Decomposing the Price-Earnings Ratio
The York Management School
University of Reading – ICMA Centre
Abstract:
The price-earnings ratio is a widely used measure of the expected performance of companies, and it has almost invariably been calculated as the ratio of the current share price to the previous year’s earnings. However, the P/E of a particular stock is partly determined by outside influences such as the year in which it is measured, the size of the company, and the sector in which the company operates. Examining all UK companies since 1975, we propose a modified price-earnings ratio that decomposes these influences. We then use a regression to weight the factors according to their power in predicting returns. The decomposed price-earnings ratio is able to double the gap in annual returns between the value and glamour deciles, and thus constitutes a useful tool for value fund managers and hedge funds.
Decomposing the Price-Earnings Ratio – Introduction
The price-earnings (P/E) effect has been widely documented since Nicholson (1960) showed that low companies having low P/E ratios on average subsequently yield higher returns than high P/E companies, and this difference is known as the value premium. A low price-earnings ratio is used as an indicator of the desirability of particular stocks for investment by many value/contrarian fund managers, and the P/E effect was a major theme in Dreman (1998). The value premium is mostly positive through time, and a large number of studies have confirmed its presence. While the continued existence of a value premium is puzzling for academics, a plausible explanation is that it provides compensation for the extra riskiness of value shares.
However, the CAPM beta does not increase as the P/E decreases; if anything, it decreases (Basu, 1977), so the risk must reside in other measures. According to Dreman and Lufkin (1997), sector-specific effects are also unable to explain the value premium, and more complex multifactor models have similarly failed to rationalise the outperformance of value stocks (see, for example, Fuller et al., 1993). Others have proposed behavioural explanations (e.g., Lakonishok, Schleifer and Vishny, 1994), ascribing the extra returns from value shares to psychological factors affecting market participants.
However, the P/E as it is commonly used is the result of a network of influences, similar to the way in which a company’s share price is influenced not only by idiosyncratic factors particular to that company, but also by movements in prices on market as a whole, and the sector in which the company operates. A large number of studies have examined the decomposition of stock returns into market-wide and sector influences, and in this paper we propose and show the usefulness of an analogous approach in deconstructing the P/E ratio.
We identify four influences on a company’s P/E, which are:
1) The year: the average market P/E varies year by year, as the overall level of investor confidence changes.
2) The sector in which the company operates. Average earnings in the computer services sector, for example, are growing faster than in the water supply sector. Companies in sectors that are growing faster in the long-term should warrant a higher P/E, so as correctly to discount the faster-growing future earnings stream.
3) The size of the company. There is a close positive relationship between a company’s market capitalisation and the P/E accorded.
4) Idiosyncratic effects. Companies examined in the same year, operating in the same sector and of similar sizes nevertheless have different P/E’s. Idiosyncratic effects, that do not affect any other company, account for this. Such effects could be the announcement of a large contract, whether the directors have recently bought or sold shares, or how warmly the company is recommended by analysts.
Using data for all UK stocks from 1975-2003, we take these four influences in turn, looking at the extent to which they affect the P/E, and how closely they are correlated with subsequent returns. We decompose the influences on each of our company/year data items, and we then run a regression to get a weight for each influence. Using these weights, we construct a new sort statistic for assigning companies to deciles, and we are able to double the difference in returns between the glamour and value deciles. Finally, we show, via a portfolio example, the practical effect of the new statistic on the values of the glamour and value deciles through time.
See full PDF here.
H/T Alpha Architect
