Economic Behavior...
Human Behavior Usually Trumps Computer Models
Jack Marrion
May 2014

Wall Street likes their computer models. The problem is that humans are not computers, and so the results of these models are often worthless because they are ignored. Here are two stories:

Wall Street’s retirement income model says retirees should withdraw a “safe” percentage of their assets each year and adjust the payout based on inflation. The initial payout percentage is based on both the expected returns of the assets being withdrawn from and the rate of future inflation assumed. Let’s pretend Wall Street can predict future returns. That still leaves us with two problems. One is that basing a payout rate today on the expected rate of future inflation is a fool’s errand because there are far to many factors affecting the inflation rate—it cannot be predicted. The second problem is that retirees do not increase their annual withdrawals at the rate of inflation.

The crystal ball models use recent inflation history as their guide. However, a model that assumes a current payout based on a 3 percent future inflation rate could provide a much higher initial payout if actual inflation is 1 percent, and a much lower initial payout if inflation is 5 percent. Even if you think you can predict market performance, you still can’t predict inflation, so your initial payout assumption is inherently flawed. In any event, recent studies have found that retirees ignore these Wall Street models anyway and do not increase their payouts as inflation occurs. Instead, they decrease their purchases—they also decrease their elective expenses as they age, regardless of inflation. The reality is due to the way retirees actually use their assets in retirement—sticking with the initial withdrawal dollar payout and not increasing it—that initial payouts should be significantly higher. As an example, assuming net annual return on assets of 3 percent on $100,000, a constant $5,000 annual withdrawal lasts an age 66 couple until age 95. Although a Wall Street inflation model with an initial 3 percent safe rate results in higher income when the couple is in their 80s and 90s, this static withdrawal method provides more income early in retirement when it is more likely to be used to enhance quality of life. The effects of inflation may be more or less offset by elective decreases in spending.

Another area in which behavior trumps computer models is on Social Security maximization. The logic is that if you don’t need the income you should delay beginning your Social Security until age 70 because the income is guaranteed to grow at 8 percent and you can’t find any other sure thing paying 8 percent. Indeed, on a pure cash basis you are money ahead by age 80 if you wait until age 70 to start taking a payout. However, a significant percentage of people take benefits at age 62 even if they don’t need the income.

The reason for the early claiming is simple: loss aversion. The retirees all have done the math and figured out if they die before age 78 or so they get less money if they wait to claim benefits. If they die early they feel they’ve gambled and lost (ignoring the reality that Social Security would have provided benefits to their spouse and minor children if they had died early, or provided an income if they became disabled). To keep the casino (Social Security) from making them suckers they’re determined to at least get back some of their bet.

You can say that from an implied investment return perspective you are silly to claim early if you don’t need to. You can talk about life expectancy and how there’s better than a two in three shot the retiree will live long enough to benefit from waiting to draw benefits. You can show all the models and data you want and it won’t do any good, because what is driving this decision is loss aversion. Instead, offer an alternative that deals with their loss aversion and talk about how they can take control of the gamble.

A person retiring today at age 62 getting $12,000 a year might get $21,750 if they waited until age 70 to start benefits (I’m keeping inflation constant on both sides).

If the person waits they get $9,750 more a year—the crossover point is age 79 (we’re assuming the retiree does not need the $12,000). Let’s say that $12,000 a year is placed in a flexible premium annuity that grows at 3 percent net interest per year and has a guaranteed lifetime income roll-up rate of 6 percent and a joint payout rate at age 70 of 4.5 percent. This would mean the cash account value would be $106,708 and the joint lifetime income would be $5,300 at age 70.

If a person uses the early Social Security money to buy an annuity, beginning at age 70 they would then receive $17,300 ($12,000 plus $5,300)—far short of the $21,750 they would have received by waiting for Social Security. However, the retiree’s early claiming decision means they have complete control of $106,708 today and the remaining balance in future years. The retiree that won’t wait to claim Social Security might be tempted with a solution that makes up for some of the foregone income and lets them keep control of the cash.

The effectiveness of retirement planning advice is not judged by how pretty the computer income model is, but by whether the retiree takes the advice. Not taking into account the behavioral biases that affect retirement decisions means the advice is much less likely to be used.

Author's Bio
Jack Marrion
provides research and consulting services to insurance companies and financial firms in a variety of annuity areas. He also serves as director of research for the National Association for Fixed Annuities and as a research fellow for Webster University. In 1994 he wrote a book to help banks market investment and insurance solutions to their small business clients. In 1996 he produced the first independent hypothetical return monthly publication comparing all index annuities on the market, and in 1997 created the first comprehensive report of index annuity sales, products and trends, "Advantage Index Product Sales & Market Report" (quarterly). His insights on the annuity and retirement income world have appeared in hundreds of publications. In 2006 the National Association of Insurance Commissioners asked him to address their annual meeting and teach regulators the realities of index annuities. He was invited back in 2009 to talk to the NAIC about the effects of aging on senior decision-making. He is a frequent speaker at industry functions. Prior to forming Advantage Compendium, Marrion was president and owner of an NASD broker/dealer with offices in nine states. Previous to that he was vice president of a life insurance company and vice president of an NYSE investment banking firm. He has a BBA from the University of Iowa, an MBA from the University of Missouri, and a doctorate from Webster University. Marrion can be reached at Advantage Compendium, Telephone: 314-255-6531. Email:

Leave a Comment