Abstract
When employers observe imperfect measures of worker effort, theorists typically assume that observed effort is unimodal and symmetrically distributed. Though observable effort may be distributed in different ways within a work day, for example, available field data on these effort distributions are rare. The symmetry assumption is largely untestable as a result. This paper presents empirical data from two experimental work environments that question the validity of such assumptions. For these piece‐rate work environments the author finds that observed effort is significantly negatively‐skew (i.e. modal>mean effort). The author's hypothesis is that on‐the‐job leisure causes this skewness in observed effort distributions. There are both theoretical and practical implications of this asymmetry. Some implications from the theoretical agency literature that we discuss include: self‐selection into rank‐order tournaments, optimal wage spreads in rank‐order tournaments, and optimal wage contracts with asymmetric information. Copyright © 2006 John Wiley & Sons, Ltd.