Policies for the future of work should be based on its past and present

July 2024, Paper: "I have no idea what the future of work will look like. Will generative AI replace jobs? Complement jobs? Exacerbate inequality through skill-biased technological change? Reduce inequality by competing with higher-skilled workers but not middle-skill workers? No economist, AI engineer, or policymaker knows the answers to these questions either. Economic research holds out the hope, albeit often not realized, of coming up with a clear and widely agreed understanding of the past, as EIG’s American Worker Project has aimed to do. But definitive guidance on the future is too much to hope for. Undaunted by this existential challenge, discussions about the “future of work” took off in 2009, as shown in Figure 1—paradoxical timing given that the biggest problem of that moment was the unnecessary absence of jobs for millions of workers who were unemployed for reasons having nothing whatsoever to do with the future of work. The problem back then was instead the failure of policymakers to learn, sufficiently, a critical lesson from the macroeconomic past: the need for adequate demand. Since then, the unemployment rate has gone down, up, and down again—while the discussion of the future of work has just gone up and up more. In the fifteen years since this conversation took off, work has changed surprisingly little, except for a shift to work from home which was largely ignored by most of the discussions prior to 2020. Yet the future is, indeed, fifteen years closer. And even if the actual data does not lend any urgency, the newspaper headlines about generative AI certainly should."