On the economics of forced labour. Did the employment of Prisoners-of-War depress German coal mining productivity in World War I?

The scholarly discourse about twentieth century forced labour has raised important questions. For example, how profitable and productive has the employment of forced labour been in different political and economic contexts? The dominant take-away from the literature is that forced labour comes with negative productivity, but positive production effects. Yet much evidence on productivity is anecdotal. To add a new quantitative take on this issue, this paper analyses the natural experiment conducted in World War I Ruhr coal mining, where, beginning with 1915, Prisoner-of-War (POW) labour was successively employed in many, but not all mines. The question to be answered is whether mines employing POW labour incurred significant labour productivity losses compared to non-POW employing mines that cannot be explained otherwise. To this end, we borrow from the treatment effects literature and implement two estimators – a baseline difference-in-difference fixed effects estimator and a doubly robust treatment effects estimator. Our study is the first to assess the productivity effects of POW employment using a full population of establishments of a particular industry. Our findings strongly support the view that the benefits from employing POW labour – i.e., the output-effect – came at the expense of a significant loss in productivity.