In this Article, we suggest that litigation can be analyzed as though it is a competitive research and development project. Developing this analogy, we present a two-stage real option model of the litigation process that involves sequential information revelation and bargaining over the surplus generated by early settlement. Litigants are risk-neutral and have no private information. The model generates results that, we believe, have analytic and normative significance for the economic analysis of litigation.

From an analytic perspective, we demonstrate that negative expected value (NEV) lawsuits are analogous to out of the money call options held by plaintiffs and that every NEV lawsuit is credible if the variance of the information revealed during the course of the litigation is sufficiently large. This finding helps explain the prevalence of a class of lawsuits that has proved puzzling to traditional, expected value-based modes of litigation analysis. The model also suggests that risk-neutral defendants can act as though they are risk-averse and that risk-neutral plaintiffs can act as though they are risk-seeking because increases in variance can increase a lawsuit's settlement option value just as it increases a call option's value without regard to the holder's degree of risk aversion. Models that presume defendants' relative risk aversion may therefore rely on an unnecessary assumption. Our model also suggests that a lawsuit's option settlement value is not a monotonically increasing function of the variance of the information revealed during the litigation. In particular, at low levels of variance a lawsuit's option settlement value may equal its traditional expected value, but as variance increases its option settlement value can display a discontinuity after which its option settlement value becomes a monotonically increasing function of variance. NEV lawsuits can also display "dead zones" - regions of variance over which the claim is not credible even though it is credible over higher or lower levels of variance. Comparative statics analysis also quantifies the extent to which a lawsuit's settlement value increases as plaintiff's litigation expenses occur later in the litigation process, as the ratio of defendant-to-plaintiff litigation expense increases and as plaintiff bargaining power increases.

From a normative perspective, we offer an "impossibility conjecture" suggesting that the mere presence of an irreducible degree of uncertainty endemic to the litigation process can be sufficient to prevent private litigation incentives from equating to socially optimal incentives, even if one adopts all other assumptions necessary to equate private and social incentives. It follows that it may be impossible to articulate normative principles of law through substantive standards that ignore the uncertainty inherent in the litigation process and the procedural environment in which the litigation occurs.