The nested logit demand system is a popular type of econometric model that has been applied in several merger cases to estimate the demand for a particular good.
It consists of a discrete choice model where consumer decisions to buy a good depend on that good’s price and characteristics (e.g. size, content, quality etc.). The nest structure allows the degree of substitutability between products belonging to the same nest to be stronger than the substitutability between products in different nests. This means that a particular consumer who chooses to buy a particular good (e.g. a Ferrari car) is more likely to choose from goods within the same category (sports car) if there is a rise in the price of the first choice good.
The nested logit demand system offers more flexibility than the standard logit model since the independence of irrelevant alternatives assumption IIA (according to which switching between products takes places in proportion to market shares) only has to hold within a nest. In addition, an advantage of the model is the limited number of parameters required for estimation. However, this comes at the cost of imposing a significant structure on the estimated parameters. In particular, the assumption that switching between goods is proportional to market shares within a nest requires an appropriate specification for each nest. EE&MC has considerable expertise in logit demand systems.