YSI – Economic History Graduate Webinar: Sophia Liu

Sophia Liu, PhD candidate from Boston University, will present her work: Evaluating the Effects of the 1918 Influenza Pandemic: The Role of the Great Depression

If you wish to attend, please register for the entire series here: https://forms.gle/LtqDgzoDtAojZg1S8

Attendees will receive by email on Wednesday morning the link to join the webinar.

Abstract:
Previous work by Almond (2006) argued that the 1918 flu pandemic lowered the educational
attainment of birth cohorts directly affected by the health shock. However, neither Almond nor any of the subsequent literature critiquing his original paper considered that the treated cohorts reached the cusp of the decision to attend high school in 1932-33, at the nadir of the Great Depression. One view is that the Depression increased high school attendance because the opportunity cost – the value of employment for young teenagers – had fallen sharply. However, another hypothesis is that secondary workers, including young teens, increased their labor supply – the so-called added worker effect – in response to the unemployment of household heads. If these economic effects interacted with health status, it is possible that the long-run impact of the pandemic on educational attainment, as estimated by Almond, may have been exacerbated or mitigated.
To explore the interaction effect, I first construct a sample of the relevant birth cohorts linked to the 1920, 1930, and 1940 complete count censuses, using the machine learning approach pioneered by Feigenbaum (2016). My linkage rates are similar to other studies using this approach, and the final sample contains 141,627 individuals, more than sufficient for econometric analysis. Crucially, the linkage establishes where individuals were born, where they lived in 1930; and, through the final link, where they lived in 1940 (and 1935), so I can match geographic variation in the relative severities of the influenza pandemic and the Great
Depression to individual locations. The severity of the Great Depression is captured at different geographies by county-level retail sales, state-level employment, and county or city level unemployment from the 1930 unemployment census. The relative severities of the two events, especially when measured at the county or city level, are sufficiently independent to provide identification of the interaction effect. A rich set of household-level variables is included as additional control variables.
My preliminary findings indicate that, overall, the Depression moderated the effects of the
pandemic on the treated cohorts; that is, if the Great Depression had been less severe (or not happened at all), Almond and others would have found larger negative effects on educational attainment. My results indicate the importance of considering how studies of the long-run effects of pandemics and other “natural disasters” need to take account of subsequent economic shocks that may interact with them, depending on the severity of the subsequent shocks and their timing during the lifecycles of the treated cohorts.