International Trade and Income in Malawi: A Co-integration and Causality Approach

Greenwell Matchaya, Pius Chilonda, Sibusiso Nhelengethwa

ReSAKSS/IWMI

Outline

• Introduction• Objectives • Methodology• Results• Conclusions

Introduction

• Exports, Imports and incomes are related

• It is important to understand the nature of the relationship for purposes of policy

• Many factors affect each of these including fiscal programs, trade policy, exchange policy etc

Objectives • To investigate causal relationships between exports, imports,

and economic growth in Malawi over the period 1961-2010.

• Draw some recommendations for policy based on the findings

Methodology• These relationships are examined using the Johansen frameworks for co-

integration • The Vector Error Correction (VECM) framework is further used to provide

estimates for both short-run and long-run dynamics in the series under study.

• Impulse responses are generated to foresee the future dynamics

• Data:

Descriptives

Descriptives Single unit root (i.e. on levels) Second unit root (i.e. on

changes) Series DF ADF PP DFGLS DF ADF PP Y -1.42 -2.54 -1.61 -1.57 -8.28*** -5.71*** -50.53*** X -1.71 -1.94 -1.66 -1.28 -6.89*** -5.12*** -42.98*** M -1.03 -2.74 -1.27 -1.82 -6.27*** -4.82*** -41.82***

Results-residual based co-integration

Series DF PP

Residuals for logY -3.54*** -19.76***

Residuals for logX-2.90* -14.27***

Residuals for logM-3.84*** -24.39***

Results –Granger CausalityEquation Excluded Chi2 df prob >

chi2DlogY DlogX 6.2** 2 0.05DlogY DlogM 2.3 2 0.30DlogY ALL 11.4** 4 0.02DlogX DlogY 0.89 2 0.63DlogX DlogM 7.1** 2 0.03DlogX ALL 12.5*** 4 0.01DlogM DlogY 5.5* 2 0.06DlogM DlogX 1.5 2 0.46DlogM ALL 8.3* 4 0.08AutocorrelationTest

Lagrange Multiplier

Ch2=6.87Prob>chi2

=0.65 Residual normality test

Kurtosis Chi2=3.52Prob >Chi2

=0.32VAR-stability All Eigen-values

lie in unit circle

Co-integration –results

Beta Coef. Std. Err.

_ce1

logY 1 .

logM (omitted)

logX 0.55* 0.31

trend -0.03** .01

_cons -6.13

_cons 0.69

Results- co-integration

• GDP and exports are co-integrated in the long run and that the relationship runs from exports to the GDP – export-led economic growth.

Hypothesized co-integratingH0

Number of relationshipsH1

Test statistic Critical values (95 %)

Max. eigenvalue

Trace Max. eigenvalue

Trace

r=0 r>0 26.6** 22.51** 20.30 29.68

r<=1 r>1 14.3** 18.26** 13.23 15.41

r<=2 r=3 3.00 1.17 3.5 3.76

Causal directions

Imports

GDP

exports

Impulse responses •

-.1

0

.1

.2

0 2 4 6 8step

irf

irfstore: logEXPORTS -> logrealGDP

0

.05

.1

.15

.2

0 2 4 6 8step

irf

irfstore: logIMPORTS -> logrealGDP

0

.5

1

1.5

0 2 4 6 8step

irf

irfstore: logrealGDP -> logEXPORTS

0

.5

1

1.5

0 2 4 6 8step

irf

irfstore: logrealGDP -> logIMPORTS

Conclusions • The general qualitative story• Exports are good for economic growth• No evidence of import led growth• Policies that can encourage ag exports are desired

– These could include :– increase in agricultural productivity– Value addition – Diversification of production, – flexible but stable exchange rate – Reduction of Tariff and non Tariff trade barrier– Reduction of transaction costs etc