Monday, February 6, 2017

Bank deregulation

John Cochrane has a post today about the potential for Dodd-Frank reform.  The short version is: it would be a good development if we could find a compromise where banks are less regulated in exchange for requiring higher capital requirements.

I agree with this, but really, I think this is simply the what any real deregulation looks like.  I think the GSEs really help to bring clarity to this issue.  They are similar here to commercial banks.  They have very low capital requirements, which worries a lot of observers.  Along with those capital requirements is de facto government backing for their debt and issued securities.

So, banks have FDIC insurance, which provides safety for depositors.  This means that as banks take in deposits (which are liabilities for banks), the interest rates required by those depositors don't react to the risks created by leverage, since FDIC insurance is basically like a credit default swap for the depositors.  Any non-financial corporation would naturally remain less leveraged, outside of crisis situations, because they would have to pay higher interest rates for debt as their leverage increases.  Since FDIC insurance undermines this basic market source of moderation, banks have to be regulated so that they don't become too leveraged.  They are forced to keep a minimum level of capital.

But, the minimum level of capital required has to be less than what the natural market level of capital would have been.  Otherwise, what's the point?  And, banks would simply be out-competed by non-regulated substitutes if they were forced to hold capital levels above the natural market level.

I'm not an expert on the repo market, but it seems to me that we created a similar problem by inventing the accounting fiction that repos are not loans, which also makes the interest rate on repo financing non-responsive to leverage risk.

In any case, there is a natural pairing between capital requirements and regulatory restraints.  As regulation and public insurance of various kinds ratchet up, capital requirements naturally decline.  If banks were completely deregulated, they would naturally hold more capital.

I wonder if commercial banks, as they exist today, are an anachronism anyway.  If they were deregulated, I wonder if they would mostly go away.  Is the central intermediation that they engage in - borrowing short and lending long - even necessary today?  Is the regulatory framework that supports this intermediation actually maintaining a risk in the economy that isn't even necessary today?  Would money market funds, REITs, investment banks, and a host of other financial agents rise up in ways that would match asset-liability maturities within each institution?

This is really clear when thinking about the GSEs.  If they didn't have federal backing, their business model would probably be ineffective.  And, if their capital requirements were set too high, their business model would probably be ineffective.  Both of those factors need to be in place for the GSEs to exist, and those factors come as a pair.  What point is there of having capital requirements if there is no guarantee?  The bondholders would not be accepting a discounted interest rate in that case, and they would be perfectly able to demand their own bankruptcy terms if there was no federal backing.  Imagine how awkward it would be to have the capital requirement without federal backing.  If the bondholders held securities that had become impaired, or where there was a probability of future default, they would naturally be in communication with the boards of the GSEs to manage the firms in everyone's best interest.  Imagine if, without giving them any support, the federal government stepped in and imposed a settlement on them that neither the equity holders or the bondholders had invoked.  What would be the point?  The capital requirement without the guarantee would do nothing but add risk.  That's what is so strange about how so many policymakers still pretend like there was never a government guarantee on GSE debt.  Of course there was.  There had to be.  And, it was invoked in 2008, if there was ever any doubt.

In the book, I walk through the argument that there should not be GSE debt, but that there should be a government guarantee of GSE MBSs.  It would be a public good which really can only be provided publicly.  That's a story for another day.  But, clearly, in the case of the GSEs we can see that, in and of itself, regulatory insurance and capital requirements go together, and they can induce institutional forms that increase systematic risk.  Public insurance increases systematic risk.  Maybe that is counterintuitive.

It seems to me that many people assume that deregulation somehow helps insiders and powerful institutions, and that, if we did deregulate banks, those same people would quite naturally and quickly, shift toward getting mortgages from privately funded REITs, investment banks, and MBS investors and toward putting their short term savings into money markets that invested in commercial paper and short term treasuries, and it would never occur to them that they were intimately involved in the devolution of power from previously regulated institutions.

Cochrane's idea of deregulating and raising capital requirements seems like a step in the direction of simply deregulating, which is all for the best.  But, if there was only some deregulation, then the capital requirements would still have to remain lower than what the completely deregulated market would provide.  In fact, it seems like what we see in the marketplace of substitutions for commercial banks are many forms of saving that don't particularly use leverage at all.  This would be even more the case if we stopped giving debt tax advantages over equity.


  1. Great post, and deep thought-provoking observations. Well, deep by my standards.

    There might evolve a free-nmarket model in which financial institutions buy insurance, or depositors buy insurance. Think AIG, but for depositors.

    So, the private sector would insure bonds or deposits, or mortgages (PMI).

    Maybe such a model would be robust and sturdy.

    My guess is such models work until they don't (see AIG).

    If my sentiments are accurate, we could have a good private-insurance model that every 50 years or so suffers catastrophic collapse, snowballing recessions-depressions etc.

    Maybe a last-stop government back-up is needed somewhere. Keep everything private, but leave authority for a central bank to print money and buy bad loans etc.

    After a few decades of financial reporting, I wonder if financial types have the gravitas, authority, self-discipline required for a permanent financial sector. Egads, what Long-Term Capital Management managed to do--just one money manager.

    I like say to ask, "Why do engineers become engineers?" They love engineering and they need money. Doctors, teachers etc.

    Why do people work on Wall Street?

    They love money and they love money.

    On side note:

    “Britain has changed since 1998.

    Back then, it only took workers about three years to save enough money for a down-payment on a house. Now it takes 20 years, on average, according to the Resolution Foundation, which published a landmark report on income, housing, and inequality in Britain last week.”


    That is from Tyler Cowen today.

    More and more, this problem of property is working its way into living costs, and our financial systems. Our banks are hugely exposed to property, which is propped up artificially by zoning. Banks love lending on property.

    Large global capital flows, migration---but rigid supplies of housing,

    1. The distinction isn't whether deregulated markets are perfect. They can't be. The distinction is whether rules and regulations make them more safe or less safe. There are a lot of regulations that make us less safe.
      LTCM is on them. It was a one-off situation, and it seems like it was handled well. I think the idea that moral hazard led to the housing bust is extremely over-stated.

  2. I think I agree about moral hazard and housing. After all, a lot of people were kicked out of their homes and lost their deposits or equity.

    I am more thinking about moral hazard on the institutional side. The ability, like LCTM, to leverage 100 to one to expand profits.

    Huge upside if you bet right, and you simply walk away if you bet wrong.

    Anyway, I agree on private financial insurance. I am just saying that ultimately the central bank has to be able to print money.

  3. I think I agree about moral hazard and housing. After all, a lot of people were kicked out of their homes and lost their deposits or equity.

    I am more thinking about moral hazard on the institutional side. The ability, like LCTM, to leverage 100 to one to expand profits.

    Huge upside if you bet right, and you simply walk away if you bet wrong.

    Anyway, I agree on private financial insurance. I am just saying that ultimately the central bank has to be able to print money.

  4. Add on: We saw the S&L tumble in the 1980s. Interesting story. Guys would buy a small S&L, set up a boiler-room, run national ads offering highest deposit rates. Huge inflows of money. Then lend the money to friends and run away.

    AIG said they insured bonds. Well, except when bond values fall as a group.


    Bear Stearns, Lehman.

    Not sure what to make of it.

    Private deposit insurance makes sense, but how to stop banks from buying insurance from fly-by-night operators?

    1. There are many ways for these arrangements to be made. In banking, it seems like the clearinghouses of the 19th century might be a model that would work. Sort of a bank co-op. It would naturally solve TBTF because nobody would be willing to insure a bank that was too large.

  5. Kevin:

    The People's Bank of China periodically buys bad loans, and solves the problem that way. The banking system goes forward. Probably this leads to some corruption.

    But since the PBOC is below its inflation target anyway, I guess this works. The Chinese banking system will never collapse. Westerners do not have a clue about that.

    Actually, I never thought bank bailouts were all that bad (even in the U.S.), as long as shareholders and bondholders get wiped out. Then, there are plenty of consequences for investors to a bank failure. They lose all their money. There is no moral hazard.

    I see no reason for unsophisticated depositors to lose money, which could incur bank runs.

    I suppose another bank system would be only one regulation, that banks buy deposit insurance from a licensed insurer. This might lead to something like a "Big Five" of national bank insurers, something like we have a Moody's, S&P and Fitch today. A few recognized outfits.

    So depositors would be told they pay 1% of deposits for insurance, more at riskier banks, and less at less risky banks etc.

    But even this Big Five would have to backed up by a central bank.

    The private financial sector is built for big storms, really big storms---but not once-a-century storms.

    In any event, I agree with your assessments that regulations upon regulations and reserve requirements and stress tests and other do-goody regs like CRA probably accomplish little, and may be counterproductive.

    Curiously enough, if you go to the ABA website, they seem to have long ago accepted Dodd-Frank.

    1. The reason they need insurance is because there is an asset - liability mismatch. Money market funds don't have FDIC. There was just one fund that "broke the buck", right? If there wasn't FDIC, you wouldn't put short term deposits in banks, but you wouldn't have to. I don't see why unleveraged money markets couldn't invest your deposit in short term treasuries and serve the same function as banks. There are plenty of investors willing to buy RMBS (or there were). Why do we need to subsidize the system that mismatches assets and liabilities?

  6. Kevin--

    Yes I have been wondering about that too. Banks intermediate...but should they?

    But people have $10 trillion now in deposits at US banks. Could all that capital be otherwise deployed?

  7. AIG and 2008 clearly showed that a private entity can't offer financial insurance for truly catastrophic financial events because correlations get too high. And if the public sector provides that insurance, it has to regulate. Side note - if the Fed implements an effective NGDPLT (which it should), then as Harry Chernoff stated, I think we will eventually have a crisis anyway and no private financial insurer would be able to survive that. Second side note - I absolutely don't want the Fed buying bad loans (unless it only buys them from me! :-)
    LTCM is an interesting case. The Fed basically coordinated an expedited Chapter 11. That's something that our system should provide for all borrowers and lenders. That includes mortgages.

    1. I don't know. I mean, certainly there are examples, like AIG, that support your point. But, I think George Selgin and Lawrence White make a decent argument, and clearly, under private insurance, capital levels would be higher and TBTF would solve itself, because an insurer would have to commit to failing if the TBTF bank failed. A bank too large would have a hard time getting insurance.

      I think it might be a step in a better direction to have public re-insurance of private insurers.

      I don't know if it will make it past review, but in the book, I argue that the GSEs should retain their guarantee business as a public utility, preferably as part of the Federal Reserve. Insurance against systematic risk should be a public function, and where there is a framework that produces exposure to systematic risk as purely as the GSE guarantee business does, it should operate as a public utility.

    2. Insurance against systemic risk should be a public function. I agree. I like proposals that harness the risk assessment abilities of the private sector too. One concept I've liked for bank regulation is this. X% of a bank's capital stack should come from a convertible debt security. This security would be quickly converted to equity under certain circumstances (another form of expedited Chapter 11). If the security traded publicly, then regulators could just leave alone the banks where the security was trading close to par. Right now, the banks react like lemmings to regulator guidance (which probably actually increases systemic risk). Maybe this security could be used to pay employee bonuses with a requirement that they not sell for 3 or 5 years. Maybe X% is a function of the bank's size. I want there to be FDIC insurance. I don't want to have to read a bank's financial statements before opening a checking account. Huge deadweight loss if everyone had to read bank balance sheets. But I want FDIC losses to be minimal and to be fully covered by fees paid by the banks.