Being wealthy doesn't mean having all the money in the world, just not having to worry about it.

My Mind Boggles at the Math

Question: I’ve followed you, as well as Andrew Tobias, for many years (some good and some very bad years, I might add). I have a retirement account of $750,000. This is good for two people who saved and didn’t overspend through the years. I understand that if I spend 4-5% of the balance each year, it should last 30 years.

Except … I don’t understand. If I make 4% on the remainder after taking out the 4% to spend for that year, why would it last almost indefinitely. And what if I make 6% on the money???

If this is too silly for you to explain, no problem.

Answer: Not silly at all: many people wonder why a portfolio with an average return of 4% per year can’t sustain 4% withdrawals per year and last forever. There are three problems with this line of thinking:

  1. Because inflation has been the rule since the Federal Reserve was founded, the retiree needs at least 4% PLUS the increase in living costs he or she is facing. So let’s at least realize that drawing 4% a year in retirement in an inflationary environment means the 4% return must be the REAL return after inflation in that retiree’s cost of living. Fortunately, the increase in spending for a typical retiree is usually around 1% per year less than the reported rate of inflation, in large part because people enjoy spending experiences less and less as they get older. (For example, travel and restaurant dining become less appealing as we age.) But even if we assume a 4% return after inflation in that person’s spending each year, we have a second problem:
  2. Volatility reduces returns below the annual average. To keep it simple, assume an investment with an average return of 0% and no volatility, meaning it earns exactly nothing every year. Obviously if we started with $1 million, it would just stay there as long as there was no return and no withdrawals. But let’s say the investment gained 20% one year and lost 20% the next, still averaging 0%, but with what is clearly 20% volatility in returns. Earning 20% the first year, the $1 million would make $200,000 and grow to $1.2 million. But then that $1.2 million would lose 20% — $240,000 — the second year and drop to $960,000. After two years the fund would have lost $40,000 overall, or 4% of the original amount, which is an average loss of 2% per year. I chose 20% volatility because that is the approximate volatility of the US stock market. This means that it would take a 6% average return in stocks to allow only 4% annual withdrawals as a result of the 2% lost to volatility. A portfolio of stocks and bonds would be less volatile but also have a lower expected return to start. Nonetheless, any variability in the annual returns means that you can’t draw 4% from an investment earning an average of 4% and have it remain stable over time.
  3. In some ways this is a restatement of item 2, but it should also be noted that taking out a flat amount after a down year for the investments means the percentage being withdrawn goes up that year. Of course, the percentage being withdrawn goes down if the investments go up, but the effect isn’t symmetrical. If you’re drawing $40,000 from a $1 million investment, you’re taking out 4% (40,000/1,000,000 = 4%). If the investment dropped 20% in a bear market, to $800,000, then your $40,000 withdrawal that year is 5% of your investment (40,000/800,000 = 5%). If the investment rose 20% in a bull market, to $1.2 million, your withdrawal that year is 3.33% (40,000/1,200,000 = 3.33%). So if the volatility of the investment is 20%, as it has been for the US stock market over time, it is effectively like taking out 5% half the time and 3.33% the other half of the time, which is an average of 4.17%, not just 4% (1/2 * 5% + 1/2 * 3.33% = 4.17%).

I’ve oversimplified and in the process introduced trivial errors, but the basic point should be clear: to draw out 4% from an investment whose returns are not precisely the same percentage each year and with an expectation that your overall cost of living will rise due to inflation, you will need a whole lot more than a 4% annual return.

As I’ve pointed out in several previous posts, I really dislike the “safe withdrawal rate” (SWR) approach for several reasons. It ignores the unpredictability of investment returns. It ignores the unpredictability of personal spending. It ignores the unpredictability of life span. And for someone who isn’t really comfortable with math, it is easy to completely misunderstand and misapply it. It is a theoretical model that was only intended by William Bergen, who first referenced it, and most subsequent revisionists as a way of giving people a general idea of about how much they could expect to safely withdraw from their portfolios over time. As an actual way to handle your retirement, it is a monstrosity.

Now, it is quite self-serving on my part, but also fair, to say that someone who is confused by the mathematics of safe withdrawal rates shouldn’t be making his or her decision about how much to spend in retirement without expert advice. But for those who insist on doing so, I’ve had several previous posts suggesting reasonable approaches, such as “How Much Can You Safely Withdraw from Your Portfolio Each Year in Retirement?”