Retirement on Beach

So Just How Much Should You Be Spending in Retirement?

Retirement on Beach

A couple of months ago, we published an article addressing the question of how much you should be saving for retirement. Today, we will examine the other side of the coin, once again utilizing research provided by Dimensional Fund Advisors—this time by Peng Chen, the CEO of Dimensional Asia ex-Japan. The paper addressing this topic that he published with two other researchers can be found here.

The essential problem can be summarized as follows: We have a pool of investments from which the retiree will be making regular withdrawals, but we do not know the sequence of returns that the investments will earn nor do we know how many years the portfolio will be required to support the withdrawals. The ideal outcome is that the portfolio value (excluding the legacy that the retiree wishes to leave to his heirs or charity) will hit zero at the time of death. While the severity of the problem of the portfolio crashing before death needs no explanation, a large unspent surplus after death suggests the possibility that the retiree lived a much more spartan lifestyle than was necessary, which is also not an optimal outcome.

One commonly proffered spending rule is to start with a percentage of the portfolio (say 4%) and increase that amount every year with inflation. There are several problems with this spending rule. First, after a long period of time, the amount of the withdrawals may become completely disconnected from the retiree’s actual needs. While some expenses tend to increase faster than inflation (especially those associated with health care), other expenses (such as travel and entertainment) tend to slow down with advancing age. Second, a small change in the initial amount can have a large impact on the final result. For example, we found that for a 60/40 portfolio starting on 1/1/1969, a 4% withdrawal assumption allowed the portfolio to last through the end of 2012 with a healthy surplus while a 4.5% withdrawal assumption caused the portfolio to crash in 2003.

An alternative to the fixed dollar amount withdrawal is the fixed percentage withdrawal, also known as the endowment model. An example of the endowment model would be the retiree who withdraws 5% of the portfolio value each year (based on Jan 1st value and divided into 12 monthly withdrawals), regardless of how different that withdrawal is from the prior year amount. For very high net worth investors who wish to leave a bequest, this can be an ideal approach, as it essentially prevents the portfolio from ever crashing. The downside, of course, is the fluctuation in the annual withdrawal amount that will accompany the volatility of the market.

The ideal solution may be a spending rule that combines the two above. For example, the current year withdrawal could be calculated as the average of the prior year’s withdrawal adjusted for inflation and 5% of the current portfolio value. Furthermore, if the retiree wishes to take his remaining life expectancy into account, the IRS’s RMD table can be helpful. For example, a 75-year-old is estimated to have about 23 years of remaining life expectancy, which indicates a 4.3% withdrawal amount. This number could be used in our average calculation in place of the straight percentage of the portfolio.

Chen’s criticism of the commonly used methods for projecting the success or failure of a retirement asset allocation and withdrawal strategy centers around their failure to address mortality and their exclusive focus on the probability of failure while ignoring the risk of spending far less than would have been feasible. Even when a portfolio crash occurs, it makes a huge difference whether it occurred three months or ten years before the retiree’s death. To quantify the degree of success or failure of a particular withdrawal strategy, Chen introduces a new measure called the “Withdrawal Efficiency Rate” (WER) that is calculated from the formula below:

The denominator, SSR, is determined as the spending rate that brings the portfolio to zero at the time of death. In other words, it is the maximum spending rate that would have been possible, had the retiree known when he was going to die and the returns that his portfolio would have received. It is easily calculated with a spreadsheet. Calculating the numerator, CEW, is quite a bit more complicated. To simplify it, CEW answers the question of what constant withdrawal amount (as a percentage of the portfolio) would have provided the same level of utility (happiness) to the retiree as the stream of withdrawals that he actually took. The crux idea here is that for most people, the pain of taking $1,000 less than what was feasible counts more than the pleasure of taking $1,000 more, because as the economists would say, they have a decreasing marginal utility function.

To summarize, the goal of any asset allocation and withdrawal strategy should be to maximize WER, and typical values of WER range between 50% and 80%. As expected, Chen found that the static withdrawal strategy of taking a fixed dollar amount had a significantly lower value of WER than a dynamic strategy that takes into consideration the changes in portfolio value and life expectancy.

The key takeaway is that while using a tool such as IFA’s Retirement Plan Analyzer is quite helpful in determining how much to withdraw over the next year, it must be re-run every year, taking into account the changes that have occurred in your portfolio and the changes that have occurred in your life. This is how you can have an optimal dynamic withdrawal strategy as opposed to a sub-optimal static strategy. To find an asset allocation that is appropriate for you, please take IFA’s Risk Capacity Survey.