Productivity of Unionized and Non-Unionized Workers in the Construction Sector: A Review of the Literature

In their seminal work, Brown and Medoff (1978) argue that unions could have both positive and negative impacts on productivity. Negative examples include restrictive rules imposed on working arrangements or resistance to new technologies. Positive examples include higher levels of capital per worker, more training, lower worker turner, and the attraction of better workers through higher wages. Brown and Medoff (1978) argue, along with a number of authors in the following decades, that in the end the impact of unions on productivity is an empirical question. This report reviews the literature on the topic, focusing mostly on empirical works. While early works found a large positive impact of unions on productivity, later results are more conflicted. More recently, new methods of causal inference such as Regression Discontinuity Design have shed light on the topic, finding either a positive impact or inconclusive results. One main take away from the literature is that null effects can often hide heterogeneous impacts. For example, unions could affect productivity differently in different countries or industries, or depending on the composition of the workforce in different plants. Developments of more detailed datasets at the firm level, as well as matched employee-employer datasets should allow for better and more precise results in the future.