The world of quantitative finance is never boring. At least not from the research perspective. Today I want to explore some new research from one of the top quant shops on the small company effect and see how an investor can apply it simply to an existing quant approach. Specifically, I want to investigate if adding a quality screen to a small cap quant value portfolio increases performance.

AQR is one of the top quant shops in the business. They do some great quant research. Their latest paper, besides having one of the best titles I’ve seen for a finance research paper – “Size matters if you control your junk”, investigates the small company effect to see if it’s real or not. Despite being one of the original factors in the original efficient market hypothesis it’s come under much scrutiny as of late with many going so far as saying it doesn’t really exist. You can read all the gory details in AQR’s paper if you like but the general summary is that the small company effect is real and very strong especially once you control for quality. Low quality stocks really kill small cap performance. Lets take that as a given and see if this can help us lowly individual investors in any way. The summary graph from the paper is presented below.

AQR Control Your Junk Summary Table Jan 2015

A while back I posted on how adding a quality metric can increase the performance of value portfolios. I took a quant value screen and added a simple debt filter to screen out companies that increased their debt significantly. This simple change added over 4% of alpha to the already stellar performance of the quant value portfolio over the SP500. Based on AQR’s paper I wanted to see of I took my quality screen a bit further and limited the universe of companies in the value screen to small cap companies. Now, the work done in AQR’s paper to define what quality means is quite extensive. I’m in the process of trying to build a detailed ranking system in P123 that uses most their quality factors. But before I dove into that I wanted to see if a simpler quality ranking system would work.  In my experience if these effect are robust then even rough estimates of the parameters work quite well. Since I already had it, I decided to use a ranking system based on Piotroski factors. The F score that Piotroski pioneered is quite famous in value investing circles and does make up very strong, and very concentrated, value portfolios. I simply used some of the fundamental parts of the Piotroski system that look for financial strength. I ranked companies by improvement over the last 12 months in the following parameters: gross margin, operating cash flow, debt to total assets, current ratio, asset turnover, return on assets, and equity issuance. All factors were equally weighted. It’s a rough estimate of stock quality. Let’s see how it all turns out.

Small cap value plus quality quant port performance Jan 2015

Well, not bad at all. Pretty much confirms what the AQR paper is saying. Small cap stocks outperform quite strongly. And the results of this simple backtest shows that the effect is just as powerful when combined with a value factor. Now I’m thinking it is definitely worth the effort to try and re-create the quality metrics presented in the AQR paper. More to come on that front in the future.

In summary, the small cap effect is very strong as shown in the new AQR paper. And when combined with value and quality factors the resultant performance of the portfolios is quite impressive, far outpacing their respective benchmarks.

 


4 Comments

Andrew · January 27, 2015 at 1:02 pm

That’s amazing. Thank you for following up on the paper with your post. You distill a lot of information and double-check it, which is great.

If you ever have the time, could you test adding a momentum filter? It would be neat to see a strategy combining everything (value, momentum, size minus junk) into one!

    paul.novell@gmail.com · January 27, 2015 at 1:35 pm

    Thanks Andrew. As they say, it’s in there…Table 8 of the AQR paper does exactly that, add momentum to the mix. Basically, adding momentum to the mix increase performance by a little bit. This effect has been getting smaller over time and is practically statistically insignificant since 2000. But that could be a temporary anomaly. I have found the same. I did look at small caps vs all stocks in my trending value portfolios (value plus momentum) and small caps do worse than all stocks with or without a quality filter. Having said that adding a quality filter to trending value to weed out the really bad stocks does increase performance.

    Paul

Comments are closed.