Better Returns By Reducing Outliers

Safety First: NASCAR driver Johanna Long and her helmet (Credit: Gearheads)

Revisiting Our Revised Security Selection Method

Since June 2017, I’ve run a Marketplace service which has an “investing with a helmet on” approach: I present concentrated portfolios where each security is hedged. To construct these portfolios, I need securities that generate strong enough returns, on average, to overcome the hedging cost. As I’ve written before, the names that rise to the top of my system’s daily ranking tend to outperform the SPDR S&P 500 ETF (SPY), on average. That continued to be the case with top names presented last year, albeit less spectacularly: After 112 weeks of presenting my top names to my Marketplace subscribers, the average top names cohort has returned 7.03% over the next six months, vs. SPY’s average of 6.09%. We (my team and I) implemented a change to that security selection process last November, which I detailed here then (“Finding Alpha In Security Selection“).

In my November article, I made a prediction about our new security selection method:

My prediction is that, among top names calculated the new way, we will have fewer positive and negative performance outliers, but time will tell.

So far, my prediction has come to pass, as I’ll show below. To do so, I’ll compare the results of our new security selection method to the old one from the first date for which we have top names generated using both methods, November 20th.

Our Top Names From November 20th (Old Method)

Below is a screen capture showing our top ten names calculated the old way on November 20th. I’ve highlighted the top row, for Galapagos (GLPG), as I’ll refer to it to illustrate how the names were selected. The rest of the top ten that day included Apple (AAPL), Edwards Lifesciences (EW), Phillips 66 (PSX), Teledyne Technologies (TDY), Taiwan Semiconductor (TSM), CVS (CVS), Sherwin-Williams (SHW), Royal Gold (RGLD), and Humana (HUM).

Screen capture via author.

The number I’ve circled in red at the bottom right was the average cost of hedging each of these names against a >20% decline over the next ~6 months, which was approximately 2.68% of position value. We’ll come back to the significance of that later, but first let’s focus in on Galapagos for a moment.

Screen capture via author.

The way our old security selection method worked was essentially this: we started by looking at the average 6-month return of a security over the long term (ten years, in the case of securities that have been public that long; for stocks that haven’t been public that long we use a long term proxy; and for ETFs that haven’t been public that long we use the time since inception as the long term). In the case of Galapagos, since it hadn’t been public 10 years, we used Johnson & Johnson (JNJ) as a long-term proxy. The next thing we looked at is the total return over the most recent 6-month period. The mean of the two appears in the column titled “Exp Return” above.

Essentially, we started by assuming mean reversion in returns. The next step was to use options market data to gauge whether we should raise or lower that return estimate. The first way we did that was to attempt to hedge the name against a >9% decline over the next several months with an optimal, or least expensive, collar capped at its Exp Return. If we weren’t able to find an optimal collar using that cap, we lowered the cap figure until we found one. In the case of Galapagos, we were able to find one, so the “w/Cap Drop” figure is the same as the “Exp Return” figure.

The next gauge of options market sentiment we used here was to see if it was possible to hedge the security against a >9% decline with an optimal put. Since it was, we adjusted its return up commensurate with the average outperformance we’d seen in securities that were also hedgeable with puts (AHP) at a 9% threshold. You can see the result of that in the higher figure shown in the column titled “w/AHP”.

Our Top Names From November 20th (New Method)

Here’s a screen capture showing our top 10 names calculated the new way on November 20th. The top 10 calculated this way were Anixter (AXE), Equity LifeStyle Properties (ELS), Home Depot (HD), iShares Silver Trust (SLV), iShares U.S. Medical Devices (IHI), Allstate (ALL), Toyota (TM), SPDR Gold Trust (GLD), Equity Residential (EQR), and Marsh & McLennan (MMC). Here I’ve highlighted the top name, Anixter, to illustrate the new selection method.

Recall that the average cost of hedging the top 10 calculated the old way against a >20% decline over the next ~6 months was about 2.68% of position value. As you can see above, here it was about 1.27% of position value, less than half as much.

Here’s a closer look at the top name, AXE, to illustrate the difference between the new selection method and the old.

The selection process starts out the same as the old way: the Exp Return is calculated the same way, and adjusted down, if necessary in the w/Cap Drop column. Then it’s adjusted up, if indicated in the w/AHP column. The difference here is that the new security selection method gives an additional boost to names that are also hedgeable with puts >6% declines, as AXE is here. Essentially, it gives preference to less risky names, as determined by the ability to hedge them against small declines with optimal puts.

This preference for less risky names is why I predicted we’d see fewer positive and negative outliers with this approach.

Performance Of Both Cohorts So Far

Here’s the performance of the top 10 names calculated the old way as of Monday’s close.

And here’s the performance of the top 10 names calculated the new way over the same time frame.

Note that the top 10 names calculated the old way have the greater outliers in positive performance: GLPG up 22.8% so far, and AAPL up 17.39%, versus the top two names in the top 10 calculated the new way: AXE, up 15.02%, and GLD, up 7.37%. But the top 10 names calculated the old way also had the greater negative outliers: EW, down 6.27% so far, and PSX down 16.49%. In contrast, the worst two performers from the top 10 calculated the new way were TM, down 0.22%, and EQR, down 4.32%.

The top 10 names calculated the old way had higher highs and lower lows, and averaged 2.38% so far while the top 10 names calculated the new way were up 4.74%.

Wrapping Up

A risk of this security selection method (as well as every other security selection method) is that one more of the names you select will decline significantly. That’s why I suggest hedging. If you were fairly aggressive in terms of risk tolerance, and hedged these names against a >20% decline, note that your returns net of hedging cost would be lower than the numbers shown above. Instead of being up 2.38% so far, with the top 10 names calculated the old way, you’d be down about 0.3% so far, after subtracting the 2.68% hedging cost. With the top 10 names calculated the new way, you’d be up 3.47%, after subtracting the 1.27% hedging cost. So the top 10 names calculated the new way are outperforming so far both in terms of gross returns and returns net of hedging cost.

See This Week’s Top Names

Neither AXE nor GLPG are among our current top 10 names. To see this week’s top names and hedged portfolios, you can sign up for a free two-week trial to my Marketplace service here.

Disclosure: I/we have no positions in any stocks mentioned, and no plans to initiate any positions within the next 72 hours. I wrote this article myself, and it expresses my own opinions. I am not receiving compensation for it (other than from Seeking Alpha). I have no business relationship with any company whose stock is mentioned in this article.

Be the first to comment

Leave a Reply

Your email address will not be published.