Valuation-Informed Indexing #251

by Rob Bennett

Michael Kitces has written an article that advances our understanding of safe withdrawal rates in an important way. I greatly encourage everyone with an interest in retirement planning to read it. It is called The Racheting Safe Withdrawal Rate — A More Dominant Version of the 4 Percent Rule? Unfortunately, the article also evidences a gravely flawed understanding of the implications of the last 34 years of peer-reviewed research in this field.

The purpose of the long-discredited 4 percent rule was to help aspiring retirees cope with the biggest problem of stock investing — the huge amount of return volatility. For those accumulating assets for retirement, volatility is a significant negative but not an entirely devastating one. Most investors understand that there will be high-return years and low-return years and presume that they will cancel each other out to the extent needed for the average long-term return of 6.5 percent real to apply once again. Retirees are in different circumstances. A series of poor returns in the early years of a retirement can cut off enough compounding to cause even retirement plans calling for seemingly modest annual withdrawals to fail. The phenomenon is highly counter-intuitive but can be deadly for those who fail to take it into consideration.

The 4 percent rule is rooted in the finding that there has never in history been a time when a retirement plan calling for an 80 percent stock allocation would have failed had the retiree been taking out no more than 4 percent of the initial portfolio value (in inflation-adjusted terms) for 30 years (covering an age-65 retiree until age 95).

I discovered the flaw in the logic behind the 4 percent rule back in 2002. The fact that a 4 percent withdrawal has always survived 30 years does not render it safe. The valuation level that applies on the day the retirement begins is the biggest factor determining whether a withdrawal rate will survive or not. For retirements beginning at times of insanely low valuations, a 9 percent withdrawal has a 95 percent chance of surviving 30 years. For retirements beginning at times of insanely high valuations, any withdrawal rate higher than 1.6 percent carries some risk. It is not possible to calculate the safe withdrawal rate accurately without taking valuations into account. The 4 percent withdrawal worked at all times in the historical record but retirements using that withdrawal rate at times of high valuations were at high risk of failing, according to the historical data.

Kitces’ article adds a wrinkle that makes all the sense in the world.

Advisors who recommended use of the 4 percent rule at the top of the bubble were giving dangerous advice. The safe withdrawal rate at that time was 1.6 percent. Retirements using a 4 percent withdrawal had only a one in three chance of surviving 30 years, according to the historical data. There is a strong likelihood that we will be seeing millions of failed retirements in days to come as a result of our failure to correct the errors in the studies that made the 4 percent rule so popular for a time.

That said, the 4 percent rule is in a general sense a highly conservative rule. When we are experiencing bubble-level valuations, withdrawals of 4 percent are too high. But at all other times, using a 4 percent rule unnecessarily delays retirements by many years. At times of fair-value stock prices, a withdrawal rate of 5 percent is perfectly safe.

Kitces’ suggestion is that retirees who do not feel comfortable using the 4 percent (the rule has fallen out of favor in recent years because the low returns experienced since 2000 have reminded investors of the dangers of ignoring the reality that poor-return years are not random but occur in groups when the boom/bust cycle caused by the advocacy of Buy-and-Hold strategies enters its bust stage) see what sort of returns turn up in the early years of a retirement and then adjust their withdrawal rate upward in the event that several good return years push their portfolio value up to 150 percent of its initial value.

The core idea here is valid. Retirements fail because of big hits taken in the early years of a retirement. Even a retirement that called for a 4 percent withdrawal that started in 1996 (when prices were insanely high) will almost certainly survive 30 years because we did not see a crash until 2008, more than 10 years after the starting date for the retirement. Compounding is a powerful force. If a retirement enjoys 10 years of compounding, it is in good enough shape to withstand a major hit. If the hit comes in the early years of the retirement, the compounding cushion has not yet had time to develop to the extent needed to protect the retirement from the effects of a big hit.

The flaw in the Kitces’ concept is his idea of using a 4 percent withdrawal as the starting point. Shiller showed that valuations affect long-term returns. An obvious implication is that risk is not stable but variable. There can be no standard starting withdrawal rate for a research-based retirement plan. In some cases, the starting withdrawal rate should be 2 percent. In others, it should be 8 percent. In all cases, ratcheting makes sense if the early years of the retirement produce good results. But it is a terrible mistake to use 4 percent or any other percent as a standard starting point. The valuations factor is the biggest factor of all. It must be taken into consideration in all cases.

Imagine a retirement that began in 1996 and that employed a 4 percent withdrawal. That was a dangerous retirement plan because valuations were so high in 1996. However, as noted above, that retirement eventually became safe because we did not see a crash until 2008. However, had the retiree employed the logic behind the ratcheting concept to increase his withdrawal rate to something even higher than 4 percent, he would have increased the risk of retirement failure to a point where even the 12-year delay until a crash occurred might well not have saved it.

We are learning. These are exciting times for investment analysis. But we still have a long way to go before anyone in this field can properly come to think of himself or herself as a true “expert.” Our arrogance and our fear of speaking openly about the dangers of overvaluation are holding us back at a time when we should be making progress by leaps and bounds.

Rob Bennett’s bio is here.