Insights


The Importance of Dispersion in Sector Rotation Strategies

  • Dispersion exhibits autoregressive properties that may be useful for tactical sector positioning.
  • The autoregressive properties of dispersion appear to be present in various geographies and over multiple timeframes.
  • Higher realised dispersion may suggest higher future dispersion over the short run.

Senior Quantitative Research Analyst, ETF Model Portfolio Solutions at SPDR EMEA & APAC
Senior ETF Strategist
Quantitative Research Analyst, ETF Model Portfolio Solutions at SPDR EMEA & APAC
Head of Quantitative Research and Analysis, SPDR ETF Model Portfolio Solutions, EMEA & APAC

Introduction to Sector Rotation Strategies

Many investors employ sector rotation strategies, whereby investments are reallocated from one sector to another depending on the changing macroeconomic outlook or market conditions. While momentum-based measures remain the most popular way to implement rotation strategies, other measures (such as business cycle indicators and value-based metrics) are also common.

With the wide array of ETF vehicles tracking equity sectors in the market, investors have many options for implementing the strategies of their choice. Regardless of the investment strategy investors wish to apply, an important ingredient of success for any rotation strategy is return dispersion.

What is Return Dispersion?

In essence, return dispersion measures the degree to which an index’s constituents perform differently from each other. Applied to the context of our analysis, return dispersion measures the variation of performance between sectors. Another way of viewing dispersion is that it measures the degree of uncertainty, and thus risk, associated with holding a particular investment position. The more variable the return of a holding, the riskier it is. In theory, higher levels of dispersion give rise to more plentiful opportunities for investors to successfully execute a rotation strategy.

Notably, most of the dispersion measures that investors focus on are based on historical, overlapping time series. This approach can represent an important drawdown, since such realised dispersion measures may be entirely unrelated to the future. To be sure, the past is hardly ever a reliable forecast of the future; however, the question here is whether current (realised) dispersion numbers are in any way linked to the past, on average, and whether any useful information can be deduced from examining them. To answer this question, we have run a host of statistical tests to assess the time series properties of realised dispersion.

What is Current (Realised) Dispersion Telling Us?

For much of the year so far, dispersion has stayed at elevated levels, by historical standards, owing to the nervousness in the markets. Commentators ascribe this to the war in Ukraine as well as the tightening of monetary policies by major central banks. While the most recent dispersion level has fallen somewhat, to around its median, it may spike up again if market uncertainty persists. In the next section, we explore the properties of realised dispersion measures and whether any useful information can be gleaned from them.