Henry Blodget

Will the "Value Effect" and "Size Effect" Persist? Do They Even Exist?

In my "dumbest column of the year" post, I suggested that DFA and other smart investors have capitalized on the "value effect" to design better passive funds.  Astute readers pointed out that the value effect is a theory, not a fact, and that it might soon disappear--leaving those who bought "value tilted" or "fundmental-weighted" funds holding an underperforming bag.

This is a complex topic, one that still produces violent arguments in the halls of academia and elsewhere.  Fama, French, DFA, and others attribute some of the superior long-term performance of value and small stocks over large growth stocks to higher risks: stocks get cheap because the companies are in distress (or riskier), and rational investors demand a higher return.  If this view is correct, the effects should persist over the long term, even if they disappear for decades at a time.

One contrary view, held by John Bogle and others, is that the value effect is merely a short-term trend that, like other such trends, will eventually become over-bought.  Having "discovered" that value stocks outperform growth stocks, this theory goes, investors will bid up the prices of value stocks, and this will reduce or eliminate future higher returns (or, more pertinently, result in lower returns relative to growth stocks). 

Common sense and analyses of past data suggest that the latter view--temporary phenomenon--is certainly valid over periods that most investors consider long-term: 1-5 years.  Other analyses, however, show that the value effect has, on average, been persistent over many decades and in many markets, validating the former theory, too.  The same can be said for the size effect.

In the coming weeks, I will assemble some of the most important work on this topic (and please feel free to weigh in with studies, thoughts, and comments, either in the comments section or via email).  For today, I will simply highlight some excellent charts produced by Index Funds Advisors.  These show the frequency with which small and value stocks have outperformed large and growth stocks (scroll down to the bar charts, Figs 9-9 through 9-16), as well as the relative performance of small vs. large over selected time periods (Fig 9-17).  The charts show that:

• Small value outperformed large growth in 58% of 1-year periods from 1927 to 2006.
• Small value outperformed large growth in 97% of 20-year periods over the same 80 years.
• Large value outperformed large growth in 58% of 1-year periods and 92% of 20-year periods.
• The "size effect" showed perfect decile by decile performance correlation over 80 years (smallest 10% outperform second-smallest 10%, etc.)  Over shorter--but still long--periods, however, this performance has completely reversed.
• From 1965-1968, 1975-1983, 1992-1993, and 2002-2006, the smallest 10% of stocks wildly outperformed the largest 10%.
• In all the interim periods, small stocks got absolutely crushed.

These charts illustrate why both camps in the "persistent" versus "temporary" camp have an important point.  Over the truly long-term (80 years), the value and small effects appear to be undeniable.  Over interim periods long enough to feel like eternity, however, small and value stocks get overextended and experienced painful reversion to (and beyond) the mean.

Should investors try to take advantage of either pattern? Both?  Should investors "tilt" portfolios toward value/small, but also try to time the mean-reversion by changing ideal portfolio weights depending on relative valuation?  The latter would depend on whether the mean-reversion can be meaningfully predicted...

March 15th 2007, http://www.investmentintelligencer.com/2007/03/will_the_value_.html