Working Papers

Trust the Experts? Relative Performance of Inflation Expectations, 1946-2022


I study long-run series of individual and professional inflation forecasts from the University of Michigan Survey of Consumers and Livingston Survey of professional economists. I find that the average professional forecast generally outperforms the average consumer forecast. However, that superior performance is attributable exclusively to periods of low and stable inflation. During periods of high inflation and inflation regime change—both from low and stable to high inflation, and from high inflation to disinflation—the average consumer forecast is more accurate and rational (unbiased, with serially uncorrelated errors) and efficient (fully exploits available information). I find 3.5% inflation to be a critical threshold above which consumers rationally predict inflation. Professional forecasters affiliated with commercial banks and labor organizations generally outperform other professionals during periods of price stability, with the latter also exhibiting less bias and more rationality and efficiency than other professionals during periods of inflation regime change.

Skin in the Game: Liability Insurance, Contingent Capital, and Financial Stability


I exploit historical border discontinuities before the U.S. National Bank Act of 1863 to investigate the effects of liability insurance, extended shareholder liability, and branching on bank activity and financial stability within contiguous county pairs. I find that while double liability and branching lowered the probability of bank failure, public and mutual liability insurance generally elevated the probability of failure in both crisis and non-crisis years. Banks with double liability experienced smaller declines in deposits, note circulation, and lending during crises, while mutually insured banks experienced larger declines. I also find that liability insurance and double liability significantly affected ex ante risk-taking; insured banks substituted deposits and interbank borrowing for note issuance, increased exposure to real estate and interbank lending, and reduced cash reserves, while banks with double liability were less levered, less exposed to real estate, less reliant on deposits for funding, and maintained higher cash reserves. Finally, though I find no evidence of a trade-off between stability and credit provision, there is evidence of trade-offs between stability and industrial development.

Slavery, Inequality, and Development: Evidence from the Georgia Experiment


From 1735 to 1751, the Board of Trustees of the Province of Georgia imposed the only ban on slavery among the North American colonies. Exploiting the historical boundary between the 88 counties of Trustee Georgia and the 71 counties that were appended to the colony after 1751, I analyze the effects of this initial institutional difference on subsequent differences in slave dependence, land inequality, income, and poverty. I find that counties that had been covered by the initial Trustee ban subsequently had lower slave population density, fewer farms holding more than 10 slaves, and have higher income and lower poverty rates today. I further find that while counties affected by the ban did not have significant differences in pre-Civil War land inequality, productivity, industrial development, or educational investment, their economic output was significantly more diversified and less reliant upon the production of cash crops. Finally, I demonstrate that controlling for pre-war output diversification significantly reduces the estimated relationship between Trusteeship and current income. Results therefore suggest that the effects of initial differences in labor institutions can persist even where those differences are not determined by geography, and that a primary channel of persistence is the path-dependence of early economic specialization.

Liability and Innovation: Evidence from the World Fairs, 1851-1876


General adoption of the limited liability corporation during the nineteenth century has often been credited with mobilizing passive investment capital for entrepreneurial ventures of the “Second Industrial Revolution.” Utilizing a dataset of innovations presented at the World’s Fair in 1851 and Centennial Exhibition in 1876, this paper examines the effects of the limitation of shareholder liability on the volume and sectoral distribution of industrial innovation. I find that the introduction of limited liability was associated with a higher marginal, but not inframarginal, propensity to innovate. Though countries allowing limited liability exhibited no more or fewer inventions than countries without limited liability, a longer period of legal existence of the limited liability corporation was associated with more inventions. Moreover, I find that the effect of an additional year of limited liability on the propensity to innovate was highly concentrated in the invention categories of mining and food processing. The results of this study therefore suggest that the limited liability corporation was not a significant determinant of technological innovation in the aggregate, but was an important factor in facilitating innovation within particular industrial sectors.

On Machinery: Water Mills, Steam, and Technological Unemployment, 1700-1820


This paper exploits the historical location of water mills in the United Kingdom prior to the Industrial Revolution to examine 1) the short- versus long-run local labor market effects of labor-saving technological change; 2) the relative importance of access to coal in the growth of British textile manufacturing; and 3) the persistence of early agglomeration effects on subsequent economic development following the obsolescence of initial geographic advantages. With the invention of the Arkwright spinning frame and Cartwright power loom, locations which were more or less suitable to the generation of hydro power experienced differential changes in the cost of textile production. I therefore exploit parish-level differences in hydraulic suitability for a water wheel to instrument for historical water mill location using the second moment of river flow, and then estimate average changes in textile employment after 1768 and 1785 in parishes with a water mill versus without. Preliminary results indicate that while overall employment in textile manufacturing increased in parishes with a water mill, relative to parishes without, employment in particular occupations declined, and vagrancy counts rose. Results also show that the association between water mills and post-water frame and power loom employment changes is attenuated after 1820 by including various measures of access to coal as independent variables, suggesting that relative abundance of coal may have been a determining factor in the growth of textile manufacturing in later, but not initial, stages of the Industrial Revolution in Britain.