“How much do you need to retire? (Or, if you are already retired, how much can you safely spend?) These are the most important financial questions that our clients face. Although each family’s situation may call for a different answer, there are some guidelines that will serve most investors well.
“In October 1994, William Bengen published an article in the Journal of Financial Planning in which he analyzed (using Ibbotson data) how much an investor with 50% in the S&P 500 and 50% in intermediate-term Treasury notes could safety withdraw without depleting his savings. He found that for all 30-year periods covered within the data available to him, an investor could withdraw 4% of principal initially and increase that amount with inflation without running out of money. Subsequent to the publication of his study, it turned out that investors who had started drawing on their retirement assets in the mid-1960s would have run out of money in less than 30 years after the ravages of the 1966-1982 period (when inflation outpaced stock and bond returns).
“The problem with the conclusions of that study is the doubt that stocks and bonds will match their long-term historical returns in the future. This is especially true in the case of bonds where the historical return of 5.4%/year total return in the Ibbotson series is mathematically improbable with interest rates as low as they are now.
Underlying assumptions about returns and inflation
“The ‘safe’ rate of withdrawal from retirement savings depends crucially on the return and inflation assumptions you make. From 1926-2011, the average return from equities was 10%/year and from intermediate-term Treasury bonds, 5.4%/year, while inflation averaged 3%/year.
“Going forward, I consider it prudent to assume lower returns than these. I believe it reasonable to assume that bond total returns will just keep up with inflation as is now the case. This is far worse for investors than the situation from 1926-2011 when bond total returns beat inflation by 2.4%/year. ...
“For the purposes of discussion I believe it preferable to make conservative assumptions, especially since economic growth may well have slowed down permanently from 3%/year above inflation to 2%/year. Therefore, I have assumed that stock total returns will outpace inflation by 5%/year.
Monte Carlo simulation results
“I took monthly total return data for the S&P 500 Index and the Barclays U.S. Aggregate Bond Index for January 1976-March 2013 (Steele Mutual Fund Expert). The average returns for stocks and bonds during this period were 12% and 8%, which is higher than projected. I therefore subtracted a constant amount from each month’s returns in order to re-center them around 5% for stocks and 0% for bonds, which are the inflation-adjusted average returns I wanted to assume.
“I then randomly selected different monthly pairs of stock/bond real return data from the adjusted data to make 35-year simulated results. ... I did this 1,000 times in order to get a feel for the potential range of outcomes. (No future result is guaranteed, of course.) This is called a Monte Carlo simulation.
“For each simulated result, I determined how long the savings would last for different rates of spending. The spending rate is a percentage of initial principal, and the calculations in the table assume that annual spending increases at 3%/year. ...
“As one would expect, on average the greater the allocation to stocks, the longer you could live off of the money. However, an average result is just a 50% likelihood. Most people want greater certainty than that. I therefore also calculated the length of retirement that could be funded with 90% probability. The results are shown in the table below.
“As one would expect, on average the greater the allocation to stocks, the longer you could live off of the money. However, an average result is just a 50% likelihood. Most people want greater certainty than that. I therefore also calculated the length of retirement that could be funded with 90% probability. The results are shown in the table below.
Returns and drawdowns are in terms of purchasing power (i.e., adjusted to take inflation into account). The worst market drawdown was actually 50%, but the impact of inflation means that from the market top in March 2000 until the market bottom in February 2009 (based on monthly closing total returns), an investor in the S&P 500 Index would have lost 70% in purchasing power.
*10th percentile, meaning there is a 90% chance your money will last this long
Implications
“The optimum portfolio for those who want to be 90% confident of their retirement plans is 50%-75% stocks and 50%-25% bonds. The rare chance of a stock market disaster makes a retiree safer if he includes some bonds in his portfolio for protection. On the other hand, because stocks have the greater potential to keep ahead of inflation, investing in 100% bonds is also less likely to be optimal than a blended stock/bond portfolio.
“The good news for investors is that it should be possible to improve investment performance compared to the benchmarks, particularly for bond market investments. The current interest rate climate favors high yield bond funds and floating-rate bond funds over investment-grade bonds, and especially over Treasuries. Moreover, high yield and floating-rate bond funds should prove to hold up better when (if) interest rates begin to rise. Of course, both high-yield and floating-rate bond funds expose investors to significant potential risks, as we saw in 2008. They are suitable only if held in conjunction with the use of a timing model to reduce risk, such as the high yield bond fund signals we utilize for our clients and that we report in the newsletter.”
Marvin Appel, Systems & Forecasts, 800-829-6229, www.systemsandforecasts.com, 5/8/13
