Saturday, July 2, 2011

Retirement Planning and Worst-Case Scenarios

Reading a new article in the July 2011 Journal of Financial Planning, “A Safer Safe Withdrawal Rate Using Various Return Distributions” by Manoj Athavale and Joseph M. Goebel, has me thinking about low sustainable withdrawal rates. 

The popular rule of thumb in the United States is that a 4% initial withdrawal rate from your assets at retirement, with this amount adjusted for inflation in subsequent years, will allow for at least 30 years of withdrawals with a stock allocation in the range of 50 to 75 percent.

Athavale and Goebel suggest that 2.52% is a better approximation for a safe withdrawal rate.

Despite not having happened yet in the United States, such a low sustainable withdrawal rate does resonate as a possibility with me. As I explained in Pfau (2010), the United States enjoyed a remarkable 20th century for asset returns, and things didn’t go nearly as well in other developed countries.

I will now talk about results which I had to cut from the paper because of space constraints. In the years since 1926 and for a 50/50 asset allocation, the 4% rule would have failed retirees in 10 of the 17 developed countries more than 25% of the time. These countries include the Netherlands, Norway, Australia, Ireland, Spain Belgium, Germany, France, Italy, and Japan. Remarkably, the 4% rule would have failed more than 70% of the time in Spain and Italy. A 50/50 allocation would have allowed for a maximum sustainable withdrawal rate of only 0.24% for a Japanese person entering retirement in 1937. Now, the historical situation obviously points to a black swan for those retirees, but the broader message is: the only truly safe withdrawal rate is 0%.

In these terms, it is really just a matter of time before a situation arises in which a cohort of American retirees experience a 2.52% withdrawal rate.  As more time passes, the only possible direction that the lowest withdrawal rate experienced can move is downward. It’s possible that the Americans who will experience this have already retired. I'm looking at you, retirees in 2000.

So, in the process of planning for retirement, what kind of withdrawal rate should we plan for?

Honestly, I think there is a better way to think about this whole issue. 

Because, first, deciding on your “safe withdrawal rate” means you must develop a “wealth accumulation target” for your retirement date. But as I argue in Pfau (2011b), it is very difficult to know if you are on track for meeting your wealth accumulation target.  The idea is based too much on assuming a fixed rate of return for the compounding growth on your wealth. Over short periods of time, returns vary greatly. Individuals who save for retirement over a 30 year period using a fixed savings rate will have all ended up with very different wealth accumulations (in real terms) by their retirement date. What’s more, the amount of wealth they have even 5 years before retirement provides little predictive power about their ultimate outcome.  This is related to the “portfolio size effect” that the big changes in your wealth occur just before retirement when your wealth is the largest and changes by the biggest amount for a given percentage return. Michael Kitces also gets at the heart of this problem with his discussion of the logic of compounded returns. With the portfolio size effect, you can’t really predict very well about how much wealth you will end up with even 5 years before retirement, because you can’t predict your returns over those 5 years and those 5 years of returns may have a bigger impact on your total wealth than much of what happened in the first 15 or 20 years of your career, for instance.

If you can’t really predict your wealth accumulation at retirement, this also makes it hard to know how much you will be able to withdraw using your “safe withdrawal rate.”

But in Pfau (2011a), I suggest doing away with wealth accumulation targets and safe withdrawal rates.  We can’t control how much wealth we end up with at retirement. We can’t control what the maximum sustainable withdrawal rate we will experience in our retirement will end up being. But we don’t need to control them.

What we can control is how much we save and how we invest! The characteristics of the following analysis can be modified to fit your personal characteristics, but to give an idea of the “save savings rate” concept described in Pfau (2011a), consider someone who earns a constant real salary over a 30 year career who wants to save enough so that her wealth at retirement would have been sufficient to cover 40 years of withdrawals equal to 50% of her salary (Social Security would be added on top of this). She also doesn’t spend a lot of time worrying about her investments, but simply uses a 60/40 asset allocation into low-cost index funds (I am assuming zero costs for this example) rebalanced annually. In the worst-case scenario provided thus far by US history (since 1871) such a person would have needed to use an 18.63% savings rate for each of these 30 years to have always had enough wealth to be able to enjoy the sustainable retirement under the terms she wanted. I didn’t say anything about what her wealth accumulation or withdrawal rates are, but whatever the wealth was and whatever the withdrawal rate needed to be, they worked out to be fine.

Now this analysis is vulnerable to the same problem as the safe withdrawal rate.  Just like the US retirees in the future may lose in the cosmic lottery of suffering from a new worst-case withdrawal rate, in the future there may also be retirees who will find that they needed to save at a higher rate than their "safe savings rate" to enjoy the retirement they desire.

The savings rate is something you can control now.  If you are worried about new worst-case scenarios and not having enough, then you save even more than the “safe savings rate”.  But you have to find the proper balance.  How painful is it to save more? While it might be ideal to save with a 2.52% withdrawal rate in mind, it may not be practical. You don’t wish to deprive yourself and your family too much now, because the chances are small that you will experience this worst-case scenario.  And you also have to think about what will happen if the worst-case scenario did come to pass in retirement.  Your wealth might be depleted while you are still alive and you may be left with just Social Security benefits. So think about the sacrifices it takes to save more now, and think about what you want to protect against in the future, evaluate the trade-off and just do the best you can. I think the benchmark you should be working from is not to target wealth needed for a safe withdrawal rate, but to think about whether you want to save more (or less) than your “safe savings rate”.

The more I think about safe withdrawal rates and how low they might be, the less useful the concept comes to be in my mind. There’s not much we can do about it if the fate of our retirement year leads to disaster. Instead, think about how much you can save and how long you want to work. Pfau (2011b) provides more guidelines about whether you are on the path for a sustainable retirement for a wide range of possibilities.

Further Discussions

Mike Piper mentions this post in "Should You Own Stocks in Retirement?"