Zach Ward, Associate Professor of Economics, Baylor University

Social mobility over multiple generations is often lower than predicted by two-generational data, suggesting that traditional estimates fail to capture long-run mobility. Using novel US data (1850-1940) that links over 1.7 million individuals across four generations, we find that the economic status of great-grandchildren is strongly tied to that of their great-grandparents, with a transmission estimate of 0.43. This suggests low long-term equality of opportunity in American history.  Yet this four-generation estimate is just 13% higher than predicted by two-generation estimates once we adjust for measurement error by using multiple observations per ancestor. If we ignore measurement error, as is standard in the historical literature, the gap between four- and two-generation estimates increases to 120%, creating a false impression that two-generational estimates substantially miss the mark on long-run mobility.