Levinsohn and Petrin's (2003) Methodology Works under Monopolistic Competition

Markups, returns to scale and productivity can be uncovered from regressing output on inputs. However, econometric identification of theses parameters may be problematic due the simultaneity problem. A common solution is the IV method. However, usual instruments are only weakly correlated to the explanatory variables. Levinsohn and Petrin (2003) propose using a commonly observable variable (intermediate input) to control for unobserved productivity. Their methodology is based on the following key result: under the assumption of perfect competition, the intermediate input's demand function is a monotonic function of productivity. However, firms in most industries enjoy some degree of market power such that perfect competition may not be a desirable assumption for most empirical studies. This paper contributes to the literature by showing the monotonicity condition holds under monopolistic competition.