Given the size effect and the value effect, it would appear that the worst investment style is large cap growth. I spent some time looking at returns on the French Fama research portfolios and both effects seem to hold, but with some interesting twists.

Firstly, the annualised returns for the period 1927-1006:

The size effect does not appear to be strong for growth stocks, but does appear to be strong for value stocks. This was a surprise as one of the commonest explanations for the size effect is that smaller companies are more likely to be high growth.

I think the explanation for this is twofold. Firstly, smaller companies have room to grow even in a low growth industry. Secondly, investor psychology. Growth investors tend to be adventurous types who are willing to buy smaller companies, value investors tend to be more conservative and prefer larger companies.

Almost every piece of advocacy for investing in smaller companies tends to talk about growth, and almost all advocacy of value investing pushes “established” companies. This leave small cap value neglected.

The value effect was also fairly consistent. Looking at the eight ten year periods covered by the data (which ran from 1927 to 2006 inclusive), value outperformed growth in all but two for large caps, and in all but one for small caps. Value also out performed growth in 50 out of 70 years for large caps and 51 for small caps.

The size effect was consistent for neutral and value portfolios, again holding for seven out of eight ten year periods. It held for 47 out of 70 years for the neutral portfolio and 50 out of 70 for the value portfolio.

The size effect was less consistent for growth portfolios, only true for five out of the eight periods and only 39 out of 70 years.

One rather odd result was that the combined effect of size and growth (i.e., comparing large cap growth to small cap value), was only moderately consistent: five out of eight ten year periods and 47 out of 70 years. However, in almost all periods where the small cap value portfolio was the under-performer, it almost always did only a little worse. The biggest exceptions are 1927-1931 and 1995-1998.

We have some evidence, we have credible explanations: forget large cap and growth and focus on small cap value.