by johnr » Mon Sep 16, 2013 11:35 am
Hi Anat
Respondents for discrete choice models can be any agent who makes a decision, not just consumers. The agents can be businesses, governments, etc. Provided a discrete choice is the unit of analysis, the theory remains the same. Note that numerous studies on business choices have been conducted over the years. Indeed, the econometrics of discrete choice models, outside of the behavioural interpretations, is such that any discrete outcome can be modelled independent of the type of agency or agents being modelled. For example, we have applied discrete choice models to stock exchange data where the outcomes were whether a firm was observed to have gone bankrupt, merged, or continued unchanged over the period of the data. If you are looking at a behavioural model of SME, then you simply need to assume that firms act in a behaviourally similar manner to individuals - that is they are utility maximisers.
The typical constraint in getting data on business choices is identifying the key decision makers from the firms and getting them to respond. Data collection is often, in our experience, much more expensive, however if one has the resources, then there is no reason to do this.
John