Functional Flexibility, Latent Heterogeneity and Endogeneity in Aggregate Market Response Models

摘要

We focus on flexibility, latent heterogeneity and endogeneity in aggregate market response models which previous reviews have considered either incompletely or not at all. Ignoring these issues could lead to biased estimates of the effects of marketing variables and finally erroneous implications for marketing decision making. We recall the main characteristics of several more frequently applied parametric market response functions. In the next chapters we review relevant studies indicating both methods applied and results obtained. We start by presenting flexible aggregate market response models. Their non-parametric component is often specified as multilayer perceptron, spline regression, or kernel regression. We then deal with latent heterogeneity both across households and across retail chains or stores. In the next section we explain endogeneity in aggregate market response models focusing on instrumental variables estimation techniques. We finish by summarizing the implications of this overview and offering an outlook on open research problems.