Dollarization & sovereign risk in latin america

SOVEREIGN RISK AND DOLLARIZATION: THE CASE OF ECUADOR
SOVEREIGN RISK AND DOLLARIZATION: THE CASE OF ECUADOR
Abstract
Through policy decision of governments and through other unofficial means, many countries have moved away from their local currency to seek protection against inflation provided by a hard currency, the dollar. Among the promises of dollarization is the reduction of sovereign risk associated with developing country debts and subsequently an increase in the rate of economic growth as the cost of financing economic growth declines. This however may not occur as the loss of the policy tool could lead to inflexibility as the economy is not able to respond to shocks, resulting in an increase in sovereign risk. This paper explores the effect of dollarization on sovereign risk in Ecuador and concludes that dollarization does not decrease sovereign risk. JEL Classification Codes: E44; F31; G19; O54. Keywords: Sovereign Risk; Dollarization; Ecuador. Introduction:

Throughout the 20th century, Panama remained the only country in Latin America to use the U.S. dollar as its official currency. However, by the beginning of the twenty-first century, the U.S. dollar status in Panama will no longer be a peculiarity in Latin America. Following the two oil shocks of the 1970’s, many Latin American economies experienced financial distress that necessitated help from the international financial institutions. In the early nineteen eighties, the U.S. Treasury lead the rescue effort by helping those countries issue the so-called Brady bonds. However, during that decade, Latin America enjoyed neither the level of recovery that was expected, nor the economic prosperity experienced by the Asian Tigers (Taiwan, Hong-Kong, Singapore, South Korea). Thus, in policy circles, the eighties would be called the lost decade for Latin America. In the early 1990s, motivated by the relative economic success of the Asian Tigers, and convinced by policy advices coming from the U.S. Treasury, the World Bank and the IMF (the so-called Washington Consensus), Latin America followed the lead of Chile and embraced neo-liberal policies. This policy package included privatization of public enterprises, removal of trade barriers, and capital account liberalization. The brisk increased in capital flow caused by the liberalization of financial markets and the potential for a “sudden-stop” constituted an element of financial fragility that would manifest itself in the mid to late nineties. One of the common characteristic of the crisis experienced in the second half of the nineties in Latin America is the move of the public away from their local currency in order to seek the protection against inflation provided by a hard currency, the dollar. However, this process known as “de-facto dollarization” posed a challenge to macroeconomic policy makers throughout Latin America. As the demand for the local currency decreased relative to supply, the risk of its devaluation increased and monetary policy became less effective. By the end of the 1990s many analyst and policy makers suggested that eliminating their own currency and adopting the U.S. dollar as legal tender would remedy Latin America’s monetary instability and force a solution to the chronic fiscal imbalance. Key among the promises of dollarization is the reduction of the sovereign risk associated with developing country debts. That in turn would lead to decreased cost of financing economic growth. In this paper we analyze the experience of Ecuador, and test the hypothesis that official dollarization is not a determining factor in the reduction of sovereign risk. If this is true, the implications adds another perspective to the debate on the effects of dollarization that could be profound given the consensus from the few studies of this relationship that there is a link between sovereign risk and economic performance. As pointed out by Nogues and Grandes (2001), the sovereign risk is important because of the crucial role it plays in determining financial costs and real business cycles. Finding out the determinants of sovereign risk and financial cost could help to determine future policy prescriptions especially as they relate to exchange rates and dollarization in countries like Ecuador. Furthermore, research in this area is relatively recent with only a few economists emphasizing the relevance of country risk on the dynamics of the economy. In the second section we review briefly the literature on sovereign risk and dollarization in order to provide justification for our work. In the third section we analyze the determinants of sovereign risk. In the fourth section we present the result of the econometric analysis. The conclusion and policy implications are presented in the fifth section. II - Literature Review

In recent years research on the effect of dollarization has been conducted by Bogetic (2000), Alesina and Barro (2001), Beckerman (2001), Edwards (2001), Willett (2001), Edwards and Magendzo (2003), Levy-Yeyati and Sturzenegger (2003), Jacome H (2004), and Ize and Levy-Yeyati (2005). Other papers have looked specifically at dollarization and country risk. In a recent summary paper by Nogues and Grandes (2001), the papers on country risk were divided into three groups. The first group that includes Avila (1998), Grandes (1999), and Rodríguez (1999) found a significant (negative) correlation between the large macroeconomic aggregates and the risk premium. In the second group, Rubinstein (1999) argued that once exchange risk disappears with dollarization, there will be a lower country risk with higher growth and employment rate. In the third group, the relationship between capital flows, contagion effects and the incidence on sovereign risk is analyzed (Calvo and Reinhart, 1996; and Calvo, 1999). Other researchers have discussed dollarization in terms of its link to country and devaluation risk. Some authors concluded that in the absence of balance sheet effects, a full dollarization of an economy increases its country risk. However, when balance sheet effects are present, the full dollarization could reduce country risk. For them dollarization ultimately is a fiscal issue. In the paper by Edwards and Magendzo (2001), although countries that dollarized were found to have lower inflation rates than countries with a national currency, the dollarized countries were found to have lower growth rates and higher volatility. Jacome H’s (2004) research focused on de facto dollarization. Although Ecuador officially adopted the U.S. dollar as its currency in March 2000, de facto dollarization was significant before 2000. De facto or unofficial dollarization is the holding by residents of assets and liabilities denominated in a foreign currency, in this case the U.S dollar. Ize and Levy-Yeyati (2005: 4) noted that for example, in the first quarter of 1998, one third of on-shore deposits were in U.S. dollars, and that by 1999 the share had increased to close to 50%. Jacome H (2004) research concludes that unofficial dollarization in conjunction with other factors (institutional weakness and rigidities in public finances) amplified the financial crisis in Ecuador in the 1990s. Jacome H (2004) found that in Ecuador in the 1990s institutional factors restricted the government’s ability to prevent the escalation of the banking crisis. He also found that public finance rigidities during dollarization limited the government’s capacity to correct existing imbalances and the deteriorating fiscal stance associated with the costs of the unofficial dollarization. In the end Jacome H (2004) claimed that dollarization led to increased demand for foreign currency and accelerated the currency crisis because it reduced the effectiveness of financial safety nets. A summary paper on Ecuador’s dollarization experience that was prepared by Abrego, Flores, Pivovarsky, and Rother (2006) indicated that although six years are not long enough to judge whether dollarization was right for the Ecuadorian economy, the period can still be used to assess how well dollarization is serving the economy. Using a regression of the Ecuador sovereign spreads on 10-Yr Treasury and several dummy variables, including one for dollarization, they found that dollarization do not have a significant effect on spreads. Like most of the other research, Abrego, Flores, Pivovarsky, and Rother (2006) put forward the idea of a trade-off between lower and stable interest rate versus the economy’s inflexibility and inability to adequately respond to shocks after dollarization. However, their analysis did not include any fiscal or domestic macroeconomic variable as a determinant of Ecuador’s sovereign spread. III - The Determinants of Sovereign Risk

Edwards (1986) assumes risk neutrality of lenders and describe the rate of return on sovereign debt in an economy with high capital mobility as an arbitrage condition: From equation (1) p is the probability of default, i* as the risk free rate of return on U.S. Treasury bill, and s as the sovereign risk premium. Thus: Theoretically, the risk premium approaches infinity as the probability of default tends to one. In practice, for a given i*, the risk premium will vary with the probability of default. However, the probability of default never gets close to one, because the borrower would be denied access to the bond market as an increasing probability of default signals a financial crisis. A logistic function given in equation (3) can be used to describe p: where Xi are the determinants of the sovereign risk premium and βi are the respective coefficients. Combining (2) and (3) and applying natural logs, the equation becomes: As in Nogues and Grandes (2001), we follow the practice of the rating agencies, in the measure of country risk and in the determination of the relevant Xi variables. Ecuador’s country risk is measured by the average EMBI spread of its government bonds, published by JP Morgan: “Spread measures the risk premium over US Treasury bonds”. For the determining factors, we use four classes of variables: (1) Macroeconomic fundamentals related to the possibilities of debt repayment; (2) Contagion variables to capture the effect of instability in other financial markets on Ecuador (3) Political variables to capture the effect of unrest on financial market; and (4) Structural reform variables. By including macroeconomic variables we are extending the model used in a recent research by Abrego, Flores, Pivovarsky, and Rother (2006). The macro variable included in our analysis is the public external debt to GDP ratio. We use the public external debt to GDP ratio rather than the Debt Service to Export ratio, because it is less influenced by debt restructuring and other arrangements that may change the repayment profile without changing the amount of debt outstanding. The contagion variable that we use is the EMBI index for non-Latin countries. The contagion effect of non-Latin American emerging countries is included because of the importance of headline risk in emerging markets. We also use the 10-year U.S. Treasury rate to capture the impact of developments in the U.S. market on Ecuador’s sovereign risk. We did not the use the 30-year rate because its issuance was discontinued during the late nineties to the early 2000s due to the U.S. fiscal surplus. A dummy variable is used to control for major political unrest in Ecuador during the period of analysis. Specifically, we try to capture the effect of three presidential overthrows (Buccaram, Mahuad, and Gutierez) during the past decade on Ecuador’s sovereign risk. Another dummy variable is used to capture the significance of dollarization as a structural reform. With regards to the Public External Debt to GDP ratio variable, the expected sign of the coefficient is positive as an increase in the size of the debt increases the probability of default hence sovereign risk increases. Because investors have a tendency to project the developments in a particular emerging market to all emerging markets, whether it’s in Latin-America or not, we expect the sign of the contagion variable (EMBINONLAT) to be positive. There is some ambiguity about the expected sign of the 10-year U.S. Treasury variable. It should be positive if when U.S. interest rates are low, investors increase their demand for emerging market debt and depress their yield (substitution effect). However, that sign may be negative if the prospect of financial crises make investors ignore increasing yields offered by emerging market debt and look for the safety offered by the low yield U.S. Treasuries (flight to quality). The magnifying impact of political crisis on financial risk is well documented and the sign of the political unrest dummy should be positive. The consequence of dollarization (the variable of interest) on sovereign risk is not straightforward. From an orthodox point of view, the official dollarization of a small economy with chronic macroeconomic imbalance like Ecuador should improve the financial environment. It is equivalent to the elimination of monetary policy viewed as a source of instability, and that should improve fiscal management. However, the economy’s inflexibility and the loss of a policy tool that could be used in responding to shocks may cause an increase in sovereign risk. Consequently, the expected sign of the dollarization dummy is ambiguous. The model used in the empirical analysis is summarized below with the signs in parentheses denoting the expected effect of the variable on sovereign risk. EMBIECU = ƒ [EMBINONLAT(+),UST10(+/-), DEBT-GDP(-), POL(+), EMBINONLAT is average spread for non-Latin countries, UST10 is the rate of interest for 10-year U.S. Treasury bond. DEBT-GDP is the Public External Debt over GDP ratio, POL is a dummy for political crisis, DOL is a dummy that identifies the official dollarization period, IV – Econometric Estimation

In order to estimate Equation 5, we take the natural log of the variables on both sides of the equation, with the exception of the Debt-GDP ratio and the dummies, and we run a linear regression. Before we present the results for the estimated equation, we provide some preliminary analysis of the dataset. Chart 1, below, presents EMBI spread in level form (not natural log) for Ecuador, Argentina, Brazil, Mexico, and a Non-Latin Countries average during the 1997-2006 period. EMBI Spread for Ecu, Bra, Arg, Mex, and Non-Latin Average
EMBI
Spread
As with any empirical analysis, the test to determine whether the data were normally distributed was applied to the variables (in the form that they are used in the regression). The results are reported in Table 1 in which all five tests including the Shapiro Wilk, Anderson Darling, and the three D′Agostino tests indicated that the normality assumption is valid for all the measures at the 5% level. These results support the use of parametric statistics to investigate the research hypothesis. The plots in Charts 2-7 also bear out the suitability of the data as Charts 2 & 3 confirm that the normality assumption holds for the dependent variability and Charts 4-7 indicate that the assumption of homoscedasticity or equal variance is also relevant. Normality Tests

Test Name
Test Value Probability Level Reject H0 at Alpha = 5%?
Normal Probability Plot of Residuals of EMBIECU Table 2 shows the mean and standard deviation of the measures excluding the dummy variables. The standard deviations confirmed the high variability of all measures. This points to the crucial importance of contingency planning in the financial sector in Ecuador, and could also help to explain why the move towards dollarization occurred so rapidly: there was no emergency plan to avoid the rapid deterioration of the Ecuadorian Sucre. Interestingly, all measures tested (except POL and DOL) failed to reject the null hypothesis for equality of means when the Levene test was applied to the two data subsets. We can therefore conclude that dollarization did not affect the integrity of the sample. Descriptive Statistics
Standard
Variable Count Mean Deviation
DEBT-GDP 120 0.5221247 0.2091631 0.25 1 LEMBIECU 120 6.982147 0.5333222 6.19 Table 3 shows the summary of the estimated model in which the directions and significance of the observed relationships support the research hypothesis. Four of the five explanatory variables proved to be significant at the 5% level (LEMBINONLAT, LUST10, DEBT-GDP, and DOL) and the remaining variable POL, proved to be significant at the 10% level. As is indicated by the F-Statistic (162.759) in the ANOVA section in Table 3, the overall model is identified and significant. Based on the Adjusted R2 of 0.87, only 13% of the variation in sovereign debt in Ecuador is unexplained by the model. Estimated Model Summary
Parameter
Independent Regression Standard T-Value
Prob. of Test
Variable Coefficient b(i) Error [Sb(i)] H0:B(i)=0 Level at 5%
Estimated Model

LEMBIECU = 8.216 + 0.299*LEMBINONLAT - 0.785* DEBT-GDP+
2.40*LUST10 + 0.122*POL + 0.753*DOL
Analysis of Variance
Source DF
Square F-Ratio Level
Total (Adjusted) 119 1.0000 33.84747 0.2844325 From Table 3, the variable of interest (DOL) seems to have increased sovereign risk in Ecuador. This means that official dollarization that was intended to improve the financial environment in Ecuador did not produce by itself the intended consequences. Looking at Chart 1, we can see a sharp drop in sovereign spread from March 2000, when the dollarization policy is implemented, to six months later in September 2000. However, up to four years after the official dollarization decision, Ecuador’s sovereign spread was still higher than in the pre-crisis 97-98 period. We can blame this on what we thought earlier in this paper that with dollarization and the loss of the monetary policy tool, Ecuador was not able to respond to shocks in the economy. This inflexibility in the economy resulted in an increased pressure on sovereign risk. Of secondary importance was the interest rate variable (LUST10). This proved to be inversely related to sovereign risk, leading us to conclude that financial crisis in Ecuador caused investors to ignore increasing yield in Ecuador as they looked for safer haven for their investments. This result is consistent with that obtained by Nogues & Grandes (2001). As expected, the contagion effect is significant and sovereign risk in Ecuador varies directly with sovereign risk in non-Latin countries (LEMBINONLAT). Instability in other emerging financial markets tends to increase sovereign risk in Ecuador, due to what market professionals call headline risk. With regards to the Public External Debt to GDP ratio variable, the empirical result obtained showed that sovereign risk varies directly related to the size of the external public debt to GDP ratio. Conclusion
In 1999, four years after implementing its Real Plan, Brazil devalued its currency and readjusted its exchange rate relative to the U.S. dollar in order to stabilize the economy. In March 2000, in the midst of a political and economic crisis, Ecuador removed its national currency, the Sucre, from circulation and adopted the U.S. dollar as legal tender. In the fourth quarter of 2001, eighteen months after Ecuador has officially dollarized, Argentina’s currency board collapsed. At that time, official dollarization was considered on the menu of policy prescriptions for Argentina. However, by 2002, Argentina’s government opted for a policy mix consisting of asset freeze and currency devaluation. Due to that financial crisis, Argentina experienced high levels of EMBI spread for four years. Curiously, it is the same amount of time it took for Ecuador’s sovereign spread to return to pre-crisis level. Moreover, during the last two years in our sample (2005-2007) Ecuador’s Sovereign spread has been higher than those of Argentina, Brazil and Mexico, a situation that resemble the first two years in the study period (96-97). These observations support the view that even though official dollarization may be appealing to policy-makers in the middle of a currency crisis, ultimately sovereign risk is function of the soundness of the government’s fiscal position and is not a function of the currency regime. Since Ecuador officially dollarized in March 2000, no country in South- America has followed what some expected to be the trend during the first decade of the twenty-first century. Since 2005, Argentina, an important member of Mercosur with strong ties to Brazil, definitely emerged from the Tango crisis. Moreover, the more fragile economies of the Andean region like Colombia, Peru, and Bolivia, believed to be the prime candidates for official dollarization during the nineties, managed to maintain a certain level of economic performance and preserved their national currencies in the presence of de-facto dollarization. If in policy-making circles official dollarization is no longer viewed as a panacea, future academic research on the determinants of the dollarization decision and its consequences on economic growth and the distribution of wealth should find more attentive ears. References
Abrego, L., E. Flores, A. Pivovarsky, & B. Rother (2006), “Ecuador”, Western Alesina, A & R. J. Barro (2001), “Dollarization”, American Economic Review, #91 (2), 381-85 Avila, J. (1998), “Riesgo Argentino y Ciclo Económico”. Cema Working Papers Beckerman, P. (2001) “Dollarization & Semi-Dollarization in Ecuador”, Policy Bogetic, Z. (2000) “Official Dollarization: Current Experiences and Issues”. CATO Journal, 20(2), 179-213. Calvo, G. (1999), “Contagion in Emerging Markets: When Wall Street is a Carrier”. Summer Camp, UTDT. Available at www.utdt.edu Calvo, G. and Reinhart, C. (1996), “Capital Flows to Latin America: Is There Evidence of Contagion Effects”, World Bank Policy Research Working Paper 1619. Edwards, S. (2001) “Dollarization and Economic Performance: An Empirical Investigation”, NBER Working Paper 8274 Edwards, S. & I. Magendzo (2003), “A Currency of One’s Own? An Empirical Investigation on Dollarization and Independent Currency Unions” NBER Working Paper Series http://www.nber.org/papers/w9514 Grandes, M. (1999), “Inversión en Maquinarias y Equipos: Un Modelo para la Experiencia Argentina 1991-1998”, Secretaría de Programación Económica y Regional, Subsecretaría de Programación Macroeconómica. Ize. A, & E. Levy-Yeyati (2005), “Financial De-Dollarization: Is It for Real?” IMF Jacome H, L. (2004), “The Late 1990s Financial Crisis in Ecuador: Institutional Weaknesses, Fiscal Rigidities, and Financial Dollarization at Work”, IMF Working Papers WP/04/12 Jameson, K (2004) “Dollarization in Ecuador: A Post Keynesian Institutional Analysis”, Department of Economics working Paper, University of Utah. Levy-Yeyati, E & F. Sturzenegger (2003) Dollarization, The MIT Press. Nogués, J. (1999), “Riesgo País y Crecimiento Económico en Argentina”, La Nación Nogues, J & M. Grandes (2001) “Country Risk: Economic Policy, Contagion Effect or Political Noise?” Journal of Applied Economics, Vol. IV, No. 1, pp. 125-162 Rodríguez, C. (1999), “Macreconomic Policies: Can We Transfer Lessons Across LDC´s?” Journal of Applied Economics 3: 169-216. Rubinstein, G. (1999), Dolarización. Argentina en la Aldea Global, Grupo Editor Willett, T. (2001), “Truth in Advertising and the Great Dollarization Scam”, Journal of Policy Modeling, pp. 279-289

Source: http://www.business.uconn.edu/finance/seminars/papers/pierre.pdf