The growing use of performance-based research funding systems has motivated increased interest in how they influence researcher behavior. This article draws both on survey and publication data to examine developments in researcher behavior and publication activity for individual researchers since the implementation of the Norwegian Publication Indicator in 2004. The Publication Indicator is a system for documenting Norwegian academic publishing, with the aim of measuring publication activity and allocating research funding according to the publishing performance. A main feature of the model is that publications are classified at two levels, where higher-level papers are valued more in terms of the model's publication points. Points are then awarded based on the fractional counts. When following the group of researchers that has been active over the entire period, 2004-12, we find that average points per researcher have fallen over the period. However, at the same time, average publication counts and number coauthors per paper have increased substantially. Essentially, productivity in publication counts has increased but productivity in fractional counts has fallen, as has average publication points. While both increased publications and collaboration would appear to be viewed as positive developments from a policy standpoint, the model's point system risks discouraging collaboration.