In my last post I introduced Quantitative Investing as a stock investment strategy with great results and low effort. Here I want to show you how you can get started getting familiar with the process involved and the basic mechanics of implementing a quant strategy by using value stocks as an example.
First, if you’re seriously interested in quant investing you need to buy yourself the book What Works on Wall Street . The book will give you all the basics on hundreds of quant strategies, how they perform, what their risk is, etc… All the strategies I that I mentioned in the last post and this one are well documented there. For this post, I will use a pure value strategy as an example. We’ve already seen how value outperforms over time. Usually, value is measured by Price to Book Value (P/B for short). In other words, stocks with low P/B ratios outperform over time. But nowadays Price to Earnings (P/E) is a more popular measure of stock valuation and just like low P/B stocks, low P/E stocks outperform over time. The table below shows the performance of large low P/E stocks vs the S&P500 overtime. I use the large stock strategy because its the best comparison to the S&P500 which is a large cap index.
A few technical details on the strategy. It starts from the universe of all stocks, then limits it to those with market caps greater than the database average. It then ranks the stocks into deciles (10% increments) based on the P/E ratio. In the book they use the inverse of P/E, the E/P ratio, but its the same thing – just remember that we want low P/E or high E/P as the measure of value. The ‘cheapest’ stocks, low P/E or high E/P, handily outperform the S&P500 – 13.52% per year vs 9.33%. And the risk adjusted return is superior as well – a sharpe ratio of 0.49 vs 0.28. Ok, lets say your convinced and you want to invest in a portfolio of low P/E stocks. What we need first is a stock screening tool that allows us to find low P/E stocks.
The best free stock screener I’ve found is the one at FINVIZ . It has most of the basics you need for implementing many quant strategies and allows excel exports. For our example large cap low P/E quant screen we only need to use one filter in the screener. We limit the market cap to large caps stocks, defined as those greater than $10B by the screener. I then sorted the stocks by ascending P/E ratio and defined a custom view, optional, to make it easier to see the results. You can see the results here . That’s it. We’re done screening. From this we get a list of 574 large cap stocks sorted by P/E from lowest to highest. The strategy calls for investing in an equal weighted portfolio of the 1st decile, the bottom 10% of the low P/E stocks. That would be 57 stocks in this case. Before I show you the example results we’ve already run into two obstacles. First, we couldn’t implement the exact definition of large stocks (those with market caps greater than the database average) in the book in the free screener. Our only choice for large cap was stocks with market caps greater than $10B. Not a huge deal in this case as the average market cap of the 574 stocks is about $40B vs $30B for the S&P500. The second limitation is that for individuals buying 57 stocks is a lot. For individuals the book recommends building portfolios of 25 or 50 stocks. So to keep expenses at a max of say 0.5% of assets per year, at $10 a trade, you’re looking at a min portfolio size of $50K to $100K. Can you now start to see some of the reasons why most investors never try these strategies? OK, now lets take a look at the top 25 stocks from my example large cap low P/E screen. Reminder: this is just an example and not a recommendation – there are much better strategies than a large cap low P/E strategy.
Now comes the hard part. Take your portfolio, say $50K, and buy equal amounts of all 25 stocks on the list, $2K each and hold for one year. Repeat once a year. Here is really where most investors get stuck. The human factor creeps in. We tell ourselves stories. We let our experiences and biases influence our decisions. Reasonable but history shows this leads to poor investment results. Look at the stock list again. POSCO, Korean Steel company – are you kidding me? Steel. Steel is in a bear market. TEF, Spanish phone company – now way, Spain is in a depression. STX, WDC, hard drive companies? No freaking way, the PC industry is dying. See what I mean? And this is why value outperforms over time. Bad news gets priced in, overly so usually, and companies turn around or at least get less worse. This leads to outperformance. The focus needs to be on the process, the overall results, and let the numbers do the work for you. That is the beauty and the promise of quant investing if you can overcome your worst enemy. Yourself.
And that’s pretty much the basics of a simple quant value strategy that’s relatively easy to implement. Not too bad right? There are some obstacles to overcome in translating the strategies from book form to implementation for individuals but the results can potentially be very rewarding. In the next post on quant investing I’ll get into some more of the basics of how to manage the portfolios and how to pick the best strategies. The we can finally get to some actual real time results from the best strategies.
11 thoughts on “ Getting Started with Quantitative Investing ”
Looking forward to the next posts.
Any ETFs do similiar things? You would think if these had such great historical results that ETFs would crop up doing it.
You would think there would be a ton right? There are couple of newish ETFs that implement some type of quant strategy; SDOG does Dogs of the DOW, MMTM is trying momentum out but in general they are few and far between. Maybe advisors don’t see a way to charge high fees from automatic investing. Even O’Shaughnessy himself only handles private accounts and has a few mutual funds that charge about 1.25% a year for a quant strategy. Please.
Financial markets are rarely completely rational. Take value investing. Despite the vast evidence that value outperforms with lower risk, look at the size of value ETFs vs growth ETFs. Growth ETFs outnumber value ETFs and are bigger in terms of assets (for example, compare MGK vs MGV).
But, if and when a large number of quant ETFs launch and get very popular it will something to keep a eye on.
Thanks Paul. Really appreciate you making it understandable. Looking forward to the next installment.
Glad it helped. Paul
Paul, do you have any historical data on performance of value portfolios selected on absolute value (eg the 20 lowest PEs on the S&P) compared with relative value – eg the lowest PEs relative to the sector or some other benchmark ?
Probably. I have reams of data. Do you mean something like low P/E stocks from an all stock universe versus low P/E stocks from a large cap universe (like the S&P500)?
What I was thinking was the lowest PE retail stocks, the lowest PE utilities, etc. Because a relatively low PE in one sector may be a relatively high PE in another sector. So just taking the lowest 20 PE stocks is likley to bias you heavily towards sectors that have generally low PEs, which may give a result that is less diversified than you might want.
There is some great sector data in What Works on Wall Street. As you might expect low P/E outperforms the sector benchmarks in all 10 sectors of the market. For example, low P/E utilities return 14.52% a year versus 11.25% for the sector as a whole. So you could build say a 50 stock portfolio by taking the lowest 5 P/E stocks in each sector. The book doesn’t have a study like this, it would be an interesting one to look at. The diversification may work or not. A heavier weighting to a lower P/E sector could be a good thing for returns as the undervalued sector returns to normal valuation. But maybe the diversification gets you lower returns but lower risk, i.e. higher sharpe ratios and drawdowns.
Comments are closed.