Twin Crises – The Contribution of Overborrowing to Interdependencies of Banking and Currency Crises

Contents

Figures

Tables

Abbreviations

Symbols

1 Preface
1.1 Basics on Twin Crises
1.2 A Short Review on Twin Crises Literature

2 Vulnerability to Twin Crises by International Overborrowing
2.1 Basics on International Overborrowing and its Investigation
2.2 The Overborrowing Syndrome
2.3 Assumptions
2.4 Model Specification: Financial and Trade Liberalization
2.5 Identification of Equilibria: Financial and Trade Liberalization
2.6 Model Specification: Uncertainty and Market Failure
2.7 Identification of Equilibria: Uncertainty and Market Failure
2.8 Model Specification: Unhedged Foreign Exchange Rate Risk
2.9 Identification of Equilibria: Unhedged Foreign Exchange Rate Risk
2.10 Policy Conclusions for Counteracting International Overborrowing
2.11 Further Theoretical Research on International Overborrowing

3 Empirical research
3.1 An All-Encompassing Approach by Kaminsky (2006)
3.2 A Case Study on Korea by Chang, Park and Yoo (1998)
3.3 An Overborrowing and -investment Approach by Ryou/ Kim (2003)

4 Final Remarks
4.1 Overview on the Issue of Coexistent Overborrowing and Twin Crises

Appendix A – Enhancement of real interest rate parity

Appendix B – Regression-tree results in Kaminsky (2006)

References

Figures

Figure 1.2.1: Systemization of third generation and twin crises literature.

Figure 2.5.1: Equilibrium in a financially repressed economy.

Figure 2.5.2: First best equilibrium in a domestically liberalized economy.

Figure 2.5.3: First best equilibrium in an internationally liberalized economy.

Figure 2.7.1: Market failure in a domestically liberalized economy.

Figure 2.7.2: Market failure in an internationally liberalized economy.

Figure 2.9.1: Market failure with unhedged credit and foreign exchange risk.

Tables

Table 2.3.1: Assumptions referring to McKinnon and Pill (1996, 1998)

Table 2.6.1: Risk and its observers referring to McKinnon/ Pill (1996)

Table 3.1.1: Indicators of currency crises

Table 3.1.2: Numbers of particular reports dependent on an occurring event

Table 3.1.3: Types of crises with different trigger mixtures

Table 3.1.4: Varieties and costs of crises

Table 3.1.5: Crises in emerging and mature markets

Abbreviations

[illustration not visible in this excerpt]

Symbols

[illustration not visible in this excerpt]

1 Preface

1.1 Basics on Twin Crises

Following Solomon (2003), Kaminski and Reinhart (1999) first labeled the phenomenon of twin crises as banking and currency crises cropping up simultaneously^[1]. The occurrence of twin crises can evoke serious recessions. From 1976 to 2002 there were 38 countries afflicted by at least one twin crisis. Popular examples are East Asia in 1997/1998 as mentioned in Sarno and Taylor (2008) and Latin America 1995-1997 following MacDonald (2008). One relatively unknown example argued by Berlemann, Hristov and Nenovsky (2002) was Bulgaria 1996/1997. Recently, Argentina^[2] and Turkey were hit by a twin crisis in 2001 as well.

Banking crises are distinguished in incidents of bank runs, banking panic and significant banking sector problems inducing serious damage to the creditworthiness of the major part in the banking sector.^[3] Basic literature on bank runs by Diamond and Dybvig (1983) describes herding behavior of creditors caused by inappropriate maturities of bank liabilities. If banks finance long-dated risky investments via short-dated debt, bankruptcy will become more likely. Thus, there will be an incentive for single creditors to withdraw their deposits. This may stimulate other creditors to put this intent into action, leading to a chain reaction of withdrawals called bank run. A banking panic is a combination of broad insolvencies in the banking sector and constrained disbursements of deposits leading to bank runs as well.

According to Burnside, Eichenbaum and Rebelo (2007) a currency crisis is an episode with massive exchange rate depreciations within a short space of time. Literature examining currency crises is divided in three generations:^[4] First generation models trace currency crises back on unsound fiscal and monetary fundamentals. Randomly altering market expectations or countercyclical policies may also trigger currency crises, as investigated in second generation models, while third Generation models extend these views. Following Schneider and Tornell (2000) they specify financial market imperfections as another origin of inconsistencies. More precisely, these papers deal with financial intermediaries and liquidity effects. Krugman (2001) enhanced those contributions by modeling influences of asset prices and exchange rates on corporate balance sheets.^[5]

This paper continues as follows: At first it will systemize literature about twin crises and give a review on several contributions. The second section deals with a basic third generation model of McKinnon and Pill who specialized on a phenomenon they call “The Overborrowing Syndrome”. Their key findings that loan guarantees may lead to a twin crisis via excessive borrowing. After this, a review on the development of overborrowing papers will complete the third section. In section four it will refer to empirical studies on overborrowing. Section five closes with final remarks.

1.2 A Short Review on Twin Crises Literature

As mentioned in Schneider and Tornell (2000), twin crises literature has arisen out of third generation contributions. The first major type of third generation models is argued by MacDonald (2008) to depend on moral hazard ^[6], while the second major type specifies bank runs as a trigger for currency crises. In a more practical and broader view following Schneider and Tornell (2004), third generation models are systemized in “bad policy” and “bad market” models, underpinning Krugman’s (1999) concern that financial market imperfections do not necessarily stem solely from banks. More precisely, weak balance sheets or massive capital outflows can indirectly^[7] stir banking problems. Thus, third generation models focus on clear twin crises and on crashing debt and equity markets entailing currency crises. Figure 1.2.1 illustrates this systemization as a radar chart.

[illustration not visible in this excerpt]

Abbildung in dieser Leseprobe nicht enthalten

Figure 1.2.1: Systemization of third generation and twin crises literature.

Referring to Kaminsky/ Reinhart (1999), theories of twin crises assume several interdependencies between banking and currency crises. The first type^[8] of theories model banking crises leading to currency crises, while the second type^[9] of theories suppose the reverse direction. The third type^[10] of theories presume shared causes of banking and currency crises.

Solomon (2003) set up a first type game theory model related to the Turkish twin crisis in 2001. His approach achieves explaining why bank runs do not necessarily involve a currency crisis. Additionally, intersecting extrinsic information sets of domestic and foreign agents are modeled, allowing for different behavior of residents and foreigners and an equilibrium selection process. For all these previous purposes, imperfectly correlated sunspot variables denoting extrinsic uncertainty are used. The multiple equilibria in his paper are banking crisis, currency crisis, twin crisis or no crisis. Solomon’s model suffers from the assumption of crises to occur exclusively caused by self-fulfilling expectations without referring to fundamentals.

The paper supposes the following progress to lead to a twin crisis: At first banks make investments with domestic and foreign deposits based on previously offered contracts. When the bank starts business, impatient domestic agents receive available dollars, while patient domestic agents receive utility services in domestic currency. The specification of the offered contracts determines whether a bank run can occur. If a bank run exhausts the central banks reserves of dollars, a currency crisis will ensue. The probability of this twin crisis equilibrium is dependent on the distribution of the imperfectly correlated sunspot variables.

Burnside, Eichenbaum and Rebelo (2000) combined a self-fulfilling currency crisis model based on Obstfeld (1986) with a basic form of their banking model in Burnside, Eichenbaum, Rebelo (1999). With this second type approach they purpose explaining the twin crises in Sweden 1991/1992, Mexico 1994, Thailand 1996/1997 and Korea 1997. Government guarantees are the fundamental giving rise to the occurrence of twin crises. The exact point of time when twin crises occur is determined by expectations of economic agents. The latter believe that a successful speculative currency attack involves expansionary monetary policy. If there is a deprecation exceeding a certain threshold, speculators attack the domestic currency. This activity boosts foreign liabilities of domestic banks. Bailout plans financed to some extend by expansionary monetary policy fuel further depreciation.

The advance of this model is the identification of a clear criterion disclosing a fragment of countries in which twin crises may become reality. On the other hand, this approach does not allow for a clear forecast of the point of time when the affected countries will encounter difficulties of twin crises. This disadvantage is due to the fact that the timing of twin crises is modeled by agents’ beliefs.

An open economy version of Diamond and Dybvig’s (1983) bank run model of the third type was introduced by Chang and Velasco (2000), referring to the impact of the Tequila Crisis in Mexico 1994 and the East Asian Crisis in 1997/1998. Their view is focused on interdependencies of three determinants bringing about twin crises: the banking system, the exchange rate regime and central bank policy. In this vein, Chang and Velasco (2000) investigate the vulnerability depending on variedly combined specifications of those criteria.

One important finding is that central banks obligated to defend a fixed exchange rate and to be lender of last resort provide highest likelihood of twin crises. In this case the incidence of international illiquidity^[11] determines the occurrence of crises. Lender of last resort policy is pointed out as a factor determining the direction of causality. If pursued, this policy could induce banking crises entailing currency crises. If not pursued, banking crises could be an accompaniment of currency crises. High bank reserve policies and extensive international reserves are one solution to avoid speculative attacks and/ or bank runs in a fixed exchange rate regime. The central disadvantage of this combination is a welfare loss burden. Flexible exchange rates and a central bank as lender of last resort is suggested as the best mix realizing the social optimum and circumventing both self-fulfilling bank runs and speculative attacks.

Chang and Velasco’s (2000) paper exhibits following weaknesses: At first, the manner of modeling demand for domestic currency has several implications. For a start it does not account for foreign currency assets, therefore implying a higher probability of illiquidity. Moreover Chang and Velasco (2000) argue their proceeding of modeling domestic currency demand via utility functions as a “brute force method”. Allowing for fundamentals would clear out oversimplifying in this regard. In addition to this, the finding of welfare advantages in a fixed exchange rate regime compared to a currency board could disappear by considering opportunity costs of holding exchange reserves. At second Chang and Velasco (2000) use a basic way of foreign credit lines. Wise extensions could be a provision for policy credibility and an international lender of last resort. Finally, in a third generation perspective, it has to be mentioned that the paper only concentrates on self fulfilling runs without considering moral hazard problems connected with sovereign guarantees to private borrowing. This aspect is treated in several contributions published by McKinnon and Pill (1996, 1997, 1998, 1999), referred to in the next section. These papers set basic milestones in respect of international overborrowing and are still increasingly cited by current publications.

2 Vulnerability to Twin Crises by International Overborrowing

2.1 Basics on International Overborrowing and its Investigation

McKinnon and Pill’s (1996), (1997), (1998), (1999) motivation of examining international overborrowing came from crises after governmental reforms in Chile 1982/1983, Great Britain 1990/1991, Sweden 1991/1992, Mexico 1994/1995, Argentina 1997 and Southeast Asia 1997. One should take into account that both developing countries and industrialized nations are prone to “the overborrowing syndrome”.

Contributions published before McKinnon and Pill’s papers emanated from the idea of trade liberalization of consumer imports expected to be transitory. Domestic agents play on the “window of opportunity”^[12], financing self-indulgent consumption by overborrowing. Whereas these earlier papers were centered on insufficient credibility of reform policies, McKinnon and Pill suspect the opposite. In fact, they focus on excess credibility of governmental reform’s success, inducing massive surge of capital inflows due to overoptimistic prospects of future revenue. Their work was excited by numerous examples of countries whose non-tradable sector was shocked by overborrowing. Another source of motivation they bring up is an empirical paper released by Reinhart and Végh (1994), stating insignificance of the deficient credibility hypothesis because capital inflows were too tremendous to deliver support.

The matter of McKinnon and Pill’s investigation are macroeconomic aspects inducing “the overborrowing syndrome”, getting granular on interdependencies of credible industrial reforms, international capital flows and market failures due to financial reforms. The major indications of overborrowing episodes argued are soaring domestic borrowing, amplification of current account deficits, slack domestic monetary regulation, real exchange rate appreciation, a capacious share of foreign deposits placed via domestic banks and high vulnerability to adverse shocks finally cumulating to an entailing financial crisis, sudden capital flight and downturn.

2.2 The Overborrowing Syndrome

Basic research on a mechanism labeled boom-bust cycles caused by international overborrowing was published by McKinnon and Pill (1996). McKinnon and Pill (1997, 1998, 1999) released further research. They argue moral hazard to be triggered by governmental bailout arrangements. Thus, industrial and financial reforms^[13] may lead to excessive short-dated capital inflows^[14] in presence of asymmetric information.

[...]

^[1] Eichengreen/ Bordo (2002) comprehend simultaneously as the same year or sequenced years.

^[2] As named in Bleaney/ Bougheas/ Skamnelos (2008), p. 696.

^[3] These events are argued as indications of banking crises by Eichengreen/ Bordo (2002), pp. 15-16.

^[4] See Kaminsky (2006), p. 504 and MacDonald (2008), pp. 313 – 328 for a brief overview.

^[5] This article is casually said to have excited fourth generation models, see MacDonald (2008) p. 326.

^[6] „A term borrowed from the insurance literature, it means that incentives of the insured party may be altered once the insurance contract is entered into. [...] The insured individual may be less concerned about security than in the absence of a contract.“, following Heffernan (2003), p. 370.

^[7] Goldstein (2005) dubs this chain of causation indirect.

^[8] E.g. see Solomon (2003), Dooley (2000), Buch/ Heinrich (1999), Krugman (1999), Krugman (1998), Miller (1998). Neeman/ Orosel (2002) present a notable model which focusses on foreign banks as too sweet-tempered lenders inducing a currency crisis.

^[9] E.g. see Burnside/ Eichenbaum/ Rebelo (2000). Burnside/ Eichenbaum/ Rebelo (2001) broaden the perspective of second type theories by focussing on the impact of distinct policies for financing the fiscal costs of twin crises.

^[10] E.g. see Shin (2005), Nissanke/ Stein (2003), Flood/ Marion (2001), Chang/ Velasco (2000),

^[11] „Defined as a situation in which the financial system’s potential short-term obligations exceed the liquidation value of its assets“, referring to Chang and Velasco (2001), p. 490.

^[12] See McKinnon/ Pill (1996), p. 10.

^[13] Following McKinnon/ Pill (1996), industrial reforms could be trade liberalization, e.g. via abandoning tariffs, privatization, deregulation and domestic structural reforms, e.g. improvements of labor legislation. A financial reform is understood as domestic or international financial liberalization and/ or the attempt to keep domestic the payments system in good order via public guarantees of banks’ deposit liabilities.

^[14] McKinnon/ Pill (1996) name these capital inflows “hot money“.