Quantitative Value, by Wesley Gray

This is an outstanding book for anyone considering value -based investing. And for anyone who isn't, to help them decide that perhaps they should. It is quite expensive yet worth the price. (Follow this link for the $19.99 copy of the book, the best price available via the author.)

(It could, however, have benefited from another round of editing, especially by someone who thinks in terms of usability.)

Here's the central notion. Value investing, introduced at Columbia Business School by Graham and Dodd, and captured in their book Security Analysis, means taking advantage of occasional market inefficiency to buy good quality stocks that are momentarily under-valued. You buy them when they are cheap, and ride them up to profit.

A big problem though is that our emotional, subjective views of what makes for a good value buy just do not work well. The authors explain why (and Dr. Gray goes on to explain this in more and even more interesting detail in his later book, DIY Financial Advisor).

The authors motivate the idea that just as a pilot runs through a checklist before takeoff, so must investors take their decision making to a model and set of procedures. This makes it less likely that we sabotage a good strategy by trying to out-thing the algorithm.

The book then takes us through the steps of building a winning algorithm. It starts with Joel Greenblatt's Magic Formula approach and addresses its obscurity problems.

The authors build off of the Magic Formula to consider and objectively analyze a variety of metrics. For example, what is the best means to determine price as a value indicator? EBIT/TEV wins, and the authors show every step of how they evaluate it against other options.

In fact, they test everything, which is part of the fun of the book. And at the end provide a detailed checklist that takes you through every screen required to identify which stocks to buy. Full transparency here: you walk away with the ability to implement a model.

There are some pet peeves of mine that weren't addressed (or not sufficiently). I wasn't always clear on when the comparisons included dividend reinvestment (e.g., when comparing to S&P 500 Total Return which does). I believe dividends completely skew the apparent value of any methodology. Plus, a reinvestment model isn't realistic for individual investors who must pay taxes even on reinvested gains.

Another is the lack of short term charts. Yes, there's value in a very long back test. But once the 1974-2011 view is complete, it would be so very useful to see the visualization of rolling five-year results. Unlike an endowment, I have an end date. I'd like to understand the efficacy of the model in time periods that I can relate to my own life events. Wouldn't most folks?

Finally, I'd like to see the authors address selling. It isn't completely clear if the model completely turns over the holdings annually or if it is buy and hold (an example of where some more editing could clarify things). It would be great to have a discussion on sell algorithms. (For example: if a value stock over time achieves its efficient valuation it is no longer a value stock. If there's annual rebalancing that liquidates, is that appropriate? Might a sell algorithm indicate to hold an asset longer? How about assets that remain on the value list for more than one rebalancing; should it stay on without limit?

All in all though, I highly recommend this book.

Quantitative Value, + Web Site: A Practitioner's Guide to Automating Intelligent Investment and Eliminating Behavioral Errors