Josh Barro has a thoughtful piece on public pensions and risk in the NYT's Upshot section. He makes many points with which I agree, most notably raising cautions about pension fund investments in "alternative investments." These are mostly private equity funds, but can also include venture capital and hedge funds. The problem with these alternative investments is that they come with unknown return distributions (essentially the pension funds have a promise that a smart investor will beat market indexes) and they come with high expenses. The public has good reason to be concerned when their pensions start to go more heavily into these alternatives to make up for funding shortfalls.

The issue where I differ is on how pensions need view the risk in the stock market. Josh notes my comment that pension funds don't need to be concerned about the short-term fluctuations in the market, only long-period averages. His counter is that many funds became underfunded when the stock market plummeted and therefore had to boost funding in the recession. He also notes that many pensions cut back funding and even raised benefits when the stock bubble in the 1990s led to considerable overfunding.

This points are correct, but they stem largely from bad projections about future returns, and I don't mean year to year, I mean long period averages. When price to earnings ratios in the market go above long-term averages, it is not possible to get historic rates of return. This means that the return projections used by pension funds in the 1990s should have been adjusted downward since there was no way on earth they would get the 7.0 percent real (10.0 percent nominal) returns that most funds were assuming.

The same story applied at the peak in the pre-recession period. They should have adjusted downward their assumption on long-term returns to a 5.0-5.5 percent real rate (8.0 percent to 8.5 percent nominal). After the market plunged they should have adjusted their return projections upward, which certainly would have been consistent with the sharp bounceback we have actually seen over the last five years. With these adjustments, pension funds would not have suddenly found themselves hugely underfunded even with the plunge in the market that we saw at the start of the recession.

None of this is 20-20 hindsight. I have been arguing this story about long period stock returns for almost twenty years, first in the context of Social Security and more recently in the context of pension funds.

Barro seems troubled by the idea that governments can benefit by investing in the stock market. It is not clear why this should be troubling, after all individuals benefit by investing in the stock market all the time. And the notion of risk arbitrage comes up in all sorts of different contexts.

For example, the claim that we made money on the TARP bailout is entirely a story of arbitrage. In a time where there was an enormous risk premium associated with lending, we made below market loans to favored banks, where the interest rate charged by the government was still above the risk free rate paid by the government. (The claim we make money on the Export-Import Bank is also a story of risk arbitrage.)

In short, we see instances of the government arbitraging risk all the time. It is not clear what policy we would be advancing if we prohibited it from doing so. We do know that such a prohibition would raise the cost of hiring public employees since they obviously value the guaranteed retirement income from a defined benefit pension.



Typos corrected -- thanks to Robert Salzberg.