Sector Detector: Introducing an Enhanced SectorCast Model
The market continues to look strong and resilient. Whether the news, economic reports, and earnings announcements are bad or good, the market has been absorbing the blows or chasing the momentum like a champion. Now it is struggling with various resistance levels from moving averages and chart formations, but continuing to hold up in this low-trading-volume environment.
Although I have been looking for more downside action to test support, the market’s resiliency has been impressive. All major averages are now back above both their 50-day and 200-day simple moving averages – and well above their exponential moving averages – which would suggest a new uptrend is firmly established. But this market has not made it quite so easy to figure out this year, so I wouldn’t back up the truck quite yet.
Enhancements to the Model: At Sabrient, we are constantly testing, refining, and enhancing our quantitative models, and I’m pleased to say that we have introduced major changes to our SectorCast model. The newly enhanced version will be used by Sector Detector starting today. The biggest change is a reduction in the importance of projected price-to-earnings ratio (PPE), plus differences in the way our proprietary relevance scoring approach is implemented for the given factors and the way the composite scores are compiled.
Another change I am making to Sector Detector is to use the iShares family of ten sector ETFs rather than using eight Sector SPDRs and two iShares. Among other reasons, the iShares hold more stocks, which is important for selecting top and bottom-ranked stocks for enhancement strategies. Because of the differences in compositions, there will be some differences in the way the iShares score versus the Sector SPDRs. For example, in Basic Materials, Monsanto (MON) composes a large percentage of the XLB but is not even a member of the IYM (iShares places it in the Consumer Goods sector).
As a result of changes to the model, the two sector ETFs that benefited the most from PPE – Financial and Energy – have taken rather large hits to their scores. But some of the other sectors score almost the same as they did in the prior version. In general, I expect to see more movement at the tails with the new scoring system, as well as greater near-term predictivity looking forward. I also like the wider distribution of scores offered by the new model.
Latest rankings: At the top, Technology (IYW) is still the highest ranked ETF with a solid score of 79. IYW fares the best (on a composite basis across its constituent stocks) in the percentage of analysts increasing earnings estimates, and it ranks high in return on equity, return on sales, and projected year-over-year change in earnings. The second place spot now goes to Industrial (IYJ) with a score of 62.
Top-ranked stocks within IYW and IYJ include Tech Data Corp (TECD), Lexmark International (LXK), Rock-Tenn Company (RKT), and Deluxe Corp (DLX).
Former cellar-dweller Telecommunications is now smack dab in the middle of the rankings. The indignity of the lowest ranking now goes to Energy (IYE) with a rather ugly score of 7. This might be a surprise, but with the exception of its reasonable valuation, it scores poorly on most metrics on a composite basis across the U.S. Energy stocks that make up the ETF. Joining IYE in the bottom two is Financial (IYF), scoring a 30 to come in just under Basic Materials.
Low-ranked stocks within IYE and IYF include Ultra Petroleum (UPL), Range Resources (RRC), Plum Creek Timber (PCL), and CIT Group (CIT).
These scores represent the view that the Technology and Industrial sectors may be relatively undervalued overall, while Energy and Financial sectors may be relatively overvalued.
Performance: I have tracked the performance of each of the prior four weekly portfolios as of the market close on Tuesday, 7/27/2010. Each portfolio assumes that the top two ETFs are entered long and the bottom two are entered short, all at the opening prices on Wednesday. Of course, for those who prefer not to sell short, this could be run as a sector rotation strategy – with perhaps the top 3 or 4 sector ETFs long.
None of the long/short portfolios have been performing very well during this July rally that has “lifted all boats.” Next week, I will begin tracking the performance of the enhanced SectorCast-ETF model, and I expect to see improved performance.
Disclosure: Author has no positions in stocks or ETFs mentioned.
About SectorCast: The rankings are based on Sabrient’s SectorCast model, which builds a composite profile of each of the ten ETFs based on bottom-up scoring of their constituent stocks. The model employs a fundamentals-based multi-factor approach including forward valuation, earnings growth prospects, analyst revisions, and various return ratios.
SectorCast has tested to be highly predictive for identifying the best (most undervalued) and worst (most overvalued) sectors, with a one-month forward look. Of course, each ETF has a unique set of constituent stocks, so the sectors represented will score differently depending upon which set of ETFs is used. For Sector Detector, I use ten iShares ETFs representing the major U.S. business sectors.
About Trading Strategies: There are various ways to trade these rankings. First, you might run a sector rotation strategy in which you buy long the top 2-4 ETFs from SectorCast-ETF, rebalancing either on a fixed schedule (e.g., monthly or quarterly) or when the rankings change significantly. Another alternative is to enhance a position in the SPDR Trust exchange-traded fund (SPY) depending upon your market bias. If you are bullish on the broad market, you can go long the SPY and enhance it with additional long positions in the top-ranked sector ETFs. Conversely, if you are bearish and short (or buy puts on) the SPY, you could also consider shorting the two lowest-ranked sector ETFs to enhance your short bias.
However, if you really don’t want to bet on which way the market is going, you could try a market-neutral, long/short trade—that is, go long the top-ranked ETFs and short the lowest-ranked ETFs. And here’s a more aggressive strategy to consider: You might trade some of the highest and lowest ranked stocks from within those top and bottom-ranked ETFs, such as the ones I identify above.
About Performance Tracking: I track each week’s set of ETFs (2 longs and 2 shorts) as a mini-portfolio over the course of four weeks. Because SectorCast does not include any technical triggers, this will give the fundamentals-based model a chance to achieve its predicted move. You might also watch just the two long positions as a separate long-only sector rotation strategy.