Momentum strategies have been popular since the original Jagadeesh and Titman article was published in 1993. Variations on the strategies have employed calculating momentum on an individual and industry basis. For instance, in a 1999 study, Moskowitz and Grinblatt produced a positive and significant excess return from a long/short strategy – buying the top three winning industries and selling the bottom three losing industries. There have been other demonstrations that momentum strategies utilizing text-based industry classifications or intra-industry approaches are also profitable. However, those studies used one of three classification systems including SICCD, NAICS, GICS1, preventing a generalization of the specific classification system as best-in-class. Which classification system is best, if any?
Industry classification, industry momentum and short-term reversal
The focus of this empirical study is to analyze the same momentum strategies across different granular levels of classification within the same industry classification system. Various levels within an individual classification system depict a different state of product market competition: the 2-digit SICCD generates 76 industries; the 3-digit SICCD generates 297 industries and Fama French contains 48 industries.
Do classification systems vary over time?
Does increasing the granularization of industry definitions impact the performance of industry momentum strategies?
What are the Academic Insights?
YES. See Figure 2 for a depiction of the evolution of market structure as reflected in changes in classification systems over time. On the left side of Figure 2 from 1963-1976, the number of 3-digit SICCD industries increased sharply and continued to rise until 1997 when it began to decline. On the right side of Figure 2, note the average number of firms per industry increases in a similar fashion but declines more abruptly. In the post-1997 period, not only does the number of industries decrease but they become more concentrated.
YES. Using a sample covering the years 1963 to 2018, the returns of industry momentum strategies changed when more granular levels of classification systems are used. This result holds over varying periods over which momentum ranking is calculated and over various holding periods. The author argues “the undiversified individual firm risks that comes from 20% single-firm (monopoly) industries under such granular level of classification” drives the results. If so, it follows that at a higher level of classification, the results stabilize. In addition, the degree of industry concentration that seemed to evolve over the sample period had a substantial impact on the performance of the industry strategies. During the industry consolidation period (1998–2018), industry momentum strategies were not profitable under a granular level of classifications such as 3-digit SICCD. While such strategies over most momentum ranking horizons and holding periods remain profitable and significant under two-digit SICCD and Fama-French (FF48) classification, a long-term reversal pattern emerged under longer momentum horizons such as nine to twelve months to rank industries. This result is consistent with other findings. For instance: when the 3-digit SICCD is used, industry concentration has a negative correlation with returns (Hou and Robinson, 2016), although this result reverses when NAICS is used (Grullon et al. 2019). More confusion emerges for the SICCD system when CRSP versus COMPUSTAT data is used, where 38% disagreement occurs at the 2-digit level and 80% at the 3-digit level (Kahle and Walking, 1996).
Why does it matter?
The paper adds to the literature on constructing industry momentum strategies. When more granular classification systems are used, performance begins to vary quite a bit and may even become negative. Understanding the link between the level of industry classification, the amount of granularity within, and the impact on performance is useful for managers building industry momentum strategies.
The most important chart from the paper
The results are hypothetical results and are NOT an indicator of future results and do NOT represent returns that any investor actually attained. Indexes are unmanaged and do not reflect management or trading fees, and one cannot invest directly in an index.
This paper finds that the level of industry classification plays a significant role in the performance of industry momentum strategies. From 1963 to 2018, it appears that using higher levels of classification generates higher positive returns, while using granular level generates results that display short- and mid-term reversal patterns. In addition, the performance of industry momentum strategies across different levels of industry classification varies significantly over different sample periods. This is likely caused by the evolution of market structure over time. This paper highlights the influence industry classifications and market structure evolution have on the performance of industry momentum strategies.
Dr. Tommi Johnsen was a past Director of the Reiman School of Finance and a tenured faculty at the Daniels College of Business at the University of Denver. She has worked extensively as a consultant and investment advisor in the areas of quantitative methods and portfolio construction. She taught at the graduate and undergraduate level and published research in several areas: capital markets, portfolio management and performance analysis, financial applications of econometrics, and the analysis of equity securities. Her publications have appeared in numerous peer-reviewed journals.
Performance figures contained herein are hypothetical, unaudited and prepared by Alpha Architect, LLC; hypothetical results are intended for illustrative purposes only. Past performance is not indicative of future results, which may vary. There is a risk of substantial loss associated with trading stocks, commodities, futures, options and other financial instruments. Full disclosures here.