Nonparametric Identification of Demand without Exogenous Product Characteristics

Kirill Borusyak and Jiafeng Chen and Peter Hull and Lihua Lei · 2025

Read Nonparametric Identification of Demand without Exogenous Product Characteristics by Kirill Borusyak and Jiafeng Chen and Peter Hull and Lihua Lei (2025) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/arxiv-2512-23211.txt.

Introduction

Kirill Borusyak, Jiafeng Chen, Peter Hull, and Lihua Lei revisit a bleak result in the econometrics of product demand. Demand models for varied products are fit to market-level data, and product characteristics may be endogenous, chosen by firms knowing demand. Earlier work suggested that without instruments built from exogenous characteristics, flexible demand models cannot be pinned down, no matter how clever the price instruments.

The authors show the discouragement was premature for the questions economists care about. Price counterfactuals, what demand would look like at different prices, are identified using recentered instruments that mix price variation from outside with the possibly endogenous characteristics themselves. The proof needs a weaker index restriction than usual plus a new condition they call faithfulness, which they argue is a technical strength test akin to completeness, not an economic assumption. The paper closes with guidance for estimating demand counterfactuals with fitted models.

Catalog record

CREATORKirill Borusyak and Jiafeng Chen and Peter Hull and Lihua Lei
YEAR2025
COLLECTIONPAPERS
FORMATTXT
FILE SIZE108 KB
VERSION1.0
LICENSEhttp://creativecommons.org/licenses/by/4.0/
RIGHTSCC BY
SOURCEarXiv

Source: https://arxiv.org/abs/2512-23211

Browse the full library · About the Tower of Babel

Full text

The opening of the work, from the library's hosted text:

Nonparametric Identification of Demand without Exogenous Product Characteristics Kirill Borusyak and Jiafeng Chen and Peter Hull and Lihua Lei Abstract . We study the identification of differentiated product demand with exogenous supply-side instruments, allowing product characteristics to be endogenous. Past analyses have argued that exogenous characteristic-based instruments are essentially necessary given a sufficiently flexible demand model with a suitable index restriction. We show, however, that price counterfactuals are nonparametrically identified by recentered instruments—which combine exogenous shocks to prices with endogenous product characteristics—under a weaker index restriction and a new condition we term faithfulness. We argue that faithfulness, like the usual completeness condition for nonparametric identification with instruments, can be viewed as a technical requirement on the richness of identifying variation rather than a substantive economic restriction, and we show that it holds under a variety of non-nested conditions on either price-setting or the index. JEL codes. C14, C36, L13 Keywords. Demand, nonparametric identification, recentered instruments, faithfulness 1. Introduction Classic parametric models of demand are often estimated with supply-side instruments for prices ( 36 ; 25 ; 14 , e.g.,) . Intuitively, exogenous variation in supply conditions across markets can trace out demand curves which can then be used to answer important counterfactuals—i.e., how quantities and welfare would change under hypothetical price changes. Modern analyses with more flexible demand models, however, often use other kinds of instruments: functions of the observed characteristics of competing products ( 10 ; 11 ; 24 , e.g.,) . It is well-known that such characteristic-based instruments can be invalid if product entry is strategic or if characteristic choice is otherwise endogenous ( 10 ; 32 ) , and they can moreover reduce robustness to model misspecification ( 5 ) . Can other instruments be found to avoid these issues? Are exogenous shocks to prices enough? In influential work, 8 argue that the answer is generally negative: i.e., that exogenous characteristic-based instruments are essentially necessary to flexibly estimate demand with market-level data. Specifically, they consider a nonparametric demand model of the form: (1) \displaystyle\sigma^{-1}(S,P)=X+\xi\equiv\delta, where S and P are J -vectors of the observed quantity shares and prices of J goods in a market, X is a J -vector of an observed product characteristic, and \xi is an unobserved J -vector of demand shocks. The demand shocks and characteristics combine linearly in an index \delta . The unknown \sigma^{-1}(\cdot) is the inverse of a demand function \sigma(d,p) giving the market shares that would be observed if \delta were set to d and P were set to p . 1 1 1 \sigma(\cdot) may also depend on other observable characteristics \tilde{X} , which we suppress here for exposition. 8 argue that observing a J -vector of exogenous price instruments Z , which shift S through P while being unrelated to \xi , is not enough to identify \sigma(\cdot) or even to identify price counterfactuals (e.g., functions of \partial\sigma/\partial p ). 2 2 2 This latter claim is most directly stated in a review, 12 : “having valid instruments for all J prices will not generally suffice for identification of […] the ceteris paribus effects of price changes.” Intuitively, the inverse demand function \sigma^{-1}(\cdot) depends on S and P separately but Z only affects the former through the latter; to trace out demand we would thus seem to need other instruments that shift market shares given prices. 3 3 3 8 also consider a model in which P enters the index and is excluded from \sigma^{-1}(\cdot) , concluding that price instruments suffice for identification, though again assuming that characteristics are exogenous. 15 show that recentered instruments identify price counterfactuals in this case, allowing endogenous characteristics. The characteristic X that enters the index uniquely fits this bill, so long as it is also exogenous. All functions of (X,Z) can then serve as instruments and identify \sigma(\cdot) provided they induce enough variation in (S,P) —a technical condition known as completeness ( 30 ) . We reexamine this setting and arrive at a different, more optimistic conclusion: exogenous shocks to prices can suffice for nonparametric identification of price counterfactuals. Intuitively, even when characteristics are endogenous, functions of X and an as-good-as-randomly assigned Z that are mean-zero given X —what 16 call recentered instruments—are still valid instruments and have identifying power. In fact, they exhaust the exogenous variation in Z\mid X and can rule out some candidate demand functions \check{\sigma}(\cdot) through a simple test: if \check{\delta}=\check{\sigma}^{-1}(S,P) correlates with a recentered instrument, then \check{\sigma}(\cdot) cannot be the true demand function. We introduce a new technical condition— faithfulness —under which all price counterfactuals are identified by the candidate demand functions that survive this test. Specifically, we show that, under faithfulness, any surviving \check{\sigma}^{-1}(\cdot) equals the true \sigma^{-1}(\cdot) up to some unknown transformation, which is inconsequential for price counterfactuals. Naturally, this argument does not fully identify the demand function, as we cannot learn counterfactuals that change the endogenous characteristics. But for identifying price counterfactuals, we show the other characteristics need not be exogenous and that conventional characteristic-based instruments may be unnecessary. 4 4 4 As we discuss below, allowing characteristics to be endogenous further relaxes the index restriction in the main text of 8 by allowing \delta to be an arbitrary function of X and \xi . The new faithfulness condition intuitively requires the price instruments to induce enough exogenous variation to make all causal effects of P detectable. Specifically, it says that functions of the form H(\delta,P) cannot be mean-independent of Z given X unless they are constant in P ; otherwise, there are causal effects of P on H(\delta,P) that are undetected by Z . This condition is useful because each candidate \check{\sigma}^{-1}(\cdot) can be written as such a function: \check{\sigma}^{-1}(S,P)=\check{\sigma}^{-1}(\sigma(\delta,P),P)\equiv\check{H}(\delta,P) . Thus, if faithfulness holds, any \check{\sigma}^{-1}(S,P) that survive the recentered instrument test must be a transformation of the true \sigma^{-1}(S,P)=\delta : i.e., \check{H}(\delta,P)=\check{H}(\delta)=\check{H}(\sigma^{-1}(S,P)) . Faithfulness is best seen as a technical restriction on the richness of identifying variation—like completeness—rather than a substantive economic assumption. The two conditions are indeed similar: completeness says that functions of the form H(S,P) cannot be mean-independent of (X,Z) unless they are constant in (S,P) . Faithfulness focuses on the richness of exogenous price variation given potentially endogenous characteristics, in line with the goal of identifying price counterfactuals. We establish several non-nested conditions under which the two technical conditions are quite close. One simple but instructive case is when \delta and X have finite support with the same number of values; we show that completeness implies faithfulness in this case, suggesting that the gap between the two assumptions can only arise in the presence of infinite-dimensional objects. We also show that when prices are exogenous (i.e., P=Z )—a setting that past analyses have characterized as surprisingly difficult for identification ( 12 ) —faithfulness follows from completeness. We further develop several conditions on either pricing or the utility index under which faithfulness also follows from completeness. On the pricing side, it is sufficient that either (i) X and Z combine in a nonparametric index \lambda such that P is independent of (X,Z) given (\lambda,\delta) or (ii) the derivatives of P with respect to Z satisfy a particular separability condition. Both conditions are compatible with firms engaging in Bertrand–Nash pricing

Read the complete text