We suggest moment estimators for the parameters of a continuous time GARCH(1,1) process based on equally spaced observations. Using the fact that the increments of the COGARCH(1,1) process are ergodic, the resulting estimators are consistent. We investigate the quality of our estimators in a simulation study based on the compound Poisson driven COGARCH model. The estimated volatility with corre...

This paper considers the Ricardian Equivalence proposition when expectations are not rational and are instead formed using adaptive learning rules. We show that Ricardian Equivalence continues to hold provided suitable additional conditions on learning dynamics are satisfied. However, new cases of failure can also emerge under learning. In particular, for Ricardian Equivalence to obtain, agents...

In an asset-pricing model, risk-averse agents need to forecast the conditional variance of a stock’s return. A near-rational restricted perceptions equilibrium exists in which agents believe prices follow a random walk with a conditional variance that is self-fulfilling. When agents estimate risk in realtime, recurrent bubbles and crashes can arise. These effects are stronger when agents allow ...

This paper investigates competition between health insurance companies under different financing regulations. We consider two alternatives advanced in recent German health care reform discussions: competition by contribution rates (health contributions) and by fees (health premia). We find that contribution rate competition yields lower company profits and higher consumer welfare than premia co...

We prove that there exist social choice correspondences that can be implemented in undominated strategies, using finite mechanisms, but can require arbitrarily many strategies to do so, even when there are only two agents and two types of each agent. This is in sharp contrast with dominant-strategy implementation, where the revelation principle implies that only one strategy for each type of ag...

We consider two Bayesian agents who learn from exogenously provided private signals, as well as the actions of the other. Our main finding is that increased interaction between the agents can lower the speed of learning: when both agents observe each other, learning is significantly slower than it is when one only observes the other. This slowdown is driven by a process in which a consensus on ...

This paper is concerned with a semiparametric Bayesian framework for estimating quantile treatment effects. A general and flexible model is specified using a generalization of the Pólya urn scheme. An exact Bayesian analysis is carried out by Markov chain Monte Carlo simulation methods. The proposed techniques are illustrated by estimating the effect of participation in the 401(k) retirement pr...

The existing superstar model (Rosen, 1981) does not require imperfect substitutes, and the convexity of total earnings with respect to talent is due to greater output for those with more talent. Our model explains why wages would increase at an increasing rate in talent. Imperfect substitutability between non-superstars and superstars with probabilistic production results in convexity in wage r...

This paper presents evidence on the contribution of public capital (infrastructure) to aggregate economic performance using panel data for the U.S. states from 1970 to 2005. Attempts to quantify the role of public or government capital in the production of aggregate output have been beset by thorny econometric issues since the early 1990s. We present estimates of the output elasticity of public...

In a recurring auction early bids may reveal bidders’ types, which in turn affects bidding in later auctions. Bidders take this into account and may bid in a way that conceals their private information until the last auction is played. The present paper analyzes the equilibrium of a sequence of firstprice auctions assuming bidders have stable private values. We show that signal-jamming occurs a...