Following up from yesterday’s blog post, it is worthwhile to also look at valuations-adjusted withdrawal rates.
A relationship that has been documented on a number of occasions (my paper in the August 2011 JFP about this builds upon a number of earlier works, and the history of these developments is reviewed quite effectively in Todd Tresidder’s “Are Safe Withdrawal Rates ReallySafe?”) is that market valuations at the retirement date are fairly closely related to that retiree’s subsequent maximum sustainable withdrawal rates.
This can be seen in the following figure by comparing the thick blue curve (these are the actual maximum sustainable withdrawal rates over 30 years for a 100% stock allocation) with the dashed blue curve (this is EY10, which is just 100/PE10… overvaluation means high PE10 and so low EY10, and vice versa). The relationship is not perfect, but the two curves move together fairly closely.
The red curve is the “valuations-adjusted” withdrawal rate. It is the actual MWR times PE10 and divided by the median PE10 over the entire period (15.71). This normalizes the withdrawal rate. MWRs tend to be quite high for those entering retirement when PE10 is quite low, and so this calculation would adjust the MWR downward. As well, MWRs are low when valuations are high (high PE10 or low EY10), and this calculation would increase the MWR. The median MWR over the whole period is about 6.5%. If EY10 was always equal to the MWR, then the red curve would basically be flat at about 6.5%.
But while the relationship is pretty good, it isn’t perfect. The red curve gets lower at times that EY10 exceeds the MWR, and the red curve gets higher when EY10 falls below MWR.
I’m not exactly sure how much value there is in looking at the red curve, but it is great if it can spark some further interesting insights from anyone, and it is at least worth a blog post on the subject.
Notice that the red curve doesn’t get far above 6.5% very often. That is because it is somewhat rare for the MWRs to be above EY10. It does happen occasionally. In particular, we can see this happen in the late 1800s, late 1940s, and early 1950s.
This is one of the reasons I am fearful for retirees in 2000. EY10 is at its lowest point in history, just a little over 2. There will need to be a break from the usual trends seen in history for the 2000 retiree’s MWR to get above 4%. Will the 4% rule still work?
Of course, this is for 100% stocks, which is not the allocation that most retirees will use. But my article I linked to above shows this same general problem for a 60/40 asset allocation as well.