We empirically investigate the effect of sales-force compensation schemes on the intertemporal allocation of sales-agents' effort. Real-world sales force compensation schemes are typically concave and non-linear, involving quotas and bonuses that depend on outputs. Such non-linearities may introduce intertemporal inefficiencies, arising from the potential for strategic timing on the part of sales-agents. We develop a structural model of sales agent behavior to empirically measure the extent of these inefficiencies. The non-linear aspects of compensation plans induce dynamic considerations into sales-agents' actions. Quantifying the empirical relevance of these dynamic considerations requires modeling the forward-looking behavior by which sales-agents allocate effort over time. We model agents as maximizing intertemporal utility, conditional on the current compensation scheme, and their expectations about how quotas would be updated based on their chosen actions. Our empirical approach is to estimate, in a first stage, the structural parameters involving the sales person's utility function, and to then simulate in a second stage, his behavior given a changed compensation profile. We utilize a rich dataset that involves complete details of sales, and compensation plans for a set of 90 sales-people for a period of 4 years at a large consumer-product company in the US. Our preliminary estimates from the data suggest significant incentives for strategic timing arising from the structure of the compensation scheme.

This is joint work with Sanjog Misra of Rochester University.