or…The Italian Leather Sofa for the community. Commenting a recent post, Luca suggested me to have a look at The Everest Formula website and…why not! I wish there will be more interaction on this blog, if you post your questions, ideas, remarks in the comment section instead of sending me a DM we will have a more shared conversation and you will help the blog from a SEO point of view. We love win-win’s here.
When I read the comment, my mind went straight to Joel Greenblatt and indeed, the website has a clear reference to his Magic Formula. I read Joel’s book, The Little Book That Beats The Market, around the year it was first published, so more than 10 years ago. If you consider my really bad memory and the fact that in my heart there is only space for one Joel, and that slot is now devoted to Joel “The Process” Embiid, you can understand why my reaction was similar to ohhhhhhhh shit 🙁. Aside from the name and the formula itself, I do not recall anything of the book; my memory is so bad that actually Joel was interviewed just weeks ago on the Master in Business podcast and…I do not remember a lot from that conversation neither, other than the fact that he definitely moved on from value investing. If you are interested in this post, check out the interview because it will give you a better context on the character and his way of thinking.
There will be two main topics in this post: Value Investing in general and how to critically judge an investment proposition, using the Everest Formula as an example.
Value Investing
Where to start…
An investment strategy based on intrinsic value makes a lot of sense, at least conceptually. You buy at 80 cents what should be worth $1 and once the valuation gap is closed, i.e. the market realise that what you hold is actually worth $1, you sell and realise the profit. This basic concept can be declined in multiple ways. The easiest one is to understand when a company has a market value (capitalisation) lower than the cash on its bank account. Given that the value of a cash balance should be undisputed (as long as crypto integralist wont take control of the planet), if you buy a share in that company you get back an equal amount of cash PLUS a slice of the value of the business: you are getting the business value, and its capacity to generate future cash flows, for free. This simple value strategy made Benjamin Graham rich. It also worked because at his time information was not as available as today: everyone knew the price of a stock but to know how much cash it had (and maybe also how many share were outstanding) you had to get access to their paper Annual Report…and hope that between the closing of the fiscal year and today that company did not spend the cash in a non-profitable way.
A lot of value investing strategy are indeed an information arbitrage. The difference between the past, where you have a successful track record, and the future, where they generate very few signals and therefore becomes unprofitable, is that the information gap is closed by technological advancements. Today it takes me 5 sec to check the cash balance of every listed stock and few lines of code to implement that strategy. Danieli is an Italian company that represented a good investment for this strategy:
It is also a small company listed in a globally disregarded Exchange. This type of opportunities still exist but they are so rare and globally dislocated that is really hard to build a relevant trading strategy around them now. (The ultimate value strategy in this category is insider trading: it might expose you to ‘different’ set of risks but they are not necessary all catastrophic, see Steve Cohen).
In a recent interview, Charlie Munger joked with Buffett that the majority of stocks they bought are now bankrupt. If you buy a company for $1 and before going bust it pays you more than $1 in dividends (that you not reinvest in the same company), your investment is profitable. These are the ‘cigarette butts’, another example of successful value investing strategy. This is also a value strategy that can be hardly coded in a rule-based fashion because the key is to make the correct assumptions about future cashflows.
Another disruptor for value strategies is what Mark Andreessen forecasted in his paper “software will eat the world”; software is the (first?) product to have zero marginal cost. Value strategy metrics were born for a different world, a world where economies of scales were possible but not infinite, a world where 10 people more working on a factory line to produce a widget were increasing production by a factor of 10. How much is worth Mark Zuck as a software engineer? How many software eng you need to replace him? Accounting rules have an hard time to properly recognise these elements and therefore traditional value metrics becomes less and less representative of the true value of a company.
This paper from Howard Marks explains really well the struggle that even a great money manager like him have to define what is value in today market. Accounting rules are a model to represent reality and they are by definition a laggard in relation to how reality evolves. Value strategies are models built on this model. When reality changes, i.e. intangibles become more valuable than tangibles, your model becomes outdated and stops working. In this sense, saying that a value model worked for the last 20 years is not really informative because the next 20 years might be different (in the model sense!).
Value is often considered in relative terms: the typical screener for value does not include absolute rules like “buy if company X P/B is less than 1” but relative ranking like “buy the first quantile of companies ranked by P/B” within a group of stocks, i.e. the S&P500. Given its relative nature, there has been periods in the past where the ‘distance’ between the first and last quantile has shrank and periods where it grew. These cycles have lasted decades: one thing is to say this strategy outperformed the index by 10% a year EVERY year and one to show that the outperformance has clustered in particular periods. No matter what is your value formula, you have to expect long periods of underperformance, periods where you will question if your strategy has still an hedge or the market changed for good and sticking with it is just foolish.
I have part of my portfolio invested in value strategies, some passive, some active (I also invest in other factors). I rebalance my portfolio quarterly to take advantage of the under/overperformance periods: I buy when the strategy underperforms and sell vice-versa. It has been mainly an underperformance in these years…but:
- cycles can be looong and I prefer to die next to Cliff Asness than Chamat (is he a proxy for growth investing?)
- if you are concerned about valuations, it makes sense to buy relatively cheap names. Value investing has a bit of risk management built in, I am giving up some upside to protect my downside (the world cannot be all value traps, innit?).
I use passive funds simply because I cannot access all the active ETFs I would like to buy, thanks to FU***NG MiFiD, otherwise I would only have a mix of Cambria and Alpha Architect funds. I like active-quant in this space because their managers do extensive research to try to mitigate all above descripted issues and publish white papers where they explain their strategies.
The Everest Formula
Whenever someone propose you an investment strategy you have to ask yourself: why? why me? why now?
I do not know Greenblatt personally but he must be an intelligent person; being smart, I am not sure if he tries to do good because he genuinely believes in it or because he knows that a good marketing stunt goes a long way to increase your public figure and your legacy. Let’s assume he is nice. I am sure when he published his first book he did not expect readers to do 30%/year in the future; he wanted to deliver the message that getting decent returns in investing is not that hard and everyone can do it. More importantly, the reader should avoid following the new kid on the block: he created a positive screener, invest in cheap stocks, while he knew he wanted the reader to have a negative screener, avoid expensive therefore fashionable and shiny stocks. If you get benchmark returns, you are already doing better than 85% of active traders…and yet no one of them will buy a book that teach them out to get ‘benchmark-like’ returns. The God of marketing needs you to die on the altar of the false promise, even if in his case is half false because for a period he must have realised those returns.
30%/year puts you in the same league as Jim Simons, if anyone managed to achieved it constantly you would know his name by now. The formula worked in a particular context for a period of time, it might deliver positive and above benchmark returns in the future but if it was that simple, I think not only all the value managers out there, but all the HF managers would stop playing with super-computers, satellite images and math PhDs and simply run that screener. In the FAQ they mention:
it is unlikely “overfitted” on the historical dataset. Overfitting is a common problem in financial algorithms, and occurs when the model is too closely fit to a limited set of data points.
My first reaction when I saw the website was: it is not that hard to create a formula that overperforms the Magic Formula in the past, you just need to change some parameters and back-test until your ingredients work better. Here they acknowledge that overfitting is a typical issue of these models but they do not explain why it does not apply here. Having five serious drawdowns does not mean you are overfitting issue-free. The only way to prove that the formula works is to provide an out-of-sample test; and as we said before, the out-of-sample period must be sufficiently long to cover different regimes (bull market, bear market, etc) so at least 10 years.
The biggest omission in their back-test are trading commissions and slippage. Trading fees might be a relic from the past now that traders have access to commission free brokers like RobinHood; the reality is that part of those commissions are now added slippage, meaning that HFT that fill your orders tend to move the market against you, especially if you are trying to buy or sell in size. Slippage is the difference between your ‘model price’ and the price you effectively get when you close your transaction. While this difference might seem trivial, just a few 0.000X%, if your algo trades frequently this difference can impact materially the final result. The fact that modelling slippage is really hard and different traders face different price conditions does not mean “so let’s assume is 0”, or worse “it is not a model issue, it is an user issue”.
WHY?
In Joel case the why might simply be that a lot of rich folks are not happy being just rich, they want to be rich AND famous. He might want to be famous by doing good, by teaching people a better behaviour using his status as a proof. He is definitely not trying to become rich with book sales, Trump does (if I have to pick the first example it comes to my mind…and no, Trump is not rich, he inherited money and squandered most of it, that’s why he needs income again).
If you are in love with value investing and want to follow a simple strategy have a look at Eddy Elfenbein. In his blog, he publish his Buy List since 15 years and everything is recorded, rules, changes, he even tells you what he is going to buy days before he actually does, you can front-run him! He can do it because the stocks he trades are so large that he would need a KimK audience before starting to move the market; he also trades so infrequently, 10 trades a year, 5 buy and 5 sell, that it does not matter anyway. He can do it because no other professional manager can follow his strategy: the first year of underperformance he will be fired because his clients and bosses would think he is a lazy ass, “only 10 trades this year?!? how did you spend the rest of the time???”. Instead of a backtest, he has a track record. He recently launched an ETF to match his portfolio, and the fees are aligned with his overperformance.
There are a lot of research shops that sell their analysis to hedge funds and money managers. They do not trade their own ideas because they know that the research is just a part of the process. Conversely they do not sell any performance based on their ideas as a proof of the quality of their analysis. To be honest the changes Everest made to Greenblatt’s formula actually might make sense, given the current predominance of tech companies and accounting rules. If they framed the website like this I would have no problem: “We believe in value investing and this formula for us is the best and easier way to tilt your stock portfolio towards value. We know that is difficult to obtain the necessary data out of free tools available on internet, so we built the model in Bloomberg; to share the platform costs, we ask you a contribution.” If you want to provide a backtest fine but at least use a real case with fees and slippage. And obviously do not publish results that are clearly due to data mining and/or would disappear after trading costs.
WHY YOU?
Why they do not trade their system themselves? Why do they need your money? There are two categories of professionals that manage ‘outside’ (client) money:
- the mutual fund manager: their alpha, when is there, is tiny compared to the benchmark. If I run a stock strategy that outperform the market 1% each year (net of fees), it would take me decades to become rich using only my funds but it will make me the best fund manager ever if I do it professionally. That’s the reason WHY YOU.
- the hedge fund managers: the majority of the fund is their money, or if they just started, their share increases every year. They use external funds because interests are skewed, in some scenarios even if the fund loses they win, but they still have skin in the game.
To generate the equivalent of the max fee they ask you ($191) they have to invest only $636 of their own in their strategy; yes, it is ‘passive income’ but the amount you get is so trivial compared to your other source of income (the strategy itself) that I do not understand why they bother…unless you know that the returns you promote are not real.
WHY NOW?
The market is pretty efficient in balancing offer and demand. The same investment in High Yield bonds can generate different results if you do it at the beginning of an economic/credit cycle, at the end of it or if you are a buy and hold investor. The yield on those bonds grow/diminish in an inverse relation to demand: since the return on the investment cannot be more than the yield you invested in, if you buy when demand is high you will have a lower return compared to someone that bought when everyone was running away, all other things equal. In this example, the rational investor should distance from this opportunity the more yields go down; what happens in reality is the reverse.
If you are in the business of selling a strategy, like Everest, it is easier to achieve your goal when your strategy is running hot. If they did not include their backtested equity line going to the moon, this strategy will appeal to no one because value investing had pretty miserable results in the last decade. I bet there are 0 value investors on TikTok. The equity line is the element that creates FOMO…and the biggest red flag on this thing. Every serious value investor is borderline suicidal by now, go check Cliff Asness or GMO if you do not believe me.
In financial markets you get extra-returns because you take extra-risks, so goes the theory but reality is not far from it. Extra-returns from factor investing do not come from risk: there is no additional risk to be compensated for if you invest in a company with a lower valuation, lower volatility or better quality. Factors are behavioural anomalies. The size of these anomalies changes over time: value stocks do not quote at a fixed discount to growth stocks. Even if you should not ‘market-time’ a factor (again, see Asness), it obviously make sense to buy when the value discount is high instead of when is low. This is why value investors are so aggressive now pushing for their strategy: because their results have been shitty in the last decade. They lost against the S&P500 but they hope at least in a regression to the mean.
Conclusion
If you have the tools to build a screener like the Magic Formula, then I would suggest you to build and test your own stuff: it helps you to understand the signals, which is the main help you can have to stick with your strategy when you are in a drawdown.
If you do not have the tools but still you want to tilt your portfolio to the value factor, there are lot of ETFs in the space, both on the ‘passive’ and active side…even if here calling something passive is quite silly because it simply means rule-based. The quality of your investment will depend on those rules, the more transparency and understanding you have the better.
If you want to donate some funds to the Everest Formula team you might get positive results back, even higher than the S&P500. Just do not hold your breath about those 30%/year, I would take the UNDER on that all my life.
I hope this post gave you some useful insights!
What I am reading now:
Follow me on Twitter @nprotasoni