69498 BHUTAN HYDROPOWER EXPORT BOOM: ITS MACROECONOMIC IMPACTS AND POLICY IMPLICATIONS September 2002 CURRENCY EQUIVALENTS Currency Unit = Ngultrum (Nu) US$1 = Nu 48.32 (September 18,2002) FISCAL YEAR July 1 to June 30 ABBREVIATIONS AND ACRONYMS BoP Balance of Payment BRER Bilateral Real Exchange Rate CPI Consumer Price Index CSO Central Statistical Organization DoP Department of Power FDI Foreign Direct Investment GDP Gross Domestic Product GoI Government of India IMF International Monetary Fund INR Indian Rupee MU Mega Unit MW Mega Watt Nu Bhutanese Ngultrum RMA Royal Monetary Authority UNDP United Nations Development Programme WPI Wholesale Price Index Vice President: Mieko Nishimizu Country Director: Marilou Uy Sector Director: SadiqAhmed Task Manager: Naoko Kojo 2 ACKNOWLEDGEMENTS This report was prepared by Naoko Kojo. The report benefited from invaluable comments from and stimulating policy discussions with a number of World Bank staff members. These include Edgardo Favaro, Elena Glinskaya, Ijaz Nabi, Garry Pursell (SASPR), Roberto Zagha, Tara Vishwanath (SASVP), Simon Bell (SASFP), and Tetsuya Konishi (ECSSD). The internal reviewer for this paper was Shahrokh Fardoust. Will em H. Buiter, European Bank of Reconstruction and Development, provided insightful comments as an external reviewer. This report would not have been completed without the excellent cooperation of many officials of Royal Government of Bhutan and Royal Monetary Authority of Bhutan. In particular, the staff would like to thank Ugyen Norbu, Karma Tashiteem (Ministry of Finance), and Panjore (Royal Monetary Authority) for their constructive comments at the workshop of December 2001 and provision of requisite data. The draft report was discussed with the Royal Government officials in October 2002. Administrative support was provided by Juliet Teodosio. The report was completed under the guidance of Sadiq Ahmed (Sector Director, SASPR) and Marilou Uy (Country Director for Bhutan). The Vice President for the South Asia Region is Mieko Nishimizu. 3 TABLE OF CONTENTS Executive Summary ................................................................................................................ 6 1. Introduction ....................................................................................................................... 7 2. Mineral Export Boom, Real Exchange Rates and Dutch Disease ......................................... 8 2.1 Spending Effect ....................................................................................................... 8 2.2 Resource Movement Effect ................................................................................... 10 2.3 Dutch Disease and Monetary Effects ................................................................ 11 3. Empirical Analysis: Bhutan's Power Exports and Macroeconomic Impacts ..................... 11 3.1 Power Sector .......................................................................................................... 12 3.2 Power Exports and Real Exchange Rates ................................................................ 23 3.3 Power Exports and Dutch Condition ....................................................................... 28 4. Policy Recommendations ................................................................................................. 38 4.1 Fiscal Policy .......................................................................................................... 39 4.2 Exchange Rate Management .................................................................................. 40 4.3 Private Sector Development ............................................. . ........... . ....... .42 5. Conclusions ..................................................................................................................... 43 Appendix 1: Measuring Real Exchange Rates ...................................................................... 44 Appendix 2: Empirical Analysis: Determinants of Bilateral Exchange Rate of Ngultrum ..... .47 References ............................................................................................................................ 53 TABLE OF BOXES Box 1. Nigeria's Dutch Disease ............................................................................................ 11 Box 2. Botswana's Experience ............................................................................................. 40 4 TABLE OF FIGURES Figure 1. Dutch Disease (Classical Definition) ........................................................................... 9 Figure 2. Bhutan: Power Exports (as percentage of total exports, GDP) ................................... 13 Figure 3. Bhutan: Volume of Sales by Chukha Hydropower Corporation (in mega units) ......... 13 Figure 4. Bhutan: Power Tariff per Unit (in ngultrum) ............................................................. 14 Figure 5. Bhutan: Real Power Tariff per Unit (in ngultrum, constant 1980 prices) .................... 14 Figure 6. Bhutan: GDP Growth Rate (in percent per annum) .................................................... 16 Figure 7. Bhutan: Index ofGDP (1980 = 100) ......................................................................... 16 Figure 8. Bhutan: Government Tax and Non-tax Revenues and Power Exports ........................ 18 Figure 9. Bhutan: Fiscal Development (as percentage ofGDP) ................................................ 18 Figure 10. Bhutan: Current Account Transactions with India (as percentage ofGDP) .............. 19 Figure 11. Bhutan: Net Capital Inflows from India and Power Export Revenues ...................... 19 Figure 12. Bhutan: Balance of Payments (as percentage ofGDP) ............................................. 20 Figure 13. Bhutan: Current Account and Balance of Payments with Third Countries ............... 20 Figure 14. Bhutan: Net Capital Inflows by Countries of Origin (as percentage ofGDP) ........... 21 Figure 15. Bhutan: Net Capital Inflows (in millions ofngrultrum, constant 1980 prices) .......... 21 Figure 16. Bhutan: Gross Rupee and Hard Currency Reserves ................................................. 22 Figure 17. Bhutan: Bilateral Real Exchange Rates 1/ ............................................................... 24 Figure 18. Price Indices: Bhutan CPI and India WPI (1980 = 100) ............................................ 26 Figure 19. Price Indices: Bhutan CPI and India CPI (1980 = 100) ............................................. 26 Figure 20. Bhutan: BRER(NU/INR) and Real Power Exports to India (in 1985 prices) ............. 27 Figure 21. Bhutan: Composition of Absorption ........................................................................ 30 Figure 22. Bhutan: Volume of Food Imports (in metric tons) and Power Income ...................... 3l Figure 23. Bhutan: Government Capital Expenditure and Grants (as percentage ofGDP) ......... 32 Figure 24. Bhutan: Government Domestic Revenues and Current Expenditure ........................ 32 Figure 25. Bhutan: Components ofGDP .................................................................................. 33 Figure 26. Bhutan: Index of Sectoral Output (1980 = 100) ....................................................... 34 Figure 27. Bhutan: Components ofGDP (as percentage of non-power GDP) ........................... 34 Figure 28. Bhutan: Exports and Non-power Exports ............................................................... 35 Figure 29. Bhutan: Exports by Country of Destination ............................................................. 35 Figure 30. Bhutan: Inflation Rates (in percentage per annum) .................................................. 37 Figure 31. Bhutan: Money Supply (M2) and Money Velocity .................................................. 37 5 EXECUTIVE SUMMARY Bhutan has shown remarkable economic performance over the last two decades. Growth during the second half of the 1990s was particularly strong, with annual GDP growth averaging 6.5 percent. A large part of this performance has been supported by generous inflows of foreign aid and buoyant electricity exports to India, which have spurred growth both directly by expanding export earnings and indirectly by stimulating investment in the construction and transport industries. Growth prospects for Bhutan look bright, supported by enormous hydropower generation potential. In particular, the scheduled commissioning of the Tala hydropower project in 2005/06 is expected to give a major boost to growth by tripling the country's power generation capacity. While the expansion in power exports has brought significant benefits to Bhutan, there has been concern that it may not be an "unmitigated blessing." Despite large net inflows of foreign capital, Bhutan's macroeconomic indicators suggest that inflation has so far been modest and under control. Over time, however, the real value of the ngultrum has appreciated steadily against the Indian rupee, eroding Bhutan's export competitiveness with India. Concern has been raised whether this is a sign of the Dutch disease caused by increased power exports to India, and whether additional foreign exchange inflows, including power export earnings and the World Bank's financial assistance, may undermine Bhutan's growth prospects by discouraging production and investment in the tradable sector such as agriculture and manufacturing. The results of the empirical analysis of the Bhutanese economy indeed confirm that the power exports to India as well as fmancial assistance have caused real appreciation of the ngultrum against the rupee over time, through increased private spending and short-term monetary disequilibrium. Despite the real appreciation, however, the study finds little evidence to suggest contraction or stagnation of the non-power tradable sector. While the performance of the tradable sector may have been stronger in the absence of the real appreciation, thus far there is no sign of the Dutch disease, or the Dutch condition, in the Bhutanese economy. Nevertheless, our empirical results indicate that further increases in power exports as well as foreign aid - regardless of which currencies they are denominated in - will likely put further upward pressure on the real value of the ngultrum against the rupee and could lead to the Dutch condition. Although there is no immediate threat that requires urgent policy attention, the country's long-term growth prospects will also depend on how effectively the Government uses the increasing power rents and foreign aid, and how well the Government manages the exchange rate. 6 1. INTRODUCTION Bhutan has shown remarkable economic performance over the past two decades. Growth during the second half of the 1990s was particularly strong, with annual GDP growth averaging 6.5 percent. A large part of this performance has been supported by generous inflows of foreign aid and buoyant electricity exports to India, which have spurred growth both directly by expanding export earnings and indirectly by stimulating investment in the construction and transport industries. Growth prospects for Bhutan look bright, supported by enormous hydropower generation potential. In particular, the scheduled commissioning of the Tala hydropower project in 2005/06 is expected to give a major boost to growth by tripling the country' s power generation capacity. While the expansion in power exports has brought significant benefits to Bhutan, there has been concern that it may not be an "unmitigated blessing. " Despite large net inflows of foreign capital, Bhutan's macroeconomic indicators suggest that inflation has so far been modest and under control. Over time, however, the real value of the ngultrum has appreciated steadily against the Indian rupee, eroding Bhutan's export competitiveness with India. The 2000 Country Assistant Strategy for Bhutan views this as a sign of the Dutch disease caused by increased power exports to India, and suggests that additional foreign capital inflows, including further expansion of power export earnings and the World Bank' s fmancial assistance, may undermine the country's growth prospects as real appreciation tends to discourage production and investment in the tradable sector such as agriculture and manufacturing. This paper examines empirically the macroeconomic impacts of Bhutan's increased power exports to India. It is the first serious attempt to address the presence of the Dutch disease in the Bhutanese economy. Previous attempts had been hampered by lack of sufficiently long time- series data.! Four sets of key questions are addressed in this study: (i) what are the features of the power export boom in Bhutan? Is it a temporary or permanent phenomenon? How large are power exports? (ii) Have power exports appreciated the real value of the ngultrum? If so, would further increase in power exports put further upward pressure on the ngultrum? (iii) What have the impacts on the tradable sector been? Are there Dutch-disease symptoms? (iv) Are these impacts sizable enough to require policy interventions? If so, what kind of policy options are then available to the Government? The results of the study may help the Royal Government of Bhutan formulate appropriate policy strategies to make effective use of power income to improve its population's living standards and well-being, while reducing heavy reliance on the foreign aid in order to achieve a self-sustained economy in the long-run. The structure of this paper is as follows. Section 2 reviews the mechanism of how a natural resource export boom can lead to real appreciation and then to contraction of the tradable sector - that is, how the Dutch disease can happen. Section 3 then examines empirically Bhutan's power exports and their impacts on the domestic economy in line with the discussions in Section 2. It first discusses the features of Bhutan's power exports and assesses the main impacts on national 1 Data sets are available upon request. 7 income, government revenues, and external positions. It then empirically investigates the impacts on the real exchange rate and analyzes the impact on the economic structure in Bhutan. Section 4 provides a set of policy implications to the Government. Section 5 concludes the paper. 2. MINERAL EXPORT BOOM, REAL EXCHANGE RATES AND DUTCH DISEASE A natural resource export boom can cause major distortions in an economy by discouraging production and investment in the non-booming exportables sector, often referred to as the "Dutch disease." 2 The extensive literature on the subject has been surveyed by Corden (1984) and Neary and van Wijnbergen (1986).3 The discussion here is concise and selective. It serves only to provide the background information necessary for the analysis in Section 3. 2.1 Spending Effect The Dutch disease can be analyzed in the context of two separate effects in the real sector - a spending effect and a resource movement effect - and a monetary effect (see Figure 1). Let us assume a natural export boom. First, there will be a rise in real income. The increased real income will raise spending on both tradables and non-tradables through the income effect, in the form of government expenditures and private spending, thereby stimulating the demand for both goodsY However, since the prices oftradables are determined on the world market, the higher domestic demand does not push up the prices of tradables. Any excess demand for tradables is met by an increased volume of imports. The domestic prices ofnon-tradables, in contrast, will rise in response to the higher demand. 6 Accordingly, there is a decline in the prices of tradables relative to the prices of non-tradables, which by definition is real appreciation of the home currency. 7 The extent of the real 2 Named because of the Netherlands' experience with the discovery of natural gas in the 1970s. 3 Also see Buiter and Purvis (1983), Corden and Neary (1982), Eastwood and Venables (1982), Neary and van Wijnbergen (1984), and van Wijnbergen (1984). Amongst the literature, a formal analysis developed by Corden and Nearly (1982) has become a standard model in the literature. For computable general equilibrium model, see Bruno and Sachs (1982). 4 Provided that both tradables and non-tradables are nonnal goods. 5 The response of expenditure to real exchange rates does not matter whether the expenditure comes from the public or the private sector. According to Vaey-Zadeh (1989), even if all the windfall income accrues to the govermnent, increases in mineral rent tend to generate a positive "confidence effect" on consumers, affecting aggregate expenditure by the private sector. 6 Provided that there is no corresponding expansion in the domestic supply ofnon-tradables. 7 Note that the (bilateral) real exchange rate of the home currency against the foreign currency, BRER, can be expressed as BRE~ == eP/ / ~NT, where e is the fixed nominal exchange rate of the home currency, ~T the prices oftradables and P/T the prices ofnon-tradables. Namely, the real exchange rate is merely the relative prices oftradables to non-tradables. A decline in BRER indicates real appreciation of the home currency. BRER is frequently used to measure the relative profitability of producing tradables compared with non-tradables. A fall in BRER (real appreciation) indicates a loss of profitability by producing tradables relative to producing non-tradables, and vice versa. 8 Figure 1. Dutch Disease (Dutch Condition) Natural resource export boom Real Sector ------------------ 1 --------------- --- ---- ------- --- 1 Monetary Sector ----------------------- Increase in real income Current account surplus BoP surplus Confid,nce effect 11 Incom, ojf0, ~ ~o, ~ n 00,b ~o, ~ '-do, ~ \0, ~o, ~ f3~ ~oo ~"o 00 ,-0,"0 ~~ ~o,~ ,-0,0," ~o, ~o, ,-0,0, ~o, ~o, ,-0,0,'\ ~o,oo ~o,o, 1/ Sale of electricity to the Department of Power, DoP, which distributes electricity domestically. Source: Data provided by the Bhutanese authorities. 13 Figure 4. Bhutan: Power Tariff per Unit (in ngultrum) 1.5 Export Price 1.0 0.5 Domestic Price (to DoP) 11 • • • • • • 0.0 : : : : : : := .;' +---,---,----,---,---,---,---,----,---,---,---,---,----,---,---4 Figure 5. Bhutan: Real Power Tariff per Unit (in constant 1980 ngultrum) 0.6 ~-------------------------------------------------------------, 0.5 0.4 0.3 0.2 0.1 Domestic price (to DoP) 0.0 +---,----r---r---r---,---,---,----r---r---,---,---,----r---r--~ 1986 1987 1988 1989 1990 1991 1992 1993 1994 1995 1996 1997 1998 1999 2000 1/ The price at which power is sold to the DoP. DoP applies different tariff rates for domestic power supply. Source: Data provided by the Bhutanese authorities. 14 The tariff for power to be exported from the Tala project, in contrast, will likely be determined at the commission level and take generation costs into account. Therefore it is expected to be closer to the market price. Payments for power sales, as well as all other imports and exports between Bhutan and India, take place in the Indian rupee according to the bilateral free trade agreement between the two countries. The Government ofIndia (GoI) has provided financial assistance (in the Indian rupee) for the construction of all major power plants, except the Basuchu project, which was built for domestic purposes only and was financed by Austria. Each GoI financing consists of 60 percent grant elements and 40 percent soft loan elements. Chukha Hydropower Corporation is a government-owned corporation that is managed by the Department of Power in the Ministry of Trade and Industries. It is not a foreign direct investment from India. The export proceeds realized by the corporation are channeled to the Government via surplus transfers, corporate taxes and dividend payments. All the investments in the power sector are recorded outside the government budget. Under the Electricity Act approved by the National Assembly in July 2001, the utility side of the Chukha Hydropower project was corporatized in July 2002. A regulatory body, the Bhutan Electricity Board, was also set up at this time. All other power plants, including those to be commissioned in the future (such as the Tala project), will be merged into Chukha Hydropower Corporation Because of Bhutan's long-term sales agreement with India, electricity export earnings from India will continue to be a permanent source of income to Bhutan. The risks of reduced demand from India is likely to be very small. At present, Bhutan supplies less than 0.5 percent ofIndia's demand. While internal demand in Bhutan will also rise substantially with progress in the Rural Electrification Program (supported by the Asian Development Bank), the rise in Bhutan's power generating capacity (when the Tala project is completed) is expected to meet both Bhutan's expected demands and a larger part of India's as well. The risks of sharp contractions in electricity supply as a result of technical and climate change are also considered to be low. 11 Bhutan's power plants are operated as run-of-the-river schemes and are environmentally benign. Unlike other natural resource export booms based on exhaustible resources (such as oil, gas, coal, copper), Bhutan's hydropower is renewable and environmentally friendly, constituting a permanent resource for exports. Looking ahead, the power sector is expected to continue to be the main engine of growth in the country and a source of significant rupee earnings. 3.1.2 Impacts on National Income Regular power exports to India since 1987 have permanently raised the level of Bhutan's national real income. 11 However, in May 2002 the power generation ofthe Kurichhu project was severely disrupted by a heavy storm, which damaged two transmission towers and two conductors. The damages also disrupted the supply of power from Kurichhu to India, causing a revenue loss of about Nu.8 million, equivalent to US$170,000 or 11 days of power generation from the Kurichhu project. 15 Figure 6. Bhutan: GDP Growth Rate (in percent per annum) 20 , -------------------------------------------------------------, 15 10 5 Figure 7. Bhutan: Index ofGDP (1980 = 100) 400 .---------------------~--------------------------------------__, 350 300 250 200 150 t L< ··· ····.. ....................... Permanent increase in real income Source: RMA Annual Reports (various issues.) 16 A plotting of the real annual growth rate of GDP in Bhutan for the period 1980-2000 shows a large spike (18 percent) in 1987, followed by a sharp decline to one percent in the following year, indicating that the regular power exports to India had a large "level effect" on GDP (Figure 6). However, the magnitude of the level effect should not be over-emphasized. First, fiscal yearl987 was 15 months long (April 1, 1987 - June 30, 1988) due to the transition from old fiscal years (April 1 - March 30) to new ones (July 1 - June 30) that occurred in 1988. Strictly speaking, therefore, the growth rate in 1987 is not comparable to that of other years. If equally appropriated, the 12-month growth rate of the fiscal year 1987 could have been around 14 percent. The index of GDP shown in Figure 7 indicates that the "accelerating effect" on the growth rate may have been marginal during the initial years after the power export had began. In the late 1990s GDP grew at an accelerating rate as the result of the tariff revisions. 3.1.3 Impacts on Government Revenues Power exports have also contributed significantly to a rise in government revenues through increases in the corporate tax revenue, surplus transfers and dividend payments from Chukha Hydropower Corporation. Figure 8 shows government tax and non-tax revenues together with power exports as percent of GDP over the last two decades. Although the Government does not publish the breakdown of revenues, it is evident from Figure 8 that there is a high correlation between government non-tax revenue and power export proceeds. The level of government non-tax revenue has risen significantly since 1987, most likely as a result of surplus transfers and dividend payments from Chukha Hydropower Corporation. In contrast, government tax revenue has risen gradually over time. The overall fiscal position was more or less balanced during the 1990s (see Figure 9). From 1998 on, however, the fiscal position began to deteriorate. The fiscal deficit widened significantly from 1.9 percent ofGDP in 1998 to 4.0 percent ofGDP in 1999, and further to 6.1 percent of GDP in 2000. This fiscal easing is viewed as a reflection of increased capital expenditures related to the construction of new power projects, which are temporary in nature and financed exclusively by foreign aid. 3.1.4 Impacts on External Positions and Reserves Because of the special relationship with India, Bhutan maintains two sets of external accounts: one with India denominated in the Indian rupee, and the other with countries other than India, or the "third countries" in the Bhutanese term, denominated in the US dollar. (i) Current Account Bhutan's current account position with India has improved markedly as a result of regular power exports. While Bhutan has continued to run a current account deficit with India throughout the past two decades, the magnitude of deficit has shrunk considerably since regular power exports began in 1987 - from the average of38.4 percent ofGDP for 1980-86 to the average of 10.2 percent ofGDP for 1987-99 (see Figure 10). Between 1995 and 1997 net exports (trade balance) were even in surplus, but larger net transfer payments kept Bhutan' s current account balance with India from turning positive. 17 Figure 8. Bhutan: Government Tax and Non-tax Revenues and Power Exports (as percentage of GDP) 20 .-------------------------~----------------------------------__, Non-tax revenue 15 10 5 Tax revenue Power exports ""I:) ",," :bv ",,'" ~ ",," ""b ~ ",,'b ~ AI:) A' A"" A'" At>< A" Ab A A'b AC) ~C'.I:) '" ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ f Source: RMA Annual Reports (various issues.) Figure 9. Bhutan: Fiscal Development (as percentage ofGDP) 60 .---------------------~------------------------------------------__, Government expenditure (inc!. 50 net lending) 40 30 II Govemment revenue (ina!. grants 1/) 20 10 f Fiscal balance -20 ~--------------------~------------------------------------------~ 1/ Grants are both from India and countries other than India. Source: RMA Annual Reports (various issues.) 18 Figure 10. Bhutan: Current Account Transactions with India (as percentage ofGDP) 40 30 20 10 Exports to India -20 -30 urrent account balance with India -40 -50 ~----------------------~----------------------------------------~ Source: RMA Annual Reports (various issues.) Figure 11. Bhutan: Net Capital Inflows from India and Power Export Revenues (as percentage of GDP) 50 40 Net catpial flows from India 11 30 20 Power export revenue 10 ~~~~~~~~~~~~~~~~~~~# ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~. ~ ~ " ~ ~ ~ " Source: RMA Annual Reports (various issues.) 1/ Consist of concessionalloans and grants. 19 Figure 12. Bhutan: Balance of Payments (as percentage ofGDP) 20 , ----------------------,--------------------------------------, 15 BoP ( wthird cOlUltries 10 5 -5 BoP wi India -10 L-____________________ ~~ __________________________________ ~ Source: RMA Annual Reports (various issues.) Figure 13. Bhutan: Current Account and Balance of Payments with Third Countries (as percentage of GDP) 20 Imports from third cOlllltries 10 -10 Cnrrent accollllt wi third cOlllltries -20 L-____________________________________________________________ ~ Source: RMA Annual Reports (various issues.) 20 Figure 14. Bhutan: Net Capital Inflows by Countries of Origin (as percentage ofGDP) 70 ~---------------------------------------------------------------, 60 6:' 50 o o '2F _+.......t_4I>_+4-1__t_-1__t_____1-_I_____1-_+__t-_I___r-+___r-_I_--+ 0 Source: Data provided by the authorities and World Bank staff estimates appreciation started before the country began to receive regular power export income in 1987, the rate of appreciation accelerated in the late 1990s, when Bhutan's power export earnings increased substantially due to tariff revisions. 3.2.4 Empirical Analysis: Determinants of Bilateral Exchange Rate ofthe Ngultrum To examine this relationship more formally, we have conducted a regression analysis using annual data for 1983-99, which is the longest period for which all variables are available. 14 The details of the empirical analysis are provided in Appendix 2. Here we present only the main findings. The regression results indicate that increases in power exports have indeed caused real appreciation of the ngultrum over time. Both the power tariff (in real terms) and the volume of power exports appear to have contributed to the appreciation, implying that future increases in either variable are likely to cause further appreciation of the ngultrum. The results also suggest that the real appreciation took place through the accumulation of international reserves (monetary effect), which have largely been affected by net capital inflows 14 While data on other variables are available for 1981-99, the monetary data are available only from 1983 onwards. 27 and power exports, rather than by power exports alone. Changes in government expenditure do not appear to have significantly affected movement of the ngultrum (that is, no significant spending effect by the Government), implying that a large part of the Government's spending may have been on tradables such as imported capital and intermediate goods instead of non- tradables. In other words, the marginal propensity to spend may have been lower on non- tradables than on tradables, or the private sector's spending on non-tradables may have been stronger than that of the public sector. Another possible explanation is that government expenditure is not a good proxy for government spending on non-tradables. In summary, the real appreciation of the ngultrum has likely been caused by power exports and the monetary disequilibrium associated with large net capital inflows. An expansion in power exports or an increase in net capital inflows in the future, regardless of which currency they are denominated in, would likely put further upward pressure on the real value of the ngultrum against the rupee. 3.3 Power Exports and the Dutch Condition We now tum to our second set of questions of this study. Has the real appreciation of the ngultrum reduced the incentive for production and investment in the tradable sector? Have there been resource shifts away from the non-power tradable to non-tradable sector? In other words, does Bhutan have the Dutch condition? To address these issues, we closely follow the discussion in Section 2. We examine the domestic economy' s adjustment to the increased power exports in light of (i) a spending effect; (ii) a resource movement effect; and (iii) a monetary effect. We show below that increased power exports (and capital inflows) have caused real appreciation mainly through increases in private sector spending (spending effect) and through monetary disequilibrium (monetary effect). However, government expenditure on non-tradables appears to have had very little effect on the appreciation of the ngultrum. Despite changes in relative prices, we have found no evidence that production of the tradable sector has declined over time. 3.3.1 Spending Effece s From Section 2 we know that the increased real income from a natural resources export boom is likely to lead to an increase in the prices of non-tradable goods relative to the prices of tradable goods (real appreciation) through a spending effect, which by definition is real appreciation of the home currency, provided that tradables and non-tradables are both normal goods. From Section 3.2, we also know that Bhutan has experienced a gradual real appreciation of its home currency over the last two decades. The empirical analysis shows, however, that the real exchange rate of the ngultrum did not respond strongly to government spending, indicating that a large part of spending by the Government may have been on tradables rather than on non- tradables, or that the private sectors' spending may have been stronger than the public sector' s. This subsection examines how the spending effect has taken place in Bhutan, in other words, how the power export income has been spent, following the rise in real income caused by power exports. 15 It would be ideal to examine the impact ofthe spending on inter-sectoral price relationships. However, effective assessment was hampered by Bhutan's weak statistical base. 28 Following Gelb (1986), we calculated how the extra power income was used within the domestic economy. Table 2 provides an indication of the base absorption structure of Bhutan averaged in four five-year periods: 1980-85, 1985-90, 1990-95 and 1995-00. The calculation was conducted as follows. First, data on private consumption and public consumption were measured as a percentage ofGDP, and investment as a percentage ofGDP was derived as a residual. Second, the components of absorption were scaled by non-power GDP and then compared with those scaled by GDP (or "norms," according to Syrquin and Chenery, 1989). The difference between these two represents the extra domestic absorption of power income. In Table 2, the breakdown of extra absorption indicates a particularly strong bias in favor of private consumption. Figure 21, which plots the breakdown of absorption as a percentage of non- power GDP for the period 1980-2000, shows two distinct private consumption booms in Bhutan, the first in 1988 and the second in 1999. From Figure 2 (p.l3), it is evident that power exports Table 2. Bhutan: Composition of Absorption (in percent) 1980-85 1985-90 1990-95 1995-00 Private consumption Actual/ non-power GDP 67.5 70.3 49.5 56.0 Norm 67.3 65.2 45.7 49.5 Public consumption Actual/ non-power GDP 23.5 20.3 20.5 25.7 Norm 23.5 19.0 18.9 22.7 Investment Actual/ non-power GDP 9.2 17.0 38.6 31.3 Norm 9.2 15.8 35.5 27.8 Absorption Actual/ non-power GDP 100.2 107.6 108.6 113.0 Norm 100.0 100.0 100.0 100.0 Breakdown of extra absorption (as % of total absorption) Private Consumption 76.8 68.1 45.6 49.5 Public Consumption 11.5 16.1 19.0 22.7 Investment 11.6 15.9 35.4 27.8 Total 99.9 100.1 100.0 100.0 Breakdown of extra absorption (as % of non-power GDP) Private Consumption 0.0 1.3 1.7 3.0 Public Consumption 0.0 1.2 3.1 3.5 Investment 0.0 7.2 7.1 9.5 Total power income (as % of non-power GDP) 0.0 9.7 11.9 16.0 Total power income (as % of total GDP) Source: Data provided by the Bhutanese authorities and World Bank staff estimates. 29 Figure 21. Bhutan: Composition of Absorption 10 ,-------------------------------------------------------------, 9 8 Private consumption 7 ~ 6 5 4 3 2 Public consumption o ~~~~~~~--~~~~~~~--~~~~~~~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ f Source: Data provided by the Bhutanese authorities and World Bank staff estimates. have given rise to these booms, probably through the confidence effect; there was a sharp rise in power income in both 1987 and 1998-99. During the time when the ratio of power exports to GDP declined (1989-95), the private sector consumed less of that power income. The trend of private consumption is also reflected in the increases in the volume of food imports (private spending on tradables). Figure 22 presents the import volumes of rice and wheat and suggest that increased power revenues have been used to increase food imports from India. In per capita terms, rice imports rose from 11 kilograms in 1983 to 17 kilograms in 1999. Rice import peaks during this time period were 20 kilograms per person in 1988 and 22 kilograms per person in 1997. The use of power income for investment rose during the early 1990s and declined in the late 1990s. Interestingly, the movement of investment is exactly opposite the movement of private consumption of extra power income. As explained above, in this calculation domestic investment is treated as a residual, and therefore it is not possible to distinguish investment as private or public capital formation. Yet there are good reasons to believe that private sector investment has responded relatively more strongly to the extra power income than public investment. First, although data are sketchy, there is a general perception that the price of land, the primary asset in the composition of wealth in agrarian societies like Bhutan, has shown a dramatic rise in the last decade, indicating that private 30 Figure 22. Bhutan: Volume of Food Imports (in metric tons) and Power Income (as percentage of GDP) 18,000 ,..-- - - - - - - - - - - - - - - - - - - - - - - - - - - - -, 14% 16,000 Rice import (left scale) 12% 14,000 ';;;' 12,000 \ 10% p:- O 0 "- 0 " .8 '" 8% .;:l bJl .~ 10,000 'i) E " '" 2 " ~ 8,000 '" 0- 6% ~ o en 0- Power exports (right scale) '" .5 -0 6,000 ~ o o 4% '" > ~ "- .... 4,000 '" '" 0 c... 2% 2,000 O~ __~~__~-~-~~-_+-~-+_-~_+-~-+_-~~-_+-+ 0% 1983 1984 1985 1986 1987 1988 1989 1990 1991 1992 1993 1994 1995 1996 1997 1998 1999 Data source: Statistical Yearbook of Bhutan, CSO (various issues), RMA Selected Economic Indicators (2001) and World Bank staff estimates. spending on non-tradables is on the significant rise. 16 Spending on another non-tradable good, labor, also appears to be increasing. Again, data on wages are not readily available, but it is commonly believed that wages in Bhutan are rising rapidly compared with those in India, attracting a large number of illegal workers into Bhutan. Second, a substantial part of government capital expenditure has been financed with grants (budgetary support), thus less power income was likely used to finance public investment. 17 Figure 23 shows the relationship of government capital expenditure with grants and with power export revenues. From Figure 23, it is clear that government capital expenditure is closely linked to the level of grants from foreign donors, but less closely linked to power income. This indicates that less extra power income has been used to finance public investment in Bhutan and that more of it has gone to finance private investment and private consumption. Notwithstanding, the overall level of private investment in Bhutan remains very low, with most investment by the public sector. In fact, the Royal Government of Bhutan has adopted a public expenditure policy, in which public investment projects take place only when matching foreign concessional financing is available. Recurrent expenditures, in contrast, are entirely covered by internal revenues, which include transfers, dividends and corporate taxes from Chukha Hydropower Corporation. See 16See Kuensel (2002). 17Note that as explained in Section 3.1.1 public investments in power projects are recorded outside the budget. 31 Figure 23. Bhutan: Government Capital Expenditure and Grants (as percentage ofGDP) 35 .-------------------~----------------------------------------__, 30 Government capital 25 20 Grants 15 10 5 Power exports Figure 24. Bhutan: Government Internal Revenues and Current Expenditure (as percentage of GDP) 30 .-------------------~----------------------------------------_, 25 Current expenditure 20 15 10 Total revenue excl. grants 5 Source: RMA Annual Reports. 32 Figure 24, which confIrms a strong one-to-one relationship between government current expenditure and total internal revenue. 3.3.2 Resource Movement Effect Finally, this subsection attempts to answer the remaining set of questions we are addressing in this section. Have there been resource movements from the tradable towards non-tradable sectors? Has there been a decline in output in the tradable sector as a result of relative price changes, that is, the Dutch condition? In the absence of sectoral data on investment and employment, it is not possible to examine resource movements from the tradable sector to the non-tradable sector. Therefore, we rely on data on sectoral production and trends of exports to examine whether relative price changes have affected the tradable sector in Bhutan. Figure 25 presents the sectoral contribution to GDP in constant 1980 prices for 1981-2000. According to Figure 25, all sectors, both tradable and non-tradable, have grown strongly in real terms. While growth of agricultural production slowed from the late 1980s onwards, agriculture remains the principal contributor to GDP, accounting for about 40 percent of value added in 2000 (see Figure 27).1 8 The modern sectors - services, construction (non-tradables), manufacturing, Figure 25. Bhutan: Components ofGDP (value added, in millions of ngultrum, constant 1980 prices) 1,400 ,--- - - - - - - - - - - - - - - - - - - - - - - - - - - - - - , 1,200 ~ ~,ooo 0- o 00 0- - 800 3 E '" .I:: "3 600 bIJ Z <.;- o ~ 400 ~ ~ 200 - - Agriculture --Mining - - - - - - . Manufa cturing - - - E lectricit y and gas - - - - Construction - - Serv ices - - - - - Government Source: RMA annual reports (various issues) and World Bank staff estimates. 18 The annual growth rate of agricultural production averaged at 6.0 percent for 1981-86 and at 2.9 percent for 1987-2000. 33 Figure 26. Bhutan: Index of Sectoral Output (1980 = 100) 11 1100 .-----------------------------------------------------------~ " ,.' 900 .' ,,. .. , 700 , J .' . 500 . , " , , " , ,., "J .. _ .. 300 , r - - -- ~ ,- - . ; 100 ~:;.; ~~-=~~ __ ~~L_ ~~ -- - -- - -- ..- -;,.- - ~~~- .~~~~~ ~ - -~~.~.~-------- - .•~> __________________________________________ --- ~ ----- Agriculture ----- Mining ....... Manufacturing - - - - Construction ----- Serv ices - .. - . Government 1/ Excludes the power sector. Figure 27. Bhutan: Components ofGDP (value added, as percentage of non-power GDP) 50 .-----------------------------------------------------------~ 40 30 20 - - - Agriculture ---Mining - - - - - .Manufacturing - - - Construction - - - Services - - - - Government Source: RMA annual reports (various issues) and World Bank staff estimates. 34 Figure 28. Bhutan: Real Exports and Real Non-power Exports (in millions of ngultrum, constant 1980 prices) 2,500 . ------------------------------------------------------------, 2,000 ~ '" " .g 0.. o 00 ~ 1,500 E '" .!:: '3 21,000 '+- o '" t: ~ Real non-power exports i 500 Source: RMA Annual Report and World Bank staff estimates. Figure 29. Bhutan: Real Exports by Country of Destination (in millions of ngultrum, constant 1980 prices) 2,500 . ------------------------------------------------------------, Real exports to 3rd countries~ ';;;' 2,000 " .g 0.. 0 00 Real power 0- exports to India ~ 1,500 2- § .!:: '3 bll ;z: '+- 1,000 0 '" t: Real non-power ~ exports to India E ..= \ II 500 o n n n n n n nI Source: RMA Selected Economic Indicators and World Bank staff estimates. 35 and mining (non-booming tradables) as well as hydropower (booming tradable) - have increased their value added over time (see Figure 26), indicating that the production structure of Bhutan has become slightly more diversified over the years. Growth of manufacturing and mining is most notable. During 1981-2000, manufacturing grew at an average rate of 13.5 percent a year, while mining grew at an even higher rate, 14.5 percent a year. Nevertheless, these two sectors are still in their infancy. Value added by manufacturing and mining remains still very small, respectively accounting for only 1.3 percent and 8.6 percent of non-power GDP in 2000. Meanwhile, trends of exports, presented in Figure 28, suggest that despite real appreciation of the ngultrum against the rupee, Bhutan's total exports have grown sharply - some 13-fold during the last two decades. A large part of this increase is attributable to power exports. However, non- power exports have also grown strongly in real terms, from an average ofNu.175 million in the 1980s to an average of more than Nul ,OOO million in the 1990s, both in constant 1980 prices (see Figure 28). Exports to countries other than India, mainly Bangladesh, which take place in US dollars, have also increased, although the magnitude is still small, accounting for about 5-6 percent of total exports (Figure 29). In conclusion, from exports and sectoral production data alone, we find no evidence to suggest that increased power exports - and foreign capital inflows - have led to a contraction of the tradable sector, despite a substantial decline in relative prices (real appreciation) over the years. There are two possible interpretations of these results. 19 One is that the non-tradable sector would have grown more rapidly in the absence of real appreciation than otherwise. Another is that, only to a small extent, tradables may have competed with non-tradables for scarce factors of production. For example, although labor is mobile between non-power tradable and non-tradable sectors, little rural labor may have been pulled away from agriculture (a tradable sector by definition, subsistence-based in many rural areas) to more profitable non-tradable sectors such as services or construction, both of which are concentrated in Thimphu. 3.3.3 Monetary Effect As discussed in Section 2, increased exports tend to raise the supply of money through balance of payments surpluses and net accumulation of reserves. Under the fixed exchange rate regime, without full sterilization of the increase in reserves, there will be a short-term monetary disequilibrium, which will cause inflation and consequently further appreciation of the home currency. As we show below, however, in the case of Bhutan there has been little monetary effect related directly to increased power exports. Although Bhutan has enjoyed large balance of payments surpluses and corresponding increases in international reserves, these were largely driven by large net foreign capital inflows, rather than by increased exports alone (Section 3.1.4). Therefore, the monetary effect is largely of net capital inflows, rather than of power exports. In the regression 19 The most general interpretation in such a result can be that the domestic sectors producing tradables are heavily protected by tariffs and quotas and are insensitive to relative price changes. The effective relative price the protected tradable sector faces is [eP T (1 + T)]/ p NT , where T is the tariff rate. Note that [eP T (1 + T) V N > ePT/ p N for positive T, meaning that the profitability oftradables is higher with the p T T tariff than without. However, since there is no trade barrier between Bhutan and India, to which a substantial proportion of manufactured goods and mineral products are exported, this interpretation is not applicable in the case of Bhutan. 36 Figure 30. Bhutan: Inflation Rates (in percentage per annum) 20 .0 ~------------------------------------------------------------, 15.0 10.0 5.0 ~~~~~~~~~~~#~~~~~~~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ " ~ ~ ~ ~ ~ ~ ~ Source: IMF International Financial Statistics and data provided by the Bhutanese authorities. Figure 31. Bhutan: Money Supply (M2) and Money Velocity 12 ,000 7.0 6.0 10 ,000 E '" ./:: '3 5.0 b1l 8,000 ;z: >, '-- 0 en Velocity of money 4.0 " t: 0 t: :2 ~ 6,000 (right scale) '-- 0 E .0 '(3 3 3.0 0 >, v 0.. 0- > ;;l 4 ,000 >, 2.0 " 0 t: :2 2,000 M2 (left scale) 1.0 o +-~---+---r--+-~---+---r--+-~---+---r--+---r--+---r--+-~---+O.O Source: IMF International Financial Statistics and World Bank staff estimates. 37 analysis of BRER in Appendix 2, what the monetary disequilibrium represented is likely the monetary effect of increased net capital inflows to Bhutan. Regardless of the source of the monetary effect, the inflation rate in Bhutan has remained moderate (see Figure 30). This is because although the money supply has grown substantially over the years, the increases have been offset by declines in the velocity of M2 (see Figure 31), creating excess liquidity in the financial system (that is, de facto sterilization). With limited investment choices, the demand for advances from the financial sector has remained stagnant. Such a build-up of liquidity in the banking sector may continue as long as surpluses continue in the external sector. 4. POLICY IMPLICATIONS We have discussed in Section 3 that the commencement of regular power exports to India brought enormous economic benefits to Bhutan - higher investment, faster growth, and rising living standards. However, in response to the large power export earnings and foreign capital inflows, the ngultrum has appreciated against the Indian rupee over time, mainly through increased private spending and monetary disequilibrium. Despite this real appreciation, however, there is little evidence to suggest contraction or stagnation of the non-power tradable sector. Although growth of the tradable sector might have been stronger in the absence of real appreciation, thus far, there is no sign of the Dutch condition in the Bhutanese economy. Bhutan 's medium-term growth prospects are generally favorable supported by large potential of exporting hydropower, which is a renewable natural resource. External vulnerabilities appear modest, with a strong commitment to financial assistance by Bhutan' s development partners. In addition, both trade and sources of foreign aid are more diversified today than in the early 1990s.20 Nevertheless, Bhutan is not completely free of risks. Fist, growth prospects of the country are dependent on how effectively the Royal Government of Bhutan manages the increasing power rents. Thus far the Government has used power rents in an effective manner, channeling substantial power exports and foreign aid to promote investment in human capital and infrastructure. It is viewed that such fiscal prudence have helped mitigate macroeconomic distortions typically seen in countries experiencing a natural resource export boom. Section 4.1 suggests a broad fiscal policy strategy for the effective use of power rents. The second risk is related to exchange rate management. With the commissioning of the Tala hydropower project in a few years, our empirical results anticipate further upward pressure on the real value of the ngultrum, which could discourage production and investment in the non-power tradable sector leading to the Dutch condition. This real appreciation is viewed as a long-term equilibrium phenomenon based on a renewable resource. With the strong expectation that the power sector will continue to dominate the country's exports, the non-power tradable sector may have a stronger incentive to get used to a strong real exchange rate than it would if the resource 20 Based on the experience ofIndia's BoP crisis, Bhutan has sought to diversify its sources offoreign aid; Japan, Denmark, Switzerland, the Netherlands, Austria and Kuwait now figure among its major donors other than India. 38 were non-renewable and the export boom were temporary. Then it will become important for the Government to control the speed of appreciation since a rapid real appreciation could distort the relative prices in the domestic economy. Section 4.2 discusses ways to contain the speed of real appreciation. The third risk is unemployment. As the economy will remain heavily dependent on exports, with few sectors competing with the power sector which creates few domestic jobs, the challenge for the Government is to create employment for Bhutan' s increasingly educated labor force by fostering private sector development. A related challenge is immigration. As domestic wages rise, there will be increasing pressure for labor to move from the neighboring countries into Bhutan. A broad strategy for private sector development is discussed in Section 4.3. 4.1 Fiscal Policies - Effective Use of Power Rents and Attaining Self-Sufficiency Despite recent fiscal easing, the Royal Government of Bhutan has maintained fiscal prudence, with domestic revenues more than covering recurrent expenditures and a maximum use of concessional financing for development expenditures. Over the last few years, more than a third of government expenditure was allocated to infrastructure, health and education. Much of that, however, was possible due to generous financial support from development partners. 2 1 While the risk of reduced foreign aid is viewed as very small, Bhutan' s long-term goal should be to reduce heavy reliance on foreign aid and to achieve self-sufficiency. This requires effective use of power income to expand the economy's future production capacity. 4.1.1 Expenditure Policy The first fiscal priority for the Government should therefore be to continue this fiscal prudence while maintaining essential expenditures for the delivery of public services. A substantial portion of power export proceeds and foreign aid should be channeled into the investment necessary to achieve balanced growth. However, caution needs to be taken so that the level of investment does not outstrip domestic implementation capacity, which would cause adverse effects on the efficiency of the capital invested (see Box 2, the experience of Botswana, which successfully used the income from diamond exports to achieve a higher growth rate). While the Government expects much of the investment to be made by the private sector (with subsidies, if any, being minimum), complementary investment needs to be undertaken by the public sector, especially in areas where strong externalities are present, such as health and education, and in other areas where large-scale financing is required, such as large infrastructure projects. Human capital formation should focus on upgrading technical skills that would help raise future labor productivity. The priority of infrastructural investment should be to improve transport in order to reduce the disadvantages of Bhutan' s landlocked and isolated location in serving regional and global markets.22 21 While the Bhutanese authorities have made concerted efforts to upgrade the country' s statistical database recently, data limitation remain especially in areas of prices, public finance and employment, hampering effective assessment of the economy on which formulation of policy strategies should depend. Strengthening of the statistical base is therefore also an important part of policy strategies for sustainable growth in Bhutan. 22 Especially, provision of efficient and reliable transport - both road and air transport - is expected to give a major boost to the agricultural sector by lowering the transport costs and assuring timely delivery, thereby strengthening the competitiveness of Bhutan's agricultural products in the world market. For example, in 39 Box 2. Botswana's Experience Botswana is one of the most successful countries in managing a natural resource export boom. The diamond boom in Botswana began in 1965 with the discovery and development of large amounts of high quality diamonds. Over the last three decades the rapid expansion of diamond exports has quadrupled the per capita GDP of Botswana. Botswana has not been affected by the Dutch disease or any other of the harmful consequences of a natural resource export boom. Its manufacturing production even doubled, albeit from a small base, to the point where it now employs three times as many workers as mining. There was a significant rise in construction prices, but there was no spillover of inflation to the rest of the economy. Agricultural production, however, seems to have suffered slightly. What did the Government of Botswana do? First, a large part of the windfall was put in foreign savings and was used only when the absorptive capacity of the economy was deemed sufficient. In determining the level of government spending, close attention was paid to the two constraints: (i) availability of skilled manpower; and (ii) the future recurrent costs of development spending in relation to revenue forecasts (Harvey, 1992). Second, very close attention was paid to the exchange rate variability. While the policy aimed at a stable real exchange rate, the nominal exchange rate was adjusted periodically to contain the inflationary pressures caused by the boom. Third, the level of public investment was maintained, but emphasis was placed on health and education, literacy, life expectancy, infrastructure and water facilities. The contribution of an improved infrastructure base (both physical and social) to economic diversification and growth lies in the stimulation and greater efficiency it promises for the future productivity and economic profitability of tradable and non-tradable goods. 4.1.2 Tax Policy Meanwhile, the Government should continue to vigorously expand its internal revenue base by broadening the tax base and streamlining the tax system to ensure its efficiency and effectiveness, with the aim of attaining fiscal self-sufficiency in the long-run. While power exports have contributed to the substantial increase in government non-tax revenue, Bhutan is still unable to raise enough internal revenue to reverse the trend of external fiscal dependence. The introduction of the Personal Income Tax in January 2002 is an important and timely step not only to expand the revenue base but also to ensure that the income disparity between the rich and poor does not widen considerably. 4.2 Exchange Rate Management the case oftimber products, the cost of transporting timber can be as high as, or even higher than, the value of timber themselves, eroding the price competitiveness ofthe Bhutanese timber. Another example is matsutake mushrooms, a time-sensitive product which needs to be delivered to the Tokyo market within 48 hours of being picked. In the absence of direct air route between Thimphu and Tokyo, and with unreliable domestic transport to reach the Thimphu Airport, exploration of the large Japanese market is severely constrained. 40 Power exports will continue to be the dominant source of export proceeds for Bhutan on a permanent basis. The associated real appreciation of the ngultrum is therefore a long-term equilibrium phenomenon. It is unavoidable and should not require automatic policy intervention. Nonetheless, too rapid a real appreciation, especially for a prolonged period, may substantially reduce incentives for production and investment in the non-power tradable sector and thus needs to be avoided. Appropriate exchange rate management should be to allow gradual real appreciation of the ngultrum over time. Under the present fixed exchange rate regime, there are broadly three possible ways to slow the speed of real appreciation: (i) to contain the growth of domestic demand for non-tradables; (ii) to expand the domestic supply ofnon-tradables; and (iii) to contain the growth of overall domestic demand by influencing disposable income. 23 The first two solutions mitigate the upward pressure on the prices ofnon-tradables by influencing domestic demand for and supply ofnon-tradables, whereas the third solution works by controlling overall domestic demand for both tradables and non-tradables. Any of these have the effect of slowing the fall in the relative prices of tradables. Specific policy options are provided below. Nominal devaluation of the ngultrum is not considered an option here. 4.2.1 To Contain the Growth of Domestic Demand for Non-tradables General policy recommendation for countries experiencing a permanent rise in exports is that governments should slow the pace of real appreciation by changing government expenditures. The rationale is that since public consumption and investment are a large part of domestic demand, lowering the proportion of government spending, especially recurrent spending, on non- tradables will mitigate the rise in domestic demand for non-tradables, thereby helping to contain the speed of real appreciation. However, this policy recommendation does not apply to the case of Bhutan straightaway. First, in Bhutan, it is the private sector rather than the public sector that tends to have a higher marginal propensity to spend on non-tradables (Section 3.2). It follows that the extent to which changes in government expenditure alone can influence the prices of non-tradables, and in turn the speed of real appreciation, is likely to be limited. Second, compressing government spending on non- tradables is not a viable option for Bhutan at this stage of development, since maintaining adequate levels of development expenditure is deemed crucial in expanding the country's future production capacity. The alternative can be to contain domestic demand by influencing private spending on non- tradables through higher taxation. Again, however, this option may not seem viable since higher taxation of non-tradables (construction, services and transport) may depress private sector investment, which will help transform the economy from aid-dependence to self-sufficiency in the long-term. 4.2.2 To Expand the Domestic Supply of Non-tradables 23 Recall that the bilateral real exchange rate of the ngultrum vis-a-vis the Indian rupee is simply the relative prices oftradables (p/ ) to non-tradables (pt \ BRER r = Pr T IP r NT , where p / is determined in the Indian market and Pt T is determined by demand and supply within Bhutan. Real appreciation (i.e., a decline in BRER) happens since, in response to increased domestic demand, the prices ofnon-tradables (PINT) rise relative to the prices oftradables (p/ ). This suggests that measures aimed at containing the rise in p/ can effectively reduce the speed and magnitude of the decline in BRER. 41 Second, expanding the supply of non-tradables will mitigate the upward pressure on the prices of non-tradables from increased demand for non-tradables, thereby slowing the rate of the ngultrum's real appreciation. This can be done indirectly by emphasizing public investment in key physical and social infrastructure sectors that would help expand the non-tradable sector's production capacity. Allowing larger labor migration into Bhutan's labor-intensive non-tradable sector, such as construction and services, is another option, since that has the effect of relaxing the factor market constraint in the non-tradable sector. 4.2.3 To Contain Rapid Growth of Disposable Income Third, the Government can slow the speed of real appreciation by influencing the growth of domestic disposable income, thereby limiting the rise in overall domestic demand. There are three possible approaches. One is through taxation, the second through sterilization, and the third through temporarily managing the size of export proceeds and net capital inflows to the economy. The last option could be done by temporarily maintaining some portion of export proceeds abroad24 ,25 or by temporarily suspending net capital inflows should real appreciation be viewed as too rapid. Such measures to restrict capital inflows should be used only in the short-term to slow the speed of real appreciation when it is deemed too rapid, but not to prevent real appreciation of the ngultrum, which is a long-term phenomenon. The power tariff and volume of power exports cannot be used as short-term policy variables since they are pre-determined at a politically fixed level with India. 4.3 Private Sector Development Private sector development holds crucial key to accelerating economic growth in Bhutan and addressing employment concerns in the medium term. Private sector development is also an effective means of addressing the excess liquidity in the financial sector. As private sector activities grow, the economy's capacity to absorb the increased power revenue will rise gradually, and the scope for higher tax collection by the Government will also improve. The authorities' long-term expectation is for the private sector to take up all commercial activity, other than projects that are too large to be financed privately. Despite the Government's efforts in privatization and deregulation, and its more recent moves to improve the stability and competitiveness of the financial sector, however, the private sector in Bhutan remains under-developed. Lending to the private sector has hardly grown. This could be one reason why a large part of power income has been used to finance private consumption rather than investment. Establishing a favorable business environment may divert part of the power income from private consumption to private investment. A broad set of policies can be implemented to stimulate private sector development. These include macroeconomic stability, trade liberalization, SME policies, labor policies, financial 24 In fact, this is not a new policy measure for Bhutan, although not as part of real exchange rate management. A large part of international reserves as well as pension fund is being held offshore with highly-rated financial institutions, individual contributions to the national pension fund are actively invested in foreign securities and equities, and the Bhutanese banks are allowed to maintain excess liquidity as deposits with commercial banks in India. 25 See, for example, Norway's State Petroleum Fund. Detailed information can be found at www.norges- bank.no/english/petroleum fundi. Also, see Fasano (2000) for cross-country experiences of oil stabilization and saving funds. 42 sector reform, provision of infrastructure, establishment of the FDI act, and improvement of the investment climate, to name a few. Most policy options for private sector development are discussed in the Private Sector Survey report, which was jointly conducted by the World Bank and UNDP in 2001 . Readers may consult the report for more in-depth analyses and sector-wide policy recommendations. 5. CONCLUSIONS This paper has empirically examined the impact of Bhutan' s increased hydropower exports on the domestic economy. It has shown that power exports have brought enormous economic benefits to Bhutan, spurring growth and raising living standards through increased export earnings. The results of this empirical analysis suggest that that power exports to India as well as net capital inflows have resulted in real appreciation of the ngultrum against the rupee over time, mainly through increased private spending and short-term monetary disequilibrium. Despite the large real appreciation, however, the study has found little evidence to suggest contraction or stagnation of the non-power tradable sector. While growth of the tradable sector might have been stronger in the absence of the appreciation, thus far there is no sign of the Dutch condition in the Bhutanese economy. Looking ahead, Bhutan's growth prospects are favorable, supported by enormous hydropower export potential. Unlike other types of natural resource exports based on exhaustible resources (oil, gas, coal, copper), Bhutan's hydropower is a renewable, permanent resource for export. Our empirical results indicate, however, that further increases in power exports as well as foreign aid - regardless of which currencies they are denominated in - will likely cause further appreciation of the ngultrum against the rupee, which could erode the competitiveness of the non-power tradable sector and lead to the Dutch condition. Although there is no immediate threat that requires urgent policy attention, the country's growth prospects will also depend on how effectively the Government uses the increasing power rents and manages the exchange rate. With the increasing role of hydropower in Bhutan's exports, there is an obvious question as to whether the Government should continue to seek policies for export diversification. While policy intervention may aid the growth of the non-power tradable sector, significant export diversification may not be a realistic aim for Bhutan. For a small country endowed with enormous hydropower export potential as well as a unique cultural heritage and delicate environment, a higher degree of specialization in power exports and tourism, similar to the strategy adopted by Brunei, may appear to be a natural outcome in the long-run. 43 ApPENDIX 1: MEASURING REAL EXCHANGE RATES This appendix describes the derivation of the bilateral real exchange rates of the ngultrum against the Indian rupee and the US dollar. A1.1 BRER(NU/INR) By definition, the bilateral real exchange rate of the Bhutanese ngultrum against the Indian rupee, BERE(NU/INR), can be expressed as: e (NuIINR) pT BRER (NuIINR) t = t p NT t t where et stands for the nominal exchange rate of ngultrum against the Indian rupee (annual average), ~T for the prices oftradables expressed in the rupee, and ~NT for the prices of non- tradables in Bhutan, in time t. Since the ngultrum is nominally fIxed to the Indian rupee at par, et is one for all periods, and therefore, BRERt (NuIINR) = ~T / ~NT . Namely, the bilateral real exchange rate of the ngultrum to the rupee is merely the relative prices oftradables and non-tradables. Since Bhutan is a small country, it takes the Indian prices oftradables, ~T (which are determined on the Indian markets) as given. The prices ofnon-tradables, ~T , are determined by domestic demand and supply in Bhutan. An increase in BRER(NU/INR) represents real depreciation of the ngultrum to the rupee, whereas a decline in BRER represents real appreciation. In this paper, we use the whole sale price index of India, WPlu , as a proxy for the prices of (non- power) tradables in India, and the consumer price index of Bhutan, CPIB •t , for the prices of non- tradables in Bhutan: WPI BRERt (NuIINR) = 1/ . (AI) CPIB,t In Equation A(l), one might argue that it would be ideal to use the Indian consumer price index, CPI, rather than the Indian whole sale price index, WPI, in deriving the BRER(NU/INR). We however believe that the WPI is a more appropriate measure, since a large part of items in the Indian CPI basket is non-tradables, particularly services, whereas the Indian WPI consists mostly of tradables, whose prices matter in calculating the bilateral real exchange rate. 44 Figure A 1. Bhutan: Bilateral Real Exchange Rates (NU/INR) (1980 = 100) 130 .-------------------------------------------------------------, 120 GDP deflator based BRER 110 100 90 80 WPIICPI based BRER 70 Source: IMF International Financial Statistics and World Bank staff estimates. As regards the prices of non-tradables in Bhutan, the consumer price index, CPIB,t, is not a perfect proxy since the consumer basket used for the CPI calculation consists of consumer goods only in Bhutan, but does not include services such as housing, transport and communication (non- tradables) that are usually included in the CPI calculation in other countries. 26 Despite this disadvantage, we view that Bhutan's CPI is a reasonably good indicator for the prices of non- tradables, since in the case of Bhutan the price of virtually all tradable goods brought into the country include high costs of transport, which is a non-tradable goods. The computation of BRER (NU/INR) using the Indian WPI and Bhutanese CPI yield a gradually declining real value of the ngultrum to the rupee, that is, gradual real appreciation of the ngultrum against the rupee, for 1980-99. We have also calculated BRER(NU/INR) using different price indices. First, as an alternative proxy for the prices oftradables, India' s consumer price index is used. Second, we have used the GDP deflator, which could be a better indicator to capture inflation. The results of the computation are presented in Figure A 1. 26The consumption basket was defined in 1979 and has not been revised since then. The basket consists of 72.3 percent offood items and 27.7 percent of non-food items, such as clothing and footwear, stationary goods, furniture and utensils, etc. 45 Figure A 1 shows three distinctive paths of the bilateral real exchange rate of the ngultrum against the rupee over the last two decades. These results are driven mainly by the choice of price indices. In contrast to the declining WPIICPI-based BRER(NU/INR), the GDP deflator-based BRER(Nu/INR) indicates real depreciation of the ngultrum to the rupee during 1989-95, followed by real appreciation. On the other hand the CPIICPI-based BRER (NU/INR) shows a relatively steady movement of the bilateral real exchange rate of the ngultrum against the rupee. Despite this confusing result, there are good reasons to believe that the real value of the ngultrum has appreciated against the rupee over time, caused largely by the rising prices ofnon-tradables in Bhutan. For example, although data are not readily available, it is generally perceived that the Bhutanese wages - the price of a non-tradable good (labor) - have been rising rapidly, thereby attracting a large number of illegal Indian workers into the country. Both land and housing prices are also on the rise. The validity of the CPIICPI-based and the GDP deflator-based BRER (NU/INR) is therefore very unlikely. Thus we continue our analysis with the understanding that the WPIICPI-based BRER(NU/INR) may still be an imperfect measurement of the bilateral real exchange rate of the ngultrum against the rupee. A1.2 BRER(Nu/US$) In a similar manner, the bilateral exchange rate of the ngultrum against the US dollar, BRER(Nu/US$), is defined as: e (Nu/US$) WPI BRERt (Nu/US$) = t us,! CPIB,f et (INR/US$) WPIUS,f = where et (Nu/US$) is the nominal exchange rate of ngultrum to the US dollar, which is equal to the nominal exchange rate of the Indian rupee to the US dollar, et (INR/US$), and WPlus"t is the US whole sale price index, which is used as a proxy for the world price of (non-power) tradables. Again, BRER(NU/INR) may not be a perfect measurement of the bilateral real exchange rate of the ngultrum to the US dollar, given the limited coverage of prices for CPI calculation. 46 ApPENDIX 2: EMPIRICAL ANALYSIS: DETERMINANTS OF BILATERAL EXCHANGE RATE OF THE NGULTRUM A2.1 Introduction This Appendix provides a detailed description of the regression analysis conducted in Section 3.2.4. We examine the relationship between the bilateral real exchange rate of the ngultrum to the Indian rupee, BRER(NU/INR), and power exports using an econometric analysis. Specifically, we investigate the following three issues : (i) whether the increased power export revenues have affected the movements ofBRER(NU/INR) significantly; (ii) what other variables have affected BRER(NU/INR); and (iii) through what channels the power exports have influenced BRER(Nu/INR). A2.2 The Model (i) BRER Dynamics The modern analyses of real exchange rate behavior distinguish between fundamentals and monetary determinants of BRERs. This distinction is captured empirically in this equation: (A2) where e* stands for the equilibrium level ofBRER, e for the actual BRER, :_ M for the 1 (nominal) supply of money in period t-1, Mt~l for the money demand in t-1 , and DEV for the nominal devaluation. e, p and ¢ are positive parameters. In the case of Bhutan, DEVt = 0 for all t, since the ngultrum is pegged to the Indian rupee at par. In Equation (A2), the observed BRER movements (~log et ) are the result of two forces: e(log ; t e -log e- 1) and - p(log M LI -log Mt~l). The first term, e(log e -log e- ; t 1) , is a partial adjustment in response to discrepancies between the equilibrium BRER, log e * , and its lagged actual value. If the equilibrium BRER is above (below) its actual value, BRER will increase (decrease) accordingly towards its equilibrium level. The speed of the adjustment will depend on the parameter e. In the case of an overvaluation ( e * > e), the adjustment will be rather slow because an "automatic" decline in e will normally require a drop in the nominal price ofnon-tradables and wages. To the extent that nominal prices show some degree of downward resistance, the adjustment will take time. The second term, - p(log M LI -log Mt~l) ' captures the impact of monetary disequilibrium on BRER and indicates that monetary pressures will cause real BRER appreciation (~log et < 0). Empirically, this impact is given by the coefficient p. If the power export boom and net capital 47 inflows from India affect the money supply process, M /_l' BRER will be affected accordingly. This would affect the dynamics ofBRER in the short-run, but not the long-run equilibrium of BRER. (ii) Long-Run Equilibrium BRER Equation (A2) suggests that to explain the behavior of the BRER (e), we need to understand what the evolution of the equilibrium BRER (e*) is. Without loss of generalization, the equilibrium BRER can be expressed in the following manner: where P/ is the real price of power, V/ is the volume of power export, G; is the government expenditure on non-tradable goods, NFAt is net foreign aid inflows, and It is the measure of import tariffs all for period t. Since there exist no trade restrictions between Bhutan and India, It= 0 for all t. The coefficients ~, ~ ,,,1,3 and ,,1,4 are expected to be negative. The effects of changes in the fundamentals on the equilibrium BRER are explained below. Real price of power (P/): An increase in the price of power is expected to lower BRER and lead to real appreciation of the ngultrum. This is mostly due to the income effect explained in Section 2.1. The export boom induces higher expenditure on tradables and non-tradables, putting pressure on the prices of non- tradables, and causes real appreciation. The key is that the response of non-tradable prices to a power export boom will exceed that of tradables. This is independent of the exchange rate system. Because in Bhutan the power company is a public sector undertaking, the direct impact of a positive price change on BRER will be smaller than if the resources were privately owned. When the price of the enclave export increases, most of the export revenue goes to the Government, in the form of an increase in the surplus transfer, corporate tax, and dividend. Consequently, a significant proportion of the Dutch-condition effects is expected to take place via changes in government expenditure, a channel captured by log G; . Besides, a direct power price effect is also expected to take place since increases in the price of power will generate a positive "confident effect" on consumers, affecting aggregate expenditure. Vaey-Zadesh (1989) makes this point using a model based on a "warehouse" approach to oil exploitation. Therefore, this analysis of the effect of the power boom on BRER considers the multiple channels through which the Dutch condition operates. Volume of power export (V;): For similar reasons, an increase in the volume of power exported to India is likely to result in a real appreciation of the ngultrum, given the other factors. Government expenditures on non-tradables (G; ): Higher power export revenues are expected to affect the level of government expenditure significantly. The dependence of G; on power export revenues would then constitutes an 48 indirect but predominant channel through which the power export will tend to affect BRER. An increase in government expenditures on non-tradables will raise both their demand and their relative prices, thereby appreciating the home currency. Net Foreign Aid Inflows (NFAt ): Higher net foreign aid inflows will result in higher disposable income and aggregate spending. If a proportion of this higher spending is on non-tradable goods, prices of non-tradables will rise, appreciating the real value of the ngultrum. (iii) Specification for Econometric Analysis Combining Equations (A2) and (A3) yields the following equation for the dynamics ofBRER, which in principle can be estimated using conventional methods: log et = (J 0 + (JI 10g}~P + (J 2 log Z: + (J3 log G:" (A4) + (J410g NFA t + (J5 log et_1 - p(logM/_I -logMt~I)+ ct ' where and & is an error term. In Equation (A4), there are at least three channels through which the power export boom can impact the level of BRER. The first channel is disposable income and is captured by the N coefficients (JI and (J 2 ' The second channel is changes in government expenditures, G t and is captured by coefficient (J3' Another channel is net foreign aid inflows, NFAt , and is captured by coefficient (J 4 ' The fourth channel relates to the monetary disequilibrium, which is of short- term in nature, and is captured by p. One problem of estimating Equation (A4), however, is the multicollinearity problem among the regressors. Although we have found no multicollinearity among ~p , V/ , NFAt and the monetary variables, we have found significant collinearity between ; G and NFAt, and M/ and NFAt . Namely, the level of government expenditure on non-tradables and the magnitude of monetary disequilibrium are significantly influenced by the level of net foreign aid Bhutan receives every year (see Figure 23 , p.32). To address this issue, we have decided to drop regressor NFAt in our estimation. A2.3 Regression Analysis We estimate Equation (A4) and its variations using annual data for the period of 1983-99, which is the longest period available for all the variables. 27 Given the imperfect measurement of the real 27 While data on other variables are available for the period 1981-99, the monetary data are available only from 1983 onwards. 49 exchange rate, BRER(NU/INR), and also the limited number of observations (17), the estimation results need to be taken as an indication only, but should not be used to draw a solid conclusion. Description of variables and data sources are provided in Section A5. The results of the empirical estimation using OLS and 2SLS are presented in Table A 1.28 The results support the hypothesis that in the short-run the BRER movements in Bhutan have responded to both the real and monetary factors. First, the results of the regression analysis suggest an important relationship between power exports to India and the bilateral exchange rate of the ngultrum against the rupee. Specifically, increases in the power export tariff (P/) and volume of power export (V/) have been significantly associated with the ngultrum's real appreciation against the rupee, implying that future increases in either variables may likely cause further real appreciation of the ngultrum to the rupee. In all Table A 1. Regression Results: Bhutan: Dynamics of the Bilateral Real Exchange Rates Equation No. (a) (b) (c) (d) (e) Method OLS OLS OLS 2SLS 2SLS Constant -0.031 -0.033 -0.026 -0.015 -0.005 (-1.510) (-1.629) (-1.188) (-0.433) (-0.144) Log BRER(t-l) 0.823 0.843 0.809 0.688 0.631 (6.093) (6.319) (5.828) (2.501) (2.160) Log Real Power Tariff (t) -0.059 -0.058 -0.059 -0.090 -0.096 (-2.898) (-2.924) (-2.853) (-1.701) (-1.705) Log Power Export Vol.(t) -0.013 -0.013 -0.013 -0.021 -0.023 (-2.420) (-2.543) (-2.462) (-1.612) (-1.615) D[Log -0.057 -0.055 Gov.Cons./GDP(t)] (-1.220) (-1.056) D[Log Total -0.021 -0.022 Gov.Exp./GDP(t)] (-0.738) (-0.672) Monetary Diseq.(t) -0.046 -0.054 -0.045 -0.052 -0.0423 (-4.573) (-4.538) (-4.407) (-3.772) (-3.530) R-squared 0.948 0.954 0.951 0.911 0.936 S.E. of regression 0.017 0.017 0.017 0.Dl8 0.020 Note: t-statistics in parentheses. 28 From the augmented Dickey-Fuller test, we found all variables, except government consumption and total government expenditure, are 1(0). Both time series of government consumption and total government expenditure exhibit l(l). To get around this problem, we used the first difference of government consumption in the regression specification so that OLS and 2SLS are consistent. 50 regressions, the coefficients of the log of the real price of power and the volume of power export are negative as expected, and are statistically significant. A formal test of the equality of the coefficients of these two variables (Wald test) rejects the null hypothesis that CJl = CJ2. According to Equation (c), a one percent increase in real power tariff will cause a 0.59 percent decline in BRERt(NulINR), or real appreciation of 0.59 percent. Similarly, an increase in the power export volume by one percent will lead to a 0.013 percent real appreciation of the ngultrum. Second, the results suggest that the ngultrum's real exchange rate has also responded significantly to the monetary disequilibrium term, (108M-l s - 108M-l d). The coefficient of monetary disequilibrium shows a negative sign, as expected, and the t-statistics indicate that the estimates are significant. As we discussed in Section 3, the increase in money supply in Bhutan has been affected by overall net capital inflows and power export earnings, rather than by power export earnings alone. According to Equation (c), a one percent increase in monetary disequilibrium will appreciate the real value of ngultrum against the rupee by 0.045 percent. Changes in government expenditures do not appear to have had a significant impact on the level ofBRER. Neither variables - government current expenditure nor total government expenditure scaled by GDP - are significantly correlated with BRER although the sign of the coefficients for are negative as expected.. There are three possible interpretations of this result. The fist is that a large part of spending by the government may have been on tradables, instead of non-tradables. The second is that the private sector' s spending on no-tradables may have been stronger than the public sector' s. The third is that neither government current expenditure nor total expenditure is a good proxy for government expenditure on non-tradables. This analysis is however subject to limitations. First, as mentioned earlier, the WPIICPI-based BRER is an imperfect measure of the real exchange rate. Second, given the limited data availability, the number of observations is small for this type of time-series analysis. Thus caution needs to be exercised in interpreting the results. Third, some of the independent variables can be measured with error. Additionally, some of them are potentially endogenous. To address issue, instrumental variables (2SLS) were also used in Equation (d). However, the 2SLS results do not differ significantly from the OLS results. This suggests no endogeneity problem. Fourth, the BRER equation does not incorporate explicitly the role of expectations about the behavior of the power exports. In principle, expected changes in power exports will affect BRER by influencing the domestic spending. This issue is not addressed in this analysis because of the difficulties in finding satisfactory data on expectations. A2.4 Conclusions In summary, the real appreciation of the ngultrum has been caused by power exports and monetary disequilibrium associated with large net capital inflows. An expansion in power exports or net capital inflows in the future would likely put further upward pressure on the real value of the ngultrum against the Indian rupee. Despite limitations, the econometric analysis suggests an important relationship between power exports to India and BRER of the ngultrum. Specifically, increases in the power export tariff and 51 the volume of power export have been associated with appreciations of the ngultrum, indicating that future increases in power exports may likely cause further appreciation of the ngultrum against the Indian rupee. Besides, the bilateral exchange rate of the ngultrum has responded strongly to the monetary disequilibrium term. If increased net capital inflows - regardless of what they are denominated in - and power exports continue to strengthen the balance of payments position with India and to increase international reserves, without full sterilization, money supply could increase beyond the demand for money. Short-term monetary disequilibrium will prevail and the ngultrum will appreciate further. However, the results indicate that changes in government expenditures did not have significant bearing on the level of BRER. A2.S Definition of Variables and Data Source The study requires annual data of various macroeconomic indicators as well as the information on power exports to India for the period 1983-99. The definition of variables and their data sources are provided below. Bilateral real exchange rate of the ngultrum to the rupee: See Section 3.2 of the main text. Also, see Appendix 1. This variable appears a stationary series. Real price of power: This is the nominal price of power deflated by the WPI in India. The data on the nominal price of power are provided by the Bhutanese authorities. Volume of power exports: Power exports to India in mega units (MU). Data are provided by the Bhutanese authorities. Government expenditure on non-tradables: This is a difficult variable to measure since dis aggregated data on government expenditure are not available for the entire period. Two variables are used: (i) government current expenditure as a proportion of GDP; and (ii) total government expenditures as a proportion ofGDP. Both data are taken from various issues of Statistical Yearbook ofBhutan published by the Bhutan Central Statistical Organization. Monetary disequilibrium: This is the difference between the log of money (Ml) and the estimated log of the long-run demand for money. This is calculated with the estimated parameters obtained from a demand for money equation, (M / p)~ = f31y f ' exp(- f33i, ), using data for 1983-99. 29 29The estimated money demand equation for Bhutan is In mf = 16.4 -1.60 In y, = O.l3il' where mId is real money demand, y, real income and i, the inflation rate for time t. Since all variables are 1(1), the Johansen co integration test was employed to estimate the money demand equation. 52 REFERENCES Bruno, M. and J. Sachs (1982), "Energy and Resource Allocation: A Dynamic Model of the 'Dutch Disease"', Review of Economic Studies, 49 (special issue): 845-59. Buiter, W. and D. Purvis (1983), "Oil, Disinflation and Export Competitiveness", in Economic Interdependence and Flexible Exchange Rates, by J. Bhandari and B. Putnam (eds.), Cambridge, MA, MIT Press, 1983. Corden, W. M. (1984), "Booming Sector and Dutch Disease Economics: Survey and Consolidation", Oxford Economic Papers, 36:359-80. Corden, W. M. and J. P. Neary (1982), "Booming Sector and De-Industrialization in a Small Open Economy", Economic Journal, 92 (December 1982): 825-48. Eastwood, R. K. and A. J. Venables (1982), "The Macroeconomic Implications of a Resource Discovery in an Open Economy", Economic Journal, 92 (June 1982): 285-99. Fasano, U. (2000), "Review of the Experience with Oil Stabilization and Savings Funds in Selected Countries", IMF Working Paper, WP/001l12, June 2000. Gelb, A. (1986), "Adjustment to Windfall Gains: A Comparative Analysis of Oil-exporting Countries", in P. Neary and S. van Wijnbergen (eds.), Natural Resources and the Macroeconomy, MIT Press, Cambridge, MA. Harvey, Charles (1992), "Botswana: Is the Economic Miracle Over?", Journal ofAfrican Economies, 1(3):335-68. Kuensel (2002), "Soaring Land Prices: Real or Mere Speculation?", April 19, 2002. Neary, J. P. and S. van Wijnbergen (1984), "Can an Oil Discovery Lead to a Recession? A Comment on Eastwood and Venables", Economic Journal, 94 (June 1984): 390-95. Neary, J. P. and S. van Wijnbergen (1986), Natural Resource and the Macroeconomy, Oxford, UK, Basil Blackwell. Syrquin, M. and H. B. Chenery (1989), "Patterns of Development 1950-83", World Bank Discussion Paper, No. 41, World Bank, Washington, DC. Vaey-Zadeh (1989), "Oil Wealth and Economic Behavior: The Case of Venezuela, 1965-81", IMF Staff Papers, 36(2): 343-84. van Wijnbergen, S. (1984), "Inflation, Employment and the Dutch Disease in Oil-Exporting Countries: A Short-Run Disequilibrium Analysis", Quarterly Journal ofEconomics, 99 (May 1984): 233-50. 53 World Bank (2000), Country Assistance Strategy for Bhutan, Report no. 20158BHU, February 22,2000. World Bank (2002), Bhutan: Private Sector Survey, June 14,2002. 54 WBll1603 L\Bhutan\Dutch Disease\Paper\Bhutan Dutch Disease Paper (Final.doc November 25 , 2002 10:24 AM 55