August 2021
Shhhhhh-orting
A new normal, P/E edition
”Natural sellers” vs “Short sellers”
The September Effect
“Sell in May, then go away”
“The trend is your friend”
These and countless other adages linger in the minds of investors seeking insights about future market returns. Although not a snappy phrase, the September Effect is likewise in the forefront of many investors’ minds. Particularly given this year’s steady progression upward that has not produced even a 5% pullback in the S&P 500, market participants seem to be concerned about an imminent correction and that the September Effect may be the willing cause.
September does generally have the worst monthly stock market returns:
Source: https://stockanalysis.com/average-monthly-stock-returns/
Past performance is not indicative of future results
But as any statistician will tell you, causation is a tricky business. Explanations for the September Effect include:
Investors are eager to take seasonal profits and harvest losses going into the end of the year.
Parent investors need to raise cash to pay for children’s school supplies.
Investors return from their summer vacations and are ready to make some moves in their portfolios.
None of these are particularly satisfying with respect to rigorous statistical analyses. I’m not convinced that investors disconnect from managing their portfolios during the summer months, particularly now that the frictions of trading have been reduced to clicks in an app. But none of that matters if a sufficiently large number of market participants believe in these explanations; a self-fulfilling prophecy can still move markets no matter how seemingly arbitrary the explanations may be.
Technical analysis is home to many such dubious indicators. My favorite “non-indicator” is the Fibonacci Retracement. With several particularly nice occurrences in nature, the Fibonacci Sequence has a sense of explanatory insights that some folks extend well beyond the logical domain of the natural world. The Fibonacci ratios of 23.6%, 38.2%, 50%, 61.8%, and 78.6% often grace many charts in hopes of finding predictive power for pivot points in price retracements. Is there really some natural force tugging on market prices in accordance with Fibonacci ratios? Or might any set of retracement ratios be “right” occasionally enough to capture the attentions of eager chartists? The power of rationalization cannot be understated.
In any case, what arguably has a more convincing explanation is that highly liquid and efficient markets will correct for any such influences over market prices. If enough investors believe in the September Effect, the game-theoretic strategy is to sell in August ahead of the anticipated September selloff and pull the September Effect ahead into August. Well then there is an “August Effect” for which the game-theoretic strategy is to sell in July. With diminishing effects in each month, at some point any actual September Effect is absorbed by the magic of efficient markets, and nothing actionable remains for investors to capture an edge.
Alright, enough already. While folks weigh and strategize on how to position themselves for the September Effect, I am focused on evaluating the unique market dynamics of the moment, guided in particular by the unprecedented pandemic and responses by policymakers, companies, and consumers.
Monte Carlo confidence
36.3% of all statistics are made up.
Okay, I made up that statistic. If you believed me though, even for just a split second, your instinct is natural - we are biased toward giving credibility to measures with numeric precision. Contrast that with the deceptive (and fictitious) marketing campaign that said “75% of dentists recommend our toothpaste (the 4th dentist couldn’t be bought).” 75% is just too clean of a number to be as convincing as 36.3%.
Now consider outputs from computer simulations. Surely those simulations will produce convincing and profitable insights! Folks will use Monte Carlo simulations to analyze possible outcomes, understand variability, and try to reduce uncertainty in a particular setting. But like the statistics in the previous paragraph, failing to understand the methodology behind the simulations will Likely lead to harmful misuses of the outputs.
Here’s a well-intentioned article from Toward Data Science, How to Use Monte Carlo Simulation to Help Decision Making:
Let’s say you are dealing with various job offers. The challenge in making a decision is that there always is some uncertainty. For instance, you do not always know exactly what you will be doing, who will be your colleagues, how this job will help in your career advancement, etc.
That’s where Monte Carlo comes in. Here are the three steps to follow to set up the experiment.
Define all the variables that go into making a decision (salary, conditions, location, work-life balance, etc.)
Define the importance (weight) of each variable. In other words, what is most important to you?
For each option, define, for every variable, a range of possibility on whatever scale you feel is appropriate (out of 10 in this example). Of course, the wider the range, the more uncertain it means you are with the grade you have given to this variable! The range represents the probability distribution of the variable.
I appreciate the interest in trying to reduce subjective “meaning of life” types of matters into objective and quantifiable metrics. But think about all the assumptions being made in these three steps! The extent to which one can even meaningfully and accurately quantify the weights associated with their optimal work-life balance is bound to make any simulation’s output questionable. If you don’t get the outcome you expected, would you accept that output, or would you reevaluate the parameters you entered into the model? The spice of life makes life meaningful, and quantifying the unquantifiable is often useless. Garbage in, garbage out.
The same can be said for many investing models. Significant assumptions are made about the applicability of the prior inputs with respect to the predictive power of future outputs. COVID-19 presents a perfect example. I wrote this in November about models:
Forecasting the future, quantifying uncertainty, making sense of data - investment math is not all that different from election math. The accuracy of investing models during this pandemic is perhaps universally in-question due to the unique nature of both the COVID-19 pandemic as well as the government’s unprecedented response. Ray Dalio, continuing to endure historic losses in Bridgewater’s flagship Pure Alpha II fund, finally ceded his usual steadfast confidence in his models when they stopped working and spent 70 hours per week along with his staff to revise their models. It’s what modelers do.
Managers trying to use pre-pandemic models to gain predictive insights about market returns during an unprecedented, once-in-a-lifetime pandemic are bound to be misled. The times when models are most predictive are usually when models are not needed, and accordingly the times when models are least predictive are usually when models are most demanded.
He who controls the inputs
Here’s Federal Reserve Chair Jerome Powell, speaking at the virtual Jackson Hole Economic Symposium:
“Longer-term inflation expectations have moved much less than actual inflation or near-term expectations, suggesting that households, businesses, and market participants also believe that current high inflation readings are likely to prove transitory and that, in any case, the Fed will keep inflation close to our 2 percent objective over time.”
For all the debate among commentators over whether inflation is transitory or structural… although the debate is interesting, the debate also misses the mark. We are still emerging from the economic impacts of a global pandemic. We still have enhanced unemployment benefits, vaccination challenges, supply chain disruptions, demand uncertainties, fiscal and monetary stimulus, and myriad other factors that are impacting inflation. Eventually the economy will normalize, eventually inflation will normalize, eventually labor markets and supply chains and the Fed’s balance sheet and everything else will normalize, each to more-or-less pre-pandemic levels.
Investors with robust long-term strategies can endure and rise above short-term inflationary pressures. I doubt those who are sincerely arguing the structural inflation case actually think today’s inflationary trends will persist for, say, 5 years and beyond. They are arguing that the current inflationary trends will persist for longer than investors and policymakers expect and that the Fed will have to raise rates sooner than expected. Investors with short-term strategies concern themselves with whether these inflation trends will persist for more than 6 months or 1 year or whatnot. With the Fed’s current outlook on raising rates being on the timeframe of 1-2 years, and sooner if economic conditions warrant, this short timeframe is much too short to support claims of long-term structural inflation. The issue is not actually about transitory inflation versus structural inflation, but rather about the duration of inflation and whether consumers can tolerate those inflationary pressures throughout that duration.
Until further notice, I am with the guy who holds inflation’s policy toolbelt.
Virtual Reality
Did you see that immersive virtual reality headsets are no longer only the realm of Beat Sabering gamers? How virtual reality will change trading for pros and everyone else.
A VR day trading headset seems perfect for Stephen Kalayjian, co-founder of Ticker Tocker and commandeer of perhaps the most immersive trading room in the business:
Source: Ticker Tocker Trading, https://www.facebook.com/TickerTockerTrading
Day traders have a natural inclination toward leveraging technology to optimize their trading. But I can promise this: humans will always lag the data synthesizing capabilities of the fastest computers, most efficient algorithms, and smartest AI technologies.
Whether it’s high-frequency trading algorithms, robo advisers, AI-powered trading bots, or yet-to-be-developed technologies, where can humans outperform computers?
Garry Kasparov, former World Chess Champion and last human to be able to beat the best chess computer-of-the-day, is uniquely qualified to answer where humans can outperform computers. As World Chess Champion at the time when IBM built Deep Blue to play chess at a level that exceeded the chess capabilities of any of its developers, Kasparov was called upon to represent all of humanity in “The Brain’s Last Stand” against the machine. Would people still want to play chess, would anyone care about being the best chess player, and would the game even be interesting in a world where computers could master the game beyond any human’s capability?
Kasparov shares his insightful answers to these and other questions in his popular TED talk Don't fear intelligent machines. Work with them:
Doomsayers predicted that nobody would touch the game that could be conquered by the machine, and they were wrong, proven wrong, but doomsaying has always been a popular pastime when it comes to technology.
What I learned from my own experience is that we must face our fears if we want to get the most out of our technology, and we must conquer those fears if we want to get the best out of our humanity. While licking my wounds, I got a lot of inspiration from my battles against Deep Blue. As the old Russian saying goes, if you can't beat them, join them. Then I thought, what if I could play with a computer -- together with a computer at my side, combining our strengths, human intuition plus machine's calculation, human strategy, machine tactics, human experience, machine's memory. Could it be the perfect game ever played?
My idea came to life in 1998 under the name of Advanced Chess when I played this human-plus-machine competition against another elite player. But in this first experiment, we both failed to combine human and machine skills effectively. Advanced Chess found its home on the internet, and in 2005, a so-called freestyle chess tournament produced a revelation. A team of grandmasters and top machines participated, but the winners were not grandmasters, not a supercomputer. The winners were a pair of amateur American chess players operating three ordinary PCs at the same time. Their skill of coaching their machines effectively counteracted the superior chess knowledge of their grandmaster opponents and much greater computational power of others. And I reached this formulation. A weak human player plus a machine plus a better process is superior to a very powerful machine alone, but more remarkably, is superior to a strong human player plus machine and an inferior process. This convinced me that we would need better interfaces to help us coach our machines towards more useful intelligence.
As high-speed algorithms identify and arbitrage away market inefficiencies in the time it takes humans to blink an eye, it is futile for human traders to try to compete in this race. Humans can add value not by trying to synthesize intraday technicals faster than computers, but rather by creatively melding the data-heavy pattern recognition strengths of algorithms with a robust long-term investment process that attends to the various high-level complexities of investing, including authentic risk management, and subjective fundamental analyses, to name a few.
Sore-os
The race to fill the top of users’ timelines are, arguably, a consequence of our increasingly short attention spans. As attentions flip to the next short-lived headline, we too easily neglect to remember and learn from the outcomes of prior newsworthy events.
Seemingly forever ago - in May(!) - I wrote about the opportunistic team at Soros who bought several Archegos stocks after their collapse, including ViacomCBS, Discovery, and Baidu. Value can be found in “divorce”-induced fire sales, but as I wrote in May, these shares may not be on a fire sale as much as they are coming back down to reality from artificially inflated prices:
Who knows what will happen, especially given the new world of meme stocks, diamond hands, and rockets, gosh stock picking is a rough way to invest. We will find out soon enough, so set your calendar reminders to check these share prices in a few months. But given that Archegos’s trading volume was a significant explanatory factor in the magnitude of share price increases, one might reasonably expect the post-implosion prices to settle back toward pre-implosion prices, perhaps prior to the Archegos-induced price increases from roughly January through mid-March. If that’s the case, Soros Fund Management might not get the post-Divorce flippable recovery that they were expecting.
Well, here you go:
Closing prices by date
Ooof.
ONE MORE THING…
Shhhhhh-orting: Hedge funds really need to learn how not to attract attention to their short positions. Heretofore, I dub this "shhhhhh-orting". AMC Stock Rises After Report of Hedge Fund Going Short
A new normal, P/E edition: "the democratization of investing brought about by the internet has expanded P/E multiples considerably over the past quarter of a century. With information accessible at the click of a button, rumors no longer weigh down valuations, as they had in the past. Thus, higher Shiller P/E values might become the norm." History Suggests a Stock Market Crash Is Likely: 5 Data Points of Concern
“Natural sellers” vs “Short sellers”: Likely scared off by wallstreetbets, and as I and others expected, short sellers have become gun shy about taking new short positions. This has skewed the normal price discovery function of short sellers and could explain the march upward without a 5% pullback year-to-date. Short-sellers have been the 'secret ingredient' to the market's Covid rally, says Jim Cramer
The information and opinions contained in this newsletter are for background and informational/educational purposes only. The information herein is not personalized investment advice nor an investment recommendation on the part of Likely Capital Management, LLC (“Likely Capital”). No portion of the commentary included herein is to be construed as an offer or a solicitation to effect any transaction in securities. No representation, warranty, or undertaking, express or implied, is given as to the accuracy or completeness of the information or opinions contained herein, and no liability is accepted as to the accuracy or completeness of any such information or opinions.
Past performance is not indicative of future performance. There can be no assurance that any investment described herein will replicate its past performance or achieve its current objectives.
Copyright in this newsletter is owned by Likely Capital unless otherwise indicated. The unauthorized use of any material herein may violate numerous statutes, regulations and laws, including, but not limited to, copyright or trademark laws.
Any third-party web sites (“Linked Sites”) or services linked to by this newsletter are not under our control, and therefore we take no responsibility for the Linked Site’s content. The inclusion of any Linked Site does not imply endorsement by Likely Capital of the Linked Site. Use of any such Linked Site is at the user’s own risk.