The exercise by an Australian state agency of coercive powers against construction industry workers has been justified by reference to claimed gains in productivity and hence national welfare. Yet the literature suggests that a more cooperative approach to union-management relations would offer better opportunities for productivity improvement. This article examines the data behind the productivity claims and finds that they were erroneous, probably due to incorrect transcription, and that the source data indicated no relative productivity gains against the identified benchmark. Despite being made aware of this, the state agency and its consultant maintained the original claims about the size of productivity and welfare gains from the use of coercive powers. Official cross-industry and time series data also showed no productivity gains arising from the use of coercive powers. However, there is some evidence that there has been a shift of income shares in the industry from labour to capital. The findings have implications for understanding the role of commissioned studies in public debate, and for regulation of the construction industry.