IV. Assessing External Vulnerability: The Case of Chile46
Author: International Monetary Fund
Publication Date:
22 Aug 2000
eISBN:
9781451807530
Language:
English
Keywords:
ISCR; CR; monetary policy shock; consumption tax; interest rate defenses; Can time; current account; Phillips curve model; forecasting inflation; Exchange rates; Inflation targeting; Consumption; Inflation; Australia and New Zealand; East Asia; Africa; stock ratio; liquidity position; BIS short-term debt data; per capita income; Chilean policy; Emerging and frontier financial markets; Current account deficits; Yield curve
Download PDF (1.9 MB)

110. This study considers the Chilean economy’s degree of external vulnerability, focussing on evidence from the 1997-99 period. Adding to interest in this subject are several related developments of the last few years. First, the occurrence of severe, largely unanticipated crises in countries in East Asia—which like Chile have been considered star performers—suggests a need for frequent re-evaluation of perceptions and methods of assessment as well. Second, there indeed has lately been considerable research and data standards development aiming to improve the assessment of vulnerability. Third, recent shocks, such as the one emanating from Russia in 1998, provide an opportunity to learn about Chile’s relative vulnerability by comparative observation.

Abstract

110. This study considers the Chilean economy’s degree of external vulnerability, focussing on evidence from the 1997-99 period. Adding to interest in this subject are several related developments of the last few years. First, the occurrence of severe, largely unanticipated crises in countries in East Asia—which like Chile have been considered star performers—suggests a need for frequent re-evaluation of perceptions and methods of assessment as well. Second, there indeed has lately been considerable research and data standards development aiming to improve the assessment of vulnerability. Third, recent shocks, such as the one emanating from Russia in 1998, provide an opportunity to learn about Chile’s relative vulnerability by comparative observation.

A. Introduction

110. This study considers the Chilean economy’s degree of external vulnerability, focussing on evidence from the 1997-99 period. Adding to interest in this subject are several related developments of the last few years. First, the occurrence of severe, largely unanticipated crises in countries in East Asia—which like Chile have been considered star performers—suggests a need for frequent re-evaluation of perceptions and methods of assessment as well. Second, there indeed has lately been considerable research and data standards development aiming to improve the assessment of vulnerability. Third, recent shocks, such as the one emanating from Russia in 1998, provide an opportunity to learn about Chile’s relative vulnerability by comparative observation.

111. The approach used here is eclectic and informal, surveying a broad range of indicators. The assessment of Chile is distinguished by these characteristics:

  • Examination of vulnerability using modern approaches, in particular the emphasis on liquidity, and greater attention to data issues.

  • Emphasis on cross-country comparison, including to advanced economies. The paper does not presume that a country such as Chile is most usefully compared to other emerging market economies. Instead, comparison is made to a group in which Chile would be centrally placed: the higher-rated emerging market countries, and a select group of advanced economies.

  • Consideration of the crisis that emanated from Russia in 1998, contrasting Chile’s experience in this period with that of other countries.

  • Considering information from a range of sources: not only countries’ official data, but also credit ratings, bond spreads, surveys of forecasts, and creditor-supplied debt data.

112. Several secondary goals of the paper can be noted. One is to identify any areas for improvement in the set of information now available for assessing Chile’s external position. Another objective is to shed light on the usefulness of modern approaches to assessing external vulnerability, by highlighting Chile-specific issues that arise in applying such approaches, and that may raise questions about the applicability of one-size-fits-all approaches. In addition, there is the question of whether the indicators normally used in assessing emerging markets and developing countries are also usefully applied to advanced economies.

113. The organization of this paper is as follows. Section B outlines the kinds of indicators to be examined and discusses the selection of comparator countries. Section C examines four summary indicators that reflect assessments made by private sector agents, emphasizing how these were influenced by the 1998 Russia shock. Section D looks at solvency indicators, and Section E addresses liquidity indicators. Section F is a Chile-specific discussion that briefly addresses factors that are qualitative or less amenable to cross-country comparison. Here also, several areas are noted where further information would deepen analysis of Chile’s external vulnerability. Section G summarizes and concludes.

B. Choice of Indicators and Sample

Quantitative indicators to be examined

114. Indicators of Chile’s external vulnerability can be grouped into three categories.

115. The first category includes summary indicators of market and other private assessments, the values of which may synthesize a great deal of underlying characteristics. Regardless of whether such assessments are always valid, they are for some purposes “what counts.” Such indicators used in this study are interest rate spreads on sovereign bonds; expected exchange rates, measured using surveys of forecasts; indices of exchange market pressure; and credit ratings. These indicators leave to others—market participants and ratings agencies—the task of weighing a vast set of potentially relevant information for each country. On the other hand, their summary nature limits their usefulness for policy purposes, since they may not directly indicate specific policy actions.

116. The other two categories include specific quantitative indicators, each of which may measure one of the underlying determinants of an economy’s vulnerability.47 The category of solvency or “fundamentals” indicators in Section IV include debt stock ratios, real exchange rate measures, export volume growth, and the current account. The category of liquidity indicators and simple “stress tests” of Section V, which have received most attention recently, are variations on the theme of the relationship of liquid assets to potential short-term liabilities.

Selection of a comparator country sample

117. A comparison of Chile with other developing countries or emerging markets might only confirm the conventional wisdom of Chile as a star within this group. It may be more useful to draw lessons from comparisons with a group of countries in which Chile is more likely to be centrally placed—in particular if this group includes some advanced economies. Chile remains outside this club in terms of per capita income, but the question is whether it otherwise has already reached a similar position in terms of external soundness, or how it could do so in the future.

118. A group of 26 comparator countries was selected using the following criteria:

(i) including a group of advanced economies—but only those which like Chile reasonably fit the “small open economy” label; (ii) also including emerging markets, but emphasizing the apparently stronger ones (based on credit ratings), (iii) excluding countries operating inside monetary unions or using currency board arrangements; and (iv) excluding countries whose currency is widely held as a form of international reserves.

119. The sample of 27 countries, including Chile, is listed below, grouped according to Standard & Poor’s long-term, foreign currency credit ratings just before the Asian crisis broke, in relation to Chile’s rating; the figures in parentheses represent each country’s ranking in terms of per capita income.48

article image

120. Chile is indeed centrally placed in this sample, with respect to both credit rating and per capita income (note the strong correspondence between ratings and income). In the analysis below, it will be useful to split the comparator countries into “higher-income” (14 countries) and “lower-income” (12 countries) subsamples, defined with respect to Chile. Reference is also made to “advanced economy” and “emerging market” countries, a distinction that corresponds fairly closely to the income-based classification.

121. As regards the exchange rate regime, other than excluding currency board and monetary union cases, it was not possible to confidently select a sample based on strict similarity to Chile’s regime. (Such regimes are in some cases ambiguous, and the management of exchange rates is often a question of degrees.) It should be kept in mind that the sample includes some variation in regimes, not only across countries but also overtime. Indeed, a number of these countries, like Chile, began the 1997-99 study period with highly managed exchange rates but have since moved to greater flexibility.

Data

122. Nearly all the data used are from publicly-available datasets, primarily International Financial Statistics, World Development Indicators, and the Joint BIS-IMF-OECD-World Bank statistics on external debt. For any given country, these data may have important shortcomings; certainly they are not necessarily superior to country-specific data that may be available from other sources. In a sample of this size, however, country-specific discussion and judgmental selection of data is not feasible; impartiality also argues for systematic use of such cross-country datasets. However, certain limitations of these data will be discussed along the way, and for the case of Chile only, reference will also be made to alternative data in a few areas.

123. Other data sources include Bloomberg, for bond yield data, and the web site of Standard & Poor’s for a detailed history of their credit ratings. Reference is also made to the survey of exchange rate forecasts appearing in FT Currency Forecaster.50

C. Market and Other Private Assessments

124. The indicators examined in this section are based on the following: credit ratings; sovereign bond spreads; an index of exchange market pressure; and revisions in exchange rate forecasts. All four indicators are first used to compare responses to a period of global stress; the first two are later used to compare Chile to other countries at a point in time.

Comparisons of responses to the 1998 crisis

125. The question here is whether Chile was affected more or less than other countries during the crisis that emanated from Russia in 1998. This episode is considered as a natural, albeit imperfect, experiment, which may reveal weaknesses not apparent during more normal times.

Credit rating changes

126. Table 1 presents revisions in the long-term, foreign currency ratings reported by Standard & Poor’s during 1997-99. Changes in rating, or in rating “outlook,” are shown according to the six-month window in which they occurred.51 The window of primary interest is the second semester of 1998 (1998H2), corresponding to the Russia crisis. Note that the 1997H2 window corresponds to the East Asian crisis, while the other four periods give an idea of the frequency of ratings revisions during more normal times.

127. Almost half the sample countries experienced a downgrade during 1997-99; Chile was one of six with no rating change, in either direction. As expected, most downgrades occurred in 1997H2 or 1998H2.52 During the other four sub-periods, downgrades were much less frequent, no more than two per semester. (Similarly, upgrades were less frequent in 1997H2 and 1998H2 than at other times.)

128. Of the many downgrades that occurred inl997H2, nearly all involved Korea or Thailand—there is no sign of a general negative effect outside East Asia in this period. On the other hand, the apparent effect of the Russia crisis, in 1998H2, was more general, with negative changes for six of the 27 countries. This group includes not only Brazil and Mexico (see Bussiere and Mulder, 1999), but also Cyprus, the Czech Republic, the Slovak Republic, and New Zealand. From this outcome (and in contrast to that for the indicators considered next) there is no suggestion that higher-income countries were less vulnerable than the lower-income economies.

Changes in bond spreads during the Russia crisis

129. Here we examine how far spreads on foreign-currency denominated bonds moved in response to the global crisis that occurred following the August 1998 events in Russia. The most dramatic events of this crisis unfolded in mid-August, but its full magnitude took time to develop, so it is useful to consider a horizon of a few months.53

Table 1.

Changes in Credit Ratings, 1997-9

(Standard & Poor’s Long-Term, Foreign Currency Ratings 1/)

article image
Source: Standard & Poor’s.

Cases marked with an asterisk (*) reflect a change in rating “outlook” only.

Downgrade of Thailand on January 8,1999 is here shown as if it occurred during 1997H2.

130. Table 2 reports average spreads on sovereign bonds during July and October of 1998. Since no Chilean sovereign bond was traded during this period, the state-owned Banco del Estado is used.54 In not one case did a bond spread decline from July to October 1998, though the range of responses was wide. For some, the increase in spread was negligible. The advanced economies—which also started off with the lowest spreads—show increases of no more than 35 basis points. Chile started with a spread ranked 9th of 26; this was the lowest among the emerging markets and similar to that of such higher-income countries as Australia, Denmark, and New Zealand. From July to October, Chile’s spread rose by about 65 basis points; while not severe, this was about twice the increase experienced by Australia and New Zealand. In terms of changes, Chile’s position was 12*, as Israel and Czech Republic, for example, showed slightly smaller increases. In terms of both levels and changes, there is a wide gap between Chile’s spread and that of most emerging markets, many of which experienced severe increases in spreads.

131. Although bond spreads suggest that Chile’s riskiness (more precisely, risk premium) was somewhat affected during the Russia crisis^ qualitatively Chile appears closer to the advanced economies than to most of the emerging market countries considered here. This is not to say that the negative effect of this crisis on Chile was economically insignificant—note the large jumps in the spreads on the bonds of Chilean private enterprises.

Exchange market pressure

132. The degree of external vulnerability may be reflected in the price of foreign exchange—that is, in various signs of “pressure” on the domestic currency during times of stress. We gauge such stress in two ways. The first, conventional, approach is to construct an index of exchange market pressure (EMP), taking a weighted average of actual currency depreciation and (a proxy for) foreign exchange market intervention during a given period. Such indices have been used for various purposes; Bussiere and Mulder (1999) have recently used one to measure the impact of crisis episodes.55

Table 2.

Sovereign Bond Spreads, July-October 1998 1/

(In basis points)

article image
Sources: Bloomberg; IP. Morgan; and Merrill Lynch.

Monthly averages, based on daily data. Sovereign bonds, where available; no data for Singapore.

indicates approximate value.

133. An EMP index was constructed as a monthly series for each country, taking a weighted average of the observed rate of currency depreciation and the reported decline in official reserves (the latter scaled by the stock of reserve money).56 Thus a positive EMP value would in general indicate market pressure against the national currency, though the measure may not be a precise one. Note that the index will not capture policy intervention in the forward exchange market, nor the role of an interest rate policy response in deflecting currency pressures. 57

134. Table 3 summarizes the behavior of EMP before and during the 1998 Russia crisis. During the last few years before this crisis, EMP measured on a monthly basis was on average close to zero in this sample, as one would expect of normal times. In August 1998, however, average EMP jumped sharply upward.58 By this measure, the countries most affected were Canada, Mexico, New Zealand, and Turkey. At this one-month horizon, there is no sign that higher-income countries were less affected than others, in contrast to the evidence found for bond spreads. Chile does show positive EMP in August 1998, but less than most others, and only a fraction of the mean or median EMP of the sample. Chile’s ranking, in terms of avoiding EMP during this crisis, is 7th of 26 countries.

135. In considering slightly longer horizons—August-September or August-October—the main change is that average EMP falls back, and even goes negative, for the higher-income countries. Chile’s (small) EMP shows this pattern as well. In contrast, average EMP for the lower-income countries does not improve over these longer horizons.

Revisions in exchange rate forecasts

136. This second indicator related to the price of foreign exchange requires a measure of expected future exchange rates. For this purpose, we use the 12-month ahead “Combined Consensus Forecasts” published monthly in FT Currency Forecaster. The idea is that a negative shock may induce a revision in the expected future level of the exchange rate, regardless of whether the shock is immediately reflected in the spot rate or (traditional measures of) official reserves, and thus the EMP.

Table 3.

Exchange Market Pressure During Russia Crisis, August-October 1998

article image
Source: Calculated from data in International Financial Statistics.

For Turkey, EMP index is adjusted for trend depreciation.

137. Table 4 shows percentage revisions in 12-month ahead expected spot rates, taking the forecasts surveyed in late July 1998 as the pre-crisis base, and considering how much the levels of these had been changed by the time new surveys taken one, two, and three months later.59

138. Most of the forecast revisions—including those for Chile—are indeed in the direction of more depreciation vis-à-vis the U.S. dollar. (This shift cannot be attributed to an improved outlook for the dollar, forecasts for which were weakening against other reserve currencies during this time.) Again, there is some sign that the higher-income countries were on average less affected than others.

139. Regarding the Chilean peso, the magnitude of the forecast revision does not stand out; in fact it is similar to the sample mean and median values. At all three horizons, Chile’s rank is 13th of the 22 countries for which forecast data are available.

Chile’s current position: credit ratings and bond spreads

140. Putting aside the question of differential reactions to common shocks, we close this section with brief reference to two summary indicators of Chile’s current relative position.

141. The first is Standard & Poor’s long-term, foreign currency ratings, as of mid-March 2000. From the sample of 26 comparator countries, 10 are still rated above Chile (this group no longer includes Czech Republic, Korea or Thailand); Israel and now also Greece and Czech Republic share Chile’s A- rating; and the remaining 13 countries have lesser ratings. Second, as regards bond spreads, in March 2000 it remained true that Chile’s spread was among the lowest in the group of emerging markets.

* * * * *

142. Chile began the 1997-99 period being evaluated by market participants as a bit behind the advanced economies—but near the top of the emerging markets—considered here, and this remains true in early 2000. As regards the impact of the 1998 Russian crisis, three of the four indicators examined (the exception being credit ratings) suggest that the lower-income countries were usually more negatively affected than the higher-income ones, though the latter were not always immune. It was revealed that Chile also was not immune to the effects of a crisis such as this one, though the other emerging markets in the comparator group tended to be more strongly affected. In terms of both bond spread and EMP responses to this shock, Chile looks more like an advanced economy.

Table 4.

Revision in One-year Ahead Exchange Rate Forecasts, August-October, 1998

(Percentage changes, from forecasts surveyed end-Jury 1998) 1/

article image
Source: Calculated from forecasts published in FT Currency Forecaster.

Based on exchange rates expressed as domestic currency units per U. S. dollar.

Figures not adjusted for trend depreciation of Turkish currency.

143. In closing this section, several interpretative limitations deserve note. First, the four indicators analyzed—all of which are related to the value of a country’s financial assets—cannot capture all aspects of vulnerability, such as the effects of external shocks on output and employment. Second, on the apparently lesser impact of the Russian crisis on Chile, these indicators do not distinguish how much may have been due to a policy response that impressed market participants, as opposed to a stronger position already in place when the crisis began. Thus in August 1998, the operational policy of interest rate targeting was temporarily abandoned by Chile, and the overnight (real) interest rate was essentially allowed to float to a level determined in the credit market, from about 8 percent in early August to over 30 percent by early September. This policy response, not captured by any of the indicators examined above, may have been more effective in stemming currency pressures. Third, the indicators examined in this section reflect the “what” rather than the “why.” The rest of this paper looks at the latter question, examining specific factors that may underlie Chile’s external position.

D. Indicators of Solvency

144. Although recent work has focussed on liquidity, solvency issues are still considered the “fundamentals.” But solvency remains inherently difficult to assess: since it is a long-run concept, judgments about solvency rest only partially on an observable past and present, depending also on conjectures about future behavior.

145. This section is organized around two groups of solvency indicators, all of which rely only on currently-available data: (i) stock data, focussing on the level of external debt; and (ii) price and flow variables, including such indicators of “competitiveness” as real exchange rate measures, performance of export volumes, and current account behavior.60

Debt stock indicators

146. Solvency of the public sector is probably a necessary, though not sufficient, condition for avoiding external crises. In Chile, the total debt of the nonfinancial public sector—both external and domestic—is clearly manageable, about 10 percent of GDP in 1999.

147. Total external debt is much higher, on the order of 50 percent of GDP or 200 percent of merchandise exports in recent years, but manageable for a country with strong growth of output and exports, moderate spreads on external debt, and sizable foreign asset holdings.

148. Comparison of Chile’s total debt stock to that of other countries in the sample is problematic. The debt data used here, from World Development Indicators (WDI) may differ in coverage or accuracy. Moreover, this dataset does not include any of the countries ranked 1st-12th in per capita income. Considering the 15 countries with data available, Chile’s debt was close to the median in 1997,61 scaled by either GDP or exports (Figure 1a). (In all charts such as these, countries are presented in order of increasing income per capita; data for Chile appear near the center.) WDI debt data for 1998 and 1999 are not yet available, but it is likely that Chile’s ranking deteriorated in those years. (Figure 1b presents these same ratios, but calculated using an external debt concept published by the BIS. While this series is available for the foil sample, it tends to understate the debt of the lower-income countries, since it does not include official credits; comparisons between higher- and lower-income countries should probably be avoided here.)

149. While Chile’s external debt ratios cannot be called low, its stock of foreign assets is believed to be considerable, although comprehensive International Investment Position data have not been published. Merely considering official reserves, and ignoring all privately-held external assets, changes the picture greatly: e.g., subtracting official reserve assets from the stock of external debt would reduce Chile’s debt ratios by about one-third.

150. Still, Chile’s gross external debt has accelerated: in just the three years through end-1999, external debt rose by close to 50 percent in U.S. dollar terms. The increases in 1997 and 1998 were roughly in line with the uncustomarily high current account deficits of those years (Figure 2). However, strong debt growth continued into 1999, despite the near-disappearance of the current account deficit. This development did not reflect any shift away from equity inflows (which increased) but rather on an accelerated accumulation of external assets by the private sector. Such portfolio adjustment is presumably mainly of a one-time nature and so would not be expected to result in rapid debt growth on an ongoing basis.

151. To the extent that it finances asset accumulation, growth of external debt does not raise solvency questions. In principle, it could be a concern as regards liquidity, but to gauge liquidity risk, extensive “micro” information would be needed, to learn whether firms doing the borrowing are the same ones increasing their external assets; how liquid are such assets; and whether firms doing the borrowing have earnings mainly in foreign exchange or have otherwise hedged themselves. (An indirect, more qualitative way of assessing the risk would involve assessment of the financial regulatory regime; section F briefly remarks on this.)

Figure 1a.
Figure 1a.

Selected Countries: External Debt (World Development Indicators), 1997

Citation: IMF Staff Country Reports 2000, 104; 10.5089/9781451807530.002.A004

Sources: Joint BIS-IMF-OECD-World Bank debt data; and IMF, International Financial Statistics.
Figure 1b.
Figure 1b.

Selected Countries: External Debt Reported in BIS Line I.C., 1997 1/

Citation: IMF Staff Country Reports 2000, 104; 10.5089/9781451807530.002.A004

Sources: Joint BIS-IMF-OECD-World Bank debt data; and IMF, International Financial Statistics.1 BIS line I.C. refers to loans from banks and debt securities issued abroad. Since credits from official sources are not included, these data tend to understate the indebtedness of developing countries especially.
Figure 2.
Figure 2.

Chile: External Debt Stock and Cumulative Current Account Deficit, 1990-99

(USS millions)

Citation: IMF Staff Country Reports 2000, 104; 10.5089/9781451807530.002.A004

Source: Central Bank of Chile.

Competitiveness: real exchange rate, export volume, and current account

152. Various indicators of Chile’s competitiveness improved in 1999, particularly with an improvement in copper prices and a sharp fall in the current account deficit. However, during 1997 and 1998, Chile’s competitiveness might have been an area of investor concern.

153. Chile experienced a sizable appreciation during the 1990s, with the appreciating trend in fact beginning in 1990. From 1990 through autumn 1997, real appreciation has been measured at some 40 to 50 percent (Figure 3).62 It is likely that some, but unlikely that most, of this real appreciation could be attributed to relative productivity gains. Valdes and Delano (1998), for example, estimate that appreciation of roughly 1 percent per year might reflect this effect, only about one-fifth of the average rate of appreciation during the 1990s. Moreover, it is not clear that the real appreciation could be attributed to improving terms of trade; while copper export prices were above the depressed levels of the mid-1980s, their path through the 1990s was if anything a declining one (Figure 4).

154. In 1996-97, two sources of potential concern developed. One was a negative shock to the terms of trade. Copper prices crashed in mid-1996; a partial price recovery was followed by a deeper crash in the second semester of 1997.63 Even if some part of this decline was perceived as temporary, it might still have concerned investors, particularly since no quick recovery was expected.64 Another concern was an acceleration of real appreciation in 1997: roughly one-third of the total 1990-97 appreciation took place in just the first nine months of 1997.

Figure 3.
Figure 3.

Chile: Real Exchange Rate Indices, January 1990-January 2000 1/

(1990=100)

Citation: IMF Staff Country Reports 2000, 104; 10.5089/9781451807530.002.A004

Sources: Central Bank of Chile; and IMF Information Notice System.1/ An increase denotes a real appreciation of the Chilean peso.2/ Annual data for 1990-96, monthly thereafter.
Figure 4.
Figure 4.

Real Copper Price, January 1957-December 1999

(U.S. cents per pound, in constant prices of 1999) 1/

Citation: IMF Staff Country Reports 2000, 104; 10.5089/9781451807530.002.A004

Source: IMF, International Financial Statistics.1/ Copper price deflated by U.S. wholesale price index.

155. Thus, by the time of the height of the Asia crisis in January 1998, several indicators might have led some investors to expect an exchange rate “correction.” However, two considerations make it unclear that Chile’s substantial real appreciation was a serious problem, likely to require a large, or abrupt, nominal exchange rate correction in the near future. First, the level of the real exchange rate in 1990 was not necessarily an equilibrium, and appreciation from that point may have been largely a movement toward equilibrium. Estimates by Cespedes and De Gregorio (1999) of Chile’s equilibrium real exchange rate imply a substantial undervaluation in 1990; not until 1996-97 do they find any overvaluation. Second, the real appreciation that occurred was evidently not governed by the exchange rate band’s limit on nominal depreciation. There was usually room available for continuous depreciation within the band (as opposed to the discrete, band-busting depreciation that draws speculative attacks). For much of the 1990s, exchange market pressure was usually in the direction of appreciation.

156. With price-based indicators not providing a clear signal, we turn to two flow indicators. There was also no obvious case that export volumes were stalling. Certainly, copper export volumes were strong: even at the real exchange rate peak in 1997, they grew by almost 20 percent. On the other hand, non-copper export volumes did slow in 1996 and especially 1997, though even in 1997 they grew by almost 5 percent (perhaps, important lagged effects of real appreciation had not yet materialized).

157. Ultimately, the solvency question comes down to the current account, and in fact the size of Chile’s deficit was a widely-expressed concern for several years. The deficit showed a steep increase in the mid-1990s, mainly accounted for by a jump in import levels that began in 1995 and was sustained until August 1998. For 1996 and 1997, Chile’s actual deficit65 was about 5 percent of GDP; it then peaked at about 7½ percent of GDP during the 12 months ended September 1998, just as several East Asian countries underwent wrenching import contractions. Such deficits were considerably larger than typical among the sample; e.g., Chile’s deficit in 1997 was more than twice the sample median (Figure 5, top panel).66

158. The immediate problem posed by Chile’s current account deficit, perhaps, was not that it was adding rapidly to the debt stock—in principle, solvency was not at issue because there was time to gradually trim the deficit to a level that was more clearly sustainable. Rather, the problem for Chile may have been that a deficit of this size meant that the economy could be more vulnerable to a “sudden stop” of capital inflows.

* * * * *

Figure 5.
Figure 5.

Selected Countries: Current Account Ratios, 1997

Citation: IMF Staff Country Reports 2000, 104; 10.5089/9781451807530.002.A004

Source: IMF, International Financial Statistics.

159. Where does Chile stand now, in early 2000? By nearly every indicator of solvency or competitiveness, Chile’s position is now much improved from just a few years earlier. A substantial if moderate real depreciation, some terms of trade improvement, an acceleration of export volumes—albeit mainly in copper—and especially the steep reduction in the current account deficit, all combine to form a picture that looks safer. Looking ahead, a positive view of solvency is also supported by the expectation of further copper price increases and continued strong growth of output. The only solvency-related indicators not to have improved are the external debt stock ratios, but part of their recent rise reflects asset accumulation. With the current account deficit now much-reduced, growth of Chile’s debt stock ratios is expected to halt.

E. The Question of Liquidity

160. The question of liquidity here centers on the level of official reserves, but there are many ways the adequacy of this can be assessed. A simple start is to seek only a relative answer, making cross-country comparisons, using a variety of scaling variables. We then turn to assessments of liquidity involving the relationship between reserves and potential short-term liabilities. Finally, we discuss various issues regarding the measurement of reserves and short-term debt, applying these questions to the case of Chile.

Scaled measures of reserves

161. Figures 6 and 7 present reserves ratios for Chile and the comparator countries, using four different scaling variables.67

  • Reserves to GDP (at actual exchange rate). GDP is an obvious candidate for scaling reserves data across countries. By this indicator, advanced and developing countries’ levels of reserves are broadly similar on average. Chile’s second-ranked ratio is 22 percent, twice the sample median.

  • Reserves to GDP (usingPPP exchange rate). Switching to PPP-based valuation of GDP alters the picture, as many lower-income countries’ ratios fall sharply.68 The average ratio for the higher-income countries is about twice the average of the lower-income countries. Chile no longer stands out; in fact, its ratio drops fairly close to the sample median.

  • Reserves to Broad Money. This indicator is perhaps most relevant in a fixed exchange rate context—so it now may be less relevant for Chile than previously. With a ratio of 55 percent, Chile holds 3rd position, again at roughly twice the median value of this ratio. By this indicator, higher-income countries tend to have smaller reserves than the lower-income ones—indeed, Australia, Canada, and New Zealand have some of the lowest ratios in the sample. It is striking that this indicator of reserve adequacy is apparently irrelevant to these highly-perceived (e.g., in terms of credit ratings or bond spreads) economies. Clearly, a very low value of this ratio is not perceived as necessarily being a problem, at least for advanced economies with flexible exchange rate regimes.

  • Reserves to Imports. This traditional indicator still receives attention, though its conceptual appeal has waned as capital accounts have been opened. (Another problem is that a high ratio might reflect mainly a low level of trade openness which might be associated with greater vulnerability.) On average, advanced economies’ import cover is lower than in lower-income countries. Chile’s 3rd place ratio of 90 percent is more than twice the sample median.

Figure 6.
Figure 6.

Selected Countries: Reserves in Comparison to GDP, 1997

Citation: IMF Staff Country Reports 2000, 104; 10.5089/9781451807530.002.A004

Source: IMF, International Financial Statistics.
Figure 7.
Figure 7.

Selected Countries: Other Reserves Ratios, 1997

Citation: IMF Staff Country Reports 2000, 104; 10.5089/9781451807530.002.A004

Source: IMF, International Financial Statistics.

162. Although these indicators of reserves can give a quite different picture for any one country, for Chile the relative position looks very strong for three of the indicators, and for the other Chile is centrally placed.

Liquidity as the relation between reserves and potential short-term liabilities

163. This type of indicator is fast becoming a standard, based on its success in predicting currency crises. The most basic form is a comparison between official reserves and the stock of short-term debt, typically expressed as a ratio. Such an indicator can have both a relative and an absolute interpretation. In relative terms, it has been shown empirically that higher values are less frequently associated with crises (though such conclusions have been drawn from samples of emerging markets, and the relationship may not apply in the same way to higher-income countries). A conceptually-based, absolute interpretation is also possible, in which the ratio is considered to represent a simple “stress simulation” for a hypothetical shock involving the complete cutoff of external inflows for a period of one year. In this case, a ratio of 100 percent can become a benchmark, one with some intuitive appeal, but also some problems.69

164. In considering such indicators, measurement problems become acute, since for many countries there is substantial uncertainty about the true level, or appropriate definition, of short-term debt. Given the size of the sample used here, the only practical option is to use the publicly-available Joint BIS-IMF-OECD-World Bank (“BIS”) data for short-term debt. (Since such data are not available for many of the sample’s higher-income countries, the sample size falls to 19.) Again, these debt data are not beyond question, and their use here is motivated by convenience and impartiality only. As regards reserves data, these are taken from the IFS. (Specific issues in the measurement of short-term debt and reserves are discussed below, emphasizing the Chilean case.)

165. Ratio of reserves to short-term debt (Figure 8, top panel). Poland and Israel are outliers at the high end, with ratios about three times that of Chile. Still, at about 150 percent, Chile’s ratio exceeds the benchmark of 100 percent, and it is the 7* highest of the 19 countries with data available.70 However, the absolute size of such ratios has no direct, intuitive interpretation, so it is useful to consider a variant with more familiar units.

166. Reserves minus short-term debt, as a percent of GDP (Figure 8, bottom panel). This formulation, which scales the gap between reserves and short-term debt by the size of the economy, changes the picture greatly for some countries: e.g., Poland and Israel no longer appear as strong outliers, while Singapore becomes one.71 Chile’s rank, however, is almost unchanged, at 6th out of 19 countries. The value for Chile is positive, by 7.3 percent of GDP.

Some Chile-specific modifications

167. Since Chile passes the simple benchmark of having reserves in excess of short-term debt, the next step is “raise the bar.”

168. One possibility is to compare reserves to the sum of short-term debt and the current account deficit over the coming year. The problem with this suggestion is judging the appropriate level of the deficit to be included. In Chile’s case, however, the question may be moot, since the excess of reserves over short-term debt is considerably larger than any deficit on current account Chile is likely to experience soon. Thus the latest-available BIS data, for-end 1999, indicate that reserves exceed short-term debt by more than JO percent of GDP.

Figure 8.
Figure 8.

Selected Countries: Reserves in Comparison to Short-term Debt, 1997 1/

Citation: IMF Staff Country Reports 2000, 104; 10.5089/9781451807530.002.A004

Sources: Joint BIS-IMF-OECD-World Bank debt data; and IMF, International Financial Statistics.1/ For Czech Republic and Slovenia, BIS short-term debt data are incomplete, so the ratios shown here are biased upward, by an unknown amount.

169. For a more challenging stress simulation, one may consider not only a year-long cutoff of external finance, but also shocks that at the same time might abruptly increase the size of the current account deficit to be financed. Consider three events which might afflict Chile:

  • A decline in copper prices. At around 40 percent of exports, and 10 percent of GDP, copper exports are clearly significant to the Chilean economy. Moreover, copper prices are highly volatile and it is not unusual for the real price of copper to change by a factor of two—in either direction—in a year or two. Currently, however, downside risk is limited by the already-weak state of copper prices, which touched a 40-year low in early 1999, and have subsequently recovered only slightly. Moreover, copper prices in 1999 seem to have been approaching a natural floor, in that prices were reportedly close to a level at which less efficient producers would be unable to cover variable costs (indeed the year was marked by many mine closures outside Chile). Nevertheless, to take an extreme scenario, consider a hypothetical decline in the LME copper price to 50 U.S. cents per pound, or about 30 percent below the 1999 average. The direct, static impact on the current account would be on the order of 2½ percent of GDP. Surely, this figure is an overestimate;72 nevertheless, it represents only a fraction of the excess of reserves over short-term debt.

  • Higher oil import prices. Another potential negative shock would be an increase in oil import prices, since Chile produces only a small share of its oil consumption. Imports of crude oil tend to be no more 1½ percent of GDP, while imports of other combustible fuels and lubricants are typically another 1 percent. As a rather extreme example, a rise in spot oil prices to US$35 per barrel would represent an approximate doubling of their 1999 level (which was broadly similar to the average level of the 1990s). If maintained for a year, the static (i.e., overestimated) impact on the current account would be roughly 2½ percent of GDP.

  • Increased foreign interest rates. Consider, for example, an increase of 300 basis points in the U.S. dollar interest rates to which Chile’s floating rate debt is tied. With total external debt on the order of 50 percent of GDP, and roughly two-thirds of that debt having a floating interest rate,73 the direct impact effect on the current account of higher interest payments would be 1 percent of GDP. Again, such an outflow is small in relation to the excess of official reserves over short-term debt. (This calculation is conservative in that it does not take into account increased receipts on Chile’s foreign assets.)

170. All these possibilities represent negative shocks large enough to be unlikely in any given year. Yet even when their effects on the current account are calculated in a clearly overstated manner, none are large enough to alter the perception of strength that comes from a simple comparison of Chile’s reserves and short-term debt levels.

171. To further check the robustness of Chile’s position, we next take a closer look at the data used above to represent reserves and short-term debt.

Taking account of reserves’ usability

172. As has become better appreciated, the concept of reserves is not self-explanatory, and for many countries the nature of reserve holdings, and their degree of usefulness, traditionally has not been transparent. However, the development of the SDDS “reserves template” has begun to counter this problem.

173. Chile began publishing reserves data on this basis with figures for end-April 2000. Overall, this new information suggests that most of Chile’s reported official reserves would in fact be available for policy purposes, even in a worst-case scenario. For example, the sum of net pre-determined and contingent short-term foreign exchange flows is negative, but it is less than 5 percent of the stock of gross reserves. Also of interest is the amount shown for liabilities denominated in foreign currency but settled by other means (such as the domestic currency): for Chile, this was reported to be a bit less than 10 percent of total reserves in April 2000.

174. For many countries, another point of interest in analyzing reserves data is the amount of foreign exchange deposited by resident financial institutions at the central bank, since some of these funds might be withdrawn quickly in a crisis. The level of such deposits at Chile’s central bank is not separately identified in the published data, but it is understood that it recently has represented only 1 to 2 percent of gross reserves.

Measures of Chile’s short-term external debt

175. As noted, the BIS data considered above are not the last word on short-term debt measurement, for Chile or any other country in the sample. A detailed study of alternative debt measures in each country would be impractical here, but a few remarks on Chile are worthwhile. Use of the officially-published short-term debt statistics would improve the view of Chile’s liquidity position, since these data indicate stocks smaller than BIS data. The Chilean authorities have studied the differences between the numbers, identifying a number of methodological differences that could account for the gap. In 1999 the gap narrowed considerably when the authorities shifted to a residual maturity basis. Of the remaining difference between Chilean and BIS data, one difference is that the former do not yet include trade credits.74 Another is that the BIS figures include foreign currency loans extended by foreign-owned banks which are resident in Chile; the Chilean statistics do not consider this as external debt.75 The BIS recently began identifying the amount of such loans in its published data; for end-1999, they constituted more than one-third of the BIS-reported short-term debt.

F. Other Factors Affecting External Vulnerability of Chile

176. This section briefly notes a range of factors, including policies, affecting Chile’s degree of vulnerability. What these factors have in common is their qualitative or judgmental nature, which does not lend itself to direct cross-country comparison—thus the focus on Chile.

177. In general, Chilean policies would seem to score well in terms of limiting external vulnerability. Some policies are of course reflected in the quantitative indicators examined in the previous sections, but mention could also be made of the following:

  • Apparent avoidance of tax or other regulatory bias in favor of short-term external borrowing. Such bias is a factor often cited to explain the origins of the 1997-98 crises in East Asia. On the contrary, Chilean policy until very recently involved an explicit bias against short-term capital inflows.

  • Strength of the banking system, including banking regulation. It often happens that currency crises and banking crises occur together. A comprehensive examination of these issues cannot be included here, but Chile’s banking system is widely considered to be strong, exceptionally so among emerging markets. For example, in early 2000 Moody’s ranked the financial strength of Chile’s banks 15th of 75 countries, similar to those of Australia, and exceeded only by Singapore among emerging markets. A further indicator of strength is the way Chilean banks have weathered the recent recession and the various shocks of 1997-99, including a substantial currency depreciation since late 1997. One factor in this performance is likely the relatively low external indebtedness of Chile’s banks: although most of Chile’s debt is private, financial institutions account for only about 15 percent of medium- and long-term external debt, for example. Banks’ exposure to foreign currency risk is a focus of regulation. The short-term foreign assets of commercial banks have exceeded their short-term foreign liabilities since 1997, with the difference at end-1999 being more than 5 percent of GDP. Recent regulatory initiatives seek to facilitate banks’ analysis of risk they incur in lending to Chilean companies.

  • Factors limiting the financial vulnerability arising from the corporate sector. Chile’s corporate sector has avoided the very high debt-equity ratios seen in East Asia and often cited as contributing to the severity of the 1997-98 crisis. Regulations also play a role, for example the requirement of a minimum credit rating for firms issuing foreign currency bonds.

  • Exchange rate flexibility, and transparency of the exchange rate regime. Since September 1999, new detail in the reporting of changes in official reserves allows a clearer picture of central bank intervention (or its absence). This step was recently complemented with data following the SDDS reserves template. Arguably, the new exchange rate regime itself may help to prevent crises.

  • Policy focus on monitoring the current account deficit. The outlook for solvency is supported by an apparent consensus among many policymakers in Chile that macro policy should closely monitor the current account deficit (i.e., aside from the real exchange rate), thinking in terms of some maximum safe, sustainable current account deficit, and taking action when this level is exceeded.

178. From the standpoint of limiting external vulnerability, it is difficult to find fault with Chile’s policy regime. However, certain areas exist in which there is room for deepening the set of information that is available for assessing vulnerability. Such information, and the better private decision making it facilitates, can in principle reduce external vulnerability. The following areas can be mentioned:

  • Data on external assets of the private sector. Since these are believed to be large, they are of interest, and could be especially useful in interpreting the significance of Chile’s also sizable external debt. Not only the size but also the liquidity of these assets could be relevant for some questions. Publication of complete International Investment Position data would be a significant step.

  • Data on foreign currency exposure by sector, or average levels of exposure. Outside the financial sector, the exposure of the rest of the private sector is unclear. It would be useful to know whether the same firms or sectors that have recently been accumulating external debt at a fast pace are also the ones which have been building up their external assets, and whether such debtors are foreign exchange earners.

  • A fully-developed medium-term fiscal framework would make clearer to markets the role the public sector sees for itself in maintaining Chile’s overall solvency, dealing with terms of trade shocks, and influencing the national saving rate. The recent initiative to consider the fiscal stance in terms of structural balance, and the suggestion to aim at a surplus on a sustained basis, could provide a useful signal.

  • Expanded coverage of external debt statistics. Data for short-term debt that included trade credits could sharpen the analysis of liquidity (though the basic picture might not be much affected, given the size of Chile’s international reserves). As regards total external debt, more complete coverage of the debt (and assets) of the military sector would be useful in analyzing the solvency of the Chilean economy.

G. Conclusions

179. Assessment of external vulnerability remains an art. After examining numerous indicators, conclusions such as the following must contain an element of judgment as well as generalization:

  • Chile aside, one finding of this study is the pattern of differences between higher- and lower-income countries, with the former tending to look stronger by most indicators. Certainly, the higher-income countries on average have higher credit ratings and lower bond spreads. Moreover, the higher-income countries also appeared less affected by the Russia shock in terms of the measured responses of bond spreads, exchange market pressure, and forecasts of their exchange rates. The higher-income countries also tend to have smaller current account deficits. However, it cannot be said that higher-income countries hold larger official reserves; for several reserves indicators, the opposite pattern holds. (Unfortunately, comparisons related to external debt stocks are frustrated by the lack of comparable data.)

  • Relative to other emerging markets, the strength of Chile’s external position is confirmed—even among the class of “better” emerging markets considered here.

  • Relative to the advanced/higher-income economies considered here, Chile remains somewhat but not too far behind, according to such summary indicators as bond spreads and private credit ratings. In terms of the relative degree of vulnerability revealed by the 1998 Russian crisis, it can be said that the crisis did not show Chile to be completely out of this league. For several of the response indicators examined, Chile looks more like a higher-income country.

  • As regards liquidity, Chile’s relatively high level of official reserves, considered against its relatively moderate level of short-term debt (both measured as shares of GDP) suggest a position of strength. Chile’s excess of reserves over (BIS-measured) short-term debt is on the order of 8 percent of GDP, which dwarfs the additional short-term financing needs that might arise from various shocks to the current account.

  • Comparisons to other countries aside, Chile’s external position in early 2000 looks stronger than in several years. The greatest weakness perceived in recent years was probably the size of the current account deficit, but this deficit nearly disappeared in 1999. The recent moderate improvement in the terms of trade is another positive factor, and some further recovery is likely. Also among Chile’s greatest strengths are the health of its banking system, the low level of public sector debt, and the coverage of short-term debt by reserves.

  • Chile’s least favorable indicator might be the total stock of external debt. Although the recent rise in external debt ratios is not likely to continue, such ratios have moved outside the low ranges, and the level of debt bears monitoring. Still, this debt should be considered against Chile’s foreign assets, and the more relevant issue may not be solvency but the management of the foreign exchange exposure of the private sector.

References

  • Berg, Andrew, and Catherine Patillo (1998) “Are Currency Crises Predictable? A Test,” Working Paper 98/154, International Monetary Fund.

    • Search Google Scholar
    • Export Citation

  • Bussiere, Matthieu and Christian Mulder (1999) “External Vulnerability in Emerging Market Economies: How High Liquidity Can Offset Weak Fundamentals and the Effects of Contagion”, Working Paper 99/88, International Monetary Fund.

    • Search Google Scholar
    • Export Citation

  • Cespedes, Luis Felipe, and Jose De Gregorio (1999) “Tipo de Cambio Real, Desalineamiento y Devaluaciones: Teoriay Evidenciapara Chile,” mimeo.

    • Search Google Scholar
    • Export Citation

  • De Gregorio, Jose, and Rodrigo Valdes (1999) “Crisis Transmission: Evidence from the Debt, Tequila, and Asian Flu Crises,” mimeo.

  • Eichengreen, Barry and Ashoka Mody (1998) “What Explains Changing Spreads On Emerging-Market Debt: Fundamentals or Market Sentiment?” NBER, Working Paper 6408.

    • Search Google Scholar
    • Export Citation

  • Frankel, Jeffrey A., and Andrew K. Rose (1996) “Currency Crashes in Emerging Markets: An Empirical Treatment,” Journal of International Economics, Vol. 41, pp.351- 366.

    • Search Google Scholar
    • Export Citation

  • Kaminsky, Graciela, Lizondo, Saul, and Carmen Reinhart (1998) “Leading Indicators of Currency Crises,” IMF Staff Papers, Vol. 45, No. 1.

    • Search Google Scholar
    • Export Citation

  • Guajardo, Jaime and Guillermo Le Fort (1999) “Cuenta Corriente and Desvios Transitorios en Terminos de Intercambioy Volumenes de Exportaciones: Chile 1985-1999” Central Bank of Chile, Working Paper No. 49.

    • Search Google Scholar
    • Export Citation

  • Milesi-Ferretti, Gian Maria (1998) “External Sustainability: A Study of Chile,” in Chile:Selected Issues, paper prepared for the International Monetary Fund Executive Board.

    • Search Google Scholar
    • Export Citation

  • Monfort, Brieuc and Christian Mulder (2000) “Should Capital Requirements for Banks Depend on Sovereign Ratings by Credit Rating Institutions?” Working Paper, International Monetary Fund.

    • Search Google Scholar
    • Export Citation

  • Sachs, Jeffrey, Aaron Tornell and Andres Velasco (1996) “Financial Crises in Emerging Markets: The Lessons from 1995,” Brookings Papers on Economic Activity: I, Brookings Institution, pp.147- 215.

    • Search Google Scholar
    • Export Citation

  • Tanner, Evan (1999) “Exchange Market Pressure and Monetary Policy: Asia and Latin America in the 1990s,” Working Paper 99/114, International Monetary Fund.

    • Search Google Scholar
    • Export Citation

  • Valdes, Rodrigo and Valentin Delano (1998) “Productividady Tipo de Cambio Real en Chile” Central Bank of Chile, Working Paper 38.

46

Prepared by Steven Phillips (WHD). The author is grateful for helpful comments from Martin Kaufman, Saul Lizondo, Norman Loayza, Christian Mulder, Francisco Nadal-De Simone, Gonzalo Sanhueza, Jeronimo Zettelmeyer, and seminar participants at the Banco Central de Chile. He is indebted to Francisco Nadal-De Simone for advice on the analysis of copper prices, and to Patricia Gillett for research assistance. The author is responsible for any errors.

47

An extensive literature has investigated the importance of many variables in either causing, or helping to predict, external crises. See, for example, Berg and Patillo (1998); Bussiere and Mulder (1999); and Kaminsky, Lizondo, and Reinhart (1998).

48

Per capita income on a PPP-adjusted basis, averaged during 1995-99.

49

Peru was given this rating only in December 1997; previously, Peru was unrated.

50

The actual levels of these forecasts are not presented.

51

The choice of a window as long as six months is motivated by the fact that such ratings are revised/confirmed not continuously but rather only occasionally.

52

On whether ratings agencies “see through the cycle,” see Monfort and Mulder, 2000.

53

A problem with very short horizons is that daily data for spreads are sometimes missing in the periods of greatest stress.

54

The remaining maturity of the Banco del Estado bond was shorter than that on most other bonds analyzed in Table 2.

55

The Bussiere and Mulder sample, in contrast to the one used here, excludes advanced economies and includes lower-rated emerging markets. The purpose of their analysis is not merely to observe EMP but to analyze its determinants.

56

Scaling reserves losses by the previous period’s reserve money stock gives meaningful units, and this allows the weighting of EMP index’s components to be uniform across countries. The weighting of the index—2:1, with the greater weight on currency depreciation—was chosen to make the variance of its two components roughly similar for most countries during the period before the crisis. The basic results discussed here are not sensitive to changes in this weighting (e.g., alternative weights of 1:1 and 4:1).

57

Another problem is the short-term noise in reserves data induced by lumpy new external borrowing and amortization of debt.

58

Over horizons of a month or two, the role of trend depreciation against the U.S. dollar is minimal, and it can be safely disregarded. The exception is Turkey, where an adjustment is made.

59

Each successive survey of 12-month ahead forecasts refers to the spot rate expected for a date that is one month further into the future, but this slight time-shift can be ignored for most currencies. Again, the exception is Turkey, the only country in the sample with high inflation, and accordingly with a steep depreciating trend to its exchange rate. As it happens, Turkey comes off well in Table 4, even without adjusting for expected trend depreciation.

60

For an earlier, more comprehensive treatment of solvency issues in Chile, see Milesi-Ferretti (1998).

61

Data for 1997 are used here, and elsewhere in this chapter, because they are consistently available for all countries in the sample. Such data also may also be useful as a “pre-Russia crisis” base (though for many variables considered, data a year or two earlier or later would show largely similar cross-country patterns).

62

That is, 41 percent according to the IMF’s Information Notice System (INS), and about 50 percent according to the series from Chile’s central bank.

63

Prices then continued to fall, bottoming in Spring 1999 at about half the mid-1997 value.

64

Since early 1998, forward market prices have pointed to some copper price recovery, but not to a fast one. For example, during the Russia crisis period of August-October 1998, when LME copper prices averaged about 72 U.S. cents per pound, prices for copper 2¼ years ahead averaged only 76 U.S. cents per pound. It can also be noted that real copper prices in 1998-99 were much below their averages of the previous 10,25, or 40 years, hinting at room for mean-reversion. If such prices do have a mean-reverting component, however, it is slow enough to make it difficult to reject the hypothesis of a unit root (even using the more powerful tests that allow for breaks in the series) in the 1957-99 period.

65

Guajardo and Le Fort (1999) construct an alternative indicator, the trend current account, abstracting from shocks to the terms of trade and the volumes of certain exports. From this perspective, the peak current account deficit occurred in 1997 rather than 1998.

66

Note that on average the higher-income economies have smaller current account deficits than the lower-income ones, scaling by GDP converted to U.S. dollar terms at actual exchange rates. Using PPP-based GDP (Figure 5, bottom panel) considerably reduces the apparent size of the lower-income countries’ deficits.

67

Based on ITS reserves data excluding gold. For most countries in the sample, these ratios would be only slightly higher if gold were included. Reserves ratios for Philippines, South Africa, and Uruguay would be most affected, but their sample rank would change little.

68

This pattern is consistent with the usual pattern of exchange rate undervaluation (in the PPP-based sense) found in developing countries.

69

One problem is that the choice of a one-year duration for the financing shock is arbitrary. A more fundamental question, where most short-term debt is owed by the private sector, is whether and how the reserves held by the public sector would come into play. (For example, short-term creditors may not be confident that the parties to which they have lent would have the domestic currency resources to purchase, at prevailing exchange rates, the foreign currency needed to amortize their debt.)

70

Regarding the rule of thumb proposed by Bussiere and Mulder (1999), which adjusts the 100 percent benchmark according to the level of the current account deficit and the extent of recent appreciation of the real exchange rate, Chile would have been near the borderline of a comfortable situation in 1997, but would easily pass this benchmark today.

71

The ratio of reserves to short-term debt can be quite sensitive to low values of its denominator. According to BIS data, both Poland and Israel have unusually low values of short-term debt (expressed in relation to GDP, for example, about 3 percent). Chile’s ratio of short-term debt to GDP is close to the median, about 12 percent of GDP.

72

For example, because the large degree of foreign ownership of Chile’s copper sector, implying that diminished profit outflows in the current account would somewhat offset the loss of export revenue. Also, such a decline in export prices would initiate a range of responses tending to trim import demand.

73

At end-1998, for example, floating rate debt represented 64 percent of the total.

74

In principle, BIS data for short-term debt captures some, but not all, trade credits.

75

The Chilean treatment is consistent with the IMF Balance of Payments Manual.

Same Series

Other IMF Content

Other Publishers

Cover IMF Staff Country Reports
Issue 104
Volume/Issue:
Volume 2000: Issue 104
Publisher:
International Monetary Fund
ISBN:
9781451807530
ISSN:
1934-7685
  • View in gallery

    Selected Countries: External Debt (World Development Indicators), 1997

  • View in gallery

    Selected Countries: External Debt Reported in BIS Line I.C., 1997 1/

  • View in gallery

    Chile: External Debt Stock and Cumulative Current Account Deficit, 1990-99

    (USS millions)

  • View in gallery

    Chile: Real Exchange Rate Indices, January 1990-January 2000 1/

    (1990=100)

  • View in gallery

    Real Copper Price, January 1957-December 1999

    (U.S. cents per pound, in constant prices of 1999) 1/

  • View in gallery

    Selected Countries: Current Account Ratios, 1997

  • View in gallery

    Selected Countries: Reserves in Comparison to GDP, 1997

  • View in gallery

    Selected Countries: Other Reserves Ratios, 1997

  • View in gallery

    Selected Countries: Reserves in Comparison to Short-term Debt, 1997 1/

Selected Countries: External Debt (World Development Indicators), 1997

Selected Countries: External Debt Reported in BIS Line I.C., 1997 1/

Chile: External Debt Stock and Cumulative Current Account Deficit, 1990-99

(USS millions)

Chile: Real Exchange Rate Indices, January 1990-January 2000 1/

(1990=100)

Real Copper Price, January 1957-December 1999

(U.S. cents per pound, in constant prices of 1999) 1/

Selected Countries: Current Account Ratios, 1997

Selected Countries: Reserves in Comparison to GDP, 1997

Selected Countries: Other Reserves Ratios, 1997

Selected Countries: Reserves in Comparison to Short-term Debt, 1997 1/