Do managers work harder in competitive industries?

Abstract

We analyze the relationship between product market competition and managerial effort in a linear principal–agent model. Firms compete in a differentiated Cournot oligopoly. We use two competition indexes: the degree of product differentiation and the number of firms. An increase in competition stemming from a lower degree of product differentiation reduces the optimal level of effort, while an increase in competition from more firms has an ambiguous effect. An increase in the number of firms reduces (increases) effort when the degree of product differentiation is low (high).

Introduction

An old belief in economics is that competition forces managers to work harder. However, Martin (1993) and Hermalin (1994) identify a general and counterintuitive effect of competition on effort in Cournot oligopolies with homogeneous products. As the number of firms increases, the quantity produced by each firm declines, which lowers the gain from cost-reducing managerial effort (this is sometimes referred to as the output effect).

Two additional issues must be considered in order to understand the impact of competition on managerial effort. First, competition can alleviate information asymmetry, as is well known from the literature on relative performance (see, e.g. Lazear and Rosen, 1981, Nalebuff and Stiglitz, 1983, Shleifer, 1985). In an environment with systemic risk, an increase in the number of firms improves information about effort. This, in turn, reduces the risk borne by the manager and makes stronger incentives possible (we denote this the information effect). Second, competition arises, not only from an increase in the number of firms, but also from a reduction in product differentiation.¹

There are two novel aspects in the present paper. First, we study the impact on effort of product market competition using two different competition indexes: the number of firms and the degree of product differentiation. Second, we analyze the effect of an increase in competition on the trade-off between output effect and information effect. Although relative performance evaluation is the most obvious way in which competition can alleviate information asymmetry, to the best of our knowledge nobody has so far examined whether its benefit offsets the negative output effect. We identify a closed form solution of managerial effort suitable of comparative statics analysis and empirical testing. The derivation of the executive compensation schedule is based both on Holmström's (1982) results on sufficient statistics and aggregate performance, and on Holmström and Milgrom (1987) [hereafter H–M] linear principal–agent model.

Our main findings are that an increase in competition stemming from a lower degree of product differentiation reduces the optimal level of effort, while an increase in competition from more firms has an ambiguous effect. An increase in the number of firms reduces (increases) effort when the degree of product differentiation is low (high). Intuition suggests that a negative output effect is at work regardless of the competition measure since the marginal benefit of effort declines with the number of firms and increases with the degree of product differentiation. However, the countervailing information effect arises only through a more precise performance measure when the number of firms increases.

The first formal work in this area is Hart (1983) that shows that in an industry with both managerial firms and owner-managed firms, an increase in the fraction of the latter reduces managerial slack. Scharfstein (1988) generalizes Hart's model and shows that the effect on effort of an increase in the number of owner-managed firms is indeterminate. Panunzi, 1994, Horn et al., 1994 show that in differentiated duopolies, competition is more likely to increase effort under Cournot competition than under Bertrand competition.

The paper most closely related to ours is Hermalin (1992) that shows that the effect of competition on effort is indeterminate and the information effect might also have an ambiguous sign. He compares contracts based on different information structures without, however, investigating whether a given information structure that allows for weaker incentives could be included in the optimal contract. Our model differs from his model on several counts. First, following Holmström (1982) the signals on which the managerial contract is based are optimally chosen. Second, Hermalin separates the information effect from the output effect by assuming that expected profits are unaffected by increased competition. In our paper, on the contrary, we focus on the trade-off between information effect and output effect. Third, following most of the literature on relative performance we assume that the conditions for the first-order approach are satisfied. Finally, we measure competition also through product differentiation.

The problem we analyze is also related to whether a more competitive (as measured by the number of firms) market structure is more likely to develop and employ a superior technology. The similarity stems from the fact that a cost-reducing process innovation is analogous to cost-reducing managerial effort. However, there are two main differences. First, the R&D literature abstracts from the potential agency problem within the firm. Second, innovation gives the winner the exclusive right to enjoy cost reduction while its rivals continue to produce with the inferior technology. The R&D literature shows that competition has ambiguous effects on investment and on the speed of technological progress, depending on the nature of the costs and benefits of R&D.²