I patted myself on the back earlier this year for what I
considered to be monumentally heroic -- and also what I considered at the time
to be successful -- efforts over the last four or five years to strip my costs
down to a bare minimum given my growing awareness of the unusual spending risk held
by an early retiree. But, wait, not so
fast… While I had cut my operating budget by at least half or more, I kept
stubbornly ignoring what I'll call hidden spending (or at least what I kept
hidden from myself); maybe you already do this kind of thing but me, I never
got around to it. This week I finally took
a look at some of the "other" things that consume retirement assets
just as surely as direct spending does but that do not get my attention as much
as they should. I also wanted, in the process, to give this hidden spending some
context especially from the perspective of an early retiree.
First let's talk about math.
The spending formula is pretty simple: A/B, numerator/denominator,
spending over something. On the other
hand, the first thing I noticed when taking a closer look at this subject of
hidden spending is that there is a lot of latitude in what gets included in A
and B -- which means there is a lot of slack in what number one can come up
with as a result of A/B. [1]
Before we get to the supposedly all-important spending
numerator, though, let's look below the line.
The denominator seems obvious enough, right? But it isn't as far as I
can tell. It took me a long time to
settle in on what number to use. One
might think "assets" is enough but I took it a little further (ask my
gf about my over-thinking…sometimes it's useful, sometimes not. Let's assume it
is useful here). What about liabilities
that offset assets or will at least consume assets in the future? Of course, that
makes sense, so net assets it is. But what
liabilities, then? Tax liabilities and bank debt, sure. But I also have softer stuff like education
liabilities on my balance sheet and I also have some 529 assets (shared with an
ex-wife) that partially offset that liability.
So I also include what I call net unfunded education liabilities. So far so good, don't you think? But one
could take it whole lot further and create an actuarially accurate liability
set and balance sheet that includes the present value of all sorts of things
like long term care expense x probability, health shocks, home repair, divorce
risk, investment losses, etc (see Ken Steiner at howmuchcanispendinretirement.com).
On the other hand, that effort might also result in totally zeroing out the denominator
which would mathematically push spending ratios to infinity or indefinable numbers
which I guess is the whole point of a true actuarial balance sheet (i.e., get completely
away from spending and withdrawal rate math) but that's not my task, at least
for today. So a simple set of near term hard
and soft liabilities is what I used: debt, taxes, net education obligations.
What about the asset side of the denominator, then? I could move towards the actuarial approach there
too and include the PV of things like social security (adjusted by a "it
might not exist in the future" factor, perhaps) or human capital (I know
mine is not zero yet but I assume that it is).
In addition, I know that if push comes to shove, I would probably monetize
my home by downsizing or using a reverse mortgage. But I don't really want to use any of that stuff
as a factor in the math…yet. For now, to
keep it "simple" I mostly use, for my denominator, what I'll call
normal monetizable financial assets (i.e., I exclude residential real estate
and the PV of social assets) and then I subtract out (see above) hard
liabilities like bank debt as well as soft liabilities such as my unfunded
education exposure and I call that result, my denominator, "net
monetizable worth" (NMW). That's
harder than it sounds maybe but I really wanted to get to a denominator that
represented a realistic net worth base from which I would truly consume for my
retirement needs over a lifetime. My old
traditional net worth might include assets like furnishings or the full value
of a house but on most days I don't
visualize monetizing that stuff and then consuming it. The closer I got to zero, though?…hmmm. Everything
is on the table before I expect my children to carry the burden.
So, now for the numerator.
I always thought that I should just count normal spending. But it dawned
on me not so long ago that anything that eats away at NMW is in the retirement game
as well and needs to be counted.
Groceries, yes, sure. Mortgage, yes. Phone bill, insurance, travel costs,
credit cards: of course. But what about
advisor fees? I hadn't included that. What about the buried expenses for ETFs
and Mutual funds or private placements? Nope, not that either. What about deferred maintenance on the house
or depreciation on a car? That stuff eats away NMW, too, but I hadn't thought
about it. All that needed to be considered (that’s all the stuff that I'm
calling hidden spending). But what about something like the excess
inflation expected in education or health spending? Maybe, but that's a little
abstract: out. Taxes? Yes, but I include that in spending or at least I'm
pretty sure I get all of it. If not, I'm not worrying about it today. For my purposes, I wanted to look at normal
consumption plus obvious other things that consume NMW and that are simple to
calculate. That means the numerator = regular
spending + advisor fees + fund fees + auto depreciation + any other obvious
stuff. Or at least until the next time I
do this.
Let's see what A/B looks like in this run.
Let's take a look at this:
- Core spend rate = 2.8%. Cool! and sustainable…but is it really? Let's look a the hidden stuff first.
- Advisor Fees? 39 bps. Relatively low, but that's because I: a) calculate that as a % of total net monetizable assets vs. a % of managed assets), b) took over a large portion of my own assets a few years ago, and c) I negotiated the remaining advisor fees down as best I could. Yes one can negotiate. Give me a few more years, though, and I'll go to 100% self managed and passive if I can.
- Embedded fund fees? These were high in 2014 but once I got it that these hurt long term success I started to strip these down too. See previous point as well. My goal is to get to a 100% very low cost passive portfolio within the next 5 years or so. 15 bps in expenses might be a goal in this area. Others have done it so I know something like this is achievable.
- Auto depreciation? I don’t usually include cars in net monetizable worth because I don't plan to monetize my car (and held long enough they depreciate to zero anyway) but I temporarily have a second car that I would sell in a heartbeat. That is monetizable and it does depreciate. That cost sucks away networth so is an expense: added.
- Deferred home maintenance? Very real but kinda abstract too. Not included today.
- Sub optimization "opportunity cost" of investing like my cash allocation? Too abstract. Not added.
- Taxes? I think I already account for most of these.
- Other? I'm sure they are there but not for this post.
The net? It looks like my consumption of NMW, when hidden
costs are added, is more like 3.6% than it is 2.8% if I account for it properly. Worse result, I guess, but still possibly
within the boundaries of sustainability, depending on how one looks at
it…something I'll get to in a bit.
So, my reward for that monumental and difficult task of
driving spending down as hard as I could to 2.8% over the last five years to
mitigate risks that now look, on a good day, like only I know about? I suppose all one can say at this point is
that I got: 1) the counter-thrill of finding out it wasn't really 2.8%, and 2)
a surprisingly unpleasant and unwelcome pushback from a non-trivial number
people around me implying that I am living a life of excessive and unhelpful
(to whom I ask?) self-denial and frugality (of course we have to realize, I
suppose, that none of them will be there to save me when I really need it; this
is a journey of one in the end). That
last comment (#2) included a financial advisor who mocked my frugality efforts
with no small amount of pseudo-professional condescension. If that right there weren't enough to fire
him on the spot, it also implied that he did not know his craft (financial
advisory) as well as he thought he did or at least as well as I did (see below).
And forget for the moment that, from my point of view, there may be a moral or perhaps
even environmental dimension (if one were to care about such things) to
over-consumption. I fired him the next
day, of course.
Here is at least one thing that he was missing on the
financial front. There is an age based
context for spending that affects early retirees in an especially severe way
that is radically underappreciated except perhaps by people that retire early. That's because an early retiree has a much
longer planning horizon than is typical.
You? You've probably seen an ungodly number of articles or research on
spending rates and withdrawal rules. The thing that most people who read this
stuff don't see or remember, though, and that most authors forget to mention is
that many of those studies are based on a 65 year old and a fixed 30 year
longevity estimate. That is a pretty
narrow slice of the world. If one were to be, say, 58, and plan for a long
healthy retirement, the horizon could be more like 40 years… or maybe even more. Whatever spend rule is based on 65/30 is totally
irrelevant if not destructive to the 58 year old with a 40+ year planning
horizon (even more so at 50, of course).
Let's put my spending efforts and hidden spending thinking into
an age based context, then. Here below
is a chart I did for another blog post. For
that post I took a bunch of free retirement tools and formulas and I attempted
to solve for spend rates that were consistent with high probability success
rates (>= 90%) , moderate risk portfolios, moderate inflation, and longevity
that was either to age 95 or stochastic distributions based on known longevity
math. I did that for each age year from 58
to 75. This is what it looked like. The circles with letters are my spending data
points which I'll describe below:
First, the lines/tools:
- Blanchett Simple: this is David Blanchett's Simple Regression Formula (for
retirements > 15 years) based on a bunch of MonteCarlo modeling where
Withdrawal = .195 - .03701(LN(years)) + .01255*sqrt(equity%) -.04471 *
(probability of success) + .507*alpha. I'm thinking this is
probably simple only to an advisor or actuary or an OCD retiree like me.
- Kolmogorov: Kolmogorov differential equations for
lifetime ruin probability as profiled in Moshe Milevsky's "7
Equations" book. I found them instantiated in a spreadsheet found
somewhere on the internet.
- Flexible Retirement Planner: free internet MonteCarlo
simulation tool.
- cFireSim.com: free internet historical rolling model
style simulator.
- FireCalc: same as cFireSim give or take a few minor
things,
- Custom Sim: I hacked out a basic excel/VB MC
simulator using historical returns and inflation but I also added in,
among other things I don't see (but want) in other simulators, a random
longevity variable using a gompertz distribution. As an early
retiree I got tired of always seeing the fixed XX-year kind of thing. Custom
Sim also has a line on the chart (black) where I forced a tactical
suppression of returns for the next 10 years, a not unreasonable
assumption at this time.
- Divide by 15/20: This is an actuarial rule of thumb created by Evan Inglis. Divide age by 20 results in a safe spend rate by age that is simple to calculate, is linked indirectly in a way to real return assumptions, keeps the risk relatively low, and more than likely leaves a legacy. Divide by 15, on the other hand, is arbitrary and designed by me just to make the chart look good rather than actually mean something. Evan used divide by 10 as an upper bound.
- ARVA (or Annually Recalculated Virtual Annuity): in
an article by Waring and Seigel (The
Only Spending Rule Article You Will Ever Need, M. Barton Waring and
Laurence B. Siegel, 2015 ) they use the Excel PMT function
to calculate spending each year based on endowment, real return, and
years. My chart has one line for "1M to age 100 at 1% real
return" and another at 3% real. These rates are arbitrarily
selected to show low real return expectations. Terminal age in the formula
(100) is a little higher than the other tools because PMT() is kind of a
blunt instrument. Maybe that's a little bit of a cheat.
The spending data points (A-C in the chart) as well as the
essential messages I wanted to convey about the early retirement "spending
context" are as follows:
A - This point is where I thought I really was until today. It took me many years and a lot of hard work
to get there. Given the age adjusted
spending chart above, I thought I was doing pretty well and that I had tamped
down hard on risk given my age. I was kind of proud of that and this is for
what I (prematurely) patted myself on the back -- only to now disabuse myself of
that notion with this blog post.
B - This is where I probably really am today with many, but
not all, of the hidden costs added in (maybe even higher if I were to be fully
honest about hidden spending). This
still does not look fatal on the chart but it is clearly higher risk than I was
expecting or wanted. Then, just to pile
on, if one were to internalize the message implied by the lowest black line,
where I try to suppress returns in the first 10 years -- something that most
pundits think will be the case going forward -- of a bunch of Monte Carlo
simulations, then point B looks like much higher risk than I want. Even the original 2.8%, if I could have managed
to keep it there -- and keep in mind that that rate would skyrocket if the
value of the portfolio falls in, say, a crash -- looks more dangerous than I
was originally thinking before I suppressed returns in a few sims. The arrow on the chart, by the way, merely describes the
move from "before the hidden costs" to after.
C - This is the spend rate where I would have been if: 1) I
had not taken over many of my own assets to mange myself and then negotiated other remaining
fees down, and 2) I had not altered the cost structure of some of the funds and
placements I owned in 2014. The point here, I guess, is that hidden costs
matter a great deal and that early retirees have a lot more risk to manage than traditional retirees. For my early
retirement age, this is very, very important and probably the biggest take-away
I have gotten from this exercise. All of
this is also the rationale for me to continue to move towards low cost passive
investing in the future. I realize here that I have gotten the low cost memo a
little late and that I may be moving over-slowly…
And all that stuff above I just ran through? That's only for
today. The future is another thing altogether.
Just for fun, here is a blurb from Ken Steiner on spending:
"if you can’t increase your current spending, you might be able to increase [decrease] it in the future if the assumptions you made about the future turn out to be too conservative [aggressive]. Unless your goals include growing your assets, you need to find the appropriate balance between spending too little and not spending enough…it is important not to let fear unduly influence this task." -Ken Steiner
And Pfau:
"The unprecedented bull market of the 1990s would have allowed the 2000 retiree to finance her planned retirement expenditures using a withdrawal rate of only 2.8%, a number which could be more in line with reality for more recent retirees." Wade Pfau.
---------------------------------------------
[1] For what it's
worth, this kind of effort is maybe part of what I am starting to call my
phenomenology of retirement. That means
forget about abstract theory for a moment. What do real live retirees
experience when they interact with retirement finance? How do they really do it and how do they feel
about it? That's more or less what I am trying to get at here.
No comments:
Post a Comment