Smart Beta ETF Construction: High versus Low Factor Exposures

Portfolio construction matters

September 2024. Reading Time: 10 Minutes. Author: Nicolas Rabener.

SUMMARY

  • Factor portfolios can be structured to offer high or low factor exposures
  • Portfolio concentration and equal weighting increase factor exposures
  • However, higher factor exposures do not necessarily lead to better returns

INTRODUCTION

Investors want outperformance, but they dislike large tracking errors to benchmarks. Naturally, the former is difficult without the latter, which is one of the reasons why most mutual funds underperform. Fund managers understand that they need to be different to generate alpha, but their investors will redeem their capital if they underperform significantly. Given this career risk, many fund managers minimize the tracking error and therefore their ability to outperform after fees.

Unfortunately, this challenging problem is not unique to active fund managers but extends to systematic products. There are plenty of smart beta products that offer low factor exposures, which essentially makes them expensive index trackers.

In this research article, we will evaluate the main drivers of factor exposures when constructing smart beta portfolios.

PORTFOLIO CONCENTRATION

Creating a factor portfolio requires many assumptions and we will evaluate portfolio concentration, weighting schemes, rebalancing frequencies, and market capitalization thresholds. Our investible universe includes all stocks trading in the U.S. stock market. We use five factors, namely momentum, quality, low volatility, size, and value, with factor definitions in line with industry standards.

The standard assumptions for the long-only factor portfolios in this analysis are using the top 30% of the stocks ranked by factors, equal-weighting, rebalancing monthly, and only considering stocks with a minimum market capitalization of $1 billion.

First, we evaluate portfolio concentration by selecting the top 10%, 20%, and 30% of stocks ranked by the factors, e.g. in the case of the value factor, stocks that trade low on price-to-earnings and price-to-book multiples. We then measure the betas of these portfolios to the factors, highlighting that more concentrated portfolios feature higher betas for all factors, which is intuitive (try Finominal’s