Should regulation be dumb? In one sense yes, in others, no. It
really depends on how well the regulators understand the risks
involved, and how much they can encourage professionalism among
profit center heads and risk managers. As those two increase,
regulation can be smart. "Follow these detailed rules to calculate
the capital you need to be solvent 99% of the time."
But when either of those two aren't true, dumb regulation may be
- Strict leverage limits, reflecting the worst outcome from
underwriting poor quality loans.
- Disallowing risky types of lending, regardless of capital
- Disallowing liabilities that can run easily.
- Disallowing products that commonly deceive buyers.
- Disallowing certain types of contracts that fuddle
- Those regulated may not choose their regulator. The highest
regulator assigns a regulator to you. The highest regulator must
evaluate the jobs that lower regulators are doing, and
eliminate/lessen regulators that do not use the powers they have
been granted, and get co-opted by those that they regulate.
If everyone were smart, things could be different. Deceiving
people would not take place, and managements would not take undue
risks. Limits could be looser, and products would be designed for
But, face it, we are dumber than we think, myself included.
Consumer choice is a good thing, though it implies that some will
be deceived, no matter where one places the line of demarcation.
Along with that, some bank will not fit the rules and go insolvent,
though it previously passed the solvency tests.
Dumb Regulation: Insurance in the US
My poster child for relatively good dumb regulation is the
insurance industry in the US. The industry is far less
free-wheeling than the banking industry, and under most
circumstances, the solvency margins are set high enough to have few
insolvencies. There is room for improvement, though:
- Make risk based capital charges countercyclical. Perhaps
tinkering with the Asset Valuation Reserve would do that.
- Have some sort of rigorous testing for capital relief from
- Ban surplus notes in related party transactions.
- Ban all forms of capital stacking, especially where the
transactions go both ways. I.e., subsidiaries can't own
securities of any companies in their corporate family. All
subsidiaries must be owned by the holding company.
- More rigorous testing for deferred tax assets.
- Assets as risky as equities, including limited partnerships,
should be a deduction from capital.
- Securitized bonds that are not "last loss" should have higher
RBC charges than comparable rated corporates, because loss
severities are potentially higher, and assets that are originated
to securitize are always lower quality than those held on balance
- A standardized summary of cash flow testing results should be
As for the banks, they need to do that and more:
- Insurance companies list all of their assets. Banks should as
- Intangible assets should be written to zero for regulatory
- Risk-based capital standards need to be tightened to at least
the level of insurance companies, if not tighter.
- Some sorts of lending to consumers should be banned. I am
talking about complex agreements, that individuals with IQs less
than 120 can't understand. Insurance policies have to be
Flesch-tested. Bank lending agreements should be the same. If
some argue that the poor need access to credit, I will say this:
the poor need to get off of credit. Credit is for the
upper-middle-class and rich. Poor people should not go into
- Standardized summaries of terms and fees must be created for
consumer lending, with large, friendly letters, and simple
language that all can read.
What I am saying is that accounting has to be more conservative,
and that regulators have to require larger amounts of capital to
support their business, particularly at the banks. Financial
products must be made simpler for consumers to understand. More
transparency is needed everywhere, and if the financial companies
complain, tell them that they will all be in the same goldfish
bowl, so no one will gain an unfair advantage.
Preventing Too Big to Fail
As part of preventing too big to fail, the Risk based capital (
) percentage should rise with the amount of risk-based capital.
Say, when RBC gets over $10 billion, the percentage of capital
needed for RBC grades up to 50% higher than the level needed at $10
billion by the time RBC gets up to $50 billion.
Here is my example of how it would work:
Marginal E/A Ratio
I have assumed that firms undertake their highest ROE projects
first, and do progressively lower ROE projects later. Now, by
raising capital requirements on bigger firms, a common response is,
"Well, then they will just take on riskier loans to compensate."
Sorry, but that dog don't hunt. If they take on riskier loans,
their RBC goes up even more rapidly, because loan quality is
reflected (or, should be reflected) in RBC formulas prior to
adjustment for bank size.
More Dumb Regulation
Dumb regulation bars certain lending practices, and raises
capital levels higher than is needed over the long run. So be it.
Smart regulation is far more flexible, and trusting that companies
and consumers know what they are doing. Unfortunately, when
financial firms fail, there are often larger repercussions. It is
better to limit regulated financial companies to businesses where
the risks are well-understood. Let the less understood risks be
borne by those outside the safety net, and bar those inside the
safety net from holding any assets in those companies.
That brings me to the Volcker Rule, which is a good example of
dumb regulation. My preferred way would be to do something similar
through adjusting the risk-based capital formulas - Equity-like
risks should be funded through a 100% allocation of equity. Few
banks would take on that level of speculation at that level of
If you need proof, look at the life insurance industry.
Companies used to hold a lot more equities prior to the tightening
of RBC rules. Now they hold little, except at a few mutual
companies that are flush with capital.
That also has preserved the insurance business in this crisis,
leaving aside mortgage and financial risks, where the state
regulators still have no idea what they are doing - that a proper
reserve level would leave most of the companies insolvent today,
but had it been implemented ten years ago, would have preserved the
companies, but eliminated much of their profits.
At the Treasury meeting with bloggers in November 2009, I
commented that the insurers were better regulated for solvency than
the banks. One of the reasons for that is that they do harder
stress tests, and they look longer-term. Life and P&C insurers
survive the process because of better RBC standards, and "scaredy
cat" state regulators. What a great system, which prior to the
crisis, was criticized as behind the times. (I suspect that if we
ever get a national regulator of insurance, there will be a big
boom and bust, much as in banking at present. It is easier to
corrupt one regulator than fifty.) The more state involvement in
bank regulation, the dumber (better) bank regulation will be.
What to Do
So, if one is trying to regulate banks for solvency, there are
seven things to do:
- Set risk-based capital formulas so that few institutions
- Make it even less likely that larger institutions fail.
- Limit the ability of financial institutions to invest in
other financial institutions.
- Regulators must benchmark the underwriting culture, and raise
red flags when underwriting is poor.
- Insure that equity is truly equity.
- Institute a code of ethics for risk managers.
- Make sure that balance sheets fairly reflect
It is almost always initially profitable to borrow short and
lend long. That said, it is a noisy trade. Who can be sure that
short rates will remain below the rates at which one invested long?
Another component of a good risk-based capital formula is that
there is no investing in assets that are longer than the
liabilities that fund the financial institution. (For wonks only:
regulated financial institutions should be matching assets versus
liabilities as their most aggressive posture. Unregulated
financials can do what they want. And no investing in unregulated
financials by regulated financials.)
One of the great subsidies banks get is the cheap source of
funds through deposits. It is only cheap because depositors know
the FDIC is there. The FDIC should raise its fees to absorb that
subsidy back to the taxpayer. Keep raising it until you see banks
begin to shift to repo and other short-term sources of funding.
As a clever old boss of mine once said, "A bank's liabilities
are its assets, and its assets are its liabilities." The idea is
this - banks that focus on their deposit franchises have something
of real value - that is hard to replicate. But any bank can invest
their funds aggressively, which will lead to defaults with higher
frequency. It is true of insurers as well, most financials die from
bad investing policies, and short-term liabilities that require
complacent funding markets.
That's why there has to be a focus on liabilities in regulating
solvency. Financial institutions, even simple ones, are opaque.
Most die from the deadly combo of illiquid assets and liquid
liabilities. Those that have funded the bank in the short run
refuse to roll over the loans at any price. Assets can't be
liquidated to meet the call on cash, and insolvency ensues. Those
that have read me for a long time know that I don't buy the
malarkey that some managements will trot out, "We're not insolvent;
we merely have a liquidity crisis." Hogwash. You took too much
risk, because the first priority of risk control is liquidity
management. Assets are only worth what you can sell them for, or,
what cash flows they can generate. If assets can't generate cash
flows or sale proceeds adequate to service liabilities, then you
are insolvent, not merely illiquid.
Cash flow testing for banks should focus on the ability of the
bank to finance itself without recourse to selling assets. To the
extent that selling assets is allowed in modeling, they must be
Treasury quality assets.
The essence of a good risk-based capital formula is that it
forces intelligent diversification, and forces adequate liquidity.
No assets should be bought that the liability structure of the bank
cannot hold until maturity. There should be no concentration of
assets by class, subclass, or credit, that would be adequate to
lead to failure.
My view is that a proper risk-based capital regime would start
with asset subclasses, and double the capital held on the largest
subclass, and 1.5X the capital on the second largest subclass.
After that, within each subclass, the top 10 credits get twice the
level of capital, the next 10 1.5x the level of capital. Having
managed assets in a framework like this, I can tell you that it
Beyond that, no modeling of asset correlations would be brought
into the modeling because risky asset correlations go to one in a
crisis. Any advantage derived from diversification should be
accepted as earned, and not capitalized as planned for.
Securitization deserves special treatment: risk based capital
should higher for securitized assets versus unsecuritized assets in
a given ratings class, because of potentially higher loss
severities, and assets that are originated to securitize are always
lower quality than those held on balance sheet. Capital charges
should be raised until banks don't want to securitize as a matter
of common practice.
In order to avoid systemic risk and contagion, banks should not
lend to or own other financial firms. That would end contagion. At
least that should be limited to a percentage of assets, or through
the RBC formula. Think of it this way, financials owning financials
is a form of capital stacking across the country as a whole. In a
stress situation it raises the odds of a deep crisis. Setting a
limit on the ability of financials to own the assets of financials
is the single most important step to avoid contagion. I would set
the limit at 5% for equity, and 20% for debt.
Most of the real risks came from badly underwritten home
mortgage debt, whether conventional, Alt-A and Jumbo, or subprime.
Underwriting standards slipped everywhere. Commercial mortgage
lending hasn't yet left its marks - there is a lot of hope that
banks can extend maturing loans rather than foreclose and take
For much but not all of this crisis, it was not a failure of
laws but a failure of regulators to do their jobs faithfully.
Regulators should have looked at indicators of loan quality, and
raised red flags when they saw standards deteriorating. Where I
worked, 2003-2007, we saw the deterioration, and were amazed that
the regulators had been neutered.
Let Equity Be Equity
Beyond that, there was a dearth of true equity, and a surfeit of
preferred stock, junior debt, trust preferreds, and particularly,
goodwill. Equity has to reflect assets that are high quality and
that are not needed to support short-term obligations from the cash
Code of Ethics for Risk Managers
One reason the banking industry is worse off than insurance, is
that they don't have many actuaries. Actuaries have a code of
ethics. They tend to be "straight arrows" telling it like it is.
Bank risk managers need the same thing, together with the rigorous
education that actuaries receive. Accept no substitutes: CFAs and
CERAs are no match for FSAs.
Reflect Derivatives Properly
Derivatives must come onto the balance sheet for regulatory
purposes, revealing leverage increases/decreases, counterparty
risk, overall sensitivity to the factors underlying the contracts.
Any instrument that can cause cash to flow at the regulated entity
should be on the regulatory balance sheet.
I would not create a prospective guarantee fund. The insurance
industry has a retrospective fund that has worked fairly well. Do
you really know what it would take to create a macro-FDIC, big
enough to deal with a large systemic risk crisis like this one?
(The FDIC, much as it is pointed out be an example, is woefully
small compared to the losses it faces, and it is not even taking on
the large banks.) It would cost a ton to implement, and I think
that large financial services firms would dig in their heels to
fight that. Also, there would be moral hazard implications -
insured behavior is almost always more risky than uninsured
Though it is not bank reform, we need to end the
Greenspan/Bernanke Put. The Fed encouraged risk-taking by the banks
by not allowing recessions to damage them. They tightened too late,
and loosened too early, and that pushed us into a liquidity trap.
Monetary policy that is too loose creates perverse incentives for
the solvency of financial institutions in the long run.
Bonuses to executives skew incentives. Bonusing a financial
executive on current earnings creates perverse incentives. It is a
form of asset/liability mismanagement, because cash flows in the
short run, while the value of the institution is a long-run issue.
Far better to incent using long dated restricted common stock. The
only trouble is, it doesn't incent as well as cash. Tough, sorry,
but that is a loss that must be accepted for the good of the system
as a whole.
Dumb regulation is good regulation. Regulators should be
risk-averse, and take actions that limit ROEs for banks in order to
promote solvency, and reduce the likelihood of liquidity crises.
The remedies that I have proposed here will do just that. May we
use them to regulate our financial sector better, for the good of
all in our nation.
Corporate Profits Continue to Impress