Quant strategies: 1H 2017 performance update

It’s time to get back to talking about quant portfolios. I haven’t posted on any quant related stuff in a while. Doesn’t mean anything. I’ve just been focused on other things. And my quant portfolios require very little maintenance so once they’re up and running there is not much to do. At least there shouldn’t be much to do. The temptation to fiddle and tweak is quite strong but usually leads to worse results in my experience. Anyway I have some more quant posts coming out over the next few weeks which will kind of re-balance the postings in the blog. Ok, lets get to the quant performance data for the first half of 2017.

Here are the 2017 YTD total return (1H 2017) and max drawdown numbers for the various quant strategies I track. For explanations of the various quant strategies see the portfolios page. All equity portfolios consist of 25 stocks and were formed at the end of 2016. No changes in the holdings since that time (except for the TAA Bond strategy and the Pure Momentum strategy which re-balance every 4 weeks).

In the table below I list various quant strategies along with their YTD performance and drawdowns. Also, listed are various benchmark indices. All performance numbers are from Portfolio123.com.

Just like the first half of the year, 2017 continues to be a tough year for the equity quant portfolios. The average performance of the 8 quant portfolios is 3.36% YTD with only 2 of the 8 outperforming the SPY and none outperforming the international markets. You could say 2017 seems like a payback year for the stellar performance of 2016 but hey the year is only half over. I’ll talk more about overall performance in a bit but let me talk about the details in the table above first.

The microcap portfolio (14.22%) and the utility value portfolio (12.04%) led the way in the first half of the year. Quite an odd combination of portfolios to lead the way. On the other side trending value and the value composite portfolio were the laggards. Both of these portfolios were had quite a few US energy companies at the beginning of the year which really hurt performance when the energy turnaround fizzled out. The foreign versions of these portfolios did quite a bit better. Foreign TV2 is up 7.68% for the year and Foreign VC is up 12.68% for the year. The other equity quant portfolios lagged the market as well but are positive for the year. On the bong side TAA Bond returned 2.09% for the first half while the more concentrated version TAA Bond 1 returned 3.47%.

Lets talk about more aggressive versions of the quant portfolios. I’ve talked about this in the past with respect to increasing performance. However, there is a risk reduction aspect as well. Sometimes the quant models get caught out, especially if a trend reverses itself or value goes on to become even more value before it turns around. After all the models are making their decisions at one point in time, on data as of Dec 31, 2016 in the cases above. Versions of these strategies that mitigate some of this timing effect can be useful. Aggressive is probably the wrong word. I mean versions of these portfolios that are both more concentrated and use some risk management in individual positions. In the table above, next to the quant portfolio, I listed what the equivalent aggressive portfolio has done in the 1H of 2017. In 7 out of the 8 portfolios performance was improved with similar drawdowns. The average performance of the aggressive versions was 7.41% for the first half. Still underperforming but much better. Something to consider for your quant portfolios.

We’ll see what the rest of 2017 holds for the quant portfolios but let me talk about measuring quant performance. Quant underperformance vs the benchmarks is one of the toughest things about quant investing. Tracking error is wonderful on the upside but really difficult on the downside. Your always wondering if it’s normal, has something changed to make your model obsolete, etc. Unfortunately, this a feature, not a bug, of quant investing. But it’s what helps sustain the outperformance over time. The table below shows the historical performance of the quant models I track.

‘Core 4’ in the table are the 4 quant portfolios I consider the core of a good quant portfolio. The 4 portfolios are XLP Value, XLU Value, TV2, and Momo.

The table shows the performance of my implementations of the various quant strategies and their performance as simulated in Portoflio123. Then the last column compares that to the published performance from the creator and source of these models, O’Shaugnessy’s What Works On Wall Street. The data shows no real deterioration in performance of the quant models over time which is what you want to see when you’re doing this kind of stuff on your own. But what I want to highlight is the base rates. As a quant investor, over the last 5 years, most of the time your are underperforming the SPY even though you are beating the index quite handily on a total return basis. Great news if you stick with these strategies but it sure does make the actual performance in the trenches quite a bit more difficult. I will add this table to the Portfolios page as well.

I’ve gone on long enough for today. In summary, 2017 is proving to be a challenging year so far for quant investing. There’s nothing unusual about this. Just par for the course in quant investing. But let’s hope for a better second half. After all, we’re human.