Denise Chisholm is a sector strategist and research analyst for Fidelity Management & Research Company, the investment advisor for Fidelity’s family of ETFs and mutual funds. She is responsible for the research of portfolio construction strategies combining sector-based mutual funds and ETFs. Chisholm also has a unique way of analyzing sectors using historical data and probabilities. ETF.com caught up with her to talk about sectors, particularly financials, as we head into next week’s Fed meeting, where the consensus is for an interest rate hike.
ETF.com: You analyze sectors using historical data. How is that different from the way others do it?
Denise Chisholm: I have a different process than most people who analyze sectors. They stand in two camps, one being a fundamental camp and the other being a more macroeconomic camp. What you find historically is that a lot of the fundamental factors don't work as well in sectors.
The approach I've chosen is to use historical data. That's mainly to layer the probabilities associated with historical data, both our internal fundamental data sets, and our macroeconomic data to integrate cycles or specific trends within sectors. I layer historical probabilities to allow history to inform my views.
ETF.com: What are some of your sector recommendations at this point in the year?
Chisholm: Let's talk about financials first because it’s not only a recommendation, but it's one of the most unique sectors historically. Most investors I talk to, both internally and externally, are reticent to buy financials since it's outperformed so much. But again, we can use history to inform our views.
If I say I'm going to base my probabilities of future performance on my conditional probabilities of past performance, you basically get 50/50 odds for outperformance in every sector but financials, where you're looking at like 62% odds.
And if I take out the times when we know there were head-fakes, it doesn't change probabilities for any other sector; they still retain 50/50 odds, but financials go to 69%.
And if you tell me further, “Well, it's different this time; they're up a lot, well beyond their historical average of just random outperformance,” I would tell you that actually increases the probabilities that they outperform more.