We extend the principal/one-agent model with countervailing incentives to a framework in which the principal deals with two agents behaving non-cooperatively and protected by limited liability. Focusing on the two-type case, we show that, beside the situation in which first best is effected even without relying on type correlation, dominant-strategy implementation yields no penalty to the principal, with respect to Bayesian-Nash implementation, when the principal faces, on the opposite, very tight constraints.

We model an agency relationship in which the agent's cost is non-monotonic with respect to type and the type is correlated with a public ex-post signal. The principal can use lotteries to exploit the type-signal correlation within the limit of the agent's liability. We establish conditions for first-best implementation, highlighting two effects on contractual design. First, the structure of the optimal lottery varies across types and, for each type, it depends on whether the cost is U shaped or reverse U shaped with respect to type. Second, as compared to the case of monotonic cost, the design of incentive compatible lotteries is easier when the cost is U shaped, more difficult when the cost is reverse U shaped. The root of the second effect is that incentives are non-monotonic either below or above some interior types. The two effects involve that non-monotonicity is unfavorable to the principal when the cost is reverse U shaped. This conclusion is at odds with the wisdom, concerning settings without correlated information, that non-monotonicity, which triggers countervailing incentives, enhances contracting.

A government delegates a build-operate-transfer project to a private firm. At the contracting stage, the operating cost is unknown. The firm can increase the likelihood of facing a low cost (the good state) by exerting effort when building the infrastructure. Once this is in place, the firm learns the true cost and begins to operate. Under limited commitment, either the firm or the government may renege on the contract. Within this context, we explore how well a contract with a state-dependent duration performs, as compared to the more standard fixed-term contract. Under full commitment, the efficient allocation is decentralized, whether the contractual term is fixed or state-dependent. Under limited commitment, in situations where break-up of the partnership is little costly for the government, the efficient allocation can be decentralized only if it is stipulated that the duration of the contract will be longer in the good state than in the bad state. This result is at odds with the prescription of the literature on "flexible- term" contracts, which recommends a longer contractual length when the operating conditions are unfavourable.

We consider electricity generation industries where thermal operators imperfectly compete with hydro operators that manage a (scarce) water stock stored in reservoirs over a natural cycle. We explore how the exercise of intertemporal market power affects social welfare and environmental quality. We show that, as compared to the outcome of spot markets, long-term contracting either exacerbates or alleviates price distortions, depending upon the consumption pattern over the water cycle. Moreover, it induces a second-order environmental effect that, in the presence of a thermal competitive fringe, is critically related to the thermal market shares in the different periods of the cycle. We conclude by providing policy insights.

A multidimensional-and-sequential screening problem arises in a framework where the agent is privately informed about expected cost and cost variability and, subsequently, learns the realized cost as well. As the principal's marginal surplus function becomes less concave/more convex, the optimal mechanism reflects progressively stronger incentives to mimic less inefficient types, and to misrepresent the cost variability relative to the expected cost. When the principal's knowledge imperfection about the cost variability is sufficiently less important than that about the expected cost, quantities are pooled with respect to the former for a high-expected-cost agent. A low-expected-cost agent is not assigned the first-best output at least in some state of nature.

We provide theoretical foundations for quality-adjusted price-cap regulation in industries where a regulated incumbent and an unregulated entrant offer vertically differentiated products competing in price and quality. We show that, whether or not the incumbent anticipates the reaction of the entrant, the optimal weights in the cap depend upon the market served by the entrant, despite the latter not being directly concerned by regulation. We further show that the cap is robust to small errors in the weights. Our findings point to the conclusion that, in partially regulated industries, regulators should use information about the whole sectors rather than on the sole regulated incumbents.

A government delegates a build-operate-transfer project to a private firm in a limited-commitment framework. When the contract is signed, parties are uncertain about the operating cost. The firm can increase the likelihood of facing a low cost by exerting some non-contractible effort while building the facility. Once the facility is in place, the firm learns the marginal cost and begins to operate. We characterize the contract which stipulates the efficient allocation. We study the financial structure and duration that secure its enforcement. To this end, we take into account that break-up of the partnership occasions a replacement cost for the government and an expropriation cost for the firm and its lender. Furthermore, both these costs are higher the earlier the contract is terminated. Enforcement is achieved as follows. The firm is instructed to invest some intermediate amount of own and borrowed funds. Under the aegis of a third party that can commit, the government provides guarantees to the lender, conditional on continuation of the partnership. Duration may be shortened, though not to the point where the initial effort of the firm is uncompensated.

In sequential screening problems it is found that, under some regularity conditions, local incentive compatibility constraints are sufficient for implementability. However, this follows from the assumption that the possible distributions of the unknown variable satisfy either first-order stochastic dominance or mean-preserving spread. That assumption is matched with private information about either the expected value or the spread of the variable. In this paper we allow for private information about both parameters. In a setting with four possible cost distributions, two with equal expected values and different spreads and two with different expected values and equal spreads, we show that there can be multiple combinations of binding incentive constraints depending on the principal's preferences. The less concave / more convex that the marginal surplus is, the more that the binding incentive constraints are related to private information about one parameter of the distribution relative to the other. Yet, screening is always two-dimensional. Local incentive constraints are sufficient, as in the literature, only when the marginal surplus is sufficiently convex. We further suggest that, in the same vein as in Consumption theory, the contractual choice can be regarded as mirroring the preference of the decision-maker for a lottery that occasions a higher (certain) cost but grants the possibility of facing more efficient (random) outcomes. Resting on this interpretation, we assess that the benefit of screening the agent in two stages, rather than in the contracting stage only, is higher when the marginal surplus is less concave / more convex.

We model a ferry market where passengers are heterogeneous in their valuation of waiting time and, unlike in previous studies, can take services from all operators. Analyzing their behaviour when two operators are active, each providing one service, we find that complex patterns of product differentiation emerge between two goods that (i) do not exactly correspond to the available services and (ii) display service frequencies as quality attributes. A (low-quality) basic good, coinciding with the cheaper service, attracts low-time-value passengers. A (high-quality) composite good, which is a bundle of the two available services, appeals to high-time-value passengers. Consequently, demand is positive for either operator so that an inefficient operator is not crowded out. In the specific case of a mixed duopoly, a price-aggressive public operator spans discipline over (but does not monopolize) the whole market; a soft one boosts "quality" (i.e., frequency) vis-à-vis a fraction of the population only, that is yet larger than under classical vertical differentiation. Policy-makers pursuing redistribution objectives should target the cheaper service, in general, privileging either a raise in its frequency (when it is low) or a cut in its price (when frequency is high), depending upon the group of passengers they wish to support.