The incredible shrinking good: Dynamic consumer attention. Consumers do not pay perfect attention to dynamic changes in a good’s composition. It is proposed that this can explain the phenomenon of products shrinking over time. A framework is described in which these consumers interact with a profit-maximizing monopoly firm producing a good of variable unit size over many periods. Two examples are presented. With a “threshold” probability of paying attention, consumers don’t notice sufficiently small changes in product composition. The firm exploits this to shrink unit size over time. With a constant probability of paying attention, the firm shrinks unit size in response to negative cost shocks, hoping consumers will be unaware of the change.

A key question regarding the design of financial markets is whether they have the ability to aggregate and reveal dispersed information. In the subjective expected utility (SEU) model with common priors and market scoring rules (Hanson 2003), full information aggregation is characterized by the class of separable securities, both in the strategic and in the non-strategic case (Ostrovsky (2012), Chen et al. (2012)). In this paper we examine the same question in the case where agents are non-strategic but ambiguity averse. First, we show by example that Arrow-Debreu securities, which are separable, do not always aggregate information. This means that even with complete markets, information may not get aggregated in the presence of ambiguity. Second, we characterize information aggregation in terms of a smaller class of securities, which we call strongly separable.

Representation Theorem for guilt adversion. Guilt is the experience of discomfort that follows when we violate a personal or social standard. If an individual's action raises his income but at the same time harms someone else, the higher income will come with a feeling of guilt. Any individual who is sufficiently averse to this discomfort may therefore refrain from taking the action in the first place. Economic experiments indicate that aversion to guilt is important in many strategic situations and can explain a vast array of behaviours, including cooperation, altruism, conformism, group favouritism and reciprocity. An increasingly popular way to model emotions such as guilt, is to include them as inputs in agents' utility functions. Particularly important for experimental work are utility functions with money and guilt as the two inputs. This paper's main objective is to put such functions on a firm footing by axiomatizing utility representations of guilt-averse preferences. Specifically, axioms are presented that are necessary and sufficient for (i) an existing linear representation, and (ii) a novel representation that is logarithmic in money and linear in guilt. We apply both models in three well-known games and show that representation (ii) yields sharp predictions consistent with experimental observations.

Policy makers and researchers often evaluate labor market efficiency using announced job openings and unemployment. It is well-known that announced job-openings only constitute a fraction of all job openings in the economy, but less is known about how this fraction varies across time and firms. To inform this discussion we construct a new database with actual hires and announced job openings using data from the Danish tax authorities and Public Employment Service during 2004-2012. Using this we document the share of hires made through announced job openings, and use a simple search matching model to derive the share of job openings being announced. We also find that the share of hires via announced openings being increasing in firm size, decreasing in employment growth, less used in start-ups than in existing firms and share exhibits an inverse u-shape in the firm' average wage level. Second, we set up and calibrate a simple search-and-matching model with two recruitment channels: announced and unannounced job openings. Using this model we show how the observed swings in the fraction of posted job openings significantly influences the perceived matching efficiency and Beveridge curve movements.

The Great Recession has sent debt levels to a post-WWII high for several advanced economies, reviving the discussion of fiscal consolidation. This paper assesses the macroeconomic implications of tax-based versus spending-based consolidation within the framework of a New Keynesian model with long-term government debt. Three results stand out: First, tax-based consolidations are inflationary whereas spending-based ones are deflationary. Second, the net benefits of inflation increase in the average maturity of outstanding debt: inflation revalues debt more efficiently, while distortions due to price dispersion remain unaffected - the maturity effect. Third, as a result, tax-based consolidations can become superior to spending cuts if the average maturity is high enough. Quantitatively, the threshold is two years for US data in 2013. The previous mechanism illustrates the importance of inflation in the consolidation process, even if raising its target rate is considered not to be an option.

I study a model that looks at causes, characteristics and consequences of loyalty schemes in a market with consumers that suffer from self-control problems (Gul and Pesendorfer, 2001). While the literature has mostly focused on loyalty schemes as tools used by firms to compete (Caminal and Claici, 2007) or increase consumers’ lifetime value (Caminal, 2012), I look at how a seller can use them to acquire private information about consumers’ preferences. From the consumers’ point of view, I show how loyalty schemes influence their decision to purchase and help them counter their self-control problems. I derive necessary and sufficient conditions for all consumers’ to optimally disclose their type. As for welfare, I show how as the level of information obtained by the seller from the loyalty scheme technology increases, first best may be restored.

Do Electoral Systems Affect the Incumbent Probability of Re-election? Evidence from Italian Municipalities. This paper studies the effects of different electoral systems on the incumbent probability of being re-elected for mayors, assessors and councillors in Italian municipalities. The features of the reform on the local electoral system, introduced by the Italian central government in 1993, allow to implement a Difference-in- Discontinuities Design, which combines a Regression Discontinuity Design with a Difference-in-Differences approach. Exploiting this empirical strategy the paper identifies the causal effect of a majoritarian system, as opposed to a proportional system, on the incumbent probability of being reelected. The results show that in majoritarian systems, where the mayors have more power to implement policy and the allocation of responsibility is concentrated in their hands, the probability of their re-election is higher. In addition, there is evidence that mayors use public spending to increase their probability of re-election, and that the effect of the majoritarian system is higher when the coalitions supporting the mayors are composed by a smaller number of parties. Finally, no evidence is found that assessors and councillors have different probability of being re-elected in the two electoral systems.

We study the educational choices of immigrant children in Italy, where the schooling system is characterized by stratification. We first show that immigrants tend to choose less demanding tracks relative to native students with similar ability. The gap is greater for male students and it mirrors an analogous differential in failure rates. We then estimate the impact of a tutoring and career counselling program offered to a randomly chosen sample of immigrant children displaying high academic potential. We find that the program was successful in reducing educational segregation: male treated students have a 5.0 percentage point lower probability of failing and 12.4 percentage point higher probability of entering an academic-oriented track, compared with control students. The effects are smaller and not significant for girls. To shed light on the mechanisms underlying these effects, we collected data on standardized test scores and psychological traits. We show that changes in academic motivation and teachers’ suggestion induced by the treatment explain a sizable portion of the effect on the high-school choice, while the effect of cognitive skills is negligible. Finally, we find evidence of positive spillovers of the intervention on immigrant peers of treated students, while there is no effect on natives.

The Effects of Binding & Non-Binding Job Search Requirements - Job search requirements constrain the effort choice of unemployment insurance recipients by enforcing a minimum number of monthly applications. Based on novel register data, this paper is the first to assess how individual search effort, job finding and job stability react to this constraint. Standard job search theory predicts that requirements affect each job seeker relative to her unconstrained effort choice. We therefore define the incremental effort which is necessary to comply with the constraint as our behavioral treatment intensity of interest. A proxy of this intensity is directly observed in our data: the difference between the individual requirement threshold and the unconstrained search effort before requirement imposition. Our empirical approach exploits that -conditional on a broad set of choice fixed effects- the match between the job seeker's effort choice and the caseworker's requirement setting behavior is arbitrary. It therefore provides exogenous variation in the treatment assignment. We find that binding search requirements, which exceed the job seeker's unconstrained effort choice, increase job finding in a substantial way. These effects are highly heterogeneous with respect to the job seeker's characteristics. They come at the cost of increased non-compliance and sanction imposition rates. Moreover, binding requirements have striking negative effects on job stability. Finally, we find that non-binding requirements can decrease effort and job finding rates. This suggests that requirements can operate as signals and thereby generate behavioral effects which are not predicted by standard job search theory.

In this paper we analyze the victim‐offender relationship by looking at the effect of past victimization on current offending and the effect of past offending on the probability to be victimized. We use a representative sample from the Netherlands and employ three different estimation approaches, starting with simple Ordinary Least Squares and where applicable, focus on different samples from our data, for which arguably there is a lower selection bias. Secondly, we employ a first‐difference and two‐step procedure where the residual of past victimization (offending) is used as an explanatory variable in the current offending (victimization). Thirdly, we use a strategy proposed by Altonji et al. (2005), which uses the selection on observables as a guide for the selection on unobservable factors. Using either of the methods, we find evidence that past offenders have higher probability to be victimized and some evidence for the effect of past victimization on the current offending. Stable personal characteristics, such as risk preferences and lifestyle choices, are likely to explain a big part of this relationship.