Jesús Fernández-Villaverde argues in Marginal Revolution that saving shapes automation’s growth boom

Jesús Fernández-Villaverde argues in Marginal Revolution that saving shapes automation’s growth boom

Marginal Revolution published Jesús Fernández-Villaverde’s argument on October 20, 2026. He argues that automation-driven growth depends both on how tasks affect the returns to machines and on how strongly capital accumulation responds to those returns. In his standard task model, capital supply determines the size of the growth boom, while task structure determines its duration and how wages evolve. His comparison of two saving-rate assumptions leads him to argue that slow growth might reflect too little capital accumulation rather than a technological bottleneck. He says automation raises the saving rate mainly by increasing capital’s share, not by making households thriftier, and that long yields price the growth boom. His benchmark calibration also points to a distributional cost: hand-to-mouth households lose a substantial equivalent amount of consumption, while temporarily keeping them on trend has a GDP cost. The source presents these as model results, not observed outcomes; it does not establish how well the calibration describes actual households or what policy would follow.

Publication: 625bcc27-e960-d9da-f944-81a721192187

attribution — pass: The post quotes Jesús Fernández-Villaverde, and the reading attributes the argument and model results to him. It correctly presents Marginal Revolution as the publisher. The title's phrase 'argues in Marginal Revolution' is slightly loose, because the text is a quoted passage that the blog republished. The summary's wording 'Marginal Revolution published ... argument' makes the attribution clear enough. The blogger's own aside, 'More bottlenecks!', is not misattributed.

From Jesús Fernández-Villaverde:

summary — uncertain: Several central economics terms are used without plain explanation: 'hand-to-mouth households' (people who spend all their income and cannot save), 'capital share', 'task model', 'long yields' and 'on trend'. A non-specialist reader cannot follow the distributional point without these. The only supplied definition, 'Benchmark', describes AI test suites, so it would mislabel 'benchmark calibration'. The summary also drops the vivid figures, such as 11.7% versus 2.2% growth and the yield jump. Smallest repair: briefly gloss these terms, remove or replace the Benchmark tooltip, and keep the growth contrast in the summary.

hand-to-mouth households lose the equivalent of two-fifths to two-thirds of their consumption

date — pass: originalPublished is 2026-10-20T18:38:49Z, and the reading's date of October 20, 2026 matches its UTC calendar day. The evidence kind is publisher-page, so no archive disclosure is needed. There is no hindsight framing. The model results are presented as model results, not as outcomes.

support — pass: Each claim's excerpt appears verbatim in the source and is fairly paraphrased. The summary's caveat that these are model results, not observed outcomes, matches the source. The question follows from the calibration result without assuming an answer.

In my benchmark calibration, hand-to-mouth households lose the equivalent of two-fifths to two-thirds of their consumption, while keeping them on trend temporarily costs at most a fifth of GDP.

media — pass: No media was selected, and no verified media candidates were supplied, so no media was invented or misidentified.