Endogeneity

The exogenous deterministic components of utility models such as price, promotion, advertising, etc., are themselves endogenous.  In fact, marketing-mix variables, such as price, not only affect firm’s demand function but are themselves endogenously set by marketing managers based on market information which may be in part unobserved to the reseacher but affects consumer demand.  This would create a situation where the marketing-mix variables could be correlated with the error terms in the latent utilities.

Failure to account for endogeneity in the deterministic componenets of utility has the potential to bias the parameter estimates of the marketing-mix variables.  Will lead to misleading results for diagnostic and optimization purposes of marketing decisions.

Three approaches to solve the problem:

1. IV (instrumental variables) approach: Use IV to “proxy” the endogenous variables.  IV are exogenous, uncorrelated with the troublesome unobserved driver, but correlated with the observed endogenous variable.  Key challenge is to identify the appropriate instruments and the potential drawback is that the estimates are inefficient if IVs are not good.

2. Impose the relationship between the observed and unobserved variables while estimating the parameters. (Estimates are efficient is this case)

3. Avoid endogeneity entirely if the marketing-mix variables are experimentally determined by firms.  Researchers can promote the use of experiementation and randomization by firms from which they obtain their data, and then the need to address the endogeneity issue is alleviated.