Comment

Abstract

Previous articleNext article FreeCommentKnut Anton MorkKnut Anton MorkHandelsbanken Capital Markets Search for more articles by this author Full TextPDF Add to favoritesDownload CitationTrack CitationsPermissionsReprints Share onFacebookTwitterLinked InRedditEmailQR Code SectionsMoreWith the emergence of cross-border banking, banks have become a key channel for the international transmission of financial conditions as well as real activity. Furthermore, with capital requirements a central tool for banking regulation, bank losses have fairly direct effects on bank lending, which is often a key determinant of business investment. That makes this paper highly topical. Indeed, as European sovereign defaults are being extensively discussed and — in the case of Greece — actually have taken place, this paper provides important insight into the current dynamics of the economies of Europe.On one level, the mechanism studied in this paper is very simple: banks invest in foreign sovereign bonds, the foreign government defaults, the default makes the bank undercapitalized, and to restore capital adequacy the bank curtails lending, thus dampening business investment. This mechanism should naturally apply to all sorts of bank losses. In this sense, the analysis is applicable to a much wider range of events than foreign sovereign defaults.The authors have sought to take the analysis to a higher level, however, by embedding the analysis of banks and business investment in a dynamic stochastic general equilibrium (DSGE) model. Within a general equilibrium framework, it is no longer obvious that a government default should weaken business investment because the supply of government bonds drops, thus lowering yields and offering an incentive for banks to substitute from government-bond investment to business lending.1 The authors show that, with their preferred parameter calibration, this effect will be dominated by the banks’ overall desire to delever after a capital loss. However, they also present an alternative calibration of adjustment costs that makes this substitution effect dominate for the Core banks, so that a government default in the Periphery helps rather than hurts growth in the Core.This sensitivity to unobservable adjustment costs illustrates the challenges involved in introducing the complexities of financial intermediation and government finances in an already complex, dynamic, and stochastic modeling framework. Although few people doubt the existence of adjustment and transactions costs, these costs take on an exaggerated importance in the model because they are used to prevent large changes in the portfolio choices among assets and liabilities (bank deposits as well as domestic and foreign government bonds) that are otherwise perfect substitutes in the model. Unfortunately, the adjustment costs can be used rather effectively to manipulate the results, as exemplified by the case just mentioned.In the real world, capital-adequacy rules are imposed to compensate for the moral hazard problem created by the likely bailout of banks in the case of crisis. In other words, banks face a truncated version of the distribution of outcomes, and requirements for capital relative to risk-adjusted assets are imposed as a way to prevent banks from taking on excessive risks. A significant part of these risks typically come from the banks’ loan portfolios. Government bonds are normally considered less risky and thus are assigned lower or no weight in the calculation of risk-adjusted assets. This is one of the reasons why banks hold government bonds. In addition, government bonds serve as collateral for central-bank funding and can, because of their liquidity, be used to quickly raise cash in case of unexpected funding squeezes. Moreover, modern banks regularly hold foreign assets, including loans to foreign clients as well as foreign government bonds. In fact, many banks have branches or subsidiaries in countries other than the one where they are headquartered.Including these features in the model of this paper would have added an additional, substantial layer of complexity on top of the already complex DSGE structure. Ad hoc tricks are used instead to emulate real-world behavior: Loans are assumed riskless, but nevertheless 100% risk weighted, whereas both foreign and domestic government bonds receive zero risk weight even though default can occur. What nevertheless motivates banks to make their loan portfolio large enough to make the capital requirement binding is the assumption that banks are more patient than entrepreneurs, yet less patient than households. Furthermore, they choose between foreign and domestic bonds as if the two were equally safe even though only Periphery bonds default in practice.Simplifying assumptions are naturally a necessity in all models. That is what makes them models. However, the kind of simplifications chosen here make the analysis look unnecessarily mechanical. The one convincing result is the finding that bank capital losses have real effects via capital-adequacy requirements. That insight is important, but could have been derived in a much simpler model without the DSGE structure. In fact, this structure seems to serve as a straightjacket preventing the authors from taking on the full analysis of bank risk, bank risk management, and bank risk regulation. A simpler dynamic structure, for example, in the three-period tradition of Diamond and Dybvig (1983), would have made this task easier. A simpler structure could also allow for other real-world features regarding government finances, such as time-varying probabilities of future default, reflected in longer-term bond prices and demand-side macro effects of budget tightening undertaken to stave off default.The internationalization of the banking industry is a relatively recent phenomenon, facilitated by developments in regulation and policy as well as technology. Banks view it as a way to exploit the returns to scale on their intangible assets such as know-how, organization, and information systems. Legal restrictions have been eased significantly in recent decades, allowing, for example, the establishment of foreign branches rather than subsidiaries. This is important to the banks because they then can use their entire capital base in all locations and also because it facilitates managerial control, even as it represents a challenge to regulators. Information about foreign clients, essential for risk management, is made easier by the establishment of foreign branches as well as modern information technology. Finally, exchange risk has been removed for cross-border activity within the euro area.Cross-border operations raise new risks as well as new opportunities. In many contexts, credit risk vis-à-vis foreign borrowers may be as important as the risk of foreign government defaults, although government defaults tend to be larger when they occur. Within the euro zone, the motivation for holding foreign government bonds is not clear from a risk perspective because domestic bonds work equally well in terms of risk weighting and are equally eligible for funding purposes, in private markets as well as with the ECB. The fact that Core banks do hold Periphery bonds thus seems rather to be explained by their higher return. From a risk perspective, the motivation is actually stronger for banks outside the euro area because they need euro denominated bonds as collateral for funding of their euro-denominated investments (including lending). This is the likely reason why the authors of this paper find the euro government bond holdings of UK and Swiss banks large enough to include them in the Core for calibration purposes. Norway, Sweden, and Iceland would have been additional candidates, though probably too small to make a difference for the results. Before crisis struck, Iceland was a rather interesting case of cross-border banking, especially because their home currency was so small.The paper ignores bank liabilities other than deposits. Although this has the rather peculiar consequence of making the capital requirement take the form of a ceiling on deposits, the authors are probably right that deposits in the model can be sensibly thought of as the total of deposits, money-market funding, and long-term bank bonds. Their omission of a bank equity market matters more because issuing new equity is an alternative to curtailing the asset side for a bank that needs to de-lever. The authors are of course right that this alternative may not be attractive to a bank’s existing shareholders in a crisis situation. However, it could still be used as a policy instrument; that is, as a requirement to banks to raise specified amounts of capital rather than satisfying the minimum ratio of capital to risk-weighted assets. The point of such a requirement would naturally be to avoid a curtailment of bank lending in the middle of crisis. It was used as a penalty for banks that failed the stress tests for the Supervisory Capital Assessment Program (SCAP) in the United States in the spring of 2009.2 For example, Bank of America was asked to raise $33.5 billion and succeeded. By the end of that year, the banks involved had raised a total of more than $125 billion of common equity. A model analysis with level requirements for Bank capital would thus be an interesting task for future research.EndnotesThe views expressed are those of the author and do not necessarily coincide with the official views of Handelsbanken Capital Markets or Svenska Handelsbanken AB (publ.), of which it is a unit. For acknowledgments, sources of research support, and disclosure of the author’s material financial relationships, if any, please see http://www.nber.org/chapters/c12787.ack.1. Yields might be expected to rise rather than fall after a default because of a loss of confidence. However, this model is constructed such that defaults occur in equilibrium only if they are completely unanticipated. Thus, the loss of market access that is specified in the model does not translate into a loss of confidence. In a more general model, a loss of confidence would furthermore provide no new incentive to invest in government bonds, only result in a premium to compensate for the loss of confidence.2. This case is referred to by Hanson, Kashyap, and Stein (2011) and further discussed by Hirtle, Schuermann, and Stiroh (2009).ReferencesDiamond, D. W., and P. H. Dybvig. 1983. “Bank Runs, Deposit Insurance, and Liquidity.” Journal of Political Economy 91 (3): 401–19.First citation in articleLinkGoogle ScholarHanson, S. G., A. K. Kashyap, and J. C. Stein. 2011. “A Macroprudential Approach to Financial Regulation.” Journal of Economic Perspectives 25 (1): 3–28.First citation in articleGoogle ScholarHirtle, B., T. Schuermann, and K. Stiroh. 2009. “Macroprudential Supervision of Financial Institutions: Lessons from the SCAP.” Federal Reserve Bank of New York Staff Report 409.First citation in articleGoogle Scholar Previous articleNext article DetailsFiguresReferencesCited by Volume 9, Number 12013 Article DOIhttps://doi.org/10.1086/669595 Views: 116 © 2013 by the National Bureau of Economic Research Crossref reports no articles citing this article.