Tuesday, October 14, 2008

Query to New Keynesians -- What do You Have Against the DOJ?

Last week, the Treasury proposed spending some of its revenue purchasing equity in struggling banks. This proposal echoed the proposals of a number of academics. However, none of the academics have discussed whether Treasury capitalization would crowd out private capitalization (one caveat noted below). This irritates one of my pet peeves with New Keynesian economics -- that they propose to use macroeconomic policy to deal with industry-level problems. Even if public policies were generally potent, I cannot imagine that such an oblique approach to a problem would have much impact.

Let's begin with the proposition that the proposed Treasury transactions will be fully neutralized by private sector transactions in the opposite direction, and explore how New Keynesians might attempt to refute it. In the simplest model, future taxes are lump sum (with a known incidence) and the economy is closed – i.e., that all potential bank stockholders are also U.S. taxpayers. In much the same way that taxpayers behave in Barro’s (1974) model, taxpayers will recognize that the Treasury has invested more in bank stocks, and has implicitly done so on their behalf because the taxpayers will reap the gains and pay the losses of those investments. As a result, taxpayers will attempt to reduce their holdings of bank stocks by the same amount that the Treasury increased them.

A number of “realistic” modifications to Barro’s (1974) model have been proposed, but they do not necessarily weaken the basic result that public transactions are at least partly offset by private transactions, and may strengthen it. Consider first the possibility that taxpayer portfolio decisions are at a corner solution, so that taxpayers desire to reduce their holdings of the equity of existing banks but cannot.[1] If bank management were responsive to shareholder demands, banks would use the cash they obtain from Treasury investment to buy back bank shares from the public. In this case, the Treasury purchases put the cash into a revolving door, without boosting bank capital. [I saw a comment by Professor Stein that seemed to suggest that bank executives would raise dividends in order to enrich themselves at the expense of other bank creditors. This may be a concern too, but I think that the Ricardian force is at least as fundamental as corporate governance issues.]

Again for the sake of argument, suppose that Treasury purchases were accompanied by (perhaps implicit) regulations restricting share repurchases and dividend payments.[2] Investors still desire to hold equity in new banks or in alternative institutions that would compete with banks and likely view that equity as (imperfectly) substitutable for equity in existing banks. In other words, the portfolio-at-corner-solution view may explain how Treasury transactions would not be precisely neutralized by private sector transactions, but it predicts that the Treasury plan reallocates capital from new entrants to the banking industry and toward the existing (and struggling) banks. Tino pointed out that smaller banks are resentful of the bailout, because it props up their larger competitors. This reallocation harms the future efficiency of the banking industry.

Another modification to the Barro (1974) model assumes that taxpayers are unaware of what the Treasury is doing, and therefore have no motivation to offset Treasury transactions. This modification may be applicable in some situations, but seems quite inapplicable today when (a) the entire country is focused on the financial crisis and public sector responses to it and (b) taxpayers have loudly voiced their displeasure with the tax liabilities they perceive to be created by the Emergency Economic Stabilization Act of 2008.

One reply to all of this might be to have the Treasury invest in the bank equity across-the-board. At worst, Barro’s model applies and the Treasury investment does nothing. With enough Treasury investment, all of the private capital will be crowded out so that Treasury investment can have a real impact at the margin. Is that amount too much even for the U.S. Treasury? Can we trust that the Treasury will locate all possible competitors to existing banks? Should the Treasury invest in Walmart too (Walmart wants to operate its own bank)?

The Role of Complementarity in the Neutrality Result
By this point, I have pushed New Keynesians to start telling me about details of the banking industry. Unfortunately, the logical escape route in this direction is narrow and unpleasant.

How might industry detail trump Barro’s analysis? Suppose that, when industry capitalization rates become low, each bank’s output is complementary with the others.[3] In this case, the most direct public policies for raising bank output would facilitate cooperation among banks by encouraging mergers or the formation of other private-sector institutions such as clearing houses or commercial paper exchanges to align each bank’s incentives with the industry-level complementarities.[4] Once banks were the proper size, the industry could otherwise be analyzed as if the complementarities were absent (with the same neutrality or non-neutrality results).

Suppose for the sake of argument that mergers and other private sector efforts were insufficient to internalize the complementarity, and that banks can be prevented from buying back shares or cutting dividends. Even so, the impact of Treasury equity purchases depends on the terms of the purchase. Professor Mankiw has proposed that the Treasury co-invest (on a non-voting basis) with private investors who decide “on their own” to make a purchase of a bank’s stock. If (some subset of) taxpayers wanted to invest, say, $20 billion in bank ABC absent the Treasury plan, then there is nothing to stop them to investing $10 billion in the presence of the plan, thereby bringing the total ABC equity sale to $20 billion. In this case, other banks in the industry are unaffected by the Treasury’s purchase, because bank ABC sells $20 billion regardless of whether the Treasury participates. Professor Mankiw’s plan does nothing to align the incentives of the private co-investors with those of the industry as a whole, and is premised on two of the mechanisms that can deliver Barro’s result (that private investors are willing to invest even absent Treasury action and that bank managers are free to make dividend decisions, etc., to the shareholders’ advantage).

In order to use complementarity to predict a significant effect of Treasury purchases on bank capitalization, we must also assume that private investments are at a corner solution. In this case, a judicious and carefully micro-managed choice of Treasury investment will raise the marginal product of capital throughout the industry, and presumably stimulate private investment.

In summary, economic theory reminds us that Treasury transactions are at least partly offset by private sector transactions. Each Treasury dollar spent on bank equity will reduce private ownership of bank equity by some multiple. More research is needed to determine whether the multiple is close to zero, close to one, or even larger.
[1] This logical possibility probably does not accord with the facts, because Bank of America announced in early October 2008 that it would issue $10 billion worth of stock.
[2] Note that these regulations are counter to the spirit of many of the academic proposals, which are to keep the public sector out of bank business decisions.
[3] Presumably the complementarity was not significant at normal capitalization rates, or else banks would have merged or cooperated with each other by contract.
[4] Among other things, I owe these examples to Fernando Alvarez.

P.S. The DOJ is short for the Department of Justice. Most Industrial Organization economist believe that it is the appropriate public forum for alleviating industry malfunctions.


Sheng said...
This comment has been removed by the author.
Sheng said...

What I am thinking is this: why does Warren Buffet think investing alongside with the government is a good idea? Not long ago he invested in PetroChina and became a partner of Chinese government. Now he threw $5 billion into Goldman Sachs (before rescue plan but he apparently anticipated the U.S. government would do something, based upon his afterward comments). Sometimes it might not a bad idea to bank on the government......

Unknown said...

I think of myself as a NK, and, quoting Tyler Cowen:

"Don't forget Mark Thoma's good analysis: 'So, by having the government take a share of any upside, the result may be less willingness of the private sector to participate in recapitalization.'"

It's not Ricardian equivalence (which is a bit murky empirically, so I follow Mankiw and discout fiscal policy by around a half), but it does reflect concern for crowding out private investors.