Artificial Intelligence & Value Investing
Perhaps the most widely known style of investing is value investing, which was founded by famed investor Benjamin Graham and is now followed most notably by billionaire Warren Buffett.
The idea of value investment says that an investor should purchase stocks considerably below their intrinsic value to enable the investor to achieve outsized gains from the security and reduce the chance of permanent capital loss due to adverse, unexpected developments at the business whose security was purchased.
The difference between the intrinsic and actual value is known as the margin of safety, which investors carefully choose when investing in company shares.
Value investing has been greatly popularized by Warren Buffett and his partner Charlie Munger, who closely adhere to it to run their conglomerate holding company, Berkshire Hathaway.
Throughout their history together, Buffett and Munger have turned Berkshire into one of the most valuable corporations in the world by consistently buying stocks and, more recently, entire businesses, for far less than they are worth, subsequently holding these investments for long periods—often decades—to reap the rewards.
What one might quickly—and correctly—notice about this process of investment is that for an investor to purchase a security for substantially less than it is worth, they must first determine with some accuracy what it is worth.
Without estimating the value of the business, the price an investor pays for a given stock has no meaning whatsoever, and the purpose of the intellectual process underlying the investment vanishes; the investor may as well buy the security at a random price.
Additionally, one might discern that in order to ascertain or even estimate the true value of a stock, an investor must analyze an unwieldy amount of data about the business itself, including its financial performance, the industry in which it operates, present macroeconomic conditions, and much more.
With the advent and subsequent proliferation of artificial intelligence, one might be inclined to ask how AI can play a part in this process given that computers can trounce even the brightest humans in the efficient analysis and categorization of large amounts of data. So, can AI be functionally applied to value investing? The answer is yes, but there appears to be at least one decided weakness in the technology.
The greatest shortcoming of applying AI to value investing is that successful value investing often requires the analysis of qualitative rather than quantitative data, such as the strength of a particular brand or the honesty and forthcomingness of a management team.
It is important to note that AI applied in value investing typically assists the investment process, as opposed to pure AI trading/investing, where the model makes the investment decision. In value investing, AI serves as a vehicle for determining a company’s intrinsic value which is important in helping actual investors make informed investment decisions.
Because of this, investment management firms have turned to using a diverse set of data sources known as alternative data, collected through different means.
For example, alternative data can include satellite images taken of a Walmart’s parking lot in order to gauge customer demand and company earnings during a certain period. While this may seem over-the-top, some investors have seen great predictability of their AI models by using these kinds of data in their investment analysis.
Other forms of data include company financial statements, earnings growth, and even Tweets made by the board. All of these sources provide information into a company’s value and are taken into account when using AI to assist the investment process.
Nevertheless, AI may not be nearly as adept at interpreting data that cannot be precisely expressed as a number or as a part in a system of numbers, and insofar as value investing depends rather substantially on this qualitative data, it is not unreasonable to question the usefulness of AI to the modern value investor.
There are often inconsistencies made by these models using qualitative data, which affects the business valuation. As such, investors such as Warren Buffett certainly do not use it and still prefer the traditional fundamental approach.
However, this may not forever be the case since AI technologies and data collection are constantly evolving. In the near future, investment firms with some form of AI involved in its investment process may prove to have a competitive advantage over firms that don’t.
Written by Jared Nussbaum & Samson Qian
Edited by Alexander Fleiss & Qilin Guo