Quant investing: building a better index

Today I wanted to take a look at how you can use quant investing to build a better stock market index. For my previous posts on quant investing see this series of posts.

In a way stock market indices are quantitative models. And due to history and other factors they are not very well constructed. For example, the Dow Jones index is comprised of 30 stocks weighted by the price of the stocks! Kind of silly in today’s day and age no? How about the grand daddy of the indices, the SP500? It’s much better than the Dow but still has some odd construction. Here’s is some of what it takes for a company to be in the SP500 (you can find all the requirements here):

  • Market capitalization greater than $4.6B
  • US companies only
  • Sum of 4 quarter earnings > 0
  • No BDCs, partnerships, LPs, MLPs, ADRs
  • Weighted by market capitalization
  • Addition to index decided by SP500 index committee

Basically, the SP500 is a quant index with some discretionary input from the committee. A few of its big limitations are the exclusion of foreign companies trading in the US, the exclusion of certain classes of companies all together, and the weighting of the index by market cap. We can do better by correcting some of these limitations and adding a value factor to the mix. As always, my reference for all this is O’Shaughnessy’s What Works On Wall Street which I’ll refer to as WWOWS.

WWOWS creates an all stock and a  large stock index, in the spirit of the SP500, by broadening the definition of an index to include small cap stocks, and the large cap index a bit looking for companies with a market cap greater than the database mean. It also allows for ADRs (foreign companies), and all company types. It then equal weights the companies instead of market cap weighting. Equal weighting has been shown to outperform market cap weighting consistently. See here. Then we’ll take it further by adding a value factor screen to the various stock universes. Now for the value tilt.

Dividends have been a reliable value factor in the stock market over time as I’ve discussed on this blog many times. I’m still a huge dividend fan but I’ve updated my views somewhat. Shareholder Yield, which adds stock buybacks to dividend yield, is a better value metric that dividend yield alone. If you want to totally geek on on Share Holder Yield see this recent study. So, we’ll sort the stocks in our new index by Shareholder Yield. These new ‘indexes’ are re-balanced to equal weight and sorted by value once a year. Lets see how all this works out in the table below.

Shareholder Yield Quant Strategy Performance May 2014

The table is sorted by risk adjusted returns, aka sharpe ratio. As you can see, the better index construction beats the SP500 and the addition of Share Holder Yield outperforms the old tried and true value indicator of dividend yield. Pretty impressive for such simple changes that basically broadened our universe of companies, sorted by value, and equal weighted the stocks in a portfolio.

This is about the simplest quant investing you can do on your own. For individual portfolios you would limit yourself to the top 25 or top 50 stocks to keep transaction costs down. The broader portfolios and dividend portfolios can be created for free with the screeners such as FINVIZ. The Share Holder Yield portfolios require fee based screeners such as SI Pro or Portfolio123.

Full Disclaimer - Nothing on this site should ever be considered advice, research or the invitation to buy or sell securities. These are my personal opinions only.

4 thoughts on “Quant investing: building a better index

  1. Considering the Sharpe ratio of this strategy compared to some of the others in your previous posts, it seems like simplicity is the main strength of this strategy.

    1. That’s right Jeff. Easily to implement strategies have a much higher probability of being adhered to. It’s a bigger deal than most investors think. Also, I think Shareholder Yield shares some properties with dividend investing, seeing a consistent return of cash every year, that also makes it easier to stick with over the long run.


    1. John, I use the ‘screener’ level in P123. Yes, I have done some looking into different timing strategies for these strategies looking to improve risk adjusted returns. All the MA rule timing strategies have produced worse risk adjusted returns and much higher turnover than simple B&H. I’ve had more success with hedging strategies but not enough to implement them in my real money portfolios as of yet. This is an active are of research for me though.


Comments are closed.