Ebook Financial liberalization and banking crises: The role of capital inflows and lack of transparency
Financial liberalizations in emerging markets are often followed by reckless lending and severe banking crises. The identification of the causes of banks’ behavior is often difficult because financial liberalizations entail several contemporaneous changes. Competition in the banking sector increases, and, at the same time, the liberalization of the current account allows capital inflows.
The existing literature has mainly stressed the first of these changes: Among others, Hellmann et al. (2000) and Allen and Gale (2000) have analyzed how competition affects banks’ incentives to risk-taking. The argument goes as follows: Competition in the market for deposits increases banks’ cost of funds and gives an incentive to select riskier projects (to shift risk on depositors).
While this argument can provide a good explanation for the Saving and Loan crisis in the US, and, in general, for banking crises in developed markets, it is unlikely to capture the effects of financial liberalizations in emerging economies. In these countries, the liberalization of the capital account, which is an essential part of the financial liberalization, implies that large amounts of funds become available to the banking system, thanks to capital inflows. Since emerging market economies are small with respect to the funds potentially available from international investors, the supply of funds becomes perfectly elastic. Hence, not only is competition unlikely to increase the cost of funds, but the cost of bank liabilities often decreases because banks are no longer constrained by low domestic saving (Henry, 2000). Competition in the loan market is also unlikely to become so fierce as to significantly decrease intermediation margins and increase incentives to risk-taking. In fact, information asymmetries remain strong after financial liberalizations and banks face little competition from financial markets (Demirguc-Kunt et al., 2004). Hence, the very mechanism on which the competition argument relies—lower profit margins—is unlikely to be at work.
Probably for this reason, the vast literature on banking crises in emerging markets considers bailout guarantees as the main cause of excessive lending to unprofitable projects. Bailout guarantees would cause moral hazard because of the lack of punishment for investors and domestic banks in case of default. Knowing that banks will be bailed out, investors would provide funding even though they know that domestic banks are financing negative-net-present-value projects.
However, the empirical evidence on the bailout guarantees explanation is mixed (Eichengreen and Arteta, 2000). Martinez Peria and Schmukler (2001) find that depositors discipline banks by withdrawing deposits and requiring higher interest rates when bank fundamentals deteriorate, if the credibility of deposit insurance is weak. Additionally, Davenport and McDill (2005) find that deposit insurance does not diminish the extent of market discipline in the US, where the credibility of bailout guarantees should be far higher. Gorton and Winton (2003) also question the relevance of the bailout guarantees explanation.
This paper proposes a new explanation for boom-bust cycles in emerging markets. It shows that capital inflows may be at the origin of overlending problems. In the model, banks do not observe whether borrowers have access to positive-net-present-value projects when they begin to lend. As a result, banks may become insolvent. In this case, having limited liability, banks refinance negative-net-present-value projects, if they have access to funds. Thus, capital inflows may cause overlending because they increase the amount of funds domestic banks can intermediate.
In equilibrium, overlending arises if investors have incomplete information about the quality of bank assets. Incomplete information may originate from the fact that, within a country, some banks are sound, while others are insolvent. It may also depend on the fact that investors are unable to judge the determinants of lending booms. Consequently, they are uncertain whether capital inflows finance profitable investment opportunities or are used to roll over bad loans. Incomplete information may explain why, in equilibrium, investors initially provide shortterm loans to domestic banks at a relatively low interest rate, and only later on require a risk premium: Investors require a risk premium for lending to domestic banks only if there is a positive probability of bank defaults. In the early phase of the lending boom, since the expected amount of bank losses is still low, the probability of having lent to an insolvent bank is small. The risk premium, which compensate for such an event, is thus relatively low. Both solvent and insolvent banks issue debt (and fund their borrowers) if investors ask to be compensated for risk. This implies that any investor can be reimbursed in full because there are other investors who want to hold bank debt at the equilibrium interest rate. Hence, investors rationally do not require any risk premium.
If some banks are expected to accumulate losses, however, the probability of not recovering the capital invested increases, and so does the risk premium that compensates investors for risk. When a substantial amount of losses has potentially been accumulated (the threshold is determined endogenously in the model), solvent banks do not find it any longer optimal to issue debt at the interest rate that would compensate investors for risk. Investors anticipate this. Hence they first ask for a risk premium, and then completely stop holding bank debt. Insolvent banks thus default.
Put differently, the insolvency of a few banks creates a lemon problem in the market for bank liabilities, like in Akerlof (1970). In the early stage, the lemon problem is not as severe as to disrupt solvent banks’ demand for funds. The market for bank liabilities can thus thrive. As bank losses grow, the lemon problem becomes more severe. Solvent banks stop issuing debt when the risk premium becomes too high. Investors, aware of this, do not lend to domestic banks any longer. Only then, insolvent banks stop the process of evergreening of loans and declare default.
I show that in this context the liberalization of capital inflows may decrease aggregate welfare. After the financial liberalization, some capital is invested in negative-net-present-value projects. Depending on the country’s investment opportunities and the availability of domestic funds, the increase in profitable projects that receive funding may not be sufficient to compensate for the wasteful investment. If the country has low growth opportunities or high domestic saving, financial liberalizations decrease the average productivity of investment and ultimately aggregate welfare.
The model suggests some policy implications for avoiding boom-bust cycles. The presence of bailout guarantees appears to be irrelevant for overlending to arise. Restrictions on capital inflows could limit (or even avoid) credit expansion to insolvent projects, but would also constrain the number of projects that are started in countries with good growth opportunities. Also attempts to lengthen the maturity of capital inflows do not help eliminate overlending problems, because long-term debt does not prevent banks from accumulating losses.
Other mechanisms, which act by improving market discipline, are preferable. Greater transparency about the effective conditions of emerging economies, higher capital requirements and diversification of bank assets could improve financial stability. If investors knew the quality of bank assets, they would not lend to banks that fund negative-net-present-value projects. Even if international financial institutions are making efforts in this direction (Fischer, 1999; Basel Committee on Banking Supervision, 2003), it is difficult to completely overcome asymmetric information. Transparency by itself may not be sufficient to improve financial stability because even a small amount of incomplete information may lead to banking crises in the longrun, if the mechanisms highlighted in the model are at work.
Another, possibly easier, way of improving financial stability is by weakening banks’ incentives to renew loans to negative-net-present-value projects. Solvent banks do not renew bad loans because this would not maximize their net wealth. Hence, financial stability could be improved by increasing bank capital in countries with severe asymmetric information. Alternatively, financial stability could be improved by creating the conditions for better diversification of bank assets. If banks are well-diversified, their solvency is less likely to depend on the outcome of a handful of loans. Banks are thus less likely to become insolvent. Incentives to renew loans to insolvent projects therefore become weaker. Banks have weaker incentives to specialize and diversify their assets in environments where asymmetric information is lower (Winton, unpublished). Transparency could thus improve financial stability thanks to the indirect effects on the structure of bank-firm relationships.
The remaining of the paper is organized as follows. Section 2 describes the related literature. Section 3 describes the model. Sections 4 and 5 describe the equilibrium with and without capital inflows. Section 6 presents the welfare analysis. Section 7 examines some extensions of the model. Section 8 describes the institutional arrangements that can improve financial stability. Section 9 concludes.
Posted in :
