Factor investing and industry concentration

The table shows that, over the entire sample period, high yield stocks earned a pre-tax return premium over low yield stocks of 1.4 per cent when portfolios were formed on an industry neutral basis and 0.7 per cent when portfolios were formed without consideration of industry. Ignoring industry in portfolio formation brings comparatively higher incremental tax benefits from dividend imputation because there is more disparity in yields between the high and low yield portfolios. Assuming 80 per cent franking and corporate tax rate of 30 per cent, imputation credits are 2.1 per cent and 0.9 per cent, respectively, for the high and low yield portfolios that ignore industry, compared to 1.9 per cent and 1.0 per cent for the high and low yield portfolios formed on an industry neutral basis. Incorporating imputation credits leads to a total return premium of 2.0 per cent of high versus low yield portfolios formed without consideration of industry (0.7 per cent pre-tax returns premium plus 1.2 per cent incremental imputation credits), and 2.3 per cent for industry neutral portfolios (1.4 per cent pre-tax returns premium plus 0.9 per cent incremental imputation credits).

Taking an industry neutral position lowers volatility for both low and high yield portfolios. For example, volatility is 17.0 per cent for high yield portfolios formed by ranking across all stocks, decreasing to 16.5 per cent for high yield portfolios formed on an industry neutral basis.

The risk to forming a portfolio without consideration of industry is highlighted by the bear and bull markets formed from the first 3.5 years of the sample and the remaining 13.5 years. In the first 3.5 years, when the whole market lost 4.2 per cent a year, the pre-tax return premium to high yield stocks formed on an industry neutral basis was 0.8 per cent (-2.6 per cent versus -3.5 per cent). But the high yield portfolio formed without consideration of industry earned returns 7.3 per cent below those of the low yield portfolio (-6.8 per cent versus 0.5 per cent).

In contrast, during the subsequent 13.5 years, the best performing portfolio was the high yield portfolio formed without consideration of industry, which earned 10.0 per cent and a 3.0 per cent pre-tax return premium over its low yield counterpart. The pre-tax return premium to the high yield portfolio formed on an industry neutral basis was 1.5 per cent (9.2 per cent versus 7.7 per cent). Stocks without analyst coverage earned below-market returns during the entire 17-year period (3.4 per cent), during the first 3.5 years (-8.3 per cent) and during the subsequent 13.5 years (6.7 per cent).

Conclusion

In ranking stocks by value-based characteristics investors need to decide whether they are trying to pick undervalued industries, or undervalued stocks within each industry. Large-sample evidence from the U.S. suggests that investors are better off ignoring industry in ranking. But this necessarily exposes the investor to the risk that their portfolio of low growth industries will underperform for a decade. Dividend yield is one value signal which is especially important in the Australian equity market due to the tax benefits of imputation. Ranking stocks by dividend yield irrespective of industry has the advantage of maximising yield and therefore tax benefits from franked dividends. But the industry concentration associated with this approach can lead to material underperformance over extended periods. On average over 17 years ending February 2024, there has been a return premium to high versus low yield portfolios of 2.3 per cent including tax benefits of imputation when portfolios are formed on an industry neutral basis. The corresponding premium when portfolios are formed without consideration of industry is approximately the same, on average, at 2.0 per cent. But consistent performance requires accounting for industry in portfolio formation. The U.S. evidence is that investors taking long-only industry bets have a better return for risk trade-off, on average, but the words “on average” are doing plenty of work. There is simply too much association between industry and financial ratios to form factor-based portfolios irrespective of industry. 

References

Ehsani, S., C.R. Harvey and F. Li, 2023. Is sector neutrality in factor investing a mistake? Financial Analysts Journal, 79, 95-117.