Winkler's Wager

Let me state upfront I know that for the most part (if at all) I am not breaking new ground here. 

Are we all basically agnostic? Or all basically believers? How much of the rejection of belief (disbelief) is just a rejection of the behavior and style and beliefs of individuals or institutions the nonbeliever (believer) finds amiss or reprehensible or simply unconvincing? 

Years ago in thinking about this topic and in preparation for this blog post specifically I polled two friends. These are each very thoughtful, highly intelligent people. One is correctly described as a strong believer in God. One is correctly described as a strong disbeliever. Separately I asked them simply "What is the probability God exists?" leaving it fairly open ended for their own interpretation. Both of these people know how probability works and why 100% and 0% are bad answers. 

The believer stated that he wanted to say 100%, knew that was technically impossible, knew that faith might be a reason to actually make it legitimate, but settled on 90% (all of this recollection conditional on if memory serves; it was 5 years ago). He gave a good explanation for his thinking to support the answer.

The disbeliever answered via email, so I have his response. After sleeping on it, he answered 20% with a thorough account of his reasoning. 

I don't want to make this about their specific answers. This was just an experiment regarding my prediction about what they would say and why they would say it. Why I completely understood what they said and why they said it (it basically matched my prediction as well), I do not feel fully compelled by either. 

Similarly, I ran this twitter poll recently:

Probability a supernatural being(s) (aka, God) exists?

— Steve Winkler (@swinkler78) March 6, 2021

Obviously, this was not a meaningful sample size. But that isn't the point as much as the split among the choices I presented gives some indication that I think resembles how people tend to think about this.

Faith = ???... belief in the face of doubt? That definition would imply that 100% and 0% are not legitimate answers. Doubt seems essential for faith to have meaning. And the existence of doubt pushes one toward the unsatisfactory middle point of 50/50.

I think this is more easily seen in the case of a believer. But it is true at the antithesis as well. Atheism (certainty there is no God) is a faith by the atheists' own definition--one cannot prove a negative (e.g., there is no God); therefore, atheism cannot be scientifically proven. 

Ask a believer and a disbeliever this question: What it would take for them to reverse their view? This leads me to believe all in this debate are "believers" ultimately. And yes I know the problems with this over simplicity.

To the point of many in the atheist community, a point Penn Jillette makes in this piece, no one is really agnostic. A person always will find a way to dismiss evidence or argument offered against the view they hold in their heart of hearts. 

I think this gets to the crux of the question. The right answer is perhaps +/-50% with faith in God or faith in not God (something beyond the material realm) pushing one off of this center point of pure agnosticism toward one of the two faiths. The existence of God is a non-falsifiable conjecture; therefore, using science or reasoning to "prove" either the existence or the nonexistence of God is futile and fallacious thinking. 

Can we at least point to arguments to guide our judgments on God's existence? It would seem this is quite hard beyond simply an exercise in persuasion for those already tempted to be on the same side of the argument--we can never change the minds of those on the other side. Yet, minds do change and in both directions. The links in the P.S. sections allude to this.

So much of this ageless debate is people talking past one another. Adjacent to this is the determinism versus free will debate. Usually there is confusion on the part of those arguing for free will between determinism and fatalism, and usually there is confusion on the part of those arguing for determinism between free will and randomness. 

Sam Harris makes a strong case for determinism but only on the back of a reductivism I don't think can be denied--yes, there are always causes . . . it is cause and effect all the way down. Yet this basically amounts to a tautology that avoids the important parts of the question. Can we hold ourselves and others responsible for actions taken? What does it mean to choose? To act? To fail to act? 

I am a dualist on the issue. When I play pool, my choice of where to aim the cue ball and how to hit it are my free will, the resulting actions of the balls on the table are pure determinism. The determinist would entreat, "Is it not just a higher order of underlying causes that lead you to 'chose freely' how to strike the cue'?" My answer is "Yes, of course, and that isn't interesting for the matter at hand." Daniel Dennett says it better. 

Similarly, believers in God and disbelievers in God tend to talk past one another. They mischaracterize the other side's position and misunderstand what the other side means. This is not helped by how poorly the believers tend to understand their own position or how dismissively the disbelievers tend to assume past the implications of their own position. Believers wish to put God in a box and disbelievers live out the story of the Apostle Thomas. Taken to logical ends most believers' understanding of God can be disproven and most disbelievers' reasoning forces them to reject all knowledge and facts about the world. Experiments to this end: 

1. Ask a believer to convince you that their belief is genuine as opposed to something that makes them feel good.
2. Ask a disbeliever to explain why their expectation that their car will get them to work tomorrow morning is not predicated on faith or many small faiths they themselves cannot prove. 

Each will often struggle: In the first case because it is hard for a believer to identify a reason for faith beyond a desire for faith; In the second case because most people do not know how an automobile works and the existence of the future is continually theoretical--I'll prove tomorrow exists . . . tomorrow.

We should not depend on the ill-equipped be the strawmen foils for our favored positions. 

Consider my work as a practitioner within the investment management profession. My most sophisticated client would find the way I explain my job to my young children and the way my young children understand what I do to be quite unimpressive and perhaps even unattractive. That doesn't invalidate the philosophies I hold or the method I employ or the track record I've achieved professionally. Likewise part of what I do and any successes or failures associated with it might be simply due to luck. My efforts and explanations are at least to some degree counterproductive, irrelevant, and orthogonal to their associated outcomes. My shortcomings, imperfections, and activities themselves within financial money management neither prove nor deny the existence of financial money management.

Perhaps the most challenging part for believers is to separate God from being simply a personification of truth, love, and perfection. 

Perhaps the most challenging part for disbelievers is to build a foundation of truth (moral, mathematical, and physical) without the identification of God as this foundation. 

Both sides accuse the other of shortcuts for the sake of certitude. Both sides make the mistake of looking to religious texts as scientific works. If you do this, you are gravely missing the point. Newtonian/Einsteinian physics can't speak to ethics or morality. Likewise, the Bible, et al. are not going to serve your quest for scientific truth. 

Along the journey of building this post over the past few years these fellow travelers were helpful: 

Tyler Cowen considered similar questions, Ross Douthat weighed in quite awesomely, and Think Again – a Big Think Podcast (125. Reza Aslan (author) – Deus Ex Hominem. 

Related: A little footnote on the population of "atheists", What is Robin Hanson's kind of atheist?, and Philosopher John Gray discussed the Seven Types of Atheism with Russ Roberts on EconTalk. 

Here is another very good, related conversation.

From this comes this insightful item: “Religion has nothing to fear from science, and science need not be afraid of religion. Religion claims to interpret the word of God, and science to reveal the laws of God. The interpreters may blunder, but truths are immutable, eternal and never in conflict.” If I could be so bold, I would like to add a corollary. Faith and religion are very poor at discovering and developing legislation to govern society and scientific facts to explain the universe and world around us. Faith and religion are very well-suited for the discovery of righteous first principles and guideposts for how to love and live among one another. Likewise science cannot teach us right from wrong but can teach us true from false.

P.S. Is God math?

And

https://fakenous.net/?p=1863

https://fakenous.net/?p=1649

https://fakenous.net/?p=1909

https://fakenous.net/?p=1929

https://fakenous.net/?p=1937

P.P.S. How should a Christian Bayesian react to the Mayans, et al? Should they heavily discount the evidence and slightly shift their prior or slightly discount the evidence and heavily shift their prior? Is this question a risk of confirmation bias?

P.P.P.S. I avoided the heresy of an adjacent issue: that perhaps believers of all types (Christians, Muslims, Buddhists, ... Mathematicians?, Universists? ("let the Universe decide..."), et al.) are all yearning and seeking to follow the same ultimate God. 

A Paradox of Choice

When well-meaning fools lament the plethora of choices we have in a free market society, they make a crucial mistake. They ignore the necessary condition for which those choices emerge and simply assume a result. 

I can imagine a very well constructed study done that proves, truly proves in a way that most all could agree, that ~99% of why a consumer chooses A versus B comes down to irrational affiliations or random luck. For example the determination of what car one chooses to buy between a Honda Accord and a very similar Toyota Camry would be shown to be nothing more than stuff like a person's prior cars happened to be Hondas or their parents bought Hondas or they buy the car they last saw an ad for or their favorite color is red and the car the dealer first showed them was red, etc.

The incorrect conclusion a well-meaning fool might jump to is that we have unnecessary redundancy in our free market economy. We should stop wasting resources on the competition because the differences between these artificial choices are illusory. Just build the Acamry and be done with it!

The fact that there are minor, subtle, and hidden differences between all such choices are lost on those drawing such conclusions. But more importantly they are making a dangerous mistake to neglect how we got to the magnificent state of the world where we have such amazing choices that a purchase decision settled by a coin flip likely leaves the consumer equally well off. 

There are no instructions from the universe for how to build the optimal mass-consumer sedan. We need a process to discover this and continually rediscover and improve upon the outcome. We need the right incentives to reward incremental success, punish incremental failure, and fully stop efforts going down a bad path. The capitalist free market is this process. It is the sine qua non for progress--as broadly defined as one can imagine it.

PS. Or maybe this is optimal?

A Greater Sage Theory

Just a few wondering thoughts on the latest techno-investing development--non-fungible tokens or NFTs.

What gives a collectible object value? Are NFTs like Beanie Babies or Picasso paintings? 

Think of this as a spectrum between pure speculation and pure intrinsic value. An object never lies entirely on one end or the other of this dimension. Where it resides is also not necessarily stable.

Fine art is "fine" in that it has a low degree of speculation relative to perceived intrinsic value. 

Gold is the ultimate financial consensual hallucination – – we can easily, reliably believe that it will have value across societies and well into the future. It is much more difficult to believe that Beanie Babies will have that quality. Picasso paintings are somewhere in between.

Scarcity is an important quality for determining marginal value, but it doesn't say much of anything about intrinsic value. This is the crux of the diamond-water paradox. I think there are two important subtypes of scarcity as it relates to collectibles: organic and manufactured.

Organic scarcity is producing 10,000 Babe Ruth cards and only 1,000 survive decades into the future. Manufactured scarcity is knowing that 10,000 Derek Jeter cards are desired but only producing 1,000 of them. 

Organic scarcity might be thought of as "authentic", but that too is in the eye of the beholder. The 1,000 Babe Ruth cards aren't any rarer in the example above given those parameters.

NFTs are a manufactured scarcity. However that is not very important except to the extent that someone values genuine authentic scarcity--the organic kind--as opposed to fabricated scarcity. Yet I can easily see an appreciation for the manufactured scarcity nature of NFTs. So don't be too quick to dismiss the limited denominator as a factor for these collectibles.

To the extent you can believe that people will continue to value the interestingness of NFTs along with the thing (art work, sports moment, etc.) that a particular NFT is associated with, you can credibly and reliably believe that that specific NFT will have value. 

At this point they are obviously deep on the speculation side of the speculation vs intrinsic value spectrum. Time will tell.

My Friends Are All Wrong For Different Reasons

I have always prided myself on the diversity of friends I have been blessed with. This goes along with an ability to be a welcomed part of many different social groups. (At least I think I have been a welcomed part . . .)

To take one example, senior year of high school my first class was Leadership with all the cool kids--the ones who were popular enough to be class officers, etc. The bulk of the day was honors classes with the smart kids--"nerds" was strictly a putdown back then. The last class was gym with all the others--misfits who in many cases lived in fringe realms and in some cases it was amazing they hadn't dropped out of school already. I don't know of anyone in my class of 600+ who could actually approach my level of diverse and deep integration. 

At my best I am a social chameleon. At my worst I'm a jingle you can't get out of your head.

Among other things, this quality has given me a chance to learn from lots of different people as well as compare key difference among groups. Painting with a broad brush, here is one thing I notice.

My friends on the right tend to naively believe the rhetoric of the politicians they admire. They think the things they hear are sincerely believed and will be sincerely pursued. 

My friends on the left have the opposite problem. They tend to naively believe that the rhetoric espoused by the politicians they admire should be ignored—that the true pursuits will be reasonable, noble, and unarguably worthwhile. 

Both can be willfully blind to the most grotesque, pandering, and ridiculous rhetoric. 

I admit this is an overgeneralization to make a point, but it is still a generalization based in truth. While there are many exceptions to this rule, in the cases where it holds I don't find the friends usually to be in contempt. They are almost always unintentionally wrong rather than willfully guilty. Political cognitive dissonance is a very real and rationally held phenomenon. 

My ultimate takeaways:

• Looking inward first, I should be on guard for both of these problems in my own thinking and behavior. 
• To change minds on the right where I agree with a friend on the policy desire, I should emphasize how politicians and political solutions fail to pursue our common cause.
• To change minds on the left where I agree with a friend on the policy desire, I should emphasize how politicians and political solutions work against our common cause.

Changing minds on the right or left where I disagree with a friend on the policy desire is much more daunting. It probably starts by inquiring as to why they want what they want in the first place. From here it is very often the case that we have the same desired ultimate outcome. It is just that we disagree about how to achieve it. Without first finding common ground, a discussion inevitably becomes an emotional argument rather than a fruitful search using logical arguments.* 

For example, trying to change a paternalist's mind on the drug war cannot constructively start until he understands that you are not advocating unconditional, rampant drug use. He will imagine chaos on the streets and label you extremely dangerous. He has been conditioned to believe that opposition to drugs because they are bad = support for the drug war. 

Likewise, you cannot begin to alter the position of a supporter of the minimum wage until you firmly establish the common ground that you both want the best for low-wage workers.

*People have always misunderstood my love of argument thinking that I love to be in arguments. 

A Hypothetical SARS-CoV-2 Study

Perhaps this already exists, but I doubt it. At least I do not expect this ambitious of a study has been attempted at the rigor I desire. And I know it is a lot to ask. 

Nevertheless, here is the rough outline of what I'd love to see done well.

Independent variables examined using county-level data for the U.S.:

• 20-day trailing average humidity
• 20-day trailing average temperature
• Latitude 
• Population density
• Stringency measure (government-mandated restrictions)
• Mobility measure during COVID relative to the same mobility measure average value for 2019
• Median income
• Proportion of population 65+
• Percentage of elderly in LTC facilities
• Population proportion by ethnic/race ancestry (hypothesizing that prior immunities are associated different geographies)
• Date of first case within y-hundred miles (adjusting for treatments, interventions, etc. changing over the timeline)

Results:

• I would like to see the cross sectional results of confirmed COVID deaths by standardized timeline (from date of first case; from date of first death; from x days past first case within y-hundred miles).
• I would also like to see the time series analysis in total and by various cohorts for confirmed COVID deaths. 

My general hypothesis is that every thoughtful observer will find the results somewhat surprising. These same, thoughtful observers come at the problem with their own biases and priors as well as some unintentional agendas (the intentional agendas are for the unthoughtful observers). I think they tend to emphasize the areas they find compelling while somewhat negligently remaining silent or quiet on the areas they actually don't disagree about but feel are overemphasized by others. To that extent there is a lot of talking past one another. I am very specifically thinking about the realms of both the libertarian/classical liberal/neoliberal/generally freedom-championing thinkers and the economists/social and public policy thinkers. 

The problem is the audience has very much grown and diversified for these thinkers. It is very hard for the casual observer to understand the nuance and the starting positions of general agreement. For example, the public has always been completely oblivious about the fact that economists agree fundamentally to the vast extent that they do.

In the case of COVID this problem has been greatly magnified. And at the same time the slippage into hyperbole has been greatly amplified too. The result is painting ourselves into corner solutions. When the narrative has been taken to the extent that many so often and so easily have taken it (myself included of course), our narratives tend to fall apart. Again, this is for the thoughtful observers. 

P.S. Yes, at first glance I can see potential problems with the independent variables, and I assume there are many more I can't yet imagine. The covariance between stringency and mobility might force one of them out of the analysis. In that case I would like to see a rigorous comparison between just these two--Phil Magness points to some of what I would expect in that voluntary mobility changes dominate policy. After all, politicians follow rather than lead.

What Stops School Choice?

We should fund students instead of systems.

That is the ubiquitous and powerful tagline from Corey DeAngelis and the larger pro-school-choice movement. I wholeheartedly agree with it. 

I also strongly agree that the pandemic was a very fortuitous turning point in the long struggle to improve educational opportunities for all and especially underprivileged students. As one of my Big Five low-hanging fruits of public policy, this is an exciting development. Yet we are still a long way from here to there. 

Understanding what stops or prevents immediate realization of the eventual winning idea, that parents should be free to send their children to the schools of their choice, should help us understand what it will take to get there eventually and hopefully sooner and sooner. 

Here is my rough approximation of the obstacles including how much they account for the prevention:

• FOOL & FOOM*  --  20%
• Local Xenophobia  --  20%
• Incumbent Interest Protecting Turf   --  60%

The first one is a shared concern of both progressives and conservatives. The second tends to be more the domain of conservatives--I'm thinking specifically of suburbanites. The last is much more in the domain of progressives. 

*Remember that FOOL (HT: Arnold Kling) and FOOM are Fear Of Others' Liberty and Fear Of Others' Mistakes. These are the only defensible position for the three causes I identify, and only on the surface are they defensible. IF we cannot trust others (other parents, etc.) to make good choices for their children's educational needs, then perhaps this is a justified obstacle. 

But if we cannot trust them for that, what can we trust them for? Feeding and sheltering those same children? Voting for the politicians that will run the school system that is instituted and operated by a government we can trust to do right by those parents and their kids? 

There is an awkward tension between placing blame upon parents for underperformance (disgraceful performance usually) of government schools while at the same time lacking trust that those same parents could or would possibly make good choices for the children in those government schools. Taking it one-step further, those who believe that the problems of government schools lie at home outside of the schools (e.g., lack of a stable home life, help with homework, etc.) should not advocate for greater and greater resources for the schools since they claim the problem isn't in the schools. If they are right, let's put those extra resources into improving things at the point of the problem.

Circling back to advancement of the school-choice cause during the past 12 months, I feel like progress is being made against all three areas of prevention. We can cure FOOL & FOOM by attacking them head on appealing to real-world examples and analogies along with asking those holding the concerns to put themselves in the shoes of those they are concerned with.

Local xenophobia takes a long time to cure. Appeals to morality only go so far. People have a natural, animalistic-like protection instinct for their children. The hypothetical Other disrupting the fragile world of one's child is a salient fear. Experience with counter cases is the best medicine, and it is a slow process.

Fighting incumbent power is like bankruptcy. It comes on slowly and then suddenly. I would like to think we are entering the sudden stage for many communities. Check out Corey's twitter feed for hope in this regard. As more successful examples emerge of funding the kid with a backpack rather than the building with the career administrators, it will be more difficult for the incumbent resistance to hold.

A Crowding Theory to Explain the Trade of the Decade

Bold title, I know. Yet while there may be some hyperbole in it, we do face a situation in equity markets quite unlike anything we've seen since the dot-com tech bubble era. 

[Disclaimer: This is not investment advice because I don't know you. I am not analyzing your situation, your appropriateness or suitability for equity or any other investment idea, your goals, or your risk tolerance. This is an attempt to explain what I believe to be the valuation existent in the market today. Read at your own risk.]

I. The Trade of The Decade

Typically with tolerable variation equities are "cheap" or "expensive" at the same time. The Great Recession was a period where equities went on sale, so to speak, for those willing and able to bear the risk. You didn't have to be brilliant; you just had to be brave. Buying and holding throughout the deepest points of the downturn (roughly Oct 2008 - Mar 2009) was a rewarding investment. Make no mistake it was not easy to do this as for years after the market bottomed it still never felt quite safe, but that is what investors generally and equity investors specifically get rewarded for--buying when no one wants to. 

At least they should be rewarded for that. And the expected reward is the "risk premium"--the expected return over and above the return for a risk-free investment. Think of it simply as the premium you should expect to earn for bearing the risk. Every investment has one. Sometimes it is relatively high like for stocks in the early 2010s. Sometimes it is relatively low like for stocks in the late 1990s. That "relative" measure could be in relation to its own history alone, the position of that asset class compared to other asset classes, or both. Current market risk premiums are the foundation behind my hypothesis that a big opportunity/challenge is in front of equity investors today.

The "trade of the decade" I allude to is a relative trade. Unlike so many highly-touted trades in history, this one is fairly simple. I wish to briefly outline it without effectively proving it so as to get to the true purpose of this post: a hypothesized reason for the opportunity to exist in the first place. After all, I am a strong supporter of semi-strong efficient markets theory (EMH).

First, some assumptions:

1. Markets are generally efficiently priced. This means all important known information is thoroughly and quickly incorporated into market prices. Therefore, an investor should not expect to be able to outperform the market. 
2. Risk tolerance (degree of risk aversion) among investors and risk outlook (the economic picture going forward) determine risk premia. The world looked very scary in March 2020 and investors' general appetites for risk were substantially lower than usual. 
3. Risk premiums today are generally low across all asset classes. A couple of ways to say this in everyday language are: equity returns going forward (next ten years) will be lower than what we have enjoyed historically especially in the last 11 years and interest rates look to be lower for longer (lower than historically and for longer than typically has been the case). Caveat: risk premiums can be low and returns look high and vice versa. I am being a bit casual with how I equate risk premiums and future returns as compared to history, but I believe the underlying point holds.

Here is the problem, opportunity, and challenge all rolled up into one. Today certain portions of the stock market look fairly expensive (high valuation) while other parts look fairly attractive (low valuation). To be more concrete about it the high/low valuations are in comparison to those specific equity subclasses' own history. However, adding to the puzzle the risk premia for those asset classes look typical for the expensive group while they likely are desirable for the attractive group. 

Even a sloppy reader at this point is growing quite frustrated by the fact that I haven't identified which groups of stocks I put into the expensive and attractive categories. That is intentional as I don't want that to be the takeaway from this post, but I will now relieve that frustration as long as you know I am NOT effectively proving my case. Take this too as assumption. 

From the perspective of valuation, large-cap stocks in the U.S. (especially growth-style stocks) are expensive compared to their own history. Small-cap stocks in the U.S. (especially value-style stocks) are cheap compared to their own history. International stocks (especially value-style stocks) are also cheap compared to their own history. One way to measure valuation is to look at the current market price compared to the earnings, the P/E ratio. The higher the ratio all else equal the more expensive the stock.

One of the best ways to see this is to look at long-term analyses of real (inflation-adjusted) price/earnings ratios. The most famous of these is Shiller's CAPE (cyclically-adjusted PE) for the S&P 500 index. 

We can then do very similar analysis on various other indices considering important subclasses to see how they compare to the S&P 500 (large-cap U.S. stocks). This analysis [summarized in the table below] is where we see these key differences. Namely, that the riskier areas of the stock market (value-style companies, smaller sized companies, international companies, et al.) look relatively inexpensive.

Risk premiums* [analysis and results not shown] both echo some of the valuation analysis as well as tell a bit of a different story. The risk premium for large U.S. stocks is near the median of where it has been over the last 15 or so years meaning that adjusted for risk these stocks look appropriately priced. The risk premium for small U.S. stocks is far above its own historical median meaning adjusted for risk these stocks look very low priced. To a lesser degree the same low price depiction can be ascribed to international value stocks. The risk premium for international growth stocks is more similar to large U.S. stocks (i.e., it is around its historic median). So we have a general range of stocks from those that look appropriately priced (safer stocks like large companies and growth-style companies) to those that still look attractively priced (riskier categories mentioned before).

The risk premium story is not as sanguine for large U.S. stocks as it may appear. Despite my assumption above that interest rates will remain low for a long time, they don't have to. And they can rise meaningfully above current levels and still be historically low. If they do rise, that will have a big impact on stock valuations. Large growth stocks in particular look very sensitive to this risk as their cash flows come far into the future. A rising interest rate means a rising discount rate applied to those cash flows, which reduces the current value of the stock.

The trade of the decade is to fade away from large U.S. stocks and increase exposure to small U.S. stocks (especially value) and international stocks (especially value). The most acute version of this is for large growth versus small value stocks in the U.S. Specifically we could identify indices like the Russell Top 200® Growth Index and compare it to indices like the Russell 2000® Value Index. The reason I call it the trade of the decade is because my more in-depth analysis focuses on 10-year expected returns and risk pricing as well as the fact that it might take a decade to fully play out. The reason you should heed caution before engaging in this trade is (1) this risk might not be for you (it indeed comes with risk not specified in this post) and (2) this is not an endorsement of reducing diversification (I still advocate exposure to large-growth stocks, etc.).

II. The Crowding Theory

IF I am correct about the different relative risk-adjusted valuations for various equity subclasses such as large growth stocks versus small value stocks, interesting questions emerge. How did this come to be? What explains it consistent with EMH?

I believe two crowding effects have brought this about. Flight to safety is the first crowding effect. Sophisticated money avoiding public equities is second. 

Recall my assumption that risk premia are determined by the general level of risk aversion among investors and the market's risk outlook. The tidal wave of the pandemic that unfolded gradually then suddenly from mid-January 2020 through to the market bottom of late March 2020 was a massive reevaluation of risk that triggered an extreme flight to safety. Note that it was not the entirety of the valuation dispersion between various groups of stocks. For some time now growth has been outperforming value, large has been outperforming small, and U.S. has been outperforming international. For U.S. equities the market's reaction to the pandemic was in fact the dominant portion of the differences we see today. 

From Vanguard: 

Difference in annualized total returns over rolling five-year periods

Source: https://advisors.vanguard.com/insights/article/growthvsvaluewillthetideschange

As the risks and worry of the pandemic grew, investors sought refuge in the safest of assets, U.S. Treasury securities. This lowered interest rates to nearly zero across the yield curve. They also derisked in other ways. For those who wanted continued equity exposer, they flocked to the safest equities in the world, large U.S. growth companies. The now much lower interest rate conditions worked in tandem to make these equities more and more attractive. The riskier aspects of the market suffered as investors tended to rotate away from them and into safety. 

So that is how risk premiums for various stocks got so extremely different, but why didn't sophisticated investors step in to absorb the difference? After all, they are supposed to have extremely long (infinite?) time horizons and not be subject to wild swings in risk. 

Sophisticated investors is a bit pejorative on my part. Here I am talking about the so-called "smart money" of institutional investors like endowments and pension funds. They like to think of themselves as cunning lions, but they bunch together like scared, vulnerable sheep. Theoretically, they should be a counterbalance to short-horizon investors who are theoretically much more sensitive to changes in risk appetite and risk conditions. 

For example, a person close to or recently entering retirement should be invested positionally to withstand the risk of market volatility. The same can be said of any investor--they should be so positioned. Yet often times they are not. And even more often they are liable to overreact to bad news leading them to drastically alter their investment positioning as an attempt to predict the future. However, I do not think this is a big effect. It is a behavioral story that isn't necessarily as irrational as it seems--extreme events like the Great Recession and COVID pandemic give us insights into our attitudes on risk tolerance not apparent before. 

At the same time the sophisticated investors themselves are subject to the same types of risk tolerance reevaluation. Rather, to be a counter-balance they need to be in the market. The reason why the “smart money” hasn’t absorbed all the excess risk premium in riskier aspects of the market already is because endowments and pensions have trended far from traditional public markets. Endowments have on average reduced their public market equity exposures by 50% in the past 50 years going from about 60% to less than 30%. They are crowding away from public equities making them unavailable to provide a counterweight. 

To be sure riskier assets have done well in the past few months--small-cap U.S. stocks were up 35% for the three months ending in January 2021. The primary catalyst were vaccine developments in November. Add to that improvement in our understanding of the true risks of the pandemic as well as rapidly improving economic fundamentals. Even still, risk premiums in risky assets (value, small cap, international stocks) are at elevated levels compared to their own history as well as their safer equity counterparts. 

As risk aversion gradually (and perhaps in sudden bursts) returns to normal levels and as the risk outlook continues to improve, my hypothesis is that those assets with outsized risk premiums will perform relatively well (high confidence) and absolutely well (moderate confidence). 

Incidentally while this bodes well for public risk assets, it likely portends poorly for alternatives (private equity, venture capital, hedge funds, et al.). That is a crowded space with not much low-hanging fruit, and what is there is very expensive to be had. 

*The reason I deliberately gloss over the risk premium results is the model I am referring to is proprietary, but more importantly the calculation of risk premia is art and science. Laden with assumptions, it can be very much argued over in fine detail. However, I believe the depiction above is well grounded and firmly supported by a wide range of reasonable underlying assumptions. For more on this topic see Research Affiliates work among many others.