article in National Tax Journal, he found that the system of taxing annuity payments doesn’t provide enough of a disincentive
to explain low rates of annuitization (although that didn’t stop
him from proposing a more efficient method of taxation in the
article). Is it because married couples have a lower incentive to
annuitize? In another article, Brown found that they do, but
they’d still be better off buying an annuity.
After exploring theoretical reasons why someone might
rationally choose not to buy an annuity according to economic
theory, Brown decided to look more closely at retirees to see
whether economic models could predict which ones actually
bought an annuity. After estimating how much value each
household should place on annuitization (for example married couples will
place slightly less value on an annuity),
it turns out that economic factors that
should make people want them more do
in fact increase demand for annuities.
One exception is that those who feel
more strongly about giving a bequest
aren’t actually less likely to annuitize.
Annuities provide a smooth income in retirement, but sometimes a retiree’s expenses
aren’t so smooth. One reason annuities
are so rare in the U.S. may be the significant spending risks we
face—particularly the risk of high health care costs. Spending
money on an annuity can put these assets out of reach. In his
second article published in American Economic Review (AER),
Brown and his co-authors find that the optimal strategy is to
simply buy insurance to protect against health risks including
long-term care costs and then annuitize the rest. If the ability
to buy market insurance is limited, annuities can still be superior to holding money in an investment account if the health
risk occurs later in life when 401(k) wealth might be depleted.
Brown concludes in his second AER article that there must
be some behavioral explanation for the low level of annuitization since he’s running out of rational economic reasons
why people don’t buy them. In his third AER paper, he explores the possibility that retirees who have been so focused
on viewing their 401(k) accounts as an investment will have
a hard time now viewing them as an income stream.
In an experiment, Brown and his co-authors ask people how
much of their retirement assets they’d invest in an annuity (they
don’t call it an annuity; they just describe what the product
does). It turns out that people really want annuities when the
product is presented in an income frame as providing $650
a month for the rest of your life. But when it was presented
as an investment that provides a regular monthly return but
stops making payments when you die, people didn’t like the
product. This is consistent with a later study that shows that
many people place a really high value on the annuity income
they get from Social Security—in many cases higher than the
amount they’d be willing to pay for a private annuity.
A possible policy solution to increase rates of annuitization
is to nudge retirees toward buying annuities. One example
might be to make partial annuitization of 401(k) assets a default
option at retirement or at a certain age. This would result in
much higher rates of annuitization among
retirees, but some have complained that
pushing lower-income households to buy
more annuities may be harmful if they
don’t live as long as other retirees. Brown
finds that while wealthier annuitants do
live longer, lower-income retirees would
still benefit significantly from annuitizing
and would certainly be better off than if
they hadn’t annuitized at all.
In recent research conducted with MIT
professor Amy Finkelstein, Brown looks
more closely at why people also don’t
buy long-term care insurance. A rational
retiree should buy a life annuity and an insurance policy that
pools the risk of unanticipated health expenses. Retirees aren’t
buying enough annuities, and they’re also not buying as many
long-term care policies as they should. The authors estimate
the significant value of long-term care insurance (especially
for women) and even point out a few problems in the industry.
In a subsequent paper, Brown finds that many don’t buy long-term care insurance because they underestimate their need
for coverage, and will rely on savings or the help of others,
and also because they either don’t trust insurance companies
or believe that coverage is too expensive.
Jeff Brown continues to study ways to improve the defined
contribution retirement system in the U.S. He’s one of the few
academics who remain productive out of a sense of curiosity and
a commitment to creating better economic policy backed up by
theory and evidence. Moreover, he has served as a mentor to
many other scholars in the field of retirement economics. His
influence on that field will be felt for a long time to come.
Michael Finke, Ph.D., is a professor and coordinator of the doctoral
program in personal financial planning at Texas Tech University.
annuities are so
rare in the U.S. may
be the significant
spending risks we
the risk of high
health care costs.