The Interactions of Strength of Governments and Alternative

The Interactions of Strength of Governments and Alternative Exchange Rate Regimes in Avoiding Currency Crises By Eric M. P. Chiu and Thomas D. Willett National Chung-Hsing University Claremont Graduate University Nov 18, 2006

Introduction n n Under a world of high capital mobility, the Bretton Woods style adjustable pegged regime (dead center) is known as the most crisis prone type of exchange rate arrangement due to the problem of “oneway speculative gamble”. This is also known as the “unstable middle” hypothesis.

Puzzles n n How far away from the middle countries need to move to substantially reduce the frequency of crises? Unstable Middle vs. Two Corners Hypothesis

Some Responses n n Frankel (2004) has recently argued that we do not have a good economic theory for why the middle itself is unstable. However, Willett (2006) suggest that combining both standard economic theory with political analysis of the incentives to delay needed adjustment, we can better understand why middle is unstable.

Main Purpose n Both the issues of why the middle is unstable and how far toward the extremes of fixed or flexible exchange rates countries need to go to substantially reduce the likelihood of currency crises depends crucially on political factors, i. e. the strength of governments, as well as standard economic considerations.

Two Innovations n n Use of IMF new de facto exchange rate regime classification as opposed to de jure classification Focus on the interactions between weak governments and alternative exchange rate regimes in influencing crisis probabilities.

Types of Weak Govt. (1) Politically unstable governments 1. n n Political instability (e. g. frequent government turnovers, riots, strikes, and lack of strong political popularity) is likely to force a government to give increased weight to short run political considerations, thereby increasing the propensity to postpone policies such as exchange rate and macro economic policy adjustments, which enhances the likelihood of crises. International Country Risk Guide (ICRG) defines government stability as government’s ability to carry out its declared program, and its ability to stay in office. The ICRG stability index ranges from 1 (the lowest level of government strength) to 12 (the highest level).

Types of Weak Govt. (2) 2. Politically divided governments (veto players) n n n Policymakers in divided governments in general face collective action problems. In particular, a country with many veto players will delay the changes in policy in responding to external shocks due to the difficulty in rectifying agreements among policymakers. A dummy variable that takes the value of one when the chief executive’s party controls the legislature, that is, a unified government. (Becks, 2003) Veto Player Approach: using “Checks” from DPI. It counts the number of veto players, actors whose approval is necessary for a shift in policy from the status quo. The higher the score, the greater is the policy constraints.

Hypotheses n n H 1: other things being equal, the more unstable and the more divided is a nation’s government; the higher is the probability of currency crises under any type of exchange rate regime. H 2: other things being equal, the effects of weak governments on the likelihood of crises will be stronger, the stickier is the exchange rate regime.

Sample & Data n n The data set comprises annual observations from 1990 to 2003 on 90 countries, including 21 industrial countries, 42 emerging markets economies, and 27 low-income developing countries. International Country Risk Guide (ICRG); Database of Political Institutions (DPI); International Financial Statistics (IFS); World Development Indicators (WDI)

Methodology n n n A probit panel model is used. Interaction dummies between political variables and exchange rate regimes are included. A one-year lag is used for all independent variables to avoid the endogeneity problem.

The Model n A probit panel model is defined as:

List of Variables n n Y: Crisis dummies G: Political variables including political instability and divided government ER: a six-way classification based on BOR’s exchange rate regimes X: control variables including: n n n Lending boom M 2/Reserve Current Account/GDP Real Effective Exchange Rate Index Election dates

Empirical Results (1) (2) (3) Stability Unified Checks STAB t-1 -0. 1582*** (0. 0698) - - Unified t-1 - -0. 5506*** (0. 2862) - Checks t-1 - - 0. 1021** (0. 0463) 710 582 649 Prob > Chi-Square 0. 0000 Pseudo R 2 0. 1523 0. 0795 0. 1390 -153. 1022 -196. 3370 -119. 8105 No. of obs. Log-Likelihood

Hard pegs Adjustabl e Forward Crawls Backward Crawls Managed Floats Independen t Floats Checks = 1 0. 0003 0. 0967 0. 2706 0. 0271 0. 0968 0. 0238 Checks = 2 0. 0017 0. 1088 0. 2530 0. 0455 0. 1026 0. 0313 Checks = 3 0. 0069 0. 1220 0. 2360 0. 0728 0. 1086 0. 0407 Checks = 4 0. 0224 0. 1362 0. 2197 0. 1111 0. 1148 0. 0522 Checks = 5 0. 0605 0. 1515 0. 2040 0. 1619 0. 1213 0. 0662 Checks = 6 0. 1365 0. 1680 0. 1890 0. 2060 0. 128 0. 0829 Checks = 7 0. 1855 0. 2605 0. 1747 0. 2424 0. 135 0. 1026

Main Findings 1. 2. As we expected, weak political institutions particularly characterized by unstable governments and divided governments increase the likelihood of currency crises under any type of exchange rate regime. More specifically, we find weak political institutions combined with adjustable pegged exchange rate system represent the highest probability of currency crises, as compared with other types of exchange rate regimes.

Investigating the anomaly of Crawls n n Dividing crawling regimes into forward-looking crawls and backward-looking crawls. Hypo: Forward looking crawls tend to be associated with efforts at exchange rate based stabilization (ERBS) policy, which require a strong government to implement it. This will make crises more likely than backward looking crawls.

ci 22==1 & crawl. L=1 Country Year Stab. L Div Checks ER regime in previous year Brazil 1999 10. 17 0 6 7 Chile 1999 10. 33 0 4 8 Colombia 1995 6. 17 0 2 8 Ecuador 1999 9. 75 0 4 8 Indonesia 1997 9. 25 1 1 10 Israel 1998 7. 17 0 3 8 Mexico 1994 7 1 2 8 Portugal 1991 6. 67 0 2 8 Russia 1998 9. 83 5 7 Sri Lanka 1998 9. 83 1 5 10 Turkey 1994 6. 33 0 3 9 Turkey 2001 9. 67 0 6 7 Uruguay 2002 10. 75 0 2 8 Venezuela 1994 4 0 5 9

Justify the Anomaly of Crawls n n Many governments overestimated the likelihood of success on ERBS. It was generally understood that extremely weak and unstable government would have little chances of successfully stabilizing. Thus it seems likely that the stronger governments of high inflation countries would be more likely to attempt ERBS. A higher than expected tendency for such programs to end in crises such as occurred with Mexico in 1994 and Brazil in 1999 could help explain the positive correlation between strength of government and the probability of crises under forward looking crawls. This is clearly an issue worthy of further investigation.

Robustness Checks 1. 2. 3. Crisis Indices: using equal weight system. Veto Player Approach: using “Checks” from DPI. It counts the number of veto players, actors whose approval is necessary for a shift in policy from the status quo. The higher the score, the greater is the policy constraints. Alternative Measures of Exchange Rate Regimes by Reinhart and Rogoff (2002).

Conclusions (1) n n The most important conclusion of our analysis is that weak political institutions significantly increase the probability currency crises. This paper demonstrates that both unstable government and divided government will make currency crises more likely particularly under Bretton Wood narrow band adjustable peg regimes than other types of exchange rate regimes.

Conclusions (2) n n These findings are consistent with the “unstable middle” hypothesis, while are not in line with the “two corners” hypothesis. It also implies that countries with weak political institutions should at least consider move away from the “dead center” of narrow band adjustable peg in order to avoid severe speculative attacks. By contrast, countries with strong political institutions are more likely to effectively manage intermediate exchange rate regimes in a stable manner.

Grouping of BOR’s Exchange Rate Regimes

R-R Regime Classification n n Hard pegs include the most rigidly peg or currency board. Adjustable parities include a pre-announced narrow horizontal band a de facto peg. Crawls include de facto and pre announced crawling pegs or bands as well as pre announced crawling band that is wide than or equal to +/- 2%. Moving bands, Managed floating, and Freely floating are treated as three individual categories. We do not consider “freely falling” regime in this paper since in reality few countries would choose to have, nor they could implement a “freely falling” exchange rate.

Simple Descriptive Statistics Variable Observations Mean Std. Dev. Min Max Stability 1086 7. 82 2. 12 1 12 Election 1158 0. 27 0. 45 0 1 Divided 1061 0. 46 0. 49 0 1 Checks 1130 3. 26 1. 89 1 7 M 2/Res 1107 7. 93 14. 05 0. 19 148. 31 CA/GDP 1085 -1. 16 9. 44 -41. 49 39. 62 Lending 1086 3. 71 18. 54 0. 014 112. 51 REER 892 102. 16 17. 96 42 205