Talking Your Book About Value (Part 3)

“A denouement is not a complete or fully resolved ending but a satisfying closure to a story.”

- To Be Told, Allender 2006

6/28/2020

The end is nigh.  As we write the final chapter in our triptych on Value, it’s worth a quick review of Parts 1 and 2.  In Part 1, we joined the discussion of the “Value of Value” and found no evidence that the relative value metric was a useful timing tool for a long Value expression, and found limited statistical support for the idea that the relative valuation of Value vs Growth provided any insight for a long/short expression.  In Part 2, we introduced the idea that the excess return associated with the Value (and Size) factors are a function of the net sale of embedded options in portfolio construction and rebalancing methodology.  So now we are positioned to address the third question:

3) Given the dynamics outlined in #1 and #2, what are the implications for investor portfolio construction?

As discussed in Part 2, we believe there is evidence for a Value factor excess return, but that this excess return is a function of a net short volatility position created via portfolio construction and rebalancing techniques.  One of the obvious implications of this association is that investors allocating to Value strategies need to be particularly sensitive to rising volatility environments, as the options that they have sold at lower volatility will mark against them.  This inconvenient truth flies in the face of “value” investors warning of bubbles and the dangers of speculative investing (to be clear, we share these general concerns, but for very different reasons).  The risk of rising volatility on value can be clearly seen in the empirical data since 2001, noting the relationship between the  Value factor and volatility.

Value Underperforms if Volatility is High or Rising

We can extend this analysis with the Ken French Data Library (1). Over time, we see a consistent pattern of the Value factor underperforming on high and rising volatility (please note we are using the Small HML factor).  This is perfectly consistent with the thesis that the “Value” excess return is driven by the sale of volatility as hypothesized in Part 2.

Unfortunately, this relationship is often obscured by the “safety” that systematic value offered in the 2000-1 DotCom Bubble; and it is to this period that proponents of the Value Factor return time and again:

We’ve seen this movie before a few times and we know how, but definitely not when, it ends. We believe that sticking with the process is the only way to achieve the long-term gains we seek (and which won’t always be provided by a long-only market that continues to levitate).

This is where long-term investors make their bones.”

                Cliff Asness, February 19,2020 (emphasis added) (2)

 

This reliance on a single major data point is troubling, not to mention statistically bankrupt. In fact, if we look at a longer history of the Value factor, as defined by Ken French in his data library, the uncertainty as to which “movie” we think we’ve seen before becomes clear.

Buying Downtrodden Equities is Not “Safe” Investing

Ironically, those who seem most devoted to “protecting” investors from mania are promoting a strategy that is virtually certain to underperform when risks rise!  The almost solitary example of the aftermath of the DotCom Bubble appears to have infected our collective psyche such that we now confuse buying downtrodden, overlevered cyclicals with “safe” investing.  No wonder the professional investor community is being trolled so successfully by Davey Day Trader Global.

A deeper analysis of this same series exposes another inconvenient fact we discussed in our whitepaper, Policy in a World of Pandemicssince 1975 it is debatable whether there is any value premium for large cap stocks.  Importantly, we see a similar period from 1926 until ~1944 when it’s arguable that a similar dry spell existed; but as this dry spell begins our database, we have no information of the Value factor performance prior to 1926.  Over the entire period, using a rolling 10-year regression (pale blue and pale orange lines above), we discover that the slope of outperformance is positive in only 56% of observed 10-year periods for Large HML and 65% of Small HML.  In other words, for any 10-year look back period there is only slightly better than coin flip odds that you observe a positive Large Value factor over the previous ten years.  With Small Value you’re basically playing “Let’s Make a Deal” with Monty Hall (and hoping there’s a car rather than a goat behind the door you pick).  To be very clear, we are not discussing the odds of any one period being negative, we are discussing the prospect of the entire 10-year period having a negative value “premium”.  Would you bet your career on a coin flip?

We can also ask ourselves a different question – “What was unique about the period from ~1944 to the early 1980s?  Obviously, this is unknowable (far too many variables), although a reasonable hypothesis can be found in observing the volatility surface.  While we do not have the history of the Value factor prior to 1926, we do have a proxy for volatility via the Dow Jones Industrial Average.  If we look at a long history of realized volatility, exactly one period sticks out for uniquely low realized volatility – the 1944 to 1975 window.

The impact of this period on the analysis of the Value factor is extraordinary.  For Large Cap HML, excluding this period changes the outcomes quite dramatically.  Without this historically atypical period, there is no Large Cap HML.

Disaggregating the datasets, and exactly as our hypothesis would lead us to expect, rising volatility is bad for “Value” and falling volatility is good.  And indeed we find that the only consistently good times to own Value is either after a volatility spike as volatility retreats or in periods of exceptionally low volatility.

With these in mind, and consistent with our original hypothesis, we are largely able to explain the large cap HML factor as a residual of the historically low volatility environment that occurred after the Great Depression and into the 1980s.  Do we have an explanation for why this volatility compression might have occurred?  Again, there are far too many variables available to establish a definitive conclusion, but it seems appropriate to note that the bookends to this low volatility period were the noteworthy Investment Company Act of 1940 and the introduction of S&P futures in 1982 (3).  It is further notable that the Small HML premium continued into the 1990s until a similar increase in volatility emerged soon after the introduction of Russell 2000 futures.  Again, the consistent result is that the transition to a higher volatilIty environment accompanied the decline in the Value factor.

It Was Different, Now It’s the Same Old

Hopefully, by now we’ve established that the Value factor is primarily a short volatility trade that increases portfolio risk.  Historical outperformance has been limited to periods of unusually low volatility or periods of rapid decline in volatility from high levels.  The immediate aftermath of the March 2020 risk selloff certainly matches this latter criteria and as we would have expected, there was a brief window of Value outperformance.  However, with volatility failing to retreat to pre-crisis levels, we have already begun to see the Value window taper off, and show signs of reversal.

The key question for the Value bulls is whether we will see volatility retreat dramatically.  Not just to the low levels of 2017, but on a permanent basis back to the highly regulated levels that prevailed in the period of Value outperformance.  At Logica, we believe this is unlikely due to issues of market structure and regulatory capture.

As we have repeatedly discussed, the widespread transition to index products (both futures and passive mutual funds/ETFs) has changed the behavior of markets.  Transactions focused on buying or selling all stocks and profitability derived from index arbitrage (again, both futures and the creation/redemption process of ETFs) rather than security selection have irrevocably changed the incentive structure on Wall Street. One of the most powerful indications of this agnostic purchase of index baskets are the dramatically higher levels of comovement of securities that we are seeing, and precisely as we would expect under these conditions.  As we noted in March, the levels of comovement now dwarf even that of the Great Depression.

Implications for Active Management

The opportunity for traditional active management to outperform, and particulary those with a focus on the Fama-French factors, is radically reduced in this environment as security selection becomes largely irrelevant, and despite claims that active managers would outperform when the torrid pace of the 2018-2020 bull market ended, we have been unsurprisingly disappointed by their performance yet again. (4)

Not to be dissuaded, proponents of Value investing are yet again calling for the Promised Land as our friends at AQR did in 2019 and again in February 2020:

“We are not into calling bottoms or factor timing. But looking at history, it appears that a golden decade for value investing might be ahead.” -- Mikhail Samonov, May 2020 (5)

 

We’re going to take the other side of this bet and again restate the obvious: at its core, by virtue of the sale of options as discussed in Part 2, and as further evidenced in this Part 3, systematic Value investing generates excess return via a  short volatility exposure.  By construction and rebalancing, it is inherently a mean reversionary strategy.  This can be seen by modeling the optimal return to a strategy that is short puts and calls, e.g. a short straddle; where maximum PNL is always generated at zero change. 

Mean reversion only works when there is a discriminating rationality that can drive outrageous valuations lower.  In a world dominated by discretionary investors governed under the regulations of the Investment Company Act of 1940, an act that recognized the national importance of professional investors placing the interests of their clients above other considerations, this rationality existed.  However, with the explosive growth of strategies that presume other investors are doing the hard work and that the only opportunity to benefit clients is to cut costs and just buy the whole basket, we have reached the inevitable conclusion that no one is standing in the way of insanity.  We are seeing this in our social lives where Cancel Culture has raised the stakes for anyone willing to stand in the way of the shaming mob, and we are seeing it in our public (and private) markets where any attempt to express rationality is met with underperformance and redemptions.

It is often stated that the most dangerous words in investing are “it’s different this time.”  We do not think it is different this time.  Instead, it was different for a period of roughly 40 years.  The Great Depression, and almost more importantly the realization of misplaced incentives in 1937-1940 (when markets again fell by over 50%), created conditions under which regulators were forced to say, “Never again.”  Competition was restricted, banks were barred from securities markets, leverage was dramatically reduced, and active trading was discouraged via high taxes on short term capital gains.  We have seen nothing in the period since the DotCom crash that suggests regulators are willing to treat the deteriorating conditions in financial markets seriously.  Instead, regulators have encouraged a process of consolidation in the name of “efficiency” that has left us with nearly unimaginable levels of systemic risk.

 

The Dash for Trash

As one might expect in an environment in which nobody is doing the work, primarily because the best strategy is to presume that others have already done it all, we are seeing a dramatic decline in corporate quality.  Per the WSJ, far more small companies are losing money.  The number of “Zombie” companies (those unable to meet debt service except via additional borrowing) is exploding.

 

 

While others claim no discernible trend in deteriorating fundamentals for “value” that might explain the underperformance, the trend is clear when genuine attempts are made to understand information below the aggregates.  True to their mission of education, our friends at Alpha Architect (www.alphaarchitect.com) were kind enough to share the following analysis with us.  If we disaggregate the S&P 1500 into Book-to-Market Deciles and track Return on Assets at the breakpoints for Value and Growth (30%ile and 70%ile), the trend in deteriorating quality for Value becomes obvious.  Note this data is trailing and does not yet reflect the unprecedented decline associated with the 2020 Coronavirus pandemic shutdowns.

In this environment, should it be any surprise that the shares of a bankrupt auto rental company have soared?  While we would like to blame the day traders, for some reason Dave Portnoy does not show up on a list of major holders.  Meanwhile, the cutting edge of asset growth remains a buyer until delisting, per the terms of the CRSP index methodology followed by Vanguard.

From CRSP (8):

DELISTINGS

A security that delists from an exchange of interest is removed from all indexes. Delistings may or may not be related to a specific corporate event.

UNRELATED TO CORPORATE ACTIONS

A company that fails to meet the exchange’s continued listing requirements (or delists voluntarily) will be removed from indexes. If the exchange’s delisting notice is confirmed by 1:30 p.m. CT on the day before the effective date of the removal (and the security is trading) the constituent will be dropped after close of that day. If not, the security will continue in indexes until it can be valued.

And this seemingly innocuous headline sits at the core of the phenomenon:

Hertz gets delisting notice from NYSE, will remain listed pending appeal

Published: June 10, 2020 at 7:29 a.m. ET

Hertz Global Holdings Inc. HTZ, -11.42% disclosed Wednesday that it received on May 26 a delisting notice from the New York Stock Exchange, which cited the car rental company's filing for bankruptcy on May 22. Hertz said it has appealed the delisting, and has requested a hearing. The stock will remain listed on the NYSE until the appeal is resolved. (9)

So how many shares did Vanguard sell into the events?  Well, none (assuming they stuck to their methodology) … instead, they likely bought shares of the bankrupt company as inflows into their funds would require them to do so.  And if the world’s largest equity holder is a buyer, you better hope you are not on the other side of a stock with 35% short interest.

Now let us flip the story and ask what happens when redemptions or rebalancing require these giants to sell?

A Conclusion is the Place Where You Got Tired of Thinking

So we conclude.  Having exhausted our minds with a deep exploration of the Value factor that brings us full circle to the raison d’etre for Logica – if everyone else wants to sell options into a market structure of rising volatility and growing uncertainty, somebody has to be the buyer.  It may as well be us.  Thank you for the opportunity.

  • Follow Wayne on Twitter
  • Follow Michael on Twitter

© 2020 by Logica

Because of the stay-at-home order and related effects of COVID-19, Logica was unable to meet the original filing deadline for its ADV, and relied on the SEC CV-19 order to file by the extended deadline.

  • LinkedIn Social Icon