Research

Assembles federal outlays from 1962 and tax expenditures from 1967 to 2024 to measure six decades of US industrial policy. Support for private investment and innovation has run at roughly 0.6–1.0% of GDP under a broad definition (1.0–1.9% with defense R&D); what changed was the instrument mix, from visible direct subsidies toward tax expenditures and credit programs that attract less scrutiny. On a comparable basis US spending is below the European and OECD average and likely below China's. The CHIPS and Science Act and the Inflation Reduction Act are a renewed commitment, not a break with the past.

Surveys the causal evidence on US industrial policy instrument by instrument and identifies an evaluation gap: instruments with evidence we would consider strong account for only about a third of measured expenditure. Federal credit programs, the instrument class that appears to have been most effective in East Asia, have been studied as small-business and export finance but not for their effects on innovation or the structure of production. Reviews evidence from other countries and the relative merits of different instruments.

During the Depression, workers moved to farms as a form of migratory insurance, largely to farms with low levels of mechanization, while the mechanized agricultural sector shed workers into low-productivity or subsistence farming. The crisis reversed the normal process by which workers displaced by farm equipment are released into more productive occupations, which suggests that macroeconomic fluctuations shape the labor-market consequences of technological change.

Uses the digitized Congressional Record and the Google Books corpus to measure the polarization of political language since 1873. Discourse in books became more polarized in the late 1990s but remained low relative to the late nineteenth and much of the twentieth century; polarization in books, not in Congress, predicts legislative gridlock.

The political economy of apartheid-era South Africa in the 1970s: how labor scarcity in the mining and manufacturing sectors split the governing coalition.

Digitized historical datasets on the South African economy from 1948 to 1994: the agricultural, mining and manufacturing censuses, TEBA mine-recruiting records, elections, strikes, forced removals, the Manpower Survey and daily financial series, with the 279 magisterial districts as the common geography. Each dataset carries a validation status, from transcribed to reproduced in analysis; every correction is recorded in a ledger that the panel builders apply, never made by hand.