I wanted to take one more look at this question of dividends
that I started to address in a previous post On Being Careful… . And it is not because I am a partisan of dividends or because I wholly buy the
party line of dividend-growth-ers. In my calmer moments I still feel that total
return is a game to aspire to especially if one has very long timeframes and
does not have a large spending requirement, requirements that if turned the
other way around might actually give me pause when contemplating concepts like
MPT and total return . I want to take a
look at this again because the "emphatic absoluteness" of the
statement by an industry expert (“there is literally no logical reason for
anyone to have a preference for dividends”) is still ringing in my ears.
It's not just that overly-self-confident and be-plumed
wrongness can sometimes irk me or that it tends to close debate and bias the discussion
in favor of the speaker, it's that it evinces a serious lack of imagination and
leaves just a few too many eddies of missed opportunity in its wake. In the case of the dividend discussion at
hand, where in the past post I leaned on generalities, here I wanted to get
more specific. Since the statement I'm reacting to was “there is literally
no logical reason for anyone to have a preference for dividends” I felt
overwhelmingly obligated to come up with at least one logical reason for a
preference. In fact, with a few minutes
of thought, I came up with five. For now.
Part of the discussion of dividends -- in the interview in
question in the previous post -- addressed the idea that a portfolio of non
dividend payers can be found that will beat dividend payers. I believe that
that is absolutely true. But for that to
be relevant to the statement "there is literally no…" or be a proof
of it, it would have to have zero selection bias and be universally and
pervasively and persistently true. I
mean I could pick myself 10 dividend payers that would beat 10 non-payers (for sure after the fact). It's all in the selection, right? I suspect there was quite a bit more rigor applied
than that to this concept by the person that made the claim, I just haven't
read his research…but you get the point. The other part of the discussion stressed
the idea that dividends have no theoretical advantage because an investor can
replicate them by selling shares. That
is quite true too, up to and not beyond a particular point. But that, in and of itself, even when true, still
does not get us to the level of "there is literally no…" because one would have to systematically identify and then
systematically eliminate all past, current, and future possible
reasons for a preference held by anyone in any situation ever. That's a
pretty big bite to chew off…if not completely impossible.
For my part, here are my thoughts on five situations where I
might actually have a preference. There
are no doubt more than that but it's all I've got right now:
1. Pure Form Doppelgangers.
If I did the math right[1], if one were to have a pure and hypothetical
instrument that retains a dividend right along side its exact doppelganger that
pays one, and then one adds a portfolio spending requirement over time, then,
all else equal, the outcome of a portfolio using the dividend retainer would
have a slight disadvantage over time to a different copy portfolio that uses
the payer…as far as I can tell. It took
me a while to figure this out since I haven't been in school for so many
decades but the portfolio that uses the dividend retainer loses a little ground
in years 2+ in proportion to the dilution of the claim on the dividend caused by
the slow bit-by bit divestiture of shares. In year 2 of n, for example, the
outcome of the two portfolios looks like it is different by D*((S2/S0)-1)
where D=dividend, S2=shares at beginning of period 2, and S0
is the beginning # shares at the start at beginning of year 1). So, if I got
this right, then it's not just that "there is literally no logical reason…"
it's that in this unique and rarefied and unrealistic case, one would always
prefer a dividend! But…these kind of
doppelgangers don't exist in nature or are rare as far as I know so maybe this
point is a tiny bit moot. That and the
fact that one could not likely build a diversified portfolio with these
wraiths. A "logical reason?" Maybe. I think so for the purposes of
this post.
2. Retirement Finance Theory. The presence of things like a spending
requirement, finite time-frames, sequence of returns uncertainty, longevity
uncertainty, legacy requirements, etc. changes the game for retirees from
standard finance theory and Modern Portfolio Theory to something else. Even Markowitz, Thaler, and Sharpe conceded
that point, each in their own way. In particular,
sequence of returns risk means that a portfolio might favor something like a
dividend, all else being equal. That
last qualifier is important, of course, but the point is that in the presence of
a bad sequence of returns, not having to liquidate shares at really bad times
has a distinct advantage. Dividends can provide some of that lift. Now it is still true that a portfolio of
non-dividend payers can be found that trounces payers even when sequence risk is realized but then the burden is on
you to both help identify the exact sub-universes of instruments you are
talking about and say by how much it out-performs and whether that out-performance
is enough to take away all of the advantage of the payer…when there is spending and
sequence risk. In practice, I think one would first
use Social Security, annuities, pensions, bonds, and part time work to mitigate
sequence risk (not to mention the psychological anxiety over total-shortfall
risk) before using a dividend paying equity but I would argue that dividend payers have a role here too especially
since they can also contribute to maintaining future purchasing power. Is this a "logical reason" for a
dividend preference? You may not be convinced but I am. At least I tried to come up with a reason.
3. Alternative/Commodity Exposure. David Swensen, of Yale Endowment fame, in his book "Pioneering
Portfolio Management" makes the
case that in managing a portfolio that has exposure to commodities -- for it's
diversifying effects and ability to nudge portfolio efficiency "up and to
the left" -- one should almost always have a logical preference for
commodity exposure that pays you rather than mere exposure to the
commodity itself. Think oil companies
rather than oil futures, for example. Similar
risk exposure and similar portfolio effects, but you get paid by one of them. I was going to try to find the page citation but
I couldn't find my copy of the book and I probably would have been too lazy to dig around
if I had. All of this may be closer to the proof-by-intimidation that I decried
earlier but when addressing the "there is literally no…" statement,
then I find that Mr. Swensen's argument does give me at least another plausible
reason for a dividend preference. I
realize that this is not exactly the retainer vs payer kind of thing, but again,
at least I made the effort to come up with "one reason" for a
dividend preference. In any case, I'll
lean on Swensen's credibility for this point. He was persuasive enough that I did what he suggested myself.
4. Trading Systems. Let's say, hypothetically, that we had a
person that works with trading systems[2].
That person knows that the trading success formula is: "Trading
Success" is a function of: a) risk-return and b) the probability of that
return. Let's say further that this
person's particular combination of solid risk-return control and pretty crummy
if not fatally flawed "probability-finding" skills will yield
somewhere around a zero expectancy inside that formula. Then, in that case, that particular person
might, in fact, have a preference for a dividend paying instrument. He or she would know that a system that can
fight consistently over to time to zero after trading costs (ignore taxes for
now) -- while it has not eliminated risk to zero it has gone pretty far down
that road -- when that system then feathers an instrument that pays a dividend
into the system's process, then that system now has an edge in the amount of
the dividend and it, the system, should probably be exploited and scaled. If, as well, the expectancy could be nudged over zero before the dividend, then even better. This is hypothetical but it would be a
demonstrably logical reason for a dividend preference. Again, this is nothing more than making an effort
to come up with at least one reason to have a dividend preference. Whether it is compelling or not depends on
the reader.
5. The Airline Rationale. Let's ignore time value and compounding for a second. If I own a $100 stock that pays 4%, I get $100 back in 25 years. If it goes bankrupt in year 26, I have at least the $100 in dividends. I realize this type of argument is not necessarily resting on good financial theory, and time value and compounding would bend the conversation, but... The "but" comes from the fact that I spent the end of the 1980s and the early 1990s doing managing consulting for the airlines. I spent a lot of time at or with Northwest, Continental, United, Air Canada, Delta, and US Air. This was a period of or just after a lot of financial instability and bankruptcies. I remember sitting in the office of the Air Canada CFO, Paul Brotto. He patiently explained to us unwashed Americans why payback metrics, the ones reviled in graduate business programs around the world, were used at the airlines. His point was that when bankruptcies loom all around you, left and right, past and future: while Net Present Value might be a really cool kid on the block, payback is cooler. Do I have a logical reason to prefer a dividend payer as a risk mitigator? Maybe you don't but I just might.
None of the above is an attempt to make the case that
dividends are perfect or better than anything else in all situations[3]. In the real hurly-burly of the world it
depends on who and when and where and how they are used. My point is that I think there are enough
real or imagined reasons for special cases where there actually and really is a
preference for a dividend that it is not logical at all to say that “there is
literally no logical reason for anyone to have a preference for dividends.” Me? At least I tried.
--------------------------------------------
[1] I'm not so sure I got it right here. I'm having a CFA take a look but that is
pending. While I'm waiting, if you happen to know this stuff and want to beat this
point down, send me a note and I'll post it as a guest blogger post.
[2] A person that might look uncomfortably like me.
[3] Buffet, btw, doesn't like to pay dividends but he is
also an institutional investor with infinite time frames. I think if he were honest with himself and contemplated
the universe from the point of view of a retiree with finite money, uncertain
longevity, spending needs, and unpredictable return series he might
reconsider. I'll end there because in
any article I read, when the name Buffet is used, I usually stop reading
because his name is so often invoked for ill-informed or even insidious reasons
that informed and open minded discussion usually ends there as well.
No comments:
Post a Comment