What Works on Wall Street: A Guide to the Best-Performing Investment Strategies of All Time [Hardcover]

The author went through the Compustat database to study the effects of eleven financial ratios on a hypothetical portfolio that was rebalanced once a year. The financial ratios or parameters are: market cap, price to earning, price to book, price to cash flow, price to sales, dividend yield, 1 year earning's growth, 5 year earning's growth, profit margin, return on equity, and price momentum. The study spans over 45 years. For each parameter he reports the portfolio return by year, and the return's arithmatic mean, standard deviation, geometric mean, Sharpe ratio, and how often it beats (or conversely underperforms) the broad market.

He then studies performance of portfolios that combine high or low values of several of the above financial parameters. And he does present some very interesting and useful results here.

Arguably one must already have some familiarity with various stock picking strategies, and some comfort with statistical analysis, to profit from this book. The book has two main weaknesses: (1) The author gives no reason to believe that the past performance will indeed guide future performance, and (2) The author gives no information on the turnover encountered by each portfolio strategy. Hence unless your investment portfolio is limited to your retirement account, you don't know whether following the conclusions from this study will really make you additional AFTER TAX money, compared to a low turnover S&P500 indexing strategy.

Nevertheless the book presents ORIGINAL reliable and FACTUAL informaion regarding how the US market behaved between 1951 and 1996. That in itself makes it more useful than most investment books. If you happen to be serious about investing in stocks, you simply can't afford to ignore these results.

The key findings of O'Shaughnessy are the slight superiority of returns for small cap stocks and larger excess returns for value stocks, and therefore the combination of both is ideal.

Of course, Ben Graham said this 70 years ago but approached the topic from the bottom up, while this book analyzes top-down.

I would however steer cleer of his idea about relative price momentum that he claims does 18.11% a year. Note that O'Shaughnessy started several funds, which did absolutely miserably and then he jumped ship having made a small fortune in book sales.

Read this book, but read Graham&Dodd as well.

The idea behind this book seems appealing: Statistical examination of stock market data over a long (>40yr) period, to determine what strategies work.
What you get, is someone tinkering with a spreadsheet and database to retrospectively determine what WOULD HAVE worked on Wall Street. Just like those guys selling sure-fire horse racing systems.
LAYOUT
Pick up this book and flip through it and you'll be surprised to notice that it's 330 or so pages are over 75% charts and tables. There are literally dozens of pages of charts of the year on year statistics the author tells us up front are useless. But the ones I found most irritating were the GIANT size bar charts spreeeeead over 3-4 pages with maybe 4 bars to a page showing what $10,000 dollars would have grown to over 40 years. Gimme a break!
STATISTICS
Good news is that you'll have no trouble understanding them. This material never gets beyond averages, and standard deviation. Oh, and someone showed him how to calculate the Sharpe ratio. Bad news is he doesn't seem to know enough stats to interpret the data.
BIBLIOGRAPHY
Rather suprising given that the book is almost all tables and charts by the author, is that he quotes 6 pages (hundreds), of sources. Who's he trying to kid? Virtually none of these are referenced in the text.
FINALLY
What finally conviced me that this is nonsense was this: The author makes much of the different performance between major stocks and the market as a whole over his 40 year analysis period. In every chapter separate results are given for major stocks vs the market. BUT right there on page 38 in table 4-4 we clearly see the results for majors vs the market steadily decreasing from +3.89% in the 50s to -0.96% in the 80s. This seems to be entirely unnoticed by the author!
CONCLUSION
Some of his comments on the psychology of Wall Street are interesting. His trading recommendations are a mixture of arbitrary decisions and standard ideas you'll get in other books (eg Ben Graham) such as:
High PERs are dangerous.
Consistent earnings growth is good.
High Sales to mkt cap is good.