(co-authored with Francisco Ruiz-Aliseda) Revise and resubmit to Journal of Economics and Management Strategy
This paper studies entry and exit decisions in perfectly competitive markets whose demand alternates between growth and decline phases at uncertain times. We introduce a stochastic process that captures these features of random market evolution, and we provide key mathematical results related to first passage times which make the characterization of entry and exit behavior quite simple and straightforward (even when the process is subject to an endogenously determined upper or lower barrier). We characterize entry and exit patterns in a dynamic competitive equilibrium, and we show why our results differ from those obtained if demand follows a diffusion process (e.g., a Geometric Brownian Motion). Despite the stochastic process of the underlying variable has a continuous sample path in both cases, we demonstrate in our setting that positive rates of entry and exit discontinuously fall to zero owing to informational overshooting. Another advantage of our framework is that it can explain discontinuities in stock prices even if sample paths are continuous or that it can be easily applied to (dis)investment timing games, as we illustrate.
Revise and resubmit to Journal of Economics and Management Strategy
This paper introduces a continuous-time game to study two ex-ante identical firms' incentives in capacity preemption. Each firm can choose small or large capacity and investment timing to enter a new industry whose demand grows until an unknown maturity date. Previous literature usually predicts that the Stackelberg leader, whether endogenously or exogenously determined, is better off by building a larger capacity than its rival. In contrast, this paper proves that in most cases the first mover's equilibrium strategy is to enter with a smaller capacity than the follower. If it had chosen the larger capacity, its follower could, and in fact would use a smaller plant to force it out of the market through a leadership contest. The large leader lacks incentive to fight for the market because the small firm can make credible commitment to stay with its higher option value of waiting to exit.
(co-authored with David Besanko) Revise and resubmit to Journal of Industrial Economics
This paper explores the trade-off between R&D cooperation and competition with learning. We develop a continuous time, two-armed bandit model in which firms can devote resources to a "safe" investment in an established market or to a risky R&D investment aimed at discovering a new product that is characterized by both "if" and "when" uncertainty. The firm that wins the race under R&D competition enjoys a period of monopoly profits. But, after this period, competition in the market for the new product occurs, which may also impair the profitability of each firm's established product. Post-patent market structure and the extent of the adverse impact of the new product on the profitability of established products play a key role in driving the tradeoff between R&D competition and R&D cooperation. Firms' incentives to invest in the non-cooperative regime differ from a consortium's investment incentives. Under non-cooperative R&D, a free rider problem can arise, which generally results in an equilibrium investment flow that is less than or equal to that of the research consortium. Expected ex ante undiscounted investment and welfare under R&D competition can also be shown to be less than what it would be under R&D cooperation. However, if the gain due to the oligopoly profit from the new product is less than the loss in profit from the established product, then the free rider problem does not arise. In this case, R&D cooperation results in the same or lower level of investment than arises under non-cooperative R&D. Thus, in contrast to the traditional literature, we show that technology spillover alone need not lead to higher R&D investment or higher social welfare under research cooperation. In fact, significant product market spillovers are necessary if the underlying R&D project exhibits both "if" and "when" uncertainty.
(co-authored with David Besanko) This paper is currently under review.
This paper studies the impact of R&D subsidies in a setting in which there is uncertainty not only about the timing of a breakthrough ("when" uncertainty), but also about whether a breakthrough is even possible ("if" uncertainty). This setting seems particularly applicable to firms engaged in fundamental scientific research. Our paper makes two broad contributions. First, we show that the way in which R&D is subsidized matters. Certain types of subsidies may crowd out private investment while other types of subsidies may stimulate private investment. Second, we show that simple subsidy mechanisms can be very effective in dealing with market failures, and under important conditions, it may be possible to implement the first-best outcome with a minimum of information. This demonstrates that even in complex dynamic settings R&D subsidies have the potential to improve social welfare. More specifically, the paper utilizes a two-armed bandit framework to model the impact of government subsidies for private R&D investment. We focus on two cases: monopoly and competitive R&D. We derive an individual firm's optimal R&D investment decision and noncooperative firms' symmetric Markov Perfect equilibrium investment strategies. If there is no shadow cost of public funding, we show that for a monopoly the first-best welfare can be attained through a pure matching subsidy that does not rely on firm beliefs about project viability. Under competition, the first-best can be attained using a combination of a (belief-free) matching subsidy and a (belief-free) unrestricted subsidy policy. The unrestricted component is needed because of the free-rider problem that arises under R&D competition. By contrast, if there is shadow cost of public funding, we show that earmark and unrestricted subsidies are never optimal for the monopoly case for a large set of parameters. In fact, numerical examples demonstrate that a pure matching policy is optimal for all cases under monopoly and when spillovers are sufficiently small under R&D competition.
(co-authored with Andras Niedermayer) Revise and resubmit to International Journal of Industrial Organization.
Inter-firm R&D collaborations through contractual arrangements have become increasingly popular, but in many cases they were broken up at the pre-committed contract expiration date without any joint discovery. We provide a rationale for pre-committing to a breakup time in the R&D agreement. Consider a research consortium initiated by a firm A with a firm B. B has private information about its benefit from a success, the difficulty of the development process, and the relevance of its expertise. In first-best, breakup never occurs: a consortium is either continued until success or never formed in the first place. In second best, a breakup rule is necessary to elicit the right type of firm B to collaborate under a fairly general condition. This condition is sufficient not only when A is restricted to offering a single, type independent contract to B, but also when A can offer a menu of contracts depending on B's type. If the costs of separating different types of B are too high, A may prefer partial pooling. Furthermore, numerical analysis shows that A only breaks up with B if B has the right type and never breaks up in case of pooling with a bad type. This stems from breakup being a screening device.
This paper introduces a two-stage model to study a duopoly competition to retain customers. Conventional wisdom tells us that competition in markets with switching cost leads to too much switching. We argue that when consumers are allowed to seek retention benefit from their existing suppliers, firms cannot refrain from offering discount in equilibrium, although their profits are lower. Even if retention seeking is costly, consumer and social welfare are unambiguously improved. As the cost of retention seeking remains positive but goes to zero, the difference between the discount and full price goes to zero, but the welfare improvement remains positive throughout.
This paper studies a simple market with N potential buyers and N potential sellers where each trader needs to incur a small participation cost to enter the trading for an indivisible good. In this case, the realized market might be asymmetric in two senses: first the numbers of traders in each side are not equal; second, the entering buyers and sellers will have different supports, either partially overlapped or completely separated. These two aspects might cause the equilibrium strategies to become kinked. However, it can be shown that the post entry efficiency loss in this simple market will be at most O((1/(N²))) even with the presence of those complexities.