What are some quantitative approaches to value investment?

Question

As a developer and statistician, I consider value investing to be a statistically sound investment strategy. I've read a few books on the area but I am still not clear on valuation measures. So I would like to hear from experts in quantitative finance about the approach and its methods. For example, I would like know the pros and cons of value investing and what valuation methods are out there. Any opinions or references would be great, too. Thanks.

Answer 1

Value has traditionally been one of the most important stock-selection signals for quantitative managers. However, since the late 2000s, following a rapid flow into quantitative investing, traditional value strategies have lost most of their predictive power and the returns generated from them have also become more volatile.

The typical approach of quantitative managers is to distill a valuation measure into a "factor". This is often done using the hedged portfolio approach pioneered by Fama and French (1993). In this approach, analysts divide the investable universe into quantiles (typically in quintiles or deciles) to form quantile portfolios. Stocks are either equally weighted or cap weighted within each quantile. Each quantile portfolio’s performance is then tracked over time. A long/short hedged portfolio is typically formed by going long the best quantile and shorting the worst quantile.

A few of the most common valuation factors are:

• Dividend yield
• Earnings quality (accruals)
• Earnings yields / P/E ratios
• Price-to-book
• Return on Equity

Many of these also come in all sorts of simple or sophisticated variations. A few related factors that are not strictly valuation-based are:

• Earnings surprise
• Analyst recommendations
• Short interest
• Insider ownership
• Turnover (volume)

These are just a few, and practically every paper published in the broad field of Behavioral Finance - Asset Pricing has been turned into a quantitative factor by someone somewhere.

Authors Ludwig Chincarini and Daehwan Kim, in their book, Quantitative Equity Portfolio Management, make a very interesting comparison of quantitative and fundamental investing:

It is inaccurate to say that fundamental managers dig deep at the solo stock level, but have no models or disciplines. It is also unfair to say that quantitative managers apply skills to so broad a set of stocks that the process is superficial at the fundamental level, and often labeled black-box, data mining nerds. This is a misrepresentation. Many quantitative investment strategies rely on factors that are based on not only solid economic principles, but also on sound fundamental intuition. At the same time, fundamental managers all use models. These may be rules-of-thumb or heuristics, and not subject to rigorous testing, but the deep implementation of the model into the security makes up for the lack of breadth. To repeat, quantitative management – lies in broadly perfecting the comprehensive portfolio system, whereas, fundamental management lies in deeply comprehending the perfect stock selection.

J.P. Morgan's US Factor Reference Book provides a good (and very lengthy) overview of the state of the art. Lots of other major sell-side institutions also have their own stock selection models, such as Deutsche Bank, Macquarie, Bernstein, and Barclays. It is not always easy to access information on their products on the public internet.