curve! Therefore it is beneficial to berestricted by the other country, or, alternatively, to impose“voluntary” export restraints.

S.300

7. Tariffs, Subsidies, Quotas, and Trade Policy7.3 Export Subsidies and Strategic Trade Policy

I Export subsidies lead to costs for the government (incontrast to tariffs).

I Subsidies increase the price competitiveness of exporters: theywill produce and export more. Since it is more profitable tosell on the world market instead domestically, also thedomestic price must rise (law of one price).

I The additional firm’s profits minus the consumer’s surplus lossmust exceed the cost of subsidies. In case of perfectcompetition this cannot happen! Small as well as largecountries will face a welfare loss.

S.301


p

y , c

D

S

p∗

c0 y0

x0

p∗ + s

c1 y1

x1

x0

a b c d

p

x

X

world demand

x0 x1

Xs

b+d

S.302


I Subsidies lead to a right shift of the export function. In caseof a small country this has no effect on the world market pricep∗ (elastic demand function). The subsidy is a kind of markupto the firm’s profits.

I Disentangling welfare effects:I Firms profit gains: +(a + b + c)I Consumer surplus loss: −(a + b)I Cost of subsidies: −(b + c + d)I Total effect (sum): −(b + d) < 0

S.303


p

y , c

D

S

p0

c0 y0

p∗ + s

c1 y1

x1

p∗

D

S

c0 y0

a b c de

p

x

X

world demand

x0 x1

Xs

b+d

p

x

X

world demand

x0

e

S.304


I In a large country, the additional exports lead to a decrease ofthe world market price from p∗ to p′. Foreign consumersbenefit from the domestic subsidy!

I The terms of trade effect is therefore negative for thedomestic country! The large country is worse off compared tothe small country.

I Disentangling welfare effects:I Firms profit gains: +(a + b + c)I Consumer surplus loss: −(a + b)I Cost of subsidies: −(b + c + d + e)I ... including terms of trade effect: −eI Total effect (sum): −(b + d + e) < 0

S.305


Export subsidies with imperfect competition:

I Many markets for technological products are characterized byoligopolies (example: aircraft, cars etc.). Governments aretempted to help the domestic industry to become more pricecompetitive on the world market.

I In some cases the world market demand carries only a limitednumber of competing firms. Subsidies may then help thatdomestic firms will stay in the market while foreign firms haveto leave it.

I If there are economies of scale or decreasing average cost forcumulative production, the subsidy could help that thedomestic firm’s output rises so that lower average costs arerealized. If then the firms have become dominant, thesubsidies could be reduced evtl. to zero.

S.306


Cournot duopoly (Brander and Spencer (1985))

I Homogenous good, one domestic and one foreign firmI Domestic quantity y , foreign quantity x , world market:

z = x + y .I Domestic firm receives a subsidity s per produced output.I Calculus of the firms:

maxxπ∗ = p(z)x − C (x)

maxyπ = [p(z) + s]y − C (y)

with the FOC

π∗x = p(z) + p′(z)x − C ′(x) = 0

πy = p(z) + s + p′(z)y − C ′(y) = 0

S.307


I From the second order condition πyy < 0 and the totalderivative of the FOC πyydy + ds = 0 we receive

dy

ds= − 1

πyy> 0

which implies a right shift of the reaction function in case of asubsidy.

I With a downwards sloped rival’s reaction curve this implies anincrease in domestic output and a decrease of foreign firm’soutput in equilibrium.

I Subsidies lead to a profit-shift favoring the domestic firm(point A shifts to B).

S.308


x

y


R

R∗

R∗


A

R

R

Rshome reaction curve

Rs

B

S.309


I By technical assumptions guaranteeing the negative slope ofthe reaction functions and the stability of the Nash solution,the foreign firm’s response to the domestic firm’s outputincrease is not too elastic. Hence, total output increases andthe price falls.

I For determining the welfare we make a simplfying assumption:Both firms sell their products to a third market so that we donot have to take consumer preferences into consideration.

I Welfare is then firm profits minus the subsidy costs:

W = [p(z) + s]y − C (y)− sy = p(z)y − C (y)

which are the profits without subsidies so that we can relabeliso-profit curves as welfare indifference curves.

S.310


I The other country will clearly respond to this trade policywith “retorsion” measures.

I Mutually optimal subsidies will decrease profits (= inefficientNash solution), see Brander/Spencer (1985).

I Recall, that a subsidies increase total output and reduce theprice. Therefore, consumers will benefit, and allocation isshifted towards the competitive solution. Brander/Spencer(1985) show, that joint welfare maximizing subsidies aregiven by:

s = (x − (c + c∗))p′, s∗ = (y − (c + c∗))p′

where c , c∗ are the consumed quantities of the imperfectcompetitive good. In a closed 2-country world we havey + x = c + c∗. If total consumption exceeds the productionof one country, the optimal subsidies are positive!

S.311


Bertrand duopoly (Eaton and Grossman (1986))

I Heterogenous goods with demand functions y(q, p) fordomestic and x(p, q) for foreign goods and q as the domesticand p as the foreign price.

I We assume a specific subsidy paid for each produced unit.

I Firm’s calculus is given by

maxpπ∗ = px(p, q)− C (x(p, q))

maxqπ = (q + s)y(q, p)− C (y(q, p))

with the FOC

π∗p = x(p, q) + pxp − C ′(x(p, q))xp = 0

πq = y(q, p) + (q + s)yq − C ′(y(q, p))yq = 0

S.312


I Again, we receive upwards sloped reaction functions q(p, s)and p(q) with a damped response to the rival’s price, i.e. thecross elasticity of price changes is less than one.

I From the second order condition πqq < 0 and the totalderivative of the FOC πqqdq + yqds = 0 we obtain

dq

ds= − yq

πqq< 0

which implies that a subsidy shifts the reaction functiontowards the origin, and the prices will decrease. The result is alower welfare!

S.313


q

p

R


R∗

R∗


W

A

R

R

Rs

Rs

Ws

B Wopt

C

Ro

Ro

S.314


I Assuming that products are sold on a third market, we canneglect consumer’s utility. Welfare is then profits minussubsidy costs:

W = (q+s)y(q, p)−C (y(q, p))−sy(q, p) = qy(q, p)−C (y(q, p))

⇒ Again, iso profit curves (without subsidies) can be relabeled aswelfare indifference curves.

⇒ Lower welfare because additional profits are more than offsetby subsidy costs.

I With negative subsidies = taxes the results reverses. Anoptimal tax would imply that the highest iso profit = welfareindifference curve is tangential to the rival’s best responsefunction (point C in the graphic).

S.315


Theorem:

a) Under Cournot competition, a subsidy of exports increaseswelfare. Joint welfare maximization is possible.

b) Under Bertrand competition, a subsidy of exports decreaseswelfare. Instead, a tax on exports raises welfare.

I Any “strategic” use of (positive or negative) subsidiesdepends very sensitive to the type of competition. “Since wedo not expect the government to be able to distinguish themarket conduct, it becomes impossible to implement thesepolicies in a welfare-improving manner” (Feenstra, 8-53).

S.316


I In case of lobbying for subsidies, lobbying costs have to betaken into account. In equilibrium, the welfare will be reduced,see Moore, M.O., Suranovic, S.M. (1993), Lobbying and Cournot-Nash

competition : Implications for strategic trade policy. Journal of

International Economics 35(3-4), 367-376.

I In case of cross-ownership, the calculated “optimal” subsidiesare much lower, since also the profits of domestic firms arereduced, see Dick, A.R. (1993), Strategic trade policy and welfare : The

empirical consequences of cross-ownership. Journal of International

Economics 25(3-4), 227-249.

S.317


Further results from the literature (see Feenstra for details):

I Consider multiple export firms under Cournot competitionwith private knowledge about their different productivities.There exist subsidy schemes such that information aboutproductivities are revealed and most productive firms receivepositive subsidies while least productive ones are taxed.

I Subsidies can be paid for building up production capacitiesinstead of subsidizing output. This works as a market entrybarrier for foreign firms.

I Subsidies can be paid for increased cumulated output in orderto experience learning curve effects (decreasing averagecosts). This provides an advantage in the world market whichis still in place when the subsidy is reduced.

I Subsidies can be paid for research and development. Thereforeit reduces fixed costs und builds up a market entry barrier.

S.318


Subsidy war:

I If (unilateral) subsidies have a domestic welfare increasingeffect then both countries will engange in this policy.Analogously to the retorsion tariff case, both countries willthen be worse off.

I “Prisoner’s Dilemma” Game: subsidizing is a dominantstrategy but the equilibrium outcome is pareto-inferior.Necessity to negotiate contracts to reduce or eliminatesubsidies.

I Example aircraft industry : Airbus (Europe) andBoeing/McDonnell-Douglas (USA). Decreasing average costswith cumulated output. In 1992, USA and Europe negotiatedto limit subsidies to 33% of the development cost.

S.319


Discussion: Can Trade Policy be Welfare Improving?

Summing up the results:

I In a world with perfect competitive markets, ...I ...free trade is welfare improving compared to autarkyI ...trade policy (tariffs, quotas, subsidies) is...

I ...welfare decreasing in case of s small countryI ...typically welfare decreasing in case of large countries plus

mutually optimal policies

(welfare improvements by unilateral measures are always“beggar-my-neighbour” policy)

S.320


I In a world with imperfect compeition, ...I ... free trade can be welfare improving by trade and

specialization effects and the additional effect of enhancingcompetitive pressure.

I ... free trade can be welfare decreasing e.g. for a small countryin case of economies of scale which faces a decline of theterms of trade.

I ... trade policy may help to induce firms to behave morecompetitive like in Cournot oligopoly with subsidies. However,this requires a socially optimal design of policy measures whichis typically not the Nash solution. Each market imperfectioncreates room for welfare improving measures.

S.321


Dynamic perspective: infant industry protection, fosteringindustrialization

I Infant industry protection would be efficient if the newindustry has potential comparative advantages which cannotbe realized due to a locked-in specialization pattern and otherimperfections like limited access to capital markets. Protectingthe infant industry would enable it to exercise learning curveeffects, effects from economies of scale and scope etc. (seegraphic, right side)

I Protection by “educational tariffs”; import-substitutingindustry has an increased price-competitiveness (importsubstitution strategy of development)

I Infant industry protection should lead to a shift of theproduction frontier, so that both countries will face a welfareimprovement when turning back to free trade (see graphic,left side).

S.322


y2infant industry

y1

A

B

AC

cumulated output

AC established firm

AC infant firm

y2

S.323


I Empirical evidence (e.g. in South America) shows that thisleads to the desired industrialization effect, but the effects ongrowth and development have been modest/disappointing.

I Success in industrialization comes along with negative effects:

I development is limited due to the lack of human capital andlack of industrial experience,

I as protectionism rises the prices, also less eficient firmsexperience an increase in competitiveness,

I a proper development strategy requires a complex mix of(trade) policy measures which could hardly be implemented bygovernments of devloping countries.

S.324


Export-oriented development strategy instead ofimport-substitution?

I Empirical evidence shows that countries with exportorientation and large trade volumes face a much faster growthand development (Hongkong, Taiwan, Singapore, Malaysia,South Corea, China, ...).

I However, export orientation and trade does not imply freetrade policy. Most countries impose strong barriers to trade orprice distortions, and have a strong regulatory industrial policy.

⇒ Mixed evidence!

S.325


Welfare improvement by trade policy due to external effects:

I Externalities (e.g. environmental externalities) create room forwelfare improvements due to regulation. Cross-borderexternalities may be a case for trade barrier policy?

A disaggregated view on welfare improvement

I Aggregated welfare functions are problematic since theypresume a representative agent.

I Trade has an impact on...I ... individual income (shift of welfare between factors)I ... relative prices (shift of welfare between consumers with

different preferences)I ... profits (shift of welfare between factors and firms)

I Is it possible that trade gains are distributed such that it isbeneficial for every agent = Pareto improving)? Yes! But thismay require a complex transfer scheme = regulation (seeMarkusen (1981), Helpman (1984))

S.326

7. Tariffs, Subsidies, Quotas, and Trade Policy7.4 Norms and Regulation

Regulating imports without tariffs, quotas, and subsidies:imported final or intermediate goods must ...

I ... have some physical/technical specifications, e.g.I electric devices must be compatible with domestic electirc

power systemI technical safety norms, food safety norms (consumer

protection), environmental safety normsI adhering specific quality standards

I ... have been produced according to specific conditions, e.g.I without child labor; according to ILO labor normsI according to environmentally sustainable conditionsI must embody at least x% material from country X , or must

contain at most y% material from country Y .

S.327


I If the foreign exporter does not comply with these norms hecannot export ⇒ therefore seen as a barrier to trade!

I Main issue in trade negotiations (such like TTIP, TPP etc.) toharmonize standards which often means to relax them tothe lower bound.

I However, the crucial question is: What are the reasons forthese norms?

S.328


I Protect domestic producers = norms have the purpose to be atrade barrier.

I Bust most norms or market regulations have been introducedas an answer to market failure, not as a barrier to trade:

I Facing (negative) externalities: environmental normsI Facing information asymmetry problems: consumer protectionI Creating benefits from standardization (network externalities,

scale effects)

I These regulations/norms reflect (via democratic voting) thepreferences of the population how to properly address thesemarket problems. Example: different preferences in US andEurope regarding bank regulation or food safety.

S.329


I Thus, these norms are not designed as a hindrance for tradeand for protecting national firms: abandoning or relaxing thesenorms in order to foster trade is therefore associated with highsocial costs!

I Problem: trade negotiations are often intransparant andmassively exposed to pressure of rent-seeking groups ⇒ lowlegitimation but with high binding power (supranationaltreaty) ⇒ danger that social costs will exceed benefits fromadditional trade.

I Again:

* Rodrik, D. (2011), The Globalization Paradox: Democracy andthe Future of the World Economy. New York and London:W.W. Norton.

* Rodrik, D. (2017), Straight Talk on Trade: Ideas for a SaneWorld Economy. Princeton University Press.

S.330


Are norms always a hindrance for trade?

I Learning to comply with production and quality norms as asource for gaining comparative advantages and to attracttasks in the GVC :

Fricke, S., Chapman, G. (2017), The Role of Standards in North-South

Trade: The Case of Agricultural Exports from Sub-Saharan African

Countries to the EU. Jena Economic Research Papers No. 2017-011

I Further:

Gandal, N., Shy, O. (2001), Standardization Policy and International

Trade. Journal of International Economics 53(2), 363-383.

S.331

8. Further Topics in Globalization and Trade8.1 Trade and Unemployment

I Davis, D.R. (1998), Does European Unemployment Prop UpAmerican Wages? National Labor Markets and Global Trade.American Economic Review 88, 478-494.

I Davis, D.R. (1998), Technology, Unemployment, and RelativeWages in a Global Economy. European Economic Review 42,1613-1633.

I Question: What happens in a trade model when we assumeimperfect labor markets with unemployment?

I For simplicity, it is assumed that unemployment is generatedby minimum wages wmin (neoclassical standard argument).

S.332


We start from a neoclassical framework:

I Input factors: unskilled labor L with factor price w and skilledlabor H with factor price r .

I Two sector model: Good Y is unskilled labor intensive, goodX is skilled labor intensive (non-reversal of intensities).

I Good Y is the numeraire, so that p = MC (r , w) is the priceof good X = terms of trade in free trade case

I Constant returns to scale

I Perfect competition in goods markets

S.333


We start with a closed economy:

I As we know from the Heckscher-Ohlin approach, there is anunambigous relation between relative factor endowmenth = H/L and relative price p = λ(h) with λ′(h) < 0.

I From the Stolper-Samuelson theorem we know that there isan unambigous relation between relative goods price p and therelative wage w = w/r , namely w = Ψ(p) with Ψ′(p) < 0.

I In a closed economy equilibrium we have for a given relativeendowment h an equilibrium price p∗ and an equilibrium wagew∗.

S.334


w h

pStolper-Samuelson Heckscher-Ohlin

w∗

p∗

S.335


I Now impose a minimum wage wmin. This makes theunskilled labor good relative more expensive than the skilledlabor intensive good.

I Hence the relative autarky price p will decrease to pu.

I However, wmin > w∗ implies pu = Ψ−1(wmin) < p∗ whichimplies hu = λ−1(Ψ−1(wmin)) > h for the employed factors.

I Since h represents the (full employed) factor endowment,hu > h implies that a part of unskilled labor is unemployed.

I It is N the employed unskilled labor, U is the unemployedamount. Then we have L = N + U andU = L− H/hu = β(hu).

S.336


w h

pStolper-Samuelson Heckscher-Ohlin

w∗

p∗

wmin

pu

U

U

β(h)

S.337


Now we turn to the free trade case: What happens with theFactor Price Equalization (FPE) in case of minimum wages?

I Consider a country with flexible wages for skilled and unskilledlabor (USA), and a country with minimum wages for unskilledlabor (Europe), where wmin induces unemployment.

I See next graphic: The FPE set in case of global flexible wagesis given by the dotted vertex. Due to minimum wages inEurope, the origin of FPE set moves. We have still fullemployment of skilled labor but unemployment U of theunskilled labor. The FPE set is smaller and shifted to the left.

S.338


OUSA

hu

man

cap

ital

OEU

labor

OuEU U

CA

B

D

S.339


I Although the European minimum wages are fixed, it ispossible to have FPE! This implies that American unskilledworkers will also be better off due to the European labormarket imperfections.

I It is possible that there exists world endowment vectors whichlead to FPE and partial specialization only in case of flexiblewages, e.g. point C, as well as endowments which will lead topartial specialization only in the minimum wage case, e.g.point D.

S.340


How trade affects the level of unemployment and welfare: (seegraphic)

I Assume two countries (USA, Europe) which are identical in allrespects except for the minimum wage for unskilled labor in Europe.The assumptions made before hold true.

I Autarky in Europe: in point A (full-employment) we have a pricerelation pu which supports the minimum wage wmin.

I However, with this price relation pu, consumers would decide forpoint A′, implying an excess demand for X and excess supply of Y .

I Excess demand should lead to higher prices pu and lower wages w(Stolper-Samuelson). Since we are already at wmin, this is notpossible! Thus the economy has to restrict the output as long asplanned production equals planned consumption.

I Unemployment has the same effect than labor outflow ⇒ outputeffects according to the Rybczynsky line.

I With homothetic preferences, the economy will produce andconsume in point E which lies on the Rybczynski line from pointA. Since we have unemployed unskilled labor, the endowmentrelation h increases but the production opportunities decrease.Hence, the Rybczynski effect holds true!

S.341


Y

X

pu

(cannot be steeper!)

A

A′

Y

X

Rybczynski effect

Rybczynski line

homothetic preferences

E

Waut

A

A′

F ′

F

Wtrade

S.342


I Autarky in USA: The tangential point with the indifferencecurve would be downhill from point A, so we would havepUSA > pu. Europe then has a comparative price advantagefor good X , USA have a comparative advantage for good Y .

I Free trade: Normally, the terms of trade would be betweenboth autarky price relations, but in Europe, p > pu is notpossible, hence we take pu as the given terms of trade.

I With pu, American firms choose point A while Americanconsumers choose A′. This results in an excess demand forgood X and an excess supply of Y . Again, a price adaptionaccording to excess demand is not possible.

I The European economy will adapt the production to point F ,and it will consume in F ′, so that the goods exchange plansare compatible (see triangles).

I As a result, the introduction of free trade will increaseunemployment!

S.343


The impact of factor accumulation:

I Assume that H increases in the USA. According to theRybczynski theorem the production of X increases while theproduction of Y is reduced. The USA demand on the worldmarket for X is reduced. Therefore, the European productionshifts from X in favor to Y , given the fixed terms of trade.We move from point E in direction of point A. The Europeanunemployed people (and the welfare) benefit from the bettereducation in the USA.

I Assume thet H increases in Europe. The minimum wagesprevent the conventional Rybczynski effect. The productionfrontier shifts to the right. The production of both goods willincrease proportionally, and the employment effect is∆N = ∆H/hu.

S.344


I Assume that L increases in the USA. Thus, its demand for Xwill increase. By the same argument we have put forwardbefore, this will increase European unemployment.Specifically, unemployment grows by the same amount as theforeign labor force ∆UEur = ∆LUSA > 0. There is no effect onglobal output.

I Assume that L increases in Europe. Output is not affected bythis. The additional labor force is completely unemployed.

I The framework has to be extended to different technologieswhich imply wage differentials.

S.345


Related topics:

I Unemployment and migration (unemployment as a cause formigration).

I Other reasons for unemployment than minimum wages.

S.346

8. Further Topics in Globalization and Trade8.2 Multinational Firms in Developing Countries

I In a Walrasian world where all factor prices equilibrate themarkets, implying full employment, the existence of MNFenlarge the FPE set and might help to enhance the efficientallocation of resources.

I Nevertheless, in the process of globalization, MNF have oftenbeen criticized for their activities especially in developingcountries. It was claimed that MNF may harm the domesticmarkets of developing countries, extract rents, and discouragedevelopment.

I If labor and capital markets are not perfect, it is possible thatMNF indeed may harm developing countries. We discuss:

Batra, R.N. (1986), A General Equilibrium Model ofMultinational Corporations in Developing Countries. OxfordEconomic Papers 38(2), 342-353.

S.347


Assumptions:

I Two countries: capital-abundant home country of the MNF,underdeveloped labor-abundant country. The latter is small,i.e. it takes the terms of trade as given.

I Two sectors in the developing country: One sector X withcompetition among MNFs, in the other sector Y only localfirms compete.

I Two non-specific factors, labor L and capital C , one specificfactor S (knowledge, technology), owned by the MNF.

I Standard neoclassical production functions.

I Full employment of capital. Labor is not necessarily fullemployed since wages are determined institutionally (w isfixed).

S.348


I Production functions:

X = X (Lx ,Cx ,S), Y = Y (Ly ,Cy )

withI Xi > 0,Xii < 0,Xij > 0 (i , j = L,C ,S , i 6= j) and

Hx ≡ XLLXCC − X 2CL > 0,

I Yi > 0,Yii < 0,Yij > 0 (analogously) andHy ≡ YLLYCC − Y 2

LC = 0.

I The assumptions regarding Hx ,Hy seem to be strange but aretrivial if we consider Cobb-Douglas production functions!

I It is assumed that MNF transmit S into the host country andutilizes it in production using local capital and labor. Later itis permitted that also capital is transfered to the developingcountry.

S.349


I The profits of the MNF are given by

π = pX − wLx − rxCx − F

= pXSS − F (from Euler equation) (32)

where F are the fixed costs of the specific factor S , p is theprice of the good (in terms of units of the good of the othersector = numeraire).

I Profit maximization of MNFs gives

w = pXL, rx = pXC (33)

while the profit maximization of the local firms give

w = YL, ry = YC (34)

S.350


I It is assumed that the capital market is imperfect (e.g. dueto information asymmetries), so that the MNF has a betteraccess to the capital market compared to local firms.

I It is reasonable to assume that banks prefer to lend money toMNF due to international reputation, more collaterals, lowerrisk etc. Thus, interest rates will differ:

αrx = ry with α > 1

I Full employment: Cx + Cy = C ,S = S

S.351


I Combining the profit maximization and full employment conditionswe have

αpXC (Lx ,Cx , S) = YC (Ly , C − Cx ) (35)

w = pXL(Lx ,Cx , S) (36)

w = YL(Ly , C − Cx ) (37)

I What happens if MNF transfer specific factors S to the developingcountry and utilizing the resources there? For simplicity, assumep = 1. Differentiating (35)-(37) and rearranging gives:

dLx

dS=αYLL(XLSXCC − XCSXCL)

D> 0

dCx

dS=αYLL(XCSXLL − XLSXCL)

D= −dCy

dS> 0

dLy

dS=αYCL(XCSXLL − XLSXCL)

D< 0

with D = −αYLLHx > 0. S.352


Result 1: The emergence of MNF in the capital-shortlabor-abundant developing country induces a shift of capital fromthe local firms to the MNF. Employment and output increases inthe MNF sector but decreases in the local sector.

I What happens with the total employment? AddingdLy/dS + dLx/dS = dL/dS gives (see Batra for details)

dL

dS=αA(ky − kx )

D

with A ≡ YCL(XCL(XCS + XLS/kx ) + SXCSXLS/Cx ) > 0 andki = Ci/Li as the capital intensity. This leads to

Result 2: If the MNF produces the relative capital-intensive good,then the emergence of the MNFs cause a decline in totalemployment in the developing country.

S.353


I Consider that the profits of the MNF are repatriated to thedeveloped country, then the remaining GDP of the developingcountry is

GDP = pX + Y − π

I Differentiating with respect to S and using (32), (33), (34)gives (for details see Batra)

dGDP

dS= w

dL

dS− S

dXS

dS+ XC (1− α)

dCx

dS< 0

I Since dL/dS is negative (see above) and α > 1 the impact onthe GDP is negative.

Result 3: The MNF cause a decline of the GDP when the transferof S is not accompanyied by a sufficient capital transfer.

S.354


Additional results:

I In the article it is shown that a tax on the non-wage incomein the developing country can help to avoid these negativeeffects.

I Furthermore, it is shown that the effects might reverse whentechnology transfer S is accompanyied by capital transfers(FDI). This is employed by assuming dC/dS = λ > 0.Recalculating the effect on the total employment gives

dL

dS=α[A(ky − kx ) + λYCLHx ]

D

where the sign depends on λ.

S.355

8. Further Topics in Globalization and Trade8.3 “Fair” Trade, Labor Standards and Ethical Consumption

What does “Fair” Trade mean?

I First of all: normative judgement!

I Specialisation and trade affects different groups in a differentway: It might be the case that unskilled worker in the Southcreate benefits for the consumers in the North under extremesocial conditions.

I Is trade responsible for this? Can trade or trade policy addressthis problem?

I More generally, trade situations and trade policy measuresmay be regarded as “unfair” if

I ... there is a very disproportional distribution of the welfaregains from international division of labor and trade

I ... they induce a welfare shift to the expense of the tradepartner (e.g. beggar-my-neighbour policy, profit shifting, rentextraction)

S.356


I Many economists therefore come to the conclusion that morecompetition, open markets, less distortional policy measuresincreases fairness, including the chance for poor workers in theSouth to “catch up” with the North.

⇒ “Free trade is fair”

⇒ ?

S.357


I Even though the globalisation process proceeds, we observe indeveloping countries which are engaged in this process, thatlarge groups (i.e. unskilled workers, rural population) do notbenefit and still produce under conditions which would beunacceptable in developed countries.

I Many economists argue, thatI ... labor force in poor countries is better off compared to the

autarky case.I ... wage differentials reflect productivity differences (and

different purchasing power of money wages)I ... wages in countries which are engaged in an export orientited

development strategy have a mich higher productivity andwage growth rate than under import substituting policy.Globalisation fosters the “catching up” process.

I ... this problem should be addressed by development policyrather than trade policy (e.g. technological progress andinstitutional reforms)

S.358


Objections:

I Most positive examples of catching up economies (Japan,“Tiger countries”, China) are not based on pure free tradepolicies but on a sophisticated mix of trade policy measuresand regulatory industrial policy.

I The theoretical arguments why free trade will be “fair” andbeneficial also for the labor force in the South, are based onmodels which presume competitive markets, full employmentetc. – conditions which are not valid.

What happens if considering market imperfections likemonopsonistic labor markets or oligopolistic intermediaries(often Northern firms)?

S.359


Fair Trade – an ethical consumption perspective

I Ethically motivated consumers in developed countries payvoluntarily more for a product, if its production in thedeveloping countries meets certain social standards.

I Producer and intermediaries can join a Fair TradeOrganization (FTO) which proves the standards and isproviding a label for the product (e.g. TransFair Coffee).

I The social standards typically include: minimum wages forworkers, long-term contracts between producer andintermediary and/or resellers for dampening price fluctuations,standards for work safety and social insurance, access tospecific long-term credits.

I The aid of ethically motivated consumers is channeledthrough trade.

S.360


What may be possible effects?

I Allocation efficiency : FTO and fair products change thesystem of relative prices. This will induce a reallocation offactors and may affect efficiency, the specialisation pattern,and welfare. Are ethical price premiums distorting allocation?

I Distribution: workers in FT relationships will be better off, butis this to the expense of other worker groups in the South?(e.g. wages of non-FT workers, unemployed workers)

I Welfare: Is there an overall positive welfare effect for theeconomy?

S.361


A first general consideration:

I If we are in a perfect Walrasian world, then an equilibriumallocation where consumers have specific (ethical) preferencesmust be efficient and Fair Trade must then be a Paretooptimum. There is no logical basis to argue that ethical prices“distort” the system of relative prices.

I Otherwise we are not in a Walrasian world = there existimperfect markets. But if markets are imperfect, then theallocation without ethical consumption can not be Paretoefficient. Then perhaps Fair Trade may help to mitigateinefficiencies.

S.362


Asymmetric information:

I Choice consequences (e.g. poor labor conditions) are notreflected in the prices. If consumers have social preferences,the price system is incomplete: prices cannot reliably reflectmarginal willingness to pay.

⇒ Allocation inefficiency

I FT labelling (among others) aim to incorporate morechoice-relevant information into the price system.

I Paradox: more globalization / division of labor ⇒ moreinformation asymmetries ⇒ efficiency?

I Only in case of pure self-interest price system is (perhaps)complete.

S.363


I Hayes, M. (2006), On the Efficiency of Fair Trade. Review ofSocial Economy 64(4), 447-468.

I Consider firms which act as monopsony on local labor markets.I A monopsomy on a labor market is a wage setting firm which

chooses a wage and employment level that maximizes profits,given a (inverse) labor supply function:

maxLπ = py(L)− w(L)L

FOC ⇒ 0 = pdy

dL− w(L)− dw

dLL

w

p=

dy

dL− dw

dL

L

w

w

p

⇒(

1 +dw

dL

L

w

)w

p=

dy

dL

which implies that real wages are below marginal productivity.S.364


I Households decide whether to accept this wage contract or tobe self-employed – which means in poor rural regions in theSouth that they “survive” on a subsistence level.

I Wages below marginal productivity depart from efficiency andlead to lower employment.

I The non-employed worker should be regarded as unvoluntaryunemployed, since lower wages would not induce more labordemand.

I The model shows that FTO work in a similar way like asubsidy on the wage or a minimum wage.

I In monopsony it can be shown that minimum wages or wagesubsidies are able to enhance both, efficiency and employment.

I A further result is, that this does not necessarily require aprice premium for ethical consumers.

S.365


I Piyapromdee, S., Hillberry, R., MacLaren, D. (2008), “Fair Trade”

Coffee and the Mitigation of Local Oligopsony Power. Research

Paper No.1057, University of Melbourne, Department of Economics.

I The paper presents a model which describes the way howcoffee is produced: Small farmers harvest coffee berries andsell them to private intermediaries. These intermediariesprocess the berries to green coffee beans. These beans areexported by local exporters to international traders which sellthem on the global coffee market to the roasters. The roastersprocess the beans to the final product and distribute themthrough resellers to the consumers. In practice, some steps ofthis process are done by vertically integrated large firms.

I The model focusses on the first part: Local intermediarieswhich demand the harvested berries are characterized asoligopsonies with market power.

S.366


I FTOs address the interaction between farmers and theseintermediaries (processors). There are commercial processorswhich maximize their profits, and fair trade processors whichmaximize a goal function which compromises their profits withthe benefits of the farmers. The weight α of the farmer’sbenefit in the goal function indicate the influence of FTO.

I A detailed analysis of the general heterogenous oligopsonymodel shows that an increasing influence of fair tradeprocessors reduce the oligopsony power of commercialprocessors. The outcome moves into the direction ofcompetition, and the extraction of farmer’s welfare is reversed.

I A calibration and estimation of the model shows, however,that the effects are positive but small.

S.367

8. Further Topics in Globalization and Trade8.4 Does Trade Lead to Inefficient Use of Nature?

Dam, T.A., Pasche, M., Werlich, N. (2017), Trade Patterns and the

Ecological Footprint – a theory-based Empirical Approach. Jena

Economic Research Papers No. 2017-005.

Basic Idea:

I Production and consumption makes use of naturalenvironment (renewable and non-renewable resources,absorptive capacity for emissions etc.).

I There is no “market” for environmental services.Nevertheless their use create costs:

I costs of appropriating natural resourcesI costs due to (strictly enforced) environmental policy

regulations

S.368


I According to neoclassical theory, countries with comparativecost advantages of environmental services will specialize onpollution-intensive goods:

I countries which are abundant of biocapacityI countries with lax environmental policy

(“Pollution Haven Hypothesis”)

I Richer countries typically have stricter environmental policythan poor countries.

I Problem that globalization leads to imbalances because richcountries shift dirty industries to poor countries which thussuffer from globalization.

I But does this imply inefficient use of nature?

S.369


I Broad measure of environmental impact: the EcologicalFootprint (EF). Unit of measure are global hectares (gha).

I The statistics separates EF of production from EF ofconsumption, and thus EF embodied in traded goods.We are interested in the net EF exports.

I Positive (negative) net EF imports (exports) means that acountry “outsources” environmental use to other countries.

I Furthermore, the biocapacity (in gha) is measured.

I The overall EF of a country is the relation of the EF ofconsumption relative to the biocapacity. A value larger than 1indicates that the consumption bundle requires more landthan this country has, meaning that the consumption patternis unsustainable.

S.370


Source: www.footprintnetwork.org

S.371


Source: Dam/Pasche/Werlich (2017)

S.372


I Assume a 2-country case: in autarky both countries eachproduce and consume a bundle of goods where in country 1the entire biocapacity is used while country 2 excessivelyexceeds its biocapacity. So the global EF is larger than 1.

I Assume that the marginal pressure on the nature is amonotonous function of the gha used in local production.

I See figure: in autarky the marginal pressures on theenvironment differ. The total pressure would be minimized ifthe marginal pressures are equalized. This requires thatcountry 1 produces more of the pollution-intensive good (andcountry 2 less), which is then traded.

I The fact that country 2 drastically exceeds its biocapacity inautarky is empirically related to the case that country 2 has alow biocapacity or a high GDP.

S.373


g(EF1) g(EF2)

gha use

Bio1 Bio2excess use

gha in consumption

(autarky: alsoin production)

A

B

C

gha in production

(trade)

S.374


I The Heckscher-Ohlin-Vanek (HOV) model predicts thatcountries with

I relative large biocapacity per capitaI low environmental policy standards

(or not stringently enforced policy)

have a comparative advantage for dirty industries and willthus specialize on pollution-intensive products.

I Hypothesis 1: countries which are abundant of environmentalgoods and/or where the use of environmental goods isrelatively cheap due to lax environmental policy, are netexporters of the EF.

I Hypothesis 2: Richer countries are more likely to become anEF net importer.

S.375


Findings for the regression of net export of ecological footprint in the extended econometric model

FE FE FE RE RE RENetExEF NetExEF NetExEF NetExEF NetExEF NetExEF

Biocapacity 0.314∗∗∗ 0.321∗∗∗ 0.343∗∗∗ 0.359∗∗∗ 0.361∗∗∗ 0.375∗∗∗

(7.64) (7.89) (8.38) (12.53) (13.17) (13.51)

EPSCPI -0.014∗∗∗ -0.012∗∗ -0.021∗∗∗ -0.012∗∗∗ -0.009∗∗ -0.020∗∗∗

(-3.06) (-2.54) (-4.38) (-2.84) (-2.12) (-4.40)

GDPCAP (ln) -0.245∗∗∗ -0.277∗∗∗ -0.253∗∗∗ -0.234∗∗∗ -0.244∗∗∗ -0.258∗∗∗

(-3.11) (-3.53) (-3.26) (-3.29) (-3.50) (-3.71)

ProdEF Green -0.042∗∗∗ -0.045∗∗∗

(-3.48) (-3.79)

ConsEF Green 0.060∗∗∗ 0.063∗∗∗

(4.34) (4.66)

Constant 1.288∗ 1.679∗∗ 1.093 0.970 1.182∗ 0.987(1.68) (2.19) (1.45) (1.39) (1.73) (1.45)

N 544 544 544 544 544 544Year effect Yes Yes Yes Yes Yes Yes

R2 within 0.208 0.227 0.237 0.207 0.226 0.237

R2 between 0.760 0.765 0.777 0.764 0.775 0.777

R2 overall 0.746 0.752 0.763 0.750 0.762 0.763

* p < 0.10, ** p < 0.05, *** p < 0.01 (one-tailed test); t statistics in parenthesesModels with year effect include unreported year-specific interceptsFE denotes fixed effects, RE denotes random effects

S.376


Summary:

I We find evidence in favor of the predictions of the HOVmodel, implying evidence for the “Pollution Haven effect”.

I On the one hand, trade is harmful for the environment becauseit fosters growth (which c.p. requires more environment).

I On the other hand, the specialization pattern could eventuallyalso reduce pressure on the environment if pollution-intensiveindustries are loacted in countries with relative largerbiocapacity (per capita).

I The conclusion is not that countries should createcomparative advantages by lax environmental policies. Notethat lax policy is strongly correlated with low GDP per capita,and a sustainable development strategy cannot rely onoverexploiting of the natural environment.

S.377

international trade - makro.uni-jena.de · 6.4mobility of human capital: brain drain, brain gain...

Documents