Wednesday, June 24, 2009

Should you claim Social Security benefits earlier if your life expectancy is shorter than average? I don't think so.

U.S. News & World Report's Emily Brandon discusses a number of different Social Security claiming strategies. Check out the details here. It's a good summary of the current literature on claiming strategies, which is developing rapidly.

But here's one place I don't entirely agree with Emily. Here's the passage, in which I'm quoted:

Benefit checks increase by about 7 to 8 percent for each year you delay claiming up until age 70. Retirees who sign up at younger ages get smaller payments over a longer period of time, while those who wait get larger checks for their remaining years. "I would argue they would be better off claiming later because then they get a higher benefit and because you are getting more insurance against outliving your assets because Social Security benefits last as long as you live," says Andrew Biggs, a resident scholar at the American Enterprise Institute and a former deputy commissioner of the Social Security Administration. Of course, if you have a reason to believe you won't live a long life, it's best to sign up right away.

It's the last line, which I've italicized, that I question.

Let's say you're part of a group with known lower-than-average life expectancies – say, African American males. You know that while a typical 65 year old might live to age 83, an African American male on average lives to only 80. On average, then, you'll collect benefits for fewer years. But here's the key: you still have a lot of uncertainty regarding how long you'll live and a higher Social Security benefit protects you against that uncertainty. The chart below, which I put together from life tables from the National Center for Health Statistics, shows that average life expectancies have only limited value; the range of outcomes also matters a lot, but people don't think very much about them. For instance, a 65-year old black male has around a 10 percent chance of living beyond age 92 – for those individuals, a higher Social Security benefit would be of great value in holding off poverty in old age. Moreover, given standard economic assumptions regarding the utility of money, such that you'd be willing to give up some income today to prevent even a small chance of falling into poverty in old age, the case for even black males – the stereotypical low life expectancy group – to delay claiming Social Security benefits looks stronger.

The broader point is that the choice you make shouldn't be premised on your life expectancy versus the average; what does it matter how long other people live? What matters is your own life expectancy and the degree of uncertainty surrounding that point estimate. The more uncertain your life expectancy is, regardless of the average value, the more a higher annuity payment will benefit you.

Another issue to consider is benefits for your surviving spouse: even if you know you'll die young, by delaying benefits you guarantee a higher income for your spouse later in life. These issues are a lot more complex than some of the usual claiming strategies portray. People shouldn't think of choosing their claiming age as a game in which they hope to survive to some "break even age" and then "make money" afterwards. Choosing your claiming age is about making sure you have money when you need it the most.


James said...

But here's the key: you still have a lot of uncertainty regarding how long you'll live and a higher Social Security benefit protects you against that uncertainty.



Fascinating stuff. Your argument seems to be, roughly, while we can make statistical claims about various segments of the population (men, women, different ethnic groups, smokers, people who are obese, etc, etc), an individual still can't be certain of their own circumstances and, because being wrong could leave you broke in old age, it's not a bet you can afford to make. However, there are clearly individuals who can be highly certain that they will die much younger than the general population (eg, 40 year old, obese male, life-long smoker, diabetes, history of heart ailments), and the article may be referring to groups such as this.

Moreover, don't all of these strategies face some prospect of the bet going wrong? For example, under "Marital Strategies" in the article you reference, we have this:

"Although it's generally better for single men to claim early because of their shorter life expectancy and for single women to delay claiming because they are likely to live longer, the opposite is true for couples wishing to boost their total payout. In general, couples can maximize their joint Social Security payouts by having the lower earner sign up as soon as possible at age 62 while the higher earner waits as long as possible to claim, ideally until age 70, according to Boston College. This strategy typically maximizes the couple's benefits. The husband is generally the higher earner, older than his wife, and has a shorter life expectancy, so the wife will get a smaller payout based on her own working record. But then she'll get bumped up to the higher survivor's benefit when her husband passes away"

Yet, this approach relies on the assumption the couple stays together: if they divorce - a statistical possibility - the spouse who took social security early made have made the wrong bet if he/she must rely on social security over a long life span (I don't know what the divorce laws say about social security as community property - this would very from state to state - but this could add yet another critical factor couples need to understand).

I think it would be very instructive to have interactive, web-based tools that allowed people to play "what if" scenarios based on their personal data (man, woman, medical conditions, ethnic group, married, not married) and payroll history to get some feel for these things. I fully appreciate, incidentally, the risks associated with these types of things, but nonetheless they could be very instructive in helping people understand possible strategies.

Andrew G. Biggs said...

James, thanks for the comments. Two quick points:

1. If you really, really knew when you'd die, then yes, you can simply plan on that and the annuity/longevity issue isn't all that important. However, I think the number of people in this category is really pretty small, and this kind of advice is coffered toward much broader groups who simply have different life expectancies from the average.

2. Characterizing things as a right/wrong best is, in my view, the wrong way to think about this. You don't think about home insurance in this way, so I'm not sure you should think about longevity insurance like this. The point is that I'd be willing to give up a statistical likelihood of receiving a little on a lifetime basis if by doing so I could guarantee that, in case I live longer than average, I won't live in poverty.

Put another way, by claiming at age 62 you're taking a 25% haircut on your benefits, which can put you at pretty significant hardship. Moreover, by doing this you're guaranteeing yourself hardship for the rest of your life. By delaying for a few years, you're guaranteeing yourself non-hardship.