I am having a bit of a disagreement with a recent article Larry Swedroe wrote.  Now Larry is a smart guy, so anytime I find myself on the opposite side of an argument, I get a little uncomfortable.  Especially when he has a powerful ally on his side, retirement researcher Wade Pfau, who I had the pleasure of meeting recently at a Bogleheads meeting.  It isn't so much the data that I have a problem with, but more the conclusions taken from the data by far too many people.  First, the data.


The 4% Rule of Retirement

I've written before about the 4% rule.  It came into existence back in the 1990s with the publication of the Trinity Study and has recently been updated with later data.  It was written to counteract the claim by many advisors that you could withdraw 6, 8, or even 10% of your portfolio each year since that's what the portfolio made on average.  The problem is the sequence of returns issue.  If your portfolio gets poor returns in the early years, then your plan to withdraw your portfolio's average return will lead to early portfolio demise, especially when considering the effects of inflation.  The overall conclusion drawn from the Trinity Study was that you could withdraw 4% a year, adjusted for inflation, of a 50/50 portfolio and have a 96% chance of it lasting 30 years.  The data from this study comes from investment returns from 1927 to 2009.  They took rolling 30 year periods, withdrew 4% a year adjusted for inflation, and determined if the portfolio survived or not.  Even a casual glance at the data reveals that a 10% withdrawal strategy only has a 44% chance of lasting 15 years, and a 0% chance of lasting 30.


Monte Carlo Simulations And Current Valuations

Larry Swedroe argued in a recent column (and he's not the first to do so) that 3% (at best) is the new 4%.  Basically, he takes current stock valuations, as measured by the PE10 methodology, and current bond valuations, as measured by their current historically low yields, and runs it through a Monte Carlo Simulator (MCS).  A MCS is handicapped by the classic, “garbage-in, garbage-out” problem, since you have to make a certain set of assumptions to use it.  Swedroe explains the methodology this way:

[With a MCS] the expected final wealth distributions are determined by two numbers: 1) the average annual return (which should be based on current valuations/yields, not historic ones); 2) the standard deviation of the average annual return. The Monte Carlo simulator will randomly select a return for each year and calculate the wealth values over the expected retirement period. This process is repeated thousands of times in order to calculate the likelihood of possible outcomes.

Larry, along with others doing similar things with various assumptions, is then led to the conclusion that 4% really isn't safe at all, and the safe withdrawal rate (SWR) is now 3% at best.  Some conservative folks (like most good savers) add a little extra margin of safety, and knock their withdrawal rate down to 2.5% in response to data like that.

The Problem With A 3% Withdrawal Rate

Using 3% instead of 4% isn't just some academic discussion.  It has very dramatic real world effects.  Using the 4% rule, if you need $100,000 in portfolio income in retirement, you need to have a $2.5M portfolio on the eve of retirement.  If you then change the number to 3%, all of a sudden you need $3.33M.  If your original plan was to save $50,000 per year and earn 5% annualized returns to get to $2.5M in 25 years, then you're now left with a dilemma.  You can either save another $17K per year, you can work an extra 5 years, or you can spend 25% less in retirement, none of which are particularly attractive.

I'm actually very concerned about making my portfolio last a long time in retirement.  These studies usually discuss a thirty year retirement, but my portfolio will probably need to last even longer.  I'm in relatively good health and plan to stay that way.  I also anticipate an early retirement.  To make matters worse, my spouse is 3 years younger than me.  According to IRS tables, I'm expected to die in 2053.  My wife, however, isn't expected to die until 2060.  Plus she eats better than I do.  So if I retire at age 55 in 2030, there's a reasonable chance we may need that portfolio to last 40 or more years, not just 30.

However, rather than arguing endlessly about whether the Trinity Study's methodology based on past returns or the newer methodologies based on valuations and MCSs is more accurate, I'd rather point out 6 reasons why I don't care if 3% is the new methodology.

1) Nobody Uses A Strict SWR Approach In Retirement

I know lots of retirees, but I don't know any of them using what I would call a “strict SWR approach” to their retirement distributions.  In practice, nobody multiplies their initial portfolio value by a certain percentage, adjusted for inflation, and then blindly withdraws and spends that exact amount year after year.  Just like during the accumulation stage, the retirees I know adjust as they go.  If the portfolio is doing well, they take a few more trips, spend more on the grandkids, and give more to charity.  If it is doing poorly, they hunker down and spend less.  It isn't rocket science.  People did it for years before the Trinity Study ever came out (not to mention that most retirees have no idea what the Trinity Study is.)  If you really look at the data, the median outcome using a 3-4% withdrawal rate shows the retiree dying with a portfolio LARGER than the one they started with.  Being overly cautious and mechanical has consequences too- i.e. spending too little.  You can't take it with you.

2) There Are Two Probabilities To Deal With

Too many people forget that they probably won't need their portfolio to last 40 years.  There are really two probabilities in play here.  First, the probability that you outlast your money due to poor investment returns.  Second, the probability that your money outlasts you due to poor health.  If the probability that your money doesn't last 30 years is 5%, and the probability of you living longer than 30 years is 20%, the real risk that you personally will run out of money is 5%*20%, or 1%.  I don't know about you, but that's a risk I'm more than willing to run.  Remember that for portfolio failure you need to encounter not one, but two “bad” outcomes- poor returns AND a long life.

3) The Risks You're Not Thinking About

William Bernstein, in his famous Retirement Calculator From Hell, figures there's about a 20% chance of something really bad happening in the next 40 years.  He includes outcomes such as hyperinflation, local military action, political failures or confiscation, and nuclear war.  There's no point in planning a portfolio that will have a 97% chance of lasting through 40 years of poor market returns if it only has an 80% chance of surviving 40 years of political, economic, and military events.

4) Remember What The Trinity Study Included

While future returns seem grim given current valuations, keep in mind the events the data in the Trinity Study already include.  They include the Great Depression, World War II, the Stagflation of the 70s, the Cold War, the Dot-Com Bubble, and the Great Recession of 2008.  If you could go through all that and 4% STILL worked, well, how much worse do you really expect it to get?

5) Don't Forget Guaranteed Income

My general recommendation for retirees is that they use guaranteed income sources for their necessities, and portfolio income/withdrawals for their wants.  Guaranteed income sources include Social Security, pensions, and immediate annuities.  If you need a minimum of $50,000 in income for your necessities, and Social Security and any pensions provide $25,000, then you should annuitize a sufficient amount of assets to provide the other $25,000 of income.  Even at today's historically low rates, a 60 year old gets 6.41% on an immediate annuity and a 70 year old gets 8.12%.  Even if you buy the annuity on two 60 year olds, it still pays 5.61%.  You can also buy inflation-adjusted immediate annuities, or simply annuitize in multiple steps in order to keep up with inflation.  At any rate, if you're willing to eliminate the possibility of leaving it to your heirs, you can get a completely safe withdrawal rate that is much higher than 4%, much less 3%.  Since your necessities are covered by guaranteed income, you can afford to take a lot more risk with the remainder of the portfolio, increasing expected returns.

6) It Isn't The End Of The World If Your Portfolio Dies Before You Do

Too many advocates for super-conservative withdrawal rates treat running out of money as a catastrophic outcome.  Although everyone is in a different situation, it probably isn't catastrophic for you.  First, aside from the obvious solution of monitoring your portfolio and annuitizing more of it if the dreaded poor early returns materialize, there are other solutions available besides eating Alpo.

First, most retirees own their home.  In fact, most Americans have far more wealth in their home than their retirement accounts.  This home equity can be accessed and spent in retirement by purchasing either a reverse mortgage, or if you'd prefer to avoid the fees, making one yourself by moving out, selling the home, and annuitizing the proceeds.

Second, the retirement plan for centuries has been to go live with your kids.  Most readers of this blog are also teaching their children how to work hard, get good jobs, save money, and invest it wisely.  If you treat them well, it seems unlikely that they will abandon you in your old age.

Last, even if you end up in a nursing home completely destitute with some time still on the clock, Medicaid will step in and pay for it.  Millions of Americans retire on Social Security and a 5 figure portfolio and I don't see very many of them living in cardboard boxes eating Alpo.

3% might be the new 4%, but that's no reason to impoverish yourself unnecessarily in retirement, or to work longer than you wish to.  Make reasonable withdrawals, monitor your portfolio performance, guarantee some of your income, and treat your children well instead.