Samuelson’s Dictum

In the years between the end of high school and the beginning of university, I tried to participate in Magic the Gathering Pro Tour. How to define “trying” is pretty tricky. I was dedicating a lot of time to it, skipping lessons here and there, but I was not skipping the Saturday night out with friends even if the next day I had to wake up at 6am to drive 3 hours to a Qualifier tournament. I was not a natural and despite what I might have told myself, I didn’t want to put in the work. Only later I realised how ‘fatalistic’ I was about certain situations of the game: if I was finding myself in a hole, I was not fighting to get out but simply accepting it as the variance of the game. Many sides to the same lazy approach.

I guess I grew up watching movies and reading books where the grind of the protagonist is left out: take the movie 21, about a group of students’ card counters at Blackjack tables. If the movie wanted to correctly depict their life, it should have been 90 minutes of training counting cards, 5 minutes at a casino with no action involved, and only 1 minute would show a victory and some joy. Not sure how many would have watched this movie. Movies have to entertain and comprehensibly the training/hard part is always shown as a 30 sec fast forward section.

If you do not have someone to explain this to you, or be smart enough to get it yourself, you end up with any sort of justification for yourself. After that experience, I had a few casual tries at poker and sports betting, with a higher level of consciousness, that indeed confirmed I had an interest but I was in any way “special” and, most importantly, I did not have any willingness to grind. Since we are at this, you might as well throw in entrepreneurship. And trading, obviously!

Throughout this journey I came to the realisation that there are some things where I can create an easier edge for me, like not selling stocks when the market is crashing. As well as holding a weird portfolio, as is different from what normies do, both in terms of assets and weightings.

Months ago (I think, I did not save a link, dammit), Kris Abdelmessih wrote in his newsletter about Samuelson’s Dictum and BAM!, can it be what I had in my mind all along, articulated in a proper, structured way? I saved some passages for later…and now I am going to analyse them. Samuelson’s Dictum has been around for decades, I found it funny that I never came across it (most likely I did but I was not paying attention to it). I am also not 100% sure my interpretation of it is the correct one: confirmation bias is always lurking around the corner, ready to jump in 😉

Nobody is bigger than the market: The market you’re invested in is far more important than the securities you select within that market.

• Paul Samuelson’s Dictum: Markets are microefficient but macro inefficient
  1. At an individual security level, market agents are able to keep prices at approximately or close to general equilibrium but there are large structural barriers to arbitrage that keep markets from finding equilibrium at the macro ie asset class level.
  2. You can’t “stock select” your way out of a bad market environment. For example, they cite that a 5th percentile US equity fund crushed the 95th percentile emerging market fund from roughly 2013 to 2018. The implication is that these markets as of 2013 must have been inefficiently priced compared to each other. That’s a macro inefficiency
  3. Original BHB study and the wave of studies which followed showed that nearly all of the variability of a portfolio comes from asset allocation. The benchmarks for the allocation sleeves explain most of the return.

I’ll delve (one day I’ll ask an AI to twist a post I wrote by adding references to all MTG mechanics. At the end of the day, I learned English by playing MTG) on the “structural barriers” mentioned in 1) later because they deserve a bigger section. In a serendipitous moment yesterday, while listening to S7E5 of the Flirting With Models podcast, Clayton Gillespie said that fundamental analysts are really good at ranking stocks, thereby confirming the micro-efficiency nature of markets without thinking about it. (again, maybe it is just the me who bought a red car and now sees red cars everywhere).

Point 2) requires an explanation. You cannot stock select your way out because each stock in an index has a beta component. Micro efficiency means that stocks are ranked efficiently between each other, even when they are listed on different exchanges. 90% (here is me putting numbers) of the price difference between BP and ExxonMobil is a function of their relative quality; the rest might be inputable to UK investors wanting to have exposure to oil and having easier access to BP compared to Exxon (and vice-versa for US investors). The mentioned difference in performance between the US and EM markets is driven by the indexes’ different composition: the US is more tilted toward tech while EM is mainly commodities and financials. Also, while BP and Exxon might be interchangeable, good luck finding a substitute for Apple. Now, the price difference between Tesla and VW might be influenced by TSLA being loved by irrational investors and/or TSLA being listed in the US where there are more “passive” investors (see my previous post) but I’d rather stick to my original point: if you remove the sector and unique names discrepancies, the two indexes would have performed quite similarly. You should read the macro inefficiency here as a function of investors loving to give mandates, explicitly or implicitly by choosing ETFs, that constrain geographical exposures. Something that confirms 1) with the only difference that here it is a self-imposed barrier.

Considering the recent crypto resurrection, a good way to look at 3) is that if you have or not crypto in your Asset Allocation will drive more returns than the coins you bought. The simple decision to allocate 2, 3 or 5% of your portfolio to crypto will make the biggest difference in returns, positive or negative, compared to a more standard AA instead of deciding to buy BTC and/or ETH and/or DOGE. I’d assume 99% of the readers of this blog do not use an investment consultant, so the “benchmark” refers to the indexes the ETFs you bought are linked to. The more you ‘constrain’ those indexes, for example by choosing the Nasdaq100 over the MSCI World for the stock sleeve, the more you restrict yourself in terms of “markets you are in”. Same if you only buy GILTs instead of a market-cap-weighted portfolio of Government Bonds hedged back to GBP (look at what happened when Madame Truss got the job if the difference is not already clear to you).

Reasons why the market is micro but not macro efficient

• 99% of cognitive and computational capacity is devoted to security selection. Why?
  • Large investment committees direct the largest pools of capital and decide the asset allocation mix that they believe can help them achieve their goals ie match their liabilities.
  • Within the buckets underneath that decisions (equities, bonds, private investments, etc) are silos that focus on security and manager selection within the bucket.
  • Much more effort is applied within the silos, than at the investment committee level.
  • Since the large allocators devote effort to finding alpha in the silos they create demand for active managers within asset types but do not create demand for broader asset allocation managers.
  • The primary barrier to macro arbitrage comes down to:
    • Mandate inflexibility (this is also driven by regulation)
    • Lack of portfolio agility because of liquidity constraints/costs
  • Private retail wealth faces the same issue from a behavioral point of view. There isn’t much deviation from standard stock/bond splits because there is little patience for weird portfolios with high tracking error.

Active investors have stronger incentives to correct (micro) inefficiencies in relative prices than to correct the overall (macro) price level. For example, active investors will ensure that the price difference between General Motors and Ford is close to efficient, but active investors may leave all stocks overvalued. Why this is the case is a super-interesting question. I guess there is also an easy answer: micro inefficiencies “close” faster than macro ones. If I am an active investor, I need the market to reflect my view to make a profit (unless I am dealing with financial instruments that have a maturity); I cannot wait for the market to mean revert in 10 years because I’ll be jobless way earlier. The financial graveyard of macro investors who thought that currency XYZ was overvalued has been overfilled since FX trading started to be a thing.

There is some research out there that tries to challenge the micro efficiency thesis bringing to the table the fact that also single stocks reach bubble valuations. The issue I have with this reasoning is that is quite hard to establish when a stock is overvalued and when the market is upfronting future gains. Take CSCO in 2000, was it in a bubble or the market was simply saying all future earning improvements are already priced in? As I presented in my previous post, there are smart people on both sides of the argument: some that see profits eventually catching up with prices, others that consider those stocks as exceptions. There is also a timing component: TSLA is considered to be “in a bubble” for more than a decade. After such a long period, the conversation ceased to be practical, since shorts have been slaughtered many times over.

My (limited) experience dealing with other non-institutional investors/savers shows that 90% are more interested in how to optimise within each asset class bucket (world stocks vs home country bias, developed vs emerging, tech or factors) than in achieving a better overall return through the asset allocation process. In other words, to use bonds to reach financial goals within max the next 10 years and stocks for everything else. Without thinking that inflation can get nasty even on a 10-year horizon and bonds are not that great to deal with that. Or, on the other side, that stocks can go and stay underwater for decades, plural (see the MSCI World in EUR).

Having some experience in dealing with investment committees (but not enough to be able to write committee correctly on the first try), I can only put my approval stamp on what Kris wrote. The asset allocation process is first and foremost driven by not getting fired: no “weird” weights, i.e. the IBM story applied to asset classes. Was there an AA that would have limited the damages of inflation in 2022? Yep. But no one got fired for not using it because inflation was a problem all allocators had; at best, you would have received a pat on your back if you did better. But you didn’t because no committee would have accepted to allocate to CTAs given their prior decade track record. Everyone is piling on Private Equity and Private Credit despite the absurd valuations and zero upside? If you don’t, you are a Neanderthal who didn’t read Swensen.

Here is my try to explain why macro inefficiencies are hard to close:

• Regulation: banks and insurance companies have to hold bonds, mainly govies and mainly from the country they operate in, if they want to be efficient from a capital point of view. This is because the local regulator determines how punitive is to hold a certain asset. Pension funds are heavily regulated in what they can and cannot do as well. The regulator considerations are not 100% aligned with risks priced by (free) markets (the Turkish regulator considers Turkish Govies as 0 risk. Or Chinese Money Market Funds are rated called AAA*, meaning they are treated like AAA securities…if you are in China). See SVB as exhibit A of issues with this framework.
• Investing in foreign markets means having to deal with FX volatility. Few have the appetite and hedging is expensive (plus, in many cases, FX is one of the components of the undervaluation but, as I said, you get the volatility today and the profits maybe in a decade).
• Access. Another regulatory nonsense preventing investors and/or brokers access to some parts of the market. How many European savers would invest in AQR Mutual Fund if they could? Would this move the needle? It is just me being pissed at this? Probably. Just forget about this point. Srsly, bond issuances are heavily compartmentalized by regulation but I think the main impact is on credit spreads…and those tend to be too tight than otherwise anyway.
• Leverage. Probably there are hedge funds that despite their weird tracking error are thriving by exploiting macro inefficiencies, but they do not have access to enough (or they do not want) leverage to close those inefficiencies. Leverage is risky by itself, employing it with very long-term strategies opens a “path risk” that is hardly justifiable.
• It is hard to determine when one asset class is over/undervalued compared to another asset class. Think about Risk Parity: the concept is to balance assets based on their risk, as represented by their volatility, correlations and other risk measures. It is already a complex strategy like this. No one is talking about a Risk-Adjusted Return Parity strategy, one that also incorporates forecasts on each asset class return; those are normally kept constant in the model (this is what I understood!). Actually RP reduces its exposure to stocks when prices crash because vol spikes…the reverse you would need to close macro imbalances.
• Retail investors need a high rate of returns to match their objectives and beat inflation but they do not like leverage. Therefore, their portfolio has a threshold in terms of returns that cuts out some assets or combination of assets.

Getting a good understanding of what is happening within a single asset class is hard and requires years of study and practice. But if you are an expert in one field, say bonds, it is easy to think that you can have a great reading of what is happening in crypto. You might be able to profitably trade tactically one class and give it all back trying the same with the others. If you rely on expert advice on each asset class, it is hard to find real expertise and it is even harder to find an honest expert who tells you “the asset is overvalued, shift your allocation to something else, let’s talk again in a couple of years”. I spent four years dealing with Private Equity guys and at a certain point my question became blunt: would you tell me if the asset is overvalued now? The good guys admitted that they cannot. Incentives matter and in this particular situation it is difficult to align them; most likely, you are better off seeking advice from Twitter profiles for free than finding advice that you are willing to pay for.

Rodrigo Gordillo wrote a post on Return Stacking as a way to exploit macro inefficiencies to beat a benchmark. I am a stan of the concept of return stacking but this feels a bit of a stretch to me.

My two (very biased) conclusions are:

• diversification between asset classes pays because you are less likely to buy an overvalued asset
• do not rebalance aggressively, assets trend and you should let them run a bit

What I am reading now:

Follow me on Twitter @nprotasoni