Lazear on Dodd-Frank and Capital

Ed Lazear has a nice WSJ oped, "How not to prevent the next financial meltdown." (Also available here via Hoover.) The main points will not be new to readers of this blog, or my much longer essay but the piece is admirable for putting the basic points so clearly and concisely.

The core problem of focusing on institutions not activities:

The theory behind so-called systemically important financial institutions, or SIFIs, is fundamentally flawed. Financial crises are pathologies of an entire system, not of a few key firms. Reducing the likelihood of another panic requires treating the system as a whole, which will provide greater safety than having the government micromanage a number of private companies.

A crisis is a run:

The risks to a system are most pronounced when financial institutions borrow heavily to finance investments. If the value of the assets falls or becomes highly uncertain, creditors—who include depositors—will rush to pull out their money. The institution fails when it is unable to find a new source of funds to meet these obligations.

Nay, a crisis is a systemic run:

A bank’s inability to pay off its creditors can be transmitted to others. The mechanism can be direct: The debtor bank defaults, and its creditors cannot repay their creditors, etc. But the mechanism can be indirect. The suspicion that similar assets held by other institutions are subject to the same downward pressure can start a run at even an unrelated financial institution.

Ok, a minor disagreement here: The dominoes theory -- I fail, I don't pay you, you fail, you don't pay Joe, Joe fails, etc. -- is popular and enshrined in much Dodd-Frank rule making. It simply did not happen. Our financial crises are simultaneous runs, not failure dominoes. I fail, your investors see that and worry you might not pay them back, so they run, and so on. Companies do understand counterparty risk! And even small equity buffers multiply -- For a domino to go from A to E, A's losses must exceed all the combined equity of A, B, C, D, and E. Domino models tend to have large single counterparty exposures and no equity.  But, this is an oped, and it's a story widely told, so I can't blame Lazear for passing it on as a possibility.

The stability of equity:

consider the contrast between the 2008 financial crisis and the dot-com crash in the late 1990s and early 2000s.

The bursting of the dot-com bubble and subsequent failure of many Internet-based companies had serious repercussions for investors, but not for the financial sector. That’s because the failed firms were financed primarily through equity, not borrowed money. Investors took big losses when the value of tech companies fell precipitously. But there were no runs.

Floating-value liabilities also are run-proof:

Mutual funds are similar. Many are large and hold assets that may be risky, but they don’t fail when the value of their assets falls. The liabilities move one-for-one with the value of the assets because the fund does not promise to pay off any fixed amount to its investors. There is no reason for a run: Getting money out first serves no purpose to investors nor does withdrawal of funds cause significant distress. The fund simply sells the assets at the market price and returns that amount to investors.

Mortgage backed securities are fine -- if held long-only in investor's portfolios. It's funding MBS by rolling over overnight debt that causes problems.

The bottom line: equity financed investment and narrowly backed deposits

These factors suggest that instead of trying to divine which firms are systemically important, banks should be required to get a larger share of the funds they invest by selling stock. Bank investment funded by equity avoids the danger of a run: If the value of a bank’s assets falls, so too does the value of its liabilities. There is no advantage in getting to the bank before others do.

deposits—the checking and saving accounts that are bank liabilities—should be invested only in short-maturity secure assets, like Treasury bills.

Good news: These views seem to be taking hold. The people who run the regulatory agencies are pretty smart, they do listen, and they understand better than we do just how unworkable the plan is for them to make sure no big highly levered bank ever loses money again:

The Federal Reserve seems to be wising up, and may require higher equity capital for the SIFIs and place less emphasis on regulation

Additionally, the international Financial Stability Board announced on July 31 that it would set aside work on designating funds or asset managers as systemically important to focus instead on whether their activities or products were systemically important.

The last point is especially important. There has been a little noticed effort underway to designate asset managers as "systemically important." Asset managers buy and sell stocks on your behalf. There is no fixed value promise and no run here. But there is a chorus that worries the asset managers might all sell, herd, or otherwise act with behavioral biases and they need to be regulated as SIFI. If you understand that a crisis is a run, and that the government should not try to prevent any asset from ever losing value, you see this is not such a great idea.