Was “Value” Just a Hot Hand Thing?

In 1861, at the outset of the Civil War in the United States, Union Pacific issued 20 shares in its IPO[1].  A year later, under the Pacific Railway Act of 1862, the railroad company received a charter from Abraham Lincoln[2] to build a railway west of the Mississippi, connecting with the Central Pacific Railroad, who were building eastward from Sacramento.  

Just over 20 years later, Dow Jones created its first stock market index.  It consisted of 11 transportation stocks, nine of them railroads, one of them being Union Pacific (UPN).  Of those initial Dow stocks in 1884, UPN, a builder of the transcontinental railroad, is the only one still listed[3] today.

Like the other securities in the index, United Pacific paid dividends, and that was what attracted potential investors.

From that point forward, Mr. Market demanded that the dividend return on stocks (in general) was higher than the return on bonds.  This made sense to everyone, because stocks were “riskier” than bonds.  Sure, there were periods where dividend yields would temporarily butt down against bond yields, or even move temporarily lower, but they always drifted back up to where they were “supposed” to be in comparison to bond yields.

Then 1958 happened.

That year, dividend yields fell below bond yields, and they stayed there for a while.  The old timers harped on about the silliness of it all, and sat back in their rocking chairs, mocking these new era investors.  They knew full well the way the world worked, and they had seen this all before.  They were almost anticipating their own schadenfreude.  

But they were wrong.

The world changed.

Dividend yields stayed generally below bond yields, forever.  

Or at least they have until now.[4]

In Against the Gods[5], Peter Bernstein describes the shift:

“…that something so unthinkable could occur has had a lasting impact on my view of life and on investing in particular. It continues to color my attitude toward the future and has left me skeptical about the wisdom of extrapolating from the past.”

This is not a post about why that happened.  There have been countless papers on this dynamic, and we link to a few of them here[6], here[7], and here[8].  This instead is an attempt of a value investor to re-examine his beliefs about the way the world works.  This is a value investor, wondering, frankly, if the value premium simply existed because it existed.  Was the value premium merely an outcome of anomaly-dredging and data mining, or was/is there an actual reason why it worked?

In other words, was the observation that “value” outperformed growth just a construct of backtesting over decades (and countries) and seeing that it worked without motivation?  Were there omitted variables perhaps driving the relationship?  Or was value just some globally-intertwined roll of the dice that kept on hitting 7 or 11 more often than crapping out, for decades?  

The folks at Research Affiliates asked this question too, in their “Reports of Value’s Death May Be Greatly Exaggerated[9]” piece.  Within, they state “because value investing has a long history of strong performance, easily traced back at least to the 1930s and solid economic footing, the authors believe that the data-mining story is an unlikely contributor to value’s recent travails.”

But what if they are wrong?  What if value just had a very hot hand until about a decade ago?  Or, like the dividend yield paradigm shift in 1958, what if there was a shift for value in 2007?

To answer the question, we must get to the crux of why value was supposed to have worked in the first place.  Sure, some quants out there may become satisfied with simple backtesting and results without theory (see the ridiculously-overpopulated factor zoo[10]), but that is where the trouble starts.  We need a theory, some theory, or it is all nonsense and ultimately unrepeatable.  

So, why is value supposed to work?  Well, there are two differing explanations.  They are opposing views that in fact form the core of the schism in academic finance that started back in the 1980s and 1990s.  Those divergent explanations, among other reasons, motivated the rise of behavioral finance.  

The efficient market guys will say that value worked (or indeed works) because value stocks are riskier. The behavioral guys will say that value worked (or works) because people are emotional, and occasionally, downright silly.

The efficient market guys will cite Bob Merton’s paper on the intertemporal capital asset model, and state that value – just like the market itself – is a “state variable of hedging concern”.  In other words, the risk of holding value stocks is the kind of non-idiosyncratic thing that impacts investors systematically, and if people don’t want that risk (especially if the value of their human capital is falling at the same time their value stocks are…) then they can shed it.  If they do accept the risk, they should be compensated for holding it.

The behavioral guys think that is all nonsense, and will cite the LSV[11] paper or one of the dozens of others that followed about how investors will naively extrapolate recent earnings growth too far in the future, bidding up the price of “glamour” stocks.  Simultaneously, these same investors will overreact to bad news and underprice “value” stocks.

This is the question that has remained open for 30 years, and the pandemic isn’t proving either of them wrong.

The efficient market guys can say:

“Yep, this is the risk we were talking about.  The threat of these rare but severe downdrafts are the reason that these heavily-levered, deeply-cyclical, economically-sensitive, value stocks will outperform over time, because they need to compensate investors for this kind of volatility when it actually shows up.”

And the behavioral guys can say:

“Yeah, right, do you see how much these idiots are paying for Tesla and Shopify?  This is what happens.  People start believing in a “new era” narrative and that no price is too high to pay for glamorous stories.  Meanwhile, value stocks are discarded as fundamentally and permanently disadvantaged, and this provides the very setup that drives the value premium over time.”

Back during the period from 1998-2000, you could have applied either of these arguments as well; although (from our perspective, perhaps naturally) the behavioral stance was stronger. But the point is that there was an ex-ante explanation for the tech bubble; and if we are in a similar environment today, we have competing reasons explaining it too.

And, despite the TSLA and SHOP references above, we aren’t talking about individual stories.  We are talking about a market-wide phenomenon.  In fact, it is our view that (at least a very) few of the stocks which are assigned to the growth bucket, aren’t overpriced glamorous shares. As we wrote in “Secular Winners and Value Investing[12]”, if we had a crystal ball in 2007, and knew the numbers that Amazon would print 10-12 years later, it would have been deep, deep value.  This, from our 2017 post:

“Amazon not only continued to surprise the consensus investor for the ensuing ten years, they even continued surprising the bulls.  That’s why the stock moved higher.  Sure, we can have the debate today about whether or not people should pay 15x or 60x for forward earnings, or wherever in between, but that is a second order argument.  The primary reason that Amazon, Facebook, Netflix and the rest of them have done well is because their fundamentals have not only done well too, but done better than people expected.”

But here in 2020, the growth investor has to be careful using individual winners to justify ownership in names they believe to be tomorrow’s winners.  Looking backwards, yes, Amazon was 100% a winner.  But looking forward, are any of these non-FAMANG names winners with 100% certainty?    

As we recently wrote in “The Times that Try Stock-Pickers’ Souls[13]”:

“It’s very convenient to draw parallels between past winners and newer companies as if it is a foregone conclusion they too will win in similar fashion.  Not everyone can be Amazon, in fact, no one else may ever be Amazon.”

And even if we do have 100% certainty about someone being a winner, there still is a price for everything.  Just as there is a right price for the game-changing new-era winner, there is a price for the downtrodden old-school loser.

And it is those downtrodden losers that form the other half of the value factor.  The unglamorous half.  The universally ugly half.  The half that has been thoroughly dominated by growth over the past dozen years, and in spectacular fashion over the past two.

We can talk all day about parallels to the tech bubble.  Some of the arguments are malarkey, and some might make some sense.  But it is a fact that the big winner mentioned above, the one with the $1.6 trillion market cap today, even got ahead of itself during the bubble.  In the 21 months from December of 1999 to September of 2001, the Amazon share price dropped 95%.  

So even “winners” can get crushed.

And back to the issue at hand, there actually was a good explanation why dividend yields moved below bond yields.  Here’s ours:

“At about the same time that the bond market entered a steep bear market, stock market volatility dampened relative to bond market volatility. Meanwhile, and perhaps there is some reflexivity here, corporate managers started thinking about capital allocation more efficiently, and thus dividend payouts fell (as these managers invested in high-return projects) and in turn capital appreciation became the key motivator for stock ownership (and performance).  All of these factors led to dividend yields falling below bond yields.”  

And maybe today with dividend yields back below bond yields, it all makes perfect sense.  As interest rates head to zero, the bond market gets stronger, the stock market again gets more volatile than the bond market, and corporate managers can’t find a more efficient use of cash than to buy back shares.  

And we have a good explanation why value isn’t dead either.  For the efficient market guys, the risk just showed up that drives the future premium, and for the behavioral guys, this is now exhibit A for what happens when people over-extrapolate the positives and negatives.  Exhibit A used to be the tech bubble, but it’s been relegated.

The paradigm hasn’t shifted.  Value’s still fine folks.[14]