Jean-Pierre Dubé

An Empirical Model of Advertising Dynamics

This paper develops a model of dynamic advertising competition, and applies it to the problem of optimal advertising scheduling through time. In many industries we observe advertising “pulsing”, whereby firms systematically switch advertising on and off at a high-frequency. Hence, we observe periods of zero and non-zero advertising, as opposed to a steady level of positive advertising. Previous research has rationalized pulsing through two features of the sale response function: an S-shaped response to advertising, and long-run effects of current advertising on demand. Despite considerable evidence for advertising carry-over, existing evidence for non-convexities in the shape of the sales-response to advertising has been limited and, often, mixed. We show how both features can be included in a discrete choice based demand system and estimated using a simple partial maximum likelihood estimator. The demand estimates are then taken to the supply side, where we simulate the outcome of a dynamic game using the Markov perfect equilibrium (MPE) concept. Our objective is not to test for the specific game generating observed advertising levels. Rather, we wish to verify whether the use of pulsing (on and off) can be justified as an equilibrium advertising practice. We solve for the equilibrium using numerical dynamic programming methods. The flexibility provided by the numerical solution method allows us to improve on the existing literature, which typically considers only two competitors, and places strong restrictions on the demand models for which the supply side policies can be obtained. We estimate the demand model using data from the Frozen Entree product category. We find evidence for a threshold effect, which is qualitatively similar to the aforementioned S-shaped advertising response. We also show that the threshold is robust to functional form assumptions for the marginal impact of advertising on demand. Our estimates, which are obtained without imposing any supply side restrictions, imply that firms should indeed pulse in equilibrium. Predicted advertising in the MPE is higher, on average, than observed advertising. On average, the optimal advertising policies yield a moderate profit improvement over the profits under observed advertising. Copyright Springer Science + Business Media, Inc. 2005

Do Switching Costs Make Markets Less Competitive

The conventional wisdom in economic theory holds that switching costs make markets less competitive. This article challenges this claim. The authors formulate an empirically realistic model of dynamic price competition that allows for differentiated products and imperfect lock-in. They calibrate this model with data from frequently purchased packaged goods markets. These data are ideal in the sense that they have the necessary variation to identify switching costs separately from consumer heterogeneity. Equally important, consumers exhibit inertia in their brand choices, a form of psychological switching cost. This makes the results applicable to the broad range of products that are distinctly identified (i.e., branded) rather than just to products for which there is a product adoption cost or explicit switching fee. In the simulations, prices are as much as 18% lower with than without switching costs. More important, equilibrium prices do not increase even in the presence of switching costs that are of the same order of magnitude as product price.

IMPROVING THE NUMERICAL PERFORMANCE OF STATIC AND DYNAMIC AGGREGATE DISCRETE CHOICE RANDOM COEFFICIENTS DEMAND ESTIMATION

The widely used estimator of Berry, Levinsohn, and Pakes (1995 )p roduces estimates of consumer preferences from a discrete-choice demand model with random coefficients, market-level demand shocks, and endogenous prices. We derive numerical theory results characterizing the properties of the nested fixed point algorithm used to evaluate the objective function of BLP’s estimator. We discuss problems with typical implementations, including cases that can lead to incorrect parameter estimates. As a solution, we recast estimation as a mathematical program with equilibrium constraints, which can be faster and which avoids the numerical issues associated with nested inner loops. The advantages are even more pronounced for forward-looking demand models where the Bellman equation must also be solved repeatedly. Several Monte Carlo and real-data experiments support our numerical concerns about the nested fixed point approach and the advantages of constrained optimization. For static BLP, the constrained optimization approach can be as much as ten to forty times faster for large-dimensional problems with many markets.

Competitive Price Discrimination Strategies in a Vertical Channel Using Aggregate Retail Data

We explore opportunities for targeted pricing for a retailer that only tracks weekly store-level aggregate sales and marketing-mix information. We show that it is possible, using these data, to recover essential features of the underlying distribution of consumer willingness to pay. Knowledge of this distribution may enable the retailer to generate additional profits from targeting by using choice information at the checkout counter. In estimating demand we incorporate a supply-side model of the distribution channel that captures important features of competitive price-setting behavior of firms. This latter aspect helps us control for the potential endogeneity generated by unmeasured product characteristics in aggregate data. The channel controls for competitive aspects both between manufacturers and between manufacturers and a retailer. Despite this competition, we find that targeted pricing need not generate the prisoners dilemma in our data. This contrasts with the findings of theoretical models due to the flexibility of the empirical model of demand. The demand system we estimate captures richer forms of product differentiation, both vertical and horizontal, as well as a more flexible distribution of consumer heterogeneity.

Improving the Numerical Performance of BLP Static and Dynamic Discrete Choice Random Coefficients Demand Estimation

The widely-used estimator of Berry, Levinsohn and Pakes (1995) produces estimates of consumer preferences from a discrete-choice demand model with random coefficients, market-level demand shocks and endogenous prices. We derive numerical theory results characterizing the properties of the nested fixed point algorithm used to evaluate the objective function of BLPs estimator. We discuss problems with typical implementations, including cases that can lead to incorrect parameter estimates. As a solution, we recast estimation as a mathematical program with equilibrium constraints, which can be faster and which avoids the numerical issues associated with nested inner loops. The advantages are even more pronounced for forward-looking demand models where Bellmans equation must also be solved repeatedly. Several Monte Carlo and real-data experiments support our numerical concerns about the nested fixed point approach and the advantages of constrained optimization.

Tipping and Concentration in Markets with Indirect Network Effects

This paper develops a framework for measuring “tipping”---the increase in a firms market share dominance caused by indirect network effects. Our measure compares the expected concentration in a market to the hypothetical expected concentration that would arise in the absence of indirect network effects. In practice, this measure requires a model that can predict the counterfactual market concentration under different parameter values capturing the strength of indirect network effects. We build such a model for the case of dynamic standards competition in a market characterized by the classic hardware/software paradigm. To demonstrate its applicability, we calibrate it using demand estimates and other data from the 32/64-bit generation of video game consoles, a canonical example of standards competition with indirect network effects. In our example, we find that indirect network effects can lead to a strong, economically significant increase in market concentration. We also find important roles for beliefs on both the demand side, as consumers tend to pick the product they expect to win the standards war, and on the supply side, as firms engage in penetration pricing to invest in growing their networks.

Balancing Profitability and Customer Welfare in a Supermarket Chain

We investigate the impact of price discrimination by a large Chicago supermarket chain. First we measure the impact of the chains current zone-pricing policy on shelf prices, variable profits and consumer welfare across its stores. Using the chains database to simulate a finer store-specific micro-pricing policy, we study the implications of this policy on profits and welfare. We show how a store-pricing policy that is constrained to offer consumers at least as much surplus as a uniform chain wide pricing policy still enables the retailer to generate substantial incremental profits.To ensure our pricing problem exhibits a well-defined optimum, we use the parsimonious, mixed-logit demand function that allows for flexible substitution patterns across brands and also retains a link to consumer theory. We discuss the issue of price endogeneity when estimating the demand parameters with weekly store-level data. Standard instrumental variables techniques used to account for such endogeneity also seem to increase the magnitudes of own-price elasticities thereby offsetting the problem encountered by previous researchers of predicted prices from a demand model exceeding those in the actual data.

Beyond the Endogeneity Bias: The Effect of Unmeasured Brand Characteristics on Household-Level Brand Choice Models

We investigate the role of potential weekly brand-specific characteristics that influence consumer choices, but are unobserved or unmeasurable by the researcher. We use an empirical approach, based on the estimation methods used for standard random coefficients logit models, to account for the presence of such unobserved attributes. Using household scanner panel data, we find evidence that ignoring such time-varying latent (to the researcher) characteristics can lead to two types of problems. First, consistent with previous literature, we find that these unobserved characteristics may lead to biased estimates of the mean price response parameters. This argument is based on a form of price endogeneity. If marketing managers set prices based on consumer willingness to pay, then the observed prices will likely be correlated with the latent (to the researcher) brand characteristics. We resolve this problem by using an instrumental variables procedure. Our findings suggest that simply ignoring these attributes may also lead to larger estimates of the variance in the heterogeneity distribution of preferences and price sensitivities across households. This could overstate the benefits from marketing activities such as household-level targeting. We resolve the problem by using weekly brand intercepts, embedded in a random coefficients brand choice model, to control for weekly brand-specific characteristics, while accounting for household heterogeneity. Overall, our results extend the finding on the endogeneity bias from the mean of the heterogeneity distribution (i.e., the price effect) to include the variance of that distribution.

Brand history, geography and the persistence of brand shares

We document evidence of a persistent “early entry” advantage for brands in 34 consumer packaged goods industries across the 50 largest U.S. cities. Current market shares are higher in markets closest to a brand’s historic city of origin than in those farthest. For six industries, we know the order of entry among the top brands in each of the markets. We find an early entry effect on a brand’s current market share and perceived quality across U.S. cities. The magnitude of this effect typically drives the rank order of market shares and perceived quality levels across cities.

Category Pricing with State-Dependent Utility

There is substantial literature documenting the presence of state-dependent utility with packaged goods data. Typically, a form of brand loyalty is detected whereby there is a higher probability of purchasing the same brand as has been purchased in the recent past. The economic significance of the measured loyalty remains an open question. We consider the category pricing problem and demonstrate that the presence of loyalty materially affects optimal pricing. The prices of higher quality products decline relative to those of lower quality when loyalty is introduced into the model. Given the well-known problems with the confounding of state dependence and consumer heterogeneity, loyalty must be measured in a model which allows for an unknown and possibly highly nonnormal distribution of heterogeneity. We implement a highly flexible model of heterogeneity using multivariate mixtures of normals in a hierarchical choice model. We use an Euler equations approach to the solution of the dynamic pricing problem which allows us to consider a very large number of consumer types.