Thursday 30 January 2020

Learning the wrong lessons; style drift; and why smart value investors underperform

There are many things that separate the great value investors from the poor to mediocre, but aside from simply being better or worse at valuing companies (table stakes for a good value investor), one of the most important is that the former tend to reason from first principles - i.e. from things that are true by definition in the long term - and implement a disciplined and consistent process informed by those principles; whereas poor to mediocre value investors attempt to draw far too many 'lessons' about how to invest from recent market experience/outcomes. Quite often, the belated incorporation of these 'lessons' into investment decisions results in untimely 'style drift', with a shift towards strategies/sectors/stocks that have worked well in the recent past, rather than those that are most likely to work in the future. This untimely vacillation all but ensures long term underperformance.

For instance, a value investing newsletter I subscribe to recently published a list of 'never sell' stocks - regardless of price. Yes, this from a service that purports to be value oriented - an investment strategy that is defined more than anything else by its price vs. value ethos. (I don't mean to pick on them - I'm a subscriber, and they are smart, open to debate, and have some great content which I enjoy reading, but this 'quality over price' ethos is something I disagree with strongly, and is so emblematic of a lot of what I've been seeing in the world of 'value' everywhere I turn lately, that I couldn't help writing something about it).

Predictably, the list of stocks is populated by high quality, resilient businesses with capable management that have delivered steady growth and high returns on capital in the past, while maintaining solid dividend payouts. However, they also have another very important thing in common: the stocks have all performed exceptionally well over the past decade not only due to their quality/business performance, but also because they started at reasonable multiples, and have experienced a decade of multiple expansion - typically from about 15x to closer to 30-50x (less for some lower-multiple industries).

The combination of solid growth, good dividends, and an incessantly rising multiple is always a heady combination in markets, and has ensured a decade of fabulous returns. However, the most important psychological outcome of this experience has been that any decision to sell/reduce such stocks over the past decade - due to the exercise of 'price discipline'/'value consciousness' - has proven - ex post - to have been the wrong one. Consequently, the lesson the newsletter has drawn from this experience is that it is a mistake to sell good businesses, regardless of price. You should buy good businesses and never sell. Why? Because the past decade of market experience proves it.

The problem is, multiples don't always rise, and the conclusion also ignores both market history and important first principles. Markets go through long, secular cycles that can last a decade or more, and strategies that have worked well over the past decade are often amongst the worst performing during the next, because expanding multiples invariably inflated returns during the prior prosperous decade. This has a corrosive effect on future returns for two reasons: (1) firstly, starting-point FCF yields are significantly less than the prior decade; and (2) in the long run, multiples are subject to mean-reversionary forces, and the higher the starting-point multiple, the more likely it is that multiples fall rather than rise in the coming decade.* (Another important contributor is that high-performing, high-growth sectors attract a lot of capital and new competitive entrants, which drive down future RoEs).

The past decade was not the only one where high quality compounders came into favour, and where it eventually came to be argued that quality ought to take absolute precedence over price. It happened during the 1960s-early 1970s with the 'Nifty Fifty' - the so called 'one decision' Blue Chip stocks, where valuations got as high as 60-80x earnings. The stocks subsequently fell 80%. It happened again in the 1990s, where not just tech stocks, but Blue Chips such as Coke and GE traded up to 40x earnings. Coke subsequently fell 50%, and took about 15 years to return to its prior levels, while GE - the archetypal 'never sell stock' of the 1990s - is currently 80% below its 2000 price levels.

During all these boom periods, good business performance was supercharged by substantial multiple expansion, and investors failed to disaggregate how much of their quality, 'compounding' return was coming from non-repeatable multiple expansion which had driven down future return potential, and how much was coming from actual underlying business performance (something which incidentally is also frequently more mean-reverting long term than people think - GE is a case in point). During each of these episodes, years and years of rapidly rising prices caused investors en mass to learn the wrong lessons: that quality was paramount and price secondary, and that you should never sell. In all cases, such stocks delivered disastrous performance in the decade that followed.

The underlying dynamics are best highlighted by a simple mathematical example that takes the emotion out of it, and thereby facilitates more rational thinking. Take a hypothetical company that trades at $15.00 and earns $1.00 (15x trailing P/E). Let's assume this is a great 'compounding' company that earns a 15% RoE; pays out 50%; and grows at 7.5% a year. If you bought this stock, and 10 years later the stock traded at 35x earnings, your realised return including reinvested dividends along the way (assuming a steady geometric rate of share price appreciation) would be a cool 19.3% a year. The stock would rise to $72.14, while $7.07 in cumulative dividends would have been received along the way. A five-bagger! Any decision to sell/trim along the way would appear to have clearly been a mistake.

However, what would be the outcome if the company delivered the exact same operational outcomes and growth during the next decade, but this time the multiple fell from 35x back to 15x? The answer is just 0.9% a year compounded, including reinvested dividends along the way. The same company delivering the same RoE and growth would experience a 'lost decade' simply due to multiple compression. Earnings would rise from $2.06 to $4.25, but with the multiple falling to 15x, the stock would be just $63.72 a decade hence - 11.7% below the price 10 years earlier. $14.58 in cumulative dividends received and reinvested during this period would salvage - very marginally - a positive total return.**

Ten years is a long, loooong time in markets. It has a tremendous impact on investor psychology. 10 years of rising prices and rising multiples is insufferable for many investors to 'miss out on', and it is equally insufferable to hold a position that goes nowhere/down for 10 full years as well (there are many 'compounder' businesses, from a book value, RoE, and dividend per share perspective in HK/China that have gone sideways/down for 13 years, as multiples have fallen from 20x to 5x; predictably, no one views these stocks as 'compounders'; merely 'value traps'). When multiples are falling, the psychology around stocks completely changes. Risks are scrutinized more than opportunities. Patience is punished rather than rewarded. People come to believe that it doesn't really matter how great the underlying business is because the stock never goes up and it lacks a catalyst - after all, 10 years of market experience proves it (at this point, people also once again argue price doesn't matter, but at the opposite end of the spectrum for opposite reasons).

If for whatever reason multiples were to contract throughout the 2020s for the high-quality, never-sell stocks referenced, I can easily imagine the very same investment newsletter reflecting in 2030 on the experience of the past decade, and the poor returns earned from many of their core long term holdings, and concluding that 'price matters'; 'a good company does not equal a good stock'; and 'when prices are high, a whole decade of great operational achievements may be neutered by a high price'. I can imagine vows being made that in the future, far more attention will be paid to price, because price ultimately determines your return, not business quality. This would no doubt inform investment allocation decisions being made at the time.

See the problem here? The fundamental issue is that lessons are being drawn from the market action, rather than first principles that will always work in the long run, as they are axiomatic (true by definition). Drawing such false lessons will cause investors to make untimely changes to their style preferences, favouring sectors, geographies, and attributes (e.g. quality) that have worked well in the recent past, and eschewing those that have experienced a long period of poor outcomes. This results in investors buying high and selling low, and in the process, delivering returns that underperform the operating results of the companies they own during their holding periods (e.g. when you sell a stock at a loss even though the company has been profitable during your period of ownership).

Furthermore, value investors of this ilk can further compound the damage they do to themselves by being late to identify a new market segment that has done very well (such as technology/software over the past decade). After missing a lot of the early gains, they go and do research on the stocks, and then belatedly realise these are actually great businesses they would like to own, and that the market's enthusiasm is justified. The only problem is that the price is a little too high for comfort. Accordingly, what they often decide to do is put those stocks on the watch list, and hope they are given an opportunity by 'Mr Market' to buy those stocks at a less expensive price in the future.

This is a very dangerous approach, because if you simply buy these expensive, rising stocks straight away, you'll at least be jumping on the momentum train and will make money so long as the current multiple expansionary trend lasts, which in the short run is quite likely (momentum is a real thing, as it is driven by the flows flywheel I have referenced in past research - performance begets flows; and flows performance - and intend to discuss in more detail in future posts).

However, if you wait until the stocks start falling to buy (after no doubt beating yourself up for 'missing' further up-moves beforehand) in a process I describe as faux contrarianism,*** there is a much higher risk that (1) the fundamentals have already started to deteriorate, unbeknownst to you, so the value investors end up buying off the smart money in regular contact with management that rode the stocks all the way up and are starting to exit - a risk amplified by the 'many eyes' and extensive brokerage coverage these stocks have often by now attracted; or (2) you have bought in the early stages of a flows-driven momentum reversal, in which case you may well end up owning the stocks all the way down during the multiple de-rating phase, while having failed to own them all the way up. This can also decimate performance, and is the fundamental driver of the well-known 'performing chasing' phenomenon that contributes to much underperformance amongst active managers.

At present, a clear sign that a value investor is being unduly influenced by what has recently worked is if they have wound up with most of their top positions in software/tech names trading at high valuations. This didn't happen overnight. Instead, they have been slowly and steadily seduced, by years of market action, into thinking these are the best stocks to own. No doubt, they started slow, and as the stocks they bought quickly moved up after purchase, they became emboldened by positive reinforcement to add more, and then even more. This goes on for a while and they end up loaded up in popular, expensive, crowded names at the peak of the cycle, despite believing themselves to be contrarian value investors that go against the crowd (the same thing happened in the late 1990s).

Furthermore, the above is particularly the case if the said investors have also underperformed over the past five years. If they have owned tech/software over the past five years, they should be massively ahead of the index. If they are not, they have loaded up their portfolio in expensive quality/growth compounders very late, and are performance chasing - even if they don't themselves yet recognise it. Unfortunately for many of these investors and their clients, they will probably only realise what has happened after a secular change in market momentum/flows dynamics occurs, and they end up substantially underperforming as the cheaper, lower-quality stocks they eschewed and once owned surge ahead. They will be left to reflect on what went wrong. No doubt, they will then pivot back to what worked well during their recent period of underperformance - cheap/unloved companies.

What investors should be doing is focusing not on what has worked in the recent past, but what must work over time on a first principles basis, and that requires a clear-headed understanding of the following: when you buy a stock, you are buying one thing and one thing alone - the right to receive the future dividends and other cash distributions the company makes. That's it. The only difference between me and Buffett/Berkshire with respect to Coke is that when Coke declares and pays a dividend, BH gets a big slug of cash deposited in its bank account, and I do not. Accordingly, any investment case that is not able to be justified based entirely on the dividends it is likely to pay out over the long term is not value investing (for instance, if a stock is on 50x earnings, it has to pay out 100% of earnings to yield 2%, and grow dividends 8% pa into perpetuity to deliver a 10% return).

Now sure - quality, competitive advantage, management, etc, contribute meaningfully to what those future cash distributions are. At a practical level, they are therefore an absolutely essential part of security analysis and the valuation process. But once a sufficiently accurate assessment of what those cash flows will likely be is made (which is the entire intellectual skill set involved in value investing), 100% of your future buy and hold forever returns will be driven by how the price you pay compares to the actual cash paid out in the future (in magnitude and timing), and the IRR generated therefrom. Double the price, and you halve your return - it's that damn simple! It is startling to see value investors arguing that 'price doesn't matter' when it determines 100% of your future return! It is a clear sign that we may be approaching the late stages of this growth/quality bubble.

Value investing sounds simple, and in many respects it is - intellectually. Many investors are drawn in by this simplicity early in life, seeing an opportunity to accumulate great riches with relative ease. However, the issue with its implementation is emotional - markets can go years and years without validating your process and delivering outcomes that converge with long term expectancies, and it is the rare investor that can resist the mentally corrosive effects of this negative reinforcement, and avoid learning the wrong lessons the market is repeatedly trying to 'teach' them. Many value investors view themselves as immune from these emotional tugs, but they are wrong. Just because you are prepared to buy a stock that has recently fallen 20% (particularly if it is associated with faux contrarianism)*** does not mean you are immune from these corrosive emotional drives, which are much subtler and run much deeper than most investors generally appreciate.

Investors also confuse ex post outcomes with ex ante probabilities - what actually happened compared to what could be realistically foreseen at the time. When you look back at a growth stock compounder that has risen and risen over the decades (such as Google), it is tempting to believe it was obvious all along, but that is often hindsight/survivorship bias. People today forget that in the late 1990s, Yahoo was one of the largest market cap companies in the world, as investors believed Yahoo was destined to be what Google has today become. Had it done so, people would today be arguing that it would have been a mistake to sell Yahoo in 1999. That's hindsight bias.

As Buffett has opined, it is often not intellect that differentiates the successful value investors from the also-rans. The analysts at the subscription newsletter I reference are smart. It's instead temperament, and an ability to resist the forces of style drift over very long periods; maintain a rigorous (and sound) process with price discipline through thick and thin; and not lose confidence in an approach merely because it has not been validated by market prices in recent times. 

It's a tough balancing act, because most of the time in life, if the world is giving you negative feedback, you probably are doing it wrong or aren't very good. If you try to play football and you keep getting knocked out and hospitalised, the world is trying to tell you something, and the world is probably right. You suck at/aren't cut out for football. There is therefore a fine line between being stubborn and ignoring feedback, and being tenacious and disciplined.

The key to success in value investing is not just discipline and consistency, but also being very good at valuing businesses, just like the key to being a good footballer is not just an ability to get back up after a hard hit, but actually having both the physicality and necessary skills to excel at the game. If you're not good at valuing businesses, you will deliver mediocre to poor results over time, regardless of your temperament. However, even if you are smart and capable enough at valuing businesses, you will still underperform over time if you allow the market to teach you the wrong lessons.


LT3000


*I discussed in a recent post how the structural force of excess savings has pushed down interest rates and returns on other financial assets (although also discussed how this is also potentially part of a very long term political cycle). However, even acknowledging the above, there is currently a record-high level of multiple dispersion - the world's most expensive stocks have never been more expensive relative to the world's cheapest.

**It actually gets even worse than the 0.9% return I mentioned, as that is prior to the inevitable occasional blow-up that happens in a highly rated, highly favoured stock. Sure, these events are unlikely, but across a portfolio of say 10-15 such stocks, it will probably happen to at least one of them. Almost on a daily basis at the moment in Australia, a highly-rated, high quality stock is selling off 20% on a profit downgrade. Gentrack - a popular SaaS story amongst 'value' and growth investors alike - has recently plummeted 70%. The de-ratings are so severe because if instead of growing, earnings fall 50%, suddenly the stock isn't on 30x forward earnings but 70x, while the future growth outlook is also revised down. Losses of 75% can happen extremely easily/swiftly with such stocks, and represent a permanent loss of capital. If only one out of your never sell portfolio of 10-15 stocks has this happen at any time over an entire decade, the overall returns from your portfolio are likely to be meaningfully subpar even if multiples don't decline.

***I call this faux contarianism. The investor sees themselves as going against the crowd in buying a stock that has recently fallen, but they are buying a popular, expensive stock that has been mostly rising over the past (say) five years; has merely become slightly less expensive/popular; and are relying on a highly-consensus assessment of the historical strengths of the business (already well known and priced in), rather than understanding and reacting to new information that may be calling that prior assessment into question. Real contrarianism involves focusing on areas of the market that are genuinely cheap and neglected, and have typically already delivered poor outcomes for many years, but have underappreciated mean-reversionary/recovery potential.