In this article we consider the issue of how a 401(k) balance can or should be paid out at retirement, from the participant’s point of view. What sorts of payment options are available and what are their pros and cons? For purposes of this article we’ll consider all the options a participant could be offered in a plan — that is, we’re not going to limit the discussion only to those that are typically offered.
Risks: inflation, asset performance and longevity
The challenge to the retired participant is to convert her account balance into a stream of income in retirement. Ideally, that stream of income should fit her anticipated spending preferences. In that regard, there are (at least) three significant risks the retiree faces – three factors that will determine how adequate her retirement income will be and whether it will last until her death:
1. Longevity risk. The participant does not know how long she will live.
2. Asset risk. The participant does not know how much money she can/will earn on her account balance between retirement and death.
3. Inflation risk. The participant does not know how much of the nominal value (purchasing power) of her account will be lost to inflation between retirement and death.
Oversimplifying (and we will be de-oversimplifying this below), with an annuity strategy, the first two risks (and conceivably the third) are in effect outsourced to an insurance company. With a draw-down strategy, the participant manages these risks on her own – an approach that gives her greater flexibility and the possibility of reaping rewards for taking these risks.
One may think of buying an annuity as simply buying the opposite side of each of these risks. If the participant dies before she reaches her life expectancy, she loses. If the asset markets do really well, she loses. If she pays for inflation protection and there is no inflation, she loses.
There are other risks and other ‘non-risk’ factors that also come into play. For instance, the risk that something catastrophic will happen or a desire to ‘do some things you always wanted to do,’ either of which may require a big, one-time expenditure. Or, the desire to provide a legacy.
As we discuss different payout strategies, the critical distinction with respect to each will be: what is happening with these risks – in effect, what are the tradeoffs the strategy implements between taking and not taking one of these risks?
In what follows it will be useful to use an example. We’re going to use a 65 year old female participant, with a $100,000 balance, in normal health.
There are a variety of annuity products available, but let’s begin with the simplest, a single premium immediate annuity (SPIA). Under an SPIA, payment of a fixed amount begins at retirement and ends at death. This sort of payout strategy generally provides the biggest first-year annual payment, so in that one sense it maximizes retirement income. The insurance company takes the longevity and asset risks. The participant takes the inflation risk. Let’s consider those risks one at a time.
Longevity risk. If the participant dies before the insurance company expected (in its life expectancy tables), then she generally has less income in retirement than she (or her heirs) would have if she had not bought an annuity. But, except in obvious circumstances (e.g., where the participant knows that she is in ill health), the participant does not know in advance that she will ‘die too soon.’ It’s fair to say that most people (if they thought about it) would be prepared to pool (insure) longevity risk, if they could do it without giving up asset risk. Generally, however (and with some exceptions we’ll discuss below), insurers take both or neither.
Asset risk. Annuities are priced based on bond rates, which are generally lower than the expected returns from, e.g., a 60/40 equity/fixed income portfolio. Buying an annuity buys security — the insurance carrier takes the risk that the assets won’t perform well enough to support the annuity payments. But some would argue that, even in retirement, some exposure to equities may produce a favorable trade-off to participants.
Inflation risk. The vanilla SPIA does not protect against inflation. Even at the current rate (about 2% per year), inflation has an effect. After 10 years, the ‘real’ (effective buying power) of the participant’s annuity payment has gone down 22%. At 3% inflation the value erodes 34% over 10 years. Perhaps more significant than this sort of steady erosion of purchasing power is the risk that, for instance, several years of very high inflation (e.g., above 10%) may occur early in retirement.
As we understand it, many insurers are (currently) reluctant to offer inflation-protected annuities, because of continuing concerns about possible future inflation and the difficulty of hedging that risk. But some insurers still do offer them.
Inflation-protected annuities provide a benefit that increases with the CPI. The first year benefit is reduced (relative to a non-inflation protected SPIA) by 25%-30% in today’s economic environment. As inflation estimates fluctuate, the 25%-30% rule of thumb will move up or down.
The higher inflation is and the longer the participant lives, the ‘better deal’ inflation protection is. Using a 3% inflation assumption and assuming a 30% reduction in the first year benefit, the inflation-indexed annuity exceeds the fixed annuity by age 78. But, again, persistent 3% inflation is not all that a participant is insuring against when she buys inflation protection – she’s also protecting against short periods of very high inflation, especially early in retirement.
There are also ‘fixed COLA’ annuities available. These simply increase the annuity payment 1% – 4% per year, regardless of how actual inflation. Increases under a fixed COLA annuity are not linked to the CPI in any way. A fixed COLA annuity is simply an SPIA that starts out with a lower payment and then increases each year by the fixed percentage, so payments ‘more or less’ track inflation to the extent the fixed increases are in line with actual inflation.
By ‘longevity annuity’ we mean a deferred annuity under which payments begin at, for instance, age 85. The market for this sort of annuity is relatively undeveloped. Longevity annuities have been proposed, however, as a way to deal with long-term longevity risk. For example, a 65-year old participant might buy an age-85 longevity annuity with a small portion of his account balance (say, 10%) and then implement a draw-down strategy for the period 65-85 without concern that there will be no money at age 85.
A longevity annuity strategy can be thought of as a hybrid (annuity + draw-down) approach, providing a long-term guarantee but preserving short-term flexibility and opportunity by maintaining a substantial proportion of retirement wealth outside of the annuity.
Guaranteed Lifetime Withdrawal Benefits
A Guaranteed Lifetime Withdrawal Benefits (GLWB) (aka Guaranteed Minimum Withdrawal Benefits (GMWB)) arrangement is a hybrid product that provides some exposure to equity performance while still providing a guarantee of income. If performance of the underlying portfolio during a year is ‘good,’ the guaranteed amount may be increased. The guaranteed amount generally may not be decreased, so good performance in effect ‘ratchets up’ the guarantee.
Here’s an example of how a GWLB might work. With her $100,000 the participant buys a GLWB with a 5% guarantee. In the first year, she gets $5,000. Assuming good performance in the first year, at the beginning of year two her ‘principal’ is worth $110,000. So in year two, the participant gets a payment of $5,500 (5% of $110,000). Let’s assume poor performance in year two, so that at the beginning of year three her ‘principal’ is worth only $90,000. The participant still gets the $5,500 guaranteed payment for year three, and for every subsequent year while alive, even if (and after) these payments and related performance exhaust his principal.
‘Excess withdrawals’ (withdrawals above the guaranteed amount) may be allowed, although such a withdrawal may affect the guarantee or trigger a penalty. On death, if there is any remaining value in the GLWB, it passes to the participant’s heirs.
A GLWB has features of an annuity (the guarantee) and features of a draw-down strategy (increased payments if portfolio performance is good, the ability to make withdrawals and a possible legacy). The fees for all these features, however, will reduce the amount of year one income the participant will get. A GLWB generally does not provide inflation protection.
By ‘draw-down strategies’ we mean any withdrawal strategy that does not involve a lifetime income guarantee. This could include, for instance, payments estimated to last over a given period of years, but the most common draw-down strategy involves the participant taking a certain percentage of her account balance each year. For instance, if the initial balance is $100,000, the participant would take $4,000 per year under a (typical) ‘4% draw down’ strategy. Subsequent withdrawals may be adjusted for inflation or the performance of the underlying asset portfolio.
Annual adjustments tied to inflation produce a stable income pattern that keeps up with the cost of living, but it can exhaust retirement assets if investment returns fall short of inflation. Adjustments based on investment performance, on the other hand, address the risk of running out of money, but they may produce an uneven stream of income that fails to keep up with inflation.
There has been a huge amount written about whether 4% is too conservative or, in the current interest rate/returns environment, too aggressive, and we are not going to go into detailed analysis of that issue. These are, as we understand it, the features of a draw-down strategy that distinguishes it from an annuity strategy:
The participant manages her own account. This may reduce fees – e.g., the fees paid to an insurance company for an annuity. But it also puts the participant in the position of making her own asset allocation decisions, using (and perhaps paying for) investment advice or investing in a target date fund.
The participant takes all longevity risk. This may bias a retired participant towards a more conservative draw-down percentage.
The participant takes all investment risk. Unless the participant has a large account (relative to her income needs), this may bias a retired participant towards a fairly conservative investment strategy.
The participant takes all inflation risk. Whatever the participant’s investment/asset allocation strategy, it must take into account inflation.
We’re going to take up the issue of fees in more detail in our next article, which will consider the payout issue from the sponsor’s point of view. But let’s note here that, whatever their virtues or defects, these different payout options involve different fees and fee structures. If the participant’s only goal is to keep fees as low as possible, a draw-down strategy looks pretty good. There are ways to get a good price on an annuity product, but as you transfer risk to the insurance company and add complexity and features, fees naturally increase, and to some extent it becomes harder to determine if in fact you are getting a good deal. Thus, participants who are particularly focused on fees (or are encouraged by regulators to focus on fees) may have a bias in favor of simpler products and retention of certain risks.
There is evidence that spending in retirement declines over time. The ‘story’ goes: for the first couple of years the retiree spends more, on ‘things she always wanted to do’ (travel, hobbies, etc.). Then she ‘settles down’ to a longer period of less doing and less spending. The story may (or may not) conclude with higher medical/long-term care costs at the end of life.
There are questions about whether this pattern of spending is what retirees ‘want’ or simply what they ‘have to do’ as their retirement resources dwindle. But it is certainly possible that, given these spending patterns, retirees’ income needs in retirement are not as smooth or as flat as an annuity or a simple draw-down provides for.
Buying retirement income before retirement
Some have suggested that participants should be encouraged to begin using at least part of their account to buy an annuity before retirement. There are products available that, generally, allow a participant, before retirement, to purchase an annuity that will not begin until actual retirement. And, generally, prior to annuitization, the participant can take money back out of the annuity.
In a 2009 paper (Automatic Annuitization: New Behavioral Strategies for Expanding Lifetime Income in 401(k)s, Iwry and Turner), Mark Iwry, now Deputy Assistant Secretary (Tax Policy) for Retirement and Health Policy at the Department of the Treasury, suggested that “the phased or incremental acquisition of deferred annuities during the plan’s accumulation phase” could help encourage participants to consider annuitization.
Annuitization during the accumulation phase, if done piecemeal over time, is, in effect, a kind of ‘averaging out’ of other assets and ‘averaging into’ the annuity. Iwry advocates avoiding “now-or-never decisions,” arguing that it is better to allow participants “to choose incremental annuitization over time, rather than being confronted with a single moment of truth when the decision of whether or not to take an annuity is thrust upon them. The stress that accompanies making such a decision is not only unpleasant but also an incubator of bad decisionmaking.”
Generally, annuitization during the accumulation phase can be thought of as an allocation to fixed income. Indeed, it’s not unlike de-risking in a DB plan. If the participant is on an asset allocation glide path, and buys an annuity benefit pre-retirement, then she should reflect this decision in investing the portion of her account that is not in the annuity.
Why don’t participants like annuities more?
It’s widely understood that participants have considerable resistance to buying annuities. At the highest level, the move away from traditional defined benefit plans, that pay an annuity benefit, to 401(k) and cash balance plans, that pay lump sums, reflects this bias. ‘Nudge’ (default) strategies that have increased participant 401(k) contributions and improved participant asset allocation have generally not been effective in getting participants to elect annuity distributions where a lump sum is available.
Policy experts have come up with a number of theories why this is. We would identify the following as factors:
There is probably a cognitive bias in favor of an account balance that you can withdraw immediately vs. a stream of income lasting into the future – some combination of the “bird in the hand” and hyperbolic discounting.
In the same vein, people may discount how important ‘living well’ in the future is relative to living well now (and having the freedom to use your account balance to do so).
Participants generally have a legitimate concern that they may have expenses (or, indeed, opportunities) that require access to their principal.
Annuities are inherently less transparent, and it is not easy (or, in many cases, even possible) to separate out the ‘mortality pooling’ benefit of annuities from the ‘fixed income’ investment decision.
Our purpose in writing this article is to survey the distribution options a participant might have available in a 401(k) plan. Obviously, most 401(k) plans do not have all of these options. Our guess is that the vast majority of plans have no annuity option and many (if not most) have very simple draw-down options. But these options (the ones we survey here) could be available in a plan; in an IRA they generally are available. In our next article we will discuss the issues these options present for sponsors and why many sponsors are reluctant to complicate their plans with them.
For participants, we come back to those three risks we discussed at the beginning: longevity risk, asset risk and inflation risk. Buying an annuity means giving those risks to the insurance company (in the latter respect, if the participant buys an inflation-protected annuity). In return, the participant (1) ‘risks’ not living long enough to make the annuity purchase a good deal, (2) loses exposure to equity performance and (3) may be paying for inflation protection that it turns out she does not need. Annuities also have a (to a large extent inherent) lack of transparency that complicates the participant’s decision.
It’s not our purpose to say what the right tradeoff for a participant is. Indeed, and obviously, there isn’t a one size fits all solution. Critically, the participant’s risk tolerance will be a significant factor in what choice or mix of choices she makes. By ‘risk tolerance’ we don’t mean anything psychological. Rather, we mean the amount the participant has relative to the amount she needs for a comfortable retirement (however she may define that). Where the has/needs ratio is higher, more risk can be tolerated.