A famous research paper, written in 1965 (Yaari, 65) argued that rational individuals should convert all of their retirement wealth to an annuity at the point of retirement. The underlying argument goes something like this. If you have to set up a lifetime income flow, but you don’t know how long you will live, you are forced to plan for an extremely long period of retirement, even though you will almost surely leave some money behind because you set aside money to last longer than you. On the other hand, an insurance company can pool a large number of retirees when setting aside the needed assets. Since some will die before reaching their life expectancy, they will effectively subsidize those who live beyond it. As a result, the insurance company can manage this pool of assets with a much lower reserve (as a percentage of the total liability) than you would as an individual who is in effect, self-insuring against longevity risk. This is a great deal because the insurance company can make a fair profit and you are much better off, in the sense that you are free to spend a much greater share of your money. You get more utility from your savings and still use less of it to cover the extreme cases, like living beyond 100.
About 20 years later, Franco Modigliani gave his acceptance speech after winning the Nobel Prize in economics in 1985. In that speech he noted that,
“It is a well known fact that annuity contracts, other than in the form of group insurance through pensions systems, are extremely rare. Why this should be so is a subject of considerable current interest. It is still ill-understood.”
In addition, surveys show us time and time again that retirees report a strong preference for a guaranteed stream of income, as opposed to a risky pool of assets in a retirement account. However, very few people actually exchange their risky pool of assets for the annuity. You always say you want it, but you never actually buy it. On a related note (Brown 2007) reported that by delaying the start of social security payments, workers can effectively buy a greater annuity value at extremely low cost. Time and again we see that the overwhelming majority of workers would get more income by waiting until the “full retirement age” to start collecting. However, only about 5% of the workforce does so. Again, you tell me you want a larger annuity but you don’t actually select it.
These are two facets of a problem known as the “Annuity Puzzle.” We all claim that we want to avoid risk, and we do many things that are consistent with this idea. But when we are presented with a chance to make a great risk go away through the purchase of an annuity we respond with, “Nah – I’m good!” It doesn’t seem to make sense.
However, it turns out that there is a lot of nuance lost in this presentation. Let me touch on a bit of it here. First, researchers have found that there is a huge difference between a setting in which the investor has to shop for an annuity and one in which it is offered as part of a screened set of options. For example, one paper from 2006 (Hurd, and Panis, 2006) found that when offered an annuity that the employer had already defined, 61% of retirees selected the annuity over a lump-sum option. Recent data on 18,761 employees who retired from IBM between 2000 and 2008 showed that 88% chose the full annuity option, while only 5% took all of the payout as a lump sum.
A much larger study by Benartzi, Previtero, and Thaler considered 103,000 payout decisions from 112 different plans between 2002 and 2008. In this set they found that roughly half (49%) of participants selected an annuity over a lump sum payment. However, they also found that when the account balances were small ($5000 or less) the decision maker almost always selects the lump sum option. It seems likely that this follows from the simple fact that annuitizing small amounts is simply not worth the trouble of doing so. As a result, the reported proportion of employees choosing the annuity is actually an understatement of how many would if they had enough to make that option viable.
Another payer by Brown, Kling, Millainathan, and Wrobel (2008) found that when decision makers were told that an annuity would deliver “$650 of monthly spending” 70% of decision makers selected this option. But when a similar population was offered “a return of $650 per month for life” only 21% found that option attractive. In other words, presenting an annuity as an “income stream” makes it much more attractive than presenting it as an “investment that only pays as long as the investor lives,” even though the resulting cash flows are identical. This is an example of what economists call “framing the problem.” How the option is presented matters.
Stated differently, if a consumer thinks of a choice between a lump sum and a corresponding stream of payments, the lump sum feels like a “sure thing” whereas the value of the annuity depends on how long the consumer lives. The annuity avoids a return-rate related risk when compared to investments like stocks or bonds. But the annuity still feels “less real” and less salient than a lump sum and therefore “feels more risky”. When we argue that retirees should take the annuity to reduce risk, we are ignoring the fact that risk is not a one-dimensional problem. Sequence of returns risk like we face with a stock portfolio at the start of retirement is one type of risk, but not being able to hold our money in our hands like we can with a lump sum is another type of risk. Mathematically, this may seem silly, but emotionally it is quite real.
This speaks to a different, and larger problem. The early leaders in finance research in the US and UK settled on a definition of risk which is in its essence synonymous with Variance or Standard Deviation. But this is never the definition that makes for sleepless nights for retirees. Think about two alternate scenarios. In Scenario A you give me $100,000. Exactly on year from now I will give you back either $120,000 or $90,000. Compare that with Scenario B. In this setting you give me $100,000 today and a year from now I will give you either $200,000 or $2,000,000. No one in his right mind would choose Scenario A over Scenario B, even though the variance of the return is roughly 10 times greater. When a human being (not an economist) thinks about risk, they are looking at the probability of a loss, and the magnitude of the loss – whether this has a higher variance or not.
Finally, some sort of mental accounting may be in play as well. Most retirement plans would require the retiree to, in effect “write a huge check” to purchase the annuity. From a business accounting standpoint, we are merely moving an asset from one type of account to another. But from a psychological perspective this “feels” like writing one big check in the hope of getting a series of smaller checks down the road. If you think you have something in hand, you are going to have an emotional response to “giving it up” even after we tell you over and over that what you get in return is of comparable financial value.
Final Word
Unfortunately, I don’t think any of this answers the fundamental question of whether we have an annuity puzzle or not. Most people love annuities when presented in the right way. However, it appears that the style of the presentation is much more important than it seems like it should be. As a result, there is not universal answer on whether this is a good idea or not.
Here is my “feeling” on the matter. I was born in 1963. Prices as reflected in the CPI have gone up by a factor of 10 over the past 60 years. Dividends on an S&P 500 index have grown by a factor of 33 over that same span of time, and the level of the index itself (assuming that I spent all of the dividends) has gone up by a factor of 76. An annuity bought at that time has risen in value by a factor of 1. If the payments never rise, the real value never rises either. I would prefer to live with the “risk” of the index over the guaranteed loss in purchasing power that comes with the annuity.
REFERENCES
- Benartzi, Shlomo, Alessandro Previtero, and Richard H. Thaler. “Annuitization puzzles.” Journal of Economic Perspectives 25.4 (2011): 143-164.
- Brown, Jeffrey R. “Rational and behavioral perspectives on the role of annuities in retirement planning.” (2007).
- Brown, Jeffrey R., et al. “Why don’t people insure late-life consumption? A framing explanation of the under-annuitization puzzle.” American Economic Review 98.2 (2008): 304-309.
- Hurd, Michael, and Constantijn Panis. “The choice to cash out pension rights at job change or retirement.” Journal of Public Economics 90.12 (2006): 2213-2227.
- Yaari, M. E. (1965). Uncertain lifetime, life insurance, and the theory of the consumer. The Review of Economic Studies, 32(2), 137-150.
