ABSTRACT. The Bureau of Economic Analysis (BEA) satellite accounts are a set of complementary statistics to the national economic accounts that provide detailed information on key sectors of the U.S. economy. Examples of BEA’s satellite accounts at the national level, include health care, travel and tourism, R&D, arts and culture, and outdoor recreation. In recent years, BEA has moved in the direction of developing regional counterparts of the national satellite accounts, adding regional nuance to the more expansive national data. The arts and cultural production satellite account is BEA’s first satellite account with regional statistics. This presentation describes the framework, methods and source data used in the development of state-level arts and culture production statistics. In addition, it discusses potential source data and methods for developing experimental state-level statistics on outdoor recreation.

ABSTRACT. Spatially fragmented production is an important feature of regional growth and development. A large literature finds that Global Value Chains (GVCs) are a key source of competitive advantage for multinational firms and for regional manufacturers. GVC practices are often cited as a key reason for the patterns of outsourcing and performances of firms (e.g., firms such as GM and Chrysler filed for bankruptcy while Toyota and Honda did not). Typical value-chain analyses are based on case studies and small datasets of a select number of firms, countries or regions. We use U.S. Customs data on imports of vehicle manufacturers (VMs) assembling cars and trucks in the U.S. The data covers all import transactions at the HTS 10-digit level such as ‘Vulcanized gaskets’ and ‘washers and other seals’. We measure how many suppliers of a given component a VM imports from and how frequently a VM switches suppliers of that component. Using this disaggregated firm-level microdata, we examine two theories of vertical governance of supplier/customer transactions: (a) Transaction-based theories (Williamson, Grossman-Hart-Moore) that argue that all firms undertaking a transaction with similar attributes govern the transactions in the same way, and (b) Organization-focused theories (Dyer, Jacobides, Helper and Henderson) that argue that attributes of the transactors affect governance. Our regression analyses covering the period 2007-2015 provide some supporting evidence for both theories. For example, we find that the product attributes matter irrespective of the type of governance. At the same time, we find that Japanese-owned VMs have fewer suppliers and their relationships with their suppliers are longer-lasting. Characteristics of GVC governance including whether firms purchase on a spot basis instead of using relational contracts, the longevity of relationships among operations, the use of just-in-time operations by suppliers, and spillovers between supplies and customers have implications for the regional economy.

ABSTRACT. California's recent five-year drought was the worst on the historical record. Farmers reacted by increasing their use of groundwater. In response, California passed the Sustainable Groundwater Management Act, the first statewide attempt to regulate groundwater. Groundwater usage does not have to be reported, so detailed analysis of the possible new rules has been difficult. I solve this problem by using field-level crop data along with estimates of crop water-intensity and surface water deliveries from an important agricultural county in the Central Valley. Using data on aquifer levels over time and my estimates of groundwater pumping, I model the spatial evolution of groundwater levels, determining to what extent groundwater is common property. While pumping depresses groundwater levels more near the well site than elsewhere, I estimate that private pumping accounts for only 0.7% of the change in private groundwater levels. I then propose a structural model of crop choice and water usage. The model makes explicit the importance of the recent boom in land devoted to permanent crops, and illustrates the role of groundwater as a buffer stock against drought. I use the model to evaluate the impact of potential groundwater rules like Pigouvian taxes, and compare them to the status quo. I find that the optimal Pigouvian tax, while quite sizable, only has a small effect on welfare.

ABSTRACT. This paper examines the impact of changing the range of expenditures which serve as the basis for the Supplemental Poverty Measure (SPM) poverty thresholds, and expanding the estimation sample upon which the thresholds are based. Currently, the thresholds are based on the 30-36th percentile range of expenditures for food, clothing, shelter, and utilities (FCSU) incurred by consumer units with two children. This is in contrast to a percentage of the median of FCSU expenditures recommended by the National Academy of Sciences Panel in 1995 (Citro and Michael 1995). Moving back to the median has advantages, including that thresholds and resources would be more consistently defined. Fewer expenditures at the median of the FCSU distribution must be augmented to account for the value of in-kind benefits than at the lower end of the distribution. Further, if in the future, health care needs were to be accounted for in the SPM thresholds, spending at the median would be more representative of spending on private health insurance compared to the lower end of the FCSU distribution which is more likely to either have no health insurance or public insurance for which they do not pay. Additionally, this paper explores the possibility of expanding the estimation sample, whose expenditures underlie the SPM thresholds, from consumer units with exactly two children to either consumer units with one or more children or to all consumer units. Both options would result in larger sample sizes and would reflect the current population distribution more fully. Thresholds based on changing the range of expenditures and estimation sample are produced along with poverty statistics.

ABSTRACT. While most would agree that poverty thresholds in New York City should be higher than poverty thresholds in rural Alabama, there is much less consensus on the issue of how these differences in the cost of living should be reflected in the thresholds for the Supplemental Poverty Measure (SPM). Currently the SPM thresholds produced by the Bureau of Labor Statistics are adjusted by the Census Bureau for differences in the cost of rent and utilities using an index developed using data from the American Community Survey (ACS). This index is applied to only the shelter portion of the threshold. This approach does not take into consideration differences in amenities across jurisdictions.

Although many economists would agree that amenities should be incorporated in the construction of this index, there is no commonly accepted methodology for taking these amenities into account. This paper make a first effort to do this by arbitrarily reducing the “weight” of the median rent index by half before estimating SPM rates. The paper then examines the poverty rates for major demographic groups with this alternative geographic adjustment focusing on the individuals for whom poverty status changes with the alternative weighting.

ABSTRACT. Since the inception of the official poverty measure (OPM) in the 1960s, several anti-poverty programs, such as the Supplemental Nutrition Assistance Program and the Earned Income Tax Credit, have been created or expanded which provide resources not accounted for in the OPM. This has led some to question the validity of the OPM and push for alternative measures of poverty such as the Census Bureau’s Supplemental Poverty Measure (SPM), which accounts for non-cash benefits, differences in cost of living, and necessary expenses (including taxes, childcare, work and medical expenses). However, questions remain as to how both measures capture material well-being and correlate with other aspects of disadvantage, such as food security, wealth, and health, especially for individuals who are classified in poverty based on one measure but not the other. This paper will provide the first estimates from the redesigned 2014 Panel of the Survey of Income and Program Participation (SIPP) to explore how the OPM compares to the SPM in regards to capturing material well-being. Using the SIPP’s robust measures of material well-being, augmented with additional administrative sources, we examine how net worth, health outcomes (including mortality), food security, child development, and lifetime earnings (captured using the Summary Earnings Record and Detailed Earnings Record) vary for those in the OPM and SPM poverty. This analysis will improve our understanding of the relationships between choice of poverty measure and material well-being.

ABSTRACT. Opioid addiction is a factor relating to the automation of jobs, the vanishing middle class, and low labor force participation within the hotspots of opioid usage; but there is little information on whether it is a cause or a symptom of deeper macroeconomic maladies, economic distress, and lower wellbeing and life satisfaction. Our hypothesis is that the job losses due to automation and increasing levels of perceived job insecurity in manufacturing in these areas is a significant factor in opioids abuse. Depression and stress levels across the country show that the same areas that are affected by higher rates of automation and opioid addiction, are also showing the highest levels of economic distress. Most federal and state policies to address the opioid crisis tends to focus on federal drug regulations/de-regulations, treatment and rehabilitation of users. While these are important considerations, it is also critical to examine the social and economic factors that affect opioid addiction in order to not just staunch the flow of this contagion but also to address the underlying causes of this crisis that may have its roots in lower life satisfaction in areas that are suffering the most from job displacement and automation as well as rising levels of job insecurity among even the employed. In this paper, we also examine the racial disparities in opioids use and the factors that are driving the difference.

ABSTRACT. The U.S. is in the midst of a drug-induced mortality epidemic marked by the introduction and spread of opioids across both rural and urban communities. Research on the geography of the epidemic has focused on the association between declining local economic opportunities and increases in drug-induced mortality since 2000, but found little purchase. Here, we divide the epidemic into two periods and examine its changing demographic and geographic aspects. In the first period, beginning in 2000, mortality rates soared among the middle-aged as prescription opioid painkillers drove the epidemic. Physical disability is associated with chronic pain, and through 2010-2013, opioid painkiller prescription rates and drug overdose deaths rose most in areas with high initial physical disability rates. Since the early 2010s, however, the number of opioid prescriptions has declined nationally and illicit opioids such as heroin and, increasingly, synthetics such as fentanyl, have been responsible for a growing share of drug overdose deaths. Drug-induced mortality has risen particularly among young adult males. The geography of recent growth in rural drug-induced mortality is markedly changed. No longer related to physical disability, it has risen in states with high urban drug mortality and dense rural settlement, all in the northeast quadrant of the U.S., suggesting that state differences in the distribution channels are currently shaping state rural drug mortality. Within states where the distribution channels are most developed, county rise in drug-induced mortality appears related to local (county) economic hardship and outmigration.

ABSTRACT. The Ohio Violent Death Reporting System collects highly detailed data on both decedents and contextual factors surrounding the death for all overdose deaths in Ohio in 2016 and 2017 which are unavailable in the CDC MCOD data. We use these restricted access data to investigate demographic and contextual factors that influence the administration of the overdose reversal drug Naloxone to those who have died of a drug overdose. We find interesting patterns according to the demographic characteristics of the decedent controlling for Naloxone availability within the community. We estimate separate models for rural and urban areas and find significant differences in the predictors of whether the decent received Naloxone. The results have important policy implications for communities looking to stem the rise in drug overdose deaths.

ABSTRACT. TBA

ABSTRACT. TBA

ABSTRACT. TBA

ABSTRACT. TBA

ABSTRACT. TBA

ABSTRACT. Historically, natural gas prices are closely linked to supply disruptions in the Gulf Coast due to the role of offshore and Gulf states production in the total domestic natural gas supply. Over the past decade, the combined share of natural gas production from the states in shale-rich areas has increased dramatically, while Gulf region production has become less important. One reasonable assumption is then that the supply disruptions due to weather events in the Gulf Coast should have a much smaller effect on natural gas prices compared to the pre-shale era. This paper uses fixed-effects panel data models to examine whether this assumption is true, based on state-level natural gas prices data from 1995 to 2016. Property losses due to natural hazards in Texas and Louisiana are used to represent supply shocks in the natural gas market from the Gulf area. Various control variables are included to filter out the effects of other supply/demand factors in the natural gas market. Results show that natural gas prices in both importing and exporting states have become less responsive to natural hazards in Texas but more sensitive to hazard events in Louisiana since the shale boom. These results are robust to the break dates used, the geographical location of states (NE, MW, South, and West) considered, and the empirical specifications employed. The increasing importance of Louisiana in natural gas pricing is perhaps due to its role as the benchmark pricing location for U.S. natural gas and its expansive pipeline networks.

ABSTRACT. This paper examines local labor market outcomes from an oil and gas boom. We examine two main outcomes across gender, race, and ethnicity: the probability of employment in the oil and gas industry and the log wages of workers employed outside the oil and gas industry. We find that men and women both gain employment in the oil and gas industry during booms, but such gains are much larger for men and are largest for black and Hispanic men. We also find positive income spillovers for workers in other industries that are similar in magnitude across demographic groups.

ABSTRACT. College students care about a number of factors when choosing their major field of study, but their employment and earning prospects are especially important factors for many. The energy industry is particularly volatile, due to significant fluctuations in energy prices and production technologies. The energy industry also needs many graduates with very specialized skills in science and engineering, and some college majors are strongly tied to the energy industry. During good times, the energy industry is a high-paying industry for college graduates with related skills. During bad times, college graduates with energy related majors likely find it more difficult to find a job using their skills and may end up unemployed or taking a job in a less related field. An important question we want to address is whether and how students respond to energy industry fluctuations by altering their college majors. In this study, we examine the effects of the 1970s and 1980s energy boom and bust on the college major choices of college graduates who were likely making college major decisions around this time. We focus on two college majors in particular, petroleum engineering and geology. These two majors are strongly related to the energy industry, especially in energy producing states. Specifically, we observe increased rates of majoring in petroleum engineering and geology during the energy boom compared to the pre-boom period. We also observe steep decreases in petroleum engineering and geology majors during the energy bust relative to the energy boom period. Given the limited previous literature related to this topic, our analysis makes an important contribution to understanding how students make college major decisions in response to industry-specific shocks. Our results strongly indicate that young people respond to industry-specific employment conditions when making college major decisions. We also emphasize that the effects we find for energy related majors are strongly concentrated among graduates born in energy states. This indicates an important location-specific element to college major decisions. These findings increase understanding of how young people make college major decisions.

ABSTRACT. Literatures relate sudden expansion of tight oil and shale gas in the US with higher employment and income. However, dis-amenities like crime rates attract little attention. This paper exploits the US shale boom as a natural experiment and compares shale-infeasible states with economically-small, relatively rural shale boom states to investigate impacts on crimes and victimization costs. I utilize the double-selection post-LASSO method for proper selection on observable confounders and the Generalize Synthetic Control (GSC) approach to control the unobserved time-varying confounding effect. Unlike, synthetic control, GSC can estimate confidence intervals for counterfactual and incorporate multiple treatment units. This method is applicable for causal inference in the regional policy setting when treatment occurs in several states at different periods. My results show a significant rise in violent crimes among treatment states. I provide falsification tests for pretreatment trend and sample selection along with the robustness checks. My results support a causal interpretation. I estimate 15.68 million (2008 dollar) worth of the annual victimization cost of the fracking boom for the treatment states.

ABSTRACT. Large-scale energy investment projects in wind, solar, and coal-replacing natural gas have increased in frequency this decade and promise to accelerate in the next decade. In addition, large energy transmission projects are needed to facilitate energy flows within and across energy producing regions. This paper describes the process and challenges associated with measuring an ambitious recent plan to rapidly deploy wind, solar, some natural gas, and a revamped array of transmission systems in the U.S., while at the same time retiring primarily coal-fired energy sources. This project, called the Interconnection Seam Study, provided the technical foundation for an economic impact evaluation using conventional IMPLAN software and methods. However, reliably measuring the construction and operational consequences presented challenges in terms of data availability and the amount of construction-related detail needed to properly model the new investments. This paper reviews the components of the Interconnection Seam Study and how those components were addressed for modeling the annualized construction and operational effects. Importantly, it highlights limits to the analysis and the concerns that they raise for the analysts, advocates, and policymakers.

ABSTRACT. Economic development incentives (EDI) are prominent in US fiscal landscape as a tool to attract or retain businesses. The annual costs of such programs can amount to billions of dollars. Given its popular use since the 1990s, this paper analyzes whether there has been any convergence in EDI use across U.S. states. The goal is to investigate if EDI competition among states for business investment and jobs has contributed to a single steady-state path. Taking advantage of a national database on EDI expenditures, we apply the novel panel convergence method developed by Phillips and Sul (2007, 2009). We find some degree of convergence: our results indicate that there are three convergence clubs during the 2008-2018 period. We further explore factors behind the heterogeneous policy setting behavior. We find that while higher manufacturing share of employment is associated with higher probability in the high EDI spending club, higher household income and higher corporate tax rates are associated with lower probabilities. Notably, states with more troubled political culture and higher mining share of employment are more likely to be in the low spending club.

ABSTRACT. While the use of business tax incentives is pervasive, the decentralized nature of state and local fiscal policy in the United States, related data challenges, and methodological difficulties leave many unanswered questions related to the use and effectiveness of these incentives. Past examinations of effectiveness have defined successful outcomes in various ways, including business attraction, job creation, investment stimulation, industry or cluster promotion, and tax base expansion. However, beyond some efforts to examine cluster creation, comparatively little has been done to directly explore the relationship between taxincentivesand longer-term changes in an economy’s economic base. Better knowledge of whether incentives can be successful at changing location decisions in a way that greater diversifiesa local economy helps get to the root of the question of whether incentives can be a tool to increase resilienceto economic downturns or shocks.By taking advantage of data on incentives and taxes from the Panel Database on Incentives and Taxes, establishment and employment data from disaggregated industry classifications from the Wholedata Establishment and Employment Database, and characteristics of the local population from the American Community Survey, this paper exploresthe relationship between incentiveuseand economic diversity between 1990 and 2015in large U.S. cities. The paper first examinesthe relationship between state-level industrial diversity andincentive use to explore the characteristics of governments that were more likely to incentivize a more diverse set of industries. Next, the paper examineswhether economic development incentives wereassociated with an increase in industrial diversity over time. More specifically, the analysisexamines whether particular incentives wereassociated with increases (or decreases) in diversity than others. The longitudinal nature of the data allowsthese relationships to be examined over time. This is important to examine, as some earlier research has found that regions may find it difficult to retain firms that they have attractedin industries peripheral to their existing industrial base. To measure these effects, an index of diversity isregressed on lagged measures of incentive use as well as economic and population controls and state fixed effects. The paperconcludesby discussing how policymakers might strategically use financial incentives to diversify an area’s economic base.Finding a positive relationship between particular incentivesandincreased diversitycould support the application of those incentives by state and local administrators to diversify economies and reap the resulting benefitsof increased economic resilience. Conversely, no relationship might indicate a strategic mismatch between the incentives offered and thecomposition of a local economy, and a negative relationship between incentive use and diversity may reflect a preference by policymakers for incentivizing incumbent industries.

ABSTRACT. We investigate state-level economic incentive policies in the U.S. State legislators are increasingly interested in evaluating tax and incentive programs. According The Pew Charitable Trusts, 29 states had incentive evaluation legislation in place in 2018, up from just 7 in 2012. Data limitations are formidable: incentive policies vary over states, industry, and time; tax expenditures are not easily tracked; and tax administrators are not equipped to gather the relevant data needed for careful program evaluation. Such limitations are evident in the databases that attempt to track the value of incentives (GJF, C2ER) offered by states. Our analysis exploits the new Panel Database on Business Incentives for Economic Development (PDIT, Bartik 2017) which tracks state tax and incentive policies by year for 45 industries, 33 US states, including the 30 main cities in the 30 largest metropolitan areas from 1990 to 2015. The PDIT specifically focuses on incentive and tax policy changes and simulates the value of an incentive on the net taxation of a hypothetical firm in a particular industry and state. We exploit distinct incentive policy regime changes to investigate potential causal links between economic outcomes including wages, employment and poverty levels.

ABSTRACT. This study sheds light on the factors that affect the location decisions of manufacturing foreign direct investment (FDI) in the United States from 2006-2015. Specifically, the contribution of this study is to analyze the relative importance of state taxes and incentives relative to other fundamental determinants of the FDI location decision. The probability of investing in U.S. counties is modeled using a Poisson specification with random effects. The results suggest that agglomeration economies are the most significant factor in the FDI manufacturing location decision. The second most important set of factors are state taxes and incentives. The corporate income tax, property tax abatement, and investment tax credit have the biggest potential influence on the location decision. Other taxes and incentives considered have no effect on the location decision of foreign-owned firms: the sales and property tax, the job creation tax credit, the research and development tax credit, and the customized job training subsidy.

ABSTRACT. This paper examines state economic development incentive program additionality or "but for" effects. Empirical work draws on a 2018 survey of approximately 250 responding firms that received at least one award for a startup, expansion, or relocation project during the period FY2010-FY2016 from a list of twenty-six state economic development incentive programs (i.e., grants, tax credits, loan assistance) offered by the Commonwealth of Virginia. The paper evaluates the role of program design features (e.g., discretionary versus automatic, relative size of incentive), firm characteristics (e.g., size of firm, industry), and locational variables (e.g., boundary location, rurality) on the size of firm deadweight assessments.

ABSTRACT. Recent studies have shown that the neighborhood where a child grows up can significantly be correlated with his/her prospects of upward income mobility. Future outcomes of children can be shaped by neighborhood characteristics of the locality such as poverty rate, crime, or quality of schools where they spend their childhood. Moving to neighborhoods just within a few miles have been found to change the outcomes of children in adulthood such as income, teenage birth rates and even incarceration rates. It should also be noted that not all neighborhood characteristics that lead to productive labor markets such as job and wage growth are also conducive for greater intergenerational income mobility. Although studies on the role of neighborhood characteristics in economic mobility have been done and extensive datasets have been available on estimates of future children’s outcomes, neighborhood characteristics, race and gender, there is still a gap in our understanding of how similar/dissimilar these neighborhood characteristics are in shaping future outcomes of children raised up in families with different economic and racial background even within the same neighborhood. This study will analyze the structural differences in the correlates of economic outcome of children grown up with poor parents and high-income parents (parents at 25th percentile and 75th percentile respectively in their national household income distribution) with different racial/ethnic backgrounds in the US. Specifically, differences in the neighborhood characteristics affecting mean predicted percentile rank of children in their national household income distribution between children raised in poor vs. high-income households with be analyzed for various race/ethnicities such as White, Black, Hispanic, Asian, Native American, and other. With an assumption of differences in outcomes for male and female, separate analysis will be done for each combination of race and gender. Seemingly Unrelated Regression (SUR) will be employed to analyze the factors affecting mean predicted percentile rank of children with poor and high-income parents for each combination of race and gender at the county level. In addition to that, structural differences in each correlate will be examined by a Chow Test where different neighborhood characteristics will be used as correlates of predicted children’s income rank. The analyses will be done using recently released county level data on parent’s background, children outcome and neighborhood characteristics by “The Opportunity Atlas”. While some of these associate factors have been addressed by Chetty and his team at a national level conditioned on race and ethnicity, a more detailed understanding of these conditional expected ranks given the level of rurality have not been addressed. This study proposes this additional extension of their analysis. The results from this study are expected to shed some light on how the interaction of neighborhood characteristics, family income level and race/ethnicity are connected with the future income of children. Moreover, the results are expected to inform policymakers how policies targeted to improve neighborhood characteristics and increase upward mobility might not have similar effect for families of different income level and race/ethnicity.

ABSTRACT. The United States' economy evolved rapidly during the 20th century, generating massive changes in technology, income levels, inequality, and occupational structure. Changes in agriculture were particularly dramatic, as the increased application capital and new biological technologies led to a dramatic decrease in agricultural labor share, accompanied by similar productivity gains. At the same time, levels of intergenerational mobility rose along with mean incomes, albeit at different rates across time and space. I examine the relationship between these two sets of stylized facts, asking whether changes in agricultural employment, productivity, and crop mix can explain changes in mobility or in the relationship between income and mobility.

I find that the income-mobility relationship is remarkably robust, and that gains in productivity and reductions in agricultural employment do relatively little to explain mobility trends, but that cotton production is associated with less mobility and a weaker income effect, consistent with evidence from the economic history literature pointing to cotton economies engendering institutional forms that depress labor market dynamism and human capital accumulation. I also find that black and white workers, as well as agricultural and non-agricultural workers, saw important differences in mobility gains from changes in the agricultural system during this period.

ABSTRACT. Does access to food stamps influence how much low-income households rely on family and friends for financial help? As eligibility rules for food stamps (SNAP) are once again being debated, the relationship between SNAP and informal networks is important to policymakers. Many low-income households do not have robust informal networks for financial help, which means that a gap in SNAP access and a gap in private aid could lead to extreme difficulty for such households and exacerbate existing wealth inequality. Even for households with strong informal networks, lack of access to SNAP could stress both the households receiving aid and individuals providing aid. Thus, knowing whether there is a causal link between these two types of aid is an important concern for policymakers and an important part of understanding low-income households' livelihood strategies. Using the Panel Study of Income Dynamics from 1996-2017, I examine the causal relationship between SNAP eligibility and usage of private aid from family and friends. I exploit a change in SNAP categorical eligibility guidelines in November 2000 in order to create a difference-in-difference model to test how access to food stamps affects private aid usage.

ABSTRACT. The middle class in the United States has been falling behind, while the number of people at high-income levels and low-income levels has been growing, contributing to increasing inequality and the hollowing out of the middle class. Additionally, places with higher percentages of the population with less than a high-school degree and places with a higher percentage of working-age (between 25 and 54) males have lower levels of labor force participation – the very demographics that have historically worked in working-class occupations. Economic development incentives are the primary policy tool for promoting local economic growth and, presumably, providing opportunities for work. However, incentives may have unintended consequences and heterogeneous effects across industries. Given the recent trends in incentives and labor force participation, this raises a question as to whether incentives are helping minimize the hollowing out or contributing to its increase. We explore the impact of incentives on the missing middle by exploiting the industry-level detail on incentives across locations from the Panel Database on Incentives and Taxes (PDIT) combined with county-level industry employment and establishment data from the Wholedata Establishment and Employment Database. We exploit the panel nature of the data and the relative intensity of incentives across industries to understand how incentives may be either or attenuating or contributing to the disappearing middle and how this may vary across space.

ABSTRACT. The struggle among low- and moderate-income households to secure affordable housing in this country expanded to historically unprecedented levels in recent years. Record foreclosures, depressed wages and constrained credit triggered a 10-year decline in homeownership rates and rapid growth in renter households and rental cost-burden. Economic and demographic dimensions of the crisis have been studied at the national level, but not for rural areas. This research helps close this knowledge gap by analyzing changes from 2005 to 2017 in rentership rates and the number of cost-burdened renter households living in rural counties.

For individuals and families, housing cost-burden is a growing impediment to upward mobility. For rural communities, unmet housing needs are a drag on future employment growth and fiscal health. Rental cost-burden is increasing at a time when many rural communities are at risk of losing federally-subsidized rental properties. This study provides with knowledge about how this continuing crisis affects rural people and places, and how it has played out differently across the country.

We use American Community Survey (ACS) annual data from 2005 to 2017 and 5-year average data from 2005-2009 and 2013-2017 to study the rise in rentership and rental cost-burden. The rent-to-income ratio approach for defining cost-burden is commonly used in research and the 30 percent threshold has long been the defining metric for identifying cost-burdened households in both research and federal housing programs.

We identify county-level characteristics associated with the change in rental cost-burden between 2005-09 and 2013-17, using OLS and generalized linear regressions with logit link for response variable. Explanatory variables include demographic, geographic, housing-market, and labor-market characteristics. Preliminary results indicate that income inequality, poverty, a share of elderly population, a share of female headed households, and high occupancy rate have significant associations with increases in rental cost-burden among rural households. Further analysis include urban-rural differential investigation and rural rental cost-burden Blinder-Oaxaca decomposition. Taking these findings into account may help increase the effectiveness of federal housing assistance programs, especially programs designed to increase the availability of affordable rental housing.