One of the biggest desires of policy makers around the world is to develop a warning system of currency crisis. Such a task requires their ability to analyze potential causes and symptoms of currency crises. Since the 1970s there have been numerous theoretical and empirical efforts to accomplish this task. Most of the empirical research focuses either on the developing Latin American and Asian countries or on the developed economies of the European Monetary System.
The recent transformation of communist economies to market economies coincided, in many of them (e.g., in Central and Eastern Europe, in the Commonwealth of Independent States and in the Baltic states), with the occurrence of currency crisis. But there is not much research of currency crises within the unusual environment of transition economies. Therefore, the predictive power of first, second or third generation models is questionable in the case of transition economies. The present work tries to cover this gap by concentrating on a subgroup of eight transition economies: Albania, Belarus, Bulgaria, Croatia, Macedonia, Moldova, Romania and Ukraine. A better understanding of the interrelationships between macroeconomic and other institutional, political and external factors provides transition economies with better understanding and guidance as to the type of policy needed to avoid currency crises.
The paper is constructed as following: the next section reviews previous theoretical and empirical research. This is followed by a discussion of the main macroeconomic problems and policy outcomes in the countries concerned. A description of the data and the variables used is provided in the fourth section. Section five presents the set of models and the results. Finally, in the last section we offer some concluding remarks.
Previous Theoretical and Empirical Research
The models of currency crises were built based on real events. The first generation models were developed after the balance of payments crisis in Mexico (1973-82), Argentina (1978-81), and Chile (1983). The second generation models were developed after speculative attacks in Europe and Mexico in the 1990s. Finally, the third generation models started after the Asian crisis in 1997-98. Transition economies in Central and Eastern Europe and the former Soviet Union offer additional empirical input that allows for re-examination of the existing theoretical models and accumulated empirical observations and verification of policy conclusions and recommendations proposed by other authors.
The research on currency crises first emerged in the economic literature in the late 1970s pioneered by Krugman (1979). According to him, under a fixed exchange rate system domestic credit creation in excess of money demand growth leads to a gradual but persistent loss of international reserves and, ultimately, to a speculative attack on the currency. The process ends with an attack because economic agents understand that the fixed exchange rate regime will ultimately collapse, and that in the absence of an attack they would suffer a capital loss of their holdings of domestic assets. Therefore, the first generation crisis occurs as a result of an expansionary macroeconomic policy incompatible with a pegged exchange rate. The collapse may happen when the "shadow floating exchange rate" becomes equal to the exchange rate peg. This is the equilibrium exchange rate prevailing after the full depletion of foreign reserves and forced abandoning of the peg.
A number of papers have extended Krugman's basic model in various directions (see, for example, Agenor, Bhandari and Flood 1991; Eichengreen, Rose and Wyplosz 1995; Blackburn and Sola 1993; Garber and Svenson 1994; Flood and Marion 1999; Edwards 1989; Grilli 1990; Cumby and Wijnbergen 1988; Harris and Raviv 1989; and Kamisky, Lizondo and Reinhart 1998). Some of these extensions concern active governmental involvement in crisis management and sterilization of reserve losses (Flood, Garber and Kramer 1996). Other extensions have shown that speculative attacks would generally be preceded by a real appreciation of the currency and a deterioration of the trade or current account balance, by an upward pressure of real wages and by higher interest rates (see survey in Garber and Svenson 1994). Extensions also include target zone models (Krugman 1991), post-collapse exchange rate systems other than permanent float, the possibility of foreign borrowing, capital controls, imperfect asset substitutability, and speculative attacks in which the domestic currency is under buying, rather than selling pressure.
The second generation models suggested by Obstfeld (1986; 1994; 1996), Cole and Kehoe (1996), Eichengreen, Rose and Wyplosz (1996), Sachs, Tornell and Velasco (1996) and Drazen (1998) are particularly useful in explaining self-fulfilling contagious currency crises. They show that a crisis may develop without a significant deterioration in the fundamentals. The idea of second generation models is based on the fact that defending exchange rate parity can be expensive (through higher interest rates) if the market believes that it will ultimately fail. This set of assumptions opens the possibility for multiple equilibria and self-fulfilling crises. Second generation models also tend to focus on political factors, such as political cost of high unemployment or foregone output.
A third generation of models gives a key role to financial structure fragility and financial institutions. Microeconomic problems, such as weak banking supervision, corruption etc., trigger capital outflows and finally currency attack. The proponents of this view use data from the Asian crisis to support the main ideas (Corsetti, Pesenti and Roubini 1998a; 1998b). Stops of capital inflows are explained as a byproduct of bank runs due to the internationally illiquid banking sector. Krugman (1999), Aghion, Banarjee and Baccetta (2000) and Aghion, Baccetta and Banarjee (2001) examine the effects of monetary policy on currency crises (such as moral hazard and resulting over-borrowing). Vaugirard's (2007) explanation of contagion involves real (trade and financial) linkages between countries. Successful .speculative attacks against the currency of a country, which exports goods that are substitutive to goods sold by a not-attacked country forces the latter also to devalue in order to maintain its competitiveness.
The empirical literature as regards currency crises is also vast. Most of the empirical studies ("warning system" approach and "stylized facts," "single-country and "multi-country") emphasize variables that were found as leading indicators of crises. All the studies were driven by the desire of authors to analyze potential causes and symptoms of currency crises and to develop a warning system of currency crisis.
The "warning system" approach is strongly associated with the work of Kaminsky, Lizondo and Reinhart (1998), Kaminsky (1998), Kaminsky and Reinhart (1999) as well as Wu, Yen and Chen (2000). The basic idea behind the "warning system" approach is that currency crises usually do not happen, i.e., that pure self fulfilling attacks are rare, but that most crises are preceded by deteriorations in the economic fundamentals of the economy.
"Stylized facts" studies focus on specific episodes of financial turmoil. While these models are less geared toward predicting the exact timing of financial crises, rather, they aim at explaining the severity of financial crises. Papers by Blanco and Garber (1986), Sachs, Tornell and Velasco (1996) or Bussiere and Mulder (1999) are notable examples for this kind of model class.
The main findings of "single country" studies (for a review see Kaminsky, Lizondo and Reinhart, 1998) is that macroeconomic indicators (foreign reserve losses, expansionary fiscal and monetary policies and high interest rate differentials) play a significant role in determining currency crises. The problem with these studies is that their results are limited since they are obtained from a small number of countries during very specific situations.
"Multi-country" studies avoid the limitations of the above single country studies (for a review see Esquivel and Larrain 1998, and Kaminsky, Lizondo and Reinhart 1998). Among the most significant determinants of currency crises are the low levels of foreign direct investment, low international reserves, high domestic credit growth, high foreign interest rates, overvaluation of the real exchange rate, output, exports, deviations of the real exchange rate from trend, equity prices, the ratio of broad money to gross international reserves etc. Esquivel and Larrin (1998) represent the first attempt to simultaneously test the main predictions of both the first and second generation models of currency crises. The explanatory variables closely associated with first generation models are seignorage, real exchange rate misalignment, current account balance, and M2/reserves ratio. As far as the second generation model is concerned, they use terms of trade shock, per capita income growth and contagion effects. Their results suggest that the insights developed by second generation models complement rather than substitute for the explanation provided by first generation models.
Empirical studies dealing with currency crises in Central and Eastern Europe are scarce, mainly for the obvious reason of the shortness of time series. Notable examples include BrOggemann and Linne (1999; 2001), Krkoska (2001), Chapman and Mulino (2000), Chiodo and Owyang (2002), Desai (2000), Kharas et al. (2001) Karfakis and Moschos (2004), Dobrinsky (2000), Chionis and Liargovas (2003) and Kemme and Roy (2006).
Bruggemann and Linne (1999; 2001) basically apply the Kaminsky-Lizondo-Reinhart (1998) framework with a few extensions to 13 Central and Eastern European countries (CEECs) and three Mediterranean countries (Cyprus, Malta and Turkey). Krkoska (2001) estimates a...