In Sunday’s New York Times, Professors Hu and Odean published an op-ed entitled “Paying for Old Age.” After reading it, I concluded that it needs a response, as I am afraid that the authors have misdiagnosed the main problem.
First, let me say that I am a big fan of Odean’s academic work, including that which uses proprietary brokerage data to provide insights into individual investor behavior. And Henry Hu is also a highly respected law and finance scholar who is well-known for his work on financial risk. I am delighted to have more super smart academics taking an interest in the important issue of guaranteed lifetime income (and I don’t mean to imply by “more super smart academics” that I consider myself super smart! Rather, this is a topic that has been researched by some of the brightest economic minds of the last 25 years, including Nobel Laureates Peter Diamond and Franco Modigliani, as well as other leading economists such as Jim Poterba, Doug Bernheim, Amy Finkelstein, Olivia Mitchell, and many others.)
In essence, Hu and Odean are suggesting that the U.S. government should issue inflation indexed life annuities directly to the public. They are not the first to propose this – the Aspen Institute had a similar proposal several years ago.
I understand the motivation of their proposal. After all, we know that counter-party risk is one reason that individuals may be concerned about entering into long-term annuity contracts with insurers, and this is especially true after the recent financial crisis that witnessed the disappearance of venerable financial institutions. Perhaps even more importantly, we know that many 401(k) plan sponsors have concerns about fiduciary liability when it comes to choosing an annuity provider. Having the government provide the annuities directly is meant to address this concern. So it is fairly straightforward to write down a simple economic model in which I can show that optimally structured government intervention appears to make people better off.
Even so, I think the proposal is off base, for several reasons.
First, I think they have misdiagnosed the reason people do not buy annuities. While concerns about counter-party risk are certainly heightened today, the demand for annuities was ridiculously low even a decade ago when most consumers were not even thinking about the issue. Nor is there much evidence that the lack of inflation protection or high prices are the primary reasons for limited demand. While these probably contribute a bit on the margin, they are only three items on a very long list of reasons that demand for annuities is limited, which I have discussed at length elsewhere.
Second, I have no confidence in the government’s ability to run this program effectively. The same holds true for the idea of the government providing a government backstop for private annuity providers. Last year I published a book (a collection of papers from an AEI conference that I organized two years ago) that includes analyses of 6 major government insurance or reinsurance programs, including programs to insure DB pensions (PBGC), bank deposits (FDIC), crops, terrorism, floods and natural catastrophes. As I discuss in my intro chapter to that book, one of the themes that came out of these analyses is that the U.S. government simply does not seem to be capable of structuring insurance programs in a manner consistent with basic economic principles. They almost universally fail to charge appropriate prices for the insurance (in two ways – premiums are too low on average, and they are not properly risk-adjusted), which distorts incentives and leads in some cases to excessive risk-taking (a form of moral hazard). They tend to create large unfunded liabilities that put taxpayer funds at risk. There are other problems as well (which you can read about if you read the book!)
So while an optimally-designed government backstop would have real value, the U.S. government has an abysmal record of optimally designing such systems. So we are immediately thrust into the world of “second best” policies where it becomes difficult to ascertain whether a poorly designed program is really better than none at all. I am extremely reluctant to run the “experiment” because, to paraphrase Milton Friedman, “there is nothing so permanent as a temporary government program.”
Third, despite the authors’ attempt to sell this as an “everybody wins” idea, I think it is pretty clear that this would crowd-out private annuity provision, and all the future innovation that may come out of it. (Disclosure: I am a trustee for TIAA, one of the world’s largest annuity providers. I have also done work over the years for many other life insurance companies. )
Finally, I believe that other market-based solutions are emerging. There has recently been a lot of discussion, and some activity, of “multi-insurer solutions,” in which a plan sponsor enters into an agreement with multiple insurers who essentially each agree to kick in to cover one another in the event that one provider experiences financial distress (of course, this does not help if the whole industry goes down.) The idea is still young and new, and I would hate to kill the innovation by starting yet another government program.
In essence, I do not think that direct government provision of annuities is necessary or desirable. The problem in the private market is not that the products do not exist, it is that people do not buy them. Nothing in this proposal will change that. We would be better off focusing our efforts on policies such as reducing fiduciary risk to plan sponsors that would like to provide annuity options to employees. Or reforming our minimum distribution requirements so that they no longer discourage annuities.
Bottomline – rather than having the government provide annuities, I would like to see the government stop discouraging the use of annuities that private providers would be happy to make available.