The 4% Rule Did Not Become a Whole Lot Easier
Membership is now required to use this feature. To learn more:View Membership Benefits
Advisor Perspectives welcomes guest contributions. The views presented here do not necessarily represent those of Advisor Perspectives.
I’ve been retired for seven years, which explains my “timed itch” to respond to Allan Roth’s recent article, The 4% Rule Just Became a Whole Lot Easier. Roth often challenges industry dogma, but in this article, he’s promised too much, made inaccurate comparisons and blurred the notion of risk. His surrender to real return projections may work in theory but will fail in a world tainted by the most significant risk of all: human behavior.
Real returns in retirement projections pass muster only because projections allow for course corrections as life events unfold. These real-return shortcuts, however, are much more dangerous when used in an actual portfolio with frozen components working inside a dynamic world.
Before I go too far, let me mention what I like about Roth’s article. I like TIPS, which have enjoyed a 20% allocation in the bond portion of my portfolio for years. I like their extremely low risk of default. Since history reports that high inflationary periods can destroy portfolios, I like their contribution to inflation protection.
Unfortunately, my complaints outweigh what I like in the article. TIPS’ inflation protection has its limits.
Turning to my complaints
Let’s start with Roth’s unfair comparisons, beginning with “safe withdrawal studies,” which leads to a summary of success or failure rates for a balanced portfolio. In Roth’s case, he compared a balanced portfolio’s 90% chance of success over the 30-year projection period to the TIPS ladder’s 100% chance of success over the same period. What he failed to mention is that those balanced portfolios within the 90% success level continue, many for years and years, adding a level of longevity protection. Roth’s TIPS ladder self-destructs at 30 years and a day. Imagine the fear in the 29th year when the frail elderly retiree suddenly realizes their portfolio is about to run out of fuel.
Roth and I disagree on his lead statement, that “the 4% rule just became a whole lot easier.” I disagree partly because he never addressed his level of commitment to this strategy. His $100,000 ladder surely doesn’t commit the whole portfolio to the strategy. He leaves wiggle room by suggesting he will likely add more to the approach. Every strategy can mitigate certain risks while, at the same time, add other risks. Roth’s strategy has this same basic flaw.
My final disagreement is in Roth’s conclusion; his strategy “produced a safe 4.3% annualized real withdrawal rate for 30 years with no risk.” But he isn’t looking hard enough for risk. Consider the risk of the bond ladder strategy; it works well unless cash flow needs break the ladder. Think of the game Jenga and the impact to the tower when the wrong block is pulled. In real life, drawdowns are not smooth and even, which adds risk to rigid withdrawal strategies. Pulling the wrong block in the Treasury ladder to make up for unexpected cash flow needs implodes this strategy by injecting interest rate risk into the mix.
The flaw of real return projections
Roth’s conclusions are based on “real return” projections, which ignore some important truths of retirement: CPI, the driver of real returns, is not an adequate measure of inflation. It ignores human behavior and serendipity, the ultimate arbiters of life’s events. And more systematically, it ignores the fact the CPI does not measure spending inflation; it measures changing prices for a theoretical basket of goods. When I see projections that assume every person’s spending increases at the same flawed rate of change, I shudder.
I was raised by a simple man who had an uncanny ability to ask the right question. When I showed him the power of Monte Carlo simulations creating 9,000 successful lives out of 10,000 runs, his response was, “Don’t give me that crap. I got one life. How’s that gonna work out?”
A few years later, I visited him in Florida and he complained his Social Security wasn’t keeping pace with inflation, which inspired me to begin a research project that resulted in my 2008 award winning paper, “The Hedonic Pleasure Index – a New Model for Spending Inflation.”1
When I explained my findings to him, he reacted, “They wrung all the fun out of inflation.” And he was right. Everything that makes us human has been removed from the CPI calculation. The Bureau of Labor Statistics (BLS) even tells us that the index measures what it takes to maintain a constant standard of living, and 40 years after the BLS embarked on extracting life, there was no joy left in Mudville.
Let me explain. In 1983, the BLS began a decades-long change to the CPI calculation method. By 1996, the Boskin Commission reported that the CPI had been overstated by 1.1% due to bias’s that existed in the Bureau’s measurement procedures.
The commission identified four sources for this estimation error: consumer product substitution, consumer outlet substitution, new goods entering the market and quality changes. As those changes found their way into their calculations, the BLS managed to shave 1.1% off reported CPI over a long enough period of time that no one noticed, least of all my peers who, like me, lived through the changes.
We ignore these changes at our peril. We know the Fed longs for a 2% inflation target and the BLS continues to search for ways to extract items not considered a price change from the index. Locking to a benchmark that is actively being toyed with is not a good long-term strategy. The BLS assumes we all respond to changing prices in the same way; by substituting products or changing where we shop. While many do respond that way, I do not. I buy what I want, because I don’t like many of the substitutes. You will never find me in the big-box stores looking in vain for the savings on aisle 183.
More insidious, however, is how new products and quality changes are measured. For most items in the basket of goods, the BLS uses hedonic regression analysis to remove the impact of quality from the price index. Its initial focus was on goods, but services are now considered for the same reason. Quality changes are not the same as price changes.
For some items, the BLS' extraction method is more intuitive than hedonic regression. Take automobiles. The BLS uses the manufacturer’s sticker price instead of hedonic regression to extract quality changes. Consider the following items that are removed from the automobile “price” index: automatic emergency braking systems, connected mobile apps and digital keys, teen safe driver technology, safe exit assist to protect cyclists, wireless smartphone connectivity and charging, stabilization and lane changing prevention and a 360-degree camera.
I can’t wait until they perfect driverless cars. And you guessed it, those improvements will not be included in the calculation of the price index either.
Consumers, of course, pay for these quality improvements; they just aren’t reflected in CPI. Yet real projections assume we don’t buy quality improvements and our behavior conforms to the BLS’ behavioral models and that all of us act the same, in lock step in a real spending frenzy. And here is where real spending falls short; it just ain’t “real”.
Conforming to the BLS’ spending models is impossible because you can’t remove the quality enhancements from your purchases. But for those who try by hanging on to their car for 20 years will feel they are falling behind their neighbors who are reaching beyond a constant standard of living by reinserting fun into their spending model.
Adding a TIPS ladder could work well for a portion of an allocation, but that would not make the 4% rule any easier. It would add another asset class with a very interesting drawdown strategy that includes a built-in self-destructing element.
I wish measuring spending inflation was a retiree’s only obstacle, but there are others:
- Retirement studies that economists use to “establish” spending patterns in retirement do not include an important sub-group of Americans: free spenders. Unfortunately, spendthrifts are rarely considered in retirement spending studies because they don’t track their spending; they just spend. The result is publicly available spending data that omits those best at spending, the spendthrifts.
- Everyone is different, but as a rule, retirees have different consumption patterns than pre-retirees. Americans over age 62 on average spend 11.7% of their budget on medical care compared to 8.5% for those under age 62 and 47.7% on housing compared to 42.4% for working-age Americans. The CPI-E for the elderly would track old folk’s inflation better than CPI-U, but even it would fail to address the most important spending factor, human behavior, and it is unlikely we will ever see products based on CPI-E.
- Risk is a multi-headed monster and Roth’s strategy is exposed to some risks while mitigating others. That brings me to the risks of serendipity.
Serendipity covers a wide range of risks that a rigid withdrawal strategy doesn’t address
Retirement risks are not limited to those of the assets, the flaws in the Index or how solid the strategy is. They include the behavior of the client and several non-financial risks that can have a profound impact on portfolio strategies, concentrated or not.
Roth’s claim that his strategy “produced a safe 4.3% annualized real withdrawal rate for 30 years with no risk” set me in full iconoclast mode. No risk? Don’t get me wrong, I love TIPS to address the risk of unexpected inflation. But my problem is not with the asset class.
Roth’s strategy fails to consider human risk. A common risk among retirees is loneliness and living alone. As the pandemic drags on, it fosters unhealthy behaviors in retirement communities that directly affect physical and mental health. The real return methodology’s perceived clean and certain solutions fail to recognize we live in an uncertain world.
The following items will impact any portfolio, let alone one designed to self-destruct in 30 years:
- Longevity – average life expectancies measure the age when only half of the measured population is expected to die.
- Deflation could devastate a TIPS portfolio, especially one constructed with TIPS purchased on the secondary market with large built-in inflation imbedded in the purchase price. Unlike TIPS bought at auction, these TIPS could lose substantial principal in the event of prolonged deflation.
- TIPS purchases assume the “TIPS to Treasury breakeven rate” is an accurate measure of future inflation. If it is not, the additional price paid for inflation protection is lost.
- Reductions in a client’s other income sources puts pressure on static models that require ladders be honored. Consider the havoc a potential 30% drop in Social Security benefits by 2030 could have on the static withdrawal model of laddered bonds.
- Portfolios are impacted by tax policy changes increasing the importance of flexible tax location and withdrawal strategies. The key word is “flexible.”
- Loss of liquidity in an unrelated asset class could require a fire sale if too much of the remaining portfolio is locked up in a bond ladder – the Jenga affect.
- The uncontrollable desire to help adult children or grandchildren at the retiree’s own expense affects many more retirees than you might expect.
- Divorce or death of a spouse, especially the one with the higher lifetime benefits may require the bond ladder be broken at the worst time in the person’s life.
- Health shocks or a broad decline in health that depletes the pocketbook also point to the flaw of inflexible withdrawal strategies.
- Cognitive decline including the inability to manage money or make wise decisions are very likely to expose the risk of CPI linked investment’s limited growth potential.
- Unstable emotions that arise from several untold occurrences require more flexibility than a portfolio of bond ladders.
Roth’s strategy adds another tool for planners to work around the edges. The CPI, however, doesn’t consider how real, live people spend money; it considers only how inanimate prices change. Planners need to acknowledge that each person is unique. Some folks are frugal, and some are spendthrifts, plain and simple.
A retired person’s risks far outnumber portfolio risk. Planners need to stop using the real return crutch and pushing products that “solve” the spending crisis with an imprecise spending tool like the CPI. Learning a client’s actual spending behavior requires work but will pay off in spades over the long haul.
Roth’s strategy would work well for a slice of the allocation, but making the claim that “the 4% rule just got a whole lot easier” is an inflated opinion.
And “no risk?” I don’t think so!
James A. Shambo, CPA, is a retired financial advisor.
1Shambo, James A. 2008. “The Hedonic Pleasure Index: An Enhanced Model for Spending Inflation.” Journal of Financial Planning 21 (11): 44–55.
Membership is now required to use this feature. To learn more:View Membership Benefits