This Selected Issues paper on Bangladesh underlies the export performance of readymade garment industry and inflation dynamics. Bangladesh has demonstrated that it is highly competitive in the world’s major garment markets. Inflation inertia, monetary factors, and exchange rate fluctuations are the main determinants of inflation in Bangladesh. Despite adoption of numerous tax policy measures during the past few years, policies implemented by the Bangladesh authorities have not been fully successful in lifting the revenue ratio to a level warranted by developmental objectives.

Abstract

This Selected Issues paper on Bangladesh underlies the export performance of readymade garment industry and inflation dynamics. Bangladesh has demonstrated that it is highly competitive in the world’s major garment markets. Inflation inertia, monetary factors, and exchange rate fluctuations are the main determinants of inflation in Bangladesh. Despite adoption of numerous tax policy measures during the past few years, policies implemented by the Bangladesh authorities have not been fully successful in lifting the revenue ratio to a level warranted by developmental objectives.

VI. The Cyclical Properties of Workers’ Remittances1

A. Introduction

1. The spectacular rise of workers remittances has sparked a wide interest in their macroeconomic properties. Remittances receipts in Bangladesh increased by 17 percent, on average, in each of the last 30 years and meanwhile constitute the largest source of foreign exchange after exports. Remittances are a particularly attractive source of foreign financing, because they are usually more stable over time than other capital flows. In addition, they are unrequited transfers, which unlike other capital flows, do not create obligations in the future.

2. Many believe—and some evidence has emerged—that remittances serve as an insurance against economic shocks. This would be good news for Bangladesh which is prone and vulnerable to a number of shocks: The export base is very narrow with 75 percent of exports concentrated in the ready-made garment sector; as an oil importer Bangladesh is vulnerable to swings in the international price; and since almost half of its territory is only inches above, or even below, sea level the country is prone to natural disasters.

3. This chapter explores to what extent remittances absorb shocks and reduce volatility in Bangladesh. It concludes that remittances are a welcome source of foreign financing and should be promoted, including through financial deepening. It also finds that remittances do respond to swings in the international oil price and help cushion its impact on the Bangladesh economy. However, there is little evidence that remittances respond to adverse domestic shocks and would come to the rescue if exports dwindled. On the contrary, remittances are found to be pro-cyclical and to increase volatility in the economy. The pitfalls of pro-cyclical capital flows such as inflationary pressures, asset bubbles, or an overvalued exchange rate, though, are largely absent.

4. This chapter is organized as follows: The following section presents some facts on remittance receipts in Bangladesh. Section C explores the link between oil prices and remittances, investigates whether remittances respond to shocks in domestic output, and shows how remittances fuel domestic demand. Section D offers some policy implications.

B. Some Facts

5. Reported workers remittances increased at an average annual rate of 17 percent over the last 30 years. Since the mid-1990s, they constitute the largest source of foreign exchange after exports (Figure 1). In 2006, workers remittances amounted to 7.7 percent of GDP, compared to 16.8 percent of GDP for merchandize exports, 1.9 percent of GDP for ODA, and 1.1 percent of GDP for FDI. Remittances through informal channels or the so-called ‛hundi’ system are estimated at about 75 percent of reported remittances.

Figure 1.
Figure 1.

Sources of Foreign Exchange, 1977–2006 1/

(In percent of GDP)

Citation: IMF Staff Country Reports 2007, 230; 10.5089/9781451804232.002.A006

1/ Fiscal years, starting July 1.

6. The rise in remittances is mirrored by a steady increase in emigrant numbers. Gross emigration of overseas workers increased from 6,000 in 1976 to 382,000 in 2006 with a quantum leap in the early nineties (Figure 2). The current stock of overseas workers is estimated at 3.9 million (2.8 percent of the population) translating into an average annual remittance of $1,230 per worker. Some 50 percent of overseas workers reside in Saudi Arabia and approximately 95 percent stay in oil exporting countries, suggesting a link between remittance flows and the oil price. The composition of emigrant workers changed little over time, ruling out a shift towards higher-end employment as a possible cause for rising remittance flows.

Figure 2.
Figure 2.

Annual Number of Emigrants, 1976–2006 1/

(In thousands)

Citation: IMF Staff Country Reports 2007, 230; 10.5089/9781451804232.002.A006

1/ Fiscal years, starting July 1.

7. The evolution of Bangladesh’s remittance receipts is broadly in line with the trend observed in global remittance flows. As a global aggregate, workers’ remittances have become the largest source of foreign exchange after exports and FDI, exceeding both ODA and portfolio investment by a wide margin. Some of the surge in workers remittances may be attributable to better recording and a shift from informal to formal channels owing to falling transaction costs and increased scrutiny after September 11. However, underpinned by population aging in the industrialized world, remittance flows are unlikely to abate soon. In the case of Bangladesh, persistent rural poverty and increasing population density will continue to drive emigration and remittances in the medium term.

8. Remittance receipts are large relative to the Bangladesh economy. Among nine countries of broadly equal population size, Bangladesh exhibits the largest remittances-to-GDP ratio after the Philippines (Figure 3). In Asia, the region with the largest remittance receipts, Bangladesh is surpassed only by the Philippines (13.8 percent of GDP), Nepal (12.2 percent of GDP), and Sri Lanka (8.9 percent of GDP).

Figure 3.
Figure 3.

Workers’ Remittances for Selected Countries, 2005 1/

(In percent of GDP)

Citation: IMF Staff Country Reports 2007, 230; 10.5089/9781451804232.002.A006

1/ Countries have broadly similar population sizes.2/ Fiscal year, starting July 1, 2005.

9. Bangladesh remittance receipts are less volatile than export receipts and ODA, confirming a pattern observed for global aggregates (IMF, 2005). The standard deviation of remittances around the trend amounts to 0.39 percentage points of GDP, compared with 0.64 percentage points of GDP for export receipts, and 0.58 percentage points of GDP for ODA receipts (Figure 4). FDI, on the other hand, is less volatile than remittances, deviating only 0.15 percentage points of GDP from the trend, on average.

Figure 4.
Figure 4.

Volatility of Foreign Exchange Flows, 1977–2006 1/

(In percentage points of GDP)

Citation: IMF Staff Country Reports 2007, 230; 10.5089/9781451804232.002.A006

1/ Fiscal years, starting July 1. Volatility is measured as the standard deviation of detrended flows-to-GDP, in percent.

C. The Macroeconomic Properties of Remittances

Oil price shocks

10. There is a presumption that remittance receipts in Bangladesh cushion against oil price shocks. For an oil importer like Bangladesh, an increase in the oil price means a real resource transfer to oil exporting nations. If the oil price hike is transitory, the oil importer usually runs down reserves or borrows abroad to avoid a contraction in consumption and investment. However, if the oil price hike is permanent, the oil importing nation eventually has to adjust to the real resource transfer and consume and invest less. In the case of Bangladesh, 95 percent of its overseas workforce is located in oil exporting countries and will likely earn and send home more when the oil price rises. This increase in foreign financing would help cushion consumption and investment against permanent increases in the oil price. In the case of transitory shocks it would be a cheap alternative to borrowing abroad or running down reserves.

11. The budget constraint of an open economy helps illustrate this point. In an open economy, real domestic demand or absorption can be written as the difference between real output and net exports: 2

A=Y(EXIM)(1)

Hence, any real demand that exceeds output generated (and sold) at home needs to be met by imports, which in turn need to be financed by borrowing, reserve depletion, grants, or workers’ remittances. If the price of imports increases and borrowing is not an option, say, because the price increase is permanent, real imports need to decline, and so does real absorption.3 However, if remittances increase with oil prices, the needed contraction in consumption and investment would be smaller.

12. There is strong evidence that remittances receipts in Bangladesh are positively correlated with oil prices. The correlation coefficient of detrended remittances and oil prices is 52 percent (Figure 5). In a regression, the coefficient of oil turns out highly significant and suggests that a one dollar increase in the oil price leads to $8 million in additional remittance receipts. The result is also significant in economic terms. In FY2005, oil prices increased by about 50 percent or $15. In the absence of additional foreign financing, and with oil imports accounting for 2.5 percent of domestic demand, this would have required a contraction of real demand by 1.25 percent. However, the same price increase also generated additional remittance receipts of $120 million, or 0.2 percent of domestic demand, offsetting about one sixth of the oil price effect on demand. Sure enough, Bangladesh did have access to other forms of foreign financing which helped smooth consumption, but those funds, to the extent that they were borrowed, have to be repaid in the future.

Figure 5.
Figure 5.

Remittances and Oil Prices, 1977–2006 1/

Citation: IMF Staff Country Reports 2007, 230; 10.5089/9781451804232.002.A006

1/ Fiscal years, starting July 1. The oil price is calculated as the average of Dubai Fateh, U.K. Brent, and West Texas Intermediate crude oil prices. Remittances have been detrended with the Hodrick-Prescott filter.

Shocks to domestic output

13. Remittance receipts may also respond positively to adverse output shocks. It is widely believed that remittances are motivated by altruism or based on some sort of insurance arrangement within the extended family (see Lucas and Stark, 1985; Rapoport and Docquier, 2005). Under these circumstances, families would attempt to smooth family consumption and increase remittances in response to country-specific shocks. Two recent cross-country studies have indeed found that remittance receipts increase following natural disasters (e.g., Yang, 2006, and Bluedorn, 2005; for the opposite finding, see Lueth and Ruiz Arranz, 2006).

14. Once more, the budget constraint of the open economy helps to frame this argument. Net exports in Equation (1), can be written as EX – IM = – NI – NT – KA+ ∆R, where NI is net income from abroad, NT is net transfers from abroad, KA is the capital account and ∆R is the change in reserves.4 Replacing this in Equation (1) and assuming that net income from abroad is negligible, gives:

A=Y+NT+KAΔR(2)

Hence, any demand that exceeds income generated at home needs to be financed by transfers from abroad, including workers remittances, or borrowing from abroad.5 If domestic output Y falters, remittances would boost NT according to the altruism/insurance hypothesis and, in this way, smooth consumption and investment.

15. There is little evidence that remittances respond to domestic shocks in Bangladesh. Figure 6 shows detrended remittances and the percentage change of remittances in years where natural disasters have occurred (shaded gray). To be considered in the graph, disasters needed to have affected at least 20 million people. Natural disasters are particularly suited to test the altruism/insurance hypotheses for a number of reasons. They are usually local events that would trigger offsetting transfers from a geographically dispersed and, hence, unaffected Diaspora. Moreover, since they are truly exogenous shocks, moral hazard is unlikely to undermine the remitters’ willingness to give, nor is there any danger that causality is running from remittances to domestic output. Out of the five disaster years, detrended remittances spiked only in one, the flood of 1988. The growth rate of remittances peaked in two of the disaster periods—1982 and 1988—but then remittances fell sharply during the flood of 1984.

Figure 6.
Figure 6.

Remittance Receipts During Natural Disasters, 1978–2006

Citation: IMF Staff Country Reports 2007, 230; 10.5089/9781451804232.002.A006

1/ Years of natural disasters are shaded gray.

16. By the same token, remittances are unlikely to cushion shocks to the country’s exports. A drop in the demand for Bangladesh exports would not only force an adjustment to imports, but would also depress domestic output. As suggested by the example of natural disasters, remittances are unlikely to come to the rescue. In addition, the main risks for Bangladesh’s exports—the loss of market share or a general price decline in the ready-made garment sector after transitory MFA safeguards expire—is a permanent shock, which insurance-type remittances would not cover anyway.

Do remittances increase volatility?

17. Remittances may actually increase volatility in Bangladesh. Another school of thought sees remittances as mainly profit driven, where the overseas worker allocates his earnings between home and host country depending on the marginal rate of return in either place. Here, the rate of return can be understood broadly to also include services rendered by the extended family in return for remittances. Instead of smoothing consumption, profitdriven remittances could actually increase economic volatility by fueling boom and bust cycles. In terms of Equation (2), workers’ remittances—which account for 90 percent of net transfers (NT)—would be positively correlated with output Y, thereby amplifying the effect of economic swings on absorption.

18. Real consumption and investment are, in fact, much more volatile than real output in Bangladesh (Figure 7). This pattern is much easier to reconcile with profit-driven remittances than with altruistic remittances. However, the swings in domestic demand could also be driven by other foreign financing flows, a possibility that is dismissed in the following few paragraphs.

Figure 7.
Figure 7.

Real Absorption and Output, 1982–2005 1/

(First difference, in billions of FY96 taka)

Citation: IMF Staff Country Reports 2007, 230; 10.5089/9781451804232.002.A006

1/ Fiscal years, starting July 1.

19. Remittances are the largest component financing domestic demand after output.6 Domestic output is obviously the main source of financing, accounting for 90–95 percent of real domestic demand during 1981–2006. The remaining sources of financing are shown in Figure 8. Reserve depletion is zero on average, as expected, but at times accounts for 1–2 percent of domestic demand. The share of foreign borrowing, at around 3 percent in 1981, has fallen steadily and is zero since the late 1990s. Similarly, other transfers (grants) haven fallen gradually from around 3 percent of domestic demand to ½ percent more recently. In contrast, workers remittances have been on an upward trend and since 1994 exceed other sources of foreign financing. Also note that the increase in volatility of domestic demand in Figure 7 coincides with the increasing importance of remittances.

Figure 8.
Figure 8.

Foreign Financing of Domestic Demand, 1981–2005 1/

(In percent of demand)

Citation: IMF Staff Country Reports 2007, 230; 10.5089/9781451804232.002.A006

1/ Fiscal years, starting July 1.2/ Capital account, errors and omissions, and net income from abroad.3/ Reserve depletion is defined as balance of payments deficit.

20. Remittances exhibit the highest cyclicality among foreign financing flows. As shown in Figure 9, remittance move with GDP in about a quarter of all years between 1981 and 2005. In contrast, the capital account is procyclical in only 13 percent of cases. Other transfers (grants) and reserves exhibit no significant cyclicality, but the coefficient of reserve depletion has the expected sign.7

Figure 9.
Figure 9.

Cyclicality of Foreign Financing Flows, 1981–2005 1/

Citation: IMF Staff Country Reports 2007, 230; 10.5089/9781451804232.002.A006

1/ Fiscal years, starting July 1. Cyclicality is defined as the correlation coefficient, in percent, between the foreign financing inflow and GDP, both detrended with the Hodrick-Prescott filter.

21. The cyclicality of remittance receipts in Bangladesh is also high by international standards. Figure 10 shows the cyclical properties of remittances for a group of 119 countries. About two-thirds of countries exhibit less cyclicality in their remittance receipts than Bangladesh. In about half of those countries’ remittances are actually countercyclical and help offset the effect of negative output shocks on consumption and investment.

Figure 10.
Figure 10.

Cyclicality of Remittnaces for Selected Countries, 1975-2005 1/

Citation: IMF Staff Country Reports 2007, 230; 10.5089/9781451804232.002.A006

1/ Cyclicality is defined as the correlation coefficient, in percent, between the foreign financing inflow and GDP, both detrended with the Hodrick-Prescott filter.

22. Remittances seem to boost consumption rather than investment. After having established the impact of remittances on domestic demand, it is interesting to know which component of demand is most affected by remittances. A large strand of mostly microeconomic literature tries to establish whether remittances are used for consumption or investment (see Rapoport and Docquier, 2005). In the former case, remittances would have at most short-term effects on growth, but could still go a long way in poverty reduction. In the latter case, remittances, by easing liquidity constraints, could move the country to a higher growth path. Figure 11 depicts the cyclical component of remittances and consumption in Bangladesh. The two series are closely correlated as evidenced by a correlation coefficient of 50 percent (significant at the 5 percent level). The correlation coefficient of remittances and investment, on the other hand, is not significant and negative.8

Figure 11.
Figure 11.

Detrended Consumption and Remittances,1981–2005 1/

(In billions of taka)

Citation: IMF Staff Country Reports 2007, 230; 10.5089/9781451804232.002.A006

1/ Fiscal years, starting July 1. Both series have been detrended by the Hodrick-Prescott filter.

23. There is no evidence so far that remittances fuel inflation or lead to Dutch disease in Bangladesh. If remittances are procyclical, there is a higher likelihood that they cause consumer price inflation, asset price bubbles, or overvalued exchange rates. The real effective exchange rate has been depreciating since 1980 (the earliest data point available) against continuously rising remittances, prima facie evidence against the Dutch disease phenomenon. However, the lack of a positive relationship between remittance flows and price indicators, including the exchange rate, the CPI, and the building construction cost index, is also confirmed using stationary series and econometric analysis.9

D. Policy Implications

24. Remittances are a welcome source of foreign financing and should be promoted. They are larger and less volatile than ODA, they are unrequited transfers which do not create debt service in the future, and they help cushion the economy against oil price shocks. According to various cross-country studies, countries can promote remittances through official channels by abstaining from dual exchange rate practices and current account restrictions, by reducing transaction costs through increased competition in the banking sector, and by fostering financial sector development more generally.

25. Remittances are no universal remedy and can not substitute for good policies at home. Remittances do not rise in response to economic shocks other than oil price shocks. On the contrary, remittances dwindle when the economy slows and amplify the economic downturn. Cross-country research also found remittances to be strongly correlated with the political and investment climate in the home country.

26. Financial deepening should leverage the impact of remittances on growth. Currently, a large part of remittances seems to be used for consumption. This increases welfare–and more so the poorer the recipient of remittances—and should not be discouraged. At the same time, government policies should support financial deepening to provide savings instruments for those who do want to save out of their remittance receipts. Abstaining from financial repression, fostering competition in the financial sector, and providing a strong regulatory environment would help in putting more remittances to productive use.

27. If remittances have not significantly aggravated economic cycles so far, it does not mean that they will not in the future. Presently, BB is experiencing upward pressure on the exchange rate owing in large parts to strong remittance inflows. This is a welcome opportunity to build up reserves to more comfortable levels, but should be accompanied by sterilization to prevent excessive money and credit growth, nonperforming loans, and inflationary pressures down the road.

References

  • Bluedorn, John C., 2005, “Hurricanes: Intertemporal Trade and Capital Shocks,” Economics Papers 2005-W2 (University of Oxford).

  • EM-DAT: The OFDA/CRED International Disaster Database, 2007, Université Catholique de Louvain, Brussels, Belgium, www.em-dat.net

  • Giuliano, Paola and Marta Ruiz-Arranz, 2005, “Remittances, Financial Development, and Growth,” IMF Working Paper 05/234 (Washington: International Monetary Fund).

    • Search Google Scholar
    • Export Citation
  • International Monetary Fund, 2005 World Economic Outlook, April 2005: A Survey by the Staff of the International Monetary Fund, World Economic and Financial Surveys (Washington).

    • Search Google Scholar
    • Export Citation
  • Lucas, Robert E. B. and Oded Stark, 1985, “Motivations to Remit: Evidence from Botswana,” Journal of Political Economy, Vol. 93, pp. 901– 918.

  • Lueth, Erik and Marta Ruiz Arranz, 2006, “A Gravity Model of Workers’ Remittances,” IMF Working Paper 06/290 (Washington: International Monetary Fund).

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Rapoport, Hillel and Fré9;déric Docquier, 2005, “The Economics of Migrants’ Remittances,” IZA Discussion Paper No. 1531 (Bonn).

  • Rajan, Raghuram G. and Arvind Subramanian, 2005, “What Undermines Aid’s Impact on Growth?” IMF Working Paper 05/126 (Washington: International Monetary Fund).

    • Search Google Scholar
    • Export Citation
  • Yang, Dean, 2006, “Coping with Disaster: The Impact of Hurricanes on International Financial Flows, 1970-2002,” NBER Working Paper 12794 (Cambridge, Massachusetts).

    • Search Google Scholar
    • Export Citation
1

Prepared by Erik Lueth (APD).

2

Domestic demand or absorption is the sum of domestic (private and public) consumption and investment.

3

The mechanism that brings this about is usually a devaluation of the currency, which reduces the purchasing power of households and discourages imports by making them more expensive. Steep devaluations can also negatively affect GDP.

4

This is derived from the balance of payments equation CA+ KA = ∆R, where CA = EX – IM + NI + NT is the current account balance.

5

Borrowing in this context means a falling net investment position and includes FDI inflows, portfolio inflows, and the running down of reserves.

6

The national accounts data are taken from the CEIC database and the balance of payments (BOP) data stem from the IMF Balance of Payments Statistics Yearbook. BOP data was converted into taka using the nominal exchange rate. Nominal data were converted into real data using the GDP deflator.

7

It is worth noting that ODA is more procyclical than remittances with a correlation coefficient of 45 percent. Since the capital account as a whole is less procyclical errors and omissions, the other major line item under the capital account, must be counter-cyclical.

8

The notions that remittances are motivated by rate of return considerations and fuel consumption are no contradiction, since the rate of return can be understood broadly to also include services rendered by the extended family in return for remittances.

9

On the contrary, one finds a negative relationship between the strength of the taka and remittance flows, suggesting that causality runs from the exchange rate to remittance flows (see Rajan and Subramanian, 2005).

Bangladesh: Selected Issues
Author: International Monetary Fund