Sunday, February 12, 2012

Tyler Cowen and Big Banks

Tyler Cowen has an excellent piece in today's NYT. It's the best idea I have heard in a long time about altering the incentives of banks in a good way. As with most good ideas, it's simple:
There is a better alternative [to other approaches to regulating bad behavior]: expanding the liability for major financial institutions. If a shareholder invests a dollar in a big bank, why not make that shareholder liable for the first $1.50 — or more — of losses as insolvency approaches? In essence, we would be making the shareholders liable for the costs that bank failures impose on society, and making the banks sort out the right mixes of activities and risks.
This idea is not new. Indeed, the early 19th century Scottish banking system had unlimited liability, and there were double liability provisions in early Canadian banking. That's not really a coincidence, as the early Canadian bankers were in fact Scots. As is well known to historians, the 19th century Scottish banking system, and the Canadian banking system (in the 19th, 20th, and 21st centuries), were successes, which is obviously important.

For references, first from my Google search, see this World Bank working paper, and this 1937 AER paper. From my archive, read this and this, and this.

One last point. You might wonder how you get the shareholders to pay up on the extended liability provisions. That's easy. They post collateral.

15 comments:

  1. I wonder how big the selloff in financial stocks would be? I would guess pretty huge. Wonder how the effects of that would work out.

    Interesting idea and it doesn't rely on (mythical) Olympian-Psychic-Infallible-Incorruptible regulators.

    ReplyDelete
    Replies
    1. I don't know. Could you make a case that it would go the other way. Establish the rule and the share price goes up?

      Delete
  2. leave it to a libertarian to propose an unconstituional law that has no respect for existing property rights.

    the relationship between bank shareholders and banks are matters of contract, which cannot be impaired, taken without just compensation, etc.

    why oh why can't we have better economists?

    when banks were given charters, no one reserved the right to alter or amend the relationship between shareholders and the bank

    what a non-starter

    ReplyDelete
  3. but isn't it a principal-agent problem? the root of the moral hazard comes from the the guys who run the bank, rather than the shareholders...

    ReplyDelete
    Replies
    1. The shareholders would now have a greater incentive to discipline the management. Principal agent problem mitigated.

      Delete
  4. The Cowen proposal amounts to a forced recapitalization of banks in the form of contingent capital call. This is fine as far as it goes, especially because it alleviates the Agency problem (Boards will more closely oversee managements). However, it leaves problem with the main part of the capital structure -- debt -- intact. Our banking system has converted from one relying primarily on deposits and somewhat on medium-term unsecured debt; to one relying heavily on s.t. collateralized funding. Collateralized funding is dangerous: as Gorton points out, it is vulnerable to self-reinforcing "runs" in the form of escalating margin calls.

    ReplyDelete
    Replies
    1. "However, it leaves problem with the main part of the capital structure -- debt -- intact."

      That's the idea. In the event of default on the debt the shareholders pay off the creditors. It's not the taxpayers paying off the creditors. So the shareholders discipline the bank so it takes on less risk.

      Delete
    2. a far better approach would be to eliminate the business judgment rule, which provides immunity to officers and directors.

      making shareholders liable will not alter how the agents act.

      Delete
    3. "a far better approach would be to eliminate the business judgment rule, which provides immunity to officers and directors."

      Bad idea I think. I'm assuming you want this to apply to everyone, not just banks. That drives all business out of the country.

      Delete
    4. no, just banks

      Cowens idea will not alter bank agent's behavior

      eliminating the business judgment rule will

      as for the idea that firms would leave, banks will do that if you follow what TC proposes.

      Delete
  5. In a system dominated by short term collateralized funding, the Fed can only alleviate systemic risk by effectively issuing a put on collateral values to creditors. Cowen's proposal, by providing more first loss protection, lowers the strike price and value of this put. Great. Still, the put exists. Why should creditors be protected by taxpayers at all? The LOLR function is intended to tide the system over until such time as lemon bank creditors can absorb losses. In a collateralized-credit system, all creditors have exposure to the same risk of escalating collateral haircuts; and therefore all creditors must be protected.

    I'm arguing for Cowen's proposal AND a ceiling on the use of collateralized credit by banks. We should go back to unsecured credit as the main vehicle for non-deposit funding.

    ReplyDelete
  6. "We should go back to unsecured credit as the main vehicle for non-deposit funding."

    Sounds bizarre to me. What do you think collateral is for?

    ReplyDelete
  7. This is the same as forcing shareholders to buy a bond with every share. Two things:

    1) if the collateral is debt for capital purposes, then the shareholders can just repackage and sell the bond portion and you've achieved nothing
    2) if it's equity, then you've just raised the equity capital requirement. If that's what you are trying to do, why don't you just raise the capital requirement?

    I really don't see any point here.

    K

    ReplyDelete
  8. This is interesting but begs a question of what happens on the secondary market.

    If I buy a share for $10 do I put up $15? What if the potential liability attached to that share is $100?

    And if the amount I have to put up in collateral is based on the potential liability of the share (based on debt outstanding per share I guess) then this means the clearing price of some bankds shares may be negative. That's not obviously a problem but it would make a lot of people uncomfortable I think.

    ReplyDelete
  9. I've now read the White paper and it is, indeed, not particularly clever. So shareholders have to post collateral which presumably will entitle the owner to interest or possibly a dividend - it's not clear. The collateral serves to pay off depositors in case of default. I.e. it serves the role of sub-debt *or* possibly preferred shares (can't tell). So the proposal amounts to requiring each shareholder to also buy a unit of sub-debt or prefs.

    If it's prefs then we have increased the equity capital of the bank. That's good, but as I said above, it's a pretty weird way to achieve that: if we want more capital, just demand more capital. If the "collateral" is sub-debt we have improved the subordination of deposits which improves recovery on default but we haven't moved the default trigger since interest on the sub-debt must be paid. But in either case, this is just amounts to a demand for more subordinate capital with the caveat that shareholders must also be the holders of the sub-debt/prefs. Which is the exactly equivalent to requiring more equity. Move along, nothing to see here.

    K

    ReplyDelete