• Retirement Plans
• Executive Compensation
• Health & Wellness
• Instant Insight
• Insight Subscription

Email Print

Retirement income – the axiomatic case for annuities

Nov 12, 2009

This is the second in a series of articles on the broad issue of "retirement income." As noted in our last article (Retirement income – fundamental issues), the phrase "retirement income" covers a host of sub-issues – annuities vs. lump sums, defined benefit (DB) vs. defined contribution (DC) plans, adequacy, "leakage" and the respective roles of the individual, the employer and the government in providing for retirement income.

In this article we lay out the axiomatic case for annuities as the investment instrument for providing retirement income. Summarizing: where, over any particular period, the objectives are to maximize lifetime income without risking the possibility that you will outlive your assets, an annuity (vs. self insurance alternatives) will always provide the greatest income. That is because "gains" from deaths at the beginning of the period can be used to increase the income of those living to the end of the period. This article simply un-packs what, given those premises, is axiomatically "true."

When we come to discuss policy initiatives, we will see that policymakers, to some extent at least, buy this case. And there are proposals in Congress that explicitly seek to encourage the use of annuities.

Yet the fact is that very few plan participants – either in defined contribution plans or, where there is a choice, in DB plans – elect annuities. That fact of participant behavior cannot be disregarded. And so we conclude this article with a discussion of possible bases for that behavior.

Longevity risk – the unpleasant choices for those blessed with long life

The focus of this article is the risk of living – the danger that that risk presents and the ways to reduce that danger.

Longevity "risk" is not a very intuitive topic. People generally want to live long and resent having to pay to avoid that "danger." So let's be really clear, right at the start, about what we're talking about: "longevity risk" is the risk of being old and poor. Not, say, age 70 and poor – retirement savings can prevent that, and, for that matter, even if the individual has no retirement savings, he or she can (in many cases) go back to work at 70. Instead, we're talking about being, say, 85 or 90 and poor. And not poor because of inadequate savings during the individual's working life. But poor because the individual lived longer than he or she expected. The retirement savings have been spent, and the individual now has nothing.

This is a risk that almost all people face. In this article we consider financial strategies for dealing with that risk.

What are the choices?

We are going to consider three different general strategies for dealing with longevity risk:

• "ignoring" it, that is, depending on someone else (family, Social Security, state welfare systems) to finance it;
• self-insuring;
• or buying an annuity.

Much of our analysis will focus on the relative efficiency of self-insuring vs. buying an annuity, but let's begin with the first strategy – in effect not providing financially for the possibility that you will outlive your retirement assets. In that case, if you live "too" long, then you will wind up old and poor. What happens then?

Ignoring longevity risk

If an individual is old and doesn't have any assets, one (or more) of three things happen. Either he or she: (1) moves in with his or her children or other family; (2) lives off of Social Security; and/or (3) lives on some combination of state welfare programs (e.g., a state-provided nursing home). Let's consider each of these alternatives in turn.

1. Living with family. The retiree may have a large and strong family, with a tradition, for instance, of grandparents living with parents and children. In this case, not only does he or she have a viable and non-financial alternative to longevity insurance, but he or she is also likely to have a positive motive in favor of preserving retirement assets for a legacy. If the individual does not have family members he or she can realistically live with, then alternative 1 is not a solution to longevity risk.

2. Living off Social Security. Let's assume the individual is entitled to the maximum Social Security benefit. In 2009 that benefit is about $2,300 per month. If the individual lives in a low cost-of-living community and is prepared to get by on "very little," then alternative 2 may work.

3. State welfare. Finally, some states and municipalities have quite livable publicly funded "old age" homes. And there are, varying with the state, a variety of sources of assistance and special "seniors" benefits on which an older person, with no resources, can depend. At least at this time, in this country, there is still a safety net.

Realistically, if the individual is going to ignore longevity risk, i.e., not insure (either with an insurance company or self-insure) against outliving personal retirement resources, he or she is probably counting on some combination of alternatives (2) and (3) and perhaps, depending on family circumstances, alternative (1).

Moreover, an individual can always "self-insure" for some period of time, that is, make sure that he or she doesn't spend so much that he or she runs out of money by, e.g., age 75. Many people just assume they'll live until, for instance, 85 and spend on that basis. They are, in effect, self-insuring for the period age 65 to 85.

We can't help remarking, at this point, the similarity of the longevity risk issue to some of the issues currently being debated with respect to health care. In both cases, some individuals may have a strategy of burdening the social welfare system if the unexpected happens. With respect to longevity risk, the unexpected is living too long. With respect to health care, the unexpected is a health crisis. And, as we discussed in our prior article, at some point, as the individual ages, it may be hard to distinguish between retirement expenses and health care expenses, and the unexpected becomes inevitable.

The point of the foregoing is to make clear that, if you don't either self-insure against the possibility of outliving your assets or buy an annuity, then you are opting for one or more of these three alternatives: living with family; subsisting on Social Security; or living on public assistance of one sort or another. Now let's consider the alternative: providing financially for the risk that you will live a long time.

Self-insurance and annuities

The rest of this article is primarily concerned with the relative merits of providing for longevity risk either by self-insuring or by buying an annuity. Let's begin by defining what we mean by "self-insurance" and an "annuity."

What is self-insurance?

Let's start with the ridiculous. If you really wanted to self-insure against any possibility of outliving your retirement resources, you literally couldn't spend a dime. Because you "might" live to, say, 1,000. That's absurd. So, let's start with a more realistic assumption. Let's assume that you are primarily concerned about the risk of living to age 95. If you live past 95, then somebody is just going to have to take you in. But up to 95, you're going to make sure, out of your resources, that you have adequate income. In that case, self-insuring simply means spending your resources at a rate that (more or less) insures that you will have income through age 95.

Pretty much everyone who doesn't buy an annuity self-insures for some period. That is, you don't (typically) turn 65, quit your job and move in with your kids. For some, the period of self-insurance may be shorter than for others. And you might argue that in many cases this behavior is not really self-insurance as such, but simply a disregarding of the possibility that your money might run out before you die. That is, this sort of person simply chooses to spend his or her resources at whatever rate he or she likes, completely without regard to whether he or she might outlive them.

We assume this sort of person – the "reckless retiree" – is in a minority. For purposes of this article we're going to assume that the typical individual, at 65, assumes that he or she is going to live into his or her 80s and is prepared to consider, at least, self-insuring for that period – say, 20 years. So that the individual will spend his or her resources at such a rate that they will last at least until age 85.

What is an annuity?

As we use the term in this article, an "annuity" is a guarantee by some other entity (e.g., a pension plan or an insurance company) of a specific income "for life," that is, with payments beginning on some date and continuing until you die.

Comparing alternatives

Let's now quantify the financial alternatives available for mitigating longevity risk – self-insurance vs. annuity.

Let's assume you have $1 million, are age 65 and in average good health. You want to insure, via self-insurance or an annuity, for the possibility that you will live longer than your "life expectancy." Let's consider four alternatives: (1) you self-insure for the period age 65 to 85 and disregard longevity risk thereafter; (2) you self-insure through age 95; (3) you self-insure for the period age 65 to 85 and buy an annuity for the period after age 85; or (4) you buy an annuity beginning at age 65 (the number provided here is for a DB plan, not a retail annuity).

The following table summarizes the results for each alternative.

Explaining these results

Let's first review what each of the items in the table "mean," before considering why they come out the way they do.

Self-insurance age 65 to 85 – means you spend all your money, at a rate of $6,501 per month, to age 85. If you live past that age, someone else has to pay your living expenses.

Self-insurance age 65 to 95 – means you spend all your money, at a rate of $5,267 per month, to age 95. As discussed, this is more or less the equivalent of fully self-insuring longevity risk, that is, survival past age 95 can be realistically ignored.

Self-insurance age 65 to 85 + annuity 85-death – means you take some of your $1 million (about 6-7% to be precise) and buy, at age 65, an annuity that is payable for life beginning at age 85. This annuity is relatively cheap, because it doesn't pay anything unless you at least survive to age 85. You then take what's left over and spend it, at a rate of $6,091 per month, to age 85.

Immediate annuity beginning at age 65 – means just a regular old garden variety life annuity. As indicated, our immediate annuity number is based on what you would get out of a DB pension plan using current "standard" actuarial assumptions. Insurance company and regulatory overhead will add costs to annuities and bring the annuity income number down somewhat.

Why are annuities such a better choice?

If your primary objectives are to maximize retirement income and minimize the chance of outliving your savings, then the numbers in the above table clearly indicate: you should get an annuity.

Why are annuities such a better deal? Here's the explanation in a nutshell. Given those two objectives, maximizing retirement income and minimizing the chance of outliving your savings, annuities are the most efficient investment because – all other things being equal – they share the risk of outliving savings within the annuity pool. Those who die "early" pay for those who die "late," in the same way that, with car insurance, premiums paid by those who don't have accidents pay for those who do.

Given as primary objectives the maximization of retirement income and the minimization of longevity risk, everyone wins. Because everyone in the group secures a lifetime retirement income, and that security costs less than if each individual self-insured against the risk of outliving their savings. That cost savings translates into a bigger retirement income.

*     *     *

That is, as we describe it, the "axiomatic" case for annuities. Reviewing it, you would think that everyone would buy one. But the fact is, very few individuals do. Consider – only 2% of the income of current retirees comes from private annuities. (Increasing Annuitization in 401(k) Plans with Automatic Trial Income; Gale, Iwry, John and Walker; 2008.)

In what follows we consider the reasons why retirement plan participants don't use annuities. We begin with a couple of problems as to the general nature of which there is agreement – individuals with compromised longevity and carrier insolvency risk. We then turn to explanations of participant behavior.

Compromised longevity

The foregoing, "axiomatic" case for annuities holds for someone in average good health. Clearly, if you realistically expect not to live to an average life expectancy – that is, your longevity risk is, for some reason, compromised – then you should generally not buy an annuity. It's a bad bet. If, as you expect, you die "early," you will wind up subsidizing those who die late.

Solvency risk

If you take the annuity approach, then a third party will be paying you an income for life. Generally that third party is either an insurance company or an employer-sponsored pension plan. Both insurance companies and (corporate) pension plans are subject to solvency rules – requiring them to fund benefits at certain minimum levels. Moreover, many states provide insolvency funds for insurance companies that go "bankrupt," and the Pension Benefit Guaranty Corporation generally insures corporate pension benefits up to certain minimum levels.

This is not the place to undertake a thorough examination of insolvency risks with respect to annuities. Suffice it to say, if you are taking an annuity option, you will want to evaluate the possibility that, under some circumstances, the third party annuity provider might not be able to make payments.

Explanations focusing on aspects of participant behavior

Concerns about compromised longevity and carrier insolvency risk cannot, however, explain the overwhelming rejection of annuities by participants. So let's now consider more general explanations for this behavior. Probably the most common explanation is some variation on the theme that "participants have an inadequate understanding of annuities" and so, in effect, are not making rational choices. For instance, Gale, Iwry, John and Walker remark that "consumers are unfamiliar with [annuity] products, often have misperceptions or biases against them, or may be unwilling or unable to make the effort required to make sensible choices."

In recent years, the field of “behavioral economics” has provided insights as to why individuals do not always act in accordance with “rational” principles. Researchers have identified a handful of “cognitive biases” in people that explain how, in the real world, people end up making choices at odds with what a “purely rational” actor would do. While we recognize the fruitfulness of this perspective, we are resistant to the idea that low uptake on annuities is simply a result of irrational decision-making.

We'd like to try to answer this question – why don't participants take annuities – with a starting assumption that such widespread participant behavior is not necessarily irrational – that our axiomatic analysis has missed something. We suggest several possible explanations.

Individuals view early death as "losing" in an annuity system

If you look at the actuarial tables, over 15% of a typical age 65 annuity cohort will die in the first ten years. They will "lose" – by dying early – a significant portion of the value of their benefit. We put "lose" in quotes because the axiomatic argument does not regard this as a loss. The participant is dead and so does not need any more income. And the participant could have had no fixed legacy intention with respect to the "lost" money – if (as was in fact more likely) he or she had lived to a normal life expectancy, there would have been no money "left over" to leave to heirs and beneficiaries.

Our guess, however, is that people don't think like that. Just as a stockholder or homeowner is reluctant to sell at a loss, human beings are reluctant to take a substantial risk that money they worked hard for will simply go to strangers, even after they're dead. So the idea that members of the cohort who die early will increase the income of those dying late just isn't that appealing. Humans retain, and act on, a proprietary view of their money, even after their (anticipated) death.

It’s hard to find a parallel in the insurance world, where a substantial premium (the cost of the annuity) is paid up front, with the risk that the policy becomes instantly worthless.

Annuities limit flexibility

Simply, if money is "locked up" in an annuity, there is no (or, at least, less) money available for emergencies or, for that matter, for something special.

 Individuals place a lower value on life after 85

That's putting the proposition harshly. More nuanced: a 65 year old does not regard the financial challenge presented by living past 85 as a particularly high priority. Perhaps an analog to this is the reluctance of young people (say, individuals in their 20s) to buy health insurance. In both cases, the likelihood of catastrophe – a health catastrophe for a young person or a longevity catastrophe for an old person – is small enough that the individual feels comfortable ignoring it. (This argument reflects the behavioral economics notion of “hyperbolic discounting.”)

And, as with health care, there's probably an unstated assumption: If something happens (e.g., if I live to 95), the safety net (Social Security or the welfare system) will take care of me. Thus, as with health care, there may be an assumed transfer of longevity risk, for many, from the individual to society as a whole.

This explanation suggests an obvious policy solution (again, as the health care problem does as well). Since individuals are reluctant to provide voluntarily for this risk, they should be required to do so (e.g., via taxpayer financed Social Security and Medicare benefits).

*     *     *

In our next article we will discuss various concrete policy initiatives that address retirement income issues.