Abstract
This paper reviews the impact of currency movements in Asia over the past year on India’s competitiveness, and examines the outlook for India’s exports. First, real exchange rate indicators are used to assess the competitiveness of India’s exports from several perspectives. Second, alternative econometric methods are used to estimate the equilibrium real exchange rate and measure the extent of misalignment of the rupee. The results, based on these backward–looking assessments, indicate that the rupee was not substantially overvalued in 1997/98, and that the likely impact of currency realignments in the region on India’s exports appears to be small.
III. Exports and Competitiveness1
A. Introduction
1. Since the onset of the Asian financial crisis in July 1997, the rupee has depreciated by about 19 percent against the dollar. By contrast, in real effective terms, the rupee has remained broadly unchanged since July 1997, and has appreciated by about 8 percent since the unification of the exchange rate and the shift to a managed float in March 1993. Currency realignments in the wake of the Asian crisis and the continued sluggish performance of exports in dollar terms have hence heightened concerns that the rupee has become overvalued.
2. This chapter focuses primarily on the following question: have currency realignments in Asia over the past year had an impact on India’s competitiveness and on the outlook for exports? Recognizing that no single indicator offers a complete and satisfactory assessment of competitiveness, and that in examining export competitiveness, price-based competitiveness indicators should not be the only part of the analysis, this chapter reviews a broad range of methodologies from both a technical and econometric point of view. To put recent movements of the rupee in perspective, section B provides a historical overview of developments in selected external indicators as well as in India’s trade and exchange regimes. Section C assesses alternative measures of competitiveness. Section D discusses various approaches to measuring exchange rate misalignment, and assesses recent levels of the rupee by estimating an equilibrium real effective exchange rate (REER). Section E examines the likely impact on India’s exports of recent exchange rate developments. Finally, section F contains concluding remarks and discusses policy issues. The main conclusions are as follows:
Most of the appreciation of the rupee in real terms in recent years appears to be an equilibrium phenomenon reflecting movements in economic fundamentals. There is also evidence of a Balassa-Samuelson effect.2
A number of econometric techniques suggest that the rupee was not substantially overvalued in 1997/98.3 However, the results should be interpreted with caution given the inherent impression of econometric estimates, and the fact that more recent events that could have affected competitiveness are not yet fully captured in the data.
The recent sluggish performance of exports in dollar terms appears to have been related mainly to nonprice factors. Moreover, the likely impact on India’s exports of currency realignments in the region appears to be small.
B. Historical Perspective
3. Until the 1980s, the rupee was significantly overvalued as external sector policies were used to promote the objective of industrialization through import substitution. With the resulting subsidization of imports, quantitative restrictions on imports and high import tariffs were maintained to prevent substantial balance of payments deficits. Import restrictions were accompanied by rationing of foreign exchange, price and capital account controls. The lack of a realistic alignment of the exchange rate led to major distortions in the relationship between domestic and foreign prices, that, coupled with export controls, penalized exports.
4. In the early 1980s, a progressive liberalization of the trade and exchange regimes began. Following an IMF–supported stabilization program in 1981, India gradually switched from import substitution to export promotion policies. By 1985, India had adopted a medium-term export and import policy intended to provide a strong base for improving the competitiveness of its exports, with an emphasis on the liberalization of essential imports of capital goods and intermediate inputs. To support this objective, the rupee was allowed to depreciate sharply (in nominal and real terms) beginning in 1985 (Chart III.1). With relatively conservative macroeconomic policies in place, movements in the REER reflected primarily policies to liberalize the exchange regime: comparison of the REER and the nominal effective exchange rate during the second half of the 1980s shows that the driving force behind movements in the REER was the nominal depreciation of the rupee. The policy of bringing the rupee to more realistic levels and the partial easing of the trade and exchange restrictions contributed to rapid export growth (see Chart III. 1), which in turn helped expand the tradable goods sector.4
INDIA: EFFECTIVE EXCHANGE RATES, EXPORTS, AND THE CURRENT ACCOUNT, 1979/80–1997/98 1/
Citation: IMF Staff Country Reports 1998, 112; 10.5089/9781451818536.002.A003
Sources: Data provided by the Indian authorities; IMF, Information Notice System; and IMF, World Economic Outlook.1/ April-March fiscal year.2/ Effective exchange rates from the IMF’s Information Notice System (INS).3/ Merchandise exports only.5. India’s outward orientation increased during the 1990s. The stabilization and reform efforts initiated in 1991 in response to a severe balance of payments crisis were accompanied by a sizable depreciation of the rupee. The trade element of the reforms involved a further liberalization of the import regime by removing quantitative restrictions on imports, with the exception of consumer goods. Moreover, import tariffs were lowered across the board in a phased manner.5 The unification of the exchange rate in March 1993 made the rupee convertible for current account transactions and brought the REER in line with fundamentals. This move was accompanied by a shift to a managed float of the exchange rate, and followed by steps toward full capital account convertibility. A surge in private capital inflows—particularly portfolio and foreign direct investment—that started in late 1993 led to pressure on the rupee to appreciate and to a substantial buildup in reserves. Reflecting the adjustment in the exchange rate and the more open trade and exchange regimes, India’s share in world merchandise exports rose from 0.5 percent in 1991 to 0.65 percent in 1997.
6. Given the distortions in the exchange rate prior to 1993, it is more meaningful to discuss developments in India’s competitiveness from 1993 to date, thus treating 1993 as a structural break. Indeed, the authorities have often used the level of the REER prevailing in March 1993—as computed by the Reserve Bank of India (RBI); see section C—as a benchmark equilibrium rate to gauge developments in competitiveness, with the recognition that the equilibrium rate may have gradually shifted.6 However, this recognition is not widely shared by market participants (especially exporters and the corporate sector), and there are many who view the appreciation of the rupee in real effective terms since 1993 as an erosion of competitiveness that should be rectified by a weaker rupee. The following analysis shows that it is not appropriate to compare the current level of the REER against its level in March 1993 to assess developments in competitiveness: although the rupee appeared to be consistent with an equilibrium in the external position in 1993, the statistical evidence (see section D below) points to an equilibrium real appreciation since then.
C. Alternative Measures of Competitiveness
7. The IMF’s Information Notice System (INS) calculations of India’s REER do not suggest a loss of competitiveness since the onset of the currency crisis in Asia in July 1997. Despite the significant depreciation of currencies in Asia, the depreciation of the rupee vis-à-vis the dollar since then has maintained the REER of the rupee broadly at its level prior to the turmoil in Asia (Chart III.2).
INDIA: EXCHANGE RATE DEVELOPMENTS, 1993–1998
Citation: IMF Staff Country Reports 1998, 112; 10.5089/9781451818536.002.A003
Sources: Data provided by the Indian authorities; and IMF, Information Notice System.8. However, the INS calculations of India’s effective exchange rates are not without shortcomings. First, they rely on a weighting scheme that takes into account trade statistics over the period 1988–90, which may not accurately reflect India’s current trade patterns. Second, they may not fully capture recent developments in competitiveness because they exclude the ASEAN–4 countries (Indonesia, Malaysia, the Philippines, and Thailand), as these countries have bilateral competitiveness weights of less than one percent.7 These countries have played an increasingly import role in India’s competitiveness in recent years that is not captured in the INS weights. Motivated by these shortcomings, this section addresses the following question: what do alternative measures of competitiveness indicate? This question is addressed by examining two alternative sets of real effective exchange rates.
CPI-based REER indices
9. To assess the discrepancy created by the omission of the ASEAN–4 countries from India’s REER calculations, India’s CPI–based REER (from the INS) is compared with the following CPI–based REER indices, which have bilateral competitiveness weights that include the ASEAN–4 countries:8 (i) an REER index constructed using the full set of bilateral competitiveness weights (referred to as INS–FULL) from the latest revision to the INS carried out over 1994–96 that includes all countries with bilateral competitiveness weights below one percent (thus including the ASEAN–4 countries); (ii) an REER index based on the standard (truncated) INS weights, but selectively adding the ASEAN–4 countries in the calculations, with competitiveness weights calculated on the basis of Direction of Trade Statistics (DOTS) for 1994–96, which provide a reasonably accurate reflection of current trading patterns; and (iii) India’s REER index as computed by the RBI, which covers a broader range of countries than the INS (36 in total, including the ASEAN–4 countries), with bilateral competitiveness weights based on trade statistics over the period 1975–91.9
10. The above indices have a number of advantages as well as shortcomings:
The methodology used by the INS to compute the REER is superior because it takes into account third-market competition. However, as noted above, countries with bilateral competitiveness weights of less than one percent (such as the ASEAN-4 countries in the case of India’s effective exchange rates) have been excluded from the calculations. Therefore, in periods of currency turmoil, the INS indices might not accurately track developments in competitiveness if they exclude the countries whose currencies undergo large fluctuations. Moreover, calculations of weights in the INS are based on trade statistics that do not capture the changes in India’s trade patterns during the 1990s.10
The methodology for constructing the REER index based on DOTS trade shares for 1994–96 is ad hoc: due to computational constraints, the only weights adjusted using DOTS trade shares were the ones for the ASEAN-4 countries. As a result of the significant depreciations of the ASEAN-4 currencies over the past year, the REER derived this way would, by design, show a more pronounced appreciation (or a less pronounced depreciation) of the rupee since July 1997 than the INS index.
The RBI’s REER does not capture any third-market competition, a particularly relevant issue for Indian exports following the severe depreciation of some competitor currencies. Moreover, although the RBI’s REER includes a broader range of countries (including the ASEAN-4 countries), its bilateral competitiveness weights are based on outdated trade statistics. These weights under represent the countries of the European Union, and overemphasize the importance of the U.S. in India’s trade.11
11. Despite its shortcomings, the INS index appears to capture reasonably well recent developments in competitiveness: Table III.1 shows that the differences between the four CPI-based REER indices discussed above are small. By all four REER measures, the rupee has appreciated by 8–12 percent from March 1993 to date, and has remained broadly unchanged since July 1997 when the turmoil in Asia began (Chart III.3). As expected, the REER indices that include the ASEAN-4 countries (most notably INS-FULL and DOTS) show a slightly less pronounced depreciation in recent months.
India: CPI-based Real Effective Exchange Rates, 1993–98
(In percent)
Percent changes through May 1998 for INS, INS-FULL, and DOTS; through April 1998 for RBI.
India: CPI-based Real Effective Exchange Rates, 1993–98
(In percent)
| Change from March 1993 to date 1/ | Change from July 1997 to date 1/ | |
|---|---|---|
| INS | 8.2 | −1.5 |
| INS-FULL | 8.4 | −0.8 |
| DOTS | 8.4 | −0.9 |
| RBI | 12.1 | 12.1 |
Percent changes through May 1998 for INS, INS-FULL, and DOTS; through April 1998 for RBI.
India: CPI-based Real Effective Exchange Rates, 1993–98
(In percent)
| Change from March 1993 to date 1/ | Change from July 1997 to date 1/ | |
|---|---|---|
| INS | 8.2 | −1.5 |
| INS-FULL | 8.4 | −0.8 |
| DOTS | 8.4 | −0.9 |
| RBI | 12.1 | 12.1 |
Percent changes through May 1998 for INS, INS-FULL, and DOTS; through April 1998 for RBI.
INDIA: CPI-BASED REAL EFFECTIVE EXCHANGE RATES, 1997–1998 1/
Citation: IMF Staff Country Reports 1998, 112; 10.5089/9781451818536.002.A003
Sources: IMF, Information Notice System, Direction of Trade Statistics; and staff estimates.1/ For India, the WPI is used in all REER indices.REER indices based on CPI, WPI, and export unit values
12. In order to assess competitiveness based on a broader set of indicators and to detect any Balassa-Samuelson effects, the following REER measures (which are based on different price indices) are compared: the CPI-based REER computed by INS, a WPI-based REER, and an REER based on relative export unit values. The rationale for using wholesale prices in the computation of REERs is that they contain a larger traded goods component than consumer prices, and that they are not as much influenced by price controls and indirect taxes. Similarly, export unit values capture a large component of trade in goods.12
13. It appears that the real appreciation of the rupee since 1993 (based on INS calculations) may not have jeopardized export competitiveness in recent years. The divergent trends in the CPI-based and WPI-based REER, on the one hand, and the REER based on relative export unit values, on the other (Chart III.4), point to the direction of relatively faster productivity growth in India’s tradable goods sector than in other countries (the Balassa-Samuelson effect).13 In India, the gap in productivity growth between the tradable and nontradable goods sectors appears to have grown more than in partner countries, thus confining the price increases that led to the real appreciation of the rupee primarily to the nontradable goods sector. As a result, relative export prices have fallen more rapidly relative to broader price indices in India than elsewhere, following the accelerated liberalization of the trade and exchange regimes beginning in the early 1990s.
INDIA: REAL EFFECTIVE EXCHANGE RATE INDICES, 1980–1998
Citation: IMF Staff Country Reports 1998, 112; 10.5089/9781451818536.002.A003
Sources: Data provided by the Indian authorities, IMF, Information Notice System; IMF, International Financial Statistics; and staff estimates.1/ Annual data.D. The Level of the Rupee
14. The appreciation of the rupee in real effective terms in recent years has been cited as one of the main factors behind the slowdown in India’s exports in dollar terms. The task of isolating the effect of the REER on exports from all other influences becomes particularly difficult in an environment of gradual liberalization of the trade and exchange regimes. Nonetheless, it appears that the recent performance of exports has primarily reflected nonprice factors (Box III.1). Against this background, this section examines whether the rupee was overvalued in 1997/98 by measuring the extent of its misalignment in real effective terms in relation to an estimated equilibrium REER.14 To increase confidence in the findings, the degree of misalignment was measured using three different methods:
A Hodrick-Prescott (HP) filter was used based on the premise of relative purchasing power parity over the long run, and on the empirical evidence showing the REER exhibiting mean-reverting properties; that is, deviations of the REER from equilibrium are temporary, although sometimes relatively long lived.15 According to this method, the rupee was slightly overvalued (1½ percent) at end-March 1998, and most of its misalignment at the onset of the Asian crisis has been subsequently corrected (Chart III.5).16
The Beveridge-Nelson decomposition method was used to separate permanent from temporary movements in the REER (the latter associated with misalignments). This method is not designed to estimate an “equilibrium” REER in an econometric sense, but rather to decompose movements in the REER into a permanent (stochastic trend) component and a temporary (cyclical) component.17 According to this method, the rupee was about one percent above its stochastic trend at end-March 1998 (Chart III.6).
The long-run cointegrating relationship between the REER and its determinants was estimated to assess whether the appreciation of the rupee since 1993 has been an equilibrium phenomenon, and to measure the extent of misalignment (if any) of the rupee in 1997/98. Regressions using annual data found an empirical long-run link between India’s REER and the following fundamental variables: (i) trade policy (proxied by the sum of exports and imports over GDP), capturing the impact on India’s REER of the shift in the trade regime; (ii) the external environment (proxied by India’s terms of trade); and (iii) domestic supply factors (productivity differential between the tradable and nontradable goods sectors). The statistical evidence suggests that: (i) most of the appreciation of the rupee since 1993 has been an equilibrium phenomenon reflecting movements in economic fundamentals; and (ii) the rupee was overvalued by 2–3 percent in 1997/98 (Chart III.7).18
INDIA: LONG-RUN REAL EFFECTIVE EXCHANGE RATE AND DEGREE OF MISALIGNMENT, 1993–98
Citation: IMF Staff Country Reports 1998, 112; 10.5089/9781451818536.002.A003
Sources: IMF, Information Notice System; and staff estimates.INDIA: LONG-RUN REAL EFFECTIVE EXCHANGE RATE AND DEGREE OF MISALIGNMENT, 1993–98
Citation: IMF Staff Country Reports 1998, 112; 10.5089/9781451818536.002.A003
Sources: IMF, Information Notice System; and staff estimates.1/ 3-month moving average.INDIA: LONG-RUN REAL EFFECTIVE EXCHANGE RATE AND DEGREE OF MISALIGNMENT, 1979/80–1997/98 1/
Citation: IMF Staff Country Reports 1998, 112; 10.5089/9781451818536.002.A003
Sources: IMF, Information Notice System; and staff estimates.1/ April-March fiscal year.The Performance of India’s Exports
India’s exports remained weak in dollar terms in 1997/98. Exports grew by 2½ percent in dollar terms in 1997/98, compared with 4½ percent in the previous year, far below the growth rates recorded in recent years. However, the performance of exports has been robust in volume terms: partner country import deflators from WEO suggest that export volumes rose by 9 percent in 1996/97 and by about 11 percent in 1997/98 (see Chart III.1).
There have been marked differences in export performance across sectors. A number of export sectors—accounting for nearly half of total exports—have performed relatively well, most notably chemicals and related products, gems and jewelry, and engineering goods. However, traditional sectors, such as agricultural products, garments, and leather goods have performed poorly.
The performance of exports has primarily reflected nonprice factors. A number of export sectors have been affected by weaker world demand, as well as specific factors such as the virtual freeze on wheat exports to meet domestic demand, antidumping action by the European Union affecting garment exports, and stricter domestic environmental norms affecting leather exports. The performance of exports has also reflected structural factors such as infrastructure bottlenecks, most notably in electric power and port facilities.
The lack of reforms in the export-oriented small-scale sector has also constrained export growth. Small-scale units are defined as manufacturing units in which investment in plant and machinery is less than Rs 30 million (about $700,000). These units enjoy special protection from both domestic and international competition. About 800 products are reserved for production by small-scale units. The small-scale sector’s products account for roughly half of all exports. These restrictions have made it difficult to take advantage of economies of scale, and have hence hindered the competitiveness of exports. A 1997 official report by the Hussain Committee assessed “…the case for reservations is fundamentally flawed… The policy for reservations has crippled the growth of several industrial sectors, restricted exports, and has done little for the promotion of small-scale industries.”
15. The patterns of misalignment estimated by the HP filter and the Beveridge-Nelson decomposition method broadly share their main features. In particular, both methods capture reasonably well the two episodes of turbulence in the foreign exchange market in recent years: the depreciation of the rupee vis-à-vis the dollar in late 1995 and early 1996, and again from August 1997 onward. Moreover, in most instances the misalignment of the rupee has been confined within 5 percentage points around its equilibrium value.19
16. The cointegration analysis is the most appealing of the three methods because it has a solid theoretical foundation. However, lack of higher frequency data for a number of explanatory variables for India’s REER prevented the empirical analysis from assessing short-run movements of the REER. The empirical analysis was also unable to establish a long-run link between fiscal policy (namely government consumption or investment) and the REER.20 This is not an unexpected result, given that fiscal policy is likely to only have a short-run effect on the REER that may not be captured by annual data. Moreover, the analysis failed to show a long-run link between capital inflows and the REER. Several factors may have accounted for this result. First, strong capital inflows are a relatively recent phenomenon in India, that may not have been sufficiently protracted for cointegration analysis to establish a long-run link between the inflows and the REER. Second, despite the surge since 1993, capital inflows remain relatively small compared to GDP.21 Third, intervention by the RBI in response to capital inflows may have prevented cointegration analysis from detecting a statistically significant impact of capital inflows on the REER. See the annex to this chapter for a detailed discussion of the statistical tests and the estimation results.
E. Impact on Export Performance
17. The likely impact on India’s exports of currency realignments in the region appears to be small. Using the export elasticities estimated in Srinivasan (1998),22 if the estimated overvaluation of the rupee persists at about 2 percent, it would imply a decline in export volumes by 0.6 percent in 1998/99, and by one percent over the long run (relative to where the export volumes would otherwise have been). However, the depreciation of the rupee vis-à-vis the dollar since end-March 1998 (a period outside the empirical analysis) has resulted in a decline of the REER of about one percent over the same period, likely bringing the REER closer to its equilibrium value. Hence, the above estimates could be seen as the upper end of the range of the likely impact on export performance of recent currency movements.
18. It is worthwhile distinguishing between disequilibrium movements in export volumes, as estimated above, and equilibrium movements in export volumes in response to equilibrium shifts in fundamental determinants. Cointegration analysis estimated the equilibrium appreciation of India’s REER since 1993/94 at around 6 percent (that is, an actual appreciation of 8½ percent less an estimated overvaluation of around 2½ percent). In view of the long-run elasticities noted above, this equilibrium appreciation of the REER implies a decline in export volumes over the long run (other things constant) of around 3 percent (relative to where export volumes would otherwise have been). This equilibrium decline in export volumes accounts for a relatively small part of the actual decline since 1993/94, thus pointing to the importance of nonprice factors in explaining recent export performance.
19. Third-market effects of currency depreciations are notoriously difficult to quantify. An attempt was made to identify countries in the region that could challenge India with increased competition in third markets. To this end, the composition of India’s exports was compared against the composition of exports of the ASEAN-4 countries, China, and Pakistan. In particular, detailed trade data (based on 2—digit Standard International Trade Classification data) were used to compute the correlation between the commodity shares of the total exports of India and those of each of these countries. The magnitude of the estimated correlation coefficient—which in effect measures similarities in the structure of exports—was then used as an indication of the degree of competition for exports between India and these countries. The analysis suggests that India competes in the same commodities primarily with Pakistan, China, and Thailand (Table III.2). In view of regional currency movements over the past year, and assuming no major regional currency realignments in the near term, the main threat to India’s exports appears to be from Thailand, which has a similar basket of exports to common markets, mainly the U.S. and the European Union.
Selected Asian Economies: Principal Exports, 1996
Includes footwear for Malaysia.
Includes electronics and computer software for India; includes electronics and electronics components for all other countries.
Non-oil/gas exports for Indonesia.
Correlation between the commodity shares of total exports of India and the respective country, based on 2-digit Standard International Trade Classification (SITC) data for 1993–95 from the TARS database. Coefficients close to one suggest that the two countries compete in the same commodities.
Percent change since June 30, 1997. Positive figures indicate appreciation of the rupee vis-à-vis the respective currency.
Selected Asian Economies: Principal Exports, 1996
| India | Indonesia | Malaysia | Philippines | Thailand | Pakistan | China | |||
|---|---|---|---|---|---|---|---|---|---|
| (In millions of U.S. dollars) | |||||||||
| Textiles and clothing 1/ | 7,408 | 6,329 | 2,767 | 2,328 | 4,066 | 3,021 | 37,150 | ||
| Agricultural and fishery products | 6,952 | 7,438 | 1,539 | 1,320 | 9,145 | 1,152 | 11,950 | ||
| Of which: Rice | 1,009 | … | -- | -- | 2,012 | 504 | 1 | ||
| Gems and jewelry | 4,707 | 423 | 610 | 62 | 2,042 | … | 1,444 | ||
| Engineering goods 2/ | 4,471 | 4,222 | 10,008 | 9,292 | 9,379 | … | 22,147 | ||
| Chemicals and chemical products | 3,072 | 244 | 2,678 | 353 | 1,792 | 87 | 8,879 | ||
| Leather and leather goods | 1,441 | 87 | 41 | 121 | 400 | 259 | 921 | ||
| Other | 4,104 | 20,876 | 60,536 | 5,690 | 27,156 | 3,417 | 68,581 | ||
| Total exports 3/ | 32,154 | 39,619 | 78,180 | 19,166 | 55,992 | 8,440 | 151,073 | ||
| (In percent of total exports) | |||||||||
| Textiles and clothing 1/ | 23.0 | 16.0 | 3.5 | 12.1 | 7.3 | 35.8 | 24.6 | ||
| Agricultural and fishery products | 21.6 | 18.8 | 2.0 | 6.9 | 16.3 | 13.6 | 7.9 | ||
| Of which: Rice | 3.1 | … | -- | -- | 3.6 | 6.0 | -- | ||
| Gems and jewelry | 14.6 | 1.1 | 0.8 | 0.3 | 3.6 | … | 1.0 | ||
| Engineering goods 2/ | 13.9 | 10.7 | 12.8 | 48.5 | 16.8 | … | 14.7 | ||
| Chemicals and chemical products | 9.6 | 0.6 | 3.4 | 1.8 | 3.2 | 1.0 | 5.9 | ||
| Leather and leather goods | 4.5 | 0.2 | 0.1 | 0.6 | 0.7 | 3.1 | 0.6 | ||
| Total of principal exports | 87.2 | 47.3 | 22.6 | 70.3 | 47.9 | 53.5 | 54.6 | ||
| Memorandum items: | |||||||||
| Correlation coefficient 4/ | 1.00 | 0.32 | 0.08 | 0.36 | 0.68 | 0.80 | 0.72 | ||
| Bilateral exchange rate | |||||||||
| (percent change) 5/ | -- | 388.3 | 38.7 | 34.4 | 34.1 | −4.0 | −15.9 | ||
Includes footwear for Malaysia.
Includes electronics and computer software for India; includes electronics and electronics components for all other countries.
Non-oil/gas exports for Indonesia.
Correlation between the commodity shares of total exports of India and the respective country, based on 2-digit Standard International Trade Classification (SITC) data for 1993–95 from the TARS database. Coefficients close to one suggest that the two countries compete in the same commodities.
Percent change since June 30, 1997. Positive figures indicate appreciation of the rupee vis-à-vis the respective currency.
Selected Asian Economies: Principal Exports, 1996
| India | Indonesia | Malaysia | Philippines | Thailand | Pakistan | China | |||
|---|---|---|---|---|---|---|---|---|---|
| (In millions of U.S. dollars) | |||||||||
| Textiles and clothing 1/ | 7,408 | 6,329 | 2,767 | 2,328 | 4,066 | 3,021 | 37,150 | ||
| Agricultural and fishery products | 6,952 | 7,438 | 1,539 | 1,320 | 9,145 | 1,152 | 11,950 | ||
| Of which: Rice | 1,009 | … | -- | -- | 2,012 | 504 | 1 | ||
| Gems and jewelry | 4,707 | 423 | 610 | 62 | 2,042 | … | 1,444 | ||
| Engineering goods 2/ | 4,471 | 4,222 | 10,008 | 9,292 | 9,379 | … | 22,147 | ||
| Chemicals and chemical products | 3,072 | 244 | 2,678 | 353 | 1,792 | 87 | 8,879 | ||
| Leather and leather goods | 1,441 | 87 | 41 | 121 | 400 | 259 | 921 | ||
| Other | 4,104 | 20,876 | 60,536 | 5,690 | 27,156 | 3,417 | 68,581 | ||
| Total exports 3/ | 32,154 | 39,619 | 78,180 | 19,166 | 55,992 | 8,440 | 151,073 | ||
| (In percent of total exports) | |||||||||
| Textiles and clothing 1/ | 23.0 | 16.0 | 3.5 | 12.1 | 7.3 | 35.8 | 24.6 | ||
| Agricultural and fishery products | 21.6 | 18.8 | 2.0 | 6.9 | 16.3 | 13.6 | 7.9 | ||
| Of which: Rice | 3.1 | … | -- | -- | 3.6 | 6.0 | -- | ||
| Gems and jewelry | 14.6 | 1.1 | 0.8 | 0.3 | 3.6 | … | 1.0 | ||
| Engineering goods 2/ | 13.9 | 10.7 | 12.8 | 48.5 | 16.8 | … | 14.7 | ||
| Chemicals and chemical products | 9.6 | 0.6 | 3.4 | 1.8 | 3.2 | 1.0 | 5.9 | ||
| Leather and leather goods | 4.5 | 0.2 | 0.1 | 0.6 | 0.7 | 3.1 | 0.6 | ||
| Total of principal exports | 87.2 | 47.3 | 22.6 | 70.3 | 47.9 | 53.5 | 54.6 | ||
| Memorandum items: | |||||||||
| Correlation coefficient 4/ | 1.00 | 0.32 | 0.08 | 0.36 | 0.68 | 0.80 | 0.72 | ||
| Bilateral exchange rate | |||||||||
| (percent change) 5/ | -- | 388.3 | 38.7 | 34.4 | 34.1 | −4.0 | −15.9 | ||
Includes footwear for Malaysia.
Includes electronics and computer software for India; includes electronics and electronics components for all other countries.
Non-oil/gas exports for Indonesia.
Correlation between the commodity shares of total exports of India and the respective country, based on 2-digit Standard International Trade Classification (SITC) data for 1993–95 from the TARS database. Coefficients close to one suggest that the two countries compete in the same commodities.
Percent change since June 30, 1997. Positive figures indicate appreciation of the rupee vis-à-vis the respective currency.
F. Concluding Remarks and Policy Implications
20. This chapter examined from several perspectives the impact of recent currency realignments in the region on India’s competitiveness and on the outlook for exports. A number of alternative indicators of competitiveness as well as methodologies were used to gain a better understanding and to increase confidence in the findings. The results were consistent across different methodologies.
21. The results suggest that the rupee was not substantially overvalued in 1997/98. Based on available empirical estimates of export elasticities, the impact on export performance of the strength of the rupee in real effective terms in 1997/98 appears to be small. However, these results, based on backward-looking assessment, should be interpreted with caution given the inherent imprecision of econometric estimates, the reliance on a single indicator of competitiveness (namely a CPI-based REER), and the fact that more recent events that could have affected competitiveness are not yet fully captured in the data.
22. The empirical evidence confirms that the appreciation of the rupee in real effective terms since March 1993 is largely an equilibrium phenomenon reflecting movements in economic fundamentals, namely trade policy, the external environment, and domestic supply factors. Looking ahead, further progress in liberalizing the trade and exchange regimes would likely shift further the equilibrium REER.
23. The econometric results provide sufficient support to the hypothesis of Balassa-Samuelson effects, despite the lack of data on productivity for recent years as well as the limited degrees of freedom in the estimated equation that did not allow direct testing of the 1990s in isolation (see the annex). The results suggest that productivity gains in the tradeable goods sector stemming from the structural reforms implemented during the 1990s played a role in the appreciation of the REER of the rupee during the same period. External factors (as proxied by the terms–of–trade variable) were also found to belong to the long-run cointegrating relationship between the REER and its fundamental determinants.
24. The analysis suggests that a deliberate policy of engineering a nominal depreciation of the rupee would have only limited effects on exports in the short run. Instead, such a policy could potentially trigger another bout of pressure on the rupee due to a loss of confidence, and could eventually lead to an erosion of competitiveness through adjustments in domestic prices. Hence, the RBI’s policy of not resisting market-determined movements in the rupee, and of intervening only to prevent “disorderly” depreciation, seems appropriate in light of the empirical evidence. However, were significant and sustained downward pressure on the rupee to materialize, monetary policy should be the first line of defense to resist any overshooting of the exchange rate. Demands often expressed by market participants and exporters to target the level of the REER based on the REER in March 1993 or another “neutral” rate, if adopted, would risk loss of control over inflation.
25. Historically, large gains in export volume growth have been associated with rounds of structural reform and external liberalization such as in the mid-1980s and early 1990s. Given the evidence of a relatively small effect on export performance of movements in the rupee around its equilibrium level, improvements in competitiveness could be achieved through nonprice factors, including by addressing the structural factors affecting export performance, and by promoting a more quality-conscious export sector. In particular, bold reforms in the export-oriented small-scale sector would enhance competitive forces and efficiency in the economy, and would improve productivity by internalizing economies of scale. Moreover, the lifting of the existing ceiling on investment would allow innovation in product design and production methods that would help enhance competitiveness.
References
Agénor, P. R., and A.W. Hoffmaister (1996), “Capital Inflows and the Real Exchange Rate: Analytical Framework and Econometric Evidence,” IMF Working Paper WP/96/137.
Beveridge, S., and C. Nelson (1981), “A New Approach to Decomposition of Economic Time Series into Permanent and Transitory Components with Particular Attention to the Measurement of the Business Cycle,” Journal of Monetary Economics, Vol. 7, pp. 151–174.
Calvo, G. A., C. M. Reinhart, and C. A. Végh (1995), “Targeting the Real Exchange Rate: Theory and Evidence,” Journal of Development Economics, Vol 47(1), pp. 97–133.
Chinn, M. D. (1998), “Before the Fall: Were East Asian Currencies Overvalued?” mimeo, Department of Economics, University of California, Santa Cruz.
Clark, P. B., and R. MacDonald (1998), “Exchange Rates and Economic Fundamentals: A Methodological Comparison of BEERs and FEERs,” IMF Working Paper WP/98/67.
Desruelle, D., and A. Zanello (1997), “A Primer on the IMF’s Information Notice System,” IMF Working Paper WP/97/71.
Halpern, L. (1996), “Real Exchange Rates and Exchange Rate Policy in Hungary,” Centre for Economic Policy Research, Discussion Paper No. 1366.
Johansen, S. (1988), “Statistical Analysis of Cointegration Vectors,” Journal of Economic Dynamics and Control, Vol. 12, pp. 231–254.
Johansen, S. (1995), “Likelihood-based Inference in Cointegrated Vector Autoregressive Models,” (Oxford. Oxford University Press).
Marsh, I. W., and S. P. Tokarick (1994), “Competitiveness Indicators: A Theoretical and Empirical Assessment,” IMF Working Paper WP/94/29.
Patel, U. R., and P. Srivastava (1998), “Some Implications of Real Exchange Rate Targeting in India,” Indian Council for Research on International Economic Relations, Occasional Paper No, 5.
Reserve Bank of India (1993), “The Nominal Effective Exchange Rate (NEER) and the Real Effective Exchange Rate (REER) of the Indian Rupee,” Reserve Bank of India Bulletin, July 1993, pp. 967–977.
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, (New Delhi: Oxford University Press).
ANNEX: Determinants of India’s Long-Run Equilibrium Real Effective Exchange Rate
25. This annex describes a model of India’s long-run equilibrium REER based on annual data for 1979/80–1997/98, and examines the issue of misalignment of the rupee from an econometric point of view.
Conceptual issues
26. A number of factors are expected to affect the REER in the long run. Taking into account data availability, the treatment of these factors in this chapter is discussed below:
Trade policy is proxied by the sum of exports and imports over GDP.23 Widespread import controls can support an overvalued exchange rate. Downward adjustments of the REER to more realistic levels have been accompanied by relaxation of imports controls (especially of imports of essential capital goods and intermediate inputs). The relaxation of controls promotes export competitiveness and leads to a rise in both imports and exports.
The external environment is proxied by India’s terms of trade (defined as the ratio of India’s export unit values to import unit values). An improvement in the terms of trade (shown as a rise in the ratio) would have a positive impact on the current account and would lead to an appreciation of the REER.
Domestic supply factors are defined as the productivity differential between the tradable and nontradable goods sectors for India and its major trading partners proxied by per capita real growth in manufacturing less per capita real growth in services. This term measures the Balassa-Samuelson effect: that is, higher productivity growth in the tradable than in the nontradable goods sector (relative to partner countries) leads to rises in wages and prices in the nontradable goods sector, and to a real appreciation of the REER.
Fiscal policy is defined as the ratio of government consumption to GDP. Higher government consumption relative to GDP would put upward pressure on the price of nontraded goods, leading to an appreciation of the REER. Implicit in this argument is the assumption that government spending is concentrated on nontraded goods.
Capital inflows are defined as the ratio to GDP of the sum of; net direct investment in India; net inflows of portfolio investment; net increase in other investment; and errors and omissions.24 Higher capital inflows would exert an upward pressure on the REER.
Estimation issues
27. Estimation of the long-run cointegrating relationship between India’s REER and its fundamental determinants is complicated by a number of factors: (i) the limited availability and frequency of data for most of the explanatory variables prescribed by theory complicates analysis of policy variables as well as short-term factors, and reduces the precision of econometric estimates due to the small number of degrees of freedom; (ii) the existence of widespread nonprice controls for most of the sample period lowers the information content of the data; (iii) there are no data on sectoral productivity covering the full sample period that could potentially help explain a significant part of the appreciation of the rupee in real terms in recent years;25 and (iv) the gradual change in the exchange and trade regimes over the last two decades complicates the identification of structural breaks in the data and weakens the relation between policy variables and the exchange rate.
28. In an effort to better examine the role of policy variables and short-term factors, attempts were made to estimate the long-run cointegrating relationship based on quarterly data (some constructed from annual series). While the results (not reported here) broadly confirmed the findings described in this section, they nonetheless yielded perverse signs for some explanatory variables because of the artificial way some quarterly series had to be constructed. Specifically, the terms-of-trade series was derived from the annual data, with the quarterly pattern of the export price index for India replicating the quarterly pattern of India’s WPI, and the quarterly pattern of the import price index replicating the quarterly pattern of the weighted export price index of India’s major trading partners. Similarly, the industrial production index was used to construct quarterly series for productivity growth in the manufacturing sector, while productivity growth in the services sector was derived by splicing the annual series.
Model
29. The long-term determinants of India’s REER were estimated using annual data over 1979/80–1997/98 for: trade policy (proxied by the sum of exports and imports over GDP); a proxy for the external environment (India’s terms of trade); and domestic supply factors (productivity differential between the tradable and nontradable goods sectors for India and its major trading partners proxied by per capita real growth in manufacturing less per capita real growth in services).26 27 All data were expressed in logarithms, except for the variable for the domestic supply factors, which included negative values.
Integration
30. To determine the appropriate estimation procedure, tests for nonstationarity of the above variables were carried out using the Augmented Dickey-Fuller (ADF) method, which looks for the presence of unit roots in the series. Table III.3 lists the fourth-order ADF statistics for the REER, trade policy (EXIM), terms of trade (TOT), and domestic supply factors (B-S). According to the tests, the null hypothesis of nonstationarity—that is, each variable has a unit root, or equivalently, is integrated of order one: I(1)—cannot be rejected for any of the variables. All variables are thus stationary in first differences, and cointegration analysis among the level variables is required.
ADF(4) Statistics Testing for a Unit Root
ADF(4) Statistics Testing for a Unit Root
| Variable | ||||
|---|---|---|---|---|
| REER | EXIM | TOT | B-S | |
| Ho: I(1) | −2.85 | −1.14 | −0.85 | −1.92 |
ADF(4) Statistics Testing for a Unit Root
| Variable | ||||
|---|---|---|---|---|
| REER | EXIM | TOT | B-S | |
| Ho: I(1) | −2.85 | −1.14 | −0.85 | −1.92 |
Cointegration
31. The Johansen procedure is used for the cointegration analysis.28 The Johansen procedure is a full information maximum likelihood estimation for vector autoregressive systems, and, as such, it is not concerned about the endogeneity of the explanatory variables. Nonetheless, the procedure imparts a heavy toll on the degrees of freedom and on the precision of the econometric estimates in small samples because it uses a lag structure. The procedure searches for the existence of one or more long-run cointegrating relationships between the selected variables. Table III.4 reports the statistics from the Johansen procedure. Only the final specification is reported. The procedure found evidence of the following long-run relationship between the fundamental variables:29
REER = -1.34369(EXIM)+0.3753(TOT)+0.5784(B-S)
Johansen Tests of Existence of Long-Run Relationships
Single and double asterisks denote significant test statistics at the 5 percent and 1 percent level, respectively.
Johansen Tests of Existence of Long-Run Relationships
| λ-max 1/ | Trace 1/ | ||||
|---|---|---|---|---|---|
| Ho: rank=0 | 29.77 | * | 59.19 | ** | |
| <=1 | 16.00 | 29.41 | |||
| <=2 | 10.32 | 13.42 | |||
| <=3 | 3.09 | 3.09 | |||
| Variable | |||||
| REER | EXIM | TOT | B-S | ||
| Standardized feedback coefficients α | |||||
| 0.0090 | 0.0766 | 0.3079 | 0.0063 | ||
Single and double asterisks denote significant test statistics at the 5 percent and 1 percent level, respectively.
Johansen Tests of Existence of Long-Run Relationships
| λ-max 1/ | Trace 1/ | ||||
|---|---|---|---|---|---|
| Ho: rank=0 | 29.77 | * | 59.19 | ** | |
| <=1 | 16.00 | 29.41 | |||
| <=2 | 10.32 | 13.42 | |||
| <=3 | 3.09 | 3.09 | |||
| Variable | |||||
| REER | EXIM | TOT | B-S | ||
| Standardized feedback coefficients α | |||||
| 0.0090 | 0.0766 | 0.3079 | 0.0063 | ||
Single and double asterisks denote significant test statistics at the 5 percent and 1 percent level, respectively.
32. The estimation found evidence of a long-run cointegrating relationship between India’s REER and its fundamental determinants. All estimated coefficients have the anticipated signs. The feedback coefficients suggest a relatively slow adjustment from disequilibrium. The estimated equation is used to derive the equilibrium REER. Misalignment is then defined as the difference between the actual REER and the antilogarithm of the predicted value of the REER derived from the equation above (that is, the equilibrium REER), expressed in percent of the equilibrium REER. Chart III.7 shows that most of the appreciation of the rupee in real terms since 1993 has been an equilibrium phenomenon. The estimated misalignment of the rupee in 1997/98 was 2–3 percent.30
Prepared by Dimitri Tzanninis.
The Balassa-Samuelson effect describes the process by which higher productivity growth in the tradable than in the nontradable goods sector (relative to partner countries) leads to wage and price increases in the nontradable goods sector, and to a real appreciation of the currency.
India’s fiscal year starts April 1.
However, it is difficult to detect any Balassa–Samuelson effects during the 1980s due to the overvaluation of the currency and the prevailing price and trade controls.
The effective import-weighted tariff rate has been lowered from 87 percent in 1990/91 to about 30 percent in 1997/98.
It is reasonable to view the REER in March 1993 as consistent with a stable equilibrium in the external position. The unification of exchange rates set the rupee at a competitive and sustainable level from the point of view of movements in both domestic prices and the external current account; since then, the net flow of foreign assets and the external debt have followed sustainable paths.
Countries with bilateral competitive weights below one percent are excluded from the calculations of effective exchange rates in the INS. For a description of the methodology and the data used to compute the effective exchange rate indices in the INS, see Desruelle and Zanello (1997).
The term “CPI-based REER” is used to reflect the use of consumer prices for partner countries in the calculation of the REER. For India, however, rather than the CPI, the wholesale price index (WPI) is used in the INS calculations because it is the widest and most timely price indicator.
The RBI also uses the WPI for India in its calculations of a CPI-based REER. See Reserve Bank of India (1993).
There has been a gradual increase in the share of Asian countries in India’s exports in recent years. The share of all Asian countries in India’s exports increased from 24 percent in 1990/91 to 29 percent in 1996/97, with the trend broadly continuing in 1997/98 despite declining exports to the countries primarily affected by the financial crisis. In 1997/98, exports to the subset of Southeast and East Asian countries accounted for about 20 percent of India’s exports, while imports from those countries accounted for about 15 percent of total imports.
The cumulative weights of the countries of the European Union in the RBI’s REER index are 13.9 percentage points lower than their cumulative weights in the INS. Moreover, the weight of the U.S. is 3 percentage points higher than its weight in the INS.
For a discussion of the advantages and limitations of various price and cost indices in calculating REERs, see Halpern (1996), and Marsh and Tokarick (1994).
The effect is more pronounced in the mid-1980s and early 1990s, when more drastic changes in India’s trade and exchange regimes took place. Such an effect is more likely to be important when examining countries at different levels of development, such as India and its major trading partners (the G-7 countries). The existence of the Balassa-Samuelson effect is tested directly in section D.
The measurement of the extent of misalignment is complicated by the fact that the “equilibrium” REER is an unobservable variable. In this chapter, misalignment is defined as the deviation (in percent) of the actual REER from its estimated equilibrium level. In interpreting the results, it is worthwhile noting that the notion of equilibrium used is a statistical one rather than one that would be consistent with macroeconomic balance as defined by economic theory.
The HP filter is a generalization of the trend-stationary hypothesis used in the literature on real exchange rates. Although the HP filter is devoid of theoretical underpinnings and gives excessive weight to the most recent observations (the end-period problem), it is often used because of its computational simplicity. To avoid the end-period problem, actual data for the REER through June 1998 were used but the results were truncated at end-March 1998.
See Agénor and Hoffmaister (1996), and Clark and MacDonald (1998) for the application of HP filters to estimate the extend of misalignment of REERs. Chinn (1998) used a simplified version of the trend-stationary hypothesis by employing a linear deterministic trend to calculate equilibrium REERs.
See Beveridge and Nelson (1981). Several studies have used this method mainly to compute the extent of misalignment of REERs. See, for example, Agénor and Hoffmaister (1996), and Calvo et al (1995). Patel and Srivastava (1998) used this method to compute the correlation between the temporary component of India’s REER and cyclical impulses such as inflation.
While keeping in mind the inherent imprecision of econometric estimates, the somewhat higher degree of misalignment found with this method apparently captures the aggregation effect of annual data, which are unable to identify turning points within a year. By the other two methods employed to measure the degree of misalignment, the overvaluation of the rupee in the first half of 1997/98 had, to a significant extent, been corrected by end-March 1998.
The estimated overvaluation (1–3 percent by all three methods used) in 1997/98 is not large relative to past deviations of the REER from its equilibrium. Historically, misalignments of the rupee have tended to correct themselves within two years.
Using annual data for India, Patel and Srivastava (1998) were also unable to establish a long-run link between fiscal policy (deficit-to-GDP ratio) and the REER.
For example, net portfolio and foreign direct investment inflows—which have accounted for the bulk of capital inflows in recent years—amounted to 1.2 percent of GDP in 1997/98, substantially lower than most Asian economies.
Srinivasan estimated the elasticity of India’s exports with respect to the REER in the range of (−0.304, −0.278) in the short run and (−0.537, −0.493) in the long run.
Reliable data on the extent of quantitative restrictions on imports as well as on other nontariff barriers are not available for the full sample period.
Capital inflows rather than a real interest rate differential was chosen as a fundamental determinant of India’s REER. This choice was dictated by the fact that India maintained a relatively closed capital account for most of the period under examination that did not allow the prevailing high interest rate differential to manifest itself in large-scale capital inflows.
Available data on sectoral productivity do not extend beyond 1993/94.
Preliminary regressions did not find evidence of fiscal variables (government consumption to GDP, and government investment to GDP) belonging to the long-run relationship. While this result may be consistent with fiscal policy having only a short-run effect on the REER, it may also indicate that the variables used were poor proxies for testing the hypothesis that government expenditure is relatively concentrated on nontraded goods.
A common cointegrating relationship could not be obtained when the capital inflows variable was included. By contrast, Patel and Srivastava (1998) were able to obtain such a relationship for India, but they did not elaborate on how the capital inflows variable was defined.
Estimation and testing were carried out in PcFiml. The small sample size (19 annual observations) was sufficient for the Johansen procedure to be conclusive despite the loss of degrees of freedom due to the lag structure of the estimated system. Given that the fewer the degrees of freedom the harder it is to reject the null hypothesis of no cointegrating relationship, the reported results are thus relatively powerful.
It is unclear to what extent the large misalignment of the rupee in 1991/92 and 1992/93 (around 13 percent) reflects actual misalignment or a statistical error in the standard econometric sense (that is, the combined effect of temporary and random factors). This period coincides with a sharp depreciation of the rupee in nominal terms that might have overshot the long-run equilibrium rate.
