Calendrier du 30 novembre 2023
Macroeconomics Seminar
Du 30/11/2023 de 16:00 à 17:15
PSE- 48 boulevard Jourdan, 75014 Paris, salle R1-07 (côté ENS)
BALLEER Almut (RWTH Aachen University)
Biased expectations and labor market outcomes: Evidence from German survey data and implications for the East-West wage gap
We measure individual bias in labor market expectations in German survey data and
find that workers on average significantly overestimate their individual probabilities
to separate from their job when employed as well to find a job when unemployed.
These biases vary significantly between population groups. Most notably, East Germans
are significantly more pessimistic than West Germans. We find a significantly
negative relationship between the pessimistic bias in job separation expectations and
wages, and a significantly positive relationship between optimistic bias in job finding
expectations and reservation incomes. We interpret and quantify the effects of (such)
expectation biases on the labor market equilibrium in a search and matching model
of the labor market. Removing the biases could substantially increase wages and expected
lifetime income in East Germany. The difference in biases in labor market
expectations explains part of the East-West German wage gap.
Du 30/11/2023 de 16:00 à 17:00
R3.71
VELDE François ()
“Debt restructuring in Britain in the early 18th c., winners and losers”
Behavior seminar
Du 30/11/2023 de 11:00 à 12:00
R3-46
MOISAN Frédéric (EM Lyon Business School, France)
An Experiment on a Linking Game
Economic theory provides us with models of network formation. The multiplicity of equilibrium and the possibility of large payoff inequality in these models motivate an experimental study of their validity. This paper presents an experiment based on the unilateral link formation model by Bala and Goyal (2000). In this model, an individual can choose to form links with others in order to access benefits. The value flow can be either one-way (benefits only from forming a link) or two-way (benefits from forming or receiving a link) in the game. We consider two group sizes (n = 10 and n = 50) for each value flow model (4 treatments in total). We find that the subjects’ performance is relatively close to the theoretical prediction in the two-way case for all groups. That is, agents create sparse, unequal, and small average distance networks and achieve high aggregate efficiency. In the one-way flow case, consistent with the theoretical prediction, the networks are sparse and have a larger average distance and smaller degree inequality, but the deviation from the theoretical prediction is larger as compared to the two-way model, especially when the group size is large. The poorer performance in the one-way case is due to the fact that the efficient network structure —the cycle network— is more difficult to form as compared to that of the two-way model — the star network, holding individual decision accuracy constant.