ABSTRACT. Several studies assess the impact of collegiate programs (e.g. sports, clubs, extracurriculars, etc.) on the number of applications that schools receive. Employing similar methodology, we analyze how local sales of recreational marijuana (RMJ)--as opposed to legality--affect the magnitude and quality of college applications. We argue that as RMJ sales become more ubiquitous, schools observe structural increases to the number of applications, with no evidence of changes to the quality of applicant. Furthermore, these increases disproportionately affect larger, more well-known schools than their smaller counterparts, typically through increases to out-of-state applications. Our results inform present debates about RMJ policy and indicate that RMJ legalization provides benefits to some colleges through larger applicant pools, though these benefits were strongest for first-mover legalizing states.

ABSTRACT. Local human capital growth is widely recognized as an important regional development strategy. For decades, many rural areas have lost their most skilled workers to urban areas. The dramatic increase in oil and gas development in the U.S. has offered hope of reversing this trend to many rural drilling communities. There are two possible channels through which oil and gas development can impact human capital growth. First, local oil and gas development may impact the human capital development of original residents by changing incentives for young adults to pursue higher education. Second, oil and gas development attracts many migrant workers from other communities that may change the human capital composition in drilling communities. Some evidence already exists about how oil and gas changes incentives for existing residents. Using synthetic control methods, Rickman, Wang, and Winters (2017) found that states with high levels of drilling had reduced high school and college attainment among original residents. We add to this literature using restricted American Community Survey microdata and proprietary oil and gas production data to examine both the question of human capital development among original residents and the net human capital migration effects from oil and gas development at the county level.

ABSTRACT. This paper studies the effects of college graduates' spatial sorting on regional economic development and individual subjective well-being. Exploiting survey data, census, and information about return-to-hometown incentives, we find that a greater share of return-to-hometown college-educated labor promotes local economic development and reduces regional disparity through the reinforced interaction of human capital and social capital in the hometown. Returning to hometown has negligible short-run effects on college graduates' subjective well-being but significantly improves the long-run outcomes. We show that relative income, job satisfaction, socioeconomic status, health care quality, and social capital are the main channels at work. We argue that new college graduates' preference for staying in big cities over returning to nest provides suggestive evidence of the theory of bounded rationality.

ABSTRACT. Violence has deleterious effects on human capital, race-related concerns permeate all levels of education, and terror events by white supremacist groups have increased in the United States since 2016. Examples of race-related violence in education include protests during school integration following Brown vs. Board of Education and recent bomb threats at historically black colleges and universities. Linking the Global Terrorism Database to a large, comprehensive survey of U.S. college freshman from 1984-2010, this paper asks how experiencing terrorist events by white supremacist terrorist groups in high school affects non-white students' high school GPA, test scores, college choices, and political views. Fixed effects regressions that control for state, year, race (asian, hispanic, black, etc.), and gender suggest that non-white students exposed to a white supremacist terrorism event in high school reduces the likelihood of getting a HS GPA greater than 3 (at least an A-) by 3.4 percentage points (pp), reduces ACT composite score by .388 points, reduces combined SAT by 26 points, and increases the likelihood that black students attend a historically black college or university. White terrorism events also increase the likelihood that non-white students identify as liberal by 1.1 pp and white students to be 2.7 pp more likely to identify as liberal. Further analysis will investigate the appropriateness of interpreting these effects as causal. This demonstrates how racially-motivated terrorism has long-term externalities for education and society and perhaps is a factor that is related to structural racial inequalities persisting in education.

ABSTRACT. Using unique survey data collected by Heartland Forward in Spring 2021, this paper estimates the willingness to pay for broadband characteristics across urban and rural populations in each of the 20 heartland states, as well as in a nationally representative sample. In addition to identifying differences in preferences across regions and states, this paper will also discuss issues related to broadband affordability and how they vary across space.

ABSTRACT. We study how tax policy affects the competition for venture capital by the creative class in two regions A and B. The creative class in each region produces a final good with venture capital and its own abilities or creative capital. Venture capital moves freely between the two regions and the representative creative class member in each region has access to an initial amount of venture capital. Each region taxes venture capital at a particular rate and the tax revenue is paid out as a transfer to the representative creative class member. In this setting, we perform five tasks. We begin by determining the first-best tax rates in the two regions. Second, we solve for the net price of venture capital and then express the objective function that is to be maximized in each region as a function of this price. Third, we compute the first-order necessary conditions that describe the optimal tax rates in the two regions and show that the sign of the tax rate depends on the net exporting position of the region. Fourth, for specific parameter values, we calculate the two tax response functions and discuss their properties. Finally, we compute the two equilibrium taxes as a function of the model’s key parameters and show that these taxes must be of opposite signs.

ABSTRACT. Does computerization promote or depress the growth of unincorporated entrepreneurship? On the one hand, computerization replaces routine-intensive jobs and shifts wage workers toward unincorporated self-employment. On the other hand, it complements the productivity and efficiency of wage workers and decreases the comparative advantage of unincorporated self-employment. The net effect of these offsetting forces is unclear. This paper presents the first evidence of a causal relationship between computerization and the growth of unincorporated self-employment. We first present the theoretical influence mechanisms underpinning two alternative hypotheses turning on the replacement effect and the efficiency-augmenting effect. Based on these two hypotheses, we empirically examine the impacts of computerization in the US between 1990 and 2010. We find that computerization decreases the growth of unincorporated self-employment in general, in contrast to the employment polarization that occurs among wage workers. However, for industries with lower computer adoption rates, unincorporated self-employment relative to wage employment increases. This implies that computerization relocates unincorporated self-employed workers towards low technology-augmented industries. The main conclusions are robust across alternative settings.

ABSTRACT. Between 1995 and 2021, the number of craft breweries in the United States increased from just under 800 to over 8,000. The popularity of craft beer is driven by consumer demand for beer that offers greater variety, in terms of flavor and style, than the beer traditionally offered by large multinational brewers such as Anhueser-Busch. The local ownership and the small production-scale of craft breweries also appeals to consumers, who prefer beer that has a stronger connection to place than that produced by faceless multinational corporations. Indeed, craft beer’s popularity is reflective of a broader neolocalism movement in the United States. As craft breweries became increasingly ubiquitous features on the American landscape, demand for locally grown hops increased. The result is that places with no history of hop production started producing one of beer’s key ingredients. The purpose of this paper is to examine the geographic spread of hop production within the context of the broader neolocalism movement, and to identify the challenges of growing hops outside of the crop’s core production region, the Pacific Northwest.

ABSTRACT. The input-output (IO) accounts published by the Bureau of Economic Analysis (BEA) are widely used in economic analysis, for example to study the changes in the structure of the United States (U.S.) economy and to perform economic impact assessments of specified events, activity levels, or policy changes. In addition, BEA tables are also used to generate regional tables for sub-regions of the U.S. for the purposed of regional economic modeling applications. To attain a homogeneous relationship between input structure and products, BEA redefines the commodity output of secondary production. When the input structure for a secondary product of an industry differs significantly from the input structure for the primary product of that industry, the value os secondary production is moved from the industry in which it generated to the industry in which this output is considered primary production. At the same time, the associated inputs and value added for this secondary product are reallocated from the origin industry to the destination industry as well. The detailed steps associated with this redefinition are not documented, which makes it difficult to appropriately assess the impacts on employment and income as these variables are only reported by industry, almost surely corresponding to tables before redefinitions. In our study, we first replicate the direct requirements, domestic direct requirements, and total requirements matrices from BEA’s standard Make and Use tables (tables before redefinition) as well as supplementary Make and Use tables (tables after redefinition), then compare the differences between the estimated matrices and the published matrices on BEA’s website to verify that BEA’s published requirements matrices are generated using supplementary Make and Use tables. Next, different redefinition scenarios are simulated with regard to those industries facing the largest output changes, and the logical counterpart to redefinitions of employment and income are estimated to evaluate the effect on estimates of employment and income impacts within economic impact assessments. We also create regional IO accounts to determine the implications of these changes in an open economy case.

ABSTRACT. There is a common maxim that more detailed input-output models are superior in accuracy to their less detailed counterpart models, even if some of the detailed data are imputed. However, how would the comparison change if the more detailed, disaggregated data were less accurate than their aggregated counterparts? As the share of detailed data values that require imputation increases, the disaggregated model results will also deviate from perfect accuracy. Is there a point at which using an aggregate model with greater certainty will provide more accurate results than a disaggregated model with greater imputation-based error and uncertainty? We used a simulation approach to address these questions using first-order aggregation bias (FAB) as a tool for assessing the tradeoffs between detail and accuracy. Regional accounts for four states, California, Illinois, Kentucky, Texas formed the foundation for the analyses. We treat the 67-industry regional accounts and their 15-industry aggregated counterparts as though they were perfectly accurate and generate corresponding FAB measures. We then simulate inaccuracies in the detailed output vectors that drive the regionalization algorithms constraining the disaggregated sector sums to be equal to their original accurate aggregated sector counterparts. We structured the output perturbations in two different ways, first to capture the effect of over-and underestimating output values by fixed and increasing increments, and second, to capture the effects of random perturbations. The FAB was recalculated at each simulation step to characterize the nature of the tradeoffs between model accuracy and aggregation given varying degrees of uncertainty in the underlying detailed data.

ABSTRACT. The Brazilian economy experienced a reduction in the participation of industry in the GDP since 1980 decade, however it became more pronounced as of 1990, reaching only 15% participation in the 2000s, with the years 2011 and 2012 exhibiting one of the worst levels of industrial production in the last 20 years. In order to understand the factors behind this slowdown, we have analyzed the structural factors of the economy that influenced variations in value added, labor force, and gross value of production. We used the IO - structural decomposition analysis (SDA) to decompose ten different effects of structural changes in these three main variables in the period of 1995 to 2013. The results show that industry was one of the sectors that suffered most negative impacts in the creation of jobs and started to add less value in production. These impacts, especially in the transformation industry, were caused mainly by the decrease in direct labor requirements by this sector and by the modification in the final demand patterns. Impacts of technological changes were also noted.

ABSTRACT. Modern secondary input output models provide alternative strategies for leveraging regional data in measuring economic contributions and impacts. In this study, we evaluate two alternative approaches, the analysis by parts approach and the editing production function approach, to measuring the economic contribution of farm-raised alligators on the Louisiana economy. Using data from a survey of alligator farm input costs, we evaluate the tradeoffs of these different approaches on measuring total effects through applying IMPLAN.

ABSTRACT. This paper studies the spatial distribution of households’ income inequality across 722 US local labor markets—or commuting zones—in 2005–2019. Theil’s Index and its within and between components are first measured using annual household-level data from the American Community Survey. Estimates suggest the rising income inequality pattern is mainly driven by the rising pattern of income inequality within local labor markets rather than between them. Decomposing the within component into the shared and the differentiated structure of the distribution of households’ income across local labor markets, we find that about two-fifths come from a shared distribution and three-fifths from the differentiated part. While New York’s labor market has the highest contribution to both within and between components, San Francisco’s, San Jose’s, and Seattle’s markets posed a high and increasing contribution in both as well. On the other hand, Texas and Oregon displayed an attractive reduction of income inequality in most of their local labor markets.

ABSTRACT. In this paper we develop a three-region economic geography model with workers of heterogeneous skills and mobility rates to consider how first-nature, regional differences impact both inter- and intraregional inequality. In our model, the skill premium within a region summarizes both the degree of intraregional inequality between mobile, skilled workers and immobile, unskilled workers and the interregional inequality through differences in the welfare of unskilled workers across regions. Regions with the highest skill premium have the greatest degree of intraregional inequality and provide the lowest level of welfare to unskilled workers, relative to other regions. We find that the skill premium will be higher in regions with a greater supply of unskilled labor, lower supply of housing, or are more remote. An increase in a region's housing supply or centrality will lower intraregional inequality and raise the welfare of the local, unskilled workforce. However, the magnitude of these changes are declining in the initial number of skilled workers in the region. The model is extended to consider imperfectly elastic housing supply. The larger the price elasticity of housing, the larger the range of values, such that more populated regions will host a disproportionate share of skilled workers, have lower levels of intraregional inequality, and provide higher levels of welfare for unskilled workers.

ABSTRACT. Economic development can be influenced by various policies such as improving infrastructure, changing the legal system, or increasing educational attainment. However, to the extent that culture influences economic outcomes, that would be very difficult for policy to alter. To examine culture’s role, we assess religion’s influence on historical regional economic development using 19th-century Prussian data. We find that compared to predominantly Catholic Prussian regions, Protestantism facilitated 19th-century industrialization and agricultural productivity growth. On the other hand, there was no positive significant impact of Protestantism on early 19th-century regional population growth, though there is a negative and significant effect in the latter 19th century. This result is robust to using IV regression. Protestantism’s positive impacts on industrialization and agricultural growth cannot be explained by differing education levels or by differing birthrates across regions, ruling out other indirect effects of Protestantism, suggesting other cultural roles.

ABSTRACT. The difference-in-difference (DID) framework is now a well-accepted method in quasi-experimental research. However, DID does not consider treatment-induced changes to a network linking treated and control units. Our instrumental variable network DID methodology controls first for the endogeneity of the network to the treatment and, second, for the direct and indirect role of the treatment on any network member. Monte Carlo simulations and an estimation of the drought impact on global wheat trade and production demonstrate the performance of our new estimator. Results show that DID disregarding the network and its changes leads to significant underestimates of overall treatment effects.

ABSTRACT. Growing concerns about PFAS (per or poly fluoroalkyl substances) pollution in drinking water in the last several years has spurred many states into legislative action (PFAS is a byproduct of production of materials like Teflon cookware and firefighting foam). For instance, New Hampshire has instituted limits for four specific types of PFAS chemicals in its waters after several high-profile incidents within the state. While research on the epidemiological consequences of PFAS ingestion by humans is at relatively early stages (limiting the information on the health effects of exposure), there are few studies on how individuals assess and value the risks that may be associated with elevated levels of PFAS ingestion. Using data collected from a contingent valuation survey of residents in New Hampshire, this is the first study we know of to estimate how much individuals are willing-to-pay for treatment of PFAS in public drinking water systems. This study is somewhat unique in that lacking better epidemiological data (e.g. the type one might have when studying arsenic) it is important that respondents are presented with both information from the state of medical knowledge on cause-effect of PFAS ingestion, and with what essentially constitutes uncertainty on the benefits of water treatment to reduce or eliminate PFAS exposure. Both state and local governments are eager to obtain this kind of data as treatment costs could be extensive for public water supply systems.

ABSTRACT. The 2010 Deepwater Horizon (DWH) oil spill was a major shock to the marine environment within the Gulf of Mexico that also had implications for the economy of the region, which is closely tied to the marine resources that the Gulf of Mexico provides (Sumaila et al., 2012). While there have been oil spill events in the Gulf region prior to 2010, a technological hazard as severe as the DWH oil spill was unprecedented. The proximity to the location where the oil drilling rig exploded varies by state that borders the Gulf and the extent to which oil reached the waters or beaches of individual states varied significantly, largely dependent on ocean currents. This research investigates the variation in impacts of the 2010 DWH oil spill on the commercial fishing supply chain across five Gulf states (Alabama, Florida, Louisiana, Mississippi, Texas).

The regional economy of the five Gulf states depends on commercial fisheries and associated industries to a great extent. After the explosion approximately between 4.16 and 6.24 million barrels (Crone and Tolstoy, 2010) of oil was released and it spread to a larger extent and reached almost nearshore areas of the coastal states of Gulf of Mexico. Impacts on the commercial fishing industry came as a collateral damage of the oil spill. The direct impact of the oil spill on the environment and the ecosystem and the indirect impact through people’s changed perception about Gulf seafood, potentially threatened the supply chain of commercial fishing. Along with the environmental issues, any change in fishing regulations, higher fuel costs, lower seafood prices etc. can implicitly impact industries that are a part of the commercial fishing supply chain.

To determine the impacts of the oil spill on the commercial fishing supply chain in the coastal states along the Gulf of Mexico, we estimate the average treatment effect on the treated (ATT) specifically for the four industries that comprise the commercial fishing supply chain using a Synthetic Difference-in-Difference (SDID) model. We utilize the Quarterly Census of Employment and Wages (QCEW) data from the United States Bureau of Labor Statistics (BLS) for this paper. We investigate three economic indicators (Quarterly Establishment Count, Monthly Employment Level, Average Weekly Wage) from the following industries - Fishing, Seafood Product Preparation and Packaging, Fish and Seafood Merchant Wholesalers, and Fish and Seafood Markets.

The primary contribution of this paper will be to estimate the causal effect of the DWH oil spill on each of the Gulf states separately and compare the results to identify the states and industry sectors within the states that experienced the largest effects. Identifying the sectors in the supply chain that were mostly affected due to oil spill and the temporal trend associated with each affected state can potentially highlight the risk areas following an oil spill. This can be helpful for future assessment of informed decision making to support stakeholders associated with the relevant sectors and response readiness.

ABSTRACT. It is well established in the health literature that females have a higher life expectancy than males. Differences in life expectancy are attributed to factors such as employment, rurality, poverty, race, environment, and geographic location. However, they can also be due to biological and social differences between sexes. The literature has established that the difference between male and female life expectancy has been decreasing. However, what is causing this change is yet to be studied. In this paper, we analyze which factors influence the difference between male and female life expectancy in the United States, particularly over time. While individual characteristics such as race, income, physical activity, and health are essential to understanding differences in life expectancy, community factors such as rurality, job composition, and food access also play a role. Using spatial econometrics and data from government and academic sources, we consider how both community and individual characteristics influence the difference between male and female life expectancy.

ABSTRACT. Appalachia has been at the center of the drug crisis that has now exploded across the United States, wreaking havoc on communities and destroying lives. Other than a brief, hopeful leveling off in 2018, drug use and overdose deaths have continued to accelerate. One key focus of policies aimed at dealing with the crisis has been treatment and recovery resources, which provide critical support for those struggling with addiction. The Appalachian Regional Commission wishes to document the distribution of treatment and recovery resources within Appalachia, which is critical to assessing unmet needs. This report describes, reviews, and evaluates the usefulness of the National Survey of Substance Abuse Treatment Services (N-SSATS) database for summarizing and analyzing substance abuse treatment resources and services at the substate level that would be required for understanding the extent of unmet needs within Appalachia’s counties and subregions. Though the focus of our initial research was on a substate/subregional analysis of the Appalachian Region, the issues discussed are broadly applicable to other areas.

ABSTRACT. We study a one-period model of an aggregate economy composed of cities and regions that demand vaccines designed to fight a pandemic such as Covid-19. The supply of vaccines is the outcome of Bertrand competition between two firms A and B. The marginal cost of producing the vaccine for both firms is stochastic and drawn from a uniform distribution. In this setting, we perform three tasks. First, we describe the equilibrium pricing strategies of the two firms and then we compute their mean ex ante profits. Second, we permit both firms to conduct risky R&D and then determine the conditions under which only one firm engages in R&D and conditions under which both do. Finally, we introduce a way of mimicking the effect of increased competition and then analyze the impact of this increased competition on the incentives to conduct R&D faced by the two firms.