Publications
Massoni, S. & Teixeira, V. (2026) Court Resolution and Predictive Justice,
European Journal of Law and Economics [article]
Working Papers (available upon request)
Partial Control, Partial Blame? Programmer Control and Blame Shifting in Algorithmic Delegation, with M. Chevrier
Under revision in the Journal of Economic Psychology (minor revision)
Fairness preferences in gain and losses: evidence from an experiment, with E-W. Tchalanga
Under review in the Journal of Economic Science AssociationÂ
Ambiguity Attitudes and the Value of Partial Information, with S. Massoni
Submitted
Choosing Between Algorithmic Forecasters: What drives Delegation?, with M. Chevrier and F. Fidanoski
Measuring Perceived Inequality: an Empirical Comparison of Methods, with Y. Kaouane, E. Kemel and E.W. Tchalanga
Replication of Banovetz and Oprea (2023), American Economic Journal: Microeconomics, conducted within the I4Replication Lab, with A. Boufarsi, P. Crosetto, A-G. Maltese, and D. Mayaux.
Work in progress
General population survey in Morocco on the causes and consequences of perceived inequality, with E. Kemel
Is the Gender Pay Gap Fair? Perceptions and Policy Support in Morocco, with S. Abouri and E. Kemel
Eliciting Risk Preferences: Overcoming Probability Distortions, with M. Abdellaoui, S. Massoni and L. Page
Abstract of the papers
Partial Control, Partial Blame? Programmer Control and Blame Shifting in Algorithmic Delegation, with M. Chevrier
When a decision is delegated to an algorithm, two humans stand behind the outcome, namely the user who relies on it and the programmer who designed it. We study how the programmer's degree of control over the algorithm shapes blame shifting. In a laboratory experiment, a user can choose an allocation or delegate it to a human intermediary, to an algorithm that always implements its programmer's choice (Algo-100), or to an algorithm that implements this choice with a 50% probability (Algo-50). Recipients can then punish the user and/or the intermediary. We find no evidence that delegation rates differ across intermediaries or that algorithmic framing alone affects punishment. When an inegalitarian allocation is implemented after delegation, recipients are less likely to punish in Algo-50 than in Algo-100, and those who punish target the programmer less and the user more. Averaged over all recipients, the programmer's punishment falls while the user's does not change significantly. The relevant dimension is therefore not the label attached to the intermediary but the degree of control of the human actors behind it, which emerges as a boundary condition of blame shifting in algorithmic decision chains.
Ambiguity Attitudes and the Value of Partial Information, with S. Massoni
Decision-makers often receive predictive information that refines uncertainty without eliminating it. This paper examines why such partial refinements may receive limited value under ambiguity. We decompose ambiguity attitudes into ambiguity aversion and ambiguity-generated insensitivity, and relate these components to willingness to pay for predictive information in two incentivized experiments involving artificial Ellsberg-type uncertainty and structured natural uncertainty from a predictive-justice environment. Ambiguity aversion is strongly source-dependent, with higher levels under artificial uncertainty than under judicial uncertainty. Ambiguity-generated insensitivity remains substantial under complete ambiguity in both sources and declines when likelihood information is introduced. In the information-valuation task, willingness to pay responds more clearly to full resolution of likelihood ambiguity in the artificial source, whereas valuation is nearly flat across information types in the structured natural source. Ambiguity-generated insensitivity predicts weaker demand for predictive information, mainly at the extensive margin, while ambiguity aversion is more closely related to valuation conditional on positive demand.
Fairness preferences in gain and losses: evidence from an experiment, with E-W. Tchalanga
Do fairness judgments depend on whether people allocate gains or losses? We investigate this question using a preregistered online experiment (N=602) in which impartial spectators redistribute a gain or a loss between two workers after observing (im)perfect information about the source of the inequality (performance or luck). Our data do not allow us to reject the hypothesis that fairness views are invariant to the domain of the outcome. A notable exception arises among meritocrats when it is certain that the initial inequality is based on performance: they redistribute less in the Loss treatment than in the Gain treatment, suggesting a stronger tendency to preserve performance-based allocations in losses than in gains. Finally, we find that spectators allocate more to the high-earning worker as the probability that the initial inequality reflects performance increases, and that meritocrats become increasingly egalitarian as uncertainty about the source of inequality rises.
Measuring Perceived Inequality: an Empirical Comparison of Methods, with Y. Kaouane, E. Kemel and E-W. Tchalanga
Beyond the reality of income inequality, its perception also matters. Several methods have been developed for the measurement of perceived inequality, and mixed results are observed regarding (i) the comparison between perceived and real inequality, and (ii) its power to predict redistribution preferences. This study compares five quantitative methods for measuring perceived income distributions, including a novel method adapted from the measurement of beliefs in behavioral economics. We assess their consistency, accuracy, and ability to predict preferences for redistribution . The methods are implemented in an incentivized and choice-based within-subjects experiment. Notably, our subjects are from Morocco, and they considered inequality and redistribution in France, thereby taking a spectator perspective. Following a pre-registered plan, we use econometric methods to estimate perceived Gini indices. This allows us to test the calibration and consistency of perceived inequality between methods as well as its power to predict redistributive choices. The different methods give consistent results regarding the perception of average income, and more heterogeneous results regarding the perceived Gini. The method based on the elicitation of histograms provides the best results according to our comparisons.