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

How Much Can You Safely Withdraw from Your Portfolio Each Year in Retirement?

For around 20 years, financial advisers and academics have been actively wrestling with the question of the safe withdrawal rate from a portfolio. The seminal article in the field is by William Bengen, entitled “Determining Withdrawal Rates Using Historical Data.” It was
published by the Journal of Financial Planning in October 1994 and later chosen as one of the best articles in the JFP’s first quarter-century of publishing. Bengen, who recently retired, is justifiably famous in the advisory world for this contribution.

You can find the article in several places on the web, such as here. It resulted in a popular rule of thumb, the “4% Drawdown Rule,” which suggests you can withdraw 4% of your portfolio value in the first year of retirement and then adjust the withdrawal in subsequent years for changes in the cost of living without running out of assets in a 30-year retirement period. Bengen later raised the permissible rate to 4.5% by including small capitalization stocks in the portfolio that he used to do the calculations. (As far as I can tell, he hasn’t considered the impact of global diversification.)
I take a special pride in that article since, if you look at the references at the end of it, you’ll see Bengen lists first an earlier article of mine from the Spring of 1994 entitled “First, Let’s Kill All The Asset Allocators.” My article is today so obscure that even I can’t locate it.

Right off the bat, I have to say that I’m not comfortable with the concept of a safe withdrawal rate. There are many problems with trying to come up with a set-it-and-forget-it withdrawal rate to get you through retirement:

  1. The future may be worse than the worst of the past. Bengen made it crystal clear in the title of his article that his calculation of the safe withdrawal rate was merely a statement about history rather than a prediction. But “the worst period in history” depends on when you are looking at history, and we have no way of being sure that, 50 years from now, “the worst thirty years in stock market history” might not be the period that is about to begin now.
  2. The retirement period selected for the study, 30 years, is arbitrary. A single individual has a shorter life expectancy than the joint life expectancy of a couple that age; younger retirees have a longer life expectancy than older ones; and as each of us is an experiment of one — our own life expectancy faces massive uncertainty unless the warden has already set the date for the execution.
  3. The spending of an actual person (or couple) in retirement is never a consistent amount from year to year, and most people gradually spend less as retirement progresses and the joys of travel, entertainment, and fine dining lose some of their luster.
  4. The ability to reduce spending in response to bear markets affects the sustainable rate of withdrawal. Since drawing a specific sum regardless of the current value of the portfolio results in a reverse-dollar-cost-averaging problem, a more flexible retiree can start higher and cut spending only if necessary rather than begin with a lower draw at the beginning just to avoid the need to reduce it later.

Asset allocation is extremely important, both in the sense that the commitment to equities matters and in the sense that a reserve for emergencies and/or bear markets can allow a higher withdrawal rate.

Does your death actually mark the complete end of your financial goals? Unlikely, unless you are so crunched for money that you have no ability to think about others. Do you have any heirs? Any
charitable legacies you want to leave? Not only is 30 years arbitrary, so is the assumption that your financial goals are limited to financing your retirement.

Putting your retirement car on cruise control the day you retire and paying no attention to the road ahead is, quite simply, idiotic. Most advisers, Bengen included, consider the rule of thumb to be a place to start a discussion and a way of giving a client some idea of the neighborhood of reasonable spending. But the flaws above are important, and there have been many subsequent attempts to improve on Bengen with different approaches to annual adjustments and asset allocation, and various decision rules and triggers. Some of this has become an example of Maths Gone Wild.

The philosophy of SimplyRich is to avoid unnecessary complexity. Common sense ought to tell you that what you can safely withdraw this year has little to do with what you withdrew last year, let alone what you withdrew in your first year of retirement. Moreover, you need a system that doesn’t fall apart because of unexpected expenses or windfalls. So each year’s calculation needs to be based on that year’s facts.

I do an annual review of the withdrawal rate with my retired clients and, as a do-it-yourselfer, you should find that is often enough. A simplified approach that is quite reasonable comes from the long history of endowment funds:


That’s it. Add up your investment assets, multiply by 5%, and that is what you can withdraw in the upcoming year. If that, added to your Social Security and other income, is acceptable, you’re done. Divide by 12, set up a monthly transfer from investments to checking for that amount, and get back to real life. In the following year, do it again: your new monthly withdrawal amount will be higher or lower depending on whether the market earned enough to cover the withdrawals that took place. Over time, it should grow around the rate of inflation (since your withdrawal is in the neighborhood of the long-run historical returns of equities). If it is a little lower, that fits the expected gradual reduction in real spending of retirees over time. If it is higher, keep in mind that the right to spend windfalls isn’t an obligation to do so, and if you want to leave your withdrawal where it is, be my guest. We were setting a maximum, not a minimum.

That’s it? Really? Well, okay, not quite. You probably should start your retirement with a reserve fund of around two years’ worth of spending money in cash equivalents (TIPS, short-term Treasuries, savings accounts, money market funds). This is not a part of your investment assets (for purposes of the 5% calculation). If you have to go over your limit because of an emergency, that’s where it comes from. If you need to supplement the reduced withdrawals from the investment assets during a bear market, that’s where it comes from. You might go a lifetime without touching it or you might, after a decade or more, use it up. What then?

Maybe nothing then. If your 5% investment withdrawals are sufficient along with other income, you could just live on that. If not, then it is time to reread my previous blog post on the proper use of annuities, and consider an immediate variable annuity for a portion of the assets that will allow you to safely take out more and not worry about outliving your assets.

Alternatively, once your life expectancy has dropped below 20 years, you could calculate your withdrawal by dividing your investment portfolio by your life expectancy, which will come to higher than 5% (since 5% is 1/20). Assuming your health is normal, this would occur for a single male at 66, a single female at 69, or an opposite-gender couple at 74. You’ll still have to adjust the withdrawals each year, both to account for the portfolio change and because your life
expectancy drops by less than one year for each additional year of life. You’ll never completely run out of money this way since you never take out 100% in a single year, but your portfolio will start to decline rapidly once your life expectancy is short enough to require distributions in excess of expected returns. At that point, a bear market would reduce your distributions massively. If you have other resources, such as a home whose equity can be tapped through sale or a reverse mortgage, you can improve the cash flow.

But as I suggested in my earlier blog post, once you want to draw more than 5% per year, your portfolio’s real value is probably going to start a long decline toward zero. If longevity-risk has you concerned enough, the immediate variable annuity might be a better option than increasing your withdrawal rate directly, especially when you have reached an advanced enough age that the simplicity of periodic payments guaranteed for life is worth the loss of control over the assets.

For most retirees, 5% from investments adjusted each year and supplemented by an initial reserve fund of a couple of years that is used when necessary has an excellent chance of lasting a lifetime. Obviously, all this depends on a willingness to remain an equity investor. Anyone who keeps his or her money in bonds should expect to earn virtually nothing after inflation. You have to be an owner of equities to have a chance of real returns near the withdrawal rate. And if the next few decades are so poor for equities that they can’t reach 5% per year in real returns (far below their long-term average), how do you expect bonds to do so? There is no safe withdrawal rate: life is uncertain. But at least give yourself the best chance.