Much recent research studies the relationship between people's inherent propensity for dishonesty and various economic outcomes. To measure individual dishonesty, researchers rely on experimental games in which individuals can gain financially by lying about the outcome of one or more die rolls, coin tosses or other random outcomes. The way behavior in these games is translated into quantitative measures of dishonest is ad hoc and varies from paper to paper. In this paper, I present a simple econometric framework formalizing probabilistic lying games and how behavor in these games relate to individual differences in dishonesty. I use this to discuss different formal definitions of dishonesty and to compare different existing dishonesty measures from the literature. I show how data from probabilistic lying games can be used to construct unbiased to different formal measures of individual dishonesty. These measures are subject to classical measurement error, however, leading to attenuation bias problems when employed in regression analyses. When each individual participates in more than one round of lying games however, a simple split sample instrument is available that addresses measurement error problems. I also discuss possible extensions to the case where researchers observe the actual outcome of the die rolls, coin tosses or other outcomes in the game. Finally, I provide a set of practical recommendations for applied researchers.
Are Workers Better Matched in Large Labor Markets?
Working Paper. First version: Nov. 2012, Latest version: Feb. 2015
Link to: Working Paper
This paper examines the relationship between labor market size and job search outcomes. Much research and many policy initiatives assume that larger labor markets lead to better job search outcomes because they give workers and firms more choice
in potential jobs or employees. The empirical finding that labor market size and job finding rates are uncorrelated, however, has led researchers to question this assumption. I show, theoretically and empirically, that large labor markets may
cause workers to find jobs that are better matches given their individual skills and characteristics, even if they do not cause workers to find jobs faster. I construct a unique new data set from Denmark that combines administrative data, an
online vacancy database and detailed geographical information. I show that workers in large labor markets find jobs for which they are a better match as measured by both previous industry experience and geographical location. They also find
jobs which pay higher wages and result in longer employment spells even after controlling for spatial productivity differences among firms. The estimated effects imply that labor market size explains 6.6% of the spatial variation in wage premia,
and also suggest a high rate of return on transport infrastructure projects that increase the effective size of labor markets by increasing workers' ability to commute to distant jobs.
A Formal Model of Corruption, Dishonesty and Selection into Public Service
(with S. Barfort, A. Leth Olsen and F. Hjorth)
Working Paper. First version: Sept. 2015
Link to: Working Paper
Recent empirical studies have found that in high corruption countries, inherently more dishonest individuals are more likely to want to enter into public service, while the reverse is true in low corruption countries. In this note, we provide
a simple formal model that rationalizes this empirical pattern as the result of countries being stuck in different self-sustaining equilibria where high levels of corruption and negative selection into public service are mutually reinforcing.
Fundamentals and Optimal Institutions:
The case of US sports leagues
(with Martin Gonzalez-Eyras and Martin Rossi)
Working Paper. First version: Jan. 2017, Latest version: Jan 2020
Link to: Working Paper
To shed light on the relation between fundamentals and adopted institutions we examine institutional choice across the "Big Four" US sports leagues. Despite having very similar business models and facing the same economic and legal environment, these leagues exhibit large differences in their use of regulatory institutions such as revenue sharing or salary caps. We show, theoretically and empirically, that these differences can in part be rationalized as optimal responses to differences in one fundamental characteristic of the sports being played: how strongly win probabilities respond to hired talent. Our results thus show evidence that in professional sports existing institutions are tailored to fundamentals.
How do job seekers apply for jobs and can we help them do better?
(with J. Maibom, K. Hvidberg and A. Augsburg)
What is Keeping Unemployed Workers From Finding Employment?
(with A. Sebald, S. Altmann, D. Skandalis, R. Mahlstedt, M. Pietraszek and A. Kristensen)
Helping the Unemployed through Statistical Prediction?
(with R. Mahlstedt and M. Rasmussen)
Different Types of Peers
(with A. Bjerre-Nielsen and D. Lassen)