Risk and Capital Flight in Developing Countries

Liliana Rojas-Suarez

rivate capital flight from developing countries lenders of the risks of holding the domestic and P has been of concern to policymakers, especially external debt of indebted developing countries. since the emergence of the debt crisis and the associ While different perceptions of the risk of holding ated drastic decline in capital inflows from indus domestic assets can explain the simultaneous oc trialized countries. Capital flight has been viewed as currence of large inflows of foreign capital and of a constraint on economic growth because it implies a large capital flight from developing countries dur loss of resources that could be used for domestic ing the 1970s and early 1980s, the emergence of the investment. Moreover, it is often argued that a re debt crisis and the accompanying policy responses versal of these capital outflows could significantly reduced such differences and resulted in a decline contribute to the solution of the debt crisis, and of foreign capital inflows coupled with a continua thereby to renewed access by developing countries tion of capital flight. to international capital markets. These considera This study first provides estimates of capital tions have led the authorities to consider policies flight for a group of developing countries with re that encourage the repatriation of capital flight or at cent debt-servicing problems, and then discusses least to stop such outflows. However, identifying the the determinants of capital flight and examines policiesthat can be most effective in achieving these measures of the risk of default associated with objectives depends crucially on what factors initi holding domestic financial instruments. It considers ated the capital outflows in the first place. alternative policies for reducing capital flight, and This study reviews the factors that have been finally the main conclusions are summarized. identified as stimuiating capital flight from de veloping countries. In this analysis, capital flight is associated with the fraction of a country's stock of Estimation of Capital Flight external claims that does not generate recorded in Previous empirical studies have employed a vestment income. Such external claims therefore broad range of definitions of capital flight. Some do not generate a stream of income that can be authorsl have adopted a "narrow" approach that used to service foreign debts or to financedomestic identifies capital flight with short-term speculative investment. Capital flight is thusdistinguished from capital outflows. Others2 have adopted a "broad" "normal" outflows of capital that would be under definition that identifies capital flight with total pri taken to achieve portfolio diversification and that vate capital outflows. An alternative approach would yield a recorded flow of income. The analy based on a "derived measure" identifies capital sis in this study suggests that increased risks in the flight with the fraction of a country's stock of exter domestic economic environment are likely to be nal claims that does not yield recorded investment key factors generating capital flight. In particular, income.3 This latter definition implies that a capital two types of riskmay have been particularly impor outflow should be considered capital flight only if tant: (1) default risk associated with the expropria it limits the resources available for either ser tion of domestic assets; and (2) the risk of large vicing the country's external debt or financing de losses in the real value of domestic assets as a result velopment programs. Numerous studies have com of economic policies that lead to rapid inflation or pared and critically evaluated these alternative to large exchange rate depreciations. Indeed, it is argued that the pattern of foreign capital inflows 1See Cuddington (1986). and capital flight from developing countries in the 2See, for example, World Bank (1985), Morgan Guaranty 1970s and 1980s can be associated with changing Trust Company (1986), and Duwendag (1987). perceptions by domestic residents and foreign 3See Dooley (1986). I

©International Monetary Fund. Not for Redistribution IV RISK AND CAPITAL FLIGHT IN DEVELOPING COUNTRIES

measures,4 and no general agreement has been capital flight for this group of countries, which reached on the relative superiority of each.s This amounted to $47 billion at the end of 1978, in study uses the "derived" measure to provide up creased continuously during the period and dated estimates of capital flightin developing coun reached $184 billion at the end of 1988 (Table 1).7 tries that have faced debt-servicing problems, since Although the stock of capital flightshowed a sus this measure provides the most direct estimate of tained increase over the period, the rate of change the economy's loss of resources that could poten of capital flight did not follow a stable pattern. As is tially be used for domestic investment. further discussed below, expansionary fiscal and Empirical estimates suggest that during 1978-88 monetary policies coupled with an increasing over the stock of flight capital increased for a group of valuation of the exchange rate resulted in high rates developing countries that had faced debt-servicing of increase in the stock of capital flight during the problems (see Table 1).6 The aggregate stock of period 1978-83.8 The adoption of stabilization pro grams in some of these countries reduced somewhat the rate of increase of capital flight during 198�6, 4See, for example, Deppler and Williamson (1987), Gordon particularly in some Latin American countries and Levine (1989), and Cumby and Levich (1987). SPerhaps the most severe criticism of all the proposed mea which undertook comprehensive adjustment pro surements is contained in Gordon and Levine (1989). They ar grams in 1985 and 1986. These programs, which in gue that severe statistical problems prevent the proposed mea cluded strong contractions of the fiscal deficits and surements from adequately capturing the scale of capital flight. major devaluations of the exchange rate, resulted in 6The methodology for estimating capital flight involves com puting the stock of external claims that would generate the in come recorded in the balance of payments statistics and sub generated by subtracting the stock of external debt implied by tracting this stock from an estimate of total external claims (see the flows reported in the balance of payments from the stock of Dooley (1986)). Total external claims arc estimated by adding external debt reported by the World Bank. the cumulative capital outflows, or increases in gross claims, 7The countries included inthis group and in Tables 1 and 2 are from balance o.f payments data (which consist of the cumulative Argentina, Bolivia, Chile, Colombia, Ecuador, Gabon, Jamaica, outflows of capital recorded in the balance of payments plus the Mexico, Nigeria, Peru, the Philippines, Venezuela, and Yugoslavia. cumulated stock of errors and omissions) to an estimate of the Bin fact, for this group of countries, the rate of increase in the unrecorded component of external claims. This last estimate is stock of capital flight reached 24 percent in 1983.

Table 1. CapitAl Flight In Ratio to Total External Debt and Total External Claims for a Group of tllghiJ Indebted Developing Countries (In billions of U.S. dollars)

Ratio of Capital Ratio of Capital

External Flight to Total External Flight to Total Capital Flight1 Debt External Debt Claims2 External Claims

(1) (2) (1) + (2) (3) (1 )+(3)

1978 47.30 113.70 0.42 71.62 0.66 1979 64.14 141.76 0.45 92.17 0.70 1980 75.41 179.06 0.42 1 19.41 0.63 1981 85.16 224.26 0.38 134.16 0.63 1982 99.95 261.30 0.38 142.57 0.70 1983 123.77 285.70 0.43 164.13 0.75 1984 136.43 301.05 0.45 182.52 0.75 1985 147.54 311 .60 0.47 199.70 0.74 1986 152.67 326.39 0.47 210.30 0.73 1987 180.62 349.75 0.52 233.73 0.77 1988 184.01 360.43 0.51 239.14 0.77

Sources: World Bank, World Debt Tables, various issues; International Monetary Fund, Bolonce of Payments Yearbook, various issues; and IMF staff estimates. ' Data refer to net capital flight, that is, the unrecorded stock of capital outflows less the unrecorded stock of capital inflows. For methodology used, see fn. 6 in text. 2The stock of external claims is defined as the net stock of recorded claims on nonresidents other than direct investment plus the net stock on unrecorded claims of residents.

©International Monetary Fund. Not for Redistribution Estimation of Capital Flight

a sharp decline in the rate of growth of capital flight, which reached only 3 percent during 1986. Table 2. External Debt and Resou�e However, many of these programs were aban TNnsfer • Fhulndng C0111pNatts af doned in 1987; fiscal deficits expanded once more, Capital Flight for a Group af HlghiJ and inflation accelerated. As a result, the rate of growth of capital flight increased again a d reached ...... ,..,...... Countries. � 1� 18 percent during 1987. It decelerated m 1988 as (In billions of U.S. dollars-4nnual averages) some major countries initiated new adjustment programs. It has been argued that given the magnitude of Change in Change in I capital flight, repatriation of those capital outflows, Stock of Total or at least of the investment income that they gen Capital External Resource erate, could significantly contribute to solvi�g the Flight Debt Balance1 external debt problems faced by these countnes. A better understanding of the importance of capital 1979-82 13.16 36.90 -4.50 flight relative to the countries' e�ternal fin�ncial 1983-85 15.86 16.77 25.50 positions can be gained by analyzmg the ratios of capital flight to total external debt and total exter 1986-88 12.16 16.28 11.90 nal claims (Table 1 ). During 1978-82, the ratio of capital flight to total Sources: World Bank, World Debt Tables, Appendix I, external debt declined from 42 percent in 1978 to various issues; International Monetary Fund, Balance of 38 percent in 1982, as the large inflows of foreign Payments Yearbook, various issues; and IMF staff estimates. capital to this group of developing countries more • Defined as net exports of goods and nonfactor than offset the increase in capital flight. This trend services. reversed during 1983-88, as the ratio of capital I flight to total external debt increased from 43 per cent in 1983 to 51 percent in 1988. This increase was the result of both a continuous increase in the stock of capital flight and a reduction in the amount tion in the country's stock of international reserves or through an increase in net exports, thereby re of new private foreign lending available to de ducing the resources available to sustain economic veloping countries. growth. Data available support thi hypothesis The ratio of the stock of capital flight to total � . During 1979-82, for example, a large mflowof for external claims increased from 66 percent in 1978 eign private capital to the indebted developing to 70 percent in 1982, consistent with the accelera countries occurred, and their total external debt tion of capital flight during this period. Except for rose faster than the estimated stock of flight capital some temporary declines in 1985 and 1986, when (Table 2). As a result, the aggregate for this group comprehensive adjustment programs were succ ss _ � of countries showed a negative resource balance, ful in reducing the rate of expansiOn of capttal implying that at least some of the external inflows flight, the ratio of capital flight to total external were used to finance imports.Io This picture claims continued to increase during 1983-88, reach- changed drastically after 1982: although the total ing 77 percent in 1988. . . external debt of these countries continued to in Despite the size of these estimates of the pnvate crease (primarily because of inflows of official holdings of external assets, it has been argued that funds), capital flight resulted in net transfers of re this capital flight will have a highly adverse eff�ct sources in most of the countries. This result implies on an economy only if it generates a substantial that the continuation of capital flight after 1982 transfer of real resources.9 For example, in periods coupled with the deceleration of external loans led when capital flight was offset by an inflow of for to a net decline in imports, which imposed an im eign loans, the proceeds from exports a�d other portant constraint on growth.tt external inflows were still used to finance Imports, and therefore the impact of capital flight on growth IOJmportant exceptions are Argentina an� Vene�uela. was not necessarily severe. However, when access Rodriguez (1987) argued that in Venezuela the mcrease m ex to external credit became limited after the emer ternal debt nearly matched the increase in the stock of flight gence of the debt crisis in 1982, greater capital capital. 11 Although the resource balance for most of these countries flight had to be "financed" either through a reduc- became positive after 1982. they experienced current account deficits in part because of interest payments on their external 9See, for example, Oeppler and Williamson (1987). debt.

©International Monetary Fund. Not for Redistribution IV RISK AND CAPITAL FLIGHT IN DEVELOPING COUNTRIES

Table J. C.pltal lmportlng Developing Countries with Recent Debt-Servicing Problems: MaCfOeCOIIGIIIIe Va rWIIes Affecting Capital Flight as Suggested In Literature, 1978-88

1 978 1979 1980 1 981 1982 1983 1984 1985 1986 1987 1988

Inflation rate1 28.26 31 .54 35.94 41.80 46.51 64.93 80.35 83.16 61.05 85.66 163.91

Real GDP growth' 3.85 5.31 3.80 -0.98 0.14 -1.79 2.59 3.32 3.48 2.39 1.71

Central government

fiscal deficits2 (in percent of GDP) 2.03 1.14 2.25 4.94 4.43 3.90 3.06 2.50 3.86 4.50 3.89

Real effective exchange rate2 (1980 = 100) 81.80 89.25 100.00 107.87 94.20 86.78 93.66 90.76 67.48 55.26 56.14

Total of current account' balances (in billions of U.S. dollars) -29.62 -23.38 -27.93 -82.26 -83.28 -31.76 -1 5.50 -10.96 -33.38 -1 5.99 -18.28

Disbursements of foreign loans as ratio to foreign debt2 26.48 28.40 21.93 23.49 17.45 11.93 8.06 7.24 7.68 6.80 6.63

Domestic real interest rate2 -4.57 -7.32 -15.52 -22.97 -24.08 -15.47 21.17 26.24 -27.96 -16.00 26.53

Interest rate differential in favor of foreign assets2,3 -3.09 0.23 13.53 32.93 27.99 12.37 8.48 2.31 20.61 5.26 8.41

Sources: International Monetary Fund, International Financial Statistics, and World Economic Outlook. 1 Countries included are those clas.sified, for purposes of the World Economic Outlook, as capital importing developing countries with

recent debt-servicing problems. 2For countries included, see footnote 7 in the text. 30efined as the six-month U.S. Treasury bill rate adjusted for the observed exchange rate change minus the domestic deposit rate.

Risk and Capital Flight unsustainable fiscal and monetary policies have been identified as the major causes of capital flight. The recent literature explaining the causes of For example, if expansionary monetary and fiscal capital flight can, in general, be divided into two policies created an overvalued exchange rate, do groups. The first (see, for example, Cuddington mestic agents would expect devaluation of the ex (1987), Dornbusch (1985), and Duwendag (1987)) change rate to occur eventually, leading them to has typically based its analysis on standard port shift out of domestic assets into foreign assets. folio models where agents are assumed to allocate Moreover, a large fiscal deficit financed by mone their wealth to maximize the overall risk-adjusted tary creation would create both inflation and incen return on their portfolios. In this context, capital tives for capital flight as agents attempted to pre flight has been explained in terms of the effects of vent losses in the real value of their domestic asset domestic macroeconomic variables on the relative holdings. returns between domestic and foreign assets. Since Econometric studies relating capital flight to the domestic interest rates in a number of the indebted macroeconomic variables, which have been sug developing countries that experienced capital flight gested by the first group as being responsible for were subject to controls,12 an overvaluation of the capital flight, have yielded divergent results.13 For exchange rate, rapid inflation, and inconsistent and instance, Cuddington (1986) found that currency

t2Jn recent years, several countries, including Mexico and l3Table 3 summarizes the behavior of some of these variables Bolivia, have undertaken financial liberalization. during 1978-88. II

©International Monetary Fund. Not for Redistribution Risk and Capital Flight

overvaluation (measured as the deviation of the Khan-Haque hypothesis of capital flight by linking actual real exchange rate from its equilibrium the risk of expropriation (which in his model is level) was a significant variable explaining capital identified with high taxation) of domestically flight in Argentina, Mexico, Uruguay, and Venezu owned assets to the existence of public and publicly ela during 1974-84, while Meyer and Bastos guaranteed private foreign debt. Eaton argued that Marques (1989) concluded that the inflation rate as the stock of publicly guaranteed private foreign and the real return on domestic assets were the debt increases, the emergence of any factors that major determinants of capital flightin Brazil during increase the probability of a private borrower de 1971-88. However, one "identification" problem in faulting would also lead other residents to expect these studies is that macroeconomic policies are higher future taxes as the government assumed the likely to influence both "normal" capital outflows obligations of the insolvent borrower. Domestic and capital flight. In addition, while this approach residents would therefore have an incentive to can basically explain the outflows of capital from place their funds abroad. developing countries, it cannot explain the simul The role of fiscal rigidities in creating risks was taneous occurrence of capital flight and the in emphasized by lze and Ortiz (1987), who examined creased inflow of foreign loans during the 1970s capital flight using a fiscal framework where do and early 1980s. mestic government debt was perceived as "junior" A second approach has taken the view that the relative to external government debt. They argued residents and nonresidents of indebted developing that fiscal rigidities prevent governments in de countries have at times had differing views on the veloping countries from adjusting quickly to shocks perceived risks of holding the domestic and exter that reduce their debt-servicing capacity. As a re nal financial instruments of the indebted develop sult, a major economic shock could increase the ing countries. In particular, such differences in per perceived risk that the government would be un ceived risks have been regarded as explaining the able to service its obligations fully. Moreover, in simultaneous decision of domestic agents to fi this situation, the risk that the authorities would nance investment with foreign borrowing and to not fully service their domestic debt would gener hold their wealth in the form of foreign assets. As ally be perceived to be higher than the correspond pointed out by Lessard and Williamson (1987), ing risk of servicing foreign obligations because the while the portfolio-based approach concentrates cost of a default on foreign obligations (which on risk and return differences between domestic could lead to a reduction in trade credits) would be and foreign assets that can be held by domestic higher. These differences in perceived risks stimu residents, the "risk differential" approach empha late capital flight by reducing the risk-adjusted re sizes the differences in the perceived risks to resi turn on domestic debt, 14 and can explain the joint dents and nonresidents of holding capital in a de occurrence of increasing external debt and capital veloping country. While both approaches flight in developing countries during the late 1970s emphasize domestic policies as a major factor influ and early 1980s. encing capital flight, the risk differential approach As the previous section has shown, capital flight also emphasizes the role of rigidities in the legal continued to be a feature of developing countries and institutional frameworks of developing coun during the rest of the 1980s, but since 1982 it has tries as a channel through which adverse shocks to been accompanied by a drastic reduction in the in the economy will result in capital flight. flow of new private external credit available to Several empirical studies have found consider these countries. This study advances an hypothesis able support for this risk differential hypothesis. that aims to explain the recent joint behavior of For example, Khan and Haque (1985) argued that capital flight and foreign lending. the perceived risks of investment in developing It has been argued that the difficulties of dealing countries were larger than those of investment in with structural fiscal deficits and ri sing inflation, as industrial countries because of the "expropriation" risk. This expropriation risk reflected institutional and legal arrangements for protecting private prop 14!n contrast. Diwan (1989) argued that there are circum stances when domestic borrowers would prefer to default on erty that were weaker in developing countries than foreign debt. For example, a sharp decline in the price of ex

in industrial countries. Facing this expropriation ports would reduce the cost of defaulting on foreign debt be risk, domestic residents prefer to hold their assets cause the penalty associated with such a default (exclusion from abroad (where they earn a more secure rate of re foreign trade) would fall as export prices declined. As a result, turn) and to borrow external funds to finance do domestic residents would prefer to finance domestic investment in the export goods sector with foreign (as opposed to domestic) mestic investment. With this strategy, domestic loans. Moreover, given the total volume of savings, the avail agents make their portfolios Jess accessible to taxa ability of additional foreign resources would just crowd out the tion or expropriation. Eaton (1987) extended the domestic component of total savings, leading to capital flight. •

well as the limited access of many indebted de of default risk. However, since the emergence of veloping countries to international capital markets, the debt crisis little or no spontaneous lending to have reduced the costs of not fully servicing foreign these countries has taken place and therefore no obligations relative to the corresponding costs of representative interest rates exist. An alternative not servicing domestic obligations.1s The increase measure of the default risk can nonetheless be ob in the perceived probability of default on foreign tained from the secondary market price of external debt has been viewed as being reflected in the de debt by subtracting the LIBOR rate from the im cline of the secondary market price for external plicit yield evident in the international secondary bank debt issued by many heavily indebted de market price for external debt.lS veloping countries. As the perceived difference of Owing to the lack of data, it is not possible to use default risk between domestic and external debt regression analysis to test directly the hypothesis declined, domestic debt would no longer be consid that capital flight has responded positively to in ered junior relative to external debt.16 creases in the default risk on external obligations Although differences in perceived risk between since the emergence of the debt crisis. However, domestic and externaldebt declined after 1982, the for those countries for which sufficient data were total default risk of holding debt (either domestic available to estimate default risk, the correlation or external) issued by these countries has increased between capital flight and default risk on external as a result of the adverse developments and pol debt is positive and very high (Table 4).19 icies affecting the capacity of these countries to ser vice their debt. Therefore, the continuation of capi tal flight coupled with the decline in foreign lending Policies to Prevent or Reverse after 1982 can be explained by a "generalized" per Capital Flight ception of an increase in the default risk of holding debt issued by these countries. If this increase in risk perception has occurred, the lack of external Macroeconomic and Structural Policies creditworthiness (that prevents countries from bor As already noted, many analyses of capital flight rowing in the international capital markets) and have emphasized the role of expansionary mone the continuation of capital flight are both reflec tary and fiscal policies and institutional and legal tions of the same fundamental phenomenaP As a result, policies oriented toward improving the at tSThe implicit yield to maturity for external debt (i•) was tractiveness of holding domestic financial instru obtained from the observed secondary market price on the country's external debt (P) and the application of the following ments of indebted developing countries would also present value formula: help restore creditworthiness. If the perceived differences in the default risk of C FV holding domestic and external debt have been P = r --- + --- , (1+i•)k ( 1 +is)n practically eliminated, then the default risk of hold ing external debt would be a good proxy for the corresponding risk on domestic debt. This hypoth where the face value ( FV) is set at 100 since the discount quoted in the secondary markets applies to $100 worth of contractual esis suggests, therefore, that the level of capital debt; the contractual coupon payment (C) is the interest rate on flight since the emergence of the debt crisis should six-month U.S. Treasury bills (as a measure of the risk-free be positively related to increases in the probability in terest rate) plus the average interest rate spread agreed to by of default on external debt. If heavily indebted de the country on signature of the contract; and n is the average veloping countries still had normal access to inter maturity of the contract. The risk of default on external obligations during 1985-88 was national capital markets, the spread between their estimated by subtracting the six-month LIBOR rate from the borrowing costs and the London interbank offer calculated implicit yield on external debt. rate (LIBOR) could potentially provide a measure The risk of default on external obligations during 1982-84was approximated by using data on spreads between the loan rates charged to indebted developing countries on external bank tSfor example, see Diwan (1989). loans and LIBOR provided by the Deutsche Bundesbank 16Some have even argued that domestic debt should now be (spreads between public sector deutsche mark bonds issued by considered "senior" relative to external debt. This argument is nonresidents and LIBOR) and the Bank of England. intended to rationalize the simultaneous decline in private ex Notice, however, that the implicit yield evident in the interna ternal lending and the continuous increase in domestic debt. tional secondary market price for external debt cannot fully Although the differences in the perceived default risks of do represent the cost of borrowing, since it is derived under condi mestic and external debt have been reduced, the risks of a loss tions of credit rationing. in the real value of domestic assets as a result of a discrete 19However, as mentioned above, the lack of international devaluation or an increase in inflation still remain a concern for creditworthiness and the continuation of capital Hight appear to holders of domestic debt. be two aspects of the same problem. Therefore, the risk of de 17Gajdeczka and Oks (1990) attribute capital Hight after 1986 fault on external debt is also an endogenous variable that should to the loss of creditworthiness of these countries. be simultaneously explained with the behavior of capital Hight.

•'

because of rigidities that inhibit the government's Table 4. Selected Highly Indebted �bility to ��j�s.t qu�ckly to adverse shocks. In par Developing Countries Facing Debt ticular, ngtdtttes mduced by wage indexation Servke Problems:. Correlation Between schemes, minimum wages, laws preventing layoffs, Capital Flight and Default Rlskl c?ntrols on the prices of public goods, and ineffi Cient tax systems often prevent a government from adjusting its fiscal deficit quickly to offset the im pact of an adverse shock. Country Correlation The role of appropriate structural and macro

Argentina 0.640 econ?micyoli7ies in reducing (and reversing) capi tal fltght ts evtdent from the recent experience of Bra zil2 0.826 Chile 0.044 several Latin American countries. The stock of Colombia 0.892 �ight capit.al �n Argentina declined in 1986 (imply Ecuador 0.718 mg repatnatwn) following the adoption in June jamaica 0.824 1985 of a comprehensive adjustment program that Mexico 0.821 was initially successful in reducing inflation and in fostering an economic recovery. However, sus Nigeria 0.992 tained fiscal deficits led to renewed inflation and Peru 0.954 capital flight apparently accelerated again. In' Philippi nes 0.604 Bra zil, Venezuela 0.067 the introduction of the Cruzado plan in early 1986 was also followed by a decline in the net Yugoslavia 0.960 out flow of capital. However, inflation accelerated in 198? and estimated capital flight increased sharply Sources: Salomon Brothers, New York; Data Resources, Inc.; and Deutsche Bundesbank. dunng that year. In December 1987, the Mexican 1 Countries shown are those lor which the default risk authorities introduced a comprehensive economic could be calculated during 1982-88. program based on a social pact with labor and busi 2Data correspond to 1 982-87. ness that encompassed front-loaded fiscal and monetary corrections as well as a strong devalua tion of the Mexican peso. As a result of the success ful implementation of this program, empirical esti mates suggest that Mexico experienced a net arrangements as key factors increasing the risks of repatriation of capital flight in 1988, the first in the holding domestic assets. Sound macroeconomic 1980s. policies and appropriatestructural reforms are thus 1\:f ny countries have also used a variety of other likely to be important elements in solving the prob � �ohctes to stem or reverse capital flight. The effec lem of capital flight. In many countries, a sustained tiveness of these policies is now examined in terms reduction of the fiscal deficit is typically an impor of recent country experiences of implementing tant first step in reducing the perceived risks of such policies.zt holding domestic debt. Large fiscal deficitsfinanced with nonindexed domestic debt are likely to gener ate the expectation that the government will even Capital Controls tually use the inflation tax and/or a devaluation of n has often been argued that externalities and the currency to reduce the outstanding real value policy distortions imply that the social rate of re of the debt.20 turn on domestic investment in developing coun The consistency of exchange rate, monetary, and tries is higher than the private rate of return. As a fiscal policies is another crucial element in stem result, capital controls have been viewed as a ming capital flight. While many countries have used means of ensuring that domestic savings are in a fixed exchange rate to provide an anchor for the vested domestically. As noted in Gordon and domestic price level, problems have at times arisen Levine (1989), however, a fundamental problem with this argument is that it assumes that capital :-vhen adverse s�ocks suddenly lead to a sharp fall . m the stock of mternational reserves and thereby not mvested abroad would be invested domes tically; it therefore ignores the possible substitution c�eate the expectation of a devaluation and of pos Sible abandonment of the adjustment program. between savings and consumption. In particular, Such expectations, which can lead to increasing the announcement of capital controls might bring capital flight and additional losses of reserves, arise . 21 The e.xperiences reported here are largely based on pre VIOUS stud1es undertaken by the staff of the International Mone 20for further discussion of these issues see Calvo (1988). tary Fund. II

expectations of further government intervention additional claims on debtor countries at a discount, that would discourage domestic investment in the net inflow of funds to the countries would not favor of increased consumption. In addition, sav necessarily increase significantly, since foreign ings may be held· in various inflation "hedges" commercial banks would still have no incentive to (such as real estate or inventories) that have little engage in additional voluntary loans. impact on productive capacity. Second, since some debt-equity swap schemes Although capital controls are ineffective in pro have involved reselling discounted paper for do moting domestic investment, the experiences of mestic currency at preannounced exchange rates, some countries suggest that such controls have had the money supply can increase as the swap occurs, a short-run effect on capital flight. For example, thereby generating inflationary pressures. If more Brazil experienced a much lower ratio of capital rapid inflation occurred, it would increase the risks flight to total external debt than Mexico although of holding domestic assets and encourage addi Brazil's average inflation rate during the late 1970s tional capital flight, which might offset the initial and early 1980s was higher than Mexico's. This inflow originated by the debt-equity swap. The could potentially reflect the more restrictive capital Mexican debt-equity swap program initiated in controls in Brazil. However, when capital controls May 1986 was suspended in late 1987 in part be are effective in stemming capital flight in the short cause the authorities wanted to limit the monetary run, the removal of such controls can be accom expansion produced by these operations. In con panied by large-scale capital flight, especially when trast, the Chilean program, introduced in 1985, al expansionary macroeconomic policies lead to ris lowed domestic residents to convert external ing inflation and expectations of devaluation. As claims on Chile into tradable domestic bonds, and documented by Cuddington (1986), Argentina is a this effectively sterilized the monetary impact of good example of sharp increases in capital flight these operations.22 These experiences suggest that following the removal of capital controls, which in one of the key factors determining whether debt turn increased the Government's incentive to reim equity swaps can be successful in repatriating capi pose them. tal flight is whether they are conducted in the con Overall, experience suggests that although capi text of an appropriate monetary policy. tal controls have sometimes helped in the short run to stem capital flight, the problems associated with them may be larger than the benefits. As capital Foreign Currency Deposits controls by themselves constitute a distortion, they Local deposits denominated in foreign curren may lower domestic savings because domestic resi cies have been used as a mechanism to prevent dents have been constrained to hold less diversified capital flight in several countries (India, Mexico, portfolios. Moreover, when capital controls are Uruguay, and Turkey).23 While offering depositors perceived as a policy instrument used on a discre some potential protection from exchange risk, they tionary basis, expectations that such controls will will not be viewed as carrying much lower risk than be imposed encourage capital flight. In the long other domestic deposits if the authorities have in run, therefore, they may decrease the real re the past frozen or limited withdrawals or returns on sources available for investment. these deposits. The freezing of the dollar denominated accounts in Mexico in 1982 provides an example of such risk.24 However, these ac Debt-Equity Swaps counts, as well as some other financial instruments Debt-equity swaps can be attractive to investors indexed to foreign exchange, may usefully contrib because assets can be purchased in the debtor ute to stemming capital flight when the government countries with external claims on those countries is pursuing appropriate macroeconomic and struc- purchased at a discount in the secondary market. While this mechanism has sometimes been effec 22Even if those operations can be made noninflationary, sub tive in reversing capital flight, its ability to increase sidized swap arrangements, such as those in Chile in 1983-84, net capital inflows may be limited for two reasons. can generate large operating deficits in the central bank. First, the profitabilityof debt-equity swaps depends 23 As reported in previous studies undertaken by the staff of the International Monetary Fund, in Turkey the Dresdner Bank on the discount in the secondary market; and, if collected deposits from Turkish workers in Germany and trans this market is very thin, increases in the demand ferred these deposits to the Central Bank of Turkey where they for claims on debtor countries would raise the were kept as foreign currency deposits. As a result, workers' prices of those claims and reduce the profitability remittances rose sharply and the stock of these foreign currency deposits reached $5.9 billion by the end of 1988. of debt-equity swaps. In this situation, only a lim 24During this freeze, withdrawals from dollar-denominated ited amount of swaps could be undertaken. More accounts were paid in Mexican pesos using the controlled ex· over, even if commercial banks were willing to sell change rate . •

tural policies but faces a credibility problem.25 If ment and that could in turn have significantly in successful in eliminating exchange rate risk, indexed creased countries' debt-servicing capacity. This assets may stimulate domestic saving and contribute study has argued that two forms of risk have been to lower domestic real interest rates and, thereby, major causes of capital Hight: the risk of direct de lower the government's costs of financingits domes fault or repudiation associated with the fear of ex tic debt. However, while foreign currency-denom propriation of domestic assets and the risk of large inated deposits can potentially help stem capital losses in the real value of domestic assets arising flight with good policies, they can create additional from inflation or large exchange rate devaluations. difficulties when policies are inadequate. Rising in Moreover, the simultaneous occurrence of large in flation can lead to a rapid shift from deposits de flowsof foreign capital and large capital flight from nominated in the domestic currency to those de developing countries during the 1970s and early nominated in foreign currencies, thereby 1980s reflected the different perceptions of foreign destabilizing domestic monetary relationships. lenders and domestic residents of developing coun In Peru, for example, accelerating inflation and tries about the risks of holding domestic claims. In balance of payments deficits during the early 1980s particular, for a variety of institutional reasons, for led domestic residents to increase their holdings of eign lenders perceived a lower default risk than did bank deposits denominated in dollars, and by the domestic residents. However, since the emergence end of 1984 these deposits, which amounted to $1.7 of the debt crisis in 1982 the differences in per billion, were equivalent to more than 50 percent of ceived risks have been reduced and have resulted broad money. Convertibility of these deposits into in a decline of foreign capital inflowscoupled with foreign currency was suspended in 1985. continuation of capital flight. Indeed, the adverse developments that accom Other Policies panied the emergence of the debt crisis may have Several other policies have been suggested to en increased the perceived default risk of holding courage the repatriation of Hight capital. For exam either the domestic or external debt of highly ple, tax treaties that would allow countries to tax indebted developing countries. If this is true, the assets held abroad by domestic residents were con continuation of capital Hight in the mid- and late sidered one alternative.26 However, as noted by 1980s, coupled with the decline in private foreign Gordon and Levine (1989), it may be difficult to lending, can be explained by a generalized percep enforce such treaties unless they are implemented tion of an increase in the default risk of holding debt globally.27 Tax amnesty programs have also been issued by highly indebted developing countries. used,28 but expectations of future tax amnesties If concerns about default risks have played a ma might reduce the normal collection of taxes and jor role in the portfolio decisions of the domestic might, paradoxically, motivate increases in capital residents of heavily indebted developing countries, flight at times when amnesties are not in place. the policies with the greatest chance of stemming Another policy has been to tighten domestic capital flight are those that decrease the risks of credit to generate liquidity shortages sufficient to holding domestic assets. In this situation, sound induce some liquidation of foreign asset holdings macroeconomic policies complemented by appro for financing domestic investments or firms' work priate structural reforms would have to be key ele ing capital needs. However, repeated liquidity ments in stemming or reversing capital Hight. shortages might decrease the expected returns Once these core macroeconomic and structural from domestic investments and might instead in policies arc in place, other policies such as debt duce greater capital Hight. equity swaps or foreign currency deposits although insufficient by themselves to solve the Conclusions problem of capital flight-can potentially contrib ute to reducing it. Capital controls are likely to be Capital Hight implies a loss of resources that at best a short-run deterrent to capital Hight and, could have been used to increase domestic invest- by introducing additional distortions, could even accentuate the problem in the long run. 2Sfor a further elaboration on these issues, see Calvo and Guidotti (1990}. 26from 1967 to 1981, Brazil was negotiating a treaty with the United States, but no final agreement took place. References 27Moreover, if those treaties involve a symmetric treatment of taxes between countries, the total tax base of the domestic The Problem country may be reduced if holdings of domestic assets by for Calvo, Guillermo, "Controlling Inflation: eigners are excluded from the tax base. of Non-Indexed Debt," IMF Working Paper No. 28fn 1986 Colombia introduced a tax amnesty on previously 88/29 (Washington: International Monetary Fund, undeclared income and wealth. March 1988). II

--- · and Pablo E. Guidotti, "Indexation and Matu Eaton, Jonathan, "Public Debt Guarantees and Private rity of Government Bonds: An Exploratory Model," Capital Flight," NBER Working Paper No. 2172 (Na in Public Debt Management: Theory and History, ed. tional Bureau of Economic Research, March 1987). by Rudiger Dornbusch and Mario Draghi Gajdeczka, P., and Daniel Oks, "Domestic Deficits, Debt (Cambridge; New York: Cambridge University Overhang, and Capital Outflows in Developing Press, 1990). Countries," AMEX Bank Review, annual series en Cuddington, John T., Capital Flight: Estimates, Issues, titled Finance and the International Economy: 3 and Explanations, Princeton Studies in International (1990). Finance, No. 58 (Princeton, New Jersey: Interna Gordon, David B., and Ross Levine, "The 'Problem' of tional Finance Section, Princeton University, 1986). Capital Flight: A Cautionary Note," World Econ omy, Vol. 12 (June 1989). ___, "Macroeconomic Determinants of Capital Flight: An Econometric Investigation," in Capital International Monetary Fund, Balance of Payments Sta Flight and Third World Debt, ed. by Donald R. tistics Yearbook (Washington: International Mone Lessard and John Williamson (Washington: In tary Fund), various issues. stitute for International Economics, 1987). ___, International Financial Statistics Yea rbook Cumby, Robert, and Richard Levich, "On the Definition (Washington: International Monetary Fund), and Magnitude of Recent Capital Flight," in Capital various issues. Flight and Third World Debt, ed. by Donald R. ___, World Economic Outlook (Washington: Inter Lessard and John Williamson (Washington: . In national Monetary Fund, May 1990). stitute for International Economics, 1987). Ize, Alain, and Guillermo Ortiz, "Fiscal Rigidities, Public Debt and Capital Flight," Staff Papers, International Deppler, Michael, and Martin Williamson, "Capital Monetary Fund, Vol. 34 (June 1987). Flight: Concepts, Measurement, and Issues," in Staff Khan, Mohsin S., and Nadeem Ul Haque, "Foreign Bor Studies for the World Economic Outlook, By the Re rowing and Capital Flight: A Formal Analysis," S search Department of the International Monetary taff Papers, International Monetary Fund, Vol. 32 (De Fund (Washington: International Monetary Fund, cember 1985). August 1987). Lessard, Donald R., and Williamson, John, "The Prob Diwan, Ishac, "Foreign Debt, Crowding Out and Capital lem and Policy Responses," in Capital Flight and Flight," Journal ofInternational Money and Finance, Third World Debt, ed. by Donald R. Lessard and Vol. 8 (March 1989). John Williamson (Washington: Institute for Interna Dooley, Michael, "Country-Specific Risk Premiums, tional Economics, 1987). Capital Flight and Net Investment Income Payments Meyer A., and M.S. Bastos-Marques, "A Fuga de Capital in Selected Developing Countries" (Washington: In no Brasil," Funda�ao Getulio Vargas, Instituto ternational Monetary Fund; unpublished, March Brasileiro de Economia, Centro de Estudios Mon 1986). etarios e de Economia Internacional, No. 06/89 (Au Dornbusch, Rudiger, "External Debt, Budget Deficits, gust 1989). and Disequilibrium Exchange Rates," in Interna Morgan Guaranty Trust Company, "LDC Capital tional Debt and the Developing Countries, ed. by Flight," World Financial Markets (March 1986). Gordon W. Smith and John T. Cuddington (Wash Rodriguez, Miguel A., "Consequences of Capital Flight ington: World Bank, 1985). for Latin American Debtor Countries," in Capital Duwendag, Dieter, "Capital Flight from Developing Flight and Third World Debt, ed. by Donald R. Countries: Estimates and Determinants for 25 Ma Lessard and John Williamson (Washington: In jor Borrowers," Societe Universitaire Europeenne stitute for International Economics, 1987). de Recherches Financieres, SUERF Series No. 52A, World Bank, World Development Report (New York: 1987. Oxford University Press, 1985).