Sketch notes on 07-May-2012 article on Greenbackd: How to beat The Little Book That Beats The Market: An analysis of the Magic Formula,
Krishnamsetty reckons MFI has annual alpha of 4.5%, not the 18% suggested by Greenblatt.
Empirical Finance couldn’t replicate the results either, and found the outperformacs was more like 2% for all-caps – but massive outperformance for small-caps.
James Montier found an 8.8% outperformance in UK, with less volatility.
Wellman and Vodel found EBITDA/EV to be the best performing price ratio. EBIT/EV yielded almost identical results. ROC may be adding virtually nothing.
Montier concluded EBIT/TEV produced higher returns than the blended approach in the MFI – but with high volatility. Low EBIT/EV did well in period to 1999, but then subsequently underperformed. From 2000 (to 2005), it was the high EBIT/EV that outperformed. The ROC component of the screen prevented the massive underperformance that was seen with the pure value strategy.
My 2 cents
Montier’s analysis is very interesting. It seems that ROC "smoothes out" your error between choosing growth or value strategies. In the period to 1999, growth (i.e. high EV/EBIT) growth was the winner, although the really significant outperformance occured from about 1997 to 1999. It is difficult to know to what extent a cautious investor would have captured this outperformance, because of course you only know the optimal exit time with hindsight. It also looks like growth works in secular bull markets, whilst value works best in secular sideways markets.
We see that Lynch ran his fund from 1977 to 1990, which appears to have been a particularly good time for growth. Contrariwise, Peter Cundill, after 20 years, was still only able to match the indices. Cundill was a value investor. It was only from about 2000 that he was able to massively outperform the indices.
This raises the point that maybe it’s not "value" that’s an inherently superior strategy, but a product of market conditions; and that an investor should try to be tactical than adhere to any one strategy.
(Edits made to correct some errors)
Powered by Blogilo