II. External Competitiveness and the Real Exchange Rate in Bangladesh9
Bangladesh’s exports of ready-made garments (RMGs) have continued to perform well despite the global financial crisis. And there appears to be scope to further boost Bangladesh’s share of developed countries’ imports of RMGs.
The taka appears to be somewhat undervalued. However, various approaches to assess the equilibrium levels of the current account and the taka suggest that the REER and the current account balance remain broadly in line with macroeconomic fundamentals.
23. The strength of Bangladesh’s balance of payments in FY2009 and the first half of FY2010 raises the question as to the appropriateness of the level of the real effective exchange rate (REER) of the taka. The trend depreciation of the REER for much of the past 30 years was halted only a few years ago. The ongoing strength of remittances and the narrowing trade deficit may suggest that Bangladesh could sustain a more appreciated REER.
24. This chapter discusses recent trends in the current account and merchandise exports (Section B) and presents estimates of the equilibrium REER of the taka (Section C). In this context, the chapter highlights that the elasticity of the current account with respect to the real exchange rate—a key variable for the estimation of the extent of over/undervaluation of the taka—varies with the composition of trade, the degree of pass-through of changes in exchange rates and international prices to import and export prices and Bangladesh’s price-taking behavior are important factors. More work is needed to better understand these factors and to firm up low-income countries’ exchange rate assessments. The role of remittances in the assessment needs to be investigated further as well.
B. Trends in the Current Account and Merchandise Exports
25. Bangladesh’s current account strengthened substantially in FY2009 and looks set to remain strong over the medium term, albeit with a narrowing surplus:
The current account improved by almost 2 percent of GDP to a surplus of 2.8 percent of GDP in FY2009. The growth of merchandise imports decelerated from 26 percent in FY08 to 4 percent in FY09, owing to an abundant agricultural harvest, lower international oil and commodity prices, a deceleration of credit to the private sector, and a wait-and-see attitude in the private sector which depressed the import of capital machinery. Combined with the resilience of exports this caused the trade balance to improve by almost 1½ percent of GDP in FY2009. The inflow of private transfers, including remittances, increased by 0.8 percent of GDP. And the income balance weakened slightly.
The current account surplus is projected to remain at close to 3 percent of GDP in FY2010. However, in later years the surplus will likely narrow, to less than 1 percent of GDP in the medium term. A further narrowing of the merchandise trade deficit is projected in FY2010, reflecting continued large year-on-year declines in merchandise imports during the first quarter of FY2010. With more favorable base effects and with a gradual return of confidence in the aftermath of the global financial crisis, the year-on-year change in imports should turn positive from November onwards and this can be expected to bring down the current account surplus over time through a widening of the merchandise trade deficit. Remittances are projected to grow slower than nominal GDP starting in FY2010 reflecting the slowdown in the hiring of Bangladeshi workers since the summer of 2008 (see Section D in the previous Chapter).
26. Bangladesh broadly maintained its share of global exports between 2000 and 2008, owing to the strong performance of the ready-made garments (RMG) sector. Over the past 10 years, Bangladesh has managed to steadily raise its share of industrialized countries’ imports of knitted garment and woven garments (Table 1 and 2).
|In percent of total EU kntted garment imports|
|In percent of total US kntted garment hiports|
|In percent of total Canadian knitted garment imports|
|In percent of total Japanese knitted garment imports|
|Size of the market (in bilhns of U.S. dollars)|
|(Annual percent change)||9.0||14.6||12.9||-11.0|
|(Annual percent change)||5.8||B.8||-1.5||-11.1|
|(Annual percent change)||10.7||16.5||11.3||-7.0|
|(Annual percent change)||6.1||4.6||9.6||5.9|
Data for Canada is for January-August of respective years rather than H1.
Data for Canada is for January-August of respective years rather than H1.
|In percent of total EU woven garment imports|
|In percent of total US woven garment imports|
|In percent of total Canadian woven garment imports|
|In percent of total Japanese woven garment imports|
|Size of the market (h billions of U.S. dollars)|
|(Annual percent change)||6.9||10.5||8.6||-12.5|
|(Annual percent change)||4.6||0.1||-5.0||-14.0|
|(Annual percent change)||10.4||10.2||4.7||-7.4|
|(Annual percent change)||4.7||1.5||5.2||2.4|
Data for Canada is tor January-August of respective years rather than H1.
Data for Canada is tor January-August of respective years rather than H1.
Bangladesh’s share of global exports
Source: Direction of Trade Statistics
27. Non-RMG exports have not been able to keep pace with RMG exports. This seems to have prevented Bangladesh from keeping pace with some of its more successful regional competitors. In particular, India and Vietnam managed to increase their share of global exports in recent years thanks to their more diversified production base, more business-friendly macroeconomic and structural policies, and better-quality governance (See also last year’s SIP on Export Diversification and External Competitiveness (IMF (2008)).
Trends in world exports market share for selected Asian countries, 2008
28. RMG exports also held up better than non-RMG exports during the global crisis. RMG exports, which make up more than 75 percent of total exports, recorded a growth rate of 15 percent in FY09, mostly owing to strong performance during the second half of 2008. However, it is important to note that the ongoing gains in the global market share of Bangladesh’s RMG sector accelerated during the first half of 2009 thanks to relative strength in the lower market segment (Table 1 and 2). By contrast, Bangladesh’s other key exports (frozen food and shrimp, and leather and jute products) contracted, causing non-RMG exports to decline by 6 percent in FY09.
Bangladesh: Composition of Exports
Source: Bangladesh Bank
29. The outlook for the RMG sector remains favorable. Production capacity has continued to expand and there are signs that Japanese retailers, which have hitherto mostly sourced from China and Vietnam, are starting to source more of their textiles and garments from Bangladesh. Gaining market share in Japan would provide a considerable boost. A slight undervaluation of the taka would help exporters to increase their market share and even the playing field with competitors in other countries who may be enjoying more business-friendly macroeconomic and structural policies, and better-quality governance.
C. Estimates of the Equilibrium Real Effective Exchange Rate
30. Estimates of the equilibrium real effective exchange rate (ERER) tend to be quite sensitive to the methodology used and are particularly challenging in developing countries where the data are weaker (see, e.g., Dunaway and others (2006) and Di Bella and others (2007)). This section assesses if Bangladesh’s REER is in line with macroeconomic fundamentals. Four methodologies are used: (i) the purchasing power parity approach (PPP), (ii) the macroeconomic balance (MB) approach; (iii) the equilibrium real exchange rate (ERER) approach, and (iv) the external sustainability (ES) approach.
31. Overall, application of four different approaches suggests that the real exchange rate of the taka is broadly in line with its long-run equilibrium value (Table 3). The PPP approach suggests that the taka is somewhat overvalued. The ERER approach suggests that the taka is somewhat undervalued. The ES and MB approach also suggest that the taka is undervalued: Bangladesh’s current account balance projected for the medium term is higher than the current account “norm” derived according to the MB approach; it is also higher than the current account balance that would stabilize Bangladesh’s NFA to GDP ratio over the medium term. The extent of undervaluation in the case of the ES and MB approach crucially depends on the estimated level of the elasticity of the current account with respect to the real exchange rate (Appendix I). Applying assumptions used to derive this elasticity for developed countries suggests a relatively small elasticity. Accordingly, a real appreciation of more than 20 percent would be needed to drive the current account down toward the norm or the level that stabilizes the NFA to GDP ratio. By contrast, applying assumptions that are more fitting for the case of Bangladesh suggests a larger elasticity and hence an undervaluation of the taka of about 10 percent.
|Approach||Estimated overvaluation (in percent)|
|Purchasing power parity||8|
|External sustainability||-24 to -10|
|Macro balance||-23 to -9|
|Equilibrium exchange rate||-8|
Purchasing Power Parity Approach
32. One way to assess the deviation of a country’s real exchange rate from its long-run level is through an international comparison of price levels. According to the theory of purchasing power parity (PPP), prices of an identical consumption basket should be the same in all countries once expressed in a common currency. However, a comparison of consumer prices across countries also needs to account for nontraded goods and differences in consumption baskets. And nontradables prices tend to rise with countries’ incomes. The strong correlation between absolute price levels and income is usually attributed to the Balassa-Samuelson (B-S) effect, which relates the long-run real exchange rate to relative productivity differentials. Countries with relatively higher productivity growth in the tradables relative to the nontradables sector (compared to trading partners) tend to experience real appreciation (increase in the relative price of nontradables to tradables). In essence, higher tradables productivity pushes up wages in the tradables sector, which leads to higher wages in the nontradables sector and, consequently, to higher nontradables prices. Since tradables prices are set in international markets and do not respond to domestic market conditions, the relative price of nontradables to tradables rises in these circumstances. As the overall price level is a weighted average of tradables and nontradables prices, the higher price of nontradable goods leads to an increase in the overall CPI. Assuming that real per capita GDP differentials across countries are a reasonable proxy for relative productivity differentials, the B-S effect implies a positive correlation between relative income levels and the REER. It also suggests that as a country’s (relative) income level rises over time, its REER will appreciate.
33. A database compiled by the 2005 International Comparison Program (ICP) suggests that the taka was valued broadly in line with its purchasing power parity in 2005. The database was published in the context of the December 2007 release of the report of the 2005 ICP.10 A standard cross section regression of the price level in 146 countries relative to that in the United States on a constant and per capita GDP indicates that for every 1 percent increase in a country’s per capita GDP, its price level relative to the United States is higher, and hence its REER is more appreciated, by about 0.22 percent. This database suggests that the price level in Bangladesh in 2005 was in line with what could be expected based on the level of Bangladesh’s per capita GDP.
Cross-section Regression of Price Levels (Relative to the United States) on Per Capita Income
Source: 2005 International Comparison Program, prelimilllliyresillts. Price levels in each ooillltiy are praued by Ihe GDP deflator.
34. A tentative extension of the PPP approach suggests that, as of end-2008, the taka was overvalued by about 8 percent.11 Reflecting the substantial inflation differential with the U.S. and the stability of the nominal U.S. dollar/taka-exchange rate, by end-2008, the bilateral real exchange rate vis-à-vis the U.S. dollar had appreciated by some 10 percent compared to the average for 2005. On that basis Bangladesh would now be located above the regression line that links countries’ price levels to the their level of real per capita GDP. Because of Bangladesh’s strong growth of per capita real GDP, it has also moved to the right in the graph. In all, the taka would now be 8 percent overvalued (measured by the vertical distance from the regression line).
The Macro Balance (MB) Approach
35. The MB approach calculates the difference between an estimated equilibrium current account balance (the “norm”) and the projected current account balance. The REER adjustment that would close the gap between the estimated norm and the projected current account is then calculated using the country’s current account elasticity with respect to the real exchange rate (Appendix I).
36. A model for the current account norm fitted to an unbalanced panel of 101 developed and developing countries suggests that the taka is undervalued by between 10 and 24 percent, depending on the assumed elasticity of the current account with respect to the real effective exchange rate.12 We added Bangladesh to a large panel of developed and developing countries put together by Vitek (2009). We replicated his estimation of the current account balance as a function of relative old age dependency, relative income, relative income growth, the oil trade balance, the relative fiscal balance, initial net foreign assets, aid inflows, and the inflow of remittances (Table 4):
|Relative old age dependency||-0.11||-0.97|
|Relative income growth||-0.52||-2.23|
|Dummy* Relative income||0.00||-0.22|
|Dummy* Oil-trade balance||1.34||1.45|
|Relative fiscal balance||0.28||2.17||0.23||1.44|
|Dummy* Relative fiscal balance||-0.36||-0.95|
|Initial net foreign assets||0.05||3.60||0.05||3.01|
|Dummy* Initial net foreign assets||0.03||0.38|
|Dummy* Aid inflows||1.15||1.68|
|Dummy* Remittances inflows||1.00||1.60|
|Wald test for dummies (prob)||0.0002|
|Number of observations||1085||1085|
The estimated coefficients for the explanatory variables are broadly in line with what would be expected based on economic theory. The current account norm rises with the oil trade balance, the initial net foreign assets position and the fiscal balance. The norm declines with higher relative income growth.
The estimation suggests that the coefficients for remittances and aid flows are insignificant at commonly used levels of statistical significance. This result contradicts some recent studies (e.g. Halikias (2009)) which found that remittances are a statistically significant determinant of the current account norm.
The panel estimation results imply a medium-term current account norm of -2 percent of GDP. This is substantially weaker than the staff projection (a surplus of 0.8 percent of GDP in the medium term), suggesting undervaluation of the taka by between 10 and 24 percent.
37. Inflows of remittances and aid are relatively large for countries in South and East Asia and turn out to be significant determinants of the current account norm. The general panel estimation suggests that aid and remittances are significant only at a 35 percent level. To test whether countries in South and East Asia behave differently we include an Asia dummy for the relevant countries in the panel (Bangladesh, Cambodia, India, Indonesia, Malaysia, Maldives, Myanmar, Pakistan, and Philippines) comprising about 250 observations. A Wald test rejects the null hypothesis that the estimated coefficients for the interaction terms with the Asia dummy are jointly equal to zero. The estimates suggest that remittances, aid inflows, and the oil trade balance have larger and more significant effects on current account norms in East- and South-Asian countries than for the rest of the world. This may reflect a tendency to add a larger share of these foreign exchange inflows to central bank reserves than is the case in countries in other regions.
38. For the case of Bangladesh, taking into account the Asia-specific determinants of the current account norm would imply a negligible undervaluation of the taka. The estimation results for the panel with the Asia dummy entail a current account norm for Bangladesh of a surplus of 0.5 percent of GDP, almost equal to the staff’s medium-term projection, suggesting that the taka is in line with macro-economic fundamentals.
The Equilibrium Real Exchange Rate (ERER) Approach
39. The ERER approach involves carrying out a panel regression for the REER and its determinants. As with the MB approach, we used the database and program setup by Vitek (2009). The REER is estimated as a function of the terms of trade, relative productivity, relative government consumption and remittance flows (Table 5). The estimated coefficient are highly significant and their signs are in line with priors based on economic theory.13 We used Bangladesh’s real GDP per capita relative to its trading partners as a proxy for relative productivity. Through the mid-1990s, this measure of relative productivity was adversely affected by political instability, natural disasters, and an inward-oriented development model. Since then, as a result of improved macroeconomic policies and a gradual move toward a more outward-oriented development model, Bangladesh has started to make some gains relative to its trading partners.
Bangladesh’s real GDP per capita relative to trading partners
|Terms of trade||0.32||3.59||0.32||3.51|
|Dummy* Terms of trade||0.02||0.04|
|Dummy* Relative productivity||-0.24||-0.61|
|Relative government consumption||0.21||2.03||0.20||1.98|
|Dummy* Relative gvmt consumpt||-0.26||-0.29|
|Wald test for dummies (prob)||0.37|
|Number of observations||3098||3098|
40. The ERER approach suggests that the taka is 8 percent undervalued. Remittances are a statistically significant determinant of the REER and no region-specific effects of explanatory variables on the ERER were found. The panel used for the estimation of the ERER approach is substantially larger than the one used for the MB approach and includes considerably more developing countries, including ones that receive substantial remittances. The resulting larger tilt to (remittances-receiving) developing countries than in the smaller panel used to apply the MB approach may account for the significant effect of remittances in the basic regression and the insignificance of the interaction terms of the regional dummy variable with explanatory variables.14
The External Sustainability (ES) Approach
41. The ES approach involves estimating the adjustment in the REER needed to stabilize Bangladesh’s NFA-to-GDP ratio at a certain benchmark level. The ES approach complements the ERER methodology by focusing on the relation between the sustainability of a country’s external stock position and its flow current account position, trade balance, and real exchange rate. It consists of three steps. The first involves determining the current account balance to GDP ratio that would stabilize the NFA position at a given ‘benchmark’ value. The second step compares this NFA-stabilizing current account balance (ES-norm) with the level of a country’s underlying current account balance. And finally, the third step consists of assessing the adjustment in the real effective exchange rate that is needed to close the gap between the underlying current account balance and the ES-norm.
Bangladesh: NFA, gross foreign assets and liabilities
42. The ES approach suggests that the REER was undervalued by between 9 and 23 percent in 2008. Using data compiled by the Bangladeshi authorities on the country’s international investment position (IIP), Bangladesh’s NFA in 2008 is put at -27 percent of GDP. With the U.S. dollar value of Bangladesh’s GDP projected to grow by 7.2 percent annually, Bangladesh could run a current account deficit of 1.8 percent of GDP without causing the NFA to GDP ratio to fall. In staff’s baseline scenario, the current account surplus is projected at 0.8 percent of GDP in 2014. Depending on the elasticity of the current account balance with respect to the REER, the gap between the NFA-stabilizing current account deficit and the projected medium-term current account surplus could be closed by a 9 to 23 percent appreciation of REER.
The elasticity of the current account with respect to the real exchange rate is a key variable for the estimation of exchange rate misalignments under the MB and ES approaches. The last step in the estimation of the over/undervaluation of the real exchange rate involves computing the real exchange rate adjustment that would (i) close the gap between the estimated current account norm and the underlying current account balance (MB approach); or (ii) close the gap between the estimated NFA-stabilizing current account balance and the underlying current account balance (ES approach).
The IMF’s Consultative Group on Exchange Rates (CGER, see Lee et al. (2008)) calculates the current account elasticities for a panel of industrial and emerging market economies. The calculations are based on estimated import and export elasticities which are derived on the back of the assumption that supplies of exports and imports are perfectly elastic. Complete pass-through of changes in foreign prices to domestic prices is assumed as well. Accordingly, the current account elasticity is calculated as (export elasticity) x (export to GDP ratio) – (import elasticity – 1) x (import to GDP ratio): for a given response of export and import volumes to the real exchange rate, the impact on the trade balance and the current account will be roughly proportional to trade openness. Therefore, a country more open to trade will be able to close the current account gap with less exchange rate adjustment. Using data for FY2009, the CGER-based current account elasticity for Bangladesh is equal to 0.12 if the elasticities estimated for industrial countries are used.
The application of uniform export and import elasticities to a wide range of countries has raised some questions, in particular, given differences in the composition of exports and imports across countries. In particular, several studies suggest that developing countries may have larger trade elasticities. Therefore, similar current account gaps could be closed by smaller adjustments in the real effective exchange rate. Several attempts have been made to calculate trade balance elasticities for low-income countries and commodities exporters. However, these attempts have been hampered by data limitations and structural changes—which have been frequent in developing countries and limit the scope for using common econometric techniques.
Using current account elasticities for low-income countries and allowing for incomplete pass-through to import and export prices renders a current account elasticity for Bangladesh which is close to the CGER-based elasticity. An elasticity of 0.10 is obtained if one corrects for the incomplete pass-through observed in many low and middle income countries and applies the export and import elasticities estimated by Hakura and Billmeier (2008) for a sample of non-oil exporting low and middle income countries in the Middle East and Central Asia.
|Pass-through to exports||1||0.42||0.66|
|Pass-through to imports||-1||-0.71||-0.66|
|Trade balance elasticity||-0.11||-0.10||-0.29|
Export and import-shares in GDP are set at 19 and 25 percent, respectively, as projected for 2014.
Export and import-shares in GDP are set at 19 and 25 percent, respectively, as projected for 2014.
A new approach to calculate current account balance elasticities suggests that Bangladesh’s current account elasticity with respect to the real exchange rate may be 0.29, substantially larger than the CGER- based elasticity and the adjusted CGER-based elasticity. Tokarick (2009) presents a new approach to calculate country-specific trade elasticities, based on a general equilibrium model. He develops an international-trade model with three different goods (an exported good, a good which competes with imports, and a non-traded good) and calculates elasticities based on the detailed Global Trade Analysis Project database. The analytical approach to the computation of current account elasticities is not hampered by the data and structure constraints of the regression analysis, which, as indicated above, is especially important for developing countries.
The composition of trade and price setting behavior are important factors:
The importance of the composition of exports is underscored by the relatively strong performance of low value-added exports, including in the case of Bangladesh, during the current global crisis.
Bangladesh, like many other developing countries, is a price taker on international markets. Therefore, “small country assumptions” (i.e. infinite export demand and import supply elasticities) need to be applied in the formula for the current account balance elasticity. Moreover, the computations need to focuses on the elasticity of the trade balance in foreign currency rather than of the trade balance in domestic currency as in the CGER (as trade data in Bangladesh is reported in foreign currency). Accordingly, the current account elasticity is calculated as (export elasticity)* (export-to-GDP share)*(pass-through to export prices) - (import elasticity)*(import-to-GDP share)*(pass-through to import prices). Estimates of pass-through of 0.66 for developing and emerging markets are taken from Frankel, Parsley, and Wei (2005).
Di BellaGabrielMarkLewisAurélieMartin2007 “Assessing Competitiveness and Real Exchange Rate Misalignment in Low-Income Countries” IMF Working Paper No. 07/201 (Washington: International Monetary Fund).
DunawayS. V.LeighL.LiXiangming2006 “How Robust are Estimates of Equilibrium Real Exchange Rates: The Case of China” IMF Working Paper No. 06/220 (Washington: International Monetary Fund).
FrankelJ.ParsleyD. and WeiS.-J.2005 “Slow Pass Through Around the World: A New Import for Developing Countries?” Working Paper No. 11199 (Cambridge: National Bureau of Economic Research).
HakuraDalia S. and BillmeierAndreas2008 “Trade Elasticities in the Middle East and Central Asia: What is the Role of Oil?” IMF Working Paper No. 08/216 (Washington: International Monetary Fund).
HalikiasIoannis2009 “Remittances and the Macrobalance Approach to Exchange Rate Assessment” in Philippines: Selected Issues IMF Staff Country Report No. 08/335 (Washington: International Monetary Fund).
IMF2008Bangladesh: Selected Issues,IMF Staff Country Report No. 09/63 (Washington: IMF) also at http://www.imf.org/external/pubs/cat/longres.cfm?sk=22420.0
LeeJ.G. M.Milesi-FerrettiJ.OstryA.PratiL. A.Ricci2008“Exchange Rate Assessments: CGER Methodologies,”IMF Occasional Paper No. 261 (Washington: International Monetary Fund).
TokarickS.2009A Method for Calculating Export Supply and Import Demand Elasticities,mimeo.
Prepared by Geert Almekinders and Svitlana Maslova.
The ICP’s global report for 2005 is based on an international effort to collect comparative price data and estimate purchasing power parities (PPPs) for 146 economies, benchmarked to the year 2005. The new PPPs, which are based on national surveys that priced nearly 1,000 products and services, replace previous benchmark estimates, some dating back to the 1980s. Comparative price levels are also included.
The estimated overvaluation of the taka may be exaggerated to the extent that per capita GDP is underestimated. Bangladesh’s national accounts are based on an outdated base year (1995/96) and structural changes that have taken place since then are not adequately reflected in GDP estimates. The national accounts base year was last updated in April 2000, from 1984/85 to 1995/96. At that time, the level of GDP was revised upward by about 30 percent. Higher per capita GDP would cause a shift to the right and move Bangladesh closer to the regression line, implying a smaller overvaluation or even some undervaluation of the taka.
A single-country estimation of the current account norm for Bangladesh also produces plausible results with coefficients with the “correct” sign. This approach suggests a 1 to 3 percent undervaluation in 2014. However, the robustness of this result is hampered by limited availability of data (only 20 annual observations) and structural breaks in the data.
Estimates of a single-country vector error correction model give implausible results (“wrong” signs for remittances and openness) possibly caused by structural breaks in the data and exchange controls.
Based on the Wald test we could not reject the null hypothesis that the estimated coefficients for the explanatory variables interacted with the dummy variables are jointly equal to zero.