Adaptive Learning and Iterated Weak Dominance

This article addresses the idea that rational players should not play iteratively weakly dominated strategies by showing that when a particular type of adaptive learning process converges, then players must have learned to play strategy profiles equivalent to those that survive iterated nice weak dominance and, for certain games, equivalent to those that survive iterated weak dominance. For games satisfying the weak single crossing condition, the set of strategies that survive iterated weak dominance is small in that its bounds are pure strategy Nash equilibria. The results hold regardless of the order in which dominated strategies are eliminated.Journal of Economic LiteratureClassification Number: C72.     

Labour quality and aggregate real-wage dynamics

The real wage plays an important role in many aggregate econometrics studies of the labour market. However, average wages, like any other average quantity, can change for two reasons – either through an underlying movement in the earnings potential of a unit of labour of given quality or, alternatively, through compositional changes which alter the relative proportions of different types of labour within the employed labour force. Using data from the General Household Survey to estimate earnings functions, the paper seeks to examine the contribution of these two factors for Britain from 1979 to 1989. It is shown that in periods of rising unemployment compositional changes can have a bigger effect on real-earnings growth than any underlying change in reward to labour of give quality. This raises the question of what economists actually are measuring when considering the behaviour of the aggregate real wage.     

Bootstrap HAC Tests for Ordinary Least Squares Regression*

There is a need for tests that are derived from the ordinary least squares (OLS) estimators of regression coefficients and are useful in the presence of unspecified forms of heteroskedasticity and autocorrelation. A method that uses the moving block bootstrap and quasi‐estimators in order to derive a consistent estimator of the asymptotic covariance matrix for the OLS estimators and robust significance tests is proposed. The method is shown to be asymptotically valid and Monte Carlo evidence indicates that it is capable of providing good control of significance levels in finite samples and good power compared with two other bootstrap tests.