Too Big To Fail Needs to Go

demolishing an office tower.jpg

One of the causes of last year’s financial collapse was the adoption of the concept, 'Too Big To Fail'. Washington decided long ago that some firms are so large and so integral to the economy that the failure of one of these firms would put the entire economy at risk. So, the government insures them at no cost.

The problem with free insurance against failure is that it encourages excessive risk taking. This is the much-talked-about moral hazard problem, and it was a serious contributor to how we got to September 2008 in the first place. Since then, we’ve merged big bad financial institutions with big good financial institutions to create even larger financial firms. This has to stop.

Why would a firm grow to the size we observe?

Often, the firms’ managers tell us they merge to diversify. It is not true. Research I did with Bill English while I was at the Fed showed that large banks really didn’t diversify after they merged. They merged with firms much like themselves in similar markets.

Besides, the argument for diversification is flawed on its face. Financial theory is clear. The investor can diversify more efficiently than the firm can diversify on the investor's behalf.

Firms also claim that they are merging to obtain economies of scale. That is not true either. A reasonably large literature is available on economies of scale. This literature is clear. Economies of scale are fully exploited when firms are much smaller than the ones that are currently considered Too Big To Fail. Indeed, diseconomies may exist at the size of our largest financial firms.

Are there other reasons firms might want to become the size we see? Sure, but the participants are not likely to advertise those reasons. Firms constantly strive for market power, and size can help them achieve that market power. Of course, when firms have market power, the consumer loses.

Firms might also merge to get the free Too Big To Fail insurance. That is clearly not in the best interest of anyone except the insured firm.

The two most believable reasons that firms become Too Big To Fail are counter to the public’s interest. That’s worth repeating more forcefully. Firms that are Too Big To Fail serve no public interest. Since the public is funding the insurance, it needs to go.

Washington’s response has been counterproductive. The preferred model seems to be fewer and even larger firms subject to more government regulation. This makes no sense. There is no evidence that regulation prevents financial collapse. The firms that were involved in last September’s nightmare were all heavily regulated. Indeed, they are among the most heavily regulated firms in the world, and we still saw the most devastating financial collapse since 1929.

Additional consolidation and regulation is not only counterproductive, it approaches criminal insanity. It guarantees that we will see something like September 2009 again.

We can only speculate as to why policy makers are responding to the financial crisis by increasing regulation of a consolidated financial sector. The most generous speculation is that fewer larger firms are easier to regulate effectively. Easier, maybe, but not more effectively.

We would all be better off if there were no firms that were Too Big To Fail. So, let’s provide a strong incentive for them to voluntarily split themselves up into little, more efficient pieces. The easiest way to do this is to apply an onerous tax on any firm considered Too Big To Fail.

This would be equivalent to overpricing the Too Big To Fail insurance. If the insurance is overpriced, no one will buy it. Instead, they will divide themselves up into several smaller, hopefully more specialized, firms.

Implementing such a tax would be very easy to do, and it would be far cheaper than the alternatives. We need to get on with it before another crisis comes our way.

Bill Watkins is a professor at California Lutheran University and runs the Center for Economic Research and Forecasting, which can be found at clucerf.org.

Subjects: