A Macroprudential Approach to Financial Regulation

Journal of Economic Perspectives—Volume 25, Number 1—Winter 2011—Pages 3–28

M

any observers have argued that the regulatory framework in place

prior to the global  nancial crisis was de cient because it was largely

“microprudential” in nature (Crockett, 2000; Borio, Fur ne, and Lowe,

2001; Borio, 2003; Kashyap and Stein, 2004; Kashyap, Rajan, and Stein, 2008;

Brunnermeier, Crockett, Goodhart, Persaud, and Shin, 2009; Bank of England,

2009; French et al., 2010). A microprudential approach is one in which regulation

is partial equilibrium in its conception and aimed at preventing the costly failure of

individual  nancial institutions. By contrast, a “macroprudential” approach recog-

nizes the importance of general equilibrium effects, and seeks to safeguard the

 nancial system as a whole. In the aftermath of the crisis, there seems to be agree-

ment among both academics and policymakers that  nancial regulation needs to

move in a macroprudential direction. For example, according to Federal Reserve

Chairman Ben Bernanke (2008):

A Macroprudential Approach to

Financial Regulation

■

Samuel G. Hanson is a Ph.D. Candidate in Business Economics, Harvard University,

Cambridge, Massachusetts. From February to December 2009, Hanson was a Special Assistant

at the U.S. Department of the Treasury, Washington, D.C. Anil K Kashyap is Edward Eagle

Brown Professor of Economics and Finance and Richard N. Rosett Faculty Fellow, University of

Chicago Booth School of Business, Chicago, Illinois. Jeremy C. Stein is Moise Y. Safra Professor

of Economics, Harvard University, Cambridge, Massachusetts. From February to July 2009,

Stein was Senior Advisor to the Secretary, U.S. Department of the Treasury, and Staff Member

of the National Economic Council, both in Washington, D.C. Kashyap and Stein are both

Research Associates, National Bureau of Economic Research, Cambridge, Massachusetts.

Their e-mail addresses are

〈

[email protected]vard.edu

〉

,

〈

[email protected]

〉

,

and

〈

[email protected]

〉

.

doi=10.1257/jep.25.1.3

Samuel G. Hanson, Anil K Kashyap, and

Jeremy C. Stein

4 Journal of Economic Perspectives

Going forward, a critical question for regulators and supervisors is what their

appropriate “ eld of vision” should be. Under our current system of safety-

and-soundness regulation, supervisors often focus on the  nancial conditions

of individual institutions in isolation. An alternative approach, which has been

called systemwide or macroprudential oversight, would broaden the mandate

of regulators and supervisors to encompass consideration of potential systemic

risks and weaknesses as well.

In this paper, we offer a detailed vision for how a macroprudential regime

might be designed. Our prescriptions follow from a speci c theory of how modern

 nancial crises unfold and why both an unregulated  nancial system, as well as one

based on capital rules that only apply to traditional banks, is likely to be fragile. We

begin by identifying the key market failures at work: why individual  nancial  rms,

acting in their own interests, deviate from what a social planner would have them

do. Next, we discuss a number of concrete steps to remedy these market failures. We

conclude the paper by comparing our proposals to recent regulatory reforms in the

United States and to proposed global banking reforms.

Theories of Financial Regulation

Microprudential Regulation

Traditional microprudential regulation of banks is based on the following

logic. Banks  nance themselves with government-insured deposits. While deposit

insurance has the valuable effect of preventing runs (Diamond and Dybvig, 1983;

Bryant, 1980), it creates an incentive for bank managers to take excessive risks,

knowing that losses will be covered by the taxpayer. The goal of capital regulation

is to force banks to internalize losses, thereby protecting the deposit insurance

fund and mitigating moral hazard. Thus, if the probability of the deposit insurer

bearing losses is reduced to a low enough level, microprudential regulation is

doing its job.

To be speci c, consider a bank with assets of $100 that is  nanced with insured

deposits and some amount of capital. Suppose that the regulator can check on the

bank once a quarter. Suppose further that the volatility of the bank’s assets is such

that with probability 99.5 percent, the assets do not decline in value by more than

6 percent during a quarter. Then if the goal of policy is to reduce the probability

of bank failure (whereby capital is wiped out and there are losses to the deposit

insurance fund) to 0.5 percent, this goal can be accomplished by requiring the bank

to have capital equal to 6 percent of its assets as a cushion against losses. Notice

that in this setting, the exact form of the capital cushion is not important. It can be

common equity, but it can equally well be preferred stock, or subordinated debt, as

long as these instruments are not explicitly or implicitly insured—that is, as long as

they will in fact bear losses in a bad state.

Samuel G. Hanson, Anil K Kashyap, and Jeremy C. Stein 5

An important element of existing capital regulation is the presumption that

a bank will take immediate steps to restore its capital ratio in the wake of losses.

Returning to our example, suppose the bank starts out with capital of $6 but then

over the next quarter experiences losses of $2, so that its capital falls to $4. If the

volatility of its assets remains unchanged, in order for its probability of failure over

the subsequent quarter to stay at 0.5 percent, it would need to bring its capital ratio

back up to 6 percent. It could do so in one of two ways: either by going to the market

and raising $2 of fresh capital, or by leaving its capital unchanged and shrinking its

asset base to $66.67 (4/66.67

=

6 percent).

The basic critique of microprudential regulation can be understood as follows.

When a microprudentially oriented regulator pushes a troubled bank to restore

its capital ratio,

its capital ratio, the regulator does not care whether the bank adjusts via the numerator or

via the denominator—that is, by raising new capital or by shrinking assets

. Either way, the

bank’s probability of failure is brought back to a tolerable level, which is all that a

microprudential regulator cares about.

Such indifference to the method of adjustment makes sense if we are consid-

ering a single bank that is in trouble for idiosyncratic reasons. If that bank chooses

to shrink its assets—perhaps by cutting back on lending—others can pick up the

slack. Indeed, asset shrinkage in this case can be part of a healthy Darwinian process,

whereby market share is transferred from weaker troubled institutions to their

stronger peers. However, if a large fraction of the  nancial system is in dif culty, a

simultaneous attempt by many institutions to shrink their assets is likely to be more

damaging to the economy.

Macroprudential Regulation

In the simplest terms, one can characterize the macroprudential approach to

 nancial regulation as

 nancial regulation as an effort to control the social costs associated with excessive balance-

sheet shrinkage on the part of multiple  nancial institutions hit with a common shock

. To

make a compelling case for macroprudential regulation, two questions must be

answered. First, what are the costs imposed on society when many  nancial  rms

shrink their assets at the same time? Second, why do individual  rms not internalize

these costs? That is, why do they not raise fresh capital rather than reduce assets

when a bad shock hits? Or alternatively, why do they not build suf ciently large

capital buffers ahead of time so that they can withstand a shock without needing

either to raise capital or to reduce assets?

Generalized asset shrinkage has two primary costs: credit-crunch and  re-sale

effects. If banks shrink their assets by cutting new lending, operating  rms  nd

credit more expensive and reduce investment and employment, with contractionary

consequences for the economy. If a large number of banks instead shrink their

assets by all dumping the same illiquid securities (think of toxic mortgage-backed

securities) the prices of these securities can drop sharply in a “ re sale” of the sort

described by Shleifer and Vishny in this issue. Moreover, the  re-sale and credit-

crunch effects are intimately connected (Diamond and Rajan, 2009; Shleifer and

6 Journal of Economic Perspectives

Vishny, 2010; Stein, 2010a). If a toxic mortgage security falls in price to the point

where it offers a (risk-adjusted) 20 percent rate of return to a prospective buyer, this

will tend to drive the rate on new loans up towards 20 percent as well—since from

the perspective of an intermediary that can choose to either make new loans or buy

distressed securities, the expected rate of return on the two must be equalized. In

other words, in market equilibrium, the real costs of  re sales manifest themselves

in the further deepening of credit crunches. Ivashina and Scharfstein (2010) offer

evidence on the extent of credit contraction during the recent crisis.

Of course, to make a case for regulatory intervention, one has to explain why

these 20 percent rates of return inside the  nancial sector—which are much higher

than the outside rates on, say, Treasury securities—don’t naturally draw in enough

private capital to eliminate the return differentials. One reason why capital is immo-

bile once a crisis is underway is the “debt overhang” problem identi ed by Myers

(1977). Once a bank is in serious trouble and its debt is impaired in value, the

bank is reluctant to raise new equity even to fund investments that have a positive

net present value. This is because much of the value that is created is siphoned off

by the more senior creditors. Given the debt overhang problem, banks that act in

the interests of their shareholders will tend to  x their damaged capital ratios by

shrinking assets rather than by raising new capital, even when the latter is more

desirable from a social perspective.

If so, why don’t banks voluntarily build up adequate buffer stocks of excess

capital in good times, when debt overhang is not yet a concern, so they can absorb

losses in bad times without having to either shrink assets or raise new capital under

duress? After all, such a dry-powder strategy would allow them to exploit pro table

opportunities should a crisis arise. This question is addressed in Stein (2010a), who

extends the  re-sale model to consider banks’ initial choices of capital structure.

He shows that if short-term debt is a cheaper form of  nance than equity, banks

will tend to take on socially excessive levels of debt: while the banks capture the

bene ts of cheap debt  nance, they do not internalize all of its costs.

1

In particular,

when Bank A takes on more debt, it does not account for the fact that by doing

so, it degrades the collateral value of any assets it holds in common with another

Bank B—since in a crisis state of the world, A’s  re-selling of its assets lowers the

liquidation value that B can realize for these same assets.

2

1

The assumption that short-term debt is a cheap form of  nance represents a particular deviation from

the Modigliani–Miller (1958) capital-structure irrelevance framework. In the context of  nancial  rms,

this deviation can arise to the extent that their short-term claims are “money-like” and carry a premium

that re ects their usefulness as a transactions medium. We discuss this point in greater detail below.

2

A subtlety is that this is a pecuniary externality—that is, it works through prices. For a pecuniary

externality to cause a misallocation of resources, one requires a departure from the standard assump-

tions that deliver the fundamental welfare theorems (Geanakoplos and Polemarchakis, 1986). In Stein

(2010a), this departure is in the form of a collateral constraint: banks’ ability to raise short-term debt is

constrained by the collateral value of their assets.

A Macroprudential Approach to Financial Regulation 7

In sum, a model based on  re sales and credit crunches suggests that  nancial

institutions have overly strong incentives 1) to shrink assets rather than recapitalize

once a crisis is underway, and 2) to operate with too thin capital buffers before a

crisis occurs, thereby raising the probability of an eventual crisis and systemwide

balance-sheet contraction. Therefore, the macroprudential approach to capital

regulation aims to counterbalance these two tendencies. With this in mind, we turn

next to some of the individual items in the macroprudential toolkit.

Before doing so, however, we should emphasize that, in contrast to the tradi-

tional view,

tional view, nothing in this alternative theory relies on the existence of deposit insurance

.

In other words, in a model of crises based on  re sales, there is socially excessive

balance-sheet shrinkage, and a rationale for regulation, even absent government

deposit insurance. Thus, there is a strong presumption that macroprudential regu-

lation should apply to more than just insured deposit-takers. The broader point

(stressed by Tucker, 2010, and Kashyap, Berner, and Goodhart, forthcoming) is

that regulators need to pay attention to all the channels through which the actions

of  nancial institutions—both those who are insured and those who are not—can

cause damage.

Macroprudential Tools

We now discuss six sets of tools that can be helpful in implementing a macro-

prudential approach to  nancial regulation. Our goal here is not to provide a

comprehensive laundry list of reform proposals, but rather to show how a particular

conceptual framework provides a uni ed way of thinking about what otherwise

might seem like a hodgepodge of different  xes.

As a prelude, note that if the goal of regulation is to prevent  nancial  rms

from shrinking their balance sheets excessively in an adverse state of the world, a

simple accounting identity imposes a lot of discipline on our thinking. In particular,

when a bank is hit with a shock that depletes its capital, there are only two ways to

prevent it from shrinking its assets: 1) it can raise new capital to replace that which

was lost; or 2) it can let its ratio of capital to assets decline. Many of the tools that we

discuss are just different mechanisms for facilitating adjustment on one of these two

margins. We start with capital proposals and then broaden the discussion to other

options that have heretofore been outside of the regulatory toolkit.

Time-Varying Capital Requirements

One intuitively appealing response to the problem of balance-sheet shrinkage

is to move to a regime of time-varying capital requirements, with banks being asked

to maintain higher ratios of capital to assets in good times than in bad times. Under

such a rule, banks can draw down their buffers when an adverse shock hits and

continue operating with less pressure to shrink assets. Kashyap and Stein (2004)

argue that time-varying capital requirements emerge as an optimal scheme in a

8 Journal of Economic Perspectives

model where the social planner maximizes a welfare function that weights both

1) the microprudential objective of protecting the deposit insurance fund and

2) the macroprudential objective of maintaining credit creation during recessions.

A planner concerned with both objectives should be willing to tolerate a higher

probability of bank failure in bad times, when bank capital is scarce and credit

supply is tight, than in good times.

One challenge in designing such a regime is that, in bad times, the regu-

latory capital requirement is often not the binding constraint on banks. Rather,

as the risk of their assets rises, the market may impose a tougher test on banks

than do regulators, refusing to fund institutions that are not strongly capitalized.

3

Table 1 shows that, as of the  rst quarter of 2010, the four largest U.S. banks had

an average ratio of Tier 1 common equity to risk-weighted assets of 8.2 percent and

an average ratio of total Tier 1 capital (including preferred stock, for example)

to risk-weighted assets of 10.7 percent. These are both well above the pre-crisis

regulatory standard, which required a ratio of total Tier 1 capital to risk-weighted

assets of 6 percent for a bank to be deemed “well capitalized.” Thus, even as the

U.S. economy was emerging from a deep  nancial crisis in early 2010, the regula-

tory constraint was nonbinding.

This pattern implies that to achieve meaningful time variation in capital ratios,

the regulatory minimum in good times must substantially exceed the market-imposed standard

in bad times

. Thus, if the market standard for equity-to-assets in bad times is 8 percent,

and we want banks to be able to absorb losses of, say, 4 percent of assets without

pressure to shrink, then the regulatory minimum for equity-to-assets in good times

3

This tendency may be ampli ed by the widespread use of “Value at Risk” (VaR) models by banks. As

measured volatility and hence VaR go up in bad times, such models mechanically call for banks to hold

higher ratios of capital. Thus, banks’ own internal risk management practices might compel shrinkage

even if market funding remains available.

Table 1

Capital Ratios for Top Four U.S. Banks, 2010Q1

Bank of

America Citigroup

JPMorgan

Chase

Wells

Fargo

Weighted

average

Total risk-weighted assets

($ millions)

1,519 1,023 1, 147 988

Tier 1 common equity to

risk-weighted assets (%)

7.6 9.1 9.1 7.1 8.2

Tier 1 capital to risk-weighted

assets (%)

10.2 11.2 11.5 10.0 10.7

Sources: Data is from the websites of individual banks.

Note: This table lists the capital ratios for the four largest U.S. banks as of 2010Q1.

Samuel G. Hanson, Anil K Kashyap, and Jeremy C. Stein 9

would have to be at least 12 percent. A loss on the order of 4 percent of assets is

actually less severe than the experience of the major banks during the recent crisis;

the IMF (2010) estimates that cumulative credit losses at U.S. banks from 2007 to

2010 were on the order of 7 percent of assets. Using this  gure, one could argue for

a good-times regulatory minimum ratio of equity to assets of 15 percent. Either way,

these are high values, signi cantly higher than obtained from a microprudential

calculation that asks only how much capital is needed to avert outright failure. (We

examine the potential costs of raising capital requirements by this much below, in

the penultimate section of the paper.)

Higher-Quality Capital

Traditionally, the capital metric given the most attention by regulators has

been the ratio of total Tier 1 capital to risk-weighted assets. In addition to common

equity, total Tier 1 capital includes, among other items, preferred stock. Thus, both

equity and preferred have “counted” in the same way towards satisfying capital

requirements. From a microprudential perspective, this makes perfect sense. If the

only concern is avoiding losses to the deposit insurer in the event of bank failure,

so long as both common and preferred holders are strictly junior in priority to the

deposit insurer, they will provide the desired loss-absorption cushion.

However, in the wake of the  nancial crisis, many investors and regulators

have discussed the “quality” of a bank’s capital base and how common stock is a

“higher-quality” form of capital than preferred. While this distinction is hard to

understand from a microprudential loss-absorption perspective, it  ows naturally

from the macroprudential approach, which focuses less on a static failure scenario

and more on enabling troubled institutions to recapitalize dynamically and remain

viable as going concerns. Common equity is more friendly to the recapitalization

process than preferred stock because it is more junior and hence less problematic

in terms of the debt overhang problem described above.

To see this point, consider two banks, A and B. Both begin with total assets of

$100 and total capital of $6. But A’s capital is composed entirely of equity, while B

has $2 of equity and $4 of preferred. Now suppose both banks lose $3. To avoid

shrinking their assets, they would like to raise new capital. Suppose they do so by

trying to issue equity. This will be harder for Bank B—whose entire pre-existing

equity layer has been wiped out and whose preferred stock is, as a result, now trading

at a steep discount to its face value—for any new equity that B brings in will largely

serve to bail out the position of its more senior preferred investors.

This logic suggests that given the goal of promoting rapid recapitalization by

going-concern banks that run into trouble, it is entirely reasonable for regulators

to require that most of the capital requirement be satis ed with common equity.

Indeed, one can argue that essentially

Indeed, one can argue that essentially all

of what is now the Tier 1 requirement

should be in terms of equity, or instruments that are contractually guaranteed to

convert into equity in a bad state (see below for a discussion), while more senior

securities like preferred stock should for the most part not count.

10 Journal of Economic Perspectives

Corrective Action Targeted at Dollars of Capital, Not Capital Ratios

When regulators are vigilant, banks that fall below a designated capital threshold

may be subject to a variety of sanctions (for example, restrictions on dividends)

until they repair their capital ratios. The principle of rapid regulatory intervention

is undoubtedly a good one, but the form of the intervention matters a great deal.

If a bank is put in the penalty box until it manages to  x its capital

If a bank is put in the penalty box until it manages to  x its capital ratio

, it may well

choose to  x the ratio not by raising the numerator (capital) but by reducing the

denominator (assets).

4

A better approach is to create incentives for the bank to raise

incremental dollars

of new capital.

One way to implement this policy would be with a capital ratio requirement that

refers to the

refers to the maximum

of current and lagged assets. Imagine a bank that starts with

assets of $100 and capital of $8 at the end of year

t

, and suppose that the threshold for

corrective action is a capital ratio of 6 percent. Now assume that the bank has losses of

$4 over year

$4 over year t

+

1, so it ends the year with $4 of capital. Normally regulators would push

the bank to get its ratio back to 6 percent, which it might do by shrinking its assets to

$66.67. Under our alternative, the bank would only get out of the penalty box when

its ratio of capital to the

its ratio of capital to the maximum of year t assets or year t

+

+ 1 assets

exceeded 6 percent.

Given that year

Given that year t

assets were $100 and cannot be reduced retroactively, the bank would

have to raise $2 of new capital—it could not avoid sanctions by shrinking assets.

A dramatic illustration of this dollars-based corrective action principle comes

from the Supervisory Capital Assessment Program (SCAP), the “stress tests” that

the major U.S. banks underwent in spring 2009.

5

The output of the SCAP was, for

each bank being tested, a

each bank being tested, a dollar

target for new equity capital that had to be raised,

via equity issues or asset sales. For some of the banks involved, the numbers were

very large—for example, Bank of America was required to raise $33.5 billion. The

penalty box in this case was that any bank failing to raise the capital from the private

markets would be required to accept an equity injection from the Treasury, which

would have involved strict limits on executive compensation. Remarkably, in the

few weeks following the release of the SCAP results, the banks involved were able to

raise nearly $60 billion in new common equity; by the end of 2009 this  gure had

risen to over $125 billion.

Here is a case where a strong regulatory hand appears to have had highly

bene cial effects. Indeed, by being tough and giving banks no choice, regulators

probably made it easier for banks to do the capital raising. This is because absent

discretion, the adverse selection problem normally associated with equity issues

disappears. If a bank has a choice of whether to issue equity, its decision to do so

may signal that management believes it to be overvalued, and hence this issuance

may knock down the stock price (Myers and Majluf, 1984). But if it has no choice,

4

Hart and Zingales (2010) recommend forcing banks to issue equity whenever their credit risk (as

measured by spreads on credit default swaps) goes above a certain level. However, this does not address

the shrinkage problem, because a bank can also reduce its credit spreads by selling risky assets.

5

See Hirtle, Schuermann, and Stiroh (2009) for a fuller discussion of the lessons learned from the SCAP.

A Macroprudential Approach to Financial Regulation 11

there is no information content, and hence no negative price impact. We hope that

this lesson can be incorporated into regulatory policy going forward. As we note

below, it may be especially helpful in thinking about the phase-in of higher capital

requirements under Basel III.

Contingent Capital

A dollars-based corrective action policy amounts to an attempt to force banks

to recapitalize on the  y when they get into trouble. A closely related idea is to “pre-

wire” the recapitalization with a contingent instrument that automatically increases

a bank’s equity position when some prespeci ed contractual provision is triggered.

Two broad types of contingent capital instruments have been proposed. The  rst,

sometimes called “reverse convertibles” or “contingent convertibles,” involves a bank

issuing a debt security that automatically converts into equity if a measure of either

the bank’s regulatory capital or stock market value falls below a  xed threshold

(Flannery, 2005; French et al., 2010).

6

For example, in November 2009, Lloyds Bank

issued £7.5 billion in contingent convertible debt, with conversion to equity to be

triggered if Lloyds’ Tier 1 capital ratio falls below 5 percent.

A second type of contingent capital is “capital insurance,” which involves a bank

purchasing an insurance policy that pays off in a bad state of the world (Kashyap,

Rajan, and Stein, 2008). To address concerns about the insurer defaulting, the policy

would be fully collateralized—that is, the insurer would put the full amount of the

policy into a lock-box up front. For example, a bank might contract with a pension

fund to buy a capital insurance policy that pays $20 billion in the event that an econo-

mywide index of bank stock prices falls below some designated value any time in the

next  ve years. At initiation, the pension fund would turn the $20 billion over to a

custodian; if the bad state is not realized within  ve years, the $20 billion reverts back

to the pension fund, and if it is realized, the funds are transferred to the bank.

These designs share a common motivation. The premise is that banks view

equity capital as an expensive form of  nance—in other words, there are one or

more violations of the Modigliani–Miller (1958) conditions that make banks reluc-

tant to carry large precautionary buffers of equity. (We discuss the precise nature

of these violations in detail below.) In principle, regulation could simply mandate

that banks maintain very large equity buffers. However, it may be more ef cient to

develop a  nancing arrangement that delivers more equity only in those bad states

where it is most valuable.

If these forms of contingent capital are such a good idea, why haven’t we seen

more of them? One simple answer is that they would have to be allowed to count

towards regulatory capital requirements. Consider the following approach. The

capital requirement for a bank in good times might be set at a relatively high level,

6

There are many important design issues associated with the speci cation of the trigger in a contingent

convertible security, with both pros and cons to using a trigger based on stock prices as opposed to

regulatory accounting numbers. See McDonald (2010) for a detailed discussion of these issues.

12 Journal of Economic Perspectives

say 20 percent. Banks would then be given a choice: they could satisfy the entire

requirement with equity, or they could satisfy up to say 10 percentage points of it

with a reverse convertible so long as it was contractually guaranteed to turn into

equity in a well-de ned bad state. (As we discuss below, Swiss banking regulators

recently announced new rules of exactly this form.) The reverse convertible might

be seen as more costly than straight debt—which is why banks would not use it if it

did not count as regulatory capital—but as long as it was cheaper than equity, there

would be an ef ciency gain.

Finally, it is worth noting the close connection between contingent capital and

proposals to reform executive compensation by imposing bonus holdbacks on key

employees of  nancial  rms. For example, French et al. (2010) suggest withholding

a signi cant share of each senior manager’s total compensation for several years.

The withheld compensation would not take the form of stock or options, but would

instead be a  xed dollar amount. And, managers would forfeit their holdbacks if

the  rm were to fail or to receive extraordinary government assistance.

Structurally, this holdback proposal is similar to the capital insurance scheme

of Kashyap, Rajan, and Stein (2008), with the key difference being that it requires

 rm managers—rather than, say, a pension fund—to be the insurance provider.

The merit of this approach is that not only does the held-back compensation

create an extra, contingent capital buffer, it also helps to improve incentives

within the  rm. In particular, by making insiders bear downside risk without any

additional upside potential, it aligns their fortunes with those of taxpayers and

other creditors—and in so doing, leans against the “heads I win, tails you lose”

risk-taking incentives created by more conventional forms of stock- and pro t-

linked compensation.

Regulation of Debt Maturity

One important lesson from the recent crisis is that the distinction between

short-term and long-term debt had been given insuf cient attention by regula-

tors.

Table 2 presents a snapshot of the aggregate liability structure of the U.S.

banking system, including not only traditional commercial banks but also broker-

dealer  rms. Clearly, the majority of their debt is short-term: either in the form of

deposits or “wholesale” funding, which includes commercial paper and repurchase

(repo) agreements. While deposits are generally insured and hence not likely to

run at the  rst sign of trouble, the same is not true for wholesale funding. Indeed,

wholesale funding runs—a refusal of repo and commercial paper creditors to roll

over their loans—played a key role in the demise of Northern Rock, Bear Stearns,

and Lehman Brothers, among other high-pro le failures (Shin, 2009; Gorton and

Metrick, 2010; Duf e, 2010).

The case for regulating the use of short-term debt by  nancial  rms—above

and beyond regulating total leverage—rests on two observations. First, the ability of

short-term lenders to run leads to more fragility than in the case with an equivalent

amount of long-term debt (Diamond and Dybvig, 1983). It is hard to imagine that

Samuel G. Hanson, Anil K Kashyap, and Jeremy C. Stein 13

Northern Rock, or Bear Stearns, or Lehman would have faced the same problems

had they done most of their borrowing on a long-term basis. Second, in the pres-

ence of marketwide  re sales, the choice of debt maturity creates an externality.

When an individual bank or broker-dealer opts to  nance largely with short-term

debt, it fails to internalize that in a crisis, an inability to roll over short-term debt

will force it to liquidate assets, thereby imposing a  re-sale cost on others who

hold the same assets and who see the value of their own collateral diminished. The

result is a level of short-term  nancing that is socially excessive—hence the role for

regulation (Stein, 2010a).

Regulating the Shadow Banking System

The  re-sale risk associated with excessive short-term funding comes from not

just insured depositories, but rather, any  nancial intermediary whose combina-

tion of asset choice and  nancing structure may exacerbate a systemic  re-sale

problem. A narrow interpretation of this principle would say that regulation should

cover large, systemically signi cant nonbank institutions such as Bear Stearns and

Table 2

Liability Structure of U.S. Bank Holding Companies, 2009

$ trillion % of assets

Assets 15.927 100.0%

Liabilities

Deposits 7.502 47.1%

Short-term wholesale funding

Repurchase agreements and federal funds purchased 1.658 10.4%

Other short-term wholesale funding 0.880 5.5%

Trading liabilities 0.736 4.6%

Total 3.274 20.6%

Long-term funding

Long-term wholesale funding 1.718 10.8%

Subordinated debt and trust preferred 0.416 2.6%

Total 2.134 13.4%

Other liabilities 1.570 9.9%

Total liabilities 14.480 90.9%

Equity

Common stock 1.309 8.2%

Preferred stock 0.137 0.9%

Total equity 1.446 9.1%

Sources: The table is based on data from the FR Y-9C reports that Bank Holding Companies are required

to  le with the Federal Reserve.

Note: This table summarizes the liability structure of U.S. bank holding companies as of December 31,

2009.

14 Journal of Economic Perspectives

Lehman Brothers, who did not  nance themselves with insured deposits but who

were nevertheless subject to wholesale  nancing runs. While this speci c point is

now well appreciated, the principle has broader application. From the perspective

of credit creation and macroeconomic impact, some of the most damaging aspects

of the crisis arose not just from the problems of individual large  rms, but also from

the

the collapse of an entire market

—namely the market for asset-backed securities.

Figure 1 illustrates this collapse.

7

The market for “traditional” asset-backed secu-

rities, those based on credit-card, auto, and student loans, averaged between $50 and

$70 billion of new issues per quarter in the years prior to the crisis (total issuance for

2007 was $238 billion). However, in the last quarter of 2008, following the demise

of Lehman, issues in this category fell to just over $2 billion. The disappearance of

this market represented a major contraction in the supply of credit to consumers.

The investors who buy tranches of asset-backed securities frequently do so

by relying on short-term borrowing. Entities known as “structured investment

vehicles” or “conduits,” which in the past tended to be af liated with sponsoring

7

The discussion in the remainder of this section is a much-abridged version of material in Stein (2010b).

Figure 1

Quarterly Issuance of Asset-Backed Securities, 2000–2010Q2

Source: The data underlying this  gure come from Thompson SDC.

Notes: The  gure plots the quarterly issuance of traditional versus nontraditional asset-backed securities

(ABS). Traditional ABS includes securitizations backed by auto loans, credit card receivables, and

student loans. Nontraditional issuance includes ABS backed by subprime mortgages, collateralized debt

obligations (CDOs), and collateralized loan obligations (CLOs). While the nontraditional category

includes securitizations backed by subprime mortgages, it does not include securitizations backed by

prime mortgages, such as mortgage-backed securities guaranteed by Fannie Mae or Freddie Mac.

350

300

$ billion

250

200

150

100

50

0

Mar. 00

Mar. 01

Mar. 02

Mar. 03

Mar. 04

Mar. 06

Mar. 05

Mar. 10

Mar. 09

Mar. 08

Mar. 07

Traditional (auto, credit cards, student loans)

Nontraditional (subprime, CDOs, CLOs)

A Macroprudential Approach to Financial Regulation 15

commercial banks, hold tranches of asset-backed securities and  nance them with

commercial paper, which typically has a maturity of only days or weeks. Hedge

funds and broker-dealer  rms may  nance their holdings of asset-backed securities

with repurchase agreements, a form of overnight collateralized borrowing. Collec-

tively, these various investors who acquire asset-backed securities and  nance them

with short-term debt are often referred to as the

with short-term debt are often referred to as the shadow banking system

. Moreover,

as emphasized by Gorton (2010), Gorton and Metrick (2010), and Covitz, Liang,

and Suarez (2009), the collapse of the asset-backed securities market featured the

essential elements of a classic bank run—namely an inability of investors in asset-

backed securities to roll over short-term  nancing.

One manifestation of the withdrawal of short-term lending to the asset-backed

securities market comes from the behavior of “haircuts” in repurchase agreements.

When an investor borrows in the repo market, the investor is required to post a

margin, or down payment, known as the “haircut.” Haircuts on most highly rated

asset-backed securities were very low prior to the crisis, on the order of 2 percent.

Thus, if a hedge fund wanted to buy $1 billion of AAA-rated, auto-linked asset-

backed securities, it only needed to put up $20 million of its own capital. The other

$980 million could be borrowed on an overnight basis in the repo market; in many

cases the ultimate lenders were money-market mutual funds.

8

In the midst of the crisis, haircuts skyrocketed. Even haircuts on consumer

asset-backed securities—which were not linked to subprime problems—rose to over

50 percent. From the perspective of the hedge fund holding $1 billion of such

securities, all of a sudden it could only borrow $500 million, and instead of having to

post a $20 million down payment, had to put up $500 million. If it did not have the

cash to do so, it would be forced to liquidate its holdings. These liquidations, and

the effect they had on the level and volatility of prices, in turn justi ed the increased

skittishness of the lenders in the repo market, since their protection depends on

the collateral value of the assets they lend against. In other words, the disruption to

the asset-backed securities market may have been what Brunnermeier and Pedersen

(2009) call a “margin spiral.”

From a macroprudential perspective, it would be a mistake to focus too

narrowly on the largest  nancial institutions while paying insuf cient attention

to potential vulnerabilities in the rest of the system.

9

What concrete steps might

be taken in this regard? A useful  rst principle is that an effort should be made to

impose similar capital standards on a given type of credit exposure irrespective of

8

This is not to say that the hedge fund’s overall leverage would be 50 to 1, as in this example—only that

it could borrow aggressively against certain highly rated assets that were seen as low risk and hence as

very good collateral.

9

Some have advocated a “narrow banking” model, whereby tightly regulated banks are restricted to

core deposit-taking and lending activities, and a substantial chunk of their other business is pushed out

to a more lightly regulated periphery. While such an approach may keep certain functions (such as the

payments system) safe, the risk is that the overall process of credit creation may be made more vulner-

able, not less, to shutdown in the event of a crisis.

16 Journal of Economic Perspectives

who winds up ultimately holding the exposure—be it a bank, broker-dealer, hedge

fund, or special purpose vehicle. This task is not easy, but one tool that would help

is broad-based regulation of haircuts on asset-backed securities.

10

Consider the case where the exposure is a consumer loan. If this loan is made

by a bank, it will be subject to a capital requirement. Now suppose instead that

the loan is securitized by the bank and becomes part of a consumer asset-backed

security whose tranches are distributed to investors. The regulation we have in mind

would stipulate that whoever holds a tranche of the asset-backed security would

be required to post and maintain a minimum haircut against that tranche—with

the value of the haircut depending on the seniority of the tranche, the quality of

the underlying collateral, and so forth. Such a requirement is nothing conceptually

new and should not be dif cult to enforce; indeed, it is closely analogous to the

initial and maintenance margin requirements that are currently applicable to inves-

tors in common stocks. For models that suggest a role for haircut regulation, see

Geanakoplos (2010) and Stein (2010a).

If these requirements are well-structured, they would have two bene ts. First,

they could help to harmonize regulation across organizational forms, thereby

reducing the incentive for lending activity to migrate into the shadow banking

sector. Second, for those assets that do end up in the shadow banking system,

haircut regulation can dampen the destabilizing dynamics described above. If hair-

cuts start out at 2 percent and then jump to 50 percent in a crisis, this creates a

powerful forced-selling pressure on owners of asset-backed securities. If haircuts are

set instead at a higher value before the crisis, this forced-selling mechanism and the

vicious spiral it unleashes might be attenuated. Note that central banks, through

their discount window and emergency-lending facilities, have already developed

signi cant expertise in determining prudent values of haircuts on various kinds of

asset-backed securities.

While we have focused on regulations that address  re-sale externalities, the

problems of the asset-backed securities market arguably go beyond  re sales. Hanson

and Sunderam (2010) show that the “tranching” process, by which large fractions

of underlying collateral pools are turned into AAA-rated securities, can blunt the

incentives of investors to become informed about what they are buying—because

AAA-rated securities are ostensibly so low risk that the returns to becoming informed

are minimal. A lack of informed investors can in turn make securitization markets

more fragile when times turn bad and the need to analyze securitization cash  ows

rises. This implies that regulators should worry about the structure of securitiza-

tions—particularly the amount of AAA-rated securities being manufactured from

any given collateral pool.

10

A more general version of this observation is that a regulatory toolkit with only a capital ratio and a

single liquidity ratio will be inadequate for controlling instability arising from deposit defaults,  re sales,

and credit crunches (Kashyap, Berner, and Goodhart, forthcoming).

Samuel G. Hanson, Anil K Kashyap, and Jeremy C. Stein 17

What Are the Costs of Higher Capital Requirements?

We have argued that a macroprudential approach involves imposing substan-

tially higher capital requirements on  nancial  rms, particularly in good times.

But will these higher capital requirements lead to increased costs for borrowers? In

what follows, we focus on the long-run steady-state consequences of higher capital

requirements, setting aside the transitional issues associated with phase-in of a new

regime.

11

To preview, our reading of the theory and relevant empirical evidence

suggests that while increased capital requirements might be expected to have some

long-run impact on the cost of loans, this effect is likely to be quite small.

A Modigliani–Miller Perspective

Modigliani and Miller (1958) famously showed that under certain conditions, a

 rm’s capital structure is irrelevant for its operating decisions. In the banking context,

this would imply that the rate that a  rm charges on its loans should be

this would imply that the rate that a  rm charges on its loans should be independent

of its capital ratio. The Modigliani and Miller conditions are stringent, including

no taxes, symmetric information, rational risk-based pricing, and cash ows that are

independent of  nancial policy. Thus, they are not meant as an accurate depiction

of reality. Rather, the value of the Modigliani and Miller framework is that it forces

one to be precise about which of the conditions is violated, and this allows for a more

disciplined analysis of the effects associated with changes in capital structure.

In particular, the Modigliani and Miller paradigm exposes the  aw in the

following reasoning: “Equity is more expensive than debt because it is riskier. Thus,

if a bank is forced to rely more on equity, its overall cost of  nance will go up, and

it will have to charge more for its loans.” The fallacy here is that the risk of equity,

and hence its required return, is not a constant, but rather declines as leverage

falls.

12

Indeed, when all the Modigliani and Miller conditions hold, this effect is just

enough to offset the increased weight of the more-expensive equity in the capital

structure so that the overall cost of capital

structure so that the overall cost of capital stays  xed

as bank leverage varies.

With this caveat in mind, we discuss two deviations from Modigliani and

Miller’s idealized conditions that are likely to be relevant in the present context.

First, interest payments on corporate debt are tax deductible while dividend

payments on equity are not. This effect lends itself to easy measurement. Suppose

that new equity capital displaces

that new equity capital displaces long-term

debt in a bank’s capital structure and

that the only effect on the bank’s weighted average cost of capital comes from the

lost tax shields on the debt. If the coupon on the debt is 7 percent, and given a

corporate tax rate of 35 percent, each percentage point of increased equity raises

11

This section draws on material from our unpublished working paper, Kashyap, Stein, and Hanson

(2010).

12

In Kashyap, Stein, and Hanson (2010), we show that this holds empirically in the banking sector: in a

panel of large banks, those with less leverage have signi cantly lower values of both beta and stock-return

volatility.

18 Journal of Economic Perspectives

the weighted average cost of capital by .07

×

.35

=

.0245 percent, or 2.45 basis

points. Thus, even a 10 percentage-point increase in the capital requirement only

boosts the weighted average cost of capital—and hence loan rates—by 25 basis

points, which is a small effect.

To generate a higher  gure, consider a case where equity displaces

To generate a higher  gure, consider a case where equity displaces short-

term

debt; this can be interpreted as capturing the joint effects of an increase

in both capital and liquidity requirements. Moreover, following Gorton (2010),

Gorton and Metrick (2010), and Stein (2010a), assume that—in violation of the

Modigliani and Miller conditions—there is a non-risk-based “money” premium

on wholesale short-term bank debt that re ects its usefulness as a transactions

medium. (Commercial paper and repo are often held by money-market mutual

funds, who in turn issue checkable deposits.) An upper-bound estimate of this

money premium might be on the order of 100 basis points.

13

Now, a 10 percentage-

point increase in capital requirements raises the weighted average cost of capital

by an added 10 basis points relative to the previous taxes-only case, and we are up

to 35 basis points. This number is still quite modest.

Time Variation in Bank Capital Ratios and Lending Rates

Our calibrations based on the Modigliani–Miller paradigm suggest that the

long-run effects of higher capital requirements on loan rates should be small. A

complementary approach is to examine the historical record. Figure 2A, which is

adapted from Berger, Herring, and Szego (1995), shows the ratio of book equity to

book assets for U.S. commercial banks from 1840 to 2009. Capital ratios exceeded

50 percent in the 1840s and fell steadily for the next 100 years, reaching 6 percent

by the 1940s. Have these large  uctuations in capital ratios translated into big differ-

ences in the cost of bank credit? To address this question, we have examined the

behavior of various proxies for the markup that banks charge on loans. In a variety

of regression speci cations (not shown here), we found no reliable time-series

correlation between these markup variables and bank capital ratios. The historical

data is simply too noisy, and our proxies for loan spreads too crude, for us to draw

any con dent conclusions about whether a correlation between equity ratios and

loan rates

loan rates even exists

.

To illustrate the loose ties between loan costs and capital ratios, Figure 2B

plots capital ratios for the period 1920–2009 against two markup proxies: 1) the

net interest margin (net interest income over earning assets); and 2) the yield on

loans (interest income on loans over gross loans) minus the rate paid on deposits

(interest expense on deposits over deposits). As can be seen, there is no apparent

correlation between capital ratios and either measure of markups.

13

As a benchmark, Krishnamurthy and Vissing-Jorgensen (2010) estimate that Treasury securities

impound a money-like convenience premium of approximately 72 basis points, on top of what would be

expected in a standard risk-vs.-return asset-pricing setting.

A Macroprudential Approach to Financial Regulation 19

Figure 2

U.S. Bank Capital Ratios and Loan Spreads

Sources: Data from 1840 to 1896 is based on Berger, Herring, and Szego (1995) who use data from the

Statistical Abstracts of the United States. Data from 1896 to 1919 is based on data for “All Banks” from the All

Bank Statistics, 1896–1955 (Board of Governors of the Federal Reserve, 1959). Data from 1919 to 1933 is

based on Federal Reserve member banks and is from Banking and Monetary Statistics, 1919–1941 (Board

of Governors of the Federal Reserve, 1943). Data from 1934 to 2010 is for all insured commercial banks

and is from the FDIC’s Historical Statistics on Banking.

Notes: Figure 2A plots the ratio of book equity to book assets for U.S. commercial banks from 1840 to

2009. Figure 2B plots two measures of bank loan spreads versus equity/assets from 1920 to 2009. We

measure loan spreads using either the net interest margin (net interest income over earning assets)

or the yield on loans minus the rate paid on deposits (interest income on loans over gross loans minus

interest expense on deposits over deposits).

A: Book Equity to Assets for U.S. Banks, 1840–2009

B: Relationship between Loan Spreads and Bank Equity/Assets, 1920–2009

60%

0%

50%

40%

30%

20%

10%

1840

2010

1850

1860

1870

1880

1890

1900

1910

1920

1930

1940

1950

1960

1970

1980

1990

2000

4%

16%

14%

12%

10%

8%

6%

1%

7%

6%

5%

4%

3%

2%

1920

2010

1930

1940

1950

1960

1970

1980

1990

2000

Bank equity/assets (left scale)

Bank net interest margins (right scale)

Bank loan yield - Deposit rate (right scale)

Equity to assets

20 Journal of Economic Perspectives

But Then Why Are Banks So Determined to Operate with High Leverage?

These conclusions may appear surprising, even paradoxical. If signi cant

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 risks this poses? And why do they

deploy armies of lobbyists to  ght increases in their capital requirements? After

all, non nancial  rms tend to operate with much less leverage and indeed appear

willing in many cases to forego the tax (or other) bene ts of debt  nance altogether.

In Kashyap, Stein, and Hanson (2010), we argue that the resolution of this

puzzle has to do with the nature of competition in  nancial services. The most

important competitive edge 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 20 basis

points, it may lose most of its business, (or become much less pro table, since the

return on assets in banking is on the order of 125 basis points). Contrast this with,

say, the auto industry, where cheap  nancing is only one of many possible sources

of advantage: a strong brand, quality engineering and customer service, and control

over labor costs may all be vastly more important than a 20 basis-point difference in

the cost of capital.

One suggestive piece of evidence for this competition hypothesis comes from

the distribution of capital ratios by bank size, as illustrated in Figure 3, which covers

the period 1976–2009. There is a strong inverse relationship between bank size

and capital ratios, with the smallest banks (with assets under $100 million) having

Tier 1 risk-based capital ratios more than double those of the largest banks (with

assets over $100 billion) for most of the sample period. Whatever their root cause,

these large differences in capital ratios hint at a couple of important points. First,

they suggest that several percentage points of additional capital need not imply

prohibitively large effects on lending rates—for if they did, it would be hard to

understand how the smaller community banks have managed to stay in business.

Second, the ability of small banks to survive at higher capital levels probably re ects

something about the softer degree of competition in their core line of business. A

large literature argues that small banks tend to focus on informationally intensive

“relationship lending” and that the embedded soft information in these relation-

ships creates a degree of speci city between  rms and their lenders (Rajan, 1992;

Petersen and Rajan, 1994, 1995; Berger, Miller, Petersen, Rajan, and Stein, 2005).

To the extent that larger banks deal with larger customers where competition from

other providers of  nance is more intense, even small cost-of-capital disadvantages

are likely to prove unsustainable.

Testing the Competition Hypothesis

To further investigate the competition hypothesis, we examine the effects of

changes in state branching regulations. We test two basic predictions. First, we expect

that a regulatory shock that increases the degree of competition in a state should

Samuel G. Hanson, Anil K Kashyap, and Jeremy C. Stein 21

Figure 3

U.S. Bank Capital Ratios by Bank Size, 1976–2009

Source: The  gure is based on data from bank Call Reports.

Notes: This  gure plots capital ratios by bank size from 1976–2009. Banks are placed into size groups

based on assets in 2008Q4 dollars. Figure 3A plots book equity to book assets. Figure 3B plots Tier

1 capital ratios (Tier 1 regulatory capital over risk-weighted assets). All banks owned by a given bank

holding company are combined into a single organization for the purposes of this size classi cation.

A: Book Equity to Book Assets

B: Tier 1 Risk-Based Capital Ratios

2%

4%

6%

8%

10%

12%

14%

1976

1978

1980

1982

1984

1986

1988

1990

1992

1994

1996

1998

2000

2002

2004

2006

2008

Assets < $100m

$10B < Assets < $100B

8%

10%

12%

14%

16%

18%

20%

6%

1996

1997

1998

1999

2000

2001

2002

2003

2004

2005

2006

2007

2008

2009

$1B < Assets < $10B

$100m < Assets < $1B

Assets > $100B

22 Journal of Economic Perspectives

lead the

lead the average

capital ratio of banks in that state to decline. Second, we expect a

compression effect

: the decline in capital ratios should be largest for those banks in the

state that, prior to the shock, were operating with the highest capital ratios. Or said

differently, we expect the regulatory shock to reduce the cross-sectional dispersion

of capital ratios of banks in the given state.

To implement our tests, we take data on the year that various state banking

regulations were relaxed from Stiroh and Strahan (2003). We examine two types

of deregulation: the easing of intrastate branching restrictions and the advent of

interstate banking. Prior to 1970, two-thirds of states had restrictions on intrastate

branching which were relaxed from 1975 to 1992. Between 1982 and 1993, 48 states

entered into regional or national agreements permitting interstate banking—i.e.,

allowing out-of-state bank holding companies to own banks in their state. Given the

timing of deregulation, we focus on bank data (from bank Call Reports) over the

period from 1976 to 1994. Since our source of variation is at the state-year level, we

work with state-year aggregates. We estimate reduced-form regressions of the state-

level equity-to-asset ratio on dummies that switch on in the year that a state relaxes

its regulations, along with state and year  xed effects. We have two deregulatory

dummies:

dummies: INTRASTATE

is based on the year that a state allows intrastate branching

by mergers, while

by mergers, while INTERSTATE

is based on the year that a state enters a regional or

national interstate banking agreement.

Table 3 displays the results of these regressions where the dependent vari-

able in the  rst column is the mean equity-to-asset ratio in state

able in the  rst column is the mean equity-to-asset ratio in state s

in year

in year t

; in the

second column is the cross-sectional standard deviation of equity-to-assets; and in

the remaining columns are the cross-sectional quantiles of the equity–asset ratio.

The results in the  rst column imply that equity-to-assets falls by about 0.3 percentage

points following intrastate branching and another 0.2 percentage points following

interstate banking. Thus, equity-to-assets falls by roughly 0.5 percentage points for

the average state relaxing both restrictions. This decline can be compared to the

typical cross-sectional standard deviation of 1.08 percentage points and is economi-

cally meaningful considering that the average equity-to-assets ratio in our sample is

just over 7 percent.

The remaining columns in Table 3 show that, consistent with the notion of

a compression effect, the dispersion of capital ratios within a state falls following

deregulation, and capital ratios fall the most for those banks that were previously

in the upper tail of the distribution. This appears to have been particularly true

following the advent of intrastate banking: the capital ratios of banks in the 75

th

and

90

th

percentiles of the distribution fall by 60 and 70 basis points, respectively, versus

only a 10 basis point change at the 10

th

and 25

th

percentiles.

In sum, the data support the hypothesis that when banks are faced with more

intense competition, they gravitate towards both higher and more uniform levels

of leverage. Such competitive effects combined with our earlier Modigliani–Miller-

based calibration results suggest one reason why tougher capital regulation of

 nancial  rms is appealing: it would seem to have the potential to reduce competition

A Macroprudential Approach to Financial Regulation 23

on a dimension that creates negative externalities and systemic risk, while at the

same time not raising loan rates by much. However, the complication is that these

same competitive pressures also create powerful incentives to evade either the letter

or the spirit of the rules. Thus, the most worrisome long-run byproduct of higher

capital requirements will likely not be its effect on the cost of credit to borrowers, but

the pressure it creates for activity to migrate outside of the regulated banking sector.

A Financial Reform Report Card

To conclude the paper, we brie y compare our proposals to policy reforms that

emerged in the second half of 2010. Our focus is primarily on the recommendations

made by the Basel Committee on Banking Supervision in September 2010 as part

of the so-called “Basel III” process (see Basel Committee on Banking Supervision,

2010, for a summary). We will say less about the Wall Street Reform and Consumer

Protection Act—often called the Dodd–Frank legislation—that was signed into law

in July 2010. This is because our analysis has been centered on capital regulation

and closely related issues, the implementation of which has been taken up in more

speci c numerical detail in Basel III; conversely, we have not attempted in this paper

Table 3

Impact of Deregulation on Distribution of Equity-to-Assets within States

Dependent variable

Mean

Standard

deviation

10th

%tile

25th

%tile

50th

%tile

75th

%tile

90th

%tile

Regression 1: –0.288 –0.274 –0.099 –0.106 –0.193 –0.626 –0.682

INTRASTATE [–2.10] [–2.75] [–0.70] [–0.81] [–1.49] [–2.47] [–2.40]

Regression 2: –0.217 –0.133 0.037 –0.182 –0.278 –0.290 –0.349

INTERSTATE [–2.25] [–0.68] [0.31] [–1.46] [–2.68] [–1.96] [–1.73]

Regression 3: –0.281 –0.270 –0.100 –0.100 –0.183 –0.617 –0.671

INTRASTATE [–2.05] [–2.68] [–0.71] [–0.78] [–1.41] [–2.44] [–2.33]

INTERSTATE –0.203 –0.120 0.042 –0.177 –0.269 –0.261 –0.317

[–2.05] [–0.62] [0.34] [–1.43] [–2.62] [–1.69] [–1.47]

Sources: Based on an annual state-level panel from 1976–1994 assembled from bank Call Reports. The

deregulation dummies are based on the data in Table 1 of Stiroh and Strahan (2003).

Notes: The table shows regressions of equity-to-assets on dummies for deregulation. The INTRASTATE

dummies switch on beginning in the year when the state  rst permitted intrastate branching via mergers.

The INTERSTATE dummies switch on beginning in the year when the state entered a regional or national

interstate banking agreement. The dependent variables are alternately the asset-weighted average, standard

deviation, and quantiles of the equity-to-assets ratio within each state-year. The table reports coef cients

from 21 separate regressions (7 dependent variables each with 3 speci cations). All regressions include a

full set of state and year effects and have 969 observations (51 states × 19 years). t-statistics, in brackets, are

based on standard errors that are robust to clustering (i.e., serial correlation of residuals) at the state level.

24 Journal of Economic Perspectives

to speak to a number of the other central elements in Dodd–Frank, such as consumer

protection, regulation of over-the-counter derivatives, and resolution authority.

We have stressed the importance of requiring that  nancial  rms have both

more capital, and, crucially, higher-quality capital. On this score, the Basel III

recommendations look quite good. They would raise the minimum common equity

requirement from 2 percent of risk-weighted assets to 7 percent (this is inclusive of

a “capital conservation buffer”). While we have argued for a higher number, this is a

signi cant step in the right direction. Moreover, systemically important institutions

will be required to have an additional, as-yet-undetermined increment in terms of

equity capital, which, if it turns out to be material, would be further good news.

The major shortcoming on the equity capital front is its very slow phase-in; the

new requirements do not become fully effective until January 2019. The motivation

for this slow phase-in is the concern that if banks are asked to comply with the higher

ratios in a more compressed timeframe, they will do so by shrinking their balance

sheets rather than by raising new external equity capital, thereby causing a further

credit crunch. While we agree that this worry might be legitimate if the phase-in is

truncated and no other offsetting steps are taken, our above analysis suggests an

obvious alternative: during the phase-in period, regulators should push those banks

that are shy of the new capital standards to

that are shy of the new capital standards to raise new dollars of equity

, rather than giving

them the option to adjust via asset shrinkage. The U.S. stress tests conducted in 2009

showed that this approach can work, and if it were applied again, the phase-in period

could be made much shorter with little adverse impact on credit supply.

We also discussed the usefulness of time-varying capital requirements. Here the

Basel Committee proposes an additional

Basel Committee proposes an additional countercyclical buffer

that will range between

0 and 2.5 percent, to be implemented on a country-by-country basis. As the report

states: “The purpose of the countercyclical buffer is to achieve the broader macro

prudential goal of protecting the banking sector in periods of excess aggregate

credit growth. For any given country, this buffer will only be in effect when there is

excess credit growth that is resulting in a system wide build up of risk.” This too is a

step in the right direction, though we worry that the wording would seem to require

an af rmative  nding of “excess credit growth” for the requirement to kick in; one

can imagine that it will be politically challenging for regulators to make this case

when they ought to.

Other elements of the reform package remain less well developed. For example,

on the topic of debt maturity, the Basel Committee introduces the concept of a

“net stable funding ratio” test—a requirement that  nancial  rms’ capital structures

have a certain amount of long-term funding, which would encompass both equity

and debt with a maturity of greater than one year. However, the details regarding

the design and calibration of this rule remain to be worked out, with an “observa-

tion period” to begin in 2012 and the introduction of the standard itself put off

until 2018.

Similarly, while the Basel Committee continues to study various forms of contin-

gent capital, it has not yet reached any  nal conclusions; a review is scheduled to

Samuel G. Hanson, Anil K Kashyap, and Jeremy C. Stein 25

be completed in mid 2011. Interestingly, however, Swiss banking regulators have

chosen to move forward on their own on this front. The Final Report of the Swiss

Commission of Experts proposed that the two big Swiss banks—UBS and Credit

Suisse—be required to have 19 percent total capital by 2019. Of this, 10 percent

would have to be in common equity (a higher standard than the 7 percent under

Basel III) while the remaining 9 percent could, at the bank’s discretion, take the

form of contingent convertibles that would convert when the ratio of equity to assets

hit a predetermined trigger value (Morgan Stanley Research, 2010). This “opting”

approach to contingent capital is, both qualitatively and quantitatively, closely in

line with what we described above.

Finally, perhaps the most glaring weak spot in  nancial reform thus far—one

that cuts across both the Dodd–Frank legislation and the Basel III process—is the

failure to fully come to grips with the shadow banking system. As we have emphasized,

if one takes a macroprudential view, the overarching goal of  nancial regulation

must go beyond protecting insured depositories and even beyond dealing with the

problems created by “too-big-to-fail” nonbank intermediaries. Instead, the task is to

mitigate the  re-sales and credit-crunch effects that can arise as a consequence of

excessive short-term debt

excessive short-term debt anywhere in the  nancial system

.

While higher capital and liquidity requirements on banks will no doubt help to

insulate banks from the consequences of large shocks, the danger is that, given the

intensity of competition in  nancial services, they will also drive a larger share of

intermediation into the shadow banking realm. For example, perhaps an increasing

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 safer than they were before, but the overall

system of credit creation may not.

To safeguard the system as a whole, attention must be paid to not tilting the

playing  eld in a way that generates damaging unintended consequences. Admit-

tedly, regulating the shadow banking sector and the other parts of the  nancial

system consistently is a complex task, and one that will require a variety of speci c

tools. As one concrete  rst step, we reiterate that it would be a good idea to establish

regulatory minimum haircut requirements on asset-backed securities, so that no

investor who takes a position in credit assets is able to evade constraints on short-

term leverage.

This discussion raises a  nal question about how such regulation might be

implemented. In the United States and in Europe, macroprudential oversight has

been delegated to large councils: the Financial Stability Oversight Committee and

the European System Risk Board, respectively. Membership of both groups consists

of the heads of many regulatory organizations. Whether either council can function

effectively and avoid turf wars is an open question. But these committees will be

pivotal in determining whether existing weaknesses in the regulatory system—such

as those having to do with the shadow banking sector—can be addressed sensibly.

26 Journal of Economic Perspectives

■ We are grateful for helpful comments from Charles Goodhart, Arvind Krishnamurthy, Andrew

Metrick, Victoria Saporta, Hyun Shin, Andrei Shleifer, René Stulz, Lawrence Summers, Paul

Tucker, seminar participants at numerous institutions, and JEP editors David Autor, Chad

Jones, and Timothy Taylor. Kashyap thanks the Initiative on Global Markets and the Center

for Research on Securities Prices for research support on this project.

References

Bank of England. 2009. “The Role of Macro-

prudential Policy: A Discussion Paper.” November.

http://www.bankofengland.co.uk/publications

/other/ nancialstability/roleofmacroprudential

policy091121.pdf.

Basel Committee on Banking Supervision. 2010.

“The Basel Committee’s Response to the Financial

Crisis: Report to the G20.” October.

http://www

.bis.org/publ/bcbs179.pdf.

Berger, Allen N., Richard J. Herring, and

Giorgio P. Szego. 1995. “The Role of Capital in

Financial Institutions.” Journal of Banking and

Finance, 19(3–4): 393–430.

Berger, Allen N., Nathan H. Miller, Mitchell A.

Petersen, Raghuram Rajan, and Jeremy C. Stein.

2005. “Does Function Follow Organizational Form?

Evidence from the Lending Practices of Large and

Small Banks.”

Journal of Financial Economics, 76(2):

237–69.

Bernanke, Ben S. 2008. “Reducing Systemic

Risk.” Speech at the Federal Reserve Bank of

Kansas City’s Annual Economic Symposium,

Jackson Hole, Wyoming, August 22, 2008.

Borio, Claudio. 2003. “Towards a Macropru-

dential Framework for Financial Supervision and

Regulation?” BIS Working Paper 128.

Borio, Claudio, Craig Fur ne, and Phillip

Lowe. 2001. “Procyclicality of the Financial System

and Financial Stability: Issues and Policy Options.”

In Marrying the Macro- and Micro-Prudential Dimen-

sions of Financial Stability, BIS Papers 1, Bank for

International Settlements, pp. 1–57. http://www

.bis.org/publ/bppdf/bispap01.htm.

Board of Governors of the Federal Reserve

System. 1943. Banking and Monetary Statistics,

1919–1941. Washington: Publications Services of

the Board of Governors. Washington, DC.

Board of Governors of the Federal Reserve

System. 1959. All Bank Statistics, United States,

1896–1955. Washington: Publications Services of

the Board of Governors. Washington, DC.

Brunnermeier, Markus, Andrew Crockett,

Charles Goodhart, Avinash D. Persaud, and Hyun

Shin. 2009. “The Fundamental Principles of Finan-

cial Regulation.” Geneva Reports on the World

Economy 11.

Brunnermeier, Markus K., and Lasse H.

Pedersen. 2009. “Market Liquidity and Funding

Liquidity.” Review of Financial Studies, 22(6):

2201–38.

Bryant, John. 1980. “A Model of Reserves, Bank

Runs, and Deposit Insurance.” Journal of Banking

and Finance, 4(4): 335–44.

Caballero, Ricardo, and Alp Simsek. 2010. “Fire

Sales in a Model of Complexity.” NBER Working

Paper 15479.

Covitz, Daniel, Nellie Liang, and Gustavo

Suarez. 2009. “The Anatomy of a Financial Crisis:

The Evolution of Panic-Driven Runs in the Asset-

Backed Commercial Paper Market.” Finance and

Economics Discussion Series Working Paper 2009-

36, Board of Governors of the Federal Reserve.

Crockett, Andrew. 2000, “Marrying the Micro-

and Macroprudential Dimensions of Financial

Stability.” BIS Speeches, September 21.

Diamond, Douglas W., and Philip Dybvig. 1983.

A Macroprudential Approach to Financial Regulation 27

“Bank Runs, Deposit Insurance and Liquidity.”

Journal of Political Economy, 91(3): 401–419.

Diamond, Douglas W., and Raghuram G. Rajan.

2009. “Fear of Fire Sales and the Credit Freeze.”

NBER Working Paper 14925.

Duf e, Darrell. 2010. How Big Banks Fail and

What To Do About It. Princeton University Press.

Flannery, Mark J. 2005 “No Pain, No Gain?

Effecting Market Discipline via ‘Reverse Convert-

ible Debentures.’” In Capital Adequacy beyond Basel:

Banking, Securities, and Insurance, 171–97. Oxford

University Press.

French, Kenneth R., Martin N. Baily, John

Y. Campbell, John H. Cochrane, Douglas W.

Diamond, Darrell Duf e, Anil K Kashyap, Frederic

S. Mishkin, Raghuram G. Rajan, David S. Scharf-

stein, Robert J. Shiller, Hyun Song Shin, Matthew

J. Slaughter, Jeremy C. Stein, and René M. Stulz.

2010. The Squam Lake Report: Fixing the Financial

System. Princeton University Press.

Geanakoplos, John. 2010. “The Leverage Cycle.”

Cowles Foundation Discussion Paper 1715R.

Geanakoplos, John, and Heraklis Pole-

marchakis. 1986. “Existence, Regularity and

Constrained Suboptimality of Competitive Alloca-

tions When the Asset Market Is Incomplete.”

In Essays in Honor of Kenneth Arrow, Vol 3. ed. W.

Heller, R. Starr, and D. Starrett, 65–95. Cambridge

University Press.

Gorton, Gary B. 2010. Slapped by the Invisible

Hand: The Panic of 2007. Oxford University Press.

Gorton, Gary B., and Andrew Metrick. 2010.

“Securitized Banking and the Run on Repo.” Yale

ICF Working Paper 09-14.

Hanson, Samuel, and Adi Sunderam. 2010. “Are

There Too Many Safe Securities? Securitization

and the Incentives for Information Production.”

http://www.people.fas.harvard.edu/~sunderam

/papers/sunderam_JMP.pdf.

Hart, Oliver, and Luigi Zingales. 2010. “A

New Capital Regulation for Large Financial

Institutions.” http://faculty.chicagobooth.

edu/luigi.zingales/papers/A_New_Capital

_Regulation_ALER_version.pdf.

Hirtle, Beverly, Til Schuermann, and Kevin

Stiroh. 2009. “Macroprudential Supervision of

Financial Institutions: Lessons from the SCAP.”

Federal Reserve Bank of New York Staff Report 409.

International Monetary Fund. 2010. Global

Financial Stability Report. April. Washington, DC:

International Monetary Fund.

Ivashina, Victoria, and David Scharfstein. 2010.

“Bank Lending During the Financial Crisis of

2008.” Journal of Financial Economics, 97(3): 319–38.

Kashyap, Anil K, Richard Berner, and Charles A.

E. Goodhart. Forthcoming. “The Macroprudential

Toolkit.” IMF Economic Review.

Kashyap, Anil K, Raghuram Rajan, and Jeremy

C. Stein. 2008. “Rethinking Capital Regulation.” In

Maintaining Stability in a Changing Financial System,

431–71. Federal Reserve Bank of Kansas City.

Kashyap, Anil K, and Jeremy C. Stein. 2004.

“Cyclical Implications of the Basel-II Capital Stan-

dards.” Federal Reserve Bank of Chicago Economic

Perspectives, vol. 28, Q1, pp. 18–31.

Kashyap, Anil K, Jeremy C. Stein, and Samuel

Hanson. 2010. “An Analysis of the Impact of

‘Substantially Heightened’ Capital Requirements

on Large Financial Institutions.” http://faculty

.chicagobooth.edu/anil.kashyap/research

/an_analysis_of_the_impact_of_substantially

_heightened.pdf.

Krishnamurthy, Arvind, and Annette Vissing-

Jorgensen. 2010. “The Aggregate Demand

for Treasury Debt.” http://www.kellogg

.northwestern.edu/faculty/vissing/htm/demand

treas_may122010_jperevision.pdf.

McDonald, Robert L. 2010. “Contingent

Capital with a Dual Price Trigger.” http://papers

.ssrn.com/sol3/papers.cfm?abstract_id=1553430.

Modigliani, Franco, and Merton H. Miller. 1958.

“The Cost of Capital, Corporation Finance and the

Theory of Investment.” American Economic Review,

48(3): 261–97.

Morgan Stanley Research. 2010. “European

Banks: €49 Billion of Swiss CoCos.” October 4.

Myers, Stewart C. 1977. “Determinants

of Corporate Borrowing.” Journal of Financial

Economics, 5(2): 147–75.

Myers, Stewart C., and Nicolas Majluf. 1984.

“Corporate Financing and Investment Decisions

When Firms Have Information That Investors Do

Not Have.” Journal of Financial Economics, 13(2):

187–222.

Petersen, Mitchell, and Raghuram Rajan. 1994.

“The Bene ts of Lending Relationships: Evidence

from Small Business Data.” Journal of Finance,

49(1): 3–37.

Petersen, Mitchell, and Raghuram Rajan.

1995. “The Effect of Credit Market Competition

on Lending Relationships.” Quarterly Journal of

Economics, 110(2): 407–44.

Rajan, Raghuram. 1992. “Insiders and

Outsiders: The Choice between Relationship

and Arms-Length Debt.” Journal of Finance, 47(4):

1367–1400.

Shin, Hyun Song. 2009. “Re ections on

Northern Rock: The Bank Run that Heralded

the Global Financial Crisis.” Journal of Economic

Perspectives, 23(1), 101–119.

Shleifer, Andrei, and Robert W. Vishny. 2010.

“Unstable Banking.” Journal of Financial Economics,

28 Journal of Economic Perspectives

97(3): 306–18.

Stein, Jeremy C. 2010a. “Monetary Policy

as Financial-Stability Regulation.” http://www

.economics.harvard.edu/faculty/stein/files

/MonetaryPolicyAsRegulation-8-2010.pdf.

Stein, Jeremy C. 2010b. “Securitization, Shadow

Banking, and Financial Fragility.” Daedalus, 139(4):

41–51.

Stiroh, Kevin J., and Philip E. Strahan. 2003.

“Competitive Dynamics of Deregulation: Evidence

from U.S. Banking.” Journal of Money, Credit, and

Banking, 35(5): 801–28.

Tucker, Paul. 2010. “Shadow Banking, Financ-

ing Markets and Financial Stability.” Speech at a

Bernie Gerald Cantor (BGC) Partners Seminar,

London, January, 21, 2010.