Performing Department
Agri and Applied Economics
Non Technical Summary
The 1980s farm crisis had a large impact on agricultural households and on the rural communities in which they were residing. Communities in the Midwest were particularly hit hard. Land values had fallen roughly 30% on average from 1980-1984, with the largest decreases seen in the Midwest. In 1985, half of the commercialized farms were insolvent (US Comptroller General 1985). The crisis hastened the restructuring of the farm industry, accelerating the decline in the number of farms. The crisis also restructured the social fabric of affected communities. Literature in sociology and psychology have shown that the farm crisis increased political activism, depressive symptoms, and marital instability in affected households. In addition, the crisis accelerated out-migration from agriculturally dependent counties and changed the demographic composition of the farming occupation as young, educated, and commercialized farmers were most likely to have failed farms.This project proposes to investigate the impact of the farm crisis on the level of social capital in Midwestern communities in the long run. Economic models of social interactions typically have multiple equilibria, such that a shock to social interactions can change equilibrium behavior in the long run, leading to a persistent effect of the crisis. We will examine the impact of the farm crisis on current levels of social capital as well as other social barometers such as deviations in life expectancy, the extent of rural blight, policy preferences and voter turnout. A careful examination of the long run effects of a major economic crisis in agriculture on social capital in rural communities will illuminate the causes of rural decline and help policymakers better respond to future economic shocks. Which in turn, will impact occupation choice and resource management.
Animal Health Component
0%
Research Effort Categories
Basic
(N/A)
Applied
100%
Developmental
(N/A)
Goals / Objectives
The first objective of this project is to assemble a data set. While readily available data exist at theUS county level, we plan to dig deeper to introduce new measures of social decay, such as the extent of rural blight, and/or obtain more disaggregated data to look at the relationships at the neighborhood level.The second objective is to test the main hypothesis concerning the long run impact of the farm crisis. Given that the economic shock altered individuals' social behavior as the sociological and psychological literature have documented, we expect that a new long run equilibrium emerged. As such, communities that were more heavily hit by the farm crisis should have lower levels of social capital today. If, however, the shock had a temporary effect on behavior, then we would expect to see no long run differences.The third objective is to formulate a set of policy implications and disseminate findings.
Project Methods
We begin with a brief discussion of measurement. There are at least two ways to measure the impact of the farm crisis on the community. The first is the share of firms that are financially overleveraged and the second is the percent of decline in land prices. Both have been used in the literature. Another possible measure is the number of failed banks. We will, of course, explore other possible measures as we learn more about available data.The main outcome variable that we will use is a county-level measure of social capital constructed by Rapasingha et al. (2006).We will also introduce an additional measure if data are available, the extent of rural blight, as a proxy for social decay. We are also particularly interested in political preferences and life expectancy. Finally, while each of these variables is of interest to economists and policymakers, we may consider mediation analysis by looking at other outcome variables that are endogenously determined by the 1980s farm crisis, such as income inequality, land values, farm debt structure and agricultural efficiency.This project will draw on a variety of data sources: social capital by US county from 1990-2014 constructed by Rapasingha et al. (2006), vital statistics from the National Center for Health Statistics, General Social Survey, US Census, US Agricultural Census, and historical archival material as well as other Wisconsin-specific data such as the Wisconsin Longitudinal Survey.General Social Survey, and the US Agricultural Census:The primary method used will be regression analysis. The crisis was unexpected and depended upon a confluence of forces, that depended upon the interaction between policies determined at the national and international level and local agricultural and economic factors. Some counties felt the crisis more deeply than others due to exogenous factors. We will therefore use a difference-in-differences strategy, following Orman (2019). We will compare the changes in social capital between now and before the farm crisis in affected and unaffected agricultural dependent rural counties in the US. We will assume that but for the 1980s crisis, the social capital variables in these counties would have had similar trends after controlling for host of regional and crop-type trends.The first pitfall concerns threats to causal identification. Our measure of the intensity of the 1980s farm crisis could be correlated with both observable and unobservable factors that are they themselves correlated with our community level outcomes. One possible strategy is to use weather shocks as instrumental variables. Instrumental variables should have no impact on the outcome except through their relationship with the potentially endogenous regressor. These weather shocks should affect farmer income, making it more or less likely that farms fail during this crisis period but should have no effect on social capital (in normal times, these shocks would just affect transitory and not permanent income).Any study on historical persistence must grapple with the concern of misattributing the uncovered relationship to historical persistence. Historical variables could be correlated with ex interim or contemporary factors that drive outcomes and are not easily controlled for because of measurement difficulties, a bad control problem or multicollinearity. Therefore, it is imperative in this literature to provide a convincing theory of change and persistence as well as empirical evidence that can trace out the persistence. We will be able to partially address this issue since some outcome variables are observed at multiple points in time since the 1980s.The second pitfall is measurement error in key variables. This is particularly a concern in the measurement of the impact of the farm crisis. Again, an instrumental variables solution can be used.