Re: SWR, terminal values, TIPS, I-bnds & comm pape
Yes, I see the logic in reducing your success target in order to accomodate your assumptions. Do you have any suggestions on how much to reduce your SWR target for each percent you reduce the market return? Why 80%?
I didn't get the impression from the Bogle article that he was saying that the average real rate of return would be less than average for more than the next 10 years. I only saw his predictions for the next 10 years.
But a look at the Bogle data for the past century indicates that there were many 30 year periods that produced market returns that were well below the average (ave=10.4%).
So even if you believe that Bogle's lower limit estimate of 6% for the next 10 years may be applied for the next 30 years, FIRECALC is already including a case from 1910-1939 that is worse than that.
In his article, Bogle chose "decades" as the periods over which to analyze market returns. Although his projections (as of June 2003) were limited to the next 10 years, any retired person who expects to withdraw money from their assets for longer than 10 years is making certain assumptions about future market conditions, whether or not they do so consciously. For my part, I think that it is both prudent from a personal financial planning standpoint, and logical for reasons of fundamental economics, to assume that the reduced stock market returns predicted by Bogle will continue indefinitely past the next 10 years.
In other words, this rate of return is what I consider to be the EXPECTED (average) rate of return in the future. This implies that the actual returns during roughly half of future periods (of whatever length) will be less than this. To simulate the possible impacts of this on a person's SWR, I don't feel comfortable using historical data AS IS from a period of history when the average rate of return on stocks was about 10.5%, even though it includes some periods when the return dropped below this.
The one way in which I feel that Bogle's paper could be improved would be for him to evaluate market returns on a real return basis, which would factor out the varying effects of inflation. If this were done, I guess that his projection of a 6% to 9% return on stocks over the next decade would translate into a projected real return of 4% to 5%, which is pretty modest in comparison to the historical average real return of about 7.5%.
Looking at historical 10 year periods, the average real return on stocks was this 7.5%, and the standard deviation was about 3.5%. From the standpoint of a person drawing down assets, a high average real return is good, and a high standard deviation is bad. Even though I am pessimistically predicting that the real rate of return in the future will drop to about 5.5%, my one optimistic assumption is that market returns will be less volatile, causing the standard deviation to be less. The way that I feel most comfortable modeling this with FIRECalc is to adjust the historical data to mathematically reduce the real return on stocks by 2%. Since I can't make a similar adjustment to reduce the standard deviation of the returns, I accept a lower "success rate" of around 80%. This has the effect of disregarding roughly the 20% of the periods (such as the 1930s) when the departure of market returns below the average was greatest. Admittedly, this selection of this "success rate" is somewhat arbitrary, but for that matter, so is the selection of a success rate of 95%, 100%, or whatever, based on a historical pattern of returns that is unlikely to be repeated.
What I have found to be interesting in running FIRECalc with various scenarios that I consider to be "reasonably conservative," is that for planning periods over 20 years, and for mixtures of stocks and TIPs (with about a 2% interest yield) ranging from 10% stocks to 80% stocks, the SWR works out to be very close to 4% per year. This will happen if a person runs FIRECalc "as is" and uses a "100% success rate."
Having a higher percentage of stocks drops the SWR slightly, and increases the "expected" terminal portfolio value substantially.