Security filters often include long-held, fundamental stock-picking wisdom: low valuation, high efficiency of capital, high dividends, etc. However, these sturdy roots of security analysis belie continuous improvement in tools used to identify investment factors. For example, screening applications have introduced groupings and multiple-variable weighting schemes, and performance testing platforms have introduced time-based recalibrations of factors.
These tools are traditionally used with standard valuation metrics, like the P/E ratio, and companies are often grouped into sectors or industries before ranking to acknowledge the inherent differences in group valuations across the market. However, this technique has traditionally relied upon a median perspective, which implicitly treats all groups equally and prevents a model from producing industry-weighted allocations.
Industry-based relative valuation metrics, for example, give one the option to easily discard a neutral approach. Instead of an equal distribution of stocks from a given industry based on their P/E ratio, this approach measures the distance of the company’s P/E ratio from the aggregate valuation of its industry and, therefore, the conviction of the signal.
Applying Relative Valuation Metrics
Five of seven equities in the Technology Hardware Storage & Peripherals industry in the S&P 500 Index were valued significantly lower (<85%) than the industry aggregate P/E ratio at the end of June. However, rather than forcing an equal distribution, with only one of these seven equities in the first quartile, the index-relative approach places all five names in the first quartile (lower quartiles rank better with P/E ratios). On the other end of the spectrum, the Air Freight & Logistics industry contains two equities that have very high P/E ratios relative to their industry aggregate and no equity that’s significantly below the industry aggregate. Therefore, both of these highly valued equities will go in the fourth quartile (and none will go in the first quartile) in an industry-relative approach, as shown in the table below.
Sector-Level P/E Analysis
This process takes neutralization out of the model. It introduces a sector bias but isn’t completely blind to differences in industry valuations as with a standard P/E back test.
The industry aggregate P/E ratio is an important comparison point because it sums all earnings and market capitalizations independently. This removes the noise of individual company earnings that result in negative or unnaturally high P/E ratios. The importance of this approach is clear when considering that most data providers throw out negative earnings in a portfolio or index P/E ratio.
For example, the Next Twelve Month (NTM) P/E ratio of the Leisure Products industry, which contains Hasbro with a P/E of 19.3 and Mattel with a “not applicable” P/E (given its negative earnings projection), is often misleadingly stated by market data providers as simply the only available result: 19.3. The Market Aggregate value, on the other hand, acknowledges Mattel’s negative earnings, and provides a much higher P/E ratio: 41.2.
This industry-relative valuation approach can be combined with other traditional techniques, like a multiple-variable scheme, to create a model that fits your needs. The back test below provides promising performance results for a strategy that used relative P/E ratios in conjunction with dividend yield, ROE, EBITDA margin, price-to-sales, and price-to-book factors.
When considering research approaches, it’s important to keep in mind that innovations in market tools are continuously expanding the available options. A sector-relative model that makes sector bets is one more contradiction that’s easily possible in an era of technological innovation.
Michael is responsible for guiding strategy and development for FactSet’s benchmark data feed solutions. Prior to this role, he was a Product Manager developing aggregated market statistics for countries, industries, benchmarks, and investment funds for the FactSet Market Aggregates (FMA) product group. In his prior role, Michael wrote research briefs that were cited by numerous financial publications. Michael joined FactSet in 2011 and is based in New York City. He holds a B.A. in Finance from the University of Notre Dame and is a CFA charterholder.