Skip to Content
Purdue Mitch Daniels School of Business logo
Purdue Mitch Daniels School of Business logo
Federico Rossi

Research

Interests

Pricing, market frictions, retail/social media competition.

Publications

We propose an approach to extend the standard framework of dynamic games to settings with multi-product firms. Our approach applies to industries with a large number of products offered by a small number of firms.

Price transparency initiatives are typically undertaken by third parties to ensure that consumers can compare the prices of competing offers in markets where obtaining such information is costly. Such practices have recently become widespread, yet it is unclear whether the increased price competition due to lower search costs overcomes the potential for collusion between competitors due to lower price coordination costs. Motivated by this question, we investigate the effect of mandatory price posting (on large electronic signs) by competing gas stations on their pricing behavior in the Italian motorway. When prices are posted, the average price of gasoline decreases by 1 cent per liter, which represents about 20% of stations' margins. About half the price decrease can be attributed to the introduction of a sign posting a station's own price and those of its nearest neighbors, with the other half coming from the introduction of other signs posting the prices of other stations on the same road. Despite the price reduction, however, the introduction of signs seems to have little impact on price dispersion, suggesting that price uncertainty persists even after the policy is implemented. Analysis of customer transaction data confirms this finding, showing that less than 10% of consumers use the posted prices effectively.

Despite criticisms regarding their effectiveness, reward programs today represent a prevalent and apparently successful form of marketing investment for several industries such as airlines, hotels, gas stations, and credit cards. To understand what might contribute to their success, this paper investigates the purchase behavior of consumers who participate in a reward program from the travel industry. We estimate a dynamic demand model of gasoline purchase, and compare the value that consumers attach to rewards with the value they attach to the money spent for gasoline. We find that there exists a significant portion of frequent travelers who extract more value from one dollar's worth of rewards than from one dollar spent for gasoline. Most of the effectiveness of the reward program is due to the behavior of these consumers. Their insensitivity to price induces firms to increase fuel prices by 1 euro cent per liter, which corresponds to about 10% increase in their margins. The program also contributes to soften competition, increasing margins by 4%.

We quantify the effect of consumers' price uncertainty on gasoline prices and margins on an Italian highway. We observe the change in prices triggered by a longitudinal policy-based change in consumers' price information from one in which drivers on the highway had no information on the prices of stations they encountered to one that allows consumers to observe the prices of four upcoming stations on a single price sign by the side of the highway. Using these data, we estimate a model of consumer search and purchase behavior and a corresponding model of gas station pricing. We then measure the impact of varying degrees of price information on equilibrium prices, including (i) no price information, (ii) the current policy and (iii) full price information. We also compare the current policy with an alternative policy in which stations' prices are advertised with individual price signs. We find that when consumers do not have price information, gas stations are able to charge 31% more, in terms of higher price-cost margins, than when prices are known. Our welfare analysis suggests that price information is worth 57 euro cents to consumers every time they take the highway. Relative to the current mandatory policy, advertising price on individual signs is worth 19 euro cents more to consumers.

Despite the increasing use of social media among personalities such as politicians, athletes, and entertainment celebrities, little is known about the intensity of competition that these popular users engage in to draw the attention of other users, and how their competition affects the users' engagement with social networks. In this research we propose a model where social media users supply content in return for user attention. Using Twitter data on soccer players from the National Women's Soccer League (NWSL), we estimate a demand model where users decide how to allocate their attention among players, based on their content posted on social media and their performance on the soccer field. We consider the amount of tweets mentioning a player's account as a measure for the level of attention captured by the player. On the supply side, players decide the amount of social media content posted on the platform. We show that the attention substitution between players depends on their posting activity and soccer performance, but also on personal characteristics, such as physical attractiveness and team affiliation. Our analysis suggests that the competitive pressure to capture user attention is responsible for about one out of three tweets posted by players. This additional content benefits the social network, increasing by 7% the users' activity on the platform. We also quantify the effect on user activity of a revenue-sharing model in which the platform rewards players for posting tweets.

The analysis of mergers in industries with differentiated products has traditionally focused its attention on postmerger price changes, ignoring the effect of a new competitive landscape on the characteristics of the products firms choose to offer. We propose a new analysis, which includes the product entry and assortment decision of multi-product firms, and simulates the competitive effects of a merger both in terms of changes to prices and to products offered. Using supermarket scanner data and historic information on products introduced and scrapped, we estimate a dynamic structural model of product entry and pricing, which we use to simulate firms' post-merger behavior and compute welfare effects. While solving the dynamic model is nearly unfeasible, due to the large number of products in the market, we recast the model using a different state space that significantly reduce the number of variables required. This approach implies using a nested logit model demand system, which we show provides similar results to the random-coefficient logit model previously estimated on the same data. The results show that allowing firms to change product entry and assortment in addition to prices leads in most cases to a further loss in consumer welfare due to a lower incentive of merging firms to introduce new products. The additional loss can be as high as 50% of the welfare loss predicted by traditional pricing analysis.

Past research shows that loyalty programs can generate switching costs for consumers and increase their purchase frequency. Theoretical work suggests that if switching costs are significant, firms should charge lower prices in the early periods of a program to boost market share, and increase prices in later periods, to take advantage of the ``lock-in'' effect. However, it is not clear whether these costs soften or exacerbate price competition. Using a large database of gas stations' prices in the Italian market, we study fuel prices in early and late periods of loyalty programs: the sharp price changes adopted by gas stations affiliated with the program during the introduction and termination dates of the program allow us to establish the causal relationship between the program and the pricing behavior of gas stations. We find evidence that gas stations affiliated with the program increase prices in later periods of the program, as predicted by theory. The higher prices of affiliated stations lead to an increase in prices, on average, across all other stations competing in the local market. We also find that affiliated stations reduce their prices in early periods of the program; however, this evidence is less conclusive, as our data cannot exclude other factors explaining the price reduction. We discuss implications for managers and policy makers.

 

Working Papers

Street parking offers convenience for customers driving to their favorite retail destinations. Yet, city officials often raise parking fees in an effort to reduce traffic congestion and enhance urban living. The impact of these fee hikes on local retailers remains ambiguous. On one hand, higher parking costs might deter potential customers from visiting the retail area; on the other, improved traffic flow and more efficient parking turnover could enhance consumer access. Using geolocation data and leveraging a recent policy introduced by the City of Chicago, we measure the impact of an increase in parking fees on consumer footfall at local restaurants. We find that a small increase in hourly street parking fees reduces the average footfall at restaurants without customer parking by as much as 15%, while boosting average footfall at nearby competing restaurants that offer parking by around 7%. In contrast, we do not find evidence of any positive effect of the parking fee hike on overall restaurant footfall. Our estimates also indicate that this policy change disproportionately affects the economically poorer sections of the population, reducing their access to these restaurants by twice as much as compared to other groups.

Consumer impatience, i.e. customers' desire to receive online purchases as quickly as possible, is a key factor in shopping decisions. Despite that, there is little empirical research on it, and on the effect on the market of a new technology reducing delivery time. In this research, we show that consumer impatience significantly affects market structure. Using online purchase data on pizza delivery, we show that it reduces substitution, softens price competition, and allows low-quality pizzerias to survive in the market. The introduction of a new technology, which cuts the transit time from the pizzeria to the consumer's door by 75%, would double the market share of the ten top-quality pizzerias, while a third of all pizzerias, especially those located in the downtown area, would face significant losses and likely shut down. We also show that the platform collecting delivery orders would significantly increase its profits by offering a price discrimination policy that includes a premium delivery service to the most impatient customers.

The analysis of mergers in industries with differentiated products has traditionally focused its attention on postmerger price changes, ignoring the effect of a new competitive landscape on the characteristics of the products firms choose to offer. We propose a new analysis, which includes the product entry and assortment decision of multi-product firms, and simulates the competitive effects of a merger both in terms of changes to prices and to products offered. Using supermarket scanner data and historic information on products introduced and scrapped, we estimate a dynamic structural model of product entry and pricing, which we use to simulate firms' post-merger behavior and compute welfare effects. While solving the dynamic model is nearly unfeasible, due to the large number of products in the market, we recast the model using a different state space that significantly reduce the number of variables required. This approach implies using a nested logit model demand system, which we show provides similar results to the random-coefficient logit model previously estimated on the same data. The results show that allowing firms to change product entry and assortment in addition to prices leads in most cases to a further loss in consumer welfare due to a lower incentive of merging firms to introduce new products. The additional loss can be as high as 50% of the welfare loss predicted by traditional pricing analysis.