How similar are pay structures in 'similar' departments of economics?

Using a unique panel data set spanning 21 years, we estimate a fixed-effect s model of pay determination for five Ph.D.-granting departments of economi cs in large Midwestern state universities. Despite program similarities, no two departments have comparable reward structures, which points to the per ils of generalizing across universities. We also explore the effects of faculty leaves. Although the first sabbatica l leave typically has little effect on pay, a second sabbatical is usually associated with higher pay, consistent with the proposition that a second s abbatical restores more lost human capital than the first sabbatical. With some exceptions, leaves without pay are associated with lower pay. Finally, in estimating rewards for research, we find that at four of the un iversities returns to quality overwhelm returns to quantity. At these unive rsities, an article published in the AER boosts pay by as much as 11 percen t, whereas an article in an unranked journal increases pay by at most 1%. ( C) 1999 Elsevier Science Ltd. All rights reserved.