COMMENTARY
Impact of Exchange Rate Appreciation on India’s Exports
C Veeramani
1962-87, found that the price elasticity of demand for exports was about 1.1 in the short run and about three in the long run. They noted that the periods of real exchange rate depreciations such as the 1970s and late 1980s were associated with rapid export growth, while the apprecia-
This article explores the relationship between the real effective exchange rate and exports for the period 1960-2007. Using World Trade Organisation and Reserve Bank of India data, it finds that the appreciation of the REER leads to a fall in the dollar value of India’s merchandise exports. It also forecasts the growth of merchandise exports over the medium term.
The author would like to thank Subrata Sarkar for helpful suggestions. The usual disclaimer applies.
C Veeramani (veeramani@igidr.ac.in) is at the Indira Gandhi Institute for Development Research, Mumbai.
I
With the benefit of hindsight derived from the healthy growth rate of exports since 2002, we should not, however, conclude that the entire hue and cry raised over the adverse impact of rupee appreciation on exports has been unwarranted. That exports did grow rapidly since 2002 despite the appreciation of the REER should not be taken to mean that the latter had no adverse effect on the former. The actual growth of exports might have been larger had the REER not appreciated. A conclusive answer for the relationship between the REER and exports begs a systematic analysis of the impact of the REER on exports, controlling for the influences of other determinants of exports.
The econometric analysis by Virmani (1991), Joshi and Little (1994), Srinivasan (1998) and Srinivasan and Wallack (2003) among others, showed that real exchange rate appreciation negatively affects India’s aggregate merchandise exports.2 Joshi and Little (1994), using data for the period tion in the 1960s and the first half of the 1980s were associated with slower export growth. Srinivasan’s (1998) analysis covered the period 1963-94 and reported price elasticity in the range (-0.304, -0.278) in the short run and (-0.493, -0.537) in the long run. Srinivasan and Wallack (2003) (henceforth SW), using a larger dataset covering the period 1960-98, obtain similar results. These studies also suggest a positive effect of India’s real GDP and world exports on exports.
In what follows, we analyse the relationship between the REER and exports for the period 1960-2007. As evident from the figure (p 11), the REER has been appreciating since 1999. It is important to understand whether the adverse effect of real exchange rate appreciation on exports continues to hold in the most recent years. While the studies mentioned above have dealt with the aggregate merchandise exports, we attempt a separate analysis of merchandise exports and services exports to discern whether the exchange rate has any differential impact in the two sectors. We also attempt a forecast of India’s export growth during the next five years (2008-12) under different scenarios relating to REER, world demand, and India’s GDP growth.
Data and the Model
Following SW (2003), we estimate an eclectic model, which postulates that the dollar value of India’s merchandise exports
(IX) is a function of the REER, India’s real GDP (RGDP) and world merchandise exports (WX). RGDP is a proxy for the export supply capacity and therefore, may exert a positive effect on exports. On the other hand, growth of GDP leads to higher domestic demand, which would lower exports. Thus, the sign of its coefficient depends on which effect dominates. The sign of WX is expected to be positive since higher world demand would also mean higher demand for India’s exports.
may 31, 2008
EPW
COMMENTARY
Data on the value (in $) of India’s merchandise exports and world merchandise exports are downloaded from the World Trade Organisation (WTO) web site. Real GDP data (at factor prices) is taken from the Handbook of Statistics, Reserve Bank of India (RBI). The 36-currency export weighted REER for the period 19752007 are downloaded from the RBI web site.3 Joshi and Little (1994) provides a series of REER adjusted for export incentives for the period 1960-88 (base: 1978 = 100). Applying the ratio of their incentive adjusted REER to unadjusted REER, we converted the unadjusted series (taken from RBI) to incentive-adjusted REER.4 The ratio of incentive-adjusted REER to the unadjusted REER in 1988 was 0.95. SW (2003) assumed a constant level of export incentives until reform began in 1991 and then reduced the weighting factor linearly (add 01 each year) until 1995. We followed the same procedure to construct the REER series. The final REER series used in the regression analysis is plotted in the figure.
The series on IX, REER, RGDP and WX were converted to logarithms to better approximate normality. Since regressing one non-stationary time series against for 1960-98, the period covered by SW (2003).6 Column (2) shows the results for the same period (1960-98) if the one-year lagged rather than contemporaneous values of the REER are used. It appears that the lagged values are appropriate since exports typically involve signing a sales contract before the actual export
Figure1 : Trend of the Real Effective Exchange Rate
2.3
2.2
2.1
2
1.9
1.8
1.7
Source: Author’s own calculations.
Table 1: ADF Unit Root Test for Stationarity
we include a period dummy (=1 for the period 2002-07 and 0 otherwise) among the regressors (column 4).
Inclusion of the period dummy indeed makes a big difference in that the REER t-1 now shows statistically significant negative coefficients.9 The period dummy enters the equation with a statistically
Log of REER (Adjusted for Export Incentives)
1961 1966 1971 1976 1981 1986 1991 1996 2001 2006
| Without Trend | With Trend | Conclusion | ||
|---|---|---|---|---|
| Log value of India’s exports | Level | 1.51 | -1.44 | Non-stationary |
| First difference | -5.37 | -5.55 | Stationary | |
| Log REER | Level | -1.06 | -0.78 | Non-stationary |
| First difference | -5.18 | -5.22 | Stationary | |
| Log real GDP | Level | 2.87 | -0.57 | Non-stationary |
| First difference | -6.73 | -8.12 | Stationary | |
| Log world exports | Level | -1.29 | -1.12 | Non-stationary |
| First difference | -3.97 | -4.03 | Stationary |
another can lead to spurious regression results, unit root tests for stationarity are performed on levels and first differences of the variables. The results are given in Table 1. The Augmented Dickey Fuller (ADF) test shows the existence of unit roots and therefore non-stationarity, in the levels of all variables, which is confirmed by the Philips-Perron unit root test. However, the first differences of all the variables are stationary under both the tests, which means that the original series are integrated of order 1.
The problem of stationarity in the level is addressed by estimating an equation, which includes a time trend and the one-year lagged values of exports on the right hand side along with other explanatory variables. We also run the regression on the first differences of the original variables.5
Regression Results
Table 2 (p 12) presents the results of the regressions on levels. For the sake of comparison, column (1) presents the results
Economic & Political Weekly
EPW
transaction takes place. Indeed the point estimate and t-values are higher when we use the lagged value rather than the contemporaneous value of the REER. Therefore, the analysis that follows uses the one-year lagged values of the REER as the explanatory variable.7
Results for the longer period (19602007) are shown in column (3). Both the lagged and contemporaneous values of the REER now fail to yield statistical significance.8 This may imply that some structural changes might have occurred during the more recent period leading to this result. As already mentioned, the period since 2002 has been characterised by a major acceleration in the growth rates of India’s exports [Veeramani 2007]. In order to account for this structural shift, significant positive coefficient. If the period dummy is interacted with log REERt-1, the coefficient of the interaction term (IN-TER) is positive (0.045) and statistically significant (column 6). This implies that while the short-run price elasticity was -0.230 during 1960-2001, it was lower at -0.185 during 2002-07. The correspon ding long-run elasticities are respectively -0.470 and -0.378. The error terms are stationary and there is no evidence of autocorrelation. Re-estimation of the regressions using the value of non-oil exports as the dependent variable makes little effect on the results.
The coefficient of RGDP is positive suggesting that the positive supply effect of the overall GDP growth dominates over the negative demand effect. RGDP is
COMMENTARY
potentially endogenous since its current The relationship between real statistically significant negative effect year value includes the current year value exchange rate and exports of services is of exchange rate appreciation on the of exports. However, the current export analysed using the WTO data on the exports of India’s services. Other would have no effect on the output of last exports of commercial services. This explanatory variables, such as the world Table 2: Effect of REER on Exportsexports of commercial services and
(1) (2) (3) (4) (5) (6)
Log REER -0.296 ----(2.35)**
Log REER
--0.423*** -0.033 -0.262*** -0.224** -0.230** (-4.33) (-0.35) (-2.58) (-2.19) (-2.25)
Log WX 0.433*** 0.409*** 0.495*** 0.482*** 0.482*** 0.481***
(3.05) (3.10) (4.02) (4.07) (5.39) (5.37)
Log RGDP 0.961** 0.767* 1.543*** 1.251***
(2.01) (1.80) (4.07) (3.48)
Log RGDP1.312*** 1.302***
t-1
(4.70) (4.67)
Log lagged exports 0.512*** 0.490*** 0.494*** 0.480*** 0.512*** 0.511***
(4.33) (4.24) (4.08) (4.18) (6.19) (6.18)
India’s services GDP, also fail to show significant coefficients. These results are not reported since the variables are not significant.
Forecasts of Merchandise Exports
In what follows, we attempt a growth forecast of merchandise exports during the next five years (2008-12) under different scenarios. This exercise is
Trend -0.022* -0.018* -0.031*** -0.028*** -0.029*** -0.029***
undertaken using the estimated equation
(-1.90) (-1.72) (-3.40) (-3.19) (-4.09) (-4.07) Period dummy 0.087*** 0.082***
(5) in Table 2. First, we consider the “busi
(4.51) (3.66) ness as usual” scenario, wherein we
| -(3.78) | assume that the trend growth rates of the |
| Constant -4.993 -3.17 -9.296*** -6.918*** -7.610*** -7.521*** | explanatory variables during 2002-07 |
| (-1.53) (-1.11) (-3.99) (-3.12) (-4.06) (-4.02) | would continue for the forecasting |
| R2 0.995 0.996 0.997 0.998 0.998 0.998 | period (2008-12). During 2002-07, the |
| n 38 37 46 46 46 46 | average annual growth rate of India’s |
| Period 1960-98 1960-98 1960-2007 1960-2007 1960-2007 1960-2007 | real GDP was 8.7 per cent, which is |
| (i) Robust t-statistic in parentheses. (ii) The ADF and Philips-Perron test confirm that the residuals are stationary. | assumed to continue during the next five |
| (iii) Durbin’s h statistic indicated no serial correlation. (iv) Dependent variable: log value of India’s merchandised exports. | years. The REER is assumed to appreciate |
| Table 3: Effect of REER on Exports | at the rate of 1.2 per cent per annum |
| data is available for the period 1980 (1) (2) (3) (4) | during the forecasting period (this was |
| 2007. These series are non-stationary D log REERt-1 -0.288* -0.288* -0.27 -0.108 | the annual rate of appreciation |
| in their levels and therefore we (-1.69) (-1.70) (-1.58) (-0.63) | during 2002-07). The dollar value of |
| included the lagged value of exports D log WX 0.669*** 0.668*** 0.700*** 0.739*** | world exports grew at the rate of 16.1 per |
| and time trend as regressors. We also (5.13) (5.27) (5.57) (5.85) D log RGDP 0.015 --- | cent per annum during 2002-07. This rate |
| run the regressions on first (0.04) | is assumed to continue during the |
| differences. D log RGDPt-1 --0.525 0.785* | forecasting period. |
| The 36-currency weighted REER (1.25) (1.90) Period dummy 0.042*** 0.042*** 0.036*** - | It is evident from Table 4 (p 13) that, |
| index, adjusted as well as unad (3.93) (4.77) (4.26) | under scenario 1 (business as usual), the |
| justed for export incentives, does Constant 0.006 0.007 0.004 0.005 | value of export would grow at the rate of |
| not show economically and statisti (0.58) (0.79) (-0.30) (-0.32) R2 0.51 0.51 0.54 0.45 | 25.1 per cent per year during 2008-12, |
| cally significant results when n 45 45 45 45 | which is similar to the performance |
| applied to the case of services Period 1960-2007 1960-2007 1960-2007 1960-2007 | during 2002-07. The ministry of com |
| export. It may, however, be noted (i) Robust t-statistic in parentheses. | merce and industry has set a merchan |
| that the construction of the REER (ii) The ADF and Philips-Perron test confirm that the residuals are stationary. (iii) Durbin-Watson d-statistic (2.06) indicated no serial correlation. | dise export target of $ 200 billion for |
| makes use of the shares of each (iv) Dependent variable: log value of India’s merchandise exports, first differences. | the 2008-09 fiscal year (April-March) |
| partner in India’s merchandise | from the actual export value of $ 155 |
| year. To avoid the potential endogeneity exports as weights whereas the direction | billion during 2007-08. Under the busi |
| problem, the one year lagged value of RGDP of India’s services exports is markedly dif | ness as usual scenario our forecasted |
| is considered but this has little effect on the ferent from that of the merchandise. For | figure for 2008-09 is $ 194 (applying 25.1 |
| results. As expected, the coefficient of world example, while the United States (US) | per cent growth over the actual value of |
| exports is significant and positive, underaccount for about 17 per cent of India’s | $ 155 billion in 2007-08), which is margin |
| scoring the importance of world demand in total merchandise exports, it accounts for | ally lower than the target. |
| determining India’s export growth. Another about 68 per cent of India’s exports of soft- | Scenario 2 assumes that the REER would |
| approach to account for non-stationarity of ware/information techno logy exports. | remain constant at the 2007 level (i e, the |
| the original series is to run the regression Thus, instead of the REER, we have used | rate of appreciation is assumed to be zero) |
| on its first differences (Table 3). Overall, the exchange rate of rupee against US | but other variables would change as under |
| the results are as expected and are condollar, adjusted for inflation in both the | scenario 1. It is evident from Table 4 that |
| sistent with those reported in Table 2. countries.10 Again, we do not find a | the additional gain from keeping the REER |
| 12 | May 31, 2008 Economic & Political Weekly |
INTER 0.045***
EPWCOMMENTARY
constant is relatively small since the aver-may fall to about 12 per cent per annum, age annual growth rate will be 25.6 per India’s exports will grow at the rate of 21.3 cent, which is only marginally higher than per cent per annum. the growth rate of 25.1 per cent under The major downward risk on the growth the business as usual scenario.11 In fact, rate of exports may arise from the fall in the cost of keeping the REER constant the growth rate of India’s real GDP. Sce(which may include higher inflation and nario 6 suggests that if the growth rate of
Table 4: Forecasted Growth Rates of India’s Merchandise Exports, 2008-12 (US $)
2008 2009 2010 2011 2012 Average India’s Growth Rate Share in World Exports (2008-12) by 2020 (Predicted) (%)
Scenario 1: business 178.7 222.4 278.2 348.8 437.9 as usual (22.3) (24.4) (25.1) (25.4) (25.5) (25.1) 2.90
Scenario 2: REER 178.7 223 280 352.8 445.2 (25.6) remain constant at the 2007 level (22.3) (24.8) (25.6) (26.0) (26.2) 3.07
Summary and Conclusions
To conclude, we find that the appreciation of the REER leads to a fall in the dollar value of India’s merchandise exports. However, the degree of the (negative) association between exports and the REER has declined since 2002. In the case of commercial services, we do not find a statistically significant negative effect of real exchange rate appreciation on exports. Should the RBI continue the practice of intervening in the foreign exchange market to smoothen the appreciation of exchange rate is a larger question of
Scenario 3: REER 178.7 220.6 272.5 336.6 416 (23.5)
overall macroeconomic management.
appreciates at 5% per annum (22.3) (23.4) (23.5) (23.6) (23.6) 2.41
Export promotion should not be the
Scenario 4: growth rate of world 174.2 208.4 248.9 297.1 354.3 exports falls to 10 per cent per annum (19.2) (19.7) (19.4) (19.3) (19.3) (19.4) 3.08 sole criterion in deciding for or against
Scenario 5: growth rate of world exports 175.7 213 258.4 313.5 380.4
the intervention. Artificial depreciation
falls to 12% per annum (20.2) (21.3) (21.3) (21.3) (21.3) (21.3) 3.02
comes with its cost (including higher
Scenario 6: growth rate of India’s real 178.7 217.9 264.1 319.2 385.2 GDP falls to 7% per annum (22.3) (21.9) (21.2) (20.9) (20.7) (21.1) 1.82 inflation and higher interest rates) while
Scenario 7: growth rate of world exports and
of India’s real GDP falls to 10 per cent 174.2 204.1 236.3 271.9 311.7
and 7%, respectively (19.2) (17.2) (15.8) (15.0) (14.7) (15.6) 1.94
| higher interest rate) may well outweigh | India’s real GDP falls to 7 per cent from the |
| the benefit. | present rate of 8.7 per cent per annum, |
| On the other hand, the cost (in terms of | India’s exports would fall from 25.1 per |
| export earnings) of allowing the REER to | cent to 21.1 per cent per annum. This |
| appreciate is not very high. Under scenario | means that for a 1 percentage point |
| 3, we assume that the REER will appreci | fall in the growth rate of India’s real GDP, |
| ate by 5 per cent per annum while other | the growth rate of exports will fall by 2.4 |
| variables are assumed to grow as under | percentage points. The worst scenario |
| scenario 1. It is clear that even if the REER | that we have considered is the fall |
| appreciates at the rate of 5 per cent per | in the annual growth rate of India’s real |
| annum (which is three times higher than | GDP to 7 per cent along with a fall in the |
| the actual rate of appreciation during | growth rate of world exports to 10 per |
| 2002-07), India’s exports would still | cent. Under this scenario, the average |
| grow at a healthy rate of 23.5 per cent | annual growth rate of India’s exports |
| per annum. | would be still 15.6 per cent (scenario 7). |
| A fall in the rate of growth of world | The ministry of commerce and industry |
| exports exerts a much greater downward | has set a target of increasing India’s share |
| pressure, compared with the downward | in world exports from the current level of |
| pressure from the REER appreciation, on | 1 per cent to 5 per cent by 2020. Table 4 |
| the rate of growth of India’s exports. | suggests that this target is highly |
| Nevertheless, even if the rate of growth of | ambitious. Under the business as |
| world exports falls to 10 per cent per | usual scenario, India’s share in world |
| annum from the present 16.1 per cent, | exports would increase to 2.9 per cent |
| India’s exports would grow at the rate of | by 2020. Even if the REER remains |
| 19.4 per cent. In the context of the loom | constant at the 2007 level, India’s share |
| ing recession in the US, the growth rate | in 2020 will be only 3.07 per cent. |
| of world exports is likely to slow down in | Under the best possible scenario (an |
| the near future but the growing import | nual rate of growth of world exports at |
| demand in many developing countries | 12 per cent, annual rate of growth of |
| may act as a counter force against the | India’s GDP at 9 per cent and 0 per cent |
| downward pressure emanating from the | appreciation of the REER), India’s share |
| US slowdown. Under the most likely sce | in world exports would be 3.5 per cent |
| nario that the growth rate of world exports | by 2020.12 |
| Economic & Political Weekly may 31, 2008 |
EPWthe gain in the form of higher export earnings appears to be small. Export policy should instead focus on other measures that can promote exports on a sustained basis.
The analysis suggests that the major downward risk on the growth rate of exports may arise from the fall in the growth rate of India’s real GDP. Any slowdown in India’s GDP growth would also lead to a fall in the growth rate of exports since the former implies a fall in the potential capacity for export. The most effective policies to promote exports are not really different from those that can accelerate the country’s overall economic growth.
Our analysis suggests that the strong growth of merchandise exports that India has been witnessing during the last six years is likely to continue, at least for the next five years (2008-12). We find that the additional gain from keeping the REER constant is relatively small. A fall in the rate of growth of India’s real GDP or that of world exports exerts a much greater downward pressure, compared with the downward pressure from the REER appreciation, on the rate of growth of India’s exports.
Notes
1 The 36-currency export weighted REER (base: 1993-94 = 100) published by the Reserve Bank of India. In constructing the REER index, the exchange rate is defined as foreign currency units per rupee. Thus, an increase of the REER means appreciation of rupee and a fall means depreciation.
COMMENTARY
2 For sectoral level analysis (within the merchandise) of the relationship between relative prices and exports, see Virmani et al (2006) and Sinha Roy (2007). Virmani et al covered the period 1980-2000 while Sinha Roy covered 1960-99.
3 The RBI provides the monthly series of the REER (base: 1985 = 100). We used the calendar year averages.
4 For the period 1960-1974, we interpolated the RBI series on REER applying the annual percentage changes in the series provided by Little and Joshi (1994).
5 SW (2003) followed the same approaches to address the non-stationarity in the original series.
6 The point estimates and t-values that we have obtained for the period 1960-1998 are not strictly the same as that reported by SW (2003) for two reasons. First, the data sources are different. Second, we use data for the calendar year (January-December) for all the variables (except RGDP) while SW have used the financial year (April to March) data in most cases.
7 SW (2003) used the contemporaneous values of REER in their regressions. 8 Results with the contemporaneous values are not reported to save space.
9 The REER is negative if the contemporaneous values are used but significant only at the 10 per cent level.
10 As indicators of inflation, we have used whosesale price index for India and producer price index for the US.
11 The forecasted value of export in 2008 is the same under scenarios 1 and 2 because the model uses the lagged value of REER (i e, the 2007 value), which is the actual and not different under the two scenarios.
12 The share will be 3.3 per cent if the world exports grows at the rate of 16.1 per cent, India’s GDP at the rate of 9 per cent and 0 per cent appreciation of the REER.
References
Joshi, V and, I M D Little (1994): India: Macroeconomics and Political Economy, 1964-1991, World Bank and Oxford University Press, Washington DC and New Delhi.
Sinha Roy, Saikat (2007): ‘Demand and Supply Factors in the Determination of India’s Disaggregated Manufactured Exports: A Simultaneous Error-Correction Approach’, Working Paper No 383, Centre for Development Studies, Thiruvananthapuram.
Srinivasan, T N (1998): ‘India’s Export Performance: A Comparative Analysis’ in I J Ahluwalia and I M D Little (eds), India’s Economic Reforms and Development: Essays for Manmohan Singh, Oxford University Press, Delhi.
Srinivasan T N and Jessica Wallack (2003): ‘Export Performance and the Real Effective Exchange Rate’ in Anne O Krueger and Sajjid Z Chinoy (eds), Reforming India’s External, Financial, and Fiscal Policies, Stanford University Press.
Veeramani, C (2007): ‘Sources of India’s Export Growth in Pre- and Post-Reform Periods’, Economic & Political Weekly, Vol 42, No 25, pp 2419-27.
Virmani, Arvind (1991): ‘Demand and Supply Factors in India’s Trade’, Economic & Political Weekly, Vol 26, No 6.
Virmani Arvind, B N Goldar, C Veeramani and V Bhat (2006): ‘Impact of Tariff Reforms on Indian Industry: Assessment Based on a Multi-sector Econometric Model’ in Arvind Virmani, Propelling India from Socialist Stagnation to Global Power (Vol 2, Chapter 7), Academic Foundation, New Delhi.
May 31, 2008
EPW
