AREAS OF INTEREST:  Political Economy, Applied Microeconomics, Public Economics, Family Economics.





·         Persuasion with costly signals (with Ming Li) submitted

     [version March 2016]

     Previous versions circulated with the title: Persuasive Signalling


We present a model of persuasion, where a privately-informed sender selects from a class of signals with di_erent levels of precision to persuade a receiver to accept his request, with higher precision associated with higher costs. The sender's information could be either favourable (high) or unfavourable (low). The receiver observes both the sender's choice of signal and a random realization of the signal. We show that all plausible equilibria must involve some pooling and any informative signal must be associated with an optimistic posterior. When the receiver is ex ante pessimistic or indi_erent, the only plausible equilibrium is semi-separating, with choices of precision independent of the prior. When the receiver is ex ante optimistic, then there could be multiple plausible equilibria. Finally, we investigate the sender's optimal persuasion policy{choice of precision before (commitment) or after (discretion) he learns his private information. We show that the sender is indi_erent between commitment and discretion when the prior is optimistic, prefers discretion to commitment when the prior is su_ciently pessimistic, and could either prefer discretion or commitment when the prior is neutral.



·     Dynastic Accumulation of Wealth  (with Emmanuel Thibault) 

    [version June 2014 [version June 2008]  Additional material

    Mathematical Social Sciences  (2016), pp. 66-78

    DOI information: 10.1016/j.mathsocsci.2016.03.005  


Why do some dynasties maintain the fortune of their founders while others completely squander it in few generations? What are the causal mechanisms that underlie the intergenerational transmission of wealth? Is there a role for public policies aiming at improving long-run social mobility and wealth inequalities? To address these questions, we use a basic deterministic microfounded model based on two main ingredients: the “hunger for accumulation” and the “willingness to exert effort”. The interplay between these two elements allows our dynamic model (i) to generate a variety of wealth accumulation and effort choice dynamics (ii) to provide an explanation for some macroeconomic features of social mobility and class structure as well as for the existence and the demise of the rich bourgeoisie. Furthermore, we analyze the effect on wealth accumulation of inheritance taxation and extend our setting to variable wage opportunities and exogenous shocks. Our analysis points to the crucial role of our two key ingredients, rather than of initial wealth or transitory shocks to wealth, in generating the patterns of wealth accumulation within a family lineage.




·       Psychologically-Based Voting with Uncertainty (with Ming Li),

European Journal of Political Economy, December 2015,

Special issue on Behavioral Political Economy


We analyze a psychologically-based model of voter turnout in an election with common value and uncertainty about who the best candidate is. Potential voters’ behaviour is based on anticipated regret, where voters will experience regret if they fail to vote or vote for the wrong candidate. Furthermore, such regret is stronger when the margin of victory is smaller. We characterize mixed and pure-strategy equilibria, which can be desirable, where the right candidate is always elected in all states, or undesirable, where the wrong candidate is elected in some state. Our model yields distinctive comparative statics results. First, an increase in the proportion of informed citizens may cause the winning margin for the right candidate to either rise or fall, depending on the type of equilibrium. In addition, such an increase can have a non-monotonic effect on turnout of uninformed citizens. Second, as the prior becomes more favourable towards the ex ante favoured candidate, turnouts of informed and uninformed voters both change in a non-monotonic way. Furthermore, total turnout can be positively or negatively correlated with winning margins. The distinctive implications of our model could be tested using experimental or empirical studies. In addition, given its inclusion of uncertainty, our model can be used to study, in future research, information provision and acquisition as well as other policy questions.





·         Civic Duty and Political Advertising, DOI: 10.1007/s00199-011-0655-1

Economic Theory, March 2013, Volume 52, Issue 2, pp 531-564

     The previous version circulated with the title:

      Mobilization and Persuasion of Political Advertising [version April 2010]


Should voter awareness policies and get-out-the-vote movements be promoted? This paper addresses the question using a model of political advertising that incorporates both the mobilization and the persuasion aspects of advertising. We characterize the equilibrium and conduct comparative statics analysis allowing evaluation of the effect of voter awareness policies or the activity of get-out-the-vote movements on political advertising and the information aggregated by the electoral outcome. We find that such policies or movements may lead to either an increase or a decrease in political advertising as well as in the probability that the candidate preferred by a majority of (all informed) citizens is elected.




·         A Structural Model of Turnout and Voting in Multiple Elections (with Antonio Merlo) [Working paper version, August 2007], the  revised version is published in The Journal of the European Economic Association, Vol. 9, Issue 2, April 2011, Pages 209-245


Understanding citizens' electoral behaviour (e.g., selective abstention and split-ticket voting), represents a fundamental step in the analysis of democratic institutions. In this paper, we assess the extent to which sincere voting can explain observed patterns of participation and voting in U.S. national elections. We propose a unified model of turnout and voting in presidential and congressional elections with heterogeneous voters. We estimate the model using individual-level data for eight presidential election years (1972-2000). Our main findings can be summarized as follows. First, a non-negligible fraction of the American electorate does not vote sincerely, and only a relatively small fraction of observed split-ticket voting can be explained by sincere voting. Second, there is a systematic, positive relationship between information and turnout.


·          Do Voters Vote Ideologically? (with Antonio Merlo)

      The Journal of Economic Theory, Vol 144, Issue 5, September 2009, Pages 1868-1894


In this paper we address the following question: To what extent is the hypothesis that voters vote “ideologically” (i.e., they always vote for the candidate who is ideologically “closest” to them) testable or falsifiable? We show that using data only on how individuals vote in a single election, the hypothesis that voters vote ideologically is irrefutable, regardless of the number of candidates competing in the election. On the other hand, using data on how the same individuals vote in multiple elections, the hypothesis that voters vote ideologically is potentially falsifiable, and we provide general conditions under which the hypothesis can be tested.




·         Policy Positions, Information Acquisition, and Turnout (Working paper version 2006)

     Final version published in The Scandinavian Journal of Economics, December 2006


   The objective of this paper is to investigate the relationship between policy preferences, information acquisition, and turnout. More specifically, we characterize the choice of acquiring political information and voting as a function of citizens’ policy preferences. We conduct comparative static analysis on the information technology and we investigate the impact of a perceived polarization on information acquisition and turnout decisions. We find that: (i) middle-of-the-road citizens are the most likely both to acquire political information and to abstain;  (ii) an increase in the effectiveness of information has a higher (positive) impact on turnout and  on the fraction  of the electorate who is informed than a comparable decrease in the cost of information;  (iii) following a perceived polarization both information and abstention increase.



·   A Dynamic Model of Voting (Working paper version 2004)

Final version published in the International Economic Review, May 2007

under the title ‘Candidate Valence: Evidence from US Presidential Elections’


We propose and estimate a dynamic spatial model of voting. Using individual-level data on voting decisions in two consecutive presidential elections, we identify and estimate (1) the distribution of voters' policy positions and (2) candidates' valence. In addition to providing an equilibrium interpretation of observed individual voting profiles and aggregate electoral outcomes, we use the estimated model to conduct counterfactual experiments to assess the impact of candidates' policy positions, valence, and voters' information on the outcomes of elections and to evaluate the performance of the electoral process.



Work in Progress

·         Campaign Advertising and Career Concerns

·         Fuzzy Political Campaign (with Ming Li)

·         Persuasive counter-signalling (with Ming Li)

·         Decentralization, Media, and Candidate Selection      

·         Dynamic Allocation of Favors