From the September 2012 issue of Research Magazine • Subscribe!

August 23, 2012

Balancing Retirement Risks

How much should a new retiree spend each year? This is a simple question with no simple answers. One approach is to start with a retiree’s wealth and work backward to a safe withdrawal rate. If you have $2 million in assets, estimate a safe withdrawal rate that worked in most historical periods (say 4% inflation-adjusted) and withdraw a fixed percentage (starting at $80,000) of assets each year. It is easy to run very impressive looking Monte Carlo analyses using historical asset return data to show that retirees will most likely avoid running out of money before they die. Especially if they invest a healthy share in equities.

This approach is what William Sharpe, finance professor at Stanford University, calls “financial planning in fantasyland.” The fantasy is that the simulations we conduct using these data can provide a useful estimate of retirement income sustainability. He isn’t alone. Zvi Bodie, finance professor at Boston University, has criticized the belief that equities are less risky when held for a long period of time. My own analyses of safe withdrawal rates, in the July issue of Research magazine, show that a sustained period of low real asset yields would decimate a traditional safe withdrawal strategy. In a series of articles using asset return data from Europe and Asian markets, professor Wade Pfau estimates safe withdrawal rates of only 2-3%.

So if a 4% strategy isn’t always safe, what is? If advisors aim downward, they run the risk that the client will spend too little and die with too much of their wealth unspent. Boston University economics professor Laurence Kotlikoff has criticized the traditionally low safe withdrawal strategy as self-serving because it favors asset preservation at the expense of a comfortable retirement. If advisors are compensated based on the amount of assets managed, they have no incentive to encourage higher levels of spending. They also avoid the psychic cost and liability of guiding a retiree toward poverty.

It would seem to most advisors that academics enjoy criticizing commonly used strategies but can’t agree on a better solution. While academics disagree about assumptions and methods, most would agree that focusing solely on a single safe withdrawal rate doesn’t make much sense. Economists assume that the ultimate goal is to maximize happiness in retirement. They see retirement as a game where the objective is to get as much enjoyment out of the money we have while dealing with uncertainty about longevity and asset returns. A happiness-based retirement income plan wouldn’t begin by focusing solely on assets. It would focus on what we spend.

Douglas Bernheim of Stanford and his co-authors in 2001 found that rich retirees really didn’t spend that much more as a percentage of their pre-retirement income than less wealthy retirees. The primary difference between wealthy and less wealthy retirees is that the rich die with a lot more money—even if they don’t have a particularly strong desire to leave an inheritance. This is surprising to most economists, who assume that the reason people accumulate money is to spend it. Data from the Consumer Expenditure Survey collected by the Labor Department show that wealthy retirees only spend a lower share of their income on categories such as food and clothing, but do spend significantly more on discretionary expenses like vacations and gifts.

One way to think of retirement income is as a continuum of spending that begins with what we would consider essentials (food, insurance, basic health expenses, shelter), moves up to semi-discretionary expenses (clothing, eating out, basic transportation, cable subscription) and then to discretionary expenses (vacations, gifts, hobbies). We’d all agree that the discretionary expenses are what make life worth living during retirement. But we won’t enjoy them much if we can’t eat.

Acceptance of investment risk means that returns could be higher or lower than safe investments. Translated into spending, a riskier portfolio means that we could have more money to spend on vacations and eating out. But we could also be forced to cut back if our portfolio takes a turn for the worse. That uncertainty is the essence of Bodie’s argument against traditional retirement portfolios that place a good portion of a retiree’s wealth in equities. If equities perform as they have in the past then you’ll be fine. If they don’t, then a retiree may have to cut back on the basic spending that is crucial to happiness.

Matching an investment strategy to each chunk of retirement income allows an advisor to take risks where it is appropriate and provide safety for the expenses that mean the most to a client. Basic expenses may be best funded with safe assets such as Social Security, a pension or an annuity. Even safer are instruments that provide inflation protection. The next slice of spending may be funded with income from a relatively safe portfolio, and discretionary expenses with a riskier portfolio. This makes intuitive sense since the client is then bearing risk only in the types of spending where they are willing to cut back. For those who still remember their economic theory classes, this is also consistent with the concept of declining marginal utility of consumption. Each additional dollar spent in retirement provides a little less happiness than the last.

The risk of low asset performance is most often realized later in retirement when assets are depleted while a client is still living. This is the point where a client is forced to live off of her guaranteed retirement income, also known as the floor. Following the retirement income approach means that you have protected a client’s income floor so that they can meet basic expenses. Unfortunately, flooring products such as annuities or TIPS are expensive. An advisor following a flooring strategy may be faced with recommending that the client lose a large chunk of her retirement portfolio for the promise of a very modest retirement income. This may not seem like a very good deal to the client, and may also not seem like a very good deal to the advisor not compensated for the flooring product.

A reasonable way to build a floor is to buy a product that only pays out later in life when a client is most likely to run out of money and that provides a more substantial payout due to mortality credits (many other owners won’t be around to claim their benefits). This can be accomplished through the remarkably underused advanced life deferred annuity (ALDA), or longevity insurance.

My graduate student Duncan Williams and I have begun simulating optimal deferral periods for clients given their other sources of guaranteed income, age, risk tolerance and wealth in order to determine when and at what age an ALDA is optimal. We find that only an unrealistically risk-tolerant client would not benefit from longevity insurance, and that the optimal age duration of deferral for a 65-year old is often until the mid-70s. The irony of our analyses is that an ALDA allows an advisor to recommend a riskier investment portfolio, often leading to higher simulated wealth later in life for those who experience positive investment return sequences. Those who aren’t as fortunate have a more generous income floor to rely on.

Linking a retirement investment strategy with spending can also help a client sleep at night. The prospect of poverty, however remote, looms over any strategy that involves a shortfall risk. Matching assets with spending allows both a client and an advisor to understand what part of a budget is secure, and allows taking investment risks in spending categories where the client is willing to trade off the possibility of scaling back for a better chance of increased comfort.

While planning a retirement income strategy may not be simple, matching investments to spending may be easier for most retirees to understand than the traditional portfolio-centric approach.


Page 1 of 3
Single page view Reprints Discuss this story
This is where the comments go.