I've got a short new column at MarketWatch's RetireMentors called, "Retirement: Two Different Views on the 4% Rule."
The first view is the standard Bengen, Trinity-study approach of basing safe withdrawal rates on historical worst-case scenarios.
The second view is the one I ascribe to, which is that current market conditions are much better indicators of what will be sustainable than historical worst-case scenarios. Historically, Shiller's PE10 has done a good job explaining sustainable withdrawal rates.
This second view is easy to misunderstand. I'm saying that with it, 4.2% is the best guess about the actual sustainble withdrawal rate for current retirees. It's not a safe withdrawal rate. The safe withdrawal rate would be less.
This short column didn't incorporate bond yields, which are also good indicators about sustainable withdrawal rates. The fact that bond yields are at historic lows also suggests that we will end up on the low side of that 4.2% guess.
Wade,
ReplyDeleteIf you have this data set, have you looked at using bond yields with PE10 and fitting a model to that data to get a point estimate for today's sustainable withdrawal rate? Is there a roadblock or difficulty with that model (other than current rates being outside the range)?
- Kris Carroll
Kris,
DeleteJust now I was returning to this subject. I did write an article in 2011 which included PE10, bond yields, and dividend yields, but I think the dividend yields being outside their historical range caused significant problems with that.
Just now as I look at bond yields and PE10, including the bond yield in the regression does reduce estimates a bit. I'm getting 3.9% instead of 4.2% as the best guess about what will happen. This is something I will try to refine further and write about.
Despite caveats, how many people do you suppose read this or the MarketWatch column and say: Good news! The 4% safe withdrawal rate has now increased to a sustainable 4.2% rate.
ReplyDeleteYour sustainable rate is a base case that is superior to a base case using average historical rates. SWR is intended to be a worst case rate and not a base case rate -- apples and oranges.
John, you are absolutely right. I'm seeing that this is an easy one to misinterpret, despite my caveats. A well-known personal finance journalist even asked me if I had revised all of my estimates upward, now. The answer, of course, is no. This estimate is not the "safe" withdrawal rate, it's the best guess about the actual withdrawal rate with a 50% chance for failure.
Deletevery helpful charts. It would also be useful to see not only the 53% probability, but also the 90% probabilities based on current PE10, which i am wildly guessing from your chart would be the range between 3.2% and 5.2%. To see the curve above and below the 53% data points would also be helpful for planning purposes. Thank once again Wade for a very useful article.
ReplyDeleteDerek,
DeleteThank you. It's a good idea, and really the best that can be done is to try and fit in reasonable-looking curves the way you are suggesting. This is because each of the points in the graph are not independent. They are from rolling 30-year periods. So traditional confidence intervals will not be correct. Calculating proper confidence intervals for this situation still seems to be at the cutting edge of theoretical econometrics. Also there is the issue that we don't have many datapoints when PE10 is so high, which would further confound the calculation of confidence intervals.
Larry, thanks, and that's a great assessment. Given one's Social Security and other income from outside the portfolio, it's really a matter of how much one wants to rely on dynamic spending and how much on income annuities / holding bonds to maturity.
ReplyDeleteMr. Pfau:
ReplyDeleteI have a 403b that gives me a fixed annual 7% interest rate and is 40% of my total portfolio (the rest is in equities). I use my 403b instead of bonds in my 40/60 split. What would the sustainable withdrawal be with my 403b throwing off a fixed 7%?
Chaz,
DeleteWith a fixed rate of return, this is something you can calculate easily in Excel. The answer depends on how many years you wish to sustain withdrawals, and what you believe inflation will be (assuming you want inflation-adjusted withdrawals and assuming that fixed return is nominal).
I must say that a 7% return sounds high. Are you sure that you will have access to this money? It's not something like a guaranteed growth rate for a hypothetical benefit base from which you can later annuitize, but which isn't actually money available to you, is it?
The 7% return is constitutionally guaranteed by New York State and is not an annuity.
DeleteYes, I do have access to the money.
Very interesting. I've been reading up on this a bit.
DeleteWell to get back toward answering your question... if you withdraw 7% a year, it means you are withdrawing your annual investment growth without dipping into your principal. Withdrawing more than 7% will start to spend down principal as well. How much higher you go than 7% depends on how long you want it to last. This sort of calculation can be done in Excel.
For example, if you want $100,000 to last for 30 years and it gets 7% returns, then:
=PMT(0.07,30,100000)
tells you that you can spend $8,058 per year.
Actually, if you want to withdraw at the start of the year, rather than the end of the year, you use:
Delete=PMT(0.07,30,100000,0,1)
which supports $7,531 of annual spending
Mr. Pfau:
DeleteI was trying to figure out what the safe withdrawal rate would be (30-40 years) if 40% of my fixed portion of my retirement income is throwing off 7% annually while the rest is in equities?
Moreover, would this allow me to take 6% out of my equity portion, using the guardrails outlined by Jonathan Guyton?
With a 7% return and no volatility, I'd suggest treating these two portfolio components separately. Look at what you can sustain from that, and then with the rest you can be looking at the results about withdrawal rates when using 100% stocks. The income from the 403(b) provides you with greater risk capacity, but beyond this I don't know enough about your circumstances to even begin to make a suggestion about where you would fall on the spectrum between spending more now and bearing the risk of greater cutbacks later, which is what the guardrails do.
DeleteThanks for the interesting data.
ReplyDeleteI think you suggest that the "sustainable" or "safe' withdrawal rate is actually dynamic, or a moving target. That is problematic for pre-retirees and retirees alike. The point of having a "sustainable" or "safe” withdrawal rate is to give people peace of mind that things are going to be “allright” if they stick to a simple rule.
It may be foolish to expect that a simple rule would work in a complex world. Indeed, how can we reasonably expect that today's assumptions will hold true tomorrow? Yet that is what people look for.