Javier Donna

Howard P. Marvel Scholar Assistant Professor of Economics


joint with Jose-Antonio Espin-Sanchez (Yale University)

Second round revision requested by Econometrica.

This paper was awarded the Best Paper in Regulatory Economics at the IIOC, 2016.

I presented this paper at the 2016 NBER Summer Institute Industrial Organization,
at the Barcelona GSE Summer Forum Applied Industrial Organization,
and the Empirical Microeconomics Workshop.

Please read our short essay in Global Water Forum and our short essay in VOX EU.

AbstractWater is an essential good for human life but there is controversy over whether it should be allocated using markets. In 1966, the irrigation community in Mula (Murcia, Spain) switched from a market institution, a 700 year old auction, to a system of fixed quotas with a ban on trading to allocate water from the town's river. We present a dynamic demand model in which farmers face liquidity constraints (LC) to explain why the new, non-market institution is more efficient. We show that ignoring the presence of LC biases the estimated (inverse) demand and demand elasticity downwards. We use the dynamic demand model and data from the auction period to estimate both farmers' demand for water and their financial constraints, thus obtaining unbiased estimates. In our model, markets achieve the first-best allocation only in the absence of LC. By contrast, quotas achieve the first-best allocation only if farmers are homogeneous in productivity. We compute welfare under both types of institutions using the estimated parameters. We find that the quota is more efficient than the market.
| pdf |       | Online Appendix |       | Note Additional Results |       | Interviews |        |Cite|


joint with Pablo Schenone (Arizona State University) and
Greg Veramendi (Arizona State University)

Under review.

This paper was accepted for presentation at the 2017 NBER Summer Institute Macro Perspectives.

AbstractThis paper uses networks to study price dispersion in seller-buyer markets where buyers with unit demand interact with multiple, but not all, sellers; and buyers and sellers compete on prices after they meet. Our approach allows for ex post indirect competition, where a buyer who is not directly linked with a seller affects the price obtained by that seller. Indirect competition generates the central finding of our paper: price dispersion depends on both the number of links in the network, and how these links are distributed. Networks with very few links can have no price dispersion, while networks with many links can still support significant price dispersion. We present three main theoretical results. First, for any given network we characterize the pairwise stable matchings and the prices that support them. Second, we characterize the set of all graphs where price dispersion is precluded. Third, we use a theorem from Frieze (1985) to show that the graphs where price dispersion is precluded arise asymptotically with probability one in random Poisson networks, even as the probability of each individual link goes to zero. We also provide quantitative results on the finite sample properties of price dispersion in random networks. Finally, we present an application to eBay to show that: (i) a calibration of our model reproduces the price dispersion documented in eBay quite well, and (ii) the amount of price dispersion in eBay would decrease substantially (35-45 percent as measured by the coefficient of variation) in a counterfactual analysis, where we change eBay’s network structure so that links are drawn with equal probability for all sellers and buyers.

| pdf | | Online Appendix | |Cite|


joint with Jose-Antonio Espin-Sanchez (Yale University)

Accepted for publication by The RAND Journal of Economics.

A version of this paper was my job market paper in 2012, and it was awarded the Best Paper Award at EARIE (2012) and JEI (2012)

Abstract: We study sequential auctions in which bidders demand multiple units. We collect a novel data set on sequential water auctions for the empirical study. Although water units are identical, two features from the empirical setting create a trade-off whereby units of water end up being complements or substitutes. First, there is a water loss that is only incurred for the first unit, generating a sunk cost. Second, subsequent units of water exhibit decreasing marginal returns. Units of water are complements or substitutes depending on the relative importance of the sunk cost and decreasing returns. Weather seasonality provides us with the required variation (in sunk costs relative to decreasing returns) to perform the empirical investigation. When units are complements, one bidder wins all units by paying a high price for the first unit, thus deterring others from bidding on subsequent units. When units are substitutes, different bidders win the units with positive probability and pay prices of similar magnitude, even when the same bidder wins all units. We analyze this stark pattern of outcomes not investigated in the literature before. We recover individual demand consistent with this pricing behavior and confirm it is not collusive, but consistent with non-cooperative behavior. Demand estimates are biased if one ignores these features.


joint with Pedro Pereira (Autoridade da Concorrência and CEFAGE),
Tiago Pires (University of North Carolina), and
Andre Trindade (Getulio Vargas Foundation).

I presented this paper at the 2017 Triangle Microeconomics Conference in Honor to Tiago Pires.

A preliminary draft is available upon request.

AbstractWe empirically investigate the welfare implications of intermediaries in oligopolistic markets, where intermediaries offer additional services to differentiate their products from the ones of the manufacturers. Our identification strategy exploits the unique circumstance that, in the outdoors advertising industry, there are two distribution channels: consumers can purchase the product either directly from manufacturers, or with the intermediation of retailers. Using product-level data for the whole industry, we estimate a differentiated products’ equilibrium model that includes: consumers who have preferences that are specific to each distribution channel and engage in costly search on the demand side; and two layers of activity (where manufacturers and intermediaries bar- gain over wholesale prices) with two distribution channels (where the two distribution channels compete a la Bertrand) on the supply side. The estimated model is used to simulate counterfactual scenarios, where intermediaries do not offer additional services. These counterfactuals are used to quantify the welfare effects of intermediaries.

joint with Greg Veramendi (Arizona State University)

Please read our white paper and our short essay in Harvard Business Review. Summarized by The Economist, Fortune, Yahoo! Finance, Slate, Business Travel News, CWT Solutions Group, OnCampus.

AbstractWe examine a unique database of 6.4 million flight bookings in 2014 to analyze gender differences in the demand for planning business travel. We find that female travelers book 2 days earlier than their male counterparts on average. Female travelers pay on average about $17 less per ticket (about 2%) than their male counterparts. We also find that advance booking improves with age. As age increases from 30 to 70, advance booking increases by 5 days for both genders. As trips become more frequent, advanced booking decreases. For high-frequency travelers who do over 2 trips per month, the gender gap in advanced booking becomes negligible.


joint with Pablo Schenone (Arizona State University) and
Greg Veramendi (Arizona State University)

Economic LettersJanuary 2016, 138, 81-84.

AbstractPeters and Severinov (2006) (PS henceforth) characterize a perfect Bayesian equilibrium (PBE) in a competing auctions environment, where all buyers are linked to all the sellers. PS characterize a PBE using a simple bidding rule, whereby buyers select in which auction to bid. In this note we show that when buyers are linked with a subset of the sellers (i.e. when there are search frictions), the PS bidding rule is no longer guaranteed to be efficient nor a PBE of the competing auctions game of PS. Our results indicate that researchers should be cautious when using the PS bidding rule to make inference about the behavior of buyers and sellers in a market where frictions are present such as eBay.

| pdf | |Cite|


joint with Joern Boehnke (Harvard University),
Dimitriy Masterov (eBay Research Labs), and
Greg Veramendi (Arizona State University)

AbstractWe empirically investigate the welfare implications of a reduction in the level of frictions and price dispersion in online retail markets. Our identification strategy exploits the unique circumstance that, in the online trading platform eBay, we have access to “click-stream” data, that records which listings appear in buyers’ searches, which listings buyers click on, and which listings they bid for, even if they do not end up buying that product. Click-stream data provides the information needed to reconstruct the links (which buyers interact with which sellers) in the network. We perform the welfare analysis using a three-step framework. First, we reconstruct the realized network of buyers and sellers using click-stream data. Second, we develop a tractable empirical networks’ model, and estimate its the primitives (i.e. distribution of buyers’ valuations) that characterize the model?s underlying demand preferences conditional on the realized network. Third, we use the estimated demand preferences and network structure to perform a “counterfactual” analysis. In the counterfactual analysis we compare the actual outcome (i.e. welfare under the actual level of frictions in eBay) to a counterfactual outcome, whereby we will use the estimated demand preferences and the behavioral buyer-seller model to compute the welfare under an alternative policy that would reduce the level of frictions and price dispersion in eBay.

| Extended Abstract | 

joint with Jose-Antonio Espin-Sanchez (Yale University)

AbstractWe investigate the role of punishment progressivity and individual characteristics in the determination of crime. To analyze welfare implications we model individuals' response to judges' optimal punishment in a dynamic setting. We introduce two distinctive features motivated by our empirical setting. First, judges rarely imposes maximum punishment for first time offenders. Instead, we observe low fines (or just a warning) even when crime detection technology is efficient and punishment is not costly. We account for this by allowing an unobservable (to the judge) individual state to be correlated with a public signal (the environment). This generates an optimal punishment that is conditional on individual observables. Second, judges punishments follow a progressive system: conditioning on type, recidivists are punished harsher than first-time offenders for the same crime. We account for these dynamics by introducing a persistent unobservable (to the judge) component. Depending on whether the individual committed a crime in the previous period, the judge updates her beliefs about the individual; this gives rise to progressivity in the optimal punishment system. For the empirical analysis we examine a novel trial data set from a self-governed community of farmers in Southern Spain. We find that judges vary the degree of imposed punishments based on individual characteristics—such as when the victim or the accused have a Don honorific title indicating he is a wealthy person. Recidivists are punished harsher than first time offenders.

| pdf |  |Cite|


Abstract: I investigate the effect of switching costs in the transportation market in Chicago. Demand is represented by a continuum of agents that are persistently heterogeneous and forward looking. Each period agents may choose among car use, public transportation or an outside option. Consumers incur a fixed cost every time they decide to switch from one mode of transportation to another. In subsequent periods, no switching costs are incurred if the agent chooses the same mode of transportation. I structurally estimate the dynamic discrete choice model using monthly data on ridership, traffic volume counting, and gasoline prices over the period 2001-2009. The results show that car switching cost (the cost incurred when switching from car to public transit) are higher than public transit switching cost (the cost incurred when switching from public transit to car), both in monetary terms ($21.41 against $1.66, respectively) and as a percentage of the monthly cost of using that mode of transportation (13.21% against 1.89%, respectively). Finally, I estimate gasoline price and public transit fare elasticities to assess the economic significance of the results.


Abstract: I investigate how sensitive is urban travel demand in Chicago to gasoline prices. I estimate the effect of gasoline prices on the number of vehicles circulating in the Chicago area and the substitution patterns between automobile circulation and public transportation. My approach exploits both cross sectional and time series variation in the data by including subarea disaggregation within Chicago area. I .find that public transportation use (CTA-Rail, CTA-Bus, Pace-Bus and Metra-Rail) increases and vehicle use decreases as gasoline prices increase. Single-trailer trucks are substituted by multi-trailer trucks.



joint with Rune M. Vejlin (Aarhus University) and
Greg Veramendi (Arizona State University)


joint with Tiago Pires (University of North Carolina)


Di Tella, Rafael, Javier Donna, and Robert MacCulloch. "Crime and Beliefs: Evidence from Latin America." Economics Letters, 99: 566–569. | |    | pdf |  |Cite|

Di Tella, Rafael, Javier Donna, and Robert MacCulloch. "Oil, Macroeconomic Volatility and Crime in the Determination of Beliefs in Venezuela." In: Venezuela Before Chávez: Anatomy of an Economic Collapse, chapter 13, edited by Ricardo Hausmann, and Francisco R. Rodríguez, Penn State University Press, (publish date: 2014), Hardcover ISBN: 978-0-271-05631-9.
pdf chapter |   | pdf book |   |Cite|