On Being Careful - Dividend Edition

"In theory there is no difference between theory and practice. In practice there is."

            -- Generally attributed to Yogi Berra

I think that really smart guys that have a little above average visibility in the financial blogosphere have a slightly higher bar than others when it comes to being careful in their commentary.  Me? I could sling whatever I want around my blog and it means almost nothing.  This thought came to me as I was listening to a blog/podcast at Meb Faber's site that consisted of a conversation between Meb (a blogger I like and respect and whose blog I read often -- co-founder and the Chief Investment Officer of Cambria Investment Management) and Larry Swedroe (another guy I like and respect -- principal and director of research for Buckingham, an independent member of the BAM Alliance and a contributor to many finance conversations, especially at ETF.com) on dividend strategies.   

What got me going was the title of the podcast: Episode #28: Larry Swedroe “There Is Literally No Logical Reason For Anyone To Have A Preference For Dividends” November 9, 2016.    My first reaction was not necessarily that I disagree, it was that as an early retiree I know that a few holes can be punched in the absolute apodictic certainty of the tone of that title. Let's try to lay out what I think I mean here.

What I Think They Got Right

This is what I think they got right or at least this is a general summary of the podcast. Some of this is taken directly from the interview, while other points are merely implicit in what they say. On the other hand, some points may be just my own imagination putting a personal filter on what is being said.  Here is a reader's digest version: 

  

• I agree with almost everything, at least at first glance.  Most of the points made, based on the mainstream financial literature out there, feels more or less unassailable.
• Based on financial theory as we know it now, dividends are evidently not a "factor" that can explain a pervasive and sustainable and differentiated source of a risk-based return premium.  Enterprise cash -- which a company can retain, declare as a dividend, or use for buybacks -- can be extracted by an investor in a simulated or self-created dividend by selling shares. This investor action makes no contribution to any reasonable explanation of risk-based returns.
• Diversified portfolios can be constructed (with instruments that pay no dividends) that outperform those that have dividend instruments.  Meb Faber apparently did a recent study along with Wesley Gray at Alpha Architect that showed that "yes, you could not only replicate dividends, but do much better than dividend stocks but with companies that pay no dividend."  See link below.
• When taxes are considered, the previous point gets stronger rather than weaker. 
• There are hard core partisans of dividends out there that will, with almost religious fervor, brook no dissent on the idea that dividend or dividend growth strategies are the way to go.  There is also a hard core cohort of investors, often retirees, that are totally committed to an income strategy that includes both fixed income and dividend paying equities. Many of these folks, according to Larry, are impervious to the logic of mainstream research on factors.
• Larry cedes that there may be behavioral explanations for a dividend preference. 
• Dividends, while not directly explaining a source of a risk-based return premium, may be directly or indirectly correlated to or point to real "factors" such as quality and value.  In particular, dividend paying instruments seem to have a value tilt.
• General financial theory and Modern Portfolio Theory (MPT), along with various current extensions of MPT into multi-factor models, all seem to point towards total-return strategies having a superior edge over other approaches including dividend strategies.  This is a little redundant but sets up one of my next points below.  

Ok.  So far so good.  I don't have a huge problem with any of that so why did I react to the title more emphatically than I might have otherwise?   It is because I think that Larry was incautious in describing "for whom" all of this might be true.  In other words, the implication was that it is true for all. 

What I Think They Missed

In order for me to answer the "for whom" question let's first put my own words into both Meb and Larry's mouths: they are leaning heavily, if not exclusively, on MPT and modern factor extensions to make their case.  That is the stepping off point for whatever counterpoint I might have.  So here, in bullet point form -- since I don't have the energy this afternoon for a more sustained and organized argument -- and including a "for whom" discussion, is a set of reactions to the title (should we call it a strawman?) “There Is Literally No Logical Reason For Anyone To Have A Preference For Dividends:”

• Harry Markowitz, the "father" of MPT, made the point back in '91 that MPT was intended for institutional investors (like mutual funds) with infinite time frames and no spending requirement and not necessarily for personal finance with it's life cycle requirements.  I riffed on this before in this post:  On Markowitz's "Individual vs. Institutional Investing."   "Infinite time frame and no spending" does not describe most investors and especially retirees and most very especially early retirees. Folks like Markowitz, Bill Sharpe, Richard Thaler and other have all in their own way emphatically made the point that personal finance is an incredibly difficult problem that financial theory helps but does not solve.  In the Faber-Swedroe podcast conversation the discussion is (in my opinion) addressing the theory and missing by a wide mark the practice of lifecycle personal finance where things are quite a bit different. 
• In the fifth bullet point above, I want to say that Larry's acknowledgement of behavioral sources for a dividend preference is correct but did not go far enough in terms of an explanation.  What he is perhaps talking about is the behavioral bias -- which I am at pains to remember exactly what it is named -- that causes us to freak out when the market is down and sell everything and when the market is high to reverse-panic and buy willy-nilly.  For retirees, having a floor of fixed income, of which dividend paying instruments is an important part, is a big salve to this kind of destructive bias.
• MPT, and the dividend concept discussed in the podcast, probably is relevant to young investors in that they are similar to institutional players by having very long timeframes and no immediate spending requirement. That means that in addition to institutional investors and investment funds, there are no doubt cohorts of retail investors for whom total-return and no dividend strategies might make a lot of sense.  For that matter, note that young investors (and counter-intuitively people very late in their retirement cycle...but that is another post) can also tolerate for the same reason extremely high equity allocations.  Retirees and especially early retirees (whom I consider to be a whole other animal) have on the other hand a strong case for both income strategies and muted equity allocations…at least for a while.
• The spending requirement missing in MPT is critical here.  I realize I have a bias toward my own cohort's needs and interests but if one were to, say, combine: a) a modest portfolio, b) relatively high consumption (spending), and c) a painfully sustained  drawdown of more than a few years (and maybe combined with high inflation like in the 70s) for someone with a 20 or 30 or even 40+ year time horizon, then that person (a retiree, esp early!) in that type of scenario is quite a bit over-exposed to "sequence of returns" risk and they could easily envision a scenario where their capital would be outlived. That certainly is not my plan.  Income strategies that include dividends are a good way to mitigate this risk while a retiree waits out whatever storm rolls over them. 
• There is a growing body of personal finance research by ____[1] that acknowledges that semi- or non-optimal strategies when viewed through an MPT lens -- things like: a) a floor of social income (e.g., Social Security), pensions, annuities, and yes dividend strategies, along with b) taxable and non taxable investment risk-portfolios (for purchasing power protection, discretionary spending needs, and importantly as a reserve for spending shocks), and then c) all of the above hedged with longevity insurance such as that provided by long-deferral income annuities -- provide a more likely than not optimal strategy for older investors and retirees when viewed through a practical life-cycle personal finance lens.  But then again my first point above was that mainstream theory is not focused on late lifecycle investors in the first place. 
• If hybrid and lifecycle-optimized (MPT non-optimal) strategies, as in the previous point, make sense at all then dividend stocks make sense to me in particular since they have a dual purpose: a) the income mitigates the behavioral and sequence risk I mentioned above, and b) the embedded equity premium protects, supposedly and hopefully, purchasing power in the face of rising prices.  Larry is 65 so he should know better.  I'm 58 and grew up in the 70s with a widowed mother that had a tiny portfolio to live off of for many many years to support 5 kids and a long retirement.  I watched (in retrospect...I don't know if I really realized it at the time) bad returns and inflation and spending blow her out of the water.  Her hedge in the end was her kids. I'm planning to avoid that route if I can arrange it.   Meb is young so he probably hasn't lived through a '70s style cycle. I'll give him the benefit of the doubt.   
• This is probably the weakest of my points but I'll try it anyway.  I view dividends as a risk mitigator in individual positions taken.  I think of this in almost a venture capital kind of way.  Over the years I have done a number of angel investments as part of an attempt at creating a coherent venture-portfolio-style strategy kind of thing.  Not surprisingly a lot of these failed.  The ones I liked the best, especially among the ones that failed, were the ones that recycled at least some cash (interest on convertible preferred or loans, royalties, etc.) before they died or while they were attempting to gain traction.  Admittedly this is not an option for many ventures in their early stages but the ones that did at least mitigated my risk a teeny tiny bit.  Getting paid -- even with all the tax inefficiencies and even with a lack of theory or research-based support for "factor" returns -- is a good thing…for me anyway.  If I were to have 30 or 40 years left, in theory I can hold ownership in something for that 30-40 year "forever" and get paid and get an awful lot of my money back while what's left in the pot goes to my emergency reserve as well as the legacy plan...a time past which I won't care so much about anything. 

So to summarize my reaction to “There Is Literally No Logical Reason For Anyone To Have A Preference For Dividends” with respect the concept of "to whom" it is likely to apply:

Institutional Investors - True
Investment Funds - True
Endowments - True
Family Offices - Who Knows or Cares
Very Young Investors - True
Mid Stage Investors - Probably True
Early Retirees - Emphatically Not True
Traditional Mid Stage Retirees - Probably Not True
Very Late-Life-cycle Retirees - Probably True, depending...

-----------------------------------------------

[1] I don't have specific citations here but think of credible sources like: Pfau, Blanchett, Finke, Milevskey, Boston College, Society of Actuaries, etc etc. I'll add links here as a postscript or a new post as I read or reread them later.

Podcast at mebfaber.com
Faber-Gray study:  How Much Are Those Dividends Costing You?
Your Complete Guide to Factor-Based Investing, Book by L Swedroe