The danger of depending on productivity

Weekly Updates

Federal Reserve Chair Warsh’s recent Congressional testimony made several references to the longer-term productivity gains that might arise from using artificial intelligence. Whether from genuine belief or the political expediency of suggesting something that could drive inflation and interest rates lower, Warsh advocates for the productivity pixie scattering magic dust across the economy.
Productivity is an elusive thing to target. Productivity is everything economists do not understand, attractively packaged into a single statistic. Statistical quality has deteriorated, and structural changes raise more questions than they answer. What economists do not know, especially about the labor market, is increasing every year.
The risk is that productivity numbers are falsely inflated. This false productivity does not create any disinflation benefit (and will be exposed when gaps in economists’ knowledge start to fill).
Macroeconomic productivity will also be affected by the structure of the labor force. While a specific sector of the economy can become more productive, and potentially employ fewer people, that tells us little about the economy’s productivity. The people whose jobs have been lost are likely to work in different areas of the economy, and those areas are more likely to be less productive. Looking across the new workforce structure, economy-wide productivity improvements are likely to be muted.