Almost all dynamic production systems are subject to lagged productive effects, which are an often-ignored latent source of interference in the efficiency measuring process. Existing data envelopment analysis (DEA) approaches rely on a static production environment. They can easily lead to biased evaluation results due to the erroneous assumption. To tackle this issue, this paper develops a dynamic DEA model that allows intertemporal effects in efficiency measuring. Specifically, the dynamic DEA model incorporates dynamic factors via a linear parametric formulation. Our model can be applied in place of static DEA models to a wide range of applications, such as analyzing longitudinal firm performance and productivity changes. As for the empirical efficiencies, we demonstrate how the lag parameters in the dynamic model can be estimated by the panel vector autoregressive model (PVAR). We use our methodology to evaluate advertising efficiencies of several major automobile and pharmaceutical firms in North America. The result shows that using static DEA in dynamic production can lead to both rank reversals and changes in efficiency scores.

Advertising, Data envelopment analysis, Dynamic systems, Lagged effects, Parametric programming,
ERIM Top-Core Articles
European Journal of Operational Research
Erasmus Research Institute of Management

Chen, C.M, & van Dalen, J. (2010). Measuring dynamic efficiency: Theories and an integrated methodology. European Journal of Operational Research, 203(3), 749–760. doi:10.1016/j.ejor.2009.09.001