Chapter

IV Developments Under Floating Exchange Rate Regimes

Author(s):
Kyong Huh, Benedicte Christensen, Peter Quirk, and Toshihiko Sasaki
Published Date:
May 1987
Share
  • ShareShare
Show Summary Details

Most of the developing countries that have adopted floating exchange rates have at the same time pursued supporting monetary and fiscal policies and have liberalized exchange and trade restrictions. Isolating the separate effects of floating is therefore a difficult task. Moreover, because the experience with flexible exchange rates in most of the countries surveyed is recent, inferences that may be drawn must necessarily be regarded as tentative, both because the full effects of flexibility take time to work through an economy and because most of the relevant macroeconomic data are preliminary. These qualifications apply with less force, however, to an examination of the effects of floating on the exchange market itself, that is, on the variables most directly affected by the exchange rate regime. Such an analysis is also helped by the fact that data for exchange rates and certain important exchange market transactions become available with relatively short lags.

This section examines the impact of floating and the accompanying adjustment measures, first on the volatility of the exchange rate itself and on capital account transactions, and second in a broader macroeconomic context. Seven of the countries with floating systems (Bolivia, the Dominican Republic, Jamaica, the Philippines, Uganda, Uruguay, and Zaïre) have sufficient experience with floating for at least some of these questions to be addressed. For each of these countries, the volatility of the exchange rate, vis-à-vis the U.S. dollar and in nominal and real effective terms, is compared between the pre-floating and post-floating periods. Capital account transactions are examined using the available data from the balance of payments, cross-border bank deposits, external debt, and the parallel exchange market. Pre-float and post-float developments in inflation and in output and trade are also analyzed in this section.

Exchange Market Developments

Bilateral and Effective Exchange Rates

There have been concerns that floating exchange rates in developing countries would tend to be unstable because of the relative thinness of financial markets. First, there has been a fear that the institution of floating arrangements would be followed by a free fall of the exchange rate, regardless of the stance of domestic policies, as increased uncertainty fed into the rate of domestic inflation and the two cumulated over time. The second related concern has been that the exchange rate would prove volatile, fluctuating widely in both directions in response to external shocks in commodities markets and capital flows, to internal reversible factors such as drought, and to generalized instability in domestic economic policies.

The experience has been that in all seven countries for which adequate data are generally available in the period in which a floating regime was maintained, the domestic currency initially depreciated in terms of the U.S. dollar (which was the intervention currency in most instances), and also in nominal and real effective terms (with the exception of the real rate in the Dominican Republic).19 The extent of the initial depreciation in each country reflected the magnitude of exchange rate disequilibrium preceding the float and also the extent to which the market was subsequently allowed to operate freely, and within a framework provided by accompanying monetary and fiscal policies and liberalization of exchange and trade controls. Although some of these currencies subsequently appreciated, following the initial correction, all except the currency of the Dominican Republic on balance depreciated over the entire period from just before the introduction of floating to June 1986 against the U.S. dollar and in nominal effective terms. However, in each of the five countries for which data are available to compare the pre-float unofficial market U.S. dollar exchange rate with the official rate in the early months of floating (Bolivia, the Dominican Republic, the Philippines, Uganda, and Zaïre), the domestic currency initially appreciated in the official market in relation to the exchange rate in the parallel market immediately prior to the float (Chart 1).20

Chart 1.Floating Exchange Rate Regimes: Exchange Rate Developments in Selected Developing Countries, January 1976–June 1986

Sources: Data provided by national authorities; International Currency Analysis, Inc., 1984 World Currency Yearbook (Brooklyn, New York, 1985); and Fund staff estimates.

NOTE: Increases represent appreciation of the domestic currency and vertical slashes indicate the beginning of floating of exchange rate. For Uganda, owing to the large disparity between the official and parallel exchange rate figures since end-1985, the upper chart has been drawn according to a logarithmic scale.

Developments in real effective exchange rates since the initial exchange rate correction following floating have varied widely. From the month following the float to June 1986, real effective rates continued their initial depreciation in five of the seven countries at moderate rates (less than 1 percent per month). In the other two countries, the real effective rate appreciated slowly over the corresponding periods, with average monthly rates of appreciation of 0.1 percent and 0.2 percent in Uruguay and the Dominican Republic, respectively. In Uruguay these real appreciations were not sufficient to offset the depreciation of the exchange rate at the time of transition to floating, but in the Dominican Republic, the real effective rate had appreciated slightly prior to the institution of floating under the influence of tight monetary policies pursued during most of the period (January 1985-June 1986).

To gauge the pre-float and post-float developments in exchange rate variability, four calculations were made, each being applied to bilateral rates vis-à-vis the U.S. dollar and to nominal and real effective rates. The variability of the exchange rate itself was measured by (1) mean monthly absolute percentage change; (2) average absolute percentage deviation from a fixed log-linear time trend; (3) standard deviation of monthly absolute percentage changes; and (4) standard deviation of the absolute deviations from a fitted log-linear trend.

First, with regard to bilateral rates vis-à-vis the U.S. dollar, the comparison was made for the seven currencies between the variability of the official rate prior to the float and the subsequent variability of the unified floating rate. For four of the countries examined (Jamaica, the Philippines, Uganda, and Zaïre), the trend-corrected measures (2) and (4) indicated that the variability of the floating rate in terms of the U.S. dollar was less than that of the official rate prior to the float.21 The Dominican peso showed an expected unambiguous increase in variability (given the previous long-standing parity with the U.S. dollar), the results for the Uruguayan peso were mixed, probably reflecting expectations regarding the political transition at the end of 1984, and for the Bolivian peso most measures indicated no significant outcome. Without trend-correction, most tests for the seven countries showed increased variability in terms of the U.S. dollar (see Tables 8, 9, and 10 in the Appendix).

Table 8.Floating Exchange Rate Regimes: Exchange Rate Variability in Selected Developing Countries in the Pre-Float and Floating Periods, January 1976–June 19861
Bilateral U.S. Dollar Rate, Floating Compared with Pre-Float PeriodNominal Effective Rate, Floating Compared with Pre-Float PeriodReal Effective Rate, Floating Compared with Pre-Float Period
Trend adjustedWithout trendTrend adjustedWithout trendTrend adjustedWithout trend
BoliviaI2I4I2UI2U
Dominican RepublicDDII4IU
JamaicaID3ID3I2D3
PhilippinesIUIUI2U
UgandaII4ID3IU
UruguayD5D6DD5D
ZaïreII4II4II
Sources: Tables 9 and 10.

Comparison of volatility between pre-floating and post-floating periods is on the basis of appropriate statistical tests at 95 percent confidence interval. The symbols in the table denote the following: I = improvement (reduced volatility in floating period), U = volatility unchanged, and D = deterioration (increased volatility in floating period).

No significant difference by measure (2), volatility reduced by measure (4); see page 22.

No significant difference by measure (3), volatility increased by measure (1).

No significant difference by measure (1), volatility reduced by measure (3).

No significant difference by measure (2), volatility increased by measure (4).

Volatility increased by measure (4), volatility reduced by measure (2).

Sources: Tables 9 and 10.

Comparison of volatility between pre-floating and post-floating periods is on the basis of appropriate statistical tests at 95 percent confidence interval. The symbols in the table denote the following: I = improvement (reduced volatility in floating period), U = volatility unchanged, and D = deterioration (increased volatility in floating period).

No significant difference by measure (2), volatility reduced by measure (4); see page 22.

No significant difference by measure (3), volatility increased by measure (1).

No significant difference by measure (1), volatility reduced by measure (3).

No significant difference by measure (2), volatility increased by measure (4).

Volatility increased by measure (4), volatility reduced by measure (2).

Table 9.Floating Exchange Rate Regimes: Variability of Exchange Rates in Selected Developing Countries, January 1976–June 19861
Mean Monthly Absolute Percentage Change2Variance of Monthly Absolute Percentage Change3Sample Size
Official U.S. dollar (monthly average)Nominal effectiveReal effective4Official U.S. dollar (monthly average)Nominal effectiveReal effective4Official U.S. dollar (monthly average)Nominal effectiveReal effective4
Bolivia
Jan. 1976–Aug. 198511.806.0411.601,923.66200.77295.0711611692
Sept. 1985-June 198610.758.566.52156.2486.5436.56101010
Dominican Republic
Jan. 1976–Dec. 19841.181.719.8510.0010810884
Jan. 1985-June 19861.392.152.871.524.294.61181818
Jamaica
Jan. 1976–Nov. 19830.971.781.9118.298.866.29959571
Dec. 1983–June 19862.783.002.8210.677.847.39313131
Philippines
Jan. 1976–Sept. 19840.931.442.1411.737.487.9910510581
Oct. 1984-June 19861.402.572.654.035.955.96212121
Uganda
Jan. 1976–May 198410.482.665.467,067.3682.32102.5710110177
June 1984-June 19867.836.927.11125.8268.9033.84252525
Uruguay
Jan. 1976–Oct. 19821.981.911.850.963.483.71828258
Nov. 1982-June 19866.694.303.98221.6954.2650.93444444
Zaïre
Jan. 1976–Jan. 19846.743.957.19691.41100.13113.24979773
Feb. 1984-June 19862.322.252.485.094.364.74292929
Source: International Monetary Fund, International Financial Statistics (Washington, various issues), and staff estimates.

Exchange rates are expressed as the foreign currency price of a domestic currency unit.

Volatility measure (1) (see text). In the post-floating periods, the first month in each case is the month of floating exchange rates.

Volatility measure (3), page 22.

Real effective exchange rate data for all countries are calculated from January 1978.

Source: International Monetary Fund, International Financial Statistics (Washington, various issues), and staff estimates.

Exchange rates are expressed as the foreign currency price of a domestic currency unit.

Volatility measure (1) (see text). In the post-floating periods, the first month in each case is the month of floating exchange rates.

Volatility measure (3), page 22.

Real effective exchange rate data for all countries are calculated from January 1978.

Table 10.Floating Exchange Rate Regimes: Exponential-Trend-Corrected Variability of Exchange Rates in Selected Developing Countries, Januar y 1976–June 19861
Mean Monthly Absolute Percentage Change2Monthly Variance of Percentage Change3Sample Size
Official U.S. dollar (monthly average)Nominal effectiveReal effective4Official U.S. dollar (monthly average)Nominal effectiveReal effective4Official U.S. dollar (monthly average)Nominal effectiveReal effective4
Bolivia
Jan. 1976–Aug. 1985343.26117.1521.30928,448.6914,661.521,137.6611511591
Sept. 1985-June 198611.3610.826.3777.8441.1830.91999
Dominican Republic
Jan. 1976–Dec. 198410.319.0779.5031.6910710783
Jan. 1985-June 19861.273.234.580.923.629.21171717
Jamaica
Jan. 1976–Nov. 198314.8415.754.2669.7344.0914.17949570
Dec. 1983–June 19867.605.765.0319.809.383.82303030
Philippines
Jan. 1976–Sept. 198412.728.466.82121.6353.8420.2910410480
Oct. 1984-June 19863.515.736.581.845.198.88202020
Uganda
Jan. 1976–May 198463.9770.3563.791,155.135,417.546,439.7510010076
June 1984-June 198612.8615.6712.61102.42125.7665.87242424
Uruguay
Jan. 1976–Oct. 19825.9415.454.9817.14101.4610.58818157
Nov. 1982-June 19865.619.205.9445.05108.4167.52434343
Zaïre
Jan. 1976–Jan. 198426.5224.3219.321,266.98406.05385.52969672
Feb. 1984-June 19863.564.013.954.978.697.80282828
Sources: International Monetary Fund, International Financial Statistics (Washington, various issues), and staff estimates; and International Currency Analysis, Inc., 1984 World Currency Yearbook (Brooklyn, New York, 1985).

A simple time trend was fitted for all exchange rates and for all countries. The exchange rates are expressed as the foreign currency price of a domestic currency unit.

Volatility measure (2) (page 22). In the post-floating periods, the first month in each case is the month of floating exchange rates.

Volatility measure (4) (page 22).

Real effective exchange rate data for all countries are calculated from January 1978.

Sources: International Monetary Fund, International Financial Statistics (Washington, various issues), and staff estimates; and International Currency Analysis, Inc., 1984 World Currency Yearbook (Brooklyn, New York, 1985).

A simple time trend was fitted for all exchange rates and for all countries. The exchange rates are expressed as the foreign currency price of a domestic currency unit.

Volatility measure (2) (page 22). In the post-floating periods, the first month in each case is the month of floating exchange rates.

Volatility measure (4) (page 22).

Real effective exchange rate data for all countries are calculated from January 1978.

Statistical tests conducted on the data for nominal and real effective exchange rates show a statistically significant reduction in the variability of five out of the seven nominal effective rates by measures (2) and (4), that is, the measures of variability obtained with correction for trend. The outcome of the measures for Bolivia and Uruguay was mixed. Without trend-adjustments—that is, on the basis of measures (1) and (3)—the evidence of a reduction in variability is less clear, as statistical tests indicate that variability declined in the Dominican Republic and Zaïre, remained virtually unchanged in Bolivia and the Philippines, and increased in the remaining three countries.

The outcome of the tests for the real effective exchange rate is similar to that for the nominal effective rates. In terms of the trend-adjusted exchange rate data, which may be argued to be the more relevant of the measures employed, the variability of the real effective rate increased in the floating period only in Uruguay. The data without trend-adjustment, using statistical tests for significance, indicate that the variability of the real effective exchange rate increased in Jamaica and Uruguay, declined in Zaïre, and showed no statistically significant change in the remaining four countries.

To sum up these results, statistical measures suggest that two-way variability (i.e., with trend correction) of exchange rates against the U.S. dollar has declined in the floating period, but that continuing adjustment of the exchange rate has been reflected in increased unidirectional changes (i.e., without trend correction). For the nominal and real effective exchange rates, based on experience to date, there is evidence that the exchange rate movements have been generally smoother than under the previous managed and fixed-rate regimes. This may be attributed to a number of factors, but especially to the support provided by monetary and fiscal policies: First, exchange markets in the floating periods have not been entirely free of central bank intervention to stabilize the rates, as described earlier. Second, reduced exchange rate volatility in the post-float periods may have reflected reduced instability at that time in overall economic conditions—for example, volatility in primary commodity prices. Third, it could be argued that exchange rates in a pegged-rate regime are especially susceptible to destabilizing speculative activity, and that they eventually require relatively large step adjustments.

Capital and Reserve Movements

The introduction of exchange rate flexibility and the accompanying liberalization of the exchange system may be expected to affect both official foreign exchange reserves and capital flows through the following channels. Floating the exchange rate removes the need for official intervention to defend the value of the domestic currency, so that reserve losses, official overborrowing, and external payments arrears may be avoided. Private capital flows are also likely to respond positively and directly, first, to the removal of the risk of a step devaluation of the domestic currency from an unrealistic level, and second, to liberalization of exchange controls on the repatriation of capital or investment income—inward direct investment increases, the rescheduling of debt is facilitated,22 and outward capital flight is stemmed. Third, there is the role of monetary policies. Capital reflows will, of course, require domestic interest rates to be competitive with foreign rates, but even so they may not need to be as high as before the float, since they then had to compensate for expectations of devaluation.

The measurement of many of these effects is problematical. First, they may not be apparent in the data because of the simultaneous operation of other factors. For example, increased repayments of previously contracted external debt may become due at the time floating is introduced. Second, most data for short-term capital flows are unreliable under the best of circumstances, and capital flight is by its nature not directly measurable. Such activities as smuggling (of financial assets, real assets, and commodities), under-invoicing of exports, and overinvoicing of imports are by their nature not observed by the agencies collecting balance of payments data. However, such transactions will be reflected as net errors and omissions in the balance of payments to the extent that they have recorded transactions as a counterpart. It is often assumed that net errors and omissions provide a reasonable indicator of capital flight, although being the balancing item, they also incorporate transactions unrelated to capital flight.

In all six countries, except the Dominican Republic in which capital flight had not been a problem, available data suggest that combined net short-term capital inflows and errors and omissions have increased since floating began (see Table 11 in the Appendix).23 The turnaround of capital flight appears to have been the largest in Jamaica—corresponding to almost 20 percent of annual imports. In the other countries, the short-term capital reversal was smaller (less than 5 percent of imports) and it was negligible in Uganda. Available indicators of export underinvoicing and import over-invoicing from partner country trade data also suggest diminished capital flight in the only countries for which data are available—Jamaica and the Philippines.24 Other indicators of capital flight or reflow, in the form of bank deposits held abroad by nonbank domestic residents or foreign-currency-denominated deposits held onshore, are inconclusive (see Table 12 in the Appendix). Data on foreign-currency-denominated deposits with domestic banks, which may indicate currency substitution, although not a balance of payments item, are available only for the Philippines and Uruguay. Foreign-currency-denominated deposits rose in the Philippines and fell in Uruguay during the floating period.

Table 11.Floating Exchange Rate Regimes: Net Capital Flows, 1980-86In percent of imports plus net service payments
1980198119821983198419851986Average of 1980 to Year of Float1Average of Post-Float Periods
Dominican Republic
Medium and long-term capital17.3313.0613.531.707.52–9.2710.7910.630.76
Short-term capital211.883.95–6.494.24–2.842.182.151.09
Jamaica
Medium and long-term capital18.3815.9533.022.8444.8114.0518.4222.5425.75
Short-term capital2–10.760.172.72–26.4117.4712.866.29–8.5710.63
Philippines
Medium and long-term capital11.4418.2617.9718.27–2.71–20.3012.65–20.30
Short-term capital25.13–6.23–3.02–12.19–0.61–0.16–3.38–0.16
Uganda
Medium and long-term capital7.914.753–27.173–0.18310.238.90–10.00–0.890.55
Short-term capital2–16.5011.293–4.193–3.933–9.63–2.143.58–4.740.52
Uruguay
Medium and long-term capital41.2934.0729.7951.9810.47–43.2725.4235.0511.15
Short-term capital28.23–6.28–55.38–38.42–13.73–1.11–17.81–14.21
Zaïre
Medium and long - term capital–0.93–7.77–7.69–9.79–13.93–13.51–6.55–13.72
Short-term capital28.436.9110.23–2.934.828.085.666.45
Memorandum items:
Errors and omissions
Dominican Republic46.216.09–6.299.96–0.883.02
Philippines1.14–5.89–4.26–2.04–0.106.08–2.236.085
Uruguay7.70-11.08–96.50–8.94–10.0033.29–9.476
Export underinvoicing and import overinvoicing7
Dominican Republic72.8110.8110.584.069.8
Jamaica–4.932.3172.0–23.30.2….
Philippines46.367.952.0155.6239.4115.1
Uruguay29.9–26.1181.4
Debt rescheduling8
Dominican Republic10.6925.9767.90
Uganda1.39312.0836.7932.073.06
Uruguay41.811.125.244.44
Zaïre67.8517.137.3648.3522.4718.5822.16
Source: International Monetary Fund, Direction of Trade Statistics (DOT) and International Financial Statistics (IFS) (Washington, various issues), and staff estimates.

Except for Zaïre: although the zaire was floated in 1984, for the purpose of calculating averages in this table 1984 is considered to be a post-float year.

Includes errors and omissions.

Data for fiscal year ended in June.

Comprises private short-term capital, commercial banks’ working capital, and errors and omissions.

1985 only.

Average of 1983 and 1984.

Exports (DOT) - exports (IFS) + imports (IFS) - imports (DOT).

Not included in the capital flows data above.

Source: International Monetary Fund, Direction of Trade Statistics (DOT) and International Financial Statistics (IFS) (Washington, various issues), and staff estimates.

Except for Zaïre: although the zaire was floated in 1984, for the purpose of calculating averages in this table 1984 is considered to be a post-float year.

Includes errors and omissions.

Data for fiscal year ended in June.

Comprises private short-term capital, commercial banks’ working capital, and errors and omissions.

1985 only.

Average of 1983 and 1984.

Exports (DOT) - exports (IFS) + imports (IFS) - imports (DOT).

Not included in the capital flows data above.

Table 12.Floating Exchange Rate Regimes: International Reserves, External Debt, Arrears, and Foreign Currency Deposits, 1980-86

In percent of imports plus net service payments; end of period1

19801981198219831984198519862
Bolivia
Gross official international reserves10.07.214.516.225.624.542.3
External debt outstanding3237.6195.1279.5372.8388.6292.3364.0
External payments arrears outstanding1.63.613.720.282.990.8128.8
Currency deposits abroad by nonbank residents45.442.934.944.64
Dominican Republic
Gross official international reserves11.112.69.112.219.226.523.3
External debt outstanding377.798.9136.7178.1204.9232.6258.8
External payments arrears outstanding8.117.530.635.858.257.515.3
Currency deposits abroad by nonbank residents39.746.426.863.556.363.34
Jamaica
Gross official international reserves6.85.17.26.07.911.919.1
External debt outstanding3142.2154.2204.1251.2272.1230.6235.0
External payments arrears outstanding9.33.916.53.17.15.0
Currency deposits abroad by nonbank residents7.810.512.213.616.121.0
Lebanon
Gross official international reserves41.839.974.551.423.259.725.64
External debt outstanding34.96.17.66.56.39.911.4
External payments arrears outstanding
Currency deposits abroad by nonbank residents112.4124.7135.6191.3340.1345.54
Philippines
Gross official international reserves35.225.111.411.612.421.226.4
External debt outstanding3212.3253.1283.4320.7390.3525.0562.25
External payments arrears outstanding19.431.1
Currency deposits abroad by nonbank residents5.75.99.515.521.026.2
Uganda6
Gross official international reserves0.58.514.419.710.211.59.3
External debt outstanding395.7121.54107.7115.2134.4150.9167.1
External payments arrears outstanding35.831.528.822.212.414.621.1
Currency deposits abroad by nonbank residents14.211.816.118.917.117.5
Uruguay
Gross official international reserves22.125.68.116.813.617.630.77
External debt outstanding385.8121.5212.6279.6369.6306.2304.07
External payments arrears outstanding
Currency deposits abroad by nonbank residents51.776.1120.9175.0194.7130.77
Zaïre
Gross official international reserves8.67.01.82.33.65.46.7
External- debt outstanding3184.9204.1220.9225.5233.7247.2269.4
External payments arrears outstanding24.128.445.914.49.55.54.4
Currency deposits abroad by nonbank residents22.621.923.324.628.928.3
Memorandum item:
Foreign currency deposits with domestic banks in
Philippines108.9115.9125.7126.6141.7175.4171.24
Uruguay13.515.996.696.0117.0109.182.34
Sources: Data provided by national authorities; and International Monetary Fund, International Financial Statistics (Washington, various issues), and staff estimates.

Except data for Lebanon which is in percent of imports.

Data for gross international reserves and currency deposits abroad are for end-September; data for external debt outstanding and external payments arrears outstanding are for end-June, unless otherwise indicated.

Coverage of external debt data differ among countries, and therefore the debt ratios may not be comparable between countries.

End-June figures.

End-May figures.

All data for Uganda are latest available estimates for end-1986.

Estimates.

Sources: Data provided by national authorities; and International Monetary Fund, International Financial Statistics (Washington, various issues), and staff estimates.

Except data for Lebanon which is in percent of imports.

Data for gross international reserves and currency deposits abroad are for end-September; data for external debt outstanding and external payments arrears outstanding are for end-June, unless otherwise indicated.

Coverage of external debt data differ among countries, and therefore the debt ratios may not be comparable between countries.

End-June figures.

End-May figures.

All data for Uganda are latest available estimates for end-1986.

Estimates.

Movements in gross official international reserves indicate that the need to intervene in the foreign exchange market did indeed diminish with floating, as would be expected. Reserves increased in all countries except Uganda; the increases in a number of instances were targeted under Fund programs. The data for Uganda are explained by the fact that the authorities have intervened quite heavily to support the level of the exchange rate for sustained periods. As regards the other elements of official financing “below the line,” external payments arrears have declined to date in all countries except Bolivia and Jamaica following the introduction of floating.

Absorption of Black Markets

In those countries in which exchange and trade controls have been liberalized, the illegal parallel market has been absorbed, or its scope much reduced. Absorption of such markets into the open economy has had important beneficial effects as the broader access to foreign exchange resources has contributed to better allocation of resources. Moreover, their addition to the taxable base has had positive implications for fiscal budgets.

In Bolivia, although purchasers are not required to specify their reasons for obtaining foreign exchange in the auction market, there has been from time to time after floating a thriving unofficial parallel market operated by exchange houses and by some banks, as well as by individuals transacting for themselves or for others as agents. There are three explanations for the continued existence of the parallel market in Bolivia. First, bids in the official market are required to be the equivalent of US$5,000 or a multiple thereof. Second, obtaining funds in the official market may be seen as a cumbersome process. Third, some dealers in illicit commodities prefer to operate totally outside the official sphere. With complete deregulation of the exchange and trade systems, however, arbitrage has generally functioned to ensure the same rate for legal transactions. Available information suggests that there is no black market in the Dominican Republic, the Philippines, and Uruguay, which maintain relatively liberal exchange and trade regimes (nor did the Dominican Republic have such a market prior to floating).

In Zaïre, the spread between parallel market rates and the official exchange rate has also narrowed considerably after unification and floating, to no more than 12 percent—the peak reached in January 1984. Transactions in the parallel market are related primarily to tax evasion and smuggling.

In Uganda and Zambia, which have the least liberalized restrictive systems of the group of countries surveyed here, incentives for widespread parallel market transactions have remained considerable. Following the unification of the market in Uganda, the scope for illegal transactions narrowed considerably when the auction system was operating effectively. However, the premium in the black market generally remained about 30 percent above the auction-determined exchange rate, because only current transactions, including debt servicing, were permitted to be effected through the auction market and there was demand for foreign exchange for capital flight. The spread has since widened substantially with the de facto fixing of the rate as the unofficial market began once again to play an important role. According to available information, a sizable discount remains also in the auction market in Zambia, although it has been substantially less than the discount in the pre-float period.

Macroeconomic Performance

Output and Trade

The impact on economic growth of an initial depreciation resulting from the floating of the exchange rate does not, in principle, vary from the impact of exchange rate changes which occur under other flexible exchange rate regimes, and depends critically on the support provided by demand management and supply policies. The six countries with floating exchange rates during Fund programs that are surveyed in this section have all suffered from serious balance of payments problems and overabsorption. Floating the exchange rate in conjunction with adoption of appropriate adjustment policies may therefore have an initial contractionary impact on the real economy, although this may be partly offset by improved financing arrangements with creditors, while in the medium term growth would be assisted. A real effective depreciation has an immediate negative impact on domestic absorption, as it reduces real wealth through its impact on the price level and real income, unless nominal wage increases fully accommodate the impact of the depreciation. However, the effects of the real effective depreciation in switching expenditures toward more competitive industries are ultimately growth-enhancing.25 The role of macroeconomic policies in supporting the floating exchange rate systems is crucial, and differences in performance cannot be attributed to the system per se but rather to the real effective exchange rate adjustment that took place and the full package of measures of which the adoption of the float was a part. The experience of these six countries is too short to gauge fully the growth and balance of payments responses because of the lags necessarily involved in the supply response to the change in relative prices. Moreover, the pertinent benchmark for comparison is how these countries would have performed in the absence of floating and the associated economic program supported by the Fund, which, by its nature, is not directly measurable. Also, as noted earlier, balance of payments developments cannot be attributed solely to exchange rate movements. However, in general, floating exchange rates appear to have contributed to favorable and envisaged macroeconomic effects or to have contained a further deterioration in the external position.

In most of the six countries adopting floating exchange rate arrangements, there was initially a sharp contraction of imports, reflecting the combined result of the change in relative prices, monetary and fiscal adjustment measures, and continuing use of import controls (see Table 13 in the Appendix). On the supply side, the real effective depreciation that has ensued may be expected to promote exports and import substitution through the change in relative prices. In addition, the real effective depreciation might increase exports recorded in national output figures as smuggled exports are replaced by exports through customs. However, the countries under review have suffered from structural problems as a result of their exports being concentrated in relatively few commodities that have been subject to both excess world market supply and protectionist measures (e.g., bauxite/aluminum, meat, sugar, and wool products), and diversification of exports has also been handicapped by the delayed effects of an unrealistic exchange rate. Although exports of traditional commodities have been made more profitable in those countries in which floating has led to a real effective depreciation of the currency, the stimulus has been primarily to nontraditional exports and services, which have expanded from a small base. After some delays, exports appear to be performing more positively following the real effective depreciation of the currencies and the package of economic measures which preceded or followed floating in Jamaica, the Philippines, Uruguay, and Zaïre—although they have continued to perform disappointingly to date in the Dominican Republic and Uganda.

Table 13.Growth and Foreign Trade Performance in Developing Countries with Floating Exchange Rates and Fund-Supported Programs, 1981-86Percentage change, in volume terms
198119821983198419851986
GNP/GDP
Dominican Republic3.91.74.00.4–2.22.0
Jamaica12.51.02.0–0.4–3.72.0
Philippines3.72.81.1–6.8–3.80.1
Uganda13.98.24.95.10.73.0
Uruguay2.0–9.6–5.9–2.30.73.0
Zaïre2.8–2.21.22.82.54.0
All developing countries2.21.61.34.13.13.0
Merchandise exports
Dominican Republic10.9–25.7–1.56.3–13.7–9.3
Jamaica1–4.0–10.4–3.516.9–12.4
Philippines1.15.6–5.0–2.3–2.21.2
Uganda122.123.60.50.4–6.1
Uruguay15.1–5.08.9–11.30.90.8
Zaïre–14.610.98.80.7–0.25.4
All developing countries–5.7–8.22.97.10.43.8
Merchandise imports
Dominican Republic–10.8–14.32.1–8.81.0–12.1
Jamaica122.0–3.3–8.50.9–10.97.0
Philippines–8.014.1–4.0–28.6–10.04.2
Uganda125.81.05.9–1.1–13.69.8
Uruguay–12.0–32.0–25.7–1.8–5.07.8
Zaïre–9.6–12.2–7.6–0.44.22.4
All developing countries7.1–4.2–3.22.2–0.3–0.6
Sources: International Monetary Fund, International Financial Statistics (Washington, various issues), and staff estimates.

Fiscal years.

Sources: International Monetary Fund, International Financial Statistics (Washington, various issues), and staff estimates.

Fiscal years.

In the Dominican Republic, the depreciation of the currency which preceded floating has stimulated primarily nontraditional exports and tourism, owing in part to the retention until recently of a surcharge on traditional exports. Overall export performance has been positive, but nevertheless disappointing in Jamaica in the face of a substantial real depreciation of the currency, largely because of delayed effects of the profitability of major exports (bauxite and alumina) that led to shutdowns in the industry. Political uncertainties have also recently dampened tourism receipts, following an initial stimulus. However, some positive effects have been evident in much stronger performance of nontraditional exports. In the Philippines, the peso appreciated initially after the float under the influence of tighter monetary conditions, contributing to a contraction of exports in 1985, although services receipts have shown some buoyancy as a result of a shift of transactions from the black market to the official market, and the 1986 performance has been stronger. Uruguayan exports have expanded only marginally on balance, despite the substantial depreciation of the peso since floating began in November 1982, owing to problems facing traditional trading partners in Latin America and in the Middle East and to protectionist practices in industrial country markets. There has been some diversification into nontraditional export commodities and increased repatriation of foreign exchange earnings. In Zaïre, part of the increase in recorded exports since floating began has been due to a shift from the unofficial to the official channels.

Imports at first declined substantially in the Dominican Republic, Jamaica, the Philippines, and Uruguay as a result of the domestic recession as well as, with the exception of the Philippines, the change in relative prices following the depreciation. However, estimates for 1986 show increased imports in all the countries but the Dominican Republic. There has been some evidence of import substitution of domestic products for imported products. In Zaïre, imports had been constrained by the lack of foreign exchange before floating. As a result, after floating and the liberalization of foreign exchange restrictions, imports have increased moderately in volume terms, despite the substantial depreciation. Further, in all cases, the flow of imports was affected by the availability of financing as part of the adjustment program.

Data for external current account balances show that the deficit (as a ratio of GDP) declined in all six countries, with the largest reductions occurring in Jamaica, the Philippines, and Uruguay (Chart 2). The current account deficit in relation to GDP declined only somewhat subsequent to the currency float in the Dominican Republic and Zaïre. The trade balance has accounted for the major part of the improvement in the current balance of payments in Jamaica and the Philippines, although the service and transfer balances have also made a positive contribution in the Philippines. By contrast, the service and transfer balance accounted for the larger part of the improvement in the Dominican Republic and in Uganda. The improvement in trade balances has reflected primarily the declines in imports discussed above.

Chart 2.External Current Account Developments in Selected Developing Countries, 1981-86

(In percent of GDP/GNP)

Source: International Monetary Fund, International Financial Statistics (Washington, various issues).

NOTE: Vertical slashes indicate the beginning of floating of exchange rate.

Inflation Effects

The impact on the rate of inflation of a switch to floating has depended crucially on the monetary and fiscal economic policies that have influenced the subsequent direction of the exchange rate changes. Domestic price liberalization has also constituted an important policy measure supplementing the floating of exchange rates in most cases, owing to extensive price controls (e.g., on agricultural and energy products) or price distortions resulting from public sector pricing policies that were in place. In most of these countries, the freeing of the exchange rate was accompanied by complete or partial removal of price controls and adjustments of public sector prices. Although the immediate impact of price liberalization has been an increase in prices, in the longer run this impact may be more than mitigated by an improvement in resource allocation resulting from correct price signals to enterprises and consumers, reflecting the opportunity cost of the goods affected, as well as by the stimulus to domestic production (e.g., the effect of appropriate grain pricing on agriculture).

Liberalization of price controls has taken place in Bolivia, The Gambia, the Philippines, South Africa, Uganda, and Zaïre shortly before or after the introduction of floating exchange rates. In Bolivia and the Philippines, all remaining price controls on consumer goods were removed—these had previously affected more than 40 percent of consumer goods in Bolivia26 and ten important consumer items, mainly food products, in the Philippines. Prices for all consumer goods except petroleum products and public sector tariffs were liberalized in Uganda. In the Dominican Republic, The Gambia, Guinea, Jamaica, Zaïre, and Zambia, floating of the exchange rate was accompanied by increases in administered and controlled prices, especially petroleum prices and public sector tariffs, to reflect the exchange rate adjustment. Moreover, in the Dominican Republic, The Gambia, Zaïre, and Zambia, the authorities committed themselves to adjusting administered prices in line with changes in costs more frequently and extensively than in the past.

Four of the countries for which sufficient post-float data is available (the Dominican Republic, Jamaica, the Philippines, and Zaïre) experienced a decline in the rate of inflation after an initial rise following the float (Chart 3), and inflation has decelerated through the third quarter of 1986 to a rate below the pre-float level in the Dominican Republic, the Philippines, and Zaïre. The decline in inflation in Zaïre, from a 12-month increase peaking at 123 percent toward end-1983 to about 10-15 percent a year by end-1985, was particularly striking. It resulted from the sharp tightening of monetary and fiscal policies that occurred at the same time that economic growth recovered. Recently, inflation has picked up again because of a relaxation of monetary policy. In the Dominican Republic, the decline in inflation reflected mainly the very tight stance of monetary and fiscal policies maintained during 1985, and inflation has since been reduced.

Chart 3.Consumer Price Developments in Selected Developing Countries, January 1981-September 1986

(Percentage change; 12-month increase in consumer prices)

Source: International Monetary Fund, International Financial Statistics (Washington, various issues).

NOTE: Vertical slashes indicate the beginning of floating of exchange rate.

Uganda is a good example of a country in which domestic prices at the consumer level had already reflected the black market exchange rate prior to floating.27 The gradual depreciation of the official rate and the transfer of transactions to the parallel market up to mid-1984 therefore had no visible impact on the rate of inflation. The subsequent surge of inflation from mid-1984, when the exchange markets were unified, was in response to a relaxation of fiscal policy in connection with a sharp increase in wages and salaries of civil servants, which was followed in turn by a marked increase in credit to the public sector and the rate of monetary expansion.

In Uruguay, in the immediate aftermath of the floating of the exchange rate, there was a 40 percent depreciation of the peso and the rate of inflation picked up considerably. Prior to floating, exchange rate policy had been deliberately used to dampen the rate of inflation at a time when there was a marked weakening in financial policies. The authorities’ initial attempts at adjustment following the floating were complicated by the large losses that began to be incurred by the central bank mainly as a result of its portfolio purchase scheme to shore up the domestic banking system. More recently, however, fiscal and monetary policies have been considerably tightened, with a view to bringing down inflation from the relatively high rates of the past three years.

Comparison of Performance Under Floating with Managed Flexibility

As noted earlier, in members’ economic programs supported by arrangements from the Fund that were concluded between January 1983 and December 1986, the major alternative to floating exchange rate arrangements was managed flexibility aimed at maintaining or increasing competitiveness by reference to a real effective exchange rate indicator. A before and after comparison of the macroeconomic developments under these two groups of arrangements concluded through December 1985 is shown in Table 5. It must be stressed that any differences in performance are the combined result of exchange rate developments, external events, and domestic policies, and may reflect greater scope for adjustment in the group of countries with floating exchange rates. Moreover, the sample of countries with adequate data on the experience with floating exchange rates (6 countries with 9 programs) is considerably smaller than that of countries with the man-aged arrangements (26 countries with 36 programs). At the present early stage of the experience with the independently floating systems, insufficient data exist to conduct meaningful statistical tests of the differences between the means of the samples drawn from the two periods.28

The available comparisons indicate that the external current account balance either improved or remained unchanged during the program year in all countries with floating exchange rates, while it worsened in one fifth of those countries with managed flexible exchange arrangements.29 The difference in these outcomes was probably attributable to fiscal performance, as the overall fiscal balance improved in two thirds of the countries with floating arrangements compared with only one half of the countries with flexible rates. However, measured in terms of the size of changes in relation to GDP, the median improvement was larger for the managed arrangements.

The inflation performance, measured as the change in the 12-month average increase in consumer prices before and after the approval of the Fund program (or latest available information), was better in countries with floating arrangements than in those with flexible exchange rates. There were 11 countries in the latter group which undertook a depreciation of their currencies at the beginning of the program. However, there was no difference in the inflation performance of the flexible exchange rate countries that initially depreciated their currencies and those that did not. With respect to monetary policy, no significant difference is apparent between the floating and managed flexible exchange arrangements. The liberalization of the exchange and trade systems was more extensive under floating than under flexible exchange arrangements, which might have contributed to differences in the inflation record.

19In calculating the real and nominal effective exchange rates, a composite exchange rate series was used in the period preceding unified exchange float for the Dominican Republic, Jamaica, the Philippines, Uganda, and Zaïre. The weights used in these composites were the estimated values of transactions in each of the multiple exchange markets.
20This initial strength of the floating exchange rate vis-à-vis the pre-float parallel market rate was also evident following the adoption of the new arrangements by The Gambia, Ghana, Nigeria, Sierra Leone, and Zambia.
21For measure (4), significant at the 95 percent probability level.
22In recent years, a number of creditors have insisted on exchange reforms as a precondition for debt rescheduling.
23Data for identified short-term capital flows (i.e., as distinct from errors and omissions) are available for only four of the six countries in 1986.
24These indicators, which are derived from partner country import and export data, are available only for those unrecorded transactions that have no balance of payments counterpart in errors and omissions. However, in addition to improper invoicing, they reflect other influences such as transit lags in trade and intercountry differences in definition and valuation. (See International Monetary Fund, A Guide to Direction of Trade Statistics (Draft) (Washington, March 1987.)
25For a discussion of the transmission mechanisms of exchange rate adjustment, see Mohsin S. Khan and Malcolm D. Knight, Fund-Supported Adjustment Programs and Economic Growth, Occasional Paper No. 41 (Washington: International Monetary Fund, November 1985).
26Price controls had not been successfully enforced in Bolivia.
27This effect is directly observable in data for several specific consumer goods comparing post-float prices in Uganda and Zaïre with the same goods in Sierra Leone prior to adoption of the floating interbank market (see Table 14 in the Appendix).
28See, for example, the tests of relative performance described in Morris Goldstein and Peter Monteil, “Evaluating Fund Stabilization Programs with Multicountry Data: Some Methodological Pitfalls,” Staff Papers, International Monetary Fund (Washington), Vol. 33 (June 1986), pp. 304-44.
29For definitions of these movements, see Table 5, footnote 1.

    Other Resources Citing This Publication