Please note that when I use SWR in this post it is used in a very limited sense: a withdrawal rate that survived all complete 30 or 40 year historical intervals since 1871. I didn't look at partial periods, leading to some "interesting" results where the 40 year historical SWR was higher than 30 year rate. Results to 2 decimal places are reported, for no good reason. I the difference between 4.2 and 4.1 is likely insignificant in these studies. The REHP Excel spreadsheet is available here:
http://www.retireearlyhomepage.com/re2002i.zip
The post below was originally posted at:
http://boards.fool.com/Message.asp?mid=18128601
Since intercst graciously updated the REHP spreadsheet to include
Schiller's PE figures (calculated using a 10 year earnings average),
I've been playing around with the PE switching function a bit.
While I haven't done a comprehensive analysis. It does look like there
is enough promise in the results to warrant further study.
Settings:
Front end/backend payout: 50%
CPI inflation.
Other settings as default.
This scenario has a base 30 SWR of 4.12% at a 60% stocks / 40%
commercial paper allocation.
Base 40 year SWR is 3.99% at a 76/24 allocation.
Activate the stock switching feature by setting intial allocation to 0%
and fixed income series to 6.
Using both high and low PE for switching did not result in anything
interesting. But using just the low to mid switching PE while leaving
the high PE allocation matched with the mid I've found some
interesting results.
It seems like varying the stock allocation in the higher PE years away
from 60% only lowers the 30 year SWR. But, bumping the stock allocation
in lower PE years raises the possibility of some benefit.
For instance:
Allocation for low PE Years: 90%
Allocation for mid and high PE years: 60%
Switch 30 year Notes
at PE SWR
< 8 4.12% PE so low no effect
8 4.15 PE so low little effect
9 4.15 "
10 4.15 "
11 4.41 Big jump here. 40 year SWR is 4.24%!
12 4.38 40 year SWR: 4.28
13 4.38 40 year SWR: 4.32
14 4.13 40 year SWR: 4.21! Inversion here showing effects
of the current bear market on 30 year but not 40
year withdrawal periods. Interesting!
15 4.12 40 year SWR: 4.19
16 4.15 40 year SWR: 4.20
17 4.23 40 year SWR: 4.06
18 4.31 40 year SWR: 4.09
19 4.06 40 year SWR: 4.15 The first time our stock
switching strategy falls below the static
allocation for a 30 year withdrawal period.
What should we take from this series? For one, switching did not hurt
the SWR until you got to a high PE, so attempting a modest PE
switching strategy should be fairly harmless. The 90% allocation at
lower PEs is very aggressive.
One more series:
Low PE stock allocation: 80%
High PE stock allocation: 60%
Switch 30 year Notes
at PE SWR
< 8 4.12% 40 year SWR: 3.70%: rarely in upper allocation
8 4.15 40 year SWR: 3.70
9 4.15 40 year SWR: 3.71
10 4.15 40 year SWR: 3.71
11 4.32 40 year SWR is 4.06%
12 4.30 40 year SWR: 4.17
13 4.30 40 year SWR: 4.19
14 4.14 40 year SWR: 4.19
15 4.13 40 year SWR: 4.12
16 4.15 40 year SWR: 4.12
17 4.20 40 year SWR: 4.02
18 4.25 40 year SWR: 4.05
19 4.09 40 year SWR: 4.11
This series is pretty similar to the other. Again we get no
deterioration in SWR until you switch at a historically very high PE,
leaving you over allocated to stocks when the market is overvalued. At
low PE switch there is little benefit, but no harm. Setting a switch
at 13, just under the markets long-term average might allow you to
catch some extra performance, but I can't say I would increase my
withdrawal unless the extra performance was already in the bag.
Note that both of the series above use a 60% stock allocation for the
higher PE years. If you were looking to optimize for periods longer
than 30 years, you would likely bump that up some.
Nothing conclusive here. But I think it does demonstrate that some
judicious use of market valuation doesn't hurt and can potentially give
outperformance. Using the (Price)/(10 year average Earnings) PE was key I
think, and I would be interested in looking at an even longer average.
I can't speak to the statistical relevance of the series above, and
I'd want to examine a lot more series before trusting it too much. I
definitely don't recommend that anyone set a switch at PE 11, as in
the first series above, take 4.4% and think they are '100%' safe for a
30 year withdrawal period.
Ben
PS While I was careful working on this there are a lot of numbers and
an error or two could have crept in.