“Another record year for index funds.”
“New high for passive fund inflows.”
Stop me if you’ve seen these headlines before. Yes, fee pressures are playing their part, as has the increasing impact of fiduciary responsibilities. But there’s no denying that investing—specifically asset allocation—looks a lot different today than it has in the past.
Advisors and portfolio builders regularly balance growth and value, stocks and bonds, international and domestic, and now the competing roles of active and passive. Do we really understand the behavioral differences between the two?
What I outline below is not a case for active management or against index tracking, but instead simply a case to modernize portfolio diversification and truly understand the drivers of portfolio performance. To do this, we’ll look at a statistical tool known as average pairwise correlation. That is, measuring the behavior of all stocks in a universe in relation to one another.
For those of us who grew up in the upper Midwest, “pairwise” may be a familiar term. College hockey uses pairwise rankings to compare teams that may or may not play one another in a season using a collection of stats like record against common opponents and strength of schedule. It helps predict which of two teams would win in a hypothetical head-to-head matchup. Average pairwise correlation is similar in that it possesses predictive power, describing the relationship between all teams (stocks) in a league (index). Lower correlation means the teams are behaving differently enough to make it easy to discern which team has a better chance of winning. On the flip side, higher correlation represents parity, making it more difficult to predict the winner.
When we apply average pairwise correlation to the world of investing, we’re looking at the relationship between stocks in a universe, or index. Using the same logic, higher correlation shows that the individual stocks in an index are more or less moving together. Could be up or down, but together, without explanation. All stocks are “winners” or all stocks are “losers” and there isn’t much difference in returns if you’re holding stock A or stock B—oftentimes despite one company’s better fundamentals. In this environment, indexing works well. Your low-cost, index-tracking ETF will provide broad exposure to an asset class, and you’ll feel fine knowing that whether the index goes up or down, there isn’t much space for a replacement fund to add value. In other words, it is difficult for an active manager to beat the market when all stocks are moving together indiscriminately.
When average pairwise correlation among stocks is low, it shows that the stocks in an index are moving more independently of one another—perhaps driven by company fundamentals. This dispersion of returns makes it easier to discern the “winners” from the “losers.” In this environment, active management has the upper hand; the differences in stock performance are easily explained and the less attractive stocks can intentionally be excluded from a portfolio—or at least underweighted.
Here’s what the average pairwise correlation in the U.S. equity market has looked like since 2008, with the dotted line representing the 20-year average of 0.20. You’ll see that for the better part of the past decade, the correlation of stocks in the S&P 500® has hovered above its 20-year average.
Looking at the most recent period, the following stats can be observed by examining the average pairwise correlation with active fund performance:
- During the period between August 2015 and June 2016 (shaded in red), when the average correlation spiked, a majority of actively managed large-cap funds (51%) outperformed their respective indices in just one of 11 months.
- In the 19 months that followed (green), when the average correlation dropped, there were 12 such months, with an average of 55% of actively managed funds outperforming their respective indices in those months.1
In fact, since the beginning of 2017, the correlation structures have been breaking down across many relationships in the market, including style factors, sectors and indices.
Of course, average pairwise correlation is just one factor to consider when building a portfolio. It is not the end- all-be-all for deciding between an active or passive fund, or what may be right for an individual client. But it should help bring awareness to the underlying behavior of a particular approach. When I’m out in the field, meeting with advisors and portfolio builders, it’s not uncommon that I hear an advisor questioning how much investment exposure matters, or if beta exposure is more important. If a niche strategy that behaves in a specific manner is appropriate for a particular client, or if broad market exposure is better.
I am of the belief, as are many of my colleagues in the industry, that there is a place for both active and passive funds in portfolio construction. Active and passive can complement one another, and average pairwise correlation is a discrete statistic that proves this.
As an asset allocator, if you believe in diversification, you’ll understand that active and passive are styles that go in and out of favor just as growth and value, large-cap and small, international and domestic, and so on. When stocks in an index are not correlated, there are dislocations that an active manager can capitalize on. Just like the conviction level of predicting my Minnesota Gophers will beat the Wisconsin Badgers in hockey.
1 Large-cap blend funds were compared to the Russell 1000® Index; large-cap value funds were compared to the Russell 1000® Value Index; large-cap growth funds were compared to the Russell 1000® Growth Index.