Professor Allan Meltzer famously quipped that “capitalism without failure is like religion without sin”. If some firms are protected from failure when they cannot pay their bills, then competition is skewed to favour inefficient, protected firms. Banks whose debts are guaranteed by the state receive an unfair advantage that enables them to allocate funds inefficiently, recklessly pursue risks at the expense of taxpayers, and waste resources that would be better used by firms operating without such protection.
The financial crisis of 2007–2009 wasn’t the first to illustrate that protected banking systems tend to blow up, imposing huge losses on taxpayers who are left to foot the bill. In the past three decades alone, there have been over a hundred major banking crises worldwide (Laeven and Valencia 2012). There is no topic in financial economics that has achieved a clearer consensus among researchers than the proposition that government protection of banks has contributed to the recent wave of costly bank failures around the world – failures on a scale that has never been witnessed before.
Anat Admati and Martin Hellwig’s recent book, The Bankers’ New Clothes (Princeton 2013) proposes to force banks to maintain much more of their financing in the form of equity rather than debt, so that bank shareholders rather than taxpayers will bear most or all of the downside risk of bank losses. In their well-intentioned zeal to make the case for how beneficial, simple, and costless it would be to mandate dramatic increases in bank equity ratios, Admati and Hellwig overstate the benefits and understate the costs associated with this proposed reform.
Book equity ratios vs. true risk-weighted equity ratios
Admati and Hellwig assert that accomplishing a credible increase in the proportion of bank equity capital is a simple matter of increasing minimum regulatory requirements for the ratio of the book value of equity relative to assets. Would that it were so simple, but it is not; increasing the book equity ratio in an accounting sense does not necessarily increase true bank capital ratios, as I argue in my recent work (Calomiris 2013). Bank balance sheets do not capture many of the economic losses that banks may incur. Also, accounting practices can disguise the magnitude of loan losses, and regulators eager to avoid credit crunches are often complicit in doing so. The result is that banks’ true equity ratios can be much lower than their book values indicate. Furthermore, banks’ risk choices matter, not just their equity. Both the Basel approach to risk weighting of assets and the simpler approach the authors advocate (that would abandon all risk weighting in favour of a simple equity-to-assets requirement) have a common flaw – they encourage banks to pursue hidden increases in asset risk.
For all these reasons, increasing required book-equity ratios does not necessarily translate into reducing the risk of bank failure. That does not mean that equity ratios are irrelevant; only that requiring increased book equity does not, by itself, result in higher true equity. Nor, and more importantly, do higher equity requirements ensure that banks will have higher equity relative to their risk, which is the essential goal of the regulatory reform that Admati and Hellwig envision.
Admati and Hellwig also argue that raising the ratio of equity finance in the structure of bank liabilities has few if any social costs. They dismiss the possibility that higher equity requirements for banks might be socially costly as a “bugbear…as insubstantial as the emperor’s new clothes in Andersen’s tale.” The authors go on to say, “For society, there are in fact significant benefits and essentially no cost from much higher equity requirements.” Such a policy would resolve the “fundamental conflict between what is good for banks and what is good for the broader economy.”
Cost of equity vs. risk-adjusted returns
These statements fail to represent the findings of decades of research encompassing scores of theoretical and empirical contributions in the banking and corporate finance literature. The key academic sleight-of-hand made by the authors, which is the basis for these statements, is to focus attention solely on the risk-adjusted returns expected by investors when discussing the risk-adjusted costs to banks of their capital structure choices. Admati and Hellwig incorrectly equate the two. “The cost of equity,” the authors claim, “essentially corresponds to the returns that corporations must provide to shareholders to justify the money it has received from them.” But for the banks that issue that equity, there are almost certain to be other important costs (and benefits) associated with capital structure choices that are only indirectly related to the returns expected and received by investors. And for this reason, the costs to a bank of issuing equity and the expected return received by equity investors who buy the new offering are not generally the same.
Differences between investors’ expected returns and firms’ financing costs have been shown to imply that, in general, there will be an optimal combination of debt and equity for each bank (or any other firm), which reflects a variety of considerations. One class of models focuses on the effect of the deductibility of interest payments on the optimal combination of debt and equity – firms balance the tax advantage of debt against the value preserved by holding more equity and thereby limiting the risk of financial distress. Another class of ‘signalling’ models considers how equity issuance can have adverse effects on market perceptions of firms’ investment opportunities, and lead issuers to avoid equity offerings more than they otherwise would. In still another class of models, choosing the right combination of debt and equity leads to efficient transfers of control to creditors under certain states of the world, which also encourages portfolio diversification and truthful revelation of investment outcomes, which reduce funding costs. In a fourth class of models, the right combination of equity and debt can provide incentives to manage risk more efficiently, which also reduces funding cost. Finally, in the context of banking, issuing very low-risk, short-term debt instruments in combination with sufficient equity can provide non-pecuniary liquidity benefits to the holders of the debt (especially depositors), which increases demand for the debt and allows bankers to save on funding costs.
Costs of higher bank equity requirements
Admati and Hellwig’s discussion of bank funding costs and capital structure recognizes only two benefits of debt finance: the tax deductibility of interest, and the safety-net distortions stemming from government guarantees that effectively reduce banks’ costs of subordinated debt as well as deposits. They argue that eliminating these advantages of debt finance is desirable. That claim neglects substantial empirical evidence consistent with other influences, such as signalling models. But even if tax favoured treatment of debt and safety net subsidies were the only factors favouring debt finance, and even if one could argue from a social cost-benefit analysis that it would be desirable to eliminate both safety-net subsidies and the tax deductibility of interest, it does not follow that doing so is costless.
An important implication of the various models of optimal capital structure is that forcing banks to raise their equity-to-asset ratio requirement generally will reduce banks’ willingness to lend. A large number of studies have shown that, when banks need to raise their equity-to-asset ratios, they often choose to do so by cutting back on new loans, which avoids the need to raise new equity and the high costs associated with it. For example, one recent study of the loan supply response to increases in required equity ratios in the UK reports that a one percentage-point increase in required equity ratios reduces the supply of lending to domestic nonfinancial firms by about 7% (implying an elasticity of loan supply of roughly negative 0.7).
The reduction in loan supply that comes from raising equity ratios is not just a one-time cost. A higher required equity ratio will mean that, as the banking system grows, a larger percentage of bank equity will have to be raised externally rather than through the retention of earnings. Because it is costly to raise outside equity (in large part because of the signalling and agency costs mentioned earlier), banks will face permanently higher funding costs, which in turn will permanently reduce the supply of lending relative to a world with lower equity ratio requirements.
Finding the right bank equity requirement
The existence of social costs associated with higher equity requirements does not rule out the desirability of a substantial increase in equity requirements. Indeed, most economists (including me) would be willing to accept some reduction in the supply of credit in return for the benefits of achieving greater financial stability, particularly given the current low equity ratios that banks maintain.
What is the right equity ratio to target, and what is the basis for the 25% equity-to-asset ratio proposed by Admati and Hellwig? After all, if they really believed their argument that raising the equity ratio can never have a cost, then why not advocate a 100% equity ratio?
The main basis for Admati and Hellwig’s recommendation of a 25% ratio is their view that historical experience shows that, prior to safety net protection, banks maintained that level of equity ratios. But Admati and Hellwig are too glib when making these historical comparisons, and they fail to note some important differences between banks then and now. Bank equity ratios, both in the US and abroad, have varied markedly in the past, and were not generally as high as 25% of assets. Some of the most stable banking systems – Canada’s, for example – have had relatively low equity ratios. The low equity ratios of Canadian nationwide branching banks reflected their greater portfolio diversification and other risk-lowering attributes in contrast to the much riskier single-office (unit) banks in the US. The equity ratios of US banks have varied dramatically over time, and in ways that have clearly reflected changes in their asset risk. Equity ratios relative to asset risk are the key attribute of interest in prudential regulation, not equity ratios per se. Using simple historical equity ratios from some past example as a benchmark, without taking risk into account, can significantly overstate or understate the extent to which current equity ratios of large, global banks should be increased.
I support substantially raising book equity ratio requirements, albeit by considerably less than proposed by Admati and Hellwig. In my view, raising equity, although costly, is worth the costs because the benefits of a stable banking system exceed the costs of reduced loan supply that would attend the increase in required equity ratios. My approach to reform would raise required equity to roughly 10% of assets, and would also ensure that banks maintain that ratio in actual equity relative to risk (not just book equity). Because simply mandating an increase in book equity requirements does not ensure a commensurate increase in true equity requirements, or in true equity relative to risk, higher equity ratio requirements need to be accompanied by several other measures – in particular, as I have argued elsewhere, by a market-value-triggered convertible contingent debt (CoCo) requirement. Although it is beyond our scope here to explain the logic behind this proposed requirement, the point of requiring a substantial amount of these CoCos is to create strong incentives for banks to maintain true equity at least as great as book equity, and to limit their risks so that a 10% equity ratio would be adequate.
The author is Henry Kaufman Professor of Financial Institutions at Columbia University and a visiting fellow in the research department of the International Monetary Fund. Disclaimer: The views expressed here are those of the author and do not necessarily represent those of the institutions with which he is affiliated.
Admati, Anat and Martin Hellwig (2013), The Bankers’ New Clothes, Princeton: Princeton University Press.
Calomiris, Charles W. (2013), “Reforming Banks Without Destroying Their Productivity and Value,” Journal of Applied Corporate Finance, 25, No. 4.
Laeven, Luc and Fabián Valencia (2012), “Systemic Banking Crises Database: An Update”, IMF Working Paper WP/12/163, June.