Back Matter
  • 1, International Monetary Fund
  • | 2, International Monetary Fund


The Natural Classification

This appendix summarizes the data and algorithm used to construct the Natural classification, and provides a brief summary of the main features of various de facto classifications.

The Natural classification—which classifies exchange rate regimes into fine and coarse categories as summarized in Table 8—employs monthly data on official and “market-determined” exchange rates for the period 1940-2001.55 The data on market-determined exchange rates are drawn from various issues of Pick’s Currency Yearbook, Pick’s Black Market Yearbooks, and Pick’s World Currency Report, while the official rate data are from the same sources as well as the IFS. The quotes are end-of-month exchange rates. Annual classifications are simply the modal monthly classifications for each country in each year.

Table 8.

Natural Classification Categories

article image

The procedure employed by the Natural classification to classify regimes is as follows:

  1. First, a separation is made between countries with either official dual or multiple rates or active parallel (black) markets.

  2. If there is no dual or black market, a check is done to see if there is an official pre-announced arrangement, such as peg, crawling peg, or band. If there is, the announced regime is verified by examining the mean absolute monthly change over the period following the announcement.56 If the regime is verified (according to rules analogous to those described in step 3 below), it is then classified according to the announcement.57

  3. If there is no pre-announced exchange rate path, or if the announced regime fails to be verified by the data (which is often the case), and if the twelve-month rate of inflation is below 40 percent, the regime is classified on the basis of actual exchange rate behavior:

    • If the absolute monthly percent change in the exchange rate is equal to zero for four consecutive months or more, that episode is classified (for however long its lasts) as a de facto peg (if there are no dual or multiple exchange rates in place).58

    • If the probability that the monthly exchange rate change remains within a +/−1 percent band over a rolling 5-year period is 80 percent or higher, then the regime is classified as a de facto peg or crawling peg over the entire 5-year period. If the exchange rate has no drift, it is classified as a fixed parity; if a positive drift is present, it is labeled a crawling peg; and, if the exchange rate also goes through periods of both appreciation and depreciation it is a moving peg.

    • The approach regarding de facto bands (as well as pre-announced bands) follows a parallel two-step process. Thus, if the probability that the monthly exchange rate change remains within a +/−2% band over a rolling 5-year period exceeds 80 percent, then the regime is classified as a de facto narrow band, narrow crawling band, or moving band over the entire period through which it remains continuously above the 80 percent threshold.

  4. If the twelve-month rate of inflation exceeds 40 percent, the episode is classified as “freely falling.”59

  5. The remaining regimes (those that have not already been classified by steps one through four) become candidates for “managed” or “freely” floating. To distinguish between the two, the degree of exchange rate flexibility is measured by a composite statistic, ε/Pr(ε<1%), where ε is the mean absolute monthly percent change in the exchange rate over a rolling five-year period, while the denominator flags the likelihood of small exchange rate changes. If this ratio falls inside the 99 percent confidence interval or is in the upper tail of the distribution of the floater’s group, the episode is characterized as freely floating. If the ratio falls in the lower one percent tail, the null hypothesis of freely floating is rejected in favor of the alternative hypothesis of managed float.

  6. When dual or multiple rates are present or parallel markets are active, steps two through five above are applied to the market-determined rates instead of the official exchange rates to identify the regime.

Table 9.

Main Features of Various De Facto Classifications

article image


Determinants of Exchange Rate Regime Choice

The Natural classification data show some links between de facto regime flexibility and certain macroeconomic and financial variables, such as trade openness and dollarization. However, a review of the literature suggest that it is difficult to find empirical regularities between potential exchange rate regime determinants and countries’ actual regimes that hold consistently across all countries, time periods, and regime classifications. Systematic robustness checks of the determinants of regime choice employing the Natural classification support this result.

Macroeconomic and financial characteristics of regimes

Optimum currency area (OCA) theory holds that variables such as large size and low openness to trade are likely to be associated with floating exchange rates. One reason for this may be that trade openness raises the transactions benefits from common currencies, and should be expected to lead, therefore, to a decline in the number of independent currencies. The data appear to support the OCA theory prediction that countries that trade a lot will tend to have less flexible exchange rate regimes. Advanced economies that have a high trade openness ratio have tended to have pegged regimes, while the prevalence of free floats has been notably higher in advanced countries with low external trade ratios, such as Australia, Japan, and the United States. A similar pattern holds among other developing countries, where the prevalence of managed floats has been markedly higher and pegs significantly lower in the countries that rely less on external trade. The pattern among emerging markets has been less clear, although relatively closed economies in this group have had a much higher likelihood of being in the freely falling category.

Higher dollarization appears to be associated with less flexible exchange rate regimes among emerging markets, consistent with “fear of floating.” Fear of floating appears to be strongest in highly dollarized emerging markets, where pegged regimes are more prevalent than in less dollarized countries in the group. Conversely, emerging markets with low and medium degrees of dollarization are more likely to have managed or freely floating regimes. However, fear of floating does not explain why other developing countries with high dollarization ratios appear to prefer regimes with limited flexibility to pegs. A possible explanation for this could be that many of these countries became highly dollarized following a “freely falling” episode, and lacked the credibility necessary to defend a peg. A regime with limited flexibility allowed them to obtain the benefits of a relatively stable currency, while at the same time maintaining some ability to adjust to shocks.

There is little systematic relation, however, in the degree of capital account openness across de facto regimes. Emerging markets and other developing countries tend to have more capital controls and lower capital flows, in relation to GDP, than advanced economies. Nevertheless, the variation in capital account openness appears not to be related to the flexibility of countries’ currency regimes. Among advanced economies, the volume of capital flows in countries with de facto pegged regimes tends to be higher than in those with intermediate regimes, and significantly higher than for those with freely floating regimes. The relationship is more mixed, however, for emerging markets and other developing countries, possibly because capital controls are often ineffective, so the expected inverse relation between controls and observed capital flows may not hold.

Empirical findings on factors affecting regime choice

Systematic prediction of exchange rate choice is elusive. A review of a reasonably broad collection of previous studies shows that different empirical studies using the de jure and other de facto regime classifications have often obtained different results, suggesting that it is very difficult to draw general conclusions about how countries choose their exchange rate regimes. Although certain characteristics have been shown to be important in determining exchange rate regime choice in certain groups of countries, and certain characteristics may distinguish countries in certain regimes from those in different regimes, no result appears fully robust to changes in country coverage, sample period, estimation method, and exchange rate regime classification.

Several empirical studies have analyzed the determinants of exchange rate regime choice in a cross section of countries. Among the first studies of this kind are Heller (1978), who analyzed the determinants of exchange rate regimes with data from the mid-1970s, soon after the generalized floating that followed the breakup of the Bretton Woods system, Dreyer (1978), Holden, Holden, and Suss (1979), Melvin (1985), Bosco (1987), Savvides (1990), Cuddington and Otoo (1990 and 1991), Rizzo (1998), and Poirson (2001). Some studies, such as those by Collins (1996), Edwards (1996 and 1998) and, more recently, Frieden, Ghezzi, and Stein (2000), have used random effects panel data to analyze also the determinants of changes in exchange rate regime. As such, they can be seen as somewhat related to the recent literature on predicting exchange rate crises. Nevertheless, we include these studies in our review because they report findings on the role of country characteristics that are relatively stable over time (such as openness) in determining exchange rate regime choice. Another recent study, by Berger, Sturm, and de Haan (2000), uses panel data in an attempt to identify the long-run determinants of exchange rate regime choice. Additional studies addressing changes in exchange rate regimes include Masson (2001), Klein and Marion (1994), and Duttagupta and Otker-Robe (2003).

The vast majority of previous studies have attempted to explain countries’ de jure exchange rate regime choice. A few studies have constructed and used measures of the degree of de facto flexibility on the basis of the actual observed volatility of exchange rates and reserves, including Holden, Holden, and Suss (1979) and, more recently, Poirson (2001). Table 4 summarizes the approaches and findings of these studies with regard to the impact of several variables on observed exchange rate regime choice. Most studies considered some of the optimum currency area variables, such as trade openness (typically measured as imports plus exports, divided by GDP), the size of the economy (gross domestic product in common currency), the degree of economic development (GDP per capita), and geographical concentration of trade (the share of trade with the country’s main partner). Among macroeconomic variables, several studies included inflation (whether the country’s own inflation, or inflation in excess of partner countries) and foreign exchange reserves. Many studies included an indicator of either capital controls (typically also drawn or constructed from the IMF’s Annual Report on Exchange Arrangements and Exchange Restrictions) or de facto capital openness (e.g., the ratio of foreign assets of the banking system to the money supply). Some studies included measures of volatility of domestic output, exports, domestic credit, or the real exchange rate, although no two studies seem to have looked at the same measure of volatility. A few studies considered variables related to political economy or institutional strength. Most studies analyzed some variables that were not included in any preceding (or subsequent) studies. Collectively, the studies considered more than 30 potential determinants of exchange rate regime choice. (Only the variables considered by more than one study are included in Table 4.)

No result appears to be reasonably robust to changes in country coverage, sample period, estimation method, and exchange rate regime classification. For example, openness—the most frequently analyzed variable—is found to be significantly associated with floating regimes by three studies, significantly associated with fixed exchange rates by three studies, and not significantly associated with any particular exchange rate regimes by another five studies. Per capita GDP is found to be significantly associated with floating regimes by three studies, significantly associated with fixed exchange rates by two studies, and not significantly associated with any particular exchange rate regime by another three studies.

There are a few possible exceptions, notably size of the economy and inflation. Size of the economy turns out to be positively associated with floating in almost all studies, though not always significantly. Inflation is almost always positively and significantly associated with floating. However, in the case of inflation there are serious questions regarding the functional form of the relationship. In a number of studies, the authors use the inflation rate or the inflation differential (rather than their logarithms or similar transformations), leaving open the possibility that the results might be driven by a few influential observations. Morever, Collins (1996) finds that high inflation affects exchange rate regime choice in the opposite direction than low/moderate inflation does, and significantly so.

New empirical tests using the Natural classification confirm that it is difficult to explain how countries choose their exchange rate regimes on the basis of simple empirical regularities. These results are consistent with previous work based on other exchange rate regime classifications (Juhn and Mauro (2002)). For a number of potential determinants of regime choice—including economic size, trade openness, capital controls—the variation across regimes is statistically significant. However, with the possible exception of economic size and trade openness, none of the variables is consistently significant across varying specifications in probit and multinomial logit regression analysis. This suggests that the macroeconomic, structural, and institutional variables postulated in various theories are not robust predictors of exchange rate regime choice. Of course, this does not preclude the potential importance of certain variables for certain groups of countries, in certain time periods, or across some of the regime categories.

Table 10.

Studies on Determinants of Exchange Rate Regimes (Likelihood to Float)

article image
article image

indicates that the coefficient of explanatory variable is positive and – that is negative.

indicates the coefficient is either positive or negative depending on the specification or method used.

indicates the coefficient is statistically significant in most cases.

indicates the coefficient is statistically significant in some specifications.

indicates not significant but sign not reported by the author.


Data and Regression Results for Economic Performance Analysis

This appendix describes the data used in section III and reports the detailed regression results that lie behind the key findings discussed with respect to economic performance across exchange rate regimes.

Much of the data are taken from Ghosh, Gulde, and Wolf (2003), including the de jure classification of exchange rate regimes, the three measures of economic performance (inflation, growth, growth volatility), and the control (or explanatory) variables used in the regression analysis. Each variable is covered at an annual frequency from 1970 to 1999 for up to 158 countries. The control variables are drawn from the literature and are thought to provide a suitable explanation of the variations in the performance measures. Table 11 provides a detailed description of the data. It lists each variable, provides a brief description, and notes which of the subsequent regressions feature these variables. Using this data has the advantage that the evaluation of performance under the Natural classification can be directly compared to a well-respected baseline that assesses performance across the de jure regime regimes.

Three groups of variables are not covered in the Ghosh, Gulde, and Wolf (2003) data. The first group is the Natural regime classification, available at an annual frequency from The second group is the crisis variables. The banking crisis variable is taken from Demirguc-Kunt and Detragiache (1998). They define a banking crisis to have occurred when any one of the following four conditions held: (a) non-performing loans exceeded 10 percent of banking system assets; (b) a bailout cost 2 percent or more of GDP; (c) large scale nationalization occurred; or (d) other emergency measures, such bank holidays, deposit freezes, and special guarantees had to be undertaken. The currency or balance-of-payments crisis variable is taken from Berg, Borensztein, and Pattillo (forthcoming) who define a crisis as having occurred when the weighted average of one-month changes in exchange rate and reserves is more than three (country-specific) standard deviations above the country average.

The final group of variable defines whether a country is classified as an advanced economy, an emerging market, or a developing country. Advanced countries are defined using the World Bank definition for upper income countries, following Ghosh, Gulde, and Wolf, (2003). In dividing the rest of the world into two further groups, the analytical distinction of relevance was their degree of exposure to international capital markets. Those considered to have high exposure were classified as “emerging markets” and the rest were designated “developing.”61 Table 12 lists the country composition of the advanced, emerging market and developing country groups.

To distinguish between emerging and developing economies, exposure to international capital can be determined either in a de jure sense (the extent of formal capital controls in place) or in a de facto sense (the actual exposure a country faces). In the spirit of this paper, a de facto definition was appropriate, an approach also followed by Prasad, Rogoff, Wei, and Kose (2003). Since there are no well-defined or generally accepted thresholds of exposure to international capital, the cut-off between high and low exposure can be arbitrary and was dealt with by dropping and adding countries on the margin to check the robustness of the results. In this paper, the emerging markets are defined using the Morgan Stanley Capital International classification, which designates a country as an emerging market according to a number of factors: GDP per capita, local government regulations, perceived investment risk, foreign ownership limits and capital controls, and other factors. The main motivation for using this classification is that it captures the notion that these countries have access to international capital markets. See for more information. In checking for the robustness of results presented, India and China (considered to have relatively closed capital accounts) were dropped from the emerging markets sample but the results were unchanged. Countries added to the list included those that are not on the MSCI index but do appear on other international emerging market indices and also such countries as Bahrain, Lebanon, and Tunisia that are not on any list but are thought of as relatively open to international capital markets. Again, the results were robust.

All regressions seek to identify the effects of the exchange rate regime, conditional on (or after taking into account) the influence of the conventional control variables relevant to that performance measure. All regressions also include two additional controls, which are not reported for brevity. First, common shocks across countries (such as spikes in oil prices or changes in the volatility of G-3 currencies) are controlled for through time dummies. Second, to control for unobserved country-specific characteristics that are constant over time, country dummies are included. The implication of this approach is that regime performance is judged by changes that occur within a country rather than across countries. For comparison, however, this appendix also discusses below results without country fixed-effects, hence taking into account differences across countries.

To briefly recap, the figures presented in Section III are based on these regressions. They present the coefficients on “dummy,” or categorical, variables representing the exchange rate regime. The dummy variable takes the value 1 if the exchange rate regime prevails in a country in a particular year; otherwise, it is assigned a value of zero. As is well-known, when a set of dummy variables represents the full range of possibilities (in this case, the full range of exchange rate regimes) then regression analysis requires one of the possibilities to be left out. The regime that is left out is the base against which the others are compared. Hence, the coefficients presented in figures are to be interpreted as measures of performance (relative to the excluded pegged regime) and conditional upon the other included variables in the regression.

Table 13 compares economic performance (inflation, growth and growth volatility) across regimes, contrasting the de jure classification with the Natural Classification. Table 14 evaluates inflation performance across all countries, advanced countries, emerging markets and developing countries. Three different specifications are presented: (1) the estimates with country fixed effects (on which the figures in the main text are based); (2) the same specification but without fixed effects; and (3) a specification with fixed effects but with the regime variables lagged by two years. The lagging of the exchange rate regime variables increases the likelihood, though does not ensure, that the results are reflecting the influence of regimes on performance rather than the other way around. Tables 15 and 16 are analogous, except that they examine growth and growth volatility, respectively. The different specifications show that the qualitative direction of the key results presented in the main text hold up with remarkable consistency. Where the results across specifications are not similar—as for inflation in advanced countries or inflation and volatility in emerging markets—these are discussed in the text.

Table 17 reports results for emerging markets in the 1990s to recognize that exposure to international capital markets became widespread mainly in that decade. Table 18 reports the inflation regression results, which include regime-specific announcement and duration variables. Finally, Table 19 summarizes all other robustness tests, which have been omitted for brevity.

Table 11.

Variable Description

article image
Table 12.

List of Countries

article image
Note: Emerging market economies are those that are included in the Morgan Stanley Capital International (MSCI) index. With the exception of Israel, which is in the MSCI index, advanced economies are those that are classified as upper income economies by the World Bank. All other economies constitute the other developing countries group.
Table 13.

Comparing IMF De Jure and Natural Classifications

article image
Source: Authors’ calculation.Note: Figures in parentheses are t-statistics; * significant at 10%; ** significant at 5%; *** significant at 1%.
Table 14.

Inflation Performance Across Country Groups

article image
Source: Authors’ calculation.Note: Figures in parentheses are t-statistics; * significant at 10%; ** significant at 5%; *** significant at 1%.
Table 15.

Growth Performance Across Country Groups

article image
Source: Authors’ calculation.Note: Figures in parentheses are t-statistics; * significant at 10%; ** significant at 5%; *** significant at 1%.
Table 16.

Volatility of Real GDP Growth Performance Across Country Groups

article image
Source: Authors’ calculation.Note: Figures in parentheses are t-statistics; * significant at 10%; ** significant at 5%; *** significant at 1%.
Table 17.

Emerging Markets in the 1990’s

article image
Source: Authors’ calculation.Note: Figures in parentheses are t-statistics; * significant at 10%; ** significant at 5%; *** significant at 1%.
Table 18.

Inflation Performance in Developing Countries: Announcement and Duration Effects

article image
Source: Authors’ calculation.Note: Figures in parentheses are t-statistics; * significant at 10%; ** significant at 5%; *** significant at 1%.
Table 19.

Robustness Tests

article image


  • Alesina, Alberto, and Alexander Wagner, 2003, “Choosing (and Reneging on) Exchange Rate Regimes,” NBER Working Paper No. 9809 (Cambridge, Massachusetts: National Bureau of Economic Research).

    • Search Google Scholar
    • Export Citation
  • Bailliu, Jeannine, Robert Lafrance, and Jean-Francois Perrault, 2002, “Does Exchange Rate Policy Matter for Growth?Bank of Canada Working Paper, 2002-17.