ABSTRACT. In the U.S., wind power is an important, low-cost contributor to the electricity grid, and its share of total production is steadily increasing. However, because wind is intermittent, wind power generation can be highly variable. Accordingly, an increased reliance of wind power can make it more difficult for utilities to optimize their production decisions across their portfolio of sources. Thus, accurate 12-hour ahead wind forecasts are important for utilities that use wind power as a source of electricity generation.

Over the past several years, the National Oceanic and Atmospheric Administration (NOAA) has developed and updated the High-Resolution Rapid Refresh (HRRR) model as an important weather forecasting tool. In this paper we estimate how improvements in wind forecasts help utilities better manage their internal production resources. Overall, we find that improved wind forecasts have led to significant cost savings for utilities, but that further improvements in wind forecast modeling can still provide significant positive economic impacts.

Our first step determines if forecasts of predicted electricity output are more accurate as the HRRR model evolves. We exploit the fact that when the HRRR model is updated, both the current and new versions run for several months. In this analysis we compare wind forecasts from each model (e.g., HRRR2 versus HRRR3) to recorded wind observations from weather stations located near wind turbines. Forecast accuracy is said to improve as the difference between the predicted and actual wind speed declines.

The second step assigns economic values to wind forecast errors. Here we draw on industry data to estimate cost increases when there are forecast errors. The costliest error occurs when utilities expect to produce more wind power than they actually do. Because electricity demand must be met, utilities must turn to the spot-market when they do not generate enough themselves. This is expensive, as spot-market prices tend to be much higher than the costs of production from most utility-owned sources.

ABSTRACT. Greenhouse gas (GHG) emissions from freight transportation account for about 10% of the total emissions in the U.S. and are steadily growing. While many factors contribute to the rapid growth of GHG emissions from freight sources (such as trade growth, fuel efficiency, mode share, just-in-time delivery, etc.), this paper focuses on fuel price. The largest source of energy consumed in transportation is petroleum. Since different freight transportation modes have different fuel dependencies as well as emission intensities, I will look into the impacts of petroleum price on mode share and the corresponding GHG emissions. Due to little data available on truck or rail interstate inflows/outflows, this research will apply gravity models for interstate trade flow estimation within the U.S. state-level multi-regional input-output (MRIO) framework. I will build the relationship between average fuel cost by mode ($ per ton-mile) and petroleum price using fuel price, fuel consumption, and freight activities by mode. The average fuel cost by mode will be used in the multinomial logit model to estimate mode share. The Commodity Flow Survey data and Freight Analysis Framework 4 State database from the Bureau of Transportation Statistics will be used for model calibration. This research helps reveal how much would fuel prices have to rise to cause significant intermodal shifts and affect GHG emissions.

ABSTRACT. This paper investigates the regional economic impact of the latest boom and bust in the oil and gas industry in the United States. Three geographic regions, encompassing ten U.S. states that had substantial variations in the production of oil and gas between the years 2000 and 2018, are considered. While many have undertaken to examine the local economic and demographic impacts of the oil and gas boom (Weber 2012; Weinstein 2014; Brown 2015; Munasib & Rickman 2015); Komarek 2016; Tsvetkova & Partridge 2016; Feyer et al. 2017; Marchand & Weber 2018), we also appraise the bust period post-2014 and compare the boom and bust effects. Additionally, the divergence of revenues and employment post-2014 allow us to decompose the effects to illustrate how each impacts the local economy differently.

Taking advantage of two unique data sets, we explore the relationship between the boom and bust in the oil and gas sector and regional employment. We use proprietary data on oil and gas production compiled by Enverus, which provides detailed information on every oil and gas production unit in the United States. The geographic information allows us to aggregate production data to obtain monthly and yearly oil and gas production measures at the county level. Using historic monthly average prices of oil and gas, we are able to construct monthly and yearly oil and gas revenues by county for the period of interest. We also use a county-level employment and labor earnings data set by industry produced by Economic Modeling Specialists International (EMSI). This data set compiles various data sources to construct yearly estimates of county-level employment and earnings per worker for 323 4-digit North American Industry Classification System (NAICS) industries.

Using the data on production, revenue and employment in the oil and gas sector, we document the boom and bust phases in the various regions of interest. We then explore potential spillovers into other sectors of the economy. Similar to Black et al. (2005), we find that natural resource production variation has a significant impact on general employment at the regional level. The results show that employment in the construction and retail sectors are positively associated with the oil and gas cycle. We find no statistically significant association documented between employment in the manufacturing sector and the boom and bust in the oil and gas sector. The empirical analysis shows heterogeneity in the findings between the three regions, and between rural and urban counties within each region.

ABSTRACT. The confluence of abundant natural gas, rising concerns over reducing greenhouse gas emissions, and a mix of state and federal policies are leading to shifts in the energy composition of the United States. Nowhere is this confluence more on display than in the power-utility sector, especially in rural portions of the country. Coal was approximately 50 percent of the fuel source used in power generation for many decades. However, in the mid-2000s coal's share began to decline and by 2018 represented only 27 percent of fuel used in electricity generation. Natural gas is already displacing coal in power generation because of the shale revolution in the United States. Between 2007 and 2012 it is estimated that abundant natural gas displaced 28 percent of coal-generated electricity. In 2018 over 60 percent of electric generating capacity installed was fueled by natural gas, while nearly 70 percent of retired capacity was fueled by coal.

Due to these shifts in power generation, the number of coal-fired power plants has declined across the country. Coal-fired units have shut down because of sluggish growth in electricity demand and increased competition from natural gas and renewable sources. Between 2001 and 2018, over 250 coal power plants closed. Previous research has shown higher levels of local air pollution near coal power plants. As a result, there could be improvements in local air quality following closures. Burning of coal is known to emit substantially more particulate matter relative to natural gas. As a result, the decline in coal-fired power plants is expected to reduce emissions in the local area where closures occur.

A general air pollutant of concern is particulate matter (PM2.5), where 2.5 references the size of air particles measured in micrometers. Researchers have focused on PM2.5 because of its diffuse and harmful nature, especially due to its association with higher risks of respiratory and cardiovascular issues. For example, prior research has found higher mortality risk from exposure to PM2.5. While others have investigated effects of the rise of natural gas via hydraulic fracturing and stockpiles of coal on local particulate matter, no research that we are aware of, has directly estimated the effect of coal power plant closure.

We help fill this gap in the literature by combining spatial data on air quality monitor stations, power plant emissions and closures, as well as local economic conditions. Using monthly data from air monitor stations and power plants, we estimate the effect of coal power plant closure on local air quality within 25 and 50 mile buffers of each monitoring station between 2001 and 2018 using a difference-in-difference strategy. Controlling for total power production, local economic conditions, location by year and monthly fixed effects, we find that the average effect is a 16 to 21 percent reduction in the level of particulate matter (PM2.5) three months after the plant closure. Thus, an improvement in local air quality from coal-fired power plant closures may also provide additional health benefits in these areas. Using previous estimates from the literature, the reduction in particulate matter is associated with an 8 to 15 percent decline in some mortality probabilities. Thus, one positive local externality from the closure of coal-fired power plants is less emissions and lower risk of adverse health effects.

ABSTRACT. This paper utilizes a Bayesian Hierarchical Model (BHM) to obtain improved estimates of state-level panel effects of changes in the minimum wage on teenage employment rates. Employing modern Hamiltonian Monte Carlo (HMC) sampling methods, aggregate time series data are used to formulate a dynamic model specification and an informative prior distribution for the parameters of state-level panel models. Traditional results indicate that there is considerable sampling noise in the panel data, strengthening the case for use of BHM, which mitigates the wide variability of state data while allowing for heterogeneity across states. Furthermore, we show that employing a stationary dynamic specification along with the informative prior distribution improves the estimation results and provides superior out-of-sample forecasts.

ABSTRACT. This paper expands the QoL and QoBE analysis into an evaluation of their role in mitigating recent labor demand shocks (The China Shock, Great Recession and afterwards). Focusing on micropolitan areas and regions, we explore the role both QoL and QoBE play in lessening or distributing regionally the effects of exogenous shocks to labor demand (e.g. import penetration or automation risk). In our model, we test the role of these shocks on the composition and level of employment, along with the level of compensation (both wages and proprietor’s income), when controlling for Quality of Life and the Quality of Business environment.

ABSTRACT. In a standard economic model, shifts in labor supply or demand induce changes in the market equilibrium outcome. The direction and the size of the changes in wage and employment depend on the direction and the size of these shifts. The literature on climate-related disasters shows that the estimated impacts on employment differ across different types of climate-related disasters. In some cases, the estimated impacts even differ within the same type of disasters across different studies. However, in contrast to the wide coverage of employment impacts in existing studies, only a couple of studies have examined the impact on wages. The lack of information on wage impact makes it impossible to make further economic inferences about the underlying mechanism driving the observed labor market impact. To capture the total labor market impacts, it is also important to consider both the impacts of disasters that occurred in the recent past and in the neighboring counties as these are still largely unknown for most of the disasters. Furthermore, the differences in the study areas, data resolution, and methods create additional complications preventing attempts to generalize the lessons learned from these studies. To further the understanding of the findings in the existing literature and initiate meaningful policy discussions, it is important to understand the mechanisms generating the diverse employment impacts and get a more complete picture of the labor market impact of climate-related disasters.

Using the U.S. county-level disaster data from 1991 to 2015 and Spatial Durbin Error Model (SDEM), this paper studies the disaster impact on employment and wage, and the underlying mechanism of these impacts. Eight types of climate-related disasters are considered in this study. These are drought, flooding, heatwaves, hurricane, severe storm, tornado, wildfire, and winter weather. The results show that different disasters have heterogeneous impacts on employment and wage. For example, employment tends to increase in the year with the incidences of wildfire, hurricane, and heatwaves, and decrease in the year with the incidences of flooding, and severe storms. Wage tends to increase with the incidence of heatwaves and hurricanes. The results also show that climate-related disasters have significant temporal and spatial spillover effects. Using the principal component analysis, proxies for labor demand and supply are constructed. It is found that the standard labor demand and supply analysis has the potential to explain the heterogeneity in the labor market impacts of different types of climate-related disasters. Two climate-related disasters may generate similar changes in labor market equilibrium, yet the underlying mechanism might differ. A better understanding of these differences in local labor market outcomes is important because it can help to develop effective disaster-preparedness programs and disaster-relief policies that can improve the socio-economic resilience of the local economy.

ABSTRACT. Fracking innovations revolutionized the United States oil and gas industry and facilitated a boom in energy production in states with oil and gas resources. This paper examines effects of oil and gas booms within a state on individual employment and earnings. To account for endogenous migration decisions, we instrument for oil and gas production in workers’ state of residence via the predicted percent of oil and gas employment in their state of birth. We find statistically significant and economically meaningful positive effects. The bulk of the effects accrue to workers employed outside the oil and gas industry indicating sizable spillovers.

ABSTRACT. This paper investigates the long-term relationship between early life exposure to malaria and adult socioeconomic outcomes in Brazil. The identification strategy relies on exogenous variation in the risk of malaria outbreaks in different states and seasons of the year to identify early life exposure according to the timing and location of birth. Furthermore, Brazil has undergone a successful campaign of malaria eradication during the late 1950s, which allows for comparing outcomes of birth cohorts born just prior to and just after eradication to identify the extent of in utero exposure. Instrumental variables estimates find consistent negative treatment effects of in utero exposure to malaria on socioeconomic outcomes, such as educational attainment and health status. The effects are stronger for exposure during the first trimester of pregnancy than during other periods of gestation. Additionally, consistent with previous findings, men are more likely to exhibit larger long-term effects.

ABSTRACT. Nationwide obesity is on the rise. Despite this national trend, parts of the U.S. exhibit higher rates than others, particularly in the Bible Belt. In this paper, we use a spatial Durbin model (SDM) to analyze obesity in South Carolina – a predominantly rural state with a prevalence of obesity and poverty. This allows for identifying socioeconomic features to explain obesity while differentiating local from neighboring influences, a research focus that is often overlooked in the existing literature. Preliminary results suggest that local obesity rates rise with rates in neighboring counties, implying that regional influence is important. Also, several factors affect obesity beyond the traditional dietary indicators, some of which include healthcare, recreation, demographics, and economic conditions. Because obesity is often viewed as a national epidemic, policies to reverse it are scaled in a similar fashion. But differences can be observed regionally and distinctions can be made with regard to spillover effects, as our paper submits. Policies addressing obesity should embrace a regional approach and thus account for the influence of external factors and surrounding geographies.

ABSTRACT. The relationship between diabetes and oral health has been well established. People with diabetes tend to have a number of oral health complications including tooth decay, gum disease, and periodontal disease—conditions that may result in the need for extractions.

In 2014, South Carolina implemented an adult dental benefit for its Medicaid recipients. Our study uses individual Medicaid claims data from South Carolina covering 2014–2019 to explore whether medical costs increased for adults with diabetes as a result of tooth extraction(s). Our data include claim information on all adults with diabetes who had a dental claim. Outcomes are then compared for those with tooth extractions and those without.

Additionally, we explore whether outcomes vary among the four geopolitical regions of South Carolina: Upstate, Midlands, Pee Dee, and the Lowcountry.

The findings will aid in better understanding the true costs associated with tooth extractions in this population, and whether outcomes vary by geographic location.

ABSTRACT. The introduction of a direct flight between two locations in different countries allows for faster travel and a lower cost of acquiring information, potentially facilitating acquisitions abroad. We evaluate this hypothesis by considering cross-border mergers and acquisitions (M&A) activity between China and the U.S. following China's accession into the WTO.We find that the introduction of a direct flight between Chinese and U.S. locations leads to about 87 percent higher likelihood of an M&A deal. Further, our results suggest that the introduction of a direct flight is associated with about 1 new cross-border M&A transaction annually and about $12.6 million annual increase in M&A volume. These estimates are robust to different empirical specifications. We also demonstrate that reverse causality is unlikely to drive our results. Finally, we find that following the introduction of a direct flight, these deals have lower premiums, higher probability of cash payment and higher cumulative abnormal return, while acquirers' operating performance improves subsequent to the deals.

ABSTRACT. This empirical application models food manufacturing entry and exit decisions as a bivariate Poisson process. Entry-exit counts of NAICS 311 firms spanning from 2002 to 2017, enumerated for Oklahoma’s 77 counties, are estimated as a function of demographic characteristics, variables that proxy localization and agglomeration economies, and industry structure. Spatial covariance between entry-exit events is modeled with a Gaussian Process Regression.    

ABSTRACT. Dollar General plans to open 1,000 new stores in 2020 in the United States and can already be found in 44 states. Using a unique business plan, Dollar General is low-cost retailer that makes shopping convenient by utilizing small stores, rather than the traditional big box approach, with (thus far) only speculative impacts on their competitors. Using the National Establishment Time Series dataset for the state of Florida between 2000 and 2014 we exploit the difference in timing of Dollar General openings to investigate its effect on businesses nearby. Specifically, we quantify the effects on nearby firm revenue, entry, and exit, as well as the composition of area businesses. The unique nature of the NETs data allow us to focus on the impacts of businesses within a 0.5 mile radius, 1 mile radius, and 5 mile radius, rather than using pre-defined geo-political boundaries. In addition, we investigate differential impacts in urban and rural areas.

ABSTRACT. This paper aims to measure the effects of oil and gas development on a state’s opioid overdose death rate. We collected county data for the overdose death rates for the period of 1999 to 2016 and use county employment data estimated by the Upjohn Institute to measure for changes in employment by county. We estimate a spatial panel model to account for spatial spillovers and spatial and time fixed effects. To account for oil and gas development, we follow the work of Rajbhandari (2017) and estimate proxies for change in percentage of the sectoral employment over time. We also include variables to capture changes in other relevant and significant sectors, such as farming and agriculture, construction, and manufacturing. Our results point towards an effective decrease in death rates in times of boom and increases in times of bust. We find that some sectoral changes are associated with county overdose death rates. We also find evidence of positive spatial spillover effects on neighboring counties. We include policy-related variables and find them to have a significant effect on opioid overdose death rates, though not always the expected effect.

ABSTRACT. Is a region’s occupation and industry risk profile related to its level of opioid abuse? Individuals already encountering significant morbidity and mortality risk in employment activities may exhibit less caution with respect to other life activities, including addictive substance use. This paper investigates the potential linkages between county-level opioid prescription rates and occupation and industry morbidity and mortality risks using prescription rate data compiled from the Centers for Disease Control and Prevention (CDC), along with industry and occupation risk data from the Bureau of Labor Statistics (BLS) and the National Institute for Occupational Safety and Health (NIOSH). Additional analysis attempts to disentangle issues related to endogeneity and spatial heterogeneity.

ABSTRACT. As the opioid epidemic continues, it is important to evaluate potential reasons for why certain areas of the U.S. are impacted more than others with regards to opioid misuse issues. The objective of this study is to explore the theory that areas with a higher percentage of their workforce in “high-risk” industries are more likely to endure opioid-related deaths. This relationship would be due to persons in these types of jobs being more likely to receive opioid prescriptions for pain management or workplace injuries over those who work in low risk occupations. Analysis will be conducted at the county-level for all U.S. counties, and a panel dataset covering years 2001 through 2017 will be analyzed. To determine the percentage of workers in high-risk jobs, data will be gathered from the Bureau of Economic Analysis (BEA). Information will be pulled on how many residents work in industries which are considered to have a high possibility for work-related injuries (mining, construction, manufacturing, etc.). County-level opioid-related age adjusted death rates will be created using the Centers for Disease Control and Prevention restricted-access mortality files. Additional demographic control variables will be gathered through utilizing the U.S. Census. To control for causality issues, the association between the percentage of employees in a high-risk industry will be lagged in a series of fixed effects models. This study will also evaluate if the findings change based on a county’s metropolitan status, to see if working in a rural versus urban area influences opioid-related mortality.

ABSTRACT. Any characteristic estimated for a given geography has some degree of error or uncertainty associated with it. Previous work has highlighted that spatial resolution greatly affects this error, and as a result, any further metrics calculated using the data. This paper seeks to examine how spatial resolution changes with the injection of noise in data obtained by the U.S. Census of Population and Housing. This opportunity is uniquely available with the introduction of a differential privacy algorithm to the 2010 decennial data. First, location quotients and confidence intervals obtained from the delta method are utilized to explore the relative intensity of 3 chosen characteristics for both U.S. counties and states. These are calculated using both the 2010 published data set and the new 2010 demonstration data product released by the Census Bureau. By using both data sets it is possible to draw inference on how the injection of noise affects spatial errors.

ABSTRACT. This article analyzes the determinants of the spatial youth labor participation using a Spatial Autoregressive Probit Models. Using the location data of 2165 young people, we find that the education, sex and poor conditions increases labor participation. Similarly, we found that being an Afro-descendant and informality have effects over participation in the labor market. The results regarding the race effect are robust using various specifications and determined how young people of African descent are less likely to participate in the labor market.

ABSTRACT. As a generalization, two colloquial theories have risen to evaluate the shifting culture of the United States through time. The first argues that the United States is a melting pot of religion, values, and demographics--implying a convergence of culture to a singular point. The second posits that the U.S. mirrors a salad bowl rather than a melting pot--suggesting that individuals retain their cultural identities and coexist rather than adopting ever similar ideologies to those of their peers. This investigates which of these two competing theories most accurately describes the spatial evolution of culture in the United States given that population dynamics and migration can both affect the demographic composition of a region. Using Census Bureau data from 1970-2010, I regress the longitude and latitude of a city on demographic and cultural variables specific to that city. These regression coefficient estimates are used to predict the longitude and latitude of each city. Finally, a within-city prediction error term is generated to assess how close these predicted locations are, relative to the actual location of the city. The dynamics of this error term suggest that the United States exhibits characteristics of both a melting pot and a salad bowl--therefore adding nuance to our understanding of cultural shifts in a spatial context.

ABSTRACT. The purpose of this investigation is to empirically assess whether Mexican states’ governments are engaged in competition when designing and setting their local tax policies, particularly, that related to taxing wages and salaries paid to workers laboring within each state. The current public economics literature identifies at least three types of subnational interaction when defining their corresponding fiscal policies: tax competition, mimicking behavior and, more broadly, public and investment expenditure policies. By means of implementing a spatial panel Durbin specification to the thirty-two Mexican states in a ten-years timeframe, the estimation results point to the presence of significant spatial interaction which according to recent empirical literature may be elucidating the Mexican state governments are strategically interacting when implementing their income taxes thus evidencing some sort of tax competition may be encompassed.

ABSTRACT. The healthcare and higher education industries, or "meds and eds," have been found to contribute to cities' local economic growth, in part because they are reliant on a skilled workforce and therefore have the ability to attract educated workers, thus increasing local human capital. However, most of the existing meds and eds literature thus far has focused on urban areas, so little is known regarding whether meds and eds might have a similar effect in rural areas. Additionally, the definition of "eds" is usually restricted to postsecondary educational institutions, which tend to be less prevalent in rural areas; in contrast, K-12 education is found nearly everywhere, and it still requires a relatively skilled workforce. In this study, I explore whether county-level employment in the healthcare and K-12 education industries is associated with higher levels of educational attainment and above-median wages for educated workers, thus potentially contributing to both "place prosperity" and "people prosperity" across the rural-urban continuum. I use regression analysis to determine whether higher meds and eds employment is associated with higher educational attainment at the county level, and I also calculate a "wage premium quotient" by dividing median earnings for selected meds and eds occupations by county median earnings. With a better understanding of whether meds and eds have the potential to contribute to prosperity in rural as well as urban areas, it may be possible to design more effective local economic development policies.

ABSTRACT. We follow the lead of Glaeser et al. (2017) using unconventional data to assess regional entrepreneurial activity together with regional differences in personality driving differences in business formation as advanced by Obschonka et al. (2015). In this paper, we expand recent research using virtually contemporaneous, and geographically granular, user online activity to estimate a region’s proclivity for entrepreneurship. We assess the statistical relationships between business formation, operationalized as establishment births, and the web activity associated with a user’s interest in Third Places – informal gathering locations – and arts, music and design or “Arts Spaces.” We operationalize interest in and association with third places and arts spaces by the website activity each U.S. ZIP code. Initially developed for marketing analytics by dstillery llc, these data are derived by several proprietary algorithms that create consumer profiles aligned according to common market preferences.

Controlling for regional interest in entrepreneurship related web resources, we find that both third places and art spaces can explain more than half of the variation in regional business formation. Establishing that third places and art spaces may attract the attention of would be entrepreneurs as desirable places to live, work and start business may help address the missing catalyst in regions with lower rates of start-ups and business growth

ABSTRACT. India accounts for two-thirds of global maternal deaths and the highest number of infant deaths annually. Given the poor performance of maternal and infant healthcare service indicators, India launched two maternal health programs- a conditional cash transfer program called Janani Suraksha Yojna (JSY) in 2005 and a free services program called Janani Shishu Suraksha Karyakaram (JSSK) in 2011. The programs aim to reduce maternal and infant mortality through the promotion of public institutional delivery. Both provide financial incentives to pregnant women through a reduced price effect on delivery care. Using the Indian District Household Surveys, I exploit the differences in individual eligibility rules across states to estimate the impact of the programs on healthcare utilization and measures of infant mortality. The results suggest that JSY and JSSK reduced home births and increased the use of public institutional care. While JSY also shifted women away from private to public facilities, JSSK increased the use of private care. I also find evidence that JSY reduces fetal and perinatal mortality but does not impact higher days mortality rates. Further, the spread of information about healthier pregnancies by health facilitators under the programs increases the use of antenatal services. On the one hand, JSY reduces the probability of women going back for a postpartum checkup, and on the other, free maternal care services under JSSK increase their likelihood to return for a checkup. Also, I find heterogeneity in programs’ impact with more educated and poorer women benefiting the most from them.

ABSTRACT. The performance of Kentucky, its seven neighboring states, and North Carolina relative to the United States is analyzed for the period 1980-2016 in terms of per capita income and its major components. The unit of analysis is the sub-state region, of which there are 94 in the states under consideration. The area development districts, which have been identified by the states, are taken to be the sub-state regions. Cross-sectional and temporal variation in performance is studied both within and across states.

ABSTRACT. The Great Recession left its mark on the US economy in a variety of ways: lost output due to the cyclical downturn, higher unemployment, and wage stagnation. Apart from the usual measures of the damage to the economy from the Great Recession, a general feeling of malaise persisted past the official end of the recession in 2009. Since 2010, U.S. labor productivity has grown an average of 1.1% per year, one of the lowest rates since World War II. However, the slow growth in labor productivity has not been consistent across regions. Some states have experienced strong growth in labor productivity while other states have seen declines in labor productivity. This paper analyzes regional and state changes in labor productivity, seeking to describe patterns of change and to identify the causes of gains or losses in productivity.

ABSTRACT. Moriarty Graduate Paper Competition

Field-of-study mismatch is a relatively nascent area of research. While previous research has focused on vertical education mismatch (whether one is over, or under, qualified for a job), recent analysis has turned its focus to horizontal mismatch (whether one's degree is related to their occupation). Of the new research in field-of-study, almost all use subjective measures and focus solely on an individual level. However, the impact horizontal education mismatch has on the local and regional economy is equally pertinent. This paper seeks to address the lack of empirical work on field-of-study mismatch in metropolitan statistical areas and its subsequent effect on a city’s productivity. Modifying Duncan & Duncan’s (1955) index of dissimilarity, I create an index to measure the extent of the field-of-study mismatch in metropolitan areas between 1997-2017. These values are then subsequently used to measure the index’s impact on economic growth and health. Preliminary results indicate that mismatches from occupations that require specific degrees, such as Nursing, have a larger impact on economic growth than other occupations that require less specificity, like systems analysts. The preliminary results and ideas presented in the paper are promising and urge further research in the area.