Work in progress

"Modeling Crime from a Behavioral Economics Perspective", submitted - Job market paper

The objective of this paper is to unveil the consequences of aggravating and mitigating circumstances from a behavioral economics perspective at two different times: when legislators decide law provisions and public budgets, and when judges apply these laws during trials. We develop a static optimization model of social welfare changes after criminal actions suitable to account for a group of heterogeneous offenders. The model includes monetary and non-monetary characteristics of the offenders, specific monetary sanctions, specific non-monetary sanctions, a general enforcement effort, and it satisfies the proportionality principle. Our model can formally explain many of the intuitive and counterintuitive empirical findings of previous literature from a social welfare optimization perspective.

A PDF copy of this work can be found here:

"Incentives for Green Technology Adoption and Compliance under Risk Aversion and Technological Uncertainty" (with Carmen Arguedas and José Luis Zofío), Revise & Resubmit, Environmental and Resource Economics

In this paper, we study firms' incentives to adopt environmentally friendly technologies as a response to emission taxes, in contexts of imperfect compliance and risk aversion. Previous work has analyzed technology adoption incentives and compliance issues under risk neutrality. However, the decision of whether to exceed a regulation entails risks, since polluting agents are exposed to penalties with some probability. Also, there may be uncertainty regarding the impact of green technology adoption on firms' abatement costs. Hence, preferences for risk may matter and should be accounted for. In a baseline model with certainty about future abatement costs, we find that adoption decisions are independent of risk preferences, even if the enforcement policy is so weak that induces imperfect compliance. However, in a context with uncertainty about future abatement costs, adoption incentives decrease with both risk aversion and weak enforcement, especially when the latter entails low sanctions rather than low inspection probabilities.

"Defining the Proportionality Principle: Evidence from the Public Perception in Spain" (with Sandra Rousseau), in progress

In this paper, we empirically study preferences for implementation of the proportionality principle used in law enforcement considering the case of Spain. The objective is to reveal which definition of the proportionality principle reflects public opinion best and is most appropriate for developing an optimal sanctioning strategy. The definition of the proportionality principle is the only assumption in "Modeling Crime from a Behavioral Economics Perspective" without the support of previous empirical evidence. Any further study of the results derived from that model requires gaining insight into public perception regarding the definition of the proportionality principle. Therefore, we use a survey to explore people's support for three different interpretations of the proportionality principle in Spain. Moreover, the proportionality principle plays a key role in European sanctioning policies. Thus improving insights into how the general population interprets this principle may play an important role in determining public policy support. 

"Retroactivity in Criminal Law" (with Gabriel Doménech-Pascual), in progress

This paper examines the economic rationale of the two key rules governing retroactivity in criminal law: the prohibition of ex post facto laws, and the lex mitior principle. The forer precludes the retroactive imposition or aggravation of criminal liability, while the latter mandates the retroactive application of laws decriminalizing conduct or reducing the punishment for it. The study presents a formal model analyzing how these principles impact social welfare under conditions of risk neutrality and risk aversion, as well as when the implementation of sanctions and the adjudication of cases entail significant social costs. Our model shows that, if individuals are risk averse, retroactivity is not optimal. If implementing penalties entails costs, the results depend on the function of these costs. If it is strictly convex (e.g., sanctions show diseconomies of scale, which is arguably the case with the most severe penalties) or strictly concave, non-retroactivity is socially optimal. If it is linear (as in the standard Law and Economics model of law enforcement), both retroactivity and non-retroactivity can maximize social welfare. If adjudicating cases entails social costs, the retroactivity (non-retroactivity) of either ex post facto laws or lex mitior might increase, lower or have no effect on such costs, while preserving the same level of deterrence.  In all but one scenario, retroactivity is suboptimal or non-necessary to maximize social welfare. Overall, the findings of this paper provide strong support for the prohibition of ex post facto laws, and weak or no support for the application of lex mitior. 

"Emotions, Algorithms, and Electoral Error: A Formal Model of Digital Disinformation in Political Communication Environments" (with Roberto Morales-Arsenal), in progress

Why do voters sometimes support candidates whose policies contradict their interests? We develop a formal model in which random misinformation (noise) and strategic disinformation (bias) interact with asymmetric affective polarization and loss-averse valuation, as described in Prospect Theory, to distort electoral choice. The model yields three main results. First, with unbiased rational voters, noise lowers accuracy but, by symmetry, cannot raise the probability of selecting the inferior candidate above 0.50. Second, a positive informational bias shifts perceived signals and, even with equal loss aversion for both candidates, raises that probability to roughly 0.64 under baseline parameters. Third, when polarization is asymmetric—modeled as stronger loss aversion toward the out-group or unequal reference points—and combined with higher noise and positive bias, the probability rises further and can exceed 0.70 in simulations. These findings reveal a hierarchy of distortion: noise creates random error, bias adds directional error, and asymmetric polarization amplifies both, leading to systemic departures from meritocratic selection. Disinformation is framed as a structural market failure that erodes trust and representation, informed by insights from political communication, behavioral economics, and neuroscience.