Nicolas de Roos

Working papers

Learning to coordinate: A study in retail gasoline

R&R at American Economic Review, with David Byrne. Appendix

Theories of collusion study the implementation of collusive agreements, yet are silent on how such agreements are initiated. We study the initiation of tacit collusion using a unique dataset from an urban gasoline market that contains the universe of station-level prices for 15 years. We uncover a gradual process whereby dominant firms engage in price leadership to create focal points that coordinate market prices, soften competition, and enhance margins. We show how price leaders can use price experiments to tacitly communicate collusive intentions, resolve strategic uncertainty, and create mutual understanding among rivals over a collusive strategy.

Collusion with limited product comparability

R&R at RAND Journal of Economics.

We adapt the framework of Spiegler (2016) to examine the effect of limited product comparability on the viability of collusion. Firms choose messages to influence the propensity of consumers to compare products. The cartel hinders transparency on the equilibrium path, and seeks it for optimal punishment. We provide five conditions, each sufficient to ensure obfuscation aids collusion: if firms can mix over messages or commit to messages, if messages are informative, or if an individual firm or the cartel as a whole can control comparability. We also analyse the impact of message differentiation and complexity for optimal messages, and identify a key role for the convexity or concavity of comparison probabilities in these features.

Shipwrecks and treasure hunters

R&R at Journal of Economic Dynamics and Control, with Alexander Matros, Vladimir Smirnov, Andrew Wait.

We examine dynamic search as a game in which two rivals explore (an island) for a hidden prize of known value. In every period until its discovery, the players decide how much of the unsearched area to comb. If a player finds the prize alone he wins it and the game ends. Players have a per-period discount factor and costs proportional to the area they search. First, as a benchmark for efficiency, we solve the monopoly problem. Second, in the duopoly setting we show that if players are sufficiently impatient they can inefficiently over-search in the unique symmetric Markov perfect equilibrium (SMPE) – a result akin to the tragedy of the commons. On the other hand, if players are patient: the SMPE is unique except for possibly one point; and either over- or under-search can result. Finally, with patient players, several counterintuitive results can arise: for example, players might be better off searching a larger island or looking for a less valuable prize.

Collusion with intertemporal price dispersion

submitted, with Vladimir Smirnov.

We develop a theory of optimal collusive intertemporal price dispersion. Dispersion clouds consumer price awareness, encouraging firms to coordinate on dispersed prices. Our theory generates a collusive rationale for price cycles and sales. Patient firms can support optimal collusion at the monopoly price. For less patient firms, monopoly prices must be punctuated with fleeting sales. The most robust structure involves asymmetric price cycles resembling Edgeworth cycles. Low consumer attentiveness enhances the effectiveness of price dispersion by reducing the payoff to deviations involving price reductions. However, for sufficiently low attentiveness, price rises are also a concern, limiting the power of obfuscation.

Startup search costs

with David Byrne.

Search frictions determine the impacts of policies or technologies in many markets. Workhorse economic models used to study these impacts assume constant marginal search costs: individuals pay the same search cost to become informed about an additional price (sequential search models), or the entire price distribution (non-sequential search models). This paper provides evidence from a natural experiment in retail gasoline on a new form of search costs: startup costs. We empirically document how a temporary, large exogenous shock to consumers' search incentives leads to a permanent doubling of their search intensity for gasoline prices. This result is difficult to explain with a standard search model but follows directly in an otherwise standard search model with a one-time upfront cost to start searching.

Collusion, price dispersion, and fringe competition

under revision, with Vladimir Smirnov.

We study the optimal behaviour of a cartel faced with fringe competition and imperfectly attentive consumers. Intertemporal price dispersion obfuscates consumer price comparison which aids the cartel through two channels: it reduces the effectiveness of free riding by the fringe; and it relaxes the cartel's internal incentive constraints. Our theory provides a collusive rationale for sales and Edgeworth cycles, explains the survival of a price-setting cartel in a homogeneous product market, and characterises the cartel's manipulation of its fringe rival through a simple cut-off rule.

The internet, search, and asymmetric pricing: A natural experiment in retail gasoline

with David Byrne and Daniel Tiong.

How does the Internet effect retail pricing? In contrast to previous empirical research that focuses on price dispersion and static margins, this paper examines how the Internet and web-based price clearing houses effect dynamic asymmetric pricing adjustment (e.g., "rockets and feathers"). We exploit a unique policy intervention in the context of the retail gasoline market that introduced a price clearinghouse in some markets but not others. We find stark evidence that the policy eliminated asymmetric price adjustment and increase the rate of passthrough of falling costs to retail prices. These results support search-based explanations for asymmetric price adjustment.

Markup uniformity under the multi-nomial logit

We show that the multi-nomial logit model of demand implies a constant markup of price above marginal cost for multi-product oligopolists. Further, for the random coefficients discrete choice model of demand, a multi-product oligopolist optimally sets the same markup for two products if they share the same form of consumer heterogeneity, even if product characteristics differ markedly.

Collusion with a competitive fringe: An application to vitamin C

In a recent high profile case of collusion in the market for vitamin C, the cartel initially accommodated a fringe competitor before ultimately collapsing under the competitive burden. In this paper, the cartel's decision of when to dissolve is endogenised within a dynamic model of collusion. Demand estimates and cost information from the vitamin C market are used to calibrate the model. Results are intimately linked to the dynamic nature of the model. A cartel is found to persist only while fringe competitors remain small; fringe competitors invest heavily while a cartel is operating; entry deterrence is mitigated if cartel members are able to accommodate entry; and firms of an intermediate size are the most likely to accommodate entry.