Although the study (or summary) heavily implies that the decline in unionizing causes greater income inequality, the standard correlation-is-not-causation cavea

Although the study (or summary) heavily implies that the decline in unionizing causes greater income inequality, the standard correlation-is-not-causation caveat still applies. In particular, there's a plausible argument that greater automation is a confounding variable, insofar that (a) easy-to-automate jobs tend to be more unionized than hard-to-automate jobs and (b) automation leads to outsized rewards for relatively small groups of people (investors, owners, inventors, etc.) That's not to say unions don't matter, but I'm curious if there's any research that attempts to distinguish between unionization and new technological developments.