Introduction
Investors from all over the world have always had a secret dream: making investing automatic and effective. Earning large sums of money with little or no effort has always attracted students, economists and mathematicians, not only for the likely higher returns, but also because of the exciting challenge that creating a system or a formula that works represents, as that feels just like finding the Holy Grail.
But it was only when databases containing companies' financials became available to the greater public that the quest for a working formula assumed the character of a real science instead of being considered an arcane topic.
Quantitative investing (or quantitative analysis) is an approach that uses mathematical or statistical methodologies in order to find the value of securities, and ultimately determine which ones are selling for less than their (presumed) intrinsic value.
Different methodologies
One of the pioneers of this approach was the late Benjamin Graham, the father of value investing, who once described a formula to determine the intrinsic value of a company.
The formula is:
Intrinsic Value = Earnings Multiplier * EPS
Where:
Earnings multiplier = 8.5 + 2*g and
g = minimum future earnings growth rate of the company.
The formula is based on the assumption that the price-earnings ratio of a no-growth company is 8.5.
Of course, this is just an approximation, but for some time it just worked fine, as the assumptions on which it was based were, on average, valid and measurable. Another reason why it worked is that, several decades ago, not a lot of people approached investing (and specifically, valuation) that way. Graham was one of the first people to rationalize the concept of value as being linked to earnings and cash flows.
When Warren Buffett (Trades, Portfolio) was a young man, he used to voraciously read thick stock manuals full of financial data for thousands of companies. As those manuals actually were the only "databases" available then, we should not be surprised that, armed with enough time, patience and effort, one could, for instance, spot a company selling for less than its cash on the books and profit from it, which, at that time, must have been like finding a buried treasure.
When Joel Greenblatt (Trades, Portfolio) was a young student at Wharton, he read a Forbes article about Graham's net-net stock picking formula. The method consisted of buying a stock only if it sold at less than two-thirds of its net current asset value (NCAV can be calculated by subtracting any non-current debt from working capital).
At that time, he was studying the efficient market theory, but he was growing increasingly disappointed as the theory was not resonating with him, so he started to verify if picking stocks that satisfied Graham's formula actually produced higher returns as he predicted. It did, and the rest is history.
Fast forward to 1992, famous professors E. Fama and K. French expanded the popular capital asset pricing model formula by adding size and value risks factors to the already existing market risk.
Specifically, their model stated that small-cap and value stocks outperformed markets on a regular basis compared to the ones that did not possess these features. By value stocks, they meant ones with higher book-market values (basically, stocks with low price-book ratios).
Small-cap overperformance can be explained by the fact that small companies can easily fly under the radar, so they have a higher probability of being undervalued compared to higher capitalization companies.
Regarding the value component, there's growing evidence that selecting stocks with a low price-book ratio compared to ones with high multiples does not necessarily constitute a predictive factor for stock price outperformance. This can be understood if we think to the fact that, while in the past most of companies produced their cash flows by intense use of physical assets, today the most profitable ones have usually little tangible assets and, consequently, are not capital-intensive anymore.
Quantitative analysis and its limitations
Let's now go back to the original question: Is a quantitative approach to investing still useful or should we discard it in favor of deep security analysis?
The answer is not easy, but here's my take: quantitative-based investing is not enough anymore, simply because almost everyone now has access to a computer, so there's simply no big advantage basing our stock-picking endeavors merely on statistics or financial metrics.
But if that approach doesn't produce the outstanding results it used to produce in the past, why is everyone still using it? Because it's better than throwing darts. Having a powerful quantitative tool is not the same as knowing how to use it. You still have to set the right inputs and conditions for the tool to filter out companies with bad prospects.
That's why, for example, applying Greenblatt's magic formula, which fundamentally tries to buy good companies at good prices, still gives you an advantage over many actively managed funds (at least the advantage was clearly evident until a few years ago). Greenblatt's quantitative method still selects potentially healthy companies and filters out most of the struggling ones. Yes, you can easily stumble into a value trap by using that method without any additional analysis, but, on average, you'll probably do well in the long term. As Greenblatt explained in "The Little Book That Beats the Market", if everyone would start using his method, this advantage would rapidly disappear.
Another good reason to use quantitative analysis is that it works quite well as a screening filter. Even the best investors use it to restrict the investing universe to a set of companies which have better probabilities of producing higher returns in the long term. It is naturally not the only method they use to research stocks, but, as Charlie Munger (Trades, Portfolio) once said, "the first rule of fishing is fish where the fish are."
Let's now try to list what we can and cannot achieve by using the above mentioned approach.
Quantitative analysis can:
- Filter out (most) bad companies.
- Detect past financial metric good trends.
- Combine multiple quality metrics in a single tool.
- Reduce our investable universe, consequently reducing search time.
- Increase the probability of finding a good company.
But it can't:
- Estimate future cash flows as it is based on past data.
- Detect a company's moat or tell us how solid it is.
- Tell us how good the management team is.
- Pick a future winner which doesn't have a good track record yet.
- Correctly estimate the intrinsic value of a company.
As we can see, what a quantitative tool cannot do is exactly what a good value investor is supposed to achieve.
Finally, we can't totally exclude that some smart investors (like Jim Simons (Trades, Portfolio) of Renaissance Technologies) can create a quantitative method that works consistently and produces above-average returns, but would rather be the exception than the rule.
Here's what Terry Smith, founder and chief executive of Fundsmith, said during a recent interview:
Interviewer: What is your view about the growing use of computers for stock picking? And do you think that the traditional fund manager will be replaced by computers in the future?
Terry Smith: Look, there's no doubt that you can use computers for this. The whole passive industry is basically driven by the use of computers to make the stock non-selection in fact, But beyond that into the active area [ ...] I think there's likely to be a role for human beings in active management for, maybe forever, certainly for a very long time.
The best way I could express is this, look you can do an awful lot of this stuff mechanically, but the human element comes in when you get somebody who is intelligent and very experienced [ ...] and you listen to the management present at meetings, conferences and you meet them and so on, unless you've done this, you won't get it.
Conclusion
Many investors have tried to find a quantitative formula that could be used to predict either the value of a security or predict future stock returns.
Some of them brilliantly succeeded, as they were able to exploit big inefficiencies present in the market due to overlooked stock, poor research or lack of financial databases or computing power.
Nowadays, good quantitative analysis can be conveniently used to produce slightly better-than-average market results and for stock screening purposes, consequently reducing research time.
Unfortunately, since this method is based on past data and not on the company's present business dynamics, it can't be (normally) used to spot future market winners.
Developing a deep knowledge of the company and sector, understanding its business model, studying management's track record and investigating its competitors and customers in order to estimate future cash flows and growth is still the best (and most difficult) method to outperform in the market. As Thomas Edison once said, "There's no substitute for hard work."
Read more:
- Is NVIDIA's Pending Arm Holdings Acquisition a Good Deal?
- Genmab: A Company With Downside Antibodies
- How the 'Fear of Mising Out' Can Harm Your Profits
Not a Premium Member of GuruFocus? Sign up for a free 7-day trial here.
