Apples and Pears

Be careful when making comparisons. There is much too much confusion surrounding Safe Withdrawal Rate research. Sometimes, this is intentional. At other times, it is entirely innocent. At all times, it is unfortunate.

The most common source of confusion is based on an imprecise definition. Historical Surviving Withdrawal Rates are different from their Safe Withdrawal Rates. Historical Surviving Withdrawal Rates tell us what has happened in the past. They are particular outcomes. They are certain.

Safe Withdrawal Rates are mathematical calculations. They are probabilities. With rare exception, they do not guarantee future outcomes. They tell us what is likely to happen provided that the future is similar to the past. They come with levels of confidence.

An early assumption was that the smallest Historical Surviving Withdrawal Rate from many years is a Safe Withdrawal Rate. That assumption failed. It succeeds only if we are unable to say anything about which future year will be selected. It fails if we are interested in today and/or if we must consider today’s valuations.

Another source of confusion is to ignore the timeframe. Dividend-based strategies, for example, last for a very long time. They should not be compared to a strategy that is designed to last 30 years, but not necessarily any longer.

A serious source of error is to ignore probabilities. We can define what happens with a TIPS-only portfolio almost exactly. An alternative portfolio with the same Safe Withdrawal Rate that includes stocks is different. There is always an element of risk with stocks. A 95% probability of success is not the same as a 100% chance of success.

Another serious source of error comes from glossing over the limitations of our calculators. Most of our historical sequence calculators handle the non-stock component very poorly. They treat bonds as single-year trading instruments. They replace bonds every year to get new interest rates. But they never experience any capital gains or losses. They can never lock in an interest rate.

[This includes most of my own calculators, which are modified versions of the Retire Early Safe Withdrawal Calculator, version 1.61, dated September 7, 2002. However, my Gummy series of calculators use bond data that include capital gains and losses.]

Sometimes, people forget to look at withdrawal algorithms carefully. We normally report withdrawals in constant real (inflation adjusted) dollars. Many studies report withdrawals that stay the same in nominal dollars. That is, they do not make any adjustments for inflation. Some studies base withdrawals on a percentage of a portfolio’s current balance as opposed to its initial balance. They can be helpful, but you must be very careful about withdrawal amounts. [See The 4% Shocker.]
The 4% Shocker

Don’t think that this kind of confusion is limited to Safe Withdrawal Rate research. We see it all of the time in stock market discussions. There are about one-half of a dozen definitions of the price-to-earnings ratio of the S&P500. Some are based on earnings projections. Some are based on trailing earnings. Sometimes, they include corporate losses (negative earnings). More often, they do not. Sometimes, they exclude certain items. At other times, they exclude other items. At still other times, they exclude nothing.

Sad to say, many commentators pick and choose their definitions to suit their arguments. I have seen articles in which the writers intentionally have mixed definitions to advance their arguments, knowing full well that they are misleading their readers.

Have fun.

John Walter Russell
July 23, 2005