Essays on the proposed monetary integration in the Southern African Development Community

by Mulatu F. Zerihun

A PhD Thesis Defense Presentation

Department of Economics

Faculty of Economics and Management SciencesUNIVERSITY OF PRETORIA

27 November, 2014

Supervisor: Prof. Marthinus C. BreitenbachCo-supervisor: Prof. Francis Kemegue

Broader Outline of PresentationFew Words of Praise Part I Introductory Chapters

Chapter 1Chapter 2Chapter 3

Part II Essays Essay I (Chapter 4)(Published AFJ,16(2), ERSA

WP No 319)Essay II (Chapter 5) (Under review JDSA, ERSA

WP No 600)Essay III( Chapter 6) (Accepted IJD, ERSA WP No

420) Essay IV(Chapter7) (Under review JIE, ERSA WP

N0 474)

Part III conclusions and Policy Implications

(Chapter 8)

2

Part IMotivations and contributions

The effect of MU on trade among members is statistically significant and economically important (Rose,2000)

EAC, ECOWAS, SADC bring about net-benefit (Debrun, et al. 2011)

counter - terrorism benefits in CFA Franc Zone (Elu & Price,2014)

In Africa, monetary integration has been in place since 1910 (Burgess, 2009).

The Abuja Treaty (1991) Regional economic integration, can have positive welfare effects for

the countries joining the union and leads to economic growth and poverty reduction (UNECA, 2010)

Association of African Central Bank Governors proposed the plan for the single currency to be materialized by 2023

Post EU Revival of OCA (Alesina et al., 2002; Masson and Pattillo, 2005, and Jefferis, 2007)

Gap in the literature -bias towards NNI, limited studies on SSI ( Fritz, et al , 2010)

3

One of the strongest justifications for regional integration in Africa

is greater economic independence and development. The root of Africa’s marginal role in the world economy is

intrinsically linked to the limited strides made in regional integration (UNECA, 2012).

Africa’s marginalised presence in the global economy (Adepoju, 2001)

Economies of most African countries remain detached from each other (ECA, 2010; Jovanovich, 2006)

With the advent of the African Union (AU), there have been intense renewed efforts by member states to give priority to policy coordination and economic integration, way to enhance their bargaining power via a common negotiating position in the era of globalization.

4

Motivations and contributions...

Motivations and contributions...

There are mounting doubts about the economic and political

feasibility of the monetary integration in SADC.

Previous studies - the results are mixed and inconclusive The focus of many studies have been on macroeconomic

convergence in the SADC region overlooking policy coordination aspects.

This study sheds light on the economic feasibility of the proposed

monetary union in the SADC region African Renaissance – Is monetary integration the right policy at the

right time (for SADC and AU?) To contribute empirical evidence towards the AU’s agenda of

economic integration on the continent5

Main contributions The thesis explores the potential for economic integration as well as

what is required to achieve SADC goals, which makes it relevant. A comprehensive empirical investigation of the degree of real

convergence towards the proposed monetary union in SADC using recent databases.

Application of methodologies never applied in the context of SADC before.

Scientific input for the stakeholders and policy makers in the SADC region to bring the proposed SADC monetary integration into existence.

Contribute to policy debate around Exchange rate Policy, Common Currency & OCA theory- in the broader context of SSI.

The findings from this thesis expand what we know about OCA’s in SSA. 6

Motivations and contributions...

Objectives

This study seeks to:

Establish empirically whether SADC countries can form the

proposed monetary union in the stipulated time frame and beyond;

Provide policy coordination measures to be addressed in an

attempt to move SADC economies towards successful monetary

integration;

Contribute to the literature on the economics of integration in the

SADC region.

7

Traditional Approach‘Ex-ante’

COSTS

Endogenity Approach “New”-‘Ex-post’

BENFITS

Lucas Critique

Rational Expectation

Recent Developments in Macroeconomic Theory

Trade Creation

Effectiveness & Credibility of Monetary

Policy

Factor Mobility(Mundel, 1961)

Openness to External Trade (McKinnon, 1963)

Other Economic Characteristics

SADC MONETARY INTEGRATION

Product diversification (Kenen, 1969)

Synchronisation of business Cycles

Effectiveness of Exchange Rate Adjustment

Institutional & Political Factors

Figure 1: Theoretical Framework of OCA Theory Source: Authors(based on literature review)

Co

ncep

tual F

ramew

ork

Trade Theory OCA THEORY

Open Economy Macroeconomics

Optimum Exchange Rate Regime

Fixed/Flexible Exchange Rate

Exchange Rate Based Policy Coordination

M e t h o d o l o g i e s

Linearity/NonlinearityStationary/Nonsationary

T

Cointegration Symmetry/PPP/GPPP

Qualitative assessments

Theoretical Reviews on Economic Integration

International trade theories and open economy macroeconomicshave shaped the evolution of OCA theory and of economic

integrationOCA is a region where factors of production are internally mobile

but internationally immobile- facilitate the intraregional redistribution

OCA is defined as the optimal geographic domain of a single currency, or of several currencies, whose exchange rates are irrevocably pegged and might be unified.

The modern theory of monetary integration- An alternative to OCA theory (Waltraud.S,2001)

The theory of exchange rate based policy coordination (Essay#4) (HughesHallett/Holtham/Hutson,1989;Bryant,1995:52-55)

The theory of policy coordination under uncertainty (for future study) (Canzoneri/Henderson,1991;Ghosh/Masson,1994).

9

Lucas Critique in OCA Analysis ( studies from EMU)

Endogenous OCA (Quirici, 2005) (Kapounek and Poměnková, 2012).

Rationality of Expectations and OCA AnalysisMinimize macroeconomic volatility in the presence of

asymmetric shocks –develops forward-looking behaviour among economic agents in the formation of monetary union (Torój, 2011)

10

Theoretical Reviews on Economic Integration...

OCA Theory and South-South Integration (SSI)

OCA theory is silent about SSI

SSI needs to deal with the specific monetary constraints of the

member countries

Like accumulation of debt in foreign currency, thereby most often

suffering from a restricted lender of last resort function, balance sheet

effects in the event of a currency devaluation and original sin and, as a

result, small and undiversified financial markets.

A successful SSI and may provide further perspectives for regional

monetary integration and financial market development (Fritz and

Mühlich, 2010)

A regional monetary arrangement potentially generates economies of

scale in regional financial markets. 11

Theoretical Reviews on Economic Integration...

Experiences of Economic Integration in Africa: The Case of SADC Creation of a pan-African common market has been a central

vision of African leaders (a single currency by 2023, African Central Bank to follow by 2028)

However, there is so much pessimism about economic integration in Africa (Masson and Pattillo, 2005) (Debrun, Masson and Patillo, 2010)

‘Spaghetti bowl’ (Geda & Kibret, 2008) Challenges: the EU-Africa Economic Partnership Agreements However, there are new developments;

NEPAD focus on trade facilitation through sub-regional preferential arrangements, should be followed establishing common currency areas and harmonization of macroeconomic policy (UNECA, 2010).

Recent paradigm shift- The Tripartite Free Trade Area (since 2012)

12

SADC Economic Integration: Achievements and Challenges

SADC Integration Milestones Free Trade Area (2008)

Since 2000 Intra-SADC trade has more than doubled Customs Union (2010?) Common Market (2015) Monetary Union (2016)

Real Time Gross Settlement System developed to modernise cross-border payment settlements = implemented in 12 SADC Member States

Implementation of best banking practices, norms and standards – 14 Member States - harmonise banking supervision processes.

Institutional administrative and legal framework – SADC Ministers of Finance developed and approved the Model Central Bank Law, designed to set harmonised frameworks for control of exchange policies, bank procedures and systems.

13

SADC economies have been affected by recent global uncertainty

and financial turmoil

Most of the member states have achieved those Maastricht type

convergence criteria (see Table 1 next slide)

However, simply looking at the differences between the main

macroeconomic indicators does not bring much insight into the

sources of divergence or the adequacy of a common policy

response (Kamar and Naceur, 2007).

14

SADC Economic Integration: Achievements and Challenges...

Table 1: Convergence Criteria of SADC (2012) Country

Inflation rate (5%)

Budget deficit (3% of GDP as anchor,

with a range of 1%)

Governmentdebt (<60%

of GDP)

Current account

balance (<9%of GDP)

Foreign Exchange

Reserve (> 6 months’ of

import cover )

Angola 14.0 -7.7 32.3* -3.8 13Botswana 8.1 -11.0 17.9** -5.11 31DRC 46.2 -1.6 - -13.13 < 1Lesotho 7.7 -6.8 - -1.5 5Madagascar 8.9 -2.3 33.3* -16.76 5.1Malawi 8.4 -5.4 72.5* -7.87 < 2Mauritius 2.5 3.4 60.0* 8.15 4.8

Mozambique 3.2 5.7 26.1** -11.91 6.9Namibia 9.1 -6.0 19.3* -2.19 4Seychelles 31.8 +2.6 43.9** -23.14 <2South Africa 7.1 -1.2 35.7** -3.98 6Swaziland 7.6 -3.5 13.0* -6.25 4Tanzania 12.1 -4.7 24.8** -9.42 6Zambia 13.4 -4.4 31.5** -3.29 3.6Zimbabwe a n.a. 304.3** -30.10 <1

• Challenges: Weak productive capacity, Low value added in the production chain Unsupportive business environment and Cumbersome regulatory frameworks in many member

countriesThese calls for policy coordination To address “negative” macroeconomic divergence (Priewe,

2007) Policy coordination plays a major role in the formulation of

currency union Regional monetary cooperation is a matter of policy

consideration (Gao , 2007) 16

SADC Economic Integration: Achievements and Challenges...

Part II Essays - Essay I (Chapter 4)

TESTING FOR STRUCTURAL SYMMETRY TOWARDS SADC MONETARY INTEGRATION

Research Question Do SADC countries adhere to structural symmetry towards

monetary integration? Methodology Triples Test (TT) (Randles et al.,1980,Verbugge,1997 Razzak,2001) More efficient than many other methods used in the literature

(asymptotically distribution free)* to detect symmetry Mathematical details of TT see (Razzak, 2001:232-35 and Breitenbach et

al, 2012). Requires decomposing the series into trend and cyclical components We used B-K filter (Inklaar et al., 2008 and Tapsoba, 2009). Data - Real GDP of 68 countries (from 1970 – 2010)

17

Chapter 4…

Bilateral co-movement and RICM

18

)()(

)()(

yx

xyxy

SS

Sr

)()(

)()(

2

2

yx

xy

XY SS

Sr

Here we applied a simple regression to each pair of countries and extract the R2 of the regression as a measure of the strength of the connection between the cyclical components of real GDP between the pair of countries in the region and among major trade partners.

R

R 2

2

2

2

nonSADCyXY

SADCyXY

r

r

nonSADC

SADCRICM

RICM-Computed using SADC countries’ ratio of relative intensity of co-movements (RICM) versus each other and that of major trade partners

RESULTSTable 1: SADC Countries Triples Test Statistics for Symmetry(1970-2010)

Country η* Ksi1 Ksi2 Ksi3 U-stat P-value(α=0.05)

Angola 0.0285 0.0045 0.0230 0.1103 0.8328 0.1977

Lesotho 0.0209 0.0070 0.0266 0.1107 0.5013 0.2912

Botswana 0.0330 0.0054 0.0200 0.1100 0.9087 0.1711

Madagascar -0.0473 0.0045 0.0188 0.1089 -1.406 0.0749

Malawi -0.0405 0.0047 0.0199 0.1095 -1.175 0.0401

Mauritius 0.0146 0.0077 0.0238 0.1109 0.3407 0.3632

Mozambique -0.0595 0.0045 0.0187 0.1076 -1.772 0.0314

Namibia 0.0344 0.0037 0.0158 0.1099 1.1213 0.1251

Seychelles 0.0025 0.0040 0.0131 0.1111 0.0800 0.4801

South Africa -0.0518 0.0017 0.0129 0.1084 -2.334 0.0094

Swaziland 0.0972 0.0061 0.0228 0.1017 2.5073 0.0054

Tanzania -0.0389 0.0037 0.0163 0.1096 -1.271 0.0885

Zambia -0.0422 0.0060 0.0158 0.1093 -1.131 0.1251

Zimbabwe 0.0055 0.0028 0.0127 0.1111 -0.203 0.4013

DRC -0.0556 0.0026 0.0132 0.1080 -2.129 0.0154

19

Note:H0: the univariate time series is symmetricThe figures in bold show significant p-values at 5 % level of significance and hence the null hypotheses is rejected which implies asymmetry in the

Table 2: Coefficients of determination as a measure of cyclical co-movements

Row name AGO LSO BWA MDG MWI MRT SYC MOZ NAM ZAF SWZ TZA ZMB ZWE DRC

AGO 1

LSO0.00

21

BWA0.04

8 0.0011

MDG0.00

0 0.000 0.0011

MWI0.05

8 0.002 0.005 0.0401

MRS0.01

0 0.082 0.001 0.046 0.0071

SYC0.00

5 0.214 0.002 0.028 0.001 0.0281

MOZ0.03

8 0.024 0.019 0.005 0.008 0.059 0.0311

NAM0.04

4 0.059 0.034 0.000 0.004 0.045 0.004 0.0031

RSA0.22

8 0.019 0.030 0.024 0.001 0.111 0.055 0.218 0.0371

SWZ0.02

1 0.101 0.036 0.039 0.005 0.119 0.016 0.049 0.104 0.1351

TZA0.02

9 0.052 0.023 0.000 0.110 0.021 0.000 0.117 0.078 0.028 0.0251

ZMB0.00

1 0.003 0.012 0.002 0.064 0.032 0.023 0.146 0.000 0.122 0.047 0.1161

ZWE0.07

3 0.049 0.007 0.008 0.033 0.008 0.063 0.261 0.008 0.110 0.024 0.011 0.0631

DRC0.26

6 0.001 0.037 0.000 0.001 0.002 0.020 0.058 0.046 0.189 0.009 0.002 0.004 0.0711

Row name Total all SADC all SADC relative intensity

AGO 2.922757 0.820042 0.389992LSO 1.326669 0.534191 0.674077BWA 1.881879 0.254965 0.156717MDG 1.498395 0.148454 0.109971MWI 0.712644 0.332029 0.872347MRS 0.58334 0.381326 1.887621SYC 1.33429 1.101481 4.731282MOZ 1.194135 0.433222 0.569344NAM 0.857648 0.466165 1.190766RSA 3.50578 1.197229 0.518606SWZ 0.978067 0.611418 1.667587TZA 1.068255 0.630444 1.43999ZMB 0.901105 0.608246 2.076927ZWE 1.319048 0.965 2.725618

DRC 1.410106 0.706682 1.0046321

Table 3: Comparison of SADC countries’ ratio of relative intensity of co-movements (RICM) and that of major trade partners

Chapter 4...Findings

Considering T T 10 out of 15 SADC countries exhibit symmetry in the distribution of their business cycles (real log GDP data) .

Evidence from bilateral co-movements using the coefficients of determination; the result shows a general lack of strong co-movements in the business cycles of SADC countries.

Mozambique, Zimbabwe, Zambia, Seychelles, Tanzania, Mauritius, and DRC, however, have relatively strong intensities in the co-movement of their business cycles when compared with the rest of SADC members.

These findings confirm that there is no sufficient structural symmetry among member states for the entry into the monetary union in 2018.

This implies that a common monetary policy will not be optimal for all countries in the region, at least in the short run.

Call for policy harmonisation and coordination for business cycles to become more synchronised as a precondition to monetary union, otherwise, SADC countries will face the same risk as the PIGGS

22

Assessment of Monetary Union in SADC: Evidence from Cointegration and Panel Unit Root Tests

Question: Do SADC countries form an optimum currency area?

Definitions• PPP-the ratio of two countries’ price levels, expressed in a common currency,

should be equal to unity, for all pairs of countries and at all times. GPPP- postulates that(in a multi-country setting) the fundamentals that drive

the real exchange rates will exhibit common stochastic trends. GPPP-can test for a bilateral or multilateral co-integrating relationship between

the countries in question vis-à-vis a base currency of choice

Methodology

Data Covers a sample of 11 SADC member countries 4 countries excluded-given

data limitation. Monthly data:1995:01 to 2012:08 Variables - CPI (based on 2005=100) and nominal exchange rates relative to

the US dollar

23

Part II Essays - Essay II (Chapter 5)

(Enders & Hurn, 1994; Wilson & Choy, 2007; Beirne, 2008):

GPPP holds when at least one linear combination of bilateral RERs is observed (Ogawa & Kawasaki, 2003; Beirne, 2008)

To test whether the set of countries form an OCA, following Beirne (2008) the VAR (k) is set up in the following matrix notation:

In line with the Vector Error Correction Model (VECM), we can rewrite the above as follows (in first-difference form):

24

Part II Essays - Essay II (Chapter 5)…

tmtmtttttt rrrr 111414131312

,,00 tGPPPtjrjtir

)IN(0,~t

t

k-t

z k

A 1-t

z1

Atz

t

1k-t

z k

1-t

z1

ktztz

25

Part II Essays - Essay II (Chapter 5) - Cointegration Test Results

Hypothesized

No. of CE(s)

Eigenvalue

Trace 1 % Max-Eigen 1 %Statistic Critical

ValueStatistic Critical

Value

None***  0.418328  395.8688  293.44  112.1628  75.95

r≤ 1***  0.268538  283.7060  247.18  64.73099  69.09

r≤ 2***  0.248888  218.9750  204.95  59.24356  62.80

r≤ 3  0.190979  159.7314  168.36  43.86950  57.69

r≤ 4  0.152646  115.8619  133.57  34.28672  51.57

r≤ 5  0.125896  81.57520  103.18  27.85303  45.10

r≤ 6  0.090721  53.72216  76.07  19.68641  38.77

r≤ 7  0.070556  34.03575  54.46  15.14602  32.24

r≤ 8  0.047670  18.88973  35.65  10.11057  25.52

r≤ 9  0.030071  8.779158  20.04  6.320142  18.63

r≤ 10  0.011809  2.459016   6.65  2.459016   6.65

Note: *** denotes rejection of the null

hypothesis.Trace test indicates 3 cointegrating equation(s) at the 1% level.

Max-Eigen test indicates 1 cointegrating equation(s) at the 1% level.

Table 5 Pedroni Residual Cointegration Test Result

26

Series: lnRER lnNER lnCPI 

Sample: 1995M03 2012M10, Included observations: 233Cross-sections included: 11Null Hypothesis: No cointegrationTrend assumption: Deterministic intercept and trendNewey-West automatic bandwidth selection and Bartlett kernel

(within-dimension)Statistic Prob. Weighted

Statistic Prob.

Panel v-Statistic  12.73971  0.0000  10.60199 0.000

0Panel rho-Statistic -3.523916  0.0002 -3.389909

 0.000

3Panel PP-Statistic -2.947256  0.0016 -2.840992

 0.002

2Panel ADF-Statistic -7.917885  0.0000 -7.861144

 0.000

0(between-dimension)Group rho-Statistic -2.129851  0.0166Group PP-Statistic -2.305456  0.0106Group ADF-Statistic -8.001912  0.0000

Note: The null hypothesis of ‘no cointegration’ is rejected in all the seven statistics as shown above.

Table 6 (Log) Real Exchange Rate Series Short-run Adjustment Coefficients (α)

27

RER series Adjustment coefficients (α)

Standard Error

D(South African rand)

0.016662 0.03318

D(Angolan new kwanza)

-0.941522 0.12886

D(Botswana pula) 0.037468  0.02376

D(Malawian kwacha)

0.086315 0.03880

D(Malagasy ariary ) 0.107311 0.02685

D(Mauritian rupee) -0.024887 0.01444

D(Mozambique metical)

0.042243 0.02391

D(Seychelles rupee) -0.016641 0.05774

D(Swazi lilangeni)  0.016270 0.03557

D(Tanzanian shilling)

0.022810 0.01770

D(Zambian kwacha) -0.008818 0.03676

Table 7: SADC (Log) Real Exchange Rate Normalised long run Cointegrating Equations (β- Coefficients)

28

RER series(1 Cointegrating Equation:  Log likelihood 4852.518 )

Normalised Long run Cointegrating Eqns (β- Coefficients)

Standard Error

South African rand 1.000

Angolan new kwanza 0.160 0.031

Botswana pula -0.818 0.345

Malawian kwacha -0.137 0.075

Malagasy ariary (MGA) -0.558 0.154

Mauritian rupee 1.616 0.252

Mozambique metical -0.387 0.131

Seychelles rupee -0.083 0.135

Swazi lilangeni -0.779 0.273

Tanzanian shilling -0.445 0.177

Zambian kwacha -0.156 0.087

Panel Unit Root Test Results Table 8: LAG Selection Criteria

29

Selection-order criteria Sample:1995:5-2012:10 Number of observation=208 Lag LL LR df p FPE AIC HQIC

SBIC

0 -257.16 0.02449 2.50154 2.5212.55

1 1134.5 2783. 9 0.00 4.1E-09 -10.79 -10.71-10.60

2 1186.1 103.1 9 0.00 2.7E-09 -11.20 -11.07-10.86*

3 1205.9 39.56 9 0.00 2.5E-09 -11.31 -11.11*-10.82

4 1219.2 26.72* 9 0.002 2.4E-09* -11.35* -11.09-10.72

Endogenous: ln_ realexrteExogenous: -cons

Panel Unit Root Test Results-LLC & IPS

The LLC panel unit root test rejects the null hypothesis of a unit root at the 1% level of significance when a time trend is included in the estimation .

The LLC panel unit root test also rejects the null hypothesis of a unit root in the panel at a 10% level of significance when a time trend is not included.

The IPS unit root test result shows that the panel of (log) RER series is stationary at 1% level of significance only when the time trend is included in the analysis; otherwise it is unit root .

The Pesaran’s CADF test also rejects the null hypothesis of ‘all series are non-stationary’ at 5 % level of significance, supporting the results obtained by the LLC and IPS panel unit root tests.

ECM based co-integration tests for the panel of SADC (log) RER series- rejects the null hypothesis of no co-integration for the panel as the whole at the 1% level of significance. However, with the group mean statistics we fail to reject the null hypothesis.

30

Findings and Implications (Essay#2)

GPPP holds for SADC implies that SADC is potential OCAThere are cointegrating relationships among SADC real

exchange rate seriesSlow speed of adjustment for (log) real exchange rate of

SADC member states might constrain the effectiveness of stabilization policies in the wake of external shocks, rendering SADC countries vulnerable to macroeconomic instability in the region.

However, the magnitude of the long run beta coefficients of all the real exchange rates are below one except in the case of Mauritius and they all bear negative sign except in the case of Angola and Mauritius.

We can take this evidence as supportive of monetary union in the region excluding Angola and Mauritius.

31

The investigation of non-linearities and asymmetries in macroeconomic behaviour constitutes an increasingly popular area of empirical research (Holmes, 2004)

Two Non-linear Approaches Structuralist-Snapshot-SR approach (SR stability around asymmetric

shocks) BDS TestFlexible Fourier stationary test - to test the long-run PPP among SADC

economiesThere have been methodological problems in attempts to test the validity

of PPP hypothesis: i) Pre-selection bias (Maddala and Kim, 1998)- because unknown number

and location of breaks in the time series dataCurrent available tests account only for one to two breaks. The use of dummies suggests sharp and sudden changes in the trend or

level.32

Part II Essays - Essay III (Chapter 6)-Non-Linear Approaches in Testing PPP of SADC Economies Towards MU

DGP (Pascalau, 2010; Enders & Lee ;2004, 2009):

)2(cos)2sin(y

1

210t

ttt

tt

u

TktTkt

)2(cos)2sin(y 210t tvTktTkt

stationary is series thei.eroot unit no is thererejected ,0:

breaksunknown andlinearity -non is Thererejected ,0:

0

210

ifH

ifH

Where is the number of frequencies of the Fourier function, ‘t’ is a trend term, ‘T’ is the number of usable observations in the regression analysis. And the captures structural change in the series )2(cos)2sin( 21 TktTkt

)51( kk

Using F-statistics F(k)& the statistics from Table 3 in Enders and Lee (2004, 2009); Nyong (2013)

DF

Cont’d...

Table 9: The BDS test results for the monthly RER series (Fraction of pairs

34

m

BDS Statistics of SADC Real Exchange Rates – country codes as local currency/USDHo = iid , is rejected in all cases[reject the hypothesis that the series is iid if the reported probability is small]

Country Codes

ANG BWA MDG MWI MUS MOZ ZAF SYC SWZ TZA ZMB

 2 0.169

 0.1796

 0.1650

 0.1779

 0.181

 0.175

 0.176

 0.181

 0.175

 0.174

 0.181

 3 0.2857

 0.3015

 0.2745

 0.2957

 0.303

 0.290

 0.295

 0.302

 0.294

 0.291

 0.304

 4 0.3610

 0.3827

 0.3446

 0.3711

 0.385

 0.364

 0.373

 0.381

 0.375

 0.366

 0.387

 5 0.4114

 0.4342

 0.3869

 0.4177

 0.437

 0.408

 0.423

 0.431

 0.425

 0.411

 0.440

 6 0.443

 0.465

 0.4103

 0.4444

 0.470

 0.432

 0.452

 0.461

 0.456

 0.435

 0.475

c1,n(ε

) 0.7701

 0.208

 0.243

 0.298

 0.181

 0.217

 0.262

 0.240

 0.293

 0.179

 0.562

k1(ε) 3824810

3821906

 3831102

 3823116

 3817066

 3814646

 3830134

 3825536

 3821180

 3818034

 3806902

V-Statistic

0.7033

 0.7027

0.7045

 0.7030

0.7019

 0.701

0.7042

0.7034

 0.702

 0.702

 0.700

P***<0.001, m=embedding dimension, Included observations (n): 2332

Table 10: Stationary Test with a Nonlinear Fourier Unit Root Test SADC

CountriesObs.(T)

SSRs Frequency

Angola 200 1.295 1 220.18*** -7.033***

Botswana 203 0.146 1 446.14*** -2.731***

Madagascar 210 0.227 1 418.91*** -3.841***

Malawi 199 0.311 1 360.60*** -2.731***

Mauritius 204 0.052 1 554.05*** -1.298***

Mozambique 199 0.131 1 446.58*** -3.529***

South Africa

203 0.248 1 392.53*** -3.004***

Seychelles 200 0.195 1 409.52*** -2.461***

Swaziland 203 0.311 1 369.82*** -2.653***

Tanzania 199 0.058 1 525.94*** -3.085***

Zambia 203 0.349 1 357.99*** -0.410***

^

k )(^

kF )(^

k

This condition implies that both sine and cosine terms in the difference equation should be included in the estimates model. Thus, the hypothesis is rejected using F-statistics F (k) of Table 3 in Enders and Lee (2004).

021

Findings Evidence favouring the long-run validity of PPP for the

SADC member countries included in this study. Logarithm of real exchange rate is nonlinear stationary ,

implying that deviations of real exchange rate is mean reverting towards the PPP equilibrium.

These findings are in line with the findings by Chi-Wei Su et al., 2013, which strongly support PPP equilibrium for 19 African countries.

Trade barriers, as well as interventions in the exchange markets, could be behind this nonlinear behaviour.

Using Fourier estimation method our findings in this paper support for the OCA in the region comprising the 11 countries included in the study.

As suggested by Chang (2009), these 11 SADC countries can use PPP to predict exchange rate that determine whether a currency is over or undervalued and experiencing difference between domestic and foreign inflation .

Objective

…continued

• To identify and estimate the long-run determinants of

the RER behavior in the SADC

• To assess the degree of coordination among major

policy instruments of SADC countries and whether they

have identical effects on the REER in the long run

Research Questions Is there sufficient policy coordination to induce the

proposed monetary union in 2018 among SADC economies?

Part II Essays - Essay IV (Chapter 7)EXPLORING EXCHANGE RATE BASED POLICY COORDINATION IN SADC TOWARDS MONETARY UNION

Why Policy Coordination?

To address “negative” macroeconomic divergence which can hamper growth and threats the long run convergence among member economies (Priewe, 2007)

Regional monetary cooperation is a matter of policy consideration (Gao , 2007)

Real Exchange Rate as a Policy Variable Real exchange has been witnessed as a relevant policy tool for

development (Eichengreen, 2007). RER behavior is subject to the influence of many variables: such as

monetary policy, government expenditure, terms of trade, degree of openness, and capital flows.

The exchange rate is at the heart of economic activity as it affects and is affected by all other policies, making policy coordination and harmonization essential for the success of a common currency (Kamar & Naceur,2007)

38

39

Pedroni Residual Cointegration TestSeries: LNREAL_RTE LNRESY LNOPEN LNM2 LNGEXP Sample: 1995 : 2012Included observations: 216Cross-sections included: 12Null Hypothesis: No cointegrationTrend assumption: No deterministic trendAutomatic lag length selection based on SIC with a max lag of 2Newey-West automatic bandwidth selection and Bartlett kernelAlternative hypothesis: common AR coefs. (within-dimension)

WeightedStatistic Prob. Statistic Prob.

Panel v-Statistic -0.689972  0.7549 -1.350298  0.9115Panel rho-Statistic  1.117448  0.8681  1.332275  0.9086Panel PP-Statistic -2.964929  0.0015 -2.095010  0.0181Panel ADF-Statistic -3.759298  0.0001 -2.535645  0.0056

Alternative hypothesis: individual AR coefs. (between-dimension)Statistic Prob.

Group rho-Statistic  2.581929  0.9951Group PP-Statistic -2.765844  0.0028Group ADF-Statistic -3.445359  0.0003

Methodology: Mean Group (MG) and Pooled Mean Group (PMG)

Main Equation:

ARDL (p, q,...q) dynamic specification of the form:

Error correction model (ECM) form:

Note: The parameter Phi is the error correcting speed of adjustment term. If

Phi =0, then no long run relationship is expected to take place. This parameter is expected to be significantly negative under the hypothesis that the variables show a return to long-term equilibrium.

Of particular importance is-θ- the vector which contains the long-run relationship between the dependent and explanatory variables.

lnRESY9NKF8 NFA7lnTOT6lnOPEN5lnLIQ4CAB3lnBUDG2 lnGEXP1 i0,REERln

Y0

,'

0

1

,it iti

q

j

jtij

p

j

jtiij XY

)(Y ,

1

0

'1,

1

1

'1,it itijti

q

j

ijti

p

j

ijititii XYXY

1-q1,2,...,j - and

1-p1,2,...,j - ,1

),1(

p

1jm im

p

1jm im01

ij

ikq

jk ik

iji

p

j ijiwhere

Econometric Results: PMG Estimated LR&SR Parameters

41

POOL MEAN GROUPPanel Variable (i): id Number of obs = 192Time Variable (t): time Number of groups = 12 Obs per group: min = 16 avg = 16.0 max = 16 Log Likelihood = 258.3874------------------------------------------------------------------------------D.lnreal_rte | Coef. Std. Err. z P>|z| [95% Conf. Interval]-------------+----------------------------------------------------------------ec | lnopen | -.395469 .0972017 -4.07 0.000 -.5859808 -.2049573 lngexp | .178408 .1631695 1.09 0.274 -.1413984 .4982143 lnm2 | -1.189833 .2334503 -5.10 0.000 -1.647387 -.732279 lnresy | .1918429 .0750204 2.56 0.011 .0448056 .3388803-------------+----------------------------------------------------------------SR | ec | -.2484906 .0549622 -4.52 0.000 -.3562145 -.1407666 | lnopen | D1. | .2594346 .1299576 2.00 0.046 .0047224 .5141468 | lngexp | D1. | .2563283 .1006283 2.55 0.011 .0591004 .4535562 | lnm2 | D1. | -.3083947 .2354526 -1.31 0.190 -.7698734 .1530839 | lnresy | D1. | .1327814 .0485677 2.73 0.006 .0375906 .2279723 | _cons | 1.870766 .4746648 3.94 0.000 .9404404 2.801092

SADC RER misalignment

42

1995 1996 1997 1998 1999 2000 2001 2002 2003 2004 2005 2006 2007 2008 2009 2010 2011 2012

-6

-4

-2

0

2

4

6

8

10

RER Misalignment Using Heterogenous In-tercepts

BWA LSO MDG MWI MUS MOZNAM SYC ZAF SWZ TZA ZMB

Mag

nit

ud

e o

f R

ER

mis

alig

nm

en

t in

p

erc

en

t (%

)

There is persistent RER misalignment in region, however, its deviation is not as such wider given developing countries standard. This is promising result for further policy coordination in the region

Findings - Results from PMG estimation confirm that

all the major macroeconomic policy variables can be used to achieve policy coordination in the region Trade policy has statistically significant impact in both short run & long runFiscal policy has significant impact only in the short runMonetary policy has significant impact only in the long runStock of reserves at year end/GDP (RESY) has statistically significant impact

in both short run & long run speed of adjustment (EC=-0.25) is low implying inability of SADC

economies to adjust adequately shocks. There is persistent RER misalignment in region, however, its

deviation is not as such wider given developing countries standard.

This is promising result for further policy coordination in the region. As the findings in the paper confirms persistent misalignment in

RER is an indicative for lower financial deepening and potential tendency to currency crisis in the SADC region.

43

Part III Conclusions and Policy Implications

Conclusions Findings confirm that there is no sufficient structural symmetry and

macroeconomic policy coordination among SADC member states for the entry into the monetary union in 2018.

However, according to the findings in this study 9 SADC countries can potentially constitute SADC-OCA namely; Botswana ,Madagascar, Malawi, Mozambique, Seychelles, South Africa, Swaziland, Tanzania ,Zambia. (Lesotho, DRC, and Zimbabwe are not included due to data limitations )

Angola and Mauritius-disqualify from a SADC-OCA at least for the sample period considered in this study.

To reap benefits SADC economic integration initiatives, it requires realistic time span, political will, common understandings and awareness, commitment and self-disciplined policy actions from member states and their fellow citizens.

SADC Will it be easy? No. Is it achievable? Yes! 44

There should be strong policy coordination and public awareness to increase the benefits and decrease costs associated with the proposed monetary union in the region.

These measures are the prerequisites before the prime move towards monetary integration in the region to avoid profound consequences as faced by the EMU after the EMU financial crises.

Suggestions for Further Research Identifying the stabilizing and destabilizing elements of regional

integration in SADC suggesting policy options remain valid Inter-temporal aspects of fiscal and monetary policy coordination

using dynamic game theory articulated to deal with credibility issues among SADC member countries

Investigate the relevance optimisation based dynamic stochastic general equilibrium (DSGE) models for policy analysis; that is, to suggest econometric policy advice for SADC countries to join the proposed monetary union. 45

Policy implications

Q&A, discussions, comments & suggestions welcome

AcknowledgementsProf. M.C Breitenbach and Prof. F. Kemegue Prof. Steve Koch, HOD, Dept. Economics, UPThe three anonymous External ExaminersBrown Bag PhD Seminar organisers & participants All the staff at the Department of Economics, UP My Colleagues (Christian Tipoy, Lardo , Laban, Nono)Prof . Muchie M., Dr Mashigo, Dr Aye, Dr Wu Suck LeeCEEPA,UP, NRF, ERSA, DAAD, African Econometrics Society, University of Ghana, for various fundingWB/WDI, IMF/IFS, SADC CBs for data supplyEtenesh T. Halignew, my wife, and my children…

“I say , to his seed, and hast performed thy words; for thou art righteous” NEHEMIAH

9:8

I thank you for your attention! M.F. Zerihun, UP, 27/11/14, PTA