The Long-Run and International Evidence on the Value Premium

The term "Value Premium" refers to the empirical observation that firms with low price multiples (i.e Price/Book, Price/Earnings, Price/Cash Flow) have tended to have higher returns than their high-multiple counterparts - even after controlling for risk. People give a lot of possible reasons for this - we have a bad model for controlling for risk, there are behavioral biases, or it's simply a case of data diving.

I just came across a paper by a group known as the Brandeis Institute titled "Value vs. Glamour: A Global Phenomenon" that seems to rule out the data diving story. They examine the evidence for the value premium both across time (the mid 1960's to the present) and internationally. They found that

While the degree of outperformance of value stocks vs. glamour stocks varied across data sets, what strikes us as most significant was the consistency the value premium exhibited:

• across valuation metrics, such as price-to-book, price-to-cash flow, price-to-earnings,and sales growth
• across time, which in this study applies to the 1968-2008 period for U.S. stocks,and the 1980-2008 period for non-U.S. stocks
• across regions, as the results indicated a value premium in developed markets in North America, Europe, and Asia
• across market capitalizations, as the relative outperformance of value stocks to glamour stocks was evident among both large- and small-cap stock universes.
  

The paper has a lot of nice graphs that could be useful in class. You can read the whole thing here.

HT: CXO Advisory Group