Buffett’s Alpha (2018)
Andrea Frazzini, David Kabiller, Lasse Heje Pedersen
Financial Analysts Journal 74(4), URL
In this week’s AGNOSTIC Paper, the authors use the major factor premiums to examine one of the best long-term investment track records in the world – Warren Buffett and Berkshire Hathaway. The latest annual report just came out a few days ago and (as usual) summarizes Berkshire’s performance on the first page. From 1965 to 2022, Berkshire returned 19.8% per year versus 9.9% for the S&P 500. That’s a 24,708% cumulative return for the S&P 500, and an unbelievable 3,787,464% return for Berkshire. There are some investors who achieved even better results over shorter time periods. But to the best of my knowledge, there is no 58-year track record that is even remotely comparable to Buffett.1For example, Edward Thorp or Renaissance Technologies achieved better risk-adjusted returns. But (so far) over much shorter time horizons and not at the scale of Warren Buffett’s >$200B equity portfolio.
Before we move on, let’s start with a very important note: neither this post nor the paper aims to discredit Buffett’s or Berkshire’s performance. What they did over the last six decades is absolutely outstanding and deserves incredible respect. Both financially and humanly. However, our understanding of equity markets also increased since the early years of Berkshire. With today’s knowledge, the authors explain that “[…] Buffett’s returns appear to be neither luck nor magic but, rather, a reward for leveraging cheap, safe, high-quality stocks.” But let’s dive into it step by step.
Everything that follows is only my summary of the original paper. So unless indicated otherwise, all tables and charts belong to the authors of the paper and I am just quoting them. The authors deserve full credit for creating this material, so please always cite the original source.
Setup and Idea
Given their track-record, Warren Buffett and his partner Charlie Munger unsurprisingly became the heroes of countless “value investors”. Full disclosure: I am one of them. I got interested in investing and finance after reading a book about Buffett and Berkshire before going to university. While this was my starting point back in 2015, I gradually moved to the more systematic and data-driven philosophy you know from this website. Nonetheless, I never gave up some essential parts of his philosophy.2Only invest in what you understand, be as rational as possible, regard stocks as fractional ownership in real companies, never stop learning, never depend on the kindness of a stranger, and some more.
This week’s paper is therefore very special, in the sense that it brings the two worlds together. It is no secret that Buffett and Munger, as well as many of their disciples, disagree with some ideas of academic finance and occasionally even make fun of them. Their success obviously proofs them right. However, I believe we are living in a very different world today. We know much more about equity markets than back in Buffett and Munger’s first decades with Berkshire and what was a distinct competitive advantage for them is nowadays quite well known. I mean, you hear so many people talking about long-term investing in great companies that it really became mainstream. That doesn’t mean it cannot work going forward. However, today it is probably much harder because all the Buffett-disciples compete in the same area. In fact, that’s a point Buffett and Munger themselves stressed several times.
Before we go into the paper, let me share a personal story to show you where my agnostic approach of both the Buffett and the “scientific” philosophy comes from. In late 2021, I received a prize for my master thesis from a German asset manager which mostly follows the concentrated Buffett-business-picking-philosophy.3To their credit, they also incorporate a lot of modern tools like alternative data, machine learning, and other analytics to make the traditional approaches better. Naturally, they are quite skeptical about academic finance and factor investing. I presented my thesis in a seminar which mostly consisted of such non-quant value portfolio managers. They talked about business models, management teams, competitive advantages, valuations, and so on. In addition to that, there were some quant practitioners and academics who also received the prize in other categories.
Expecting such an audience, I created my presentation accordingly and mostly illustrated the ideas of my thesis at the example of one particular company. In particular, I left out most of the geeky academic stuff. The other quants and academics didn’t do that and pretty soon there was a discussion about the Price/Book multiple for a value strategy. To make the point, I am exaggerating a little and I don’t want to offend anyone but the debate went something like this:
Non-Quant Value Manager: “Price/Book is no longer the right multiple because intangible assets are more important nowadays. Your models are too simple.”
Quant / Academic: “Valid point, but the variable still explains a lot of the cross-sectional variation in expected returns of quintile portfolios. We want to make a diversified bet on the general theme that cheap stocks outperform expensive ones. We don’t need a perfect valuation for that.”
Non-Quant Value Manager (somewhat dismissive): “But you still leave out so much information. To make money, you need to understand the company and buy it when it sells for less than its fair value. Ranking stocks by one variable doesn’t help here.”
The funny thing is: although the two characters disagree, both are right. They follow the same philosophy, fundamentally cheap stocks beat expensive ones, but implement it very differently. You probably shouldn’t put the two in the same company, but there is generally nothing wrong with their approaches. Applied correctly, both can lead to the desired outcome (outperformance). Think of it as two persons who both want to stay in good shape. One of them prefers to run while the other cycles. Two different approaches but the underlying philosophy is the same. To stay in good shape, you need to exercise.
As we will see in this week’s paper, Buffett and factor investors are no different. Both follow the same underlying ideas and just implement them very differently. To their credit, however, Buffett and Munger discovered those ideas much earlier than the research community and became billionaires while the rest was thinking about it.
Data and Methodology
To analyze Berkshire’s performance, the authors use the standard databases for the US stock market (CRSP and Compustat). For Berkshire-specific information, they use the well-known 13F filings to get the stock holdings and they also hand-collect relevant information from Buffett’s letters and the annual reports. Finally, they use the three factor model from Kenneth French’s website, the Betting Against Beta factor from Frazzini & Pedersen (2014), and the Quality Minus Junk factor from Asness et al. (2019).
The final sample ranges from 1976 to March 2017. This is of course a lot of Berkshire’s history, but not everything. Looking at the most recent annual letter, there were some pretty extreme years before 1976. For example, a 77.8% return in 1968, a 80.5% return in 1971, or a 48.7% loss in 1974. Overall, Berkshire returned 231.5% from 1965 to 1975 versus just 57.5% for the S&P 500 (11.5% vs. 4.2% per year, respectively). As strange as it may sound, these excess returns are relatively small compared to the subsequent decades. The authors’ sample may therefore slightly overestimate Berkshire’s actual performance.
Important Results and Takeaways
How good is Berkshire? Damn good…
Buffett’s track record is by all means impressive. As mentioned before, Berkshire returned 19.8% per year since he took over the company in 1965. For the authors’ sample from 1976 to March 2017, the number is 18.6% per year in excess of T-Bills (versus 7.5% for the US market). These returns also came with considerably higher volatility of 23.5% per year vs. “just” 15.3% for the US market.
I deliberately haven’t used the word risk in the previous sentence as it is well-known that Buffett don’t believes that volatility equals risk. However, if you look at some of its drawdowns (which in my opinion are a good measure for risk, or at least pain), it also becomes clear that Berkshire was anything but risk-free. For example, Berkshire lost 44% from June 1998 to February 2000 while the US market returned 32%. That’s an underperformance of 76%-points which would have been a career-disaster for almost any professional fund manager. In addition to that, I think it is really difficult for most people to stick with a stock through such a period.
With the benefit of hindsight, we know that Buffett’s reputation, his fortress-like setup with Berkshire, and his special shareholder base allowed him to stay course. In my opinion, these factors (which Buffett created deliberately!) are a very important and often underestimated part of his success. You can be the best long-term stock picker in the world, but if a 2-year drawdowns kills your operation, that’s worth nothing.
Despite higher volatility and some drawdowns, the sheer size of Buffett’s returns lead to strong risk-adjusted performance. The authors compare Berkshire’s Sharpe and information ratio to the universe of US stocks and US mutual funds that also exist since 1976. Berkshire ranks first in both categories. Even for shorter time periods, Berkshire always ranks within the first percentile of stocks and funds. There is a saying on Wall Street, “You may beat the house, but you cannot beat Warren Buffett.”. It seems to be true and the following charts summarize his outstanding results.
The Buffett Style: cheap, high-quality, and low-risk
Now to the more interesting part. How did Buffett achieve these returns? We all know that Berkshire is a collection of both public stock holdings and private companies. In a first step, the authors decompose Berkshire’s returns among the two and find the public stocks to be more important. I personally don’t think we need this distinction. It is well-known that Buffett applies the same principles to all of his investments. Ultimately, the value of his private holdings should also be reflected in Berkshire’s stock price. Therefore, I will focus on the analysis of Berkshire’s returns and mostly ignore the distinction between public stocks and private holdings.
The authors use the major factors – Market (MKT), Size (SMB), Value (HML), Momentum (UMD), Low-Risk (BAB), Quality (QMJ) – to examine Berkshire’s returns and summarize the results in the following table. If we only control for the overall stock market, Berkshire has a beta of 0.69 and an annualized alpha of 13.4%. Going to the Fama-French three factor model (plus momentum) shrinks the alpha to 11% per year and reveals significant loadings on Size and Value. Both are fully consistent with what we know about Buffett: a value investor with preference for large caps. Notably, he has no exposure to the momentum factor which is again consistent with his fundamental business-picking philosophy.
Now it gets interesting. In the remaining two columns, the authors successively add the BAB and QMJ factor. Buffett’s alpha shrinks further to 8.5% but remains significant after adding BAB. However, including both BAB and QMJ renders the alpha insignificant. Looking at the bold coefficients, we find positive factor loadings for the Market, Value, Low-Risk, and Quality. This 6-factor model still explains only 29% of the variation in Berkshire’s stock returns. But as we will see below, this is sufficient to reconstruct its success.
Interestingly, these results are very consistent with what Buffett and Munger explained over and over again. Buffett often praises the american tailwind (positive beta on the US market) and explains that he looks for predictable businesses (Low-Risk) with durable competitive advantages (Quality) that sell for reasonable prices (Value). Once again, this shows that Buffett and Munger follow the same underlying ideas as factor investors. They just did it earlier and with a much more concentrated portfolio.
Don’t practice what you preach – Buffett’s Leverage…
Before we come to the systematic replication of Berkshire, there is one further ingredient to its success: leverage. The authors estimate that Berkshire operated with an average leverage of 1.7 to 1, meaning Berkshire’s total assets were 70% larger than its equity. The authors acknowledge that this is just a rough estimate and subject to measurement error, but it is generally known that Buffett uses leverage in his operation.
Apart from standard bonds with the excellent credit rating of Berkshire, insurance float is by far the most important source of Buffett’s leverage. Berkshire owns a variety of insurance businesses and Buffett continuously invested the collected premiums in his stock holdings and private companies. Together with a disciplined underwriting policy, this has been (and still is) a very convenient and particularly cheap access to leverage. In fact, the authors estimate that the cost of Berkshire’s leverage from insurance float was about 3%-points below the T-Bill and 4.7%-points below the 10Y Treasury. Buffett and Munger therefore enjoyed cheaper debt than the US government.
Apart from bonds and insurance float, Buffett also occasionally used other creative ways to get cheap capital. For example, he uses some accelerated tax-depreciation rules and sold long-term out-the-money put options during the financial crisis 2008/09 (which never got exercised). Overall, these actions lead to a sizable amount of very cheap leverage within Berkshire and explain some of its higher volatility. Interestingly, this is not in line with what Buffett often tells the public…
If you don’t have leverage, you don’t get in trouble. That’s the only way a smart person can go broke, basically. And I’ve always said, ‘If you’re smart, you don’t need it; and if you’re dumb, you shouldn’t be using it.’Warren Buffett on Leverage, URL
Given his own use of leverage, the better way to say it is probably, “If you’re smart, you don’t need it. But it if you can handle it, it of course helps.” In any case, the authors explain that the access to cheap leverage and Buffett’s ability to avoid the associated problems are another important aspect of Berkshire’s success.
Systematizing Buffett and Berkshire
In the last part, the authors finally replicate Berkshire’s returns with the previously mentioned factor premiums. Specifically, they create a hypothetical multi-factor portfolio that invests in theoretical long-short factors as governed by Berkshire’s loadings. Their portfolio further matches Berkshire’s exposure to the US market and its volatility. The following charts summarize the performance compared to Berkshire, just its public stock holdings, and the US market.
Theoretically, the Buffett-Style portfolio performed even better. However, there are a number of caveats. First, this is an ex-post analysis. The authors act as if they know BRK’s factor loadings, market exposure, and volatility between 1976 and 2017 at any point in the sample. Obviously, that’s a big look-ahead bias. To their credit, the authors are well aware of this and mention that their analysis is rather descriptive. In the appendix, however, they show that their methodology also works out-of-sample with the appropriate time lags. So while this chart is probably too optimistic, it is generally possible to replicate Buffett’s success with the major factors.
Second, the factor portfolios are very theoretical and in this form not investable. These are hypothetical long-short strategies without considering trading costs, short-sale constraints, or any other real-world issues. Also note that the Buffett-Style portfolio ran at a leverage ratio of 5:1 which is probably hard to sustain in practice. Berkshire returns, in contrast, are a real-world track record after all fees.
Skeptics of academic finance and factor investing may be happy now. Yet another beautiful backtest that doesn’t work in the real-world… Although true, this is not the point here. As I mentioned before, neither I with my post nor the authors with their paper want to play-down Berkshire’s impressive results. He has the track record and applied the stuff before the rest of the world caught up.
However, with today’s knowledge of equity markets, we know where his returns come from and can even quantify the underlying drivers empirically. And in the interest of a balanced argumentation, levered long-short factor portfolios are implementable in today’s markets. So with enough sophistication, it is now possible to invest in such a Buffett-Style portfolio going forward. And this is what ultimately counts for investors, right?
To summarize the most important points, the table below again shows Berkshire’s performance compared to the Buffett-Style portfolios. There is also a Buffett-Style Long-Only Portfolio which doesn’t require the sophistication of shorting and leverage. While beating the US market over the full period, it lags Berkshire’s returns by about 10%-points per year. This again shows that all of the arguments only apply for the future and don’t question Buffett and Munger’s genius in creating the probably best long-term track record the world has ever seen.
Conclusions and Further Ideas
This week’s paper is (in my opinion) a very important one. It combines two areas of the investment world which still operate mostly independent from each other: “traditional” fundamental stock picking and systematic factor investing. I think the example of Buffett shows impressively that the underlying ideas are actually very similar and there is no need to fight about it. We can just agree that the two camps implement them differently.
With respect to Buffett and Berkshire, I believe there are many lessons which are also very relevant for all other investment strategies (and in some sense, life in general)…
- Being smart and having the best strategy without proper implementation is worthless.
- Even the best investors have painful drawdowns. Creating a setup that allows to survive such periods is key.
- Convenient access to cheap leverage helps to boost a good strategy even further but only if you can avoid the associated dangers of getting killed in a crisis.
- Don’t aim too high and rather focus on consistency. Compared to some hedge funds, Buffett’s Sharpe ratio of 0.79 appears not gigantic but made him one of the richest people in the world. We should keep that in mind when some fund managers aim for Sharpe ratios >1.
- Evolve or die. Buffett’s returns are outstanding but with today’s understanding of markets, we know where they came from and can replicate them with modest data crunching. Success attracts competition and I believe it is unlikely (but not impossible!) that someone will build a second Berkshire by just copying what Buffett and Munger did over the last sixty years.4You will most likely not become the next Apple by launching a smartphone in 2023. There is simply too much competition.
Finally, I agree with the authors that there is no better way to conclude than the following quote from Buffett’s 1994 letter: “Ben Graham taught me 45 years ago that in investing it is not necessary to do extraordinary things to get extraordinary results.” Although we can debate whether it is really “non-extraordinary” to find three profitable factors in the stock market before many others…
- AgPa #72: Machine-Reading of Private Equity Prospectuses
- AgPa #71: Go Where the Earnings (per Share) Are
- AgPa #70: Equal vs. Market Cap Weights
- AgPa #69: Rebalancing Luck
This content is for educational and informational purposes only and no substitute for professional or financial advice. The use of any information on this website is solely on your own risk and I do not take responsibility or liability for any damages that may occur. The views expressed on this website are solely my own and do not necessarily reflect the views of any organisation I am associated with. Income- or benefit-generating links are marked with a star (*). All content that is not my intellectual property is marked as such. If you own the intellectual property displayed on this website and do not agree with my use of it, please send me an e-mail and I will remedy the situation immediately. Please also read the Disclaimer.
|1||For example, Edward Thorp or Renaissance Technologies achieved better risk-adjusted returns. But (so far) over much shorter time horizons and not at the scale of Warren Buffett’s >$200B equity portfolio.|
|2||Only invest in what you understand, be as rational as possible, regard stocks as fractional ownership in real companies, never stop learning, never depend on the kindness of a stranger, and some more.|
|3||To their credit, they also incorporate a lot of modern tools like alternative data, machine learning, and other analytics to make the traditional approaches better.|
|4||You will most likely not become the next Apple by launching a smartphone in 2023. There is simply too much competition.|