Do price-earnings ratios explain investment decisions better than Tobin’s q? Evidence from German firm-level data

ABSTRACT

This article tests the additional information content of price-earnings ratios, with respect to Tobin’s q, in explaining firms’ investment behaviour. While Tobin’s q describes the expected future earnings related to those projected by the book value, the price-earnings ratio compares future growth of earnings based on the projection of current earnings. In other words, a high price-earnings ratio might indicate that investors are willing to rely on future earnings growth, even though current earnings are low. By using an unbalanced panel of about 500 listed firms from Germany over the period 1987–2007, we find that including the price-earnings ratio in the investment equation does not change the explanatory power of Tobin’s q. Most notably, the price-earnings ratio exerts a positive and significant impact on investment. These results are robust to the inclusion of a measure of the firm’s internal funds and of fixed effects and also to the use of different estimators.

KEYWORDS:

• Investment decisions
• Tobin’s q
• price-earnings ratios
• Germany

JEL CLASSIFICATION:

• D92
• G31

Acknowledgements

I would like to acknowledge valuable comments of Vicente Cunat, Giovanni Ferri, Sergio Ginebri, Francesco Nucci and Alberto Pozzolo. I would also like to thank participants at the 6th Workshop for Italian PhD Students in Economics (University of Salerno), at the 3rd Doctoral Meeting of Montpellier and at the XVIII International Tor Vergata Conference on Money, Banking and Finance. Special thanks to the Electronic Resources of Bocconi University for providing the access to the Worldscope database. I would also like to thank Margherita Di Paolo, Dario Focarelli and Paolo Zanghieri. Remaining errors are my own.

Disclosure statement

No potential conflict of interest was reported by the author.

Notes

¹ Hubbard (1998, 200) argues that ‘if financial frictions are unimportant, internal and external financing are perfect substitutes, and information about changes in net worth which is dated contemporaneously with q should be irrelevant for the investment decision’.

² A detailed description of these measures is provided in Section IV.

³ In what follows, OLS, robust and quantile regressions include year and industry dummies, but exclude individual fixed effects. On the contrary, the within estimator includes both time and individual dummies, but excludes industry dummies. The GMM estimator includes all fixed effects, even though firm-fixed effects will be removed as described in this section.

⁴ Estimations are conducted using the XTABOND2 program for Stata written by Roodman (2006).

⁵ The base year for the Worldscope Database is 1980, although data are best represented from January 1985. This database contains both qualitative and quantitative information on each listed firm. The qualitative information refers to a variety of characteristics that help to define the firms’ profile and includes the company’s header information, among others. On the other hand, the quantitative information include the financial statements, such as the balance sheets, the income statements and the cash flow statements, the valuation ratios, such as the profitability, the liquidity and the leverage ratios, and the security and market data that include, among others, the stock prices and the stock performances.

⁶ The reported number of observations refers to the size of the sample after excluding influential observations. In particular, observations with high values of investment rate, price earnings and average q are excluded from the original sample.

⁷ Another reason for which our result on financial constraints for small firms is different from Audretsch and Elston (2002) is that they consider the period 1970–1986, including the hypothesis that German firms in the Eighties benefitted from an increasing competition of the banking sector, establishing long-term relationships.