Demand estimation with high-dimensional product characteristics

Benjamin J. Gillen, Matthew Shum, Hyungsik Roger Moon

Research output: Contribution to journalArticle

2 Citations (Scopus)

Abstract

Structural models of demand founded on the classic work of Berry, Levinsohn, and Pakes (1995) link variation in aggregate market shares for a product to the influence of product attributes on heterogeneous consumer tastes. We consider implementing these models in settings with complicated products where consumer preferences for product attributes are sparse, that is, where a small proportion of a high-dimensional product characteristics influence consumer tastes. We propose a multistep estimator to efficiently perform uniform inference. Our estimator employs a penalized pre-estimation model specification stage to consistently estimate nonlinear features of the BLP model. We then perform selection via a Triple-LASSO for explanatory controls, treatment selection controls, and instrument selection. After selecting variables, we use an unpenalized GMM estimator for inference. Monte Carlo simulations verify the performance of these estimators.

Original languageEnglish
Pages (from-to)301-323
Number of pages23
JournalAdvances in Econometrics
Volume34
DOIs
Publication statusPublished - 2014 Jan 1

Fingerprint

Product characteristics
Demand estimation
Estimator
Inference
Product attributes
Heterogeneous consumers
Model specification
Consumer preferences
Monte Carlo simulation
GMM estimator
Proportion
Market share
Structural model

All Science Journal Classification (ASJC) codes

  • Economics and Econometrics

Cite this

Gillen, Benjamin J. ; Shum, Matthew ; Moon, Hyungsik Roger. / Demand estimation with high-dimensional product characteristics. In: Advances in Econometrics. 2014 ; Vol. 34. pp. 301-323.
@article{2393e9bd19b5440992164947b68dd33e,
title = "Demand estimation with high-dimensional product characteristics",
abstract = "Structural models of demand founded on the classic work of Berry, Levinsohn, and Pakes (1995) link variation in aggregate market shares for a product to the influence of product attributes on heterogeneous consumer tastes. We consider implementing these models in settings with complicated products where consumer preferences for product attributes are sparse, that is, where a small proportion of a high-dimensional product characteristics influence consumer tastes. We propose a multistep estimator to efficiently perform uniform inference. Our estimator employs a penalized pre-estimation model specification stage to consistently estimate nonlinear features of the BLP model. We then perform selection via a Triple-LASSO for explanatory controls, treatment selection controls, and instrument selection. After selecting variables, we use an unpenalized GMM estimator for inference. Monte Carlo simulations verify the performance of these estimators.",
author = "Gillen, {Benjamin J.} and Matthew Shum and Moon, {Hyungsik Roger}",
year = "2014",
month = "1",
day = "1",
doi = "10.1108/S0731-905320140000034020",
language = "English",
volume = "34",
pages = "301--323",
journal = "Advances in Econometrics",
issn = "0731-9053",
publisher = "JAI Press",

}

Demand estimation with high-dimensional product characteristics. / Gillen, Benjamin J.; Shum, Matthew; Moon, Hyungsik Roger.

In: Advances in Econometrics, Vol. 34, 01.01.2014, p. 301-323.

Research output: Contribution to journalArticle

TY - JOUR

T1 - Demand estimation with high-dimensional product characteristics

AU - Gillen, Benjamin J.

AU - Shum, Matthew

AU - Moon, Hyungsik Roger

PY - 2014/1/1

Y1 - 2014/1/1

N2 - Structural models of demand founded on the classic work of Berry, Levinsohn, and Pakes (1995) link variation in aggregate market shares for a product to the influence of product attributes on heterogeneous consumer tastes. We consider implementing these models in settings with complicated products where consumer preferences for product attributes are sparse, that is, where a small proportion of a high-dimensional product characteristics influence consumer tastes. We propose a multistep estimator to efficiently perform uniform inference. Our estimator employs a penalized pre-estimation model specification stage to consistently estimate nonlinear features of the BLP model. We then perform selection via a Triple-LASSO for explanatory controls, treatment selection controls, and instrument selection. After selecting variables, we use an unpenalized GMM estimator for inference. Monte Carlo simulations verify the performance of these estimators.

AB - Structural models of demand founded on the classic work of Berry, Levinsohn, and Pakes (1995) link variation in aggregate market shares for a product to the influence of product attributes on heterogeneous consumer tastes. We consider implementing these models in settings with complicated products where consumer preferences for product attributes are sparse, that is, where a small proportion of a high-dimensional product characteristics influence consumer tastes. We propose a multistep estimator to efficiently perform uniform inference. Our estimator employs a penalized pre-estimation model specification stage to consistently estimate nonlinear features of the BLP model. We then perform selection via a Triple-LASSO for explanatory controls, treatment selection controls, and instrument selection. After selecting variables, we use an unpenalized GMM estimator for inference. Monte Carlo simulations verify the performance of these estimators.

UR - http://www.scopus.com/inward/record.url?scp=84912121446&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=84912121446&partnerID=8YFLogxK

U2 - 10.1108/S0731-905320140000034020

DO - 10.1108/S0731-905320140000034020

M3 - Article

AN - SCOPUS:84912121446

VL - 34

SP - 301

EP - 323

JO - Advances in Econometrics

JF - Advances in Econometrics

SN - 0731-9053

ER -