Betting the Business

A new report by Dean Baker on “The Origins and Severity of the Public Pension Crisis” is getting a lot of attention. It’s hard not to notice that public employees’ pension benefits are under attack. Baker’s piece is a thoughtful look into the numbers and the economic forces behind the numbers, and they make a useful contribution to an informed approach to the issue. However, while he gets some things right, he gets one thing wrong, and it is useful to separate the wheat from the chaff.

Baker makes essentially three points about the shortfall between pension assets and liabilities. One is about looking back, to see how we got here. A second is about where we are now, assessing the true size of the current shortfall. The third is looking forward, to see what it will take to remedy the situation.

1. Looking back. The current pension shortfall is due to the Great Financial Crisis and the ensuing Great Recession. It results primarily from the drop in stock prices and the value of other financial assets between 2007-2009, and from the decline in state and local contributions that follow from the decline in state and local revenues caused by the recession. The underfunding that pre-existed the financial crisis is small in comparison.

2. Where we are. The shortfall equals the difference between the value of the assets and the value of the liabilities. While it is relatively easy to achieve a consensus on the value of the assets, the value of the liabilities is a tougher nut to crack. The value depends crucially on the discount rate employed. Traditionally people have used a relatively high discount rate, which lowers the calculated value of the liabilities, and lowers any measured shortfall. Many economists advocate using a much lower rate, which raises the calculated value of the liabilities, and raises any measured shortfall. Baker advocates using the higher rate.

3. The shortfall is manageable going forward. Baker calculates the share of future state product required to make up for the current shortfall. It’s a small number. We can do this, he is telling us.

I appreciate Baker’s useful calculations on points #1 and #3. Unfortunately, he is just plain wrong on point #2. Baker blithely ignore fundamental principles about risk and return, like those mentioned in a post I made on the subject before Baker weighed in.

The key issue which Baker and others trip over is how to factor risk into your evaluation of long-term expected returns. It is a fallacy to think that a long horizon allows one to more confidently expect to achieve the mean return. That only happens if stock prices are substantially mean reverting, which they are not — at least not at the time horizons and on the scale that is relevant for this debate. What is overlooked is the danger of compounded negative returns. Bad outcomes can get even worse. The Free Exchange blog at the Economist has a nice and concise post on this. Another is by Josh Barro here. One of the prominent economists on this issue, Joshua Rauh, recently testified on the matter before Congress.

One version of this mistake is known as the St. Petersburg paradox: faced with the opportunity to play a repeated fair bet, where the stakes double each round, there is a strategy available which, on its face, seems assured of winning an arbitrarily large positive payoff. Baker’s stance suggests he would buy into the St. Petersburg lottery.

Baker and others — such as Felix Salmon at Reuters — have made mention of the idea that public employee pension funds are a mutual insurance policy, as opposed to an agglomeration of 401K accounts. This is a potentially fruitful basis for thinking the problem through differently. It shows the limited import of the headline number on the balance of assets and liabilities in a public pension plan. While stock returns are not mean reverting, GDP could be–although that is debated. How to account for the social claims on GDP through time is a more difficult problem than how to value the balance sheet of a pension plan. But while that puts the headline number into context, it doesn’t change the headline number. The assets remain the assets, the liabilities remain the liabilities.The claim about a mutual insurance policy is a claim that the correct balance sheet includes other assets and liabilities not yet recognized. It is not a claim that these assets and these liabilities should be valued using the expected rate of return on stocks.

The broader perspective of a mutual insurance policy is why I think Baker’s points #1 and #3 are worth making. From a purely technical point of view focused just on valuing the assets and liabilities in the pension fund, points #1 and #3 are not completely distinct from point #2. But they highlight the limited frame of reference at hand in utilizing the answer to #2 as a complete answer to the public policy problem of what to do about pensions.

Others have pointed out that privatized defined contribution 401K accounts are an inferior way of providing retirement benefits. I think this is true. But this is a different issue from selecting the correct discount rate for evaluating the assets and liabilities of the current pension system. It speaks to what we do with the information, but not to the statement of the information itself. There is no way to get from any of the arguments in favor of the current defined benefit pension system to the current policy for selecting a discount rate on the pension’s liabilities. These are different things. The politics of the current situation has thrown them into the pot together, but good and thoughtful analysis should keep them distinct.