Today I am diving into an incredible piece of literature, not just any book, but one so dense in data and empirical evidence, that it can be seen almost as a 700 page research paper. The book I am talking about is the Fourth Edition of “What Works On Wall Street” by James P. O’Shaughnessy. The book covers a dataset of US stock market data that dates back all the way to 1926, covering a multitude of strategies and indicators that investors should use, or more importantly, which NOT to use. While the data is based on US stocks, these strategies are still useful in any market.
Even in the midst of crises and manias, we see that over the very long term, stocks tend consistently align with their fundamental value. This underscores the importance of studying history and how markets behaved during such times in the past. Or as Mark Twain stated: “History doesn't repeat itself, but it often rhymes.”. Therefore studying history and data from the past is something every investor should do to become an overall better investor.
🖊️ James Patrick O'Shaughnessy (born May 24, 1960) is an American investor and venture capitalist, currently serving as the CEO of O'Shaughnessy Ventures. He is the founder of O'Shaughnessy Asset Management, LLC (OSAM), an asset management firm that Franklin Templeton later acquired. O'Shaughnessy's areas of expertise include quantitative equity analysis, portfolio management, research decisions, and investment models.
My 🔑 takeaways:
All graphs, figures, tables, results, and strategies are from the book “What Works On Wall Street” by James Patrick O'Shaughnessy. The summary consists of multiple parts, following the structure of the book:
- The Data
- Strategies **
- The Overall Market
- Investing By Market Capitalization
- Market Leaders Outperform Large Stocks With Lower Risk
- Which Ratios Matter?
- Dividends, How Much Is Good?
- Combining Value Factors
- What Other Measures To Use?
- Looking At Price Momentum
- Which Sectors To Invest In?
- Which Strategy To Choose?
- Final Investment Lessons By O’Shaughnessy
1. The Data
Effective research begins with a robust foundation, which is the dataset. The data used in this book consists of monthly data on US publicly listed stock from 1926 to 2009. In previous editions the data consisted of a Compustat dataset dating back to 1963, thus this edition has a much more extensive database.
2. Strategies
Each of the strategies covered by O’Shaughnessy is tested on each decile, for example, the 10% of companies with the highest PE ratio, followed by the second highest 10%, and so on. Also, given the monthly nature of the data, each strategy is conducted in every month, with the final assessment being an average across the 12 portfolios. The total number of strategies covered are much too extensive to cover in detail throughout this brief summary, but I have made an attempt to highlight the most notable results.
3. The Overall Market
The simplest strategy of all might actually be “no strategy”. If $10,000 was invested on 1926, it would have been worth $23,171,851 in 2009, or a compound return of 9.78%. Even when accounting for inflation, $1,000 in 1926, or just over $10,000 in 2009 value, would have grown to over $2,000,000. Thus investing in the market without a true “strategy”, would still lead to a comfortable sum for retirement. Especially compared to how inflation would have eaten up more than 90% of the value. Now let us take a look into the growth when a real strategy is applied and how active investing can make a huge positive difference, when done correctly.
Comparative returns, December 31, 1926, to December 31, 2009
4. Investing By Market Capitalization
One of the most notable results discussed is that adding smaller stocks to the mix significantly enhances overall performance. Even though volatility increases for the lower deciles, it can be seen that every decile has a higher sharpe ratio than the first decile, thus only investing in the largest 10% of stocks is not the best strategy, even when volatility is lower. O’Shaughnessy also mentions that they have found that the strategies looking for maximum return inevitably lead to small- and mid-cap stocks since these are often least efficiently priced. Just for size reference, in 2009, 400 stocks made up 75% of the US stock market capitalization, whereas thousands made up the remaining 25%. This shows how many opportunities there are when it comes to stock picking when one is not restricted to the largest market cap stocks.
Summary results for market-cap decile analysis of all stocks universe, January 1, 1927, to December 31, 2009.
5. Market Leaders Outperform Large Stocks With Lower Risk
Even though small- and mid-cap stocks offer better returns, it does not mean that one can not outperform by investing solely in large-caps. One strategy mentioned is investing in market leaders, or large stocks on steroids. These are defined in the book as: non-utility stocks with an above-average market cap, cash flow, shares outstanding, and sales 50% higher than the average. It can also be seen from the table below that the small stocks outperform market leaders slightly, but both still outperform the S&P 500. Also, we indeed see, as mentioned, that investing in all stocks is a better investment than only investing in large stocks such as the S&P 500. Also, market leaders have the highest Sharpe ratio, indicating a very good risk-to-reward ratio and from 1964 to 2009, there are no periods larger than 7 years with negative returns.
Summary annual return and risk results data: all stocks, small stocks, large stocks, market leaders, and S&P 500; January 1, 1964, to December 31, 2009
6. Which Ratios Matter?
O’Shaughnessy further delves into individual performance metrics, analyzing the performance of each decile in detail. Frequently ratios are used, as investors commonly examine metrics like price-to-earnings, price-to-book value, and other similar indicators. While these ratios play a pivotal role in assessing investment opportunities, the challenge lies in deciding which factors truly hold significance.
The very famous price-to-earnings ratio, or PE ratio, is very popular, which is for a reason. As it is indeed found that the highest deciles (table is earnings-to-price) have higher average returns. Also consistently investing in stocks with the highest ratios will lead to a mere return of 9.37% with a much higher standard deviation, leading to a very low Sharpe ratio of just 0.02.
Summary results for earnings/price decile analysis of all stocks universe, January 1, 1964, to December 31, 2009
The same analysis is conducted for a number of other ratios such as EBITDA-to-enterprise value, price-to-cash flow, price-to-sales, and price-to-book value.
For the EBITDA-to-enterprise value ratio it holds that investing in the stocks with the lowest ratio yields very bad results.
Summary Results for EBITDA/Enterprise Value Decile Analysis of All Stocks Universe, January 1, 1964, to December 31, 2009
The opposite holds for the price-to-cash flow ratio where a lower ratio is rewarded by investors (table is cash flow-to-price). The worst decile did worse than T-bills, turning a $10,000 investment into a mere $48,471 from 1963 to 2009.
Summary Results for Net Operating Cash Flow to Price Decile Analysis of All Stocks Universe, January 1, 1964, to December 31, 2009
For the price-to-sales ratio they also find that a lower ratio is better (table is sales-to-price). However, it is very difficult to stay away from these stocks as they will often perform extremely well over a short period of time, eventually falling hard in the years after. For example, the group with the highest ratios returned 207% in the year ending in February 2000, but fell more than 73% the year after. Also from February 2000 to 2009, the group lost 91%. Thus, in the short-term these investments seem wonderful, they will make you feel like the new Wolf of Wall Street, yet these runs never last long.
Summary Results for Sales to Price Decile Analysis of All Stocks Universe, January 1, 1964, to December 31, 2009
For the price-to-book value ratio it is found that the lowest ratio is better (table is book-to-price), however, O’Shaughnessy also remarks that there are long subperiods in which low ratios do not work, especially the lowest 10%. They show exactly why a strategy should be analyzed over very long periods of time, as for example from 1967 to 1984 the group performed worse than US T-bills. Over the entire period from 1927 to 2009, the best performing deciles were actually decile 2 and 3, thus lower ratios performed much better.
Summary results for book-to-price decile analysis of all stocks universe, January 1, 1927, to December 31, 2009
7. Dividends, How Much Is Good?
Since stock dividends have historically accounted for more than half a stock’s return, investors often seek protection in stocks with high dividend yields. However, solely seeking the highest dividends is not a solid investment strategy. O’Shaughnessy covers the various dividend yields in deciles and the results show that buying the highest decile yields lower returns than deciles two, three, and four. Also, looking at changes in dividends can be a good performance measure, as is mentioned in the book:
“A study of dividend yields that my research team at O’Shaughnessy Asset Management (OSAM) conducted found that in the 4,722 instances when a company cut its dividend by between >0 and 50 percent, the stock’s average performance in the year following the cut was a modest –0.7 percent drop. However, in the 801 instances when a company cut the dividend by 50 to <100 percent, the stocks lost to their benchmarks by an average of 3.6 percent in the following year. The worst returns came from companies that suspended their dividends altogether; there were 3,329 instances when this occurred, and the stocks, lost an average 5.1 percent to their benchmark. On the other hand, stocks that increased their dividends (over the course of our study there were 46,358 instances when this occurred) beat their benchmark, on average, by 4.5 percent in the following year. The biggest winners were stocks that had paid no dividend but then started to pay one. In the 6,035 observations in which a firm initiated a dividend, the stocks, on average, went on to beat their benchmark by 9.2 percent in the following year.”.
Average annual compound return by dividend yield decile, All stocks universe, January 1, 1927, to December 31, 2009
Another measure used to value stocks is their buyback yield. Here the research clearly shows that the highest yielding stocks also have the best on average performance. Thus when both dividend and buyback yield are combined into shareholder yield, the highest deciles perform very well over time. O'Shaughnessy then concludes that “Both buyback yield and shareholder yield are good performers and should be considered by both value investors who want the best buy-back yield and shareholder yield and by growth investors who want to make certain to avoid the risks associated with the bottom deciles of each group.”.
Average annual compound return by buyback yield decile, All Stocks universe, January 1, 1927, to December 31, 2009
8. Combining Value Factors
O'Shaughnessy further combines the best performing value factors into a single measure called the Value Factor One, consisting of: price-to-book, price-to-earnings, price-to-sales, EBITDA-to-enterprise value, and price-to-cash flow. They also add shareholder yield and call the new measure Value Factor Two. They also add just buyback yield instead of shareholder yield and call this measure Value Factor Three. Each value factor takes into account the value factors and assigns points to each one. All strategies and their highest and lowest decile are shown in the figure below with their compound average annual rates of return and their standard deviations to indicate the riskiness of the strategy.
Compound average annual rates of return for the 46 years ending December 31, 2009, results of applying strategies on the All Stocks universe
Standard deviation of return for strategies from the All Stocks universe, 1964–2009 (higher is riskier)
O’Shaughnessy also shows that some low scoring companies on the value factors outperform in market bubbles, but states very clearly: “You must remember this the next time some hot sector is soaring and you are hearing all of the predictably short-term reasons why the old rules of valuation no longer apply. They always apply—and have for the entire time for which we have records of the stock market”. Therefore it is always important to look at long-term base rates, not short-term hype cycles.
Another statement worth highlighting about this segment is: “Value strategies work, rewarding patient investors who stick with them through bull and bear markets and through bubble and burst. But it’s sticking with them that is extraordinarily hard. Since we all filter today’s market performance through our decision-making process, it’s almost always the glamorous, high-expectations, high-ratio stocks that grab our attention. They are the stocks we see zooming up in price, they are the ones that our friends and fellow investors talk about, and they are the ones on which investors focus their attention and buying power. Yet they are the very stocks that consistently disappoint investors over the long term.”.
9. What Other Measures To Use?
What may come as a surprise, is that only looking at earnings gains is a losing proposition. That is why this book is such a good starter for those starting to value stocks, as controversial results may truly surprise the reader. As we see below, it is highlighted that for Large stocks, there is no real pattern in earnings per share deciles and performance, besides the lowest decile.
Average annual compound return by EPS Change (%) decile, Large Stocks universe, January 1, 1964 to December 31, 2009
Also, profit margins just by itself is not a good measure, as solely picking the highest decile even underperforms the market.
Average annual compound return by net margin decile, All Stocks universe, January 1, 1964, to December 31, 2009
Also return on equity is not the best measure, but O’Shaughnessy does state that it is best to avoid those stocks with the lowest ROE.
Average annual compound return by ROE decile, All Stocks universe, January 1, 1964, to December 31, 2009
10. Looking At Price Momentum
O’Shaughnessy also conducts tests on price momentum, as it often conveys different factors of a stock and is a better indicator than many other factors just by themself. They look at price momentum as the stocks with the highest price appreciation in the previous six- and twelve-month periods.
Summary Results for 6-Month Momentum Decile Analysis of All Stocks Universe, January 1, 1927, to December 31, 2009
Summary Results for 12-Month Momentum Decile Analysis of All Stocks Universe, January 1, 1927, to December 31, 2009
The average returns over large periods of time, from 1927 to 2009 are incredible, turning $10,000 into a whopping $572M+ for six-month price momentum or $156M+ for twelve-month price momentum. O’Shaughnessy concludes this chapter by an overall strategy: “Thus, linking pure price momentum to trading volume is a strategy worth considering. By adding points if the stock is in the seventh to tenth decile by average trading volume and subtracting points from the deciles with the highest trading volume, particularly deciles 1 and 2, you penalize `stocks with the highest trading volume and reward those with lower trading volume. This might allow you to smooth out your ride if you are using price momentum exclusively to make stock selections.”.
Now combined with shareholder yield they find one of the best performing strategies thus far. The strategy consist of: starting universe is the Small Stocks universe, three-month price momentum must exceed the median for the Small Stocks universe, six-month price momentum must exceed the median for the Small Stocks universe, then buy the 25 stocks with the highest shareholder yield. The strategy outperforms the small stock universe in almost every year.
Five-year average annual compound excess (deficient) return Small Stocks, 3 and 6 mo. >med, top 25 SY minus Small Stocks; January 1, 1928, to December 31, 2009
Also combining price momentum with price-to-book ratio creates the following strategy: market capitalization between a deflated $50 million and $250 million at time of purchase, price-to-book (or, book to price, BP) ratio in the three cheapest deciles (i.e., lowest 30 percent of the universe by price-to-book ratio), three- and six-month price appreciation greater than zero, and buy the 25 stocks with the best 12-month price appreciation. This strategy compounds annually at 18.16% according to O’Shaughnessy. The only downside is the high volatility, where the strategy had one 10-year period with a negative return of 7.23% per year.
O’Shaughnessy also shows that the strategies work in the all stock universe. O’Shaughnessy states: “Between 1928 and 2009, there are two strategies from the All Stocks universe that have perfect ten-year base rates. The first strategy requires that stocks have a book-to-price greater than the median (i.e., eliminate the half of the stocks where investors are paying the most for every dollar of book value); have six-month price appreciation greater than the median; and finally, buy the 25 stocks with the highest shareholder yield. Not only does this strategy beat the All Stocks universe in 100 percent of all rolling ten-year periods, but it also beats the universe in seven out of every ten one-year holding periods. There is a price for the consistency however, as the strategy earned an average annual compound return of 15.32 percent, some 61 basis points behind the best-performing strategy. It also has a larger maximum decline of 82 percent and a lower Sharpe ratio. Nevertheless, a 100 percent base rate over all rolling ten-year periods between 1928 and 2009 is a rarity.”. Therefore the strategy is worth highlighting.
11. Which Sectors To Invest In?
O’Shaughnessy also highlights multiple strategies per sector, however there are too many to cover here. What is important is that not every strategy works as well in each of the sectors, thus every investor who invests solely in certain sectors should check which ones apply best to their sector. Also, a significant result from the data shown below, shows that technology and telecom were not the best sectors to invest in.
Sectors Ranked by Standard Deviation of Return, Least to Most Risky, December 31, 1967, to December 31, 2009
Or as O'Shaughnessy states: “Thus, appearances can be deceiving. If you asked the average investor what the best-performing sector was over the last 42 years, my guess is that the majority would say technology or telecom. I would seriously doubt many would guess consumer staples, but when you think about it, it does make sense. Industries that make goods and services that people have to buy, regardless of economic circumstances, are bound to do well whatever the economic conditions. The consumer staples sector is also filled with companies that have wide moats and world-recognized brands. Two of the biggest business advantages that company’s can have are monopoly power or brand power with wide moats. While I am fairly certain that there are more than a few MIT students who are also working on schemes to disintermediate the tech powers-that-be, I doubt that there is a group of friends in a garage cooking up the next soft drink to knock Coke off its pedestal.”.
12. Which Strategy To Choose?
For a further deep-dive into the strategies that can be used for both growth and value investors I refer to the book Chapter 25: “SEARCHING FOR THE IDEAL GROWTH STRATEGY”, Chapter 26: “SEARCHING FOR THE IDEAL VALUE STOCK INVESTMENT STRATEGY”, and Chapter 27: “UNITING THE BEST FROM GROWTH AND VALUE”. In these chapters all value factors, ratio’s, and other performance measures are combined to form the ultimate strategy for both growth and value investors. They reach five of the best and worst strategies, which are really neat since they can be replicated by purchasing just a small number of stocks with an initial investment amount of $10,000.
The five strategies with the highest absolute returns (duplicate strategies eliminated), August 31, 1965 to December 31, 2009
The five strategies with the worst absolute returns, August 31, 1965 to December 31, 2009
What is also very important, mentioned by O'Shaughnessy is: “In the real world, many investors check their portfolio’s value daily and let the daily ups and downs inform their decisions, usually for the worse. In the real world, investors are far more frightened of short-term volatility then any rational economic model would suggest, but that very real fear must be accounted for in determining which strategy will be right for you. I have watched investors’ reactions to short-term volatility over the last 14 years, and I can tell you that it is far more predictable than the markets’ performance.”. Therefore finding the right strategy should not only be based on return, but also on risk, maximum drawdowns, base rates, financial goals, and more.
O’Shaughnessy concludes with two strategies that are overall best:
“After weighing risks, rewards, and long-term base rates, two strategies emerge as the best overall—one for investors willing to take market risk and one for very conservative investors.
The first is the Trending Value strategy that combines the best of value and the best of growth. It selects stocks from the All Stocks universe in decile 1 of Value Composite 2, and then buys the stocks with the best six-month price appreciation.
The second is the combined portfolio that draws from the consumer staples and utilities sectors and then buys the 25 stocks from consumer staples with the highest shareholder yield and the 25 stocks from the utilities sector with the best scores on Value Composite 2.
These two strategies consistently ranked at or near the top when ranked by absolute return, risk-adjusted return, downside risk, and maximum decline.”.
13. Final Investment Lessons By O’Shaughnessy
Finally, O’Shaughnessy has a few key lessons every investor should take into account.
- Always use strategies
- Ignore the short term
- Use only strategies proven over the long term
- Dig deep (study a lot)
- Invest consistently
- Always bet with the base rate (knowing how often a strategy beats the market)
- Never use the riskiest strategies
- Always use more than one strategy
- Use multifactor models
- Insist on consistency (if you do not invest yourself, make sure your manager is consistent)
- The stock market is not random
Those investors who take into account these lessons will do very well in the long-term. I truly want to thank O’Shaughnessy and the team for these wonderful insights. What was shared in this summary is all backed by their data and studies, which can be found back in “What Works On Wall Street”. I strongly recommend this book for those interested in the results in more detail and various insights into each strategy for various sizes, sectors, and investment types.
💭 “Investors can learn much from the Taoist concept of Wu Wei. Taoism is one of the three schools of Chinese philosophy that has guided thinkers for thousands of years. Literally, Wu Wei means “to act without action,” but in spirit it means to let things occur as they are meant to occur. Don’t try to put square pegs into round holes. Understand the essence of a circle and use it as nature intended. The closest western equivalent is Wittgenstein’s maxim: “Don’t look for the meaning: Look for the use!”
For investors, this means letting good strategies work. Don’t second guess them. Don’t try to outsmart them. Don’t abandon them because they’re experiencing a rough patch. Understand the nature of what you’re using and let it work. This is the hardest assignment of all. It’s virtually impossible not to insert our ego or emotions into decisions, yet it is only by being dispassionate that you can beat the market over time.”, by James O’Shaughnessy in “What Works On Wall Street”.