This is a follow on to my last post on having a fixed withdrawal in retirement as one is going through a market crash like 2008-09. This post has the same assumptions and the same process, except
this time I use a fixed 3% spend rather than 4%. This is not at all what I meant by adaptive
retirement planning but the least one could say is that considering a lower spend
is a step in the right direction. The
same setup applies: 57 years old in 2004
with $1M invested in a 60/40 low cost fund except now with a $30k spend
rate rather than 40. We look to see what happens to the effective
spend rate month by month as the portfolio varies (from things like spending, inflation and
market changes). In addition, each year we calculate a forward-looking fail rate using
montecarlo and aftcast simulators using the new info available in any given new year (age, longevity and portfolio will have changed).
The questions are the same:
1. What would happen to a fixed withdrawal rate recalculated
as % of a changing endowment at each time T during a market crash,
2. What would happen to the forward-looking fail rate
estimate (with updated info) at any given time, and
3. How would it have felt as a real live retired person and
what would I have done?
The major and minor assumptions are the same (see previous post). This is what it looks like at 3% with the effective spend on the left axis and
the fail rate estimates on the right.
The scale is the same on the right and I nudged the lower bound on the
left because I am using a lower spend rate.
I added an additional MC simulation that uses the actual spend
percentage in each year in addition to the plain vanilla 3%. That's because if I were sitting at my PC in 2009: would I
plug in a straight up 3% or would I use the spend rate I actually have at the time (i.e., the fixed
inflated one I set in 2004 as a percentage of starting portfolio in January 2009)? There is a case for doing neither but I'd
probably do both, at least for this exercise anyway. Remember that the values here for spend rate and failure rates are recalculated at the beginning of each period - monthly for the effective spend rate and annually for the fail rates.
Let's check in with the questions:
1. What would happen to a fixed withdrawal rate recalculated
as % of a changing endowment at each time T during a market crash?
The peak effective spend rate hits ~4.5% in beginning of 2009. I
would have assumed it was going to infinity at the time but 4.5% is within the range of
estimates that the pros consider acceptable especially if I can nudge spending
down a bit at the worst point. I would
have bit my nails...and then used my foreknowledge of the giant rally to come to
sit tight and enjoy my declining spend rates (ok that latter part is a joke).
2. What would happen to the forward-looking fail rate
estimate (with updated info) at any given time?
The forward looking sims (on the same scale as the last post) are
hardly worth mentioning here (around 5-8% worst case fail rates). By any measure of simulation these are within
normal planning bounds. I might have
stressed out at the time but the math did not support the stress. I might have also run sims every day in Q1
2009 just to fan the flames of my gnawing anxiety (it would have gone up to
maybe 17.4% in a worst case simulation but If I had cut my spending down to say
3.5% -- about a $600/mo. change -- in
that month the sim would have projected a more benign 6.4% fail expectation)
but this is where I have to say that looking at simulations is a research tool
not a window into reality so stop looking at simulators!
3. How would it have felt as a real live retired person and
what would I have done?
I would more than likely not have lost my mind contra the last post. What little we know of retirement finance
would have told me that I was not going to die a financial death. It may have been agonizing but I know now
that I could have sat tight especially with the "floor plus upside"
approach [see note below, Pfau 2016] and even more so if I was willing to adapt by cutting spending and adapting the plan among
other things. This situation (low expected fail rates, declining spend rates) also, given
how blasé I seem to make it now, begs the question of what I might do with
legacy planning if I am not creating a reserve for future spending shocks by my apparent under-spending. Elegant problem to have I guess.
Here is the prior chart just for comparison. I added a MC
sim that uses not just a 4% spend rate but plugs in the % based on the actual
fixed/inflated spend divided by the starting endowment in a given month. Note the minor diff in the left scale.
Are there any conclusions I could make from these two recent
posts? Yes. In a nutshell the
conclusions include, among other things, the obvious:
Spend less, maybe closer to 3% than 4, and don't lock in a high
fixed spend rate
Avoid big crashes during retirement
If you do crash, have an historic rally after that
(kidding…but it helps)
Don't over-analyze or use simulators too often
Recalculate and adapt the plan as needed given conditions
Go do research on sequence of returns risk while you are at
it
Speaking of reading, for further reading on this and other retirement finance topics, there are a million references out
there but here are a few things to look at for now:
Sensitivity of Safe Withdrawal Rates to Longevity, Marketand Failure Risk Preferences with Implications for Asset Allocation, Butler
et al.
Can I Retire Yet, Darrow Kirkpatrick
The Retirement Café, Dirk Cotton
How Much Can I Afford To Spend In Retirement, Ken
Steiner
Retirement Researcher, Wade Pfau
The Only Spending Rule Article You’ll Ever Need, Waring and
Siegel 2014.
No comments:
Post a Comment