Portail HAL EHESS (École des hautes études en sciences sociales)
Not a member yet
103682 research outputs found
Sort by
Le rire à la frontière italo-française : un outil de résistance morale et politique.
No full textInternational audienc
Government Reputation in Ramsey Taxation
No full textNational audienceWe embed honesty-based reputation into a Ramsey taxation framework with com-petitive firms and households. In a static benchmark with exogenous trust, there is a sharp cutoff below which the optimal policy sets no taxes and above which the optimal tax take rises with trust. In the dynamic model, beliefs evolve through noisy public monitoring of delivered public goods; the planner’s problem is well posed, the value is increasing and convex in beliefs, and optimal revenue is monotone in reputation with a trust threshold that is weakly below the static cutoff. With multiple broad instruments and symmetric monitoring, the dynamic force acts through the total revenue scale; the tax mix is indeterminate along an equivalence frontier. Blackwell-improving monitor-ing and greater type persistence expand the optimal scale and shift the trust threshold inward. The model delivers clear policy prescriptions for building fiscal capacity in low-trust environments and testable links between measured trust, verifiability, and revenue
Lifecycle Wages and Human Capital Investments: Selection and Missing Data
No full textInternational audienceWe derive wage equations with individual specific coefficients from a structural model of human capital investment over the life cycle. This model allows for interruptions in labour market participation and deals with missing data and attrition problems. We propose a new framework that deals with missingness at random and is based on factor decompositions that allow for flexible control of selection. Our approach leads to an interactive effect wage specification, which we estimate using long administrative panel data on male wages in the private sector in France. A structural function approach shows that interruptions negatively affect average wages. Interestingly, they also negatively affect the inter-decile range of wages after twenty years. This is only partly due to the fact that interruptions are endogenous
L'oral dans tous ses états à travers les langues du monde : enjeux linguistiques, littéraires et socioculturels
No full textInternational audienceCe colloque s’adresse aux linguistes, littéraires, anthropologues, historiens du langage et spécialistes des médias, dans les langues du monde, souhaitant explorer ces problématiques dans toute leur diversité. Il invite à confronter différentes traditions linguistiques et littéraires, afin de penser les multiples façons dont l’oral et l’écrit interagissent, se transforment et se recomposent.Nous encourageons des propositions adoptant une approche théorique, descriptive ou appliquée, ainsi que des études de cas illustrant la pluralité des relations entre oralité et scripturalité dans différentes langues et cultures
Bilevel gradient methods and the Morse parametric qualification condition
No full textWe introduce the Morse parametric qualification condition for bilevel programming. Generic semi-algebraic functions are Morse parametric in a piecewise sense. Thus, bilevel programs with a Morse parametric lower level constitute a relevant intermediate class between strongly convex and fully generic lower levels. In this framework, we study bilevel gradient algorithms with two strategies: the single-step multi-step strategy, which involves a sequence of steps on the lower-level problems followed by one step on the upper-level problem, and a differentiable programming strategy that optimizes a smooth approximation of the bilevel problem. While the first is shown to be a biased gradient method on the problem with rich properties, the second, inspired by meta-learning applications, is less stable but offers simplicity and ease of implementation
Mixtures Closest to a Given Measure: A Semidefinite Programming Approach
No full textInternational audienceMixture models, such as Gaussian mixture models (GMMs), are widely used in machine learning to represent complex data distributions. A key challenge, especially in high-dimensional settings, is to determine the mixture order and estimate the mixture parameters. We study the problem of approximating a target measure, available only through finitely many of its moments, by a mixture of distributions from a parametric family (e.g., Gaussian, exponential, Poisson), with approximation quality measured by the 2-Wasserstein or the total variation distance. Unlike many existing approaches, the parameter set is not assumed to be finite; it is modeled as a compact basic semi-algebraic set. We introduce a hierarchy of semidefinite relaxations with asymptotic convergence to the desired optimal value. In addition, when a certain rank condition is satisfied, the convergence is even finite and recovery of an optimal mixing measure is obtained. We also present an application to clustering, where our framework serves either as a stand-alone method or as a preprocessing step that yields both the number of clusters and strong initial parameter estimates, thereby accelerating convergence of standard (local) clustering algorithms
Routes to the Top
No full textInternational audienceWho makes it to the top? We use the leading socio-economic survey in Germany, supplemented by extensive data on the rich, to answer this question. We identify the key predictors for belonging to the top 1 percent of income, wealth, and both distributions jointly. Although we consider many, only a few traits matter: Entrepreneurship and self-employment in conjunction with a sizable inheritance of company assets is the most important covariate combination across all rich groups. Our data suggest that all top 1 percent groups, but especially the joint top 1 percent, are predominantly populated by intergenerational entrepreneurs
Dans l'intimité de Sapiens. Vivre il y a 40 000 ans
No full textInternational audienceQui n'a pas rêvé de découvrir au plus près le quotidien de ses ancêtres ? Entre enquête scientifique et récit immersif, Nicolas Teyssandier nous entraîne dans l'intimité de ces groupes de chasseurs-cueilleurs nomades, qui -il y a quarante mille ans -sillonnaient les steppes durant la dernière période glaciaire dans le sillage des troupeaux de rennes, de chevaux et autres bisons. Il nous fait partager leur quotidien, leurs expéditions de chasse, en faisant renaître sous nos yeux leurs techniques de fabrication d'armes de chasse et d'outils les plus abouties. Loin de l'image caricaturale de peuples isolés luttant pour leur survie, en proie à un environnement hostile, errant sans cesse à la recherche de nourriture, les Sapiens du Paléolithique récent formaient des sociétés complexes, disposant de hauts niveaux de connaissances et de savoir-faire. Nicolas Teyssandier est préhistorien, directeur de recherche au CNRS et directeur adjoint du laboratoire TRACES, basé à l'université Toulouse-Jean-Jaurès. Il a coscénarisé et a été le conseiller scientifique du film documentaire "Le fils de Néandertal ou le secret de nos origines" (Arte, 2017). Il est notamment l'auteur de Nos premières fois (La ville brûle/Champs Flammarion)
Bayesian Learning in Mean Field Games
No full textInternational audienceWe consider a mean-field game model where the cost functions depend on a fixed parameter, called \textit{state}, which is unknown to players. Players learn about the state from a a stream of private signals they receive throughout the game. We derive a mean field system satisfied by the equilibrium payoff of the game and prove existence of a solution under standard regularity assumptions. Additionally, we establish the uniqueness of the solution when the cost function satisfies the monotonicity assumption of Lasry and Lions at each state