OECD GDP, exchange rate and exports performance of India in the post reform period: an empirical analysis.

Suresh, K.G. ; Reddy, V. Nagi

Abstract

This study uses exchange rate and OECD GDP to explain the exports of India for the period 1994 to 2008. Since the ADF and Phillips-Perron tests show the first order cointegration in Indian exports, OECD GDP and exchange rate, the first difference of the log of exports of India has been regressed on first differences of the log of OECD GDP and the log of one of the four exchange rate variables at a time namely the Trade weighted Real effective exchange rate(REERT), Exports weighted Real Effective exchange rate(REERX), Trade weighted Nominal Effective exchange rate(NEERT) and Export weighted nominal effective exchange rate(NEERX). The empirical results indicate that only OECD GDP is significant in explaining the Indian exports.

I. INTRODUCTION

The neo-liberal economic policies implemented in India in the mid 1980s and 1990s tried to make India more integrated with the rest of the world. The integration happened through the reduction of barriers to trade and capital flows, and the reforms in the foreign exchange market. In the external sector, the government has either reduced or abolished the quotas and tariffs on trade flows. In the foreign exchange sector, with the reforms in the early 1990's India has shifted to the floating exchange rate system from the previous fixed rate system. In 1991 the government has again devalued the rupee and in 1992 the dual exchange rate system was implemented. This is followed by the floating exchange rate system in 1993-94.

The above mentioned reforms have given boost to the exports sector of India. This is evident from the fact that the exports sector has registered an impressive growth rate during the post reform period. As observed by Mookerjee (1997), the reduction in trade barriers as a consequences of the GATT and WTO negotiations have made the foreign market specifically the OECD region more accessible to the Indian products. According to Veeramani (2007) the average annual growth rate of merchandise exports during 1980 to 1985 was 2.39% as against 17.76% during 1986 to 1990. The % growth rates for the period 1993-97, 1999-2001 and 2002-2005 are 13.30, 10.28 and 25.29 respectively. According to the Economic surveys 2008-09 and 2009-10, the annum exports growth rates for the periods 2005-06, 2006-07, 2007-08 and 2008-09 are 23.6%, 22.6%, 29% and 13% respectively. Even though the extent of integration with the rest of the world provides ample scope for exports growth (report from Center for Development Studies, Thiruvananthapuram, Kerala, India on the Impact of global crisis on Kerala economy (2008)), "the extent of integration with the rest of the world is the potential vulnerability" in times of a crisis in global economy. With the background of the present economic crisis in the developed world especially in the OECD region, it is important to analyze the influence of the OECD region's income on the exports of India. In this context the present paper examines the relationship between the exports of India and the factors explaining the exports namely exchange rate and OECD GDP during 1994 to 2008.

II. REVIEW OF LITERATURE

In the theoretical literature, elasticity approach postulates that the devaluation or depreciation of the domestic currency will increase the exports or decrease the imports of the domestic country by decreasing the price of exports or increasing the price of imports. Reciprocally in response to an appreciation of domestic currency, the exports will fall and imports will increase. With this theoretical background, the IMF has proposed the devaluation of the domestic currency as a policy measure to correct the balance of payments problems of developing countries including India. Existing literature also depicts the presence of a lagged impact of devaluation or depreciation on trade balance, popularly known as the J curve effect.

Some of the empirical studies in the Indian context provide evidence for the presence of J curve (Bahmani-Oskooee 1985). But studies by Buluswar, Thomson and Upadhaya (1996), Himarios (1985) and Singh (2002) find no evidence of a J curve. In fact Buluswar, Thomson and Upadhaya (1996) found the presence of an "M" curve in the Indian context. Interestingly Bahmani-Oskoose (2001) found the presence of an inverse "J" curve in the Indian perspective. Arora et al. (2003) observe that the depreciation of the rupee increases the exports of India to its main trading partners, even if there is no evidence of a bilateral "J" curve.

Mukerjee (1997) observed that India's exports is sensitive to exchange rate (current value), OECD GDP and world GDP during 1960-1992 and states that there is no long run relation (cointegration) between exchange rate and exports of India. This is confirmed by Buluswar, Thomson and Upadhaya (1996). Sarkar (1994) observes the same phenomena for the period 1971 to 1991. In response to the Sarkar's (1994), Nag and Upadhaya (1994) stated that exchange rate and exports performance of India is cointegrated from 1985 onwards. Subsequent studies by Sarkar (1995, 1997) also found that exchange rate is insignificant in explaining the export performance of India. Dhaolakia and Saradhi (2000) found that export quantity is sensitive to exchange rate. He also states that the exchange rate pass-through to export price is almost complete in the post 1991 period against an incomplete pass through during 1980 to 1991 period.

For the post reform period, Presty (2008) found a long run relationship between exchange rate and exports of India during 1992 to 2007. Eckaus (2008) also observes that exchange rate influences the exports of India and China.

By differing from the contemporary studies, the present study uses both the nominal and real exchange rate variables in the study as used by Sarkar (1994, 1995, and 1997) but differing from Sarkar's studies we have used income of India's exports markets namely OECD GDP as an additional variable. For the exchange rate variable, we have used one of the four exchange rate variables at a time, namely, the two versions of nominal exchange rate (exports weighted and trade weighted) and two versions of real exchange rate variables (exports weighted and trade weighted). By using the quarterly data, we are analyzing the nexus between OECD GDP, exchange rate and exports performance of India for the period 1994 to 2008 (during this period India followed a floating exchange rate system).

III. DATA AND VARIABLES

We have used the quarterly data on exports (in million dollars) of India, GDP of OECD region and four types of exchange rate variables namely trade weighted real effective exchange rate (REERT), Exports weighted real effective exchange rate (REERX), Trade weighted nominal effective exchange rate (NEERT) and exports weighted nominal effective exchange rate (NEERX). Since OECD region accounts for about 50% of India's exports during the study period, we use the OECD GDP as a proxy for the income of India's exports market. The monthly data on four types of exchange rate variables and exports are collected from the "Handbook of Statistics on Indian economy, 2009" published by RBI. The monthly data is converted into quarterly data. Quarterly data on OECD GDP is collected from the OECD Stat data base of the OECD.

IV. STATIONARY TEST OF THE DATA

The Augmented Dickey Fuller (ADF) and Phillips-Perron tests have been used to check the stationarity characteristics of the data. The results are given in table 1. In the ADF and Phillips-Perron tests, the selection of the model (constant, constant and trend, neither) is based on the Swartz Information Criteria (SIC).

The lag length selection is automatically done by the E-views software based on the SIC criteria. The ADF test results indicate that all the variables are non stationary at level form (posses a unit root), but in first difference form all the variables are stationary. So the study variables are integrated at 1st order, I(1). This is confirmed by the Phillips-Perron test results. Since all the variables are I(1), we have taken the first difference to make it stationary for the estimation.

V. EMPIRICAL RESULTS

We have regressed the exports of India on the lagged value of exports, OECD GDP current period, 1st lag of OECD GDP and 1st lag of exchange rate. For the exchange rate we have used one of the four types of exchange rate indices at a time, namely REERT, REERX, NEERX, and NEERT. Since, theoretically the exchange rate is influenced by the exports; the 1st order lagged exchange rate is used to avoid the endogenity problem in estimation. For the OECD GDP, we have used up to the 1st lag, since the second lag found to be not significant in the estimation. The empirical results for the models using the different exchange rate indices are given in the table 2.

In model 1 using the REERT, the exchange rate coefficient is not significant, but the coefficients of OECD GDP for the current period and the first lag are significant. The estimated coefficients for the current and first lag OECD GDP variables are 4.29 and 4.85 respectively. In models 2, 3 and 4 we have replaced the REERT with REERX, NEERT and NEERX respectively. The coefficient of exchange rate is found to be insignificant in all the four models. But we obtain almost the same coefficient for the OECD GDP for the current and Ist order lagged values. The results indicate that 1% increase in OECD GDP will increase the Indian exports by about 9% in two quarters.

The values of Durbin Watson Statistics indicate that all the four models are free from the problem of autocorrelation. The VIF scores are given in appendix table one. The JB test statistics indicates the normal distribution of residuals in all the four models.

VI. CONCLUSION

We have analyzed the effect of OECD GDP and exchange rate on the exports performance of India for the period 1994 to 2008. We have used one of the four exchange rate variables at a time namely REERT, REERX, NEERX, and NEERT in the estimation. The empirical results indicate that OECD GDP is an important determinant of India's exports and for the exchange rate variables, there is no evidence that the exchange rate is affecting the exports of India during the study period. Since the aggregate exports are used as the dependent variable, we are not rejecting the possibility of the effect of exchange rate on disaggregate commodity level exports and the possible aggregation bias in the estimation. While analyzing the influence of exchange rate volatility on the firm level exports of Turkish firms, Solakoglu et al. (2008), indicated that firms may absorb modest exchange rate changes to their prices. We are not rejecting this argument in the Indian context. Since India has a well developed foreign exchange market it is not difficult for the exporters to avoid the exchange rate risk through hedging. In this context it is desirable to conduct the study at the commodity level.

Appendix
Table 1
VIF Scores of Variables

Independent variables         Model 1   Model 2   Model 3   Model 4

                                  VIF       VIF       VIF       VIF

[DELTA] lnExports(-1)            1.06      1.06      1.05      1.05
[DELTA] lnOECD GDP               1.07      1.07      1.08      1.08
[DELTA] lnOECDGDP(-1)            1.12      1.15      1.10      1.08
[DELTA] lnExchange rate(-1)      1.08      1.10      1.08      1.04

References

Arrora et al. (2003), "Bilateral J Curve between India and her Trading Partners", Applied Economics, Vol. 35, 1037-1041.

Bahmani-Oskooee, M. (1985), "Devaluation and J curve: Some Evidence from LDCs", The Review of Economics and Statistics, Vol. 67, 500-504.

Baluswar, Thomson and Upadhaya (1996), "Devaluation and Trade Balance in India: Stationarity and Co integration", Applied Economics, Vol. 28, 429-432.

Dholakia, Raveendra H., Raveendra Saradhi V. (2000), "Exchange Rate Pass-through and Volatility: Impact on Indian Foreign Trade", Economic and Political Weekly, Vol. 35, 4109-4116.

Eckaus, Richart S. (2008), "An Inquiry into the Determinants of the Exports of China and India", China & World Economy, Vol. 16, 1-15.

Economic survey 2008-09 (2009), http/indiabudget.nic.in.

Global Financial Crisis and Kerala Economy: Impact and Mitigation Measures (2008), www.cds.edu.

Himarious D. (1985), "The Effect of Devaluation on Trade Balance: A Critical Review of and Reexamination of Mile's New Results", Journal of International Money and Finance, Vol. 4, 553-563.

Handbook of Statistics on Indian economy (2009), www.rbi.org.in.

Mookerjee, Rajen (1997), "Export Volume, Exchange Rates and Global Economic Growth: the Indian Experience", Applied Economics Letters, Vol. 4, 425-429.

OECD Stat data base: www.oecd.org

Presty, Sudananda (2008), "An Analysis of Exchange Rate and Exports Growth of India" The Business Review, Cambridge Vol. 9, 139-144.

Sarkar, Prabirjit. (1995), "Indian Economy Since 1991-trade: Price and Exchange Rate Behavior", Economic and Political Weekly, Vol. 30, 1197-1201.

Sarkar, Prabirjit (1997), "Foreign Trade and Real Exchange Rate Behavior 1980-96", Economic and Political Weekly, Vol. 32, 1133-1136.

Sarkar, Pribirjith (1994), "India's Balance of Payment and Exchange Rate Behavior Since 1971 A New Approach", Economic and Political Weekly, Vol. 29, 43-48.

Nag Ashok K. and Ghanesha Upadhaya (1994), "Exchange Rate and Trade Balance", Economic and Political Weekly, Vol. 29, 819-20.

Singh, Tarlog (2004)," Testing J-curve Hypothesis and Analyzing the Effect of Exchange Rate Volatility on the Balance of Trade in India", Empirical Economics. Vol. 29, 227-247.

Solakoglu et al. (2008), "Exchange Rate Volatility and Exports: A Firm-Level Analysis", Applied Economics, Vol. 40, 921-29.

Veeramani, C. (2007), "Sources of India's Export Growth in the Pre and Post Reform Periods", Economic and Political Weekly, Vol. 42, 4109-4116.

SURESH, K. G.

ICFAI University, Dehradun, India

V. NAGI REDDY

ICFAI Business School (IBS), Hyderabad, India

Table 1
Augmented Dickey Fuller Test (ADF) and Phillips-perron (PP) Test
Results

             Level form

variable       ADF   P value          P-P   P value
                               statistics

lnExports    -2.17      0.49        -2.09      0.54
lnOECD GDP   -0.01      0.99        -0.34      0.98
lnREERT      -0.42      0.52        -2.57      0.29
lnREERX      -0.53      0.48        -0.52      0.48
lnNEERT      -1.03      0.26        -4.53      0.27
lnNEERX      -1.02      0.27       -1.002      0.28

              First Difference form

variable       ADF       P          P-P       P
                     value   statistics   value

lnExports    -5.84    0.00        -8.38    0.00
lnOECD GDP   -5.23    0.00        -5.23    0.00
lnREERT      -6.95    0.00        -6.59    0.00
lnREERX      -4.44    0.00        -6.94    0.00
lnNEERT      -4.73    0.00        -6.93    0.00
lnNEERX      -4.87    0.00        -6.34    0.00

Table 2
Results of the Regression Analysis

Dependent variable: [DELTA] lnExports

Independent                    Model 1            Model 2
variables                      Using REERT        Using REERX
                               (Standard error)   (Standard error)

Constant                       -0.01(0.018)       -0.01(0.01)
[DELTA] 1nExports(-1)          -0.31(0.14) **     -0.30(0.14) **
[DELTA] 1nOECD GDP             4.29(1.74) ***     4.27(1.74) ***
[DELTA] 1nOECDGDP(-1)          4.85(2.08) **      4.75(2.11) **
[DELTA]  1nExchange rate(-1)   -0.39(0.47)        -0.26(0.48)
[R.sup.2]                      0.27               0.26
D W Statistics                 1.74               1.77
S.E. of the model              0.08               0.08
JB Statistic (p value)         3.08(0.21)         3.01(0.21)

Independent                    Model 3            Model 4
variables                      Using NEERT        Using NEERX
                               (Standard error)   (Standard error)

Constant                       -0.01(0.01)        -0.01(0.01)
[DELTA] 1nExports(-1)          -0.30(0.14) **     -0.31(0.14) **
[DELTA] 1nOECD GDP             4.37(1.75) ***     4.47(1.75) ***
[DELTA] 1nOECDGDP(-1)          4.55(2.05) **      4.75(2.06) **
[DELTA]  1nExchange rate(-1)   -0.21(0.46)        -0.40(0.41)
[R.sup.2]                      0.26               0.27
D W Statistics                 1.76               1.79
S.E. of the model              0.08               0.08
JB Statistic (p value)         3.14(0.18)         3.48(0.17)

[DELTA]-First difference operator, *** and ** shows significance at
2% and 5% respectively, -1 shows 1st lag value