«Anil K Kashyap University of Chicago Booth School of Business Jeremy C. Stein Harvard University Economics Department Samuel Hanson Harvard ...»
An Analysis of the Impact of “Substantially Heightened” Capital
Requirements on Large Financial Institutions*
Anil K Kashyap
University of Chicago Booth School of Business
Jeremy C. Stein
Harvard University Economics Department
Abstract: We examine the impact of “substantially heightened” capital requirements on
large financial institutions, and on their customers. Our analysis yields three main conclusions.
First, the frictions associated with raising new external equity finance are likely to be greater than the ongoing costs of holding equity on the balance sheet, implying that the new requirements should be phased in gradually. Second, the long-run steady-state impact on loan rates is likely to be modest, in the range of 25 to 45 basis points for a ten percentage-point increase in the capital requirement. Third, due to the unique nature of competition in financial services, even these modest effects raise significant concerns about migration of credit-creation activity to the shadow-banking sector, and the potential for increased fragility of the overall financial system that this might bring. Thus to avoid tilting the playing field in such a way as to generate a variety of damaging unintended consequences, increased regulation of the shadowbanking sector should be seen as an important complement to the reforms that are contemplated for banks and other large financial institutions.
* This research was funded by The Clearing House Association L.L.C.
I. Introduction and Executive Summary The purpose of this paper is to examine the impact of the imposition of “substantially heightened” capital requirements on large financial institutions, and on their customers. We focus almost exclusively on the cost side of the equation, taking as given that, all else equal, increased capital requirements offer a benefit in terms of increased systemic stability. Our analysis yields three principal conclusions.
1. Phase-in effects: In the short-to-medium run, banks may be reluctant to seek new external equity to comply with increased capital-ratio requirements. Thus if the new requirements are phased in too rapidly, banks may opt to meet these requirements by slowing the growth of their assets, rather than by adding new capital. This can lead to a contractionary effect on lending activity. A large body of empirical work, which we survey below, suggests that these short-to-medium run effects can be quantitatively significant. Thus there is a strong case to be made for phasing in the new requirements sufficiently gradually that banks can generate the necessary additional capital largely out of retained earnings.
2. Long-run steady-state effects: Much less is known about the long-run steady-state effects on lending activity associated with higher capital requirements. We argue that existing empirical data and techniques do not allow us to make a meaningful direct estimate of these effects. Thus we are left to make predictions based on calibrating a particular economic model, and our predictions are only as good as the assumptions that underlie this model. Using the wellknown framework of Modigliani and Miller (1958), where the primary differences in the costs of debt and equity finance are due to differences in their tax treatment, we estimate that even proportionally large changes in the capital requirement are likely to lead to small long-run impacts on the borrowing costs faced by banks’ customers. For example, even if the minimum capital ratio is raised by ten percentage points, our methodology suggests that loan rates will increase by something on the order of just 25-45 basis points.1 Again, however, the caveat is that these estimates are model-dependent, and cannot be directly validated with a precise empirical experiment.
Interestingly, our qualitative conclusions here are close to those reached in recent work by Elliott (2009, 2010), in spite of the fact that our calibration methodologies are quite different.
3. Competition, regulatory arbitrage, and unintended consequences: The above conclusions regarding long-run effects may appear surprising, even paradoxical. If significant increases in capital ratios have only small consequences for the rates that banks charge their customers, why do banks generally feel compelled to operate in such a highly-leveraged fashion, in spite of the obvious risks this poses? By contrast, non-financial firms tend to operate with much less leverage than financial firms, and indeed often appear willing to forego the tax (or other) benefits of debt finance altogether.
We argue that the resolution of this puzzle has to do with the unique nature of competition in financial services. Unlike in many other industries, the most important (and in some cases, essentially the only) competitive advantage that banks bring to bear for many types of transactions is the ability to fund themselves cheaply. Thus if Bank A is forced to adopt a capital structure that raises its cost of funding relative to other intermediaries by only 20 basis points, it may lose most of its business.2 On the one hand, this line of reasoning suggests that substantially heightened capital requirements have the potential to meaningfully increase overall social welfare—by stopping a systemically-dangerous form of competition, with only a relatively small adverse effect on the ultimate item of interest, namely loan rates. However, the danger is that, in the face of higher capital requirements, these same forces of competition are likely to drive a greater volume of traditional banking activity into the so-called “shadow banking” sector. For example, perhaps an increasingly large fraction of corporate and consumer loans will be securitized, and in their securitized form will end up being held by a variety of highly-leveraged investors (say hedge funds) who are not subject to the usual bank-oriented capital regulation. If so, the individual regulated banks may be left safer than they were before, but the overall system of credit creation may not.
The thrust of this argument is not that capital requirements should not be raised for large banking firms. Rather, it is that in doing so, very careful attention must be paid to not tilting the playing field in such a way as to generate a variety of damaging unintended consequences. As we argue in more detail below, this is likely to involve increased regulation of the shadow banking sector as a complement to the reforms that are contemplated for banks and other large Contrast this example with, say the auto industry, where cheap financing is only one of many possible sources of advantage: a strong brand, quality engineering and customer service, control over labor and other input costs, etc., may all be vastly more important to profitability than a 20 basis point difference in the cost of capital.
financial institutions. In particular, we suggest that it would be a good idea to establish regulatory minimum margin (or “haircut”) requirements on asset-backed securities, so that any investor who takes a long position in credit assets, irrespective of their identity, cannot do so with an arbitrarily high degree of leverage.
The analysis that supports these conclusions consists of the following parts.
Section II: Conceptual framework: why is equity capital costly?
We begin by discussing the primitive forces that make equity capital more expensive to banks than either short or long-term debt finance. A crucial distinction here is that between stock and flow costs of equity finance, or equivalently between balance-sheet and new-issuance costs. Stock costs are factors like taxes and agency conflicts that make equity capital more expensive to a bank on an ongoing basis, no matter how the equity comes to be on the balance sheet (i.e., even if the equity is accumulated over time via retained earnings). Stock costs thus create a permanent wedge when regulatory capital requirements are raised, irrespective of the length of any phase-in period. By contrast, flow costs are costs specifically associated with the process of raising new external equity, and hence can be avoided via an appropriate phase-in period that allows banks to gradually accumulate equity capital via retained earnings. An oftcited source of flow costs is asymmetric information: firms don’t like to issue new public equity because this can be interpreted by the market as a negative signal, and thereby knock down their stock price.
The distinction between stock and flow costs is important in interpreting the empirical evidence. For example, a variety of data (to be discussed in Section III) make it clear that when banks experience adverse shocks to their capital, they cut lending. This evidence is thus pretty decisive in establishing that flow costs are important—i.e., that they inhibit banks from immediately offsetting a shock to capital by raising new equity externally, and therefore lead to significant effects on lending activity. But the evidence does not speak nearly as clearly to the magnitude of stock costs. Thus it is harder to make the empirical case that a properly phased-in increase in capital requirements will lead to a large permanent increase in banks’ weighted average cost of capital.
Section III. Evidence on the effects of shocks to bank capital
We survey the large empirical literature that shows how adverse shocks to bank capital can impact lending and broader economic activity. As emphasized above, this evidence is more convincing in establishing the importance of flow costs of equity, as opposed to stock costs: it clearly suggests that jacking up capital requirements in too much of a hurry could be damaging, but it would be a stretch to use it to argue that, in a long-run steady-state, permanently higher capital ratios would have a major effect on lending.
Section IV. Long-run steady-state effects
A. Calibration of magnitudes. As noted above, it is difficult to predict the long-run steady-state impact of an increase in capital requirements on loan rates based simply on an analysis of historical data. As an alternative, we adopt a model-based calibration approach. This approach requires us to take a stand on: i) what source of financing is displaced at the margin when a bank increases its level of equity capital; and ii) what the effective net cost difference is between equity and this alternative source.
In our baseline calibration, we assume that an increase in capital requirements leads banks to replace long-term debt financing with equity. This seems like a reasonable assumption, particularly given that for the largest banks, long-term debt as a fraction of assets is roughly 14%—suggesting that there is plenty of room for long-term debt to absorb the regulatory change without forcing banks to alter their use of (ostensibly cheaper) short-term debt.3 Moreover, in the spirit of Modigliani-Miller (1958), we assume that the only net difference in financing costs associated with this change in capital structure comes from the tax advantage of debt. This leads us to estimate that each one-percentage point increase in capital requirements raises the rate on bank loans by 2.5 basis points. Thus even a relatively radical ten-percentage point increase in the capital ratio would lead to only a 25 basis-point increase in loan rates.
In an alternative, more aggressive calibration, we assume that an increase in capital requirements leads banks to replace short-term wholesale debt financing with equity. (This assumption is perhaps more naturally interpreted as capturing the combined effects of a change in capital requirements along with a new liquidity requirement that caps the ratio of short-term As of 2009Q4, large U.S. BHCs – those with more than $10 billion in assets – together controlled $14.555 trillion in assets and had $2.057 trillion in long-term debt outstanding. Long-term debt includes the sum of other borrowed money with a remaining maturity of one year of more, subordinated notes, and trust preferred securities.
bank debt to total debt.) We also assume that, in addition to the tax advantage, short-term debt is cheaper for a bank than long-term debt because it offers investors a liquidity and/or surety advantage that they are willing to pay a premium for. Taken together, and erring on the side of overstating magnitudes, these two assumptions together imply that each one-percentage point increase in capital requirements now raises the rate on bank loans by 3.5 basis points. So the effect in the more aggressive scenario is clearly stronger, though still small in absolute terms.
Again, these estimates are only as good as the model that underlies them. The key assumption that we rely on in adopting the Modigliani-Miller approach is that when a bank shifts to a capital structure with more equity financing, its cost of equity falls, so that the overall effect on its weighted average cost of capital is less than would obtain in a world in which the cost of equity was a fixed constant, independent of capital structure. The reasoning is that as a bank delevers, its equity becomes less risky, so investors should rationally demand a lower risk premium for holding the equity. Although it is hard to fully validate this assumption empirically, we are able to present some supporting evidence, showing that the stock returns of less-levered banks do indeed tend to exhibit both less volatility and a lower beta with respect to the aggregate stock market.
B. Bank capital ratios across time and space. We present some descriptive statistics on bank capital ratios at different periods in history, and for banks of different types. The striking observation here is that there have been times when banks have operated with much more capital than they do today. Moreover, there does not seem to be any significant correlation in the timeseries data between capital ratios and various measures of lending spreads, such as the difference between the prime lending rate and the Treasury-bill rate. While obviously crude, these findings fit with the conclusion from our calibration exercise, namely that higher capital ratios need not imply large effects on loan rates.