Nested Logit Brand Choice
Also known as: Nested Multinomial Logit, Hierarchical Choice Model, Tree-Structured Logit, GEV Nested Logit
The nested logit model of brand choice relaxes the restrictive independence-of-irrelevant-alternatives (IIA) assumption of the standard multinomial logit by grouping similar alternatives into nests. Developed by Daniel McFadden as a member of the generalized-extreme-value (GEV) family, it allows the unobserved utilities of alternatives within the same nest to be correlated while keeping a tractable closed form. In a brand-choice setting the natural structure is a tree: consumers first effectively choose a category, sub-category, or product form and then a brand within it, with an inclusive-value term carrying the expected utility of the lower level up to the upper level. The dissimilarity parameter on each nest measures within-nest correlation and reduces to ordinary logit when it equals one. The result is a model whose substitution patterns are far more realistic than plain logit — a price cut on one brand draws disproportionately from its nest-mates — while remaining estimable by maximum likelihood. It is a workhorse for choice analysis when alternatives fall into obvious clusters.
Key highlights
- Relaxes the IIA assumption so that substitution is stronger among similar brands, yielding realistic cross-price elasticities.
- Retains a closed-form likelihood, making it far faster and more stable to estimate than simulation-based mixed logit.
- The dissimilarity parameter provides a direct, interpretable test of whether IIA holds within a branch of the tree.
- Naturally encodes a managerially meaningful hierarchy such as category-to-brand or tier-to-brand decision making.
Intuition
This section is available to Pro members. Upgrade to Pro
How it works
This section is available to Pro members. Upgrade to Pro
When to use it
Use nested logit for brand choice when the alternatives fall into natural groups of close substitutes and you suspect that the multinomial logit's proportional-substitution (IIA) assumption is violated. It is appropriate with disaggregate choice data — scanner panel, survey, or choice-experiment — where you observe the full choice set and alternative-specific attributes such as price and promotion, and where a defensible hierarchy (category to brand, tier to brand, form to brand) can be specified a priori. It is the right tool when you need realistic cross-elasticities for pricing and competitive analysis but want to stay within a closed-form, fast-to-estimate model rather than moving to mixed logit or hierarchical Bayes. It is less suitable when the correct nesting is genuinely unknown or alternatives overlap several groups (consider cross-nested or mixed logit), when individual-level heterogeneity rather than correlated alternatives is the main concern (use random-coefficients models), or when choice sets are tiny enough that IIA is harmless.
Strengths & limitations
- Relaxes the IIA assumption so that substitution is stronger among similar brands, yielding realistic cross-price elasticities.
- Retains a closed-form likelihood, making it far faster and more stable to estimate than simulation-based mixed logit.
- The dissimilarity parameter provides a direct, interpretable test of whether IIA holds within a branch of the tree.
- Naturally encodes a managerially meaningful hierarchy such as category-to-brand or tier-to-brand decision making.
- Requires the analyst to specify the nesting structure in advance; the wrong tree yields biased substitution patterns.
- Each alternative belongs to exactly one nest, so it cannot represent alternatives that are similar along several overlapping dimensions.
- Captures correlation through nests but not continuous random taste heterogeneity across consumers, unlike mixed logit.
- Sequential (level-by-level) estimation is inefficient and can give inconsistent standard errors relative to full-information maximum likelihood.
Common pitfalls
This section is available to Pro members. Upgrade to Pro
Applications
This section is available to Pro members. Upgrade to Pro
Frequently asked
What is IIA and why is it a problem for brand choice?
Independence of irrelevant alternatives (IIA) is a property of multinomial logit stating that the relative odds of choosing between two brands do not depend on what other brands are available. A consequence is proportional substitution: any change to one brand draws share from all others in proportion to their current shares. This is unrealistic for brands, because close substitutes should compete more directly than distant ones. Nested logit relaxes IIA across nests so that a price cut on one brand cannibalizes its nest-mates more heavily than alternatives in other nests, producing the differentiated substitution patterns brand competition actually exhibits.
What does the dissimilarity (inclusive-value) parameter mean?
Each nest has a dissimilarity parameter, often written lambda, that scales utilities within the nest and weights the inclusive value passed up to the nest-choice stage. It measures how correlated the unobserved utilities of the alternatives in that nest are: values near zero mean the nest-mates are very close substitutes, while a value of one means they are no more related than alternatives in other nests, which makes the model collapse to ordinary multinomial logit. For consistency with random-utility theory the parameter must lie between zero and one, and testing whether it equals one is the standard way to check whether nesting is needed.
When should I use nested logit instead of mixed logit or hierarchical Bayes?
Nested logit is the right choice when the source of IIA violation is that alternatives cluster into recognizable groups of close substitutes, and you can specify that hierarchy a priori. It keeps a closed form and estimates quickly. Mixed (random-coefficients) logit and hierarchical Bayes instead address heterogeneity in tastes across consumers and can represent flexible substitution without a predefined tree, but at the cost of simulation- or MCMC-based estimation. In practice, use nested logit when the grouping structure is clear and you want speed and interpretability, and move to mixed logit or hierarchical Bayes when individual-level taste variation is the dominant concern or no single nesting tree is defensible.
Sources
- 1.McFadden, D. (1978). Modelling the Choice of Residential Location. In A. Karlqvist, L. Lundqvist, F. Snickars, & J. Weibull (Eds.), Spatial Interaction Theory and Planning Models (pp. 75-96). North-Holland.ISBN 9780444851826
- 2.McFadden, D. (1980). Econometric Models for Probabilistic Choice Among Products. The Journal of Business, 53(3), S13-S29.
You have read it. What now?
Cite this page
ScholarGate. (2026, June 23). Nested Logit Brand Choice. ScholarGate. https://scholargate.app/marketing/nested-logit-brand-choice