Socioeconomic Transition in the Appalachia Coal Region Some Factors of Success Research commissioned by the World Bank under the Global Support to Coal Regions in Transition (P171194) Linda Lobaoa, Mark Partridgea, Oudom Heana, Paige Kellya, Seung-hun Chunga, and Elizabeth Ruppert Bulmerb a The Ohio State University b World Bank The Ohio State University Columbus OH 43212 Elizabeth Ruppert Bulmer Jobs Group, World Bank Socioeconomic Transition in the Appalachia Coal Region Some Factors of Success Table of Contents Acknowledgements . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6 Executive Summary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9 Section 1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 14 Section 2 Factors Related to Prosperity and Poverty Across Communities: A Synthesis of Related Research . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 18 General Factors Determining Prosperity and Poverty across Communities . . . . . . . . . . . . . . . . . . . . . . . 19 Communities Experiencing Natural Resource and Other Transitions: Targeted Literature Review Detailing Factors Associated with Well-Being . . . . . . . . . . . . . . . . . . . . . . . . 20 Section 3 U.S. Coal Mining: Past Trends, Future prospects . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 24 Overview of Coal Mining in the U.S. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 25 Shifting Geography of U.S. Coal Production . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 28 Long-term U.S. Coal-Mining Employment Trends and Productivity Growth . . . . . . . . . . . . . . . . . . . . . . . 30 Coal within Appalachia: Time and Spatial Trends . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 32 Section 4 Identifying Successful Post-transition Counties . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 36 Selection Criteria and Methodology . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 37 Robustness Check with Other Measures of Success . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 42 Section 5 Factors Underlying Successful Coal Transition . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 45 Human Capital . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 46 Sectoral Composition and Diversification . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 46 Population Demographics . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 48 Geographic Characteristics and Agglomeration Economies . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 48 Social Capital . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 51 Government Capacity . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 52 Health Outcomes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 52 Socioeconomic Transition in the Appalachia Coal Region Some Factors of Success 6 Section 6 A Closer Look at Successful Transition: County Case-studies . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 54 Athens County, Ohio . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 55 Noble County, Ohio . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 56 Laurel County, Kentucky . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 58 Sequatchie County, Tennessee . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 60 Ouray County, Colorado . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 61 Regional Economic Benefits of the Appalachian Development Highway System . . . . . . . . . . . . . . . . . . 62 Section 7 Conclusions and Policy Implications . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 64 References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 69 Appendix A Targeted Literature Review and Findings . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 76 Methodology: Selection of Studies . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 77 Classification of Studies . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 77 Overview of Table 1 Sections . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 79 List of Studies Reviewed . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 105 Appendix B Appalachian and U.S. Coal Industry Figures . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 107 Appendix C Methodologies Used in the Literature to Understand the Economic Impact of Coal Mining . . . . . 118 Appendix D Regression Model Description and Results . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 123 Appendix E Further Details on “successful” County Selection Robustness Analysis and Data . . . . . . . . . . . . . . . . . 128 Figures Figure 1: Appalachian Development Highway System Maps . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 149 Figure 2: Appalachian Development Highway System in Kentucky . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 150 Figure 3: Appalachian Development Highway System in Tennessee . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 151 7 Acknowledgments This study was carried out by Linda Lobao, Mark Partridge, Oudom Hean, Paige Kelly, and Seung-hun Chung, researchers at the Ohio State University, and Elizabeth Ruppert Bulmer, a Lead Economist in the World Bank’s Jobs Group. The work was guided by Michael McCormick (Climate Change Carbon Markets Group, World Bank), and benefited from support by Rachel Perks (Infrastructure, Energy and Extractive Industry Group, World Bank) and Aldo Mori (Jobs Group, World Bank). Michael Stanley and Christopher Sheldon (Infrastructure, Energy and Extractive Industry Group, World Bank) and Rohit Khanna (Infrastructure and Energy Global Programs, World Bank) provided strategic direction to the team. Luc Christiaensen (Jobs Group, World Bank), Joana Mclean Masic (Urban, DRM, Resilience and Land Global Practice, World Bank) and Harikumar Gadde (Climate Change Group, World Bank) provided helpful comments, for which the team is grateful. The World Bank commissioned this research under the Global Support to Coal Regions in Transition (P171194), with financial support from the Energy Sector Management Assistance Program (ESMAP). Socioeconomic Transition in the Appalachia Coal Region Some Factors of Success 8 Executive Summary From the early 19th century, Appalachia was the primary U.S. coal producer, but over the course of the last century, its coal industry evolved and ultimately faded from market dominance. This transition took a very long time, and spanned periods of boom and bust, new mine openings and mine closures, as well as wide-ranging economic development unrelated to coal, all of which shaped coal sector employment in the region’s coal communities. During the past 70 years, coal sector performance exhibited significant heterogeneity in the timing of coal-related business cycles and in spatial concentration. The 1950s and most of the 1960s saw the closure of many marginally-productive mines on the outer fringes of Appalachia, squeezed out by technology-aided competitors. Many Appalachian counties experienced a surge in coal employment during the 1970s, followed by mine closures during the “bust” years of the 1980s and during the 1990s with the tightening of environmental regulations. 9 Appalachia’s reign as the principal coal communities facing transition challenges. region of the U.S. has considerably waned The criteria we use to define “successful” in recent decades. The coal economy-based transition are (a) a near total phase-out of coal Appalachian region has historically been jobs from a previous level of dependence, and poorer than much of the U.S. For decades, it (b) a shift to alternative economic activities had some of the lowest per-capita income sufficient to sustain a growing population with levels and highest poverty rates in the country rising household incomes and thus improving (Lobao et al. 2016). When coal production economic well-being. began to shift west in the late 1960s, Appalachia’s mines began to struggle. By 1998, We identify the main variables associated Appalachia’s share of U.S. coal production had with successful economic transition, fallen by more than half to 41 percent, and based on evidence from a wide-ranging continued to decline, reaching 27 percent in literature review. We examine the large 2018 (EIA 2019). Meanwhile, its national share body of literature on the impact of resource- of coal employment declined from 85 percent dependence on economic development in in 1954 to 57 percent in 2018. Appalachia and elsewhere, the economics of boom and bust cycles, challenges related to the The declining fortunes of the coal industry so-called “natural resource curse” and lagging has exposed Appalachian communities to regions. We reach beyond the traditional severe negative economic shocks. Despite economics literature to include geography, declining share, Appalachia’s coal-country demography, and sociology. We measure remains quite vulnerable to the fortunes of performance on key variables across the 420 coal mining. During the coal industry’s rise, ARC counties and rank the best performing communities across Appalachia became counties. Four counties emerge as “successful” dependent on the coal economy and coal under our criteria: Athens and Noble Counties employment, whether in coal mining or in in Ohio; Laurel County, Kentucky; and Sequatchie associated coal supply chains. The severe and County, Tennessee. ongoing decline of the sector has resulted in widespread economic dislocation, requiring We check the validity of our initial results communities to adjust to new market realities. by examining the performance of these The extent and timing of the adjustment counties along other metrics of community varies from one county to the next, as do well-being, such as population growth rate, the resulting economic outcomes. The mining and coal employment shares, median Appalachian Regional Commission (ARC), household income. We also compare their which operates in 13 states and 420 counties, performance to average performance of three was established in 1965 to address these issues. sub-groups of ARC counties: those with the top-10 fastest growing populations, the 10 This study identifies Appalachian counties counties that comprise median-growers, and that have successfully transitioned from the 10 counties that are the slowest growers. dependence on coal while sustaining growth, In a second check of robustness, we assess and assesses factors that facilitate more additional factors deemed in the literature to successful community transition. The key contribute to local economic development: goal is to determine why some communities agglomeration economies, human capital, perform better to draw lessons for other economic diversification, population age Socioeconomic Transition in the Appalachia Coal Region Some Factors of Success 10 structure, proximity to metropolitan areas, relative transition success. These “deep- natural amenities, extent of urbanization, dive” assessments focus especially on the topography, inequality, unemployment, role of distance to larger urban centers, road minority population share, social capital, infrastructure, local non-mining economic government capacity, and health outcomes. We activity, local government institutional assess the extent to which these factors may capacity, and local social capital networks. have contributed to successful transition for each of the four counties. The following conclusions emerge from our combined analysis: Finally, we complement our quantitative findings with case-studies on each • Very few Appalachian counties have “successful” Appalachian county (and one managed a positive transition from coal non-Appalachian county, for comparison). dependence. Of 222 ARC counties with a Our case study approach combines qualitative high level of coal activity at some the period methods (key informant interviews), with after 1950, only four counties managed secondary sources to shed light on the to transition out of coal and remain opportunities and resources within each economically viable communities with community that likely contributed to sustained population growth. 11 • In terms of economic well-being, the level factors, making it difficult to generalize of success of the four counties is modest. approaches for other counties. Athens While the four counties have grown in County’s economic development has centered population and diversified their production, around its large public Ohio University and most have experienced significant which supports direct and indirect jobs and poverty reduction, average household generates local social capital. Noble County incomes remain low and poverty rates was able to attract a large public investment exceed national and ARC averages. to build a state prison, which has served as an economic driver. While infrastructure • Severe economic structural impediments investment is a common theme in our across Appalachia constrain growth. Being successful counties, this is not sufficient small and remote, most ARC counties have to guarantee successful transition from limited access to labor markets with more coal as much of Appalachia has received and diverse job opportunities. ARC counties significant investment, at least with have low levels of physical capital, especially respect to road infrastructure. That said, infrastructure, and high transportation improved roads helped our four successful costs. Human capital is also low, with lower counties by increasing connectivity to larger educational attainment and lower quality metropolitan areas, manufacturing chains, education and health services. and to living, tourism, and recreational opportunities. Laurel County became a • Non-structural impediments reinforce regional hub following investment to poor economic outcomes and reduce construct two major highways—including local economic resilience. Historically Interstate-75, which linked the area to coal-dependent communities exhibit northern manufacturing centers—as well less economic diversification, modest as a regional airport, a hydroelectric power manufacturing activity, problematic dam, piped water supply, and industrial patterns of “boom and bust” cycles, and low parks. Sequatchie County benefited from levels of entrepreneurship. investments in highways to access nearby Chattanooga, a large metropolitan market • Institutional capacity and social capital offering diverse job opportunities. have helped some counties transition more successfully. Local government institutional Our ability to draw specific policy lessons capacity and social capital are generally based on the economic development low across the ARC region compared to patterns in Appalachia is limited. This national averages. Our case studies highlight would require assessing general and coal- examples where local government capacity specific policies at the county, regional, state, to design, finance, and implement economic and federal levels, as well as employers’ and development initiatives in collaboration with workers’ incentives in the region, and fiscal local civil society appears to have helped and public investment stances and capacity of sustain transition impetus. particular counties or states. This is beyond the scope of our research. • Transition paths in our four “successful” Appalachian counties each have their Nonetheless, our analysis highlights some own unique features and success broad policy areas for addressing common Socioeconomic Transition in the Appalachia Coal Region Some Factors of Success 12 economic development impediments in includes designing and enforcing the right Appalachia: mix of rules and regulations and providing public goods consistent with economic • Enhance connectivity: Infrastructure development, but also coordinating with investments alone are not enough to ensure local stakeholders and regional agencies to successful community transition from coal stimulate private sector activities, such as dependence. But remote communities need through investment incentives or promoting connectivity—whether in the form of roads public-private partnerships (PPP) in priority or digital connections—to larger markets to sectors. grow, achieve scale economies, match excess labor with nearby job markets, and even to • Coordinate economic development connect people to recreation and tourism strategies: Counties that are economically opportunities. constrained need a coordinated set of economic development strategies and • Invest in human capital: Enhancing human approaches to (a) foster larger economic capital through investments in education agglomeration and/or linkages to larger and health will improve residents’ well- regional/national/global markets, and (b) being and help raise workers’ productivity, exploit natural amenities in a sustainable allowing them to compete for better, higher- manner that can attract demand for local skilled jobs and generate more added value. services. Not all communities are equally In the absence of employer labor demand, endowed, so policymakers will have to however, human capital investments may weigh tradeoffs between propping up very have negligible returns and risk frustrating small communities and investing in areas workers with heightened expectations for with higher potential. This will require jobs that are not locally available. Successful deep analysis of communities’ long- investments in education and training align term sustainability, and consideration of curricula to identified private sector needs alternative approaches, such as programs to and/or focus on building entrepreneurial facilitate out-migration. capacity. • Seek economic diversification to ease “boom and bust” cycles: Coal-dependent communities must move beyond volatile coal “boom-bust” cycles associated with long- term economic and social costs. Government and civil society must facilitate new economic activities and attract investments in new, job-creating firms that serve local or regional markets or beyond. • Build local institutional capacity: Diversifying sectors of economic activity requires institutional capacity to develop a suitable business environment. This 13 SECTION 1 Introduction Economic transition away from dependence on coal mining can be difficult and costly, but can yield significant medium- term gains. As nations move away from coal production and coal-based energy generation, the transition creates short-term economic disruption to coal communities, notably through job losses and severe economic recession. In the medium term, however, transition can generate “winners”, both locally and in other regions. Displaced coal sector workers may find jobs in more sustainable and more productive industries. Environmental degradation from mining activities can give way to restoration and conversion of natural resource assets. And both coal and noncoal regions can benefit from reduced pollution and healthier people, directly improving human capital. But local economic downturns can also persist to the point that economic decline threatens a community’s viability. The likelihood of this more pessimistic outcome increases in coal regions that are already lagging their non-coal counterparts in terms of economic well-being. Socioeconomic Transition in the Appalachia Coal Region Some Factors of Success 14 This paper examines the transition away Appalachia’s reign as the dominant coal from coal mining in the Appalachia region producer in terms of its national share of of the United States and the impact on local production began to wane in the late 1960s as communities; the aim is to identify factors coal production began to shift west. By 1998, that helped some communities transition Appalachia’s share of U.S. coal production more successfully than others. The analysis was 41 percent and fell to 27 percent in 2018 looks beyond traditional economic factors, (U.S. Energy Information Agency, EIA, 2019).2 and takes into account broader social While total U.S. coal production fell 6 percent and institutional aspects germane to the over this 20-year period, Appalachian coal community capitals literature. Whether driven production fell by 41 percent (EIA 2019). by new market realities or the imperative to These changes, along with the legacy of mitigate climate-change, coal communities poverty in Appalachia, make its coal-country in other countries will need to transition communities particularly vulnerable. to alternative economic activities, and can benefit from the experiences of coal regions in Despite a significant body of research, Appalachia or elsewhere. certain knowledge gaps remain regarding the patterns of local economic transition The Appalachia region has been significantly in Appalachia and why some communities affected by coal sector decline dating back transition more successfully than others: more than a half-century, but intensifying over the past several decades. It is a mostly • Most research looks at national rather rural region spanning the Appalachian than community-level data and outcomes. Mountain range, nearby foothills, and While numerous studies address changes in contiguous areas, such as the Shenandoah the U.S. energy industry, national studies tell Valley. The region has historically lagged in us little about how local populations adapt in terms of socioeconomic development, and affected areas. its reliance on coal likely slowed the pace of long-term growth. The Appalachian Regional • Most community studies focus on mining Commission (ARC) was established in 1965 to as an aggregate sector and/or other address these challenges. Today, ARC covers types of resource extraction. Studies parts of 13 states and includes 420 counties.1 rarely disaggregate coal mining in terms of community effects (Betz et al. 2015; Appalachia has some of the lowest per capita Lobao et al. 2016). For example, researchers income levels and highest poverty rates of examined general boom/bust cycles in the any U.S. region (Lobao, Partridge, Zhou, and energy industry and the “natural resources Betz 2016). Appalachia historically has been curse” under which natural resource- the U.S.’s primary coal producing region since intensive locations have lower long-run the early 19th century, typically accounting growth rates when averaged over boom and for four-fifths of national coal production. bust cycles (Van der Ploeg 2011). Research 1 Though we will use the terms Appalachian region and ARC interchangeably, the ARC has expanded over time and covers more  territory than traditional Appalachia. 2  EIA will refer to the U.S. Department of Energy’s Energy Information Agency’s Annual Energy Outlook. The annual outlook is one of the most respected sources of energy data and forecasts in the world. We cite the Annual Energy Outlook so often that we will simply refer to the associated citations as “EIA” and the year of the outlook. 15 on employment in the U.S. energy industry natural resources curse (Deaton and Niman devotes relatively more attention to the oil 2012). There are some exceptions, including and gas industry, which rapidly expanded Lobao et al. (2016) and Betz et al. (2015), who after 2006. analyzed changes in coal employment and effects on poverty rates, household incomes, • Existing research tends to use a case-study and other variables from 1990 to 2010. approach focused on specific communities, which cannot be easily generalized. Few • Factors that might lead to positive local systematic, quantitative studies exist outcomes in the wake of coal decline are in part because researchers have lacked not well-studied. Betz et al. (2015) note that detailed employment data that span small coal mining employment could influence communities. a wide range of community attributes, including local entrepreneurship and • Recent quantitative research into the coal employment in retail and accommodation industry’s effects on communities is rare. (for example, tourist industries), which Studies have examined the boom/bust cycles affect the degree to which communities of the 1970s and 1980s (for example, Black could transition to other economic sectors. et al. 2005) or the long-rn 20th century They find that Appalachian communities Socioeconomic Transition in the Appalachia Coal Region Some Factors of Success 16 with a larger share of coal employment had 99 coal-dependent counties. We complement lower entrepreneurship rates. Historical this quantitative approach with qualitative case studies and case studies of the region studies, and draw generalizable lessons about (Billings and Blee 2000, Duncan 2014, community adjustment and policy implications. and others) also explain how coal mining affected community development, including We focus on three analytical questions by weakening local government and critical for understanding the transition institutional capacity to address residents’ process away from dependence on a single needs. These studies point to the importance industry and for identifying the challenges of considering a broad range of community and barriers to community socioeconomic factors that might influence future revitalization. socioeconomic outcomes. • Which communities made a relatively • Limited temporal focus represents another more successful transition away from coal gap in existing literature. A majority of mining employment? studies examining community-level impacts (positive or negative) on coal or other mining • What were the socioeconomic outcomes in employment focused on a single or specified these communities? point(s) in time, rarely addressing the long- term consequences of transitioning from • What factors mattered for successful coal (with the exception of case studies). To transition? our knowledge, no studies have determined the time period needed to make successful This report is organized as follows. adjustment. For Appalachia, existing The next section reviews and synthesizes research has not established how well the literature on community well-being communities have adapted to this shift from during periods of transition. The third section coal since World War II. describes historical trends in coal industry production and employment. The fourth section This paper aims to address these knowledge presents our quantitative methodology for gaps by deepening understanding of identifying communities that transitioned factors that contribute to communities’ beyond coal, and identifies four counties that ability to transition successfully from emerge as “successful”. The fifth section reliance on coal, and offers some broad documents factors explaining why some policy implications. Drawing from the wider Appalachian communities have fared better social sciences literature on factors that than others, based on both quantitative and contribute to community well-being and qualitative analyses. The sixth section more successful transition, we empirically test the specifically describes the underlying economic relevance of these factors for Appalachia using development climate for each of five case-study statistical models that exploit community- counties, including the role of highways. The level differences in socioeconomic outcomes concluding seventh section synthesizes the from the 1950 post-War period to the present. findings to outline lessons learned and general This empirical approach enables us to track policy implications. Appalachian communities over the long term and document the degree of recovery across 17 SECTION 2 Factors Related to Prosperity and Poverty Across Communities: a Synthesis of Related Research This section synthesizes a wide range of studies to gauge factors that are important in driving community recovery in the wake of coal transition. We summarize the conclusions of this large and disparate literature by dividing it into: (a) studies on general factors that affect prosperity and poverty across U.S. communities, and (b) targeted case studies that examine communities experiencing natural resource, energy, and other industry transitions. Socioeconomic Transition in the Appalachia Coal Region Some Factors of Success 18 General Factors Determining c. Sociodemographic factors, such as race/ ethnicity, age, education, gender, and family Prosperity and Poverty structure that reflect residents’ structural vulnerability. Across Communities Community socioeconomic well-being (lower To determine factors that matter for poverty and higher income) historically has community recovery, it is useful to draw been greater in metropolitan counties, rural first from the large social science literature counties closer to metropolitan centers, and on why U.S. communities vary in terms places outside the U.S. south (Blank 2005; of prosperity and poverty. These studies Partridge and Rickman 2006). Socioeconomic identify persistent structural conditions that well-being is also associated with better- either hinder or facilitate socioeconomic well- quality employment sectors, such as higher- being, as measured by outcome indicators wage service industries (producer services) such as poverty rates, population growth, and, at least in the past, higher manufacturing income levels, inequality, job growth, and employment (Cotter 2002; Lobao and Hooks unemployment. Numerous and wide-ranging 2003; Moretti 2012; Moller, Alderson, and studies from economics (Blank 2005; Partridge Nielsen 2009). Communities with a smaller and Rickman 2006; Weber et al. 2005), sociology share of structurally vulnerable populations— (Brown and Schaftt 2011; Lobao 2004), such as those with higher educational geography (Glasmeier 2002), and regional attainment; a higher share of working-age science (Isserman, Feser, and Warren 2009) population; fewer single-parent, female- look at determinants of U.S. communities’ headed households; and lower racial/ethnic socioeconomic well-being. Much of this work minority population—generally have lower specifically addresses structural conditions poverty and higher incomes (Partridge and in rural America, making it more relevant for Rickman 2006; Lichter and Cimbaluk 2012; Voss analyzing Appalachia (See Appendix A, Table 1 et al. 2006). for a complete list of studies reviewed). Local institutional factors, such as local This research helps us identify potential governmental capacity and social capital, key structural factors that enable some can also promote community prosperity communities to fare better than others when (Blank 2005; Weber et al. 2005). While few experiencing similar declines in coal mining generalizable studies have empirically tested employment. Studies commonly recognize these relationships, they provide some evidence three sets of structural determinants of that county government administrative community well-being: capacity (Lobao et al. 2012) and local social capital (Isserman et al. 2009) are associated a. Geographic attributes, such as regional with lower poverty and greater prosperity. location, urban-rural location, distance from metropolitan areas of varying size, and The empirical analysis presented below degree of urbanization including population. assesses these potential correlates of successful transition. Based on the large b. Local economic structure, or the quantity, body of work on prosperity and poverty across quality, and mix of local employment sectors. communities, we test the hypothesis that 19 coal communities characterized by more commodity prices, and underinvestment in favorable structural conditions—geography, education (Partridge et al. 2013). In sociology, economic structure, and a smaller vulnerable Freudenburg’s (1992) classic research explains population—along with greater government similar processes that jeopardize community capacity, are likely to fare better when well-being. He notes that communities tend transitioning from coal. to become “addicted” or over-adapted to extractive industries. Peoples’ expectations and community institutions revolve Communities Experiencing Natural around the industry, making inevitable busts particularly devastating. Extractive Resource and Other Transitions: communities become over-specialized, and periodic shutdowns increase unemployment Targeted Literature Review Detailing (Freudenberg and Wilson 2002). Finally, the literature describes the problem of “path Factors Associated with Well-Being dependence”, whereby extractive industries give rise to a self-reinforcing development We draw more detailed evidence on path in communities, which displaces determinants of success and revitalization other industries, fosters less diverse local from the literature on natural resource and economies, and leads to underinvestment other transitions. While this literature also in education and human capital. The remote covers the four broad categories mentioned rural location of many extractive communities above, it concludes that natural resource- exacerbates these factors. based communities such as in Appalachia are even more constrained with regards to We narrowed our focus to review literature geographic location, economic structure, on transitioning communities that share population vulnerability, and institutional similar characteristics to Appalachian capacity. coal communities. Our aim was to identify factors that facilitate or hinder transition and Overlapping social science frameworks recovery in Appalachia. This encompasses stress that resource-dependent regions communities facing economic shocks— historically tend to fare poorer. Experts have including those stemming from energy/ often noted some key negative factors present natural resource transitions related to oil in many extractive industries: a history of and gas production—and support activities exploitative relationships, distinct phases of for coal and oil and gas industries, including development, non-local market structure, electric power generation, petroleum and susceptibility to boom-and-bust shocks refining, and natural gas distribution. We (Freudenburg and Wilson 2002). also considered studies addressing changes in small, rural communities, such as military Economists have often pointed to the base closures. Finally, we include previous “natural resource curse”. Natural resource- studies on changes in coal mining and its intensive settings appear to have lower long- effects on communities. The studies we term growth when averaging over the boom- analyzed—which comprise academic journal bust cycle, explained by lack of alternative articles, as well as government and non- labor market opportunities, volatility in governmental agency reports—vary by Socioeconomic Transition in the Appalachia Coal Region Some Factors of Success 20 methodology, thematic areas of focus, and diversification or alternative industries is impact indicators (see Appendix A).3 cyclical, and depends on commodity price movements. We draw some generalizations based on review of 37 studies.4 • The region where transition from mining occurs seems to matter. For resource • Mining and coal-dominated communities extraction, communities in Southern U.S. fair worse on welfare outcomes. Large- states appear to fare worse and Western scale quantitative studies examining states better (Freudenburg and Wilson 2002; changes in the mining sector across the U.S. Lobao et al. 2016; Stedman et al. 2012). This tend to find statistically significant negative finding could be due to relatively worse initial impacts on community poverty, incomes, structural conditions in Southern states, such employment, population growth, and other as poorer quality jobs or lower education.5 measures of well-being (Betz et al. 2015; Black et al. 2005; Cook 1995; Douglas and • Community structural characteristics Walker 2017; Freudenburg and Wilson 2002; matter for adaptation. Three key sets of Lobao et al. 2016). This finding might seem barriers across coal and mining communities obvious, but these impacts do not necessarily exist: (a) geography and degree of remoteness hold for employment in other industries such from cities (Douglas and Walkers 2017; as service sectors. Haggerty 2019; Haggerty et al. 2018; Snyder 2018); (b) availability of alternative economic • The timing of transition matters. Mining opportunities (Carley 2018; Deaton and Niman effects on community wellbeing tend to 2012; Haggerty 2014; Haggerty et al. 2018); be positive in times of price upswings and and (c) population vulnerability, reflected negative when prices are low (Freudenburg by low educational attainment (Douglas and Wilson 2002). This also holds for coal and Walker 2017; Haggerty et al. 2018) or by mining (Betz et al. 2015; Black et al. 2005; an aging workforce (Haggerty 2019). These Lobao et al. 2016). The uncertainty makes it structural barriers limit workforce upgrading difficult for communities to adjust for the through various channels; for example, “natural resource curse.” Haggerty (2014) low digital literacy limits job search and finds that the longer counties specialized in occupational mobility; and transportation is oil/gas (over a 30-year period), the lower the important for accessing college and training education and income levels. The boom/bust centers, as well as employment elsewhere cycle has implications for local support for (Jolley et al. 2019). Training, community extractive industries. As noted, communities college programs, and new professional over-adapt to extractive industries opportunities, including those in alternative (Freudenburg 1992), with residents assuming energy industries, can improve residents’ busts are temporary and that conditions will ability to adapt to change (Carley et al. 2018). return to “prosperity”. Local receptivity to 3 Note that we limit our geographic coverage to the United States.  4 Note that although the selected literature represents substantive research on the topic, it is not exhaustive of all past work. 5 It could also be due to regional variations in employment quality and other regional features of the extractive sector, but the studies  reviewed do not provide systematic comparisons within that sector. 21 • The quality of the local environment and promising return of coal jobs discourages presence of natural amenities matter. community and individual efforts to adapt Local environmental degradation hampers to change. Long-term dependence on coal future development and the ability to attract delays acceptance of transition, but in any tourism and other non-extractive industries case, populations find it difficult to move (Appalachian Law Center 2019; Haggerty et elsewhere (Haggerty 2019). al. 2018; Kelsey et al. 2016). Places with higher quality of life—including natural amenities • Rural communities are not homogeneous, as reflected in climate, topography, and and the benefits/costs of transition vary water area—are more likely to attract by social group. For example, Appalachian migrants, especially retirees (Isserman et al. communities with a greater share of coal 2009; Partridge and Olfert 2011). employment tend to have a lower share of sole proprietors, higher disability rates, • The transitioning populations are aware of and a higher share of poor people (Betz et the challenges their communities face. The al. 2015). Much has been written about the barriers acknowledged by locals are similar uneven impacts of natural gas expansion, to those indicators cited in quantitative with communities divided among those who studies. Research based on opinion surveys benefit and those who do not. Extractive (Besser et al. 2008; Graffe 2019) and focus industries employ a higher proportion groups (Carley et al. 2018) point to barriers of men, and women tend to have fewer such as lack of alternative employment, local employment opportunities. Declines low education, and boom/bust economic in extractive employment affect family cycles. In Appalachia, residents’ concerns structure, and may increase the share of about their community shifting to non-coal female-headed households (Cook 1995). employment include fear of potential job loss and business closures, detrimental effects • Local “social capital” is an important on schools and retail as families leave, lack factor for transition. Communities with of affordable housing elsewhere and high higher levels of social capital—strong inter- attachment to the community (mobility group relationships within the community— barriers) (Carley et al. 2018). tend to be more resilient. In the case of plant closures, Besser et al. (2008) find that • The prevalence of a “coal culture” can residents report less negative overall quality impede transition in Appalachia. Carley et of life where social capital is higher. al. (2018) note that coal mining generational employment has fostered a community bond • Low capacity of local governments and identity with the industry. Haggerty et represents an institutional barrier to al. (2018) point out that residents’ resistance transition. Lack of local government to change can arise when populations blame capacity in rural and small U.S. communities restrictive environmental regulations, has long been noted (Johnson et al. 1995; while they ignore the larger role of markets Lobao and Kelly 2020).6 The overriding and price competition from, for example, problem is replacing and stabilizing income natural gas. Carley et al. (2018) note that streams (Haggerty et al. 2018). Haggerty 6 This has been documented in studies of transition planning, but there are few empirical analyses of actual transitions.  Socioeconomic Transition in the Appalachia Coal Region Some Factors of Success 22 (2019) summarizes the barriers small or provides economic development grants and rural governments face when coal plants technical support to assist communities in close, including limited administrative redevelopment. Cowan (2012) notes that while leadership capacity (for example, little or rural communities typically find it harder to no planning staff); limited staff and ties to recover from a base closure than urban ones, state or regional actors, which limits ability the differences are not very large. to apply for federal and state assistance; and low fiscal autonomy, which limits • Federal policy addressing the needs of local budgeting authority. Haggerty et al. rural communities has been limited. The (2018) and Haggerty (2019) stress that small U.S. has proxied agricultural-policy for rural communities often lack access to a dedicated development, focusing on the needs of the transition fund, requiring them to substitute farm population as opposed to the broad other funds or secure external funding. They rural nonfarm population. In place of general lack adequate information to assess fiscal rural economic development, including risks and state and federal fiscal policies infrastructure and facilities, rural areas rely (Haggerty 2019). In analyzing the impacts of more on social protection transfers targeted coal-power plant closures in an Appalachian to individuals, such as disability assistance county, Jolley et al. (2019) draw similar (Bishop 2017). Haggerty et al. (2019) note conclusions, noting that local governments that in the case of coal plant closure federal have limited fiscal resilience to recover from assistance is over-prescribed, poorly targeted, lost tax revenue. and ultimately limits local flexibility to use funds to meet local needs. Various programs • Conventional economic development have offered financial resources to transition policies to attract business investment, areas. For example, the Appalachian Regional retain local businesses, and develop the Commission (ARC) has long provided grants workforce have variable effects. Benefits and other programmatic assistance. The from economic development policies federal Partnerships for Opportunity and typically appear to be modest (Daniels et Workforce and Economic Revitalization al. 2000), and strategies tend to be overly (POWER) initiative, established in 2015, focused on retaining or attracting a single offers a tool-kit of community development large employer (Haggerty et al. 2018). strategies to provide workforce training No policies/programs work everywhere, and economic development assistance in and successful models appear difficult to Appalachian coal communities. POWER aims replicate. Outcomes have been difficult to to increase regional economic diversification, evaluate and there are ambiguous results job creation, capital investment, and re- across studies. employment for displaced workers. As of fall 2019, it had provided $190 million in grants • Military base closures, while similarly spanning 239 projects in 326 ARC counties. affecting small communities, differ from However, Morris et al. (2019) note that the transitions from extractive industries. wider goal of investing $9 billion in total Military base closures follow the Federal federal aid has not been realized. Base Closure Act-mandated rigorous planning protocol for engaging local communities. The Department of Defense 23 SECTION 3 U.S. Coal Mining: Past Trends, Future Prospects Socioeconomic Transition in the Appalachia Coal Region Some Factors of Success 24 Overview of Coal Mining in the U.S. Appalachia has long been the center of America’s coal industry. Appalachian coal The 1919 map of coal fields (Figure 3.1)7 mining began in Pennsylvania in the early identifies significant coal deposits in several 19th century. By 1954, coal employment was continental U.S. regions. The large swath of heavily concentrated in the region where Ohio, Appalachian coal fields runs from northern Pennsylvania, and West Virginia meet, an Pennsylvania through central Appalachia, to area of rich bituminous coal fields (Figure 3.2). its southern boundary around Birmingham, Other dominant regions for coal production Alabama. The key competing regions to and coal employment were northeastern Appalachian coal are the western Kentucky/ Pennsylvania (characterized by high- southern Illinois coal fields and, beginning quality anthracite coal fields used in ferrous in the 1970s, the Powder River Basin (PBR) metallurgy), and southeastern Kentucky/ centered in northeastern Wyoming. southwestern West Virginia. Note the small Figure 3.1 Coal Mining Fields in the Continental United States, 1919 Source: 1919 Census of Mineral Industries, p. 254 7 Appendix B Figure 1 depicts a contemporary perspective of coal fields. A careful look shows that the maps do not tangibly vary in  terms of the location of reserves, although contemporary analysis provides a more accurate view of remaining coal reserves and its quality in each region. 25 Figure 3.2 1954 Mining Employment Intensity (each dot = 100 mine workers) Source: 1954 Census of Mineral Industries, Vol 2. Pg. VI. cluster around Birmingham, AL, an area its employment share to rise to 85 percent by endowed with both iron ore and the necessary 1954. Appalachia’s dominance began to wane coal for smelting the ore, allowing the region after the mid-1960s as western coal began to become both a coal producer and the south’s its initial ascent. By 1972, Appalachia’s share only primary iron/steel producer. of coal mining employment fell to about 80 percent, and continued a steady downward In 1919, the Appalachian coal sector trend to reach 67 percent by 2002. Although accounted for two-thirds of total U.S. it maintained employment share in the 2000- mining employment, but is down to just 2012 period, the downward trend resumed over half today. As transportation costs fell, thereafter, reaching 57 percent in 2018 (Census Appalachia’s productivity advantages over of Mineral Industries, various years; EIA, other U.S. coal mining regions rose, leading various years). Socioeconomic Transition in the Appalachia Coal Region Some Factors of Success 26 Figure 3.3 U.S. Coal Employment and Production 1,200,000 1,000,000 800,000 600,000 400,000 200,000 – 29 39 82 54 63 72 12 17 92 02 19 20 20 19 19 19 19 19 19 19 20 19 US Co l Emplo m nt US Production (1,000 tons) Source: U.S. Census Bureau, Census of Mineral Industries. Various Years. Expanding coal production accompanied the downward trend as coal production expanded national decline in coal employment. Coal by over 250 percent by 2002. Production production increased modestly in the post- plateaued until about 2012 before declining by World War I era until 1929, before declining by about 30 percent by 2017. A number of factors almost one-third by 1954 (Figure 3.3), initially explains this most recent decline, including due to the Great Depression and then from rising competition in coal export markets, substitution by less expensive fuel sources, price competition from cheaper natural such as oil and gas. This decline was also gas, and increasing competitiveness from spurred by technological changes, such as the renewable energy sources. switch to diesel locomotives. Strong economic growth after 1954, combined with rapid coal- mining productivity growth, reversed the 27 Shifting Geography of cost advantages (EIA 2001). Today, 50 to 70 coal trains leave Wyoming every day, destined U.S. Coal Production to 29 states, the largest consumers being Texas, Missouri, and Illinois.11 The Appalachian region has lost its position as the locus of national coal production (Betz The shift in coal production away from et al. 2105; EIA 1999, 2005, 2013). The 1970s saw eastern regions was relatively swift. Eastern a massive increase in western coal production, coal, mostly Appalachian, accounted for especially from the Powder River Basin (PRB) 93 percent of total U.S. coal production in of Wyoming and southeastern Montana. 1965, but fell to 74 percent by 1978, while the Environmental regulations associated with the respective shares for western coal were 4 Clean Water Act of 1972, Clean Air Act of 1970, percent in 1965 and 25 percent in 1978 (EIA and Clean Air Act of 1990 increased demand 1979). In 1990, coal production east of the for low-sulfur, less polluting western coal.8 Mississippi River accounted for 61 percent In addition to being low in sulfur, PRB coal of the total (47 percent from Appalachia), is also relatively low in mercury and arsenic, compared to 39 percent produced west of the giving it additional competitive advantages Mississippi River (EIA, 1993). In addition to over bituminous coal, in light of tightened falling transport costs, increasingly stringent environmental regulations (EIA 1979, 2001).9 environmental regulations accelerated the PRB coal is also relatively inexpensive to shift westward. By 2000, western coal share strip mine with little surface overhang, very reached 52 percent of national production thick coal seams, and high-quality coal (EIA while Appalachia’s fell to 38 percent. 2001).10 By the 1970s, these lower production costs became a compelling reason to mine Innovations in unconventional shale oil western coal rather than developing other and natural gas drilling (“fracking”) led to energy sources (EIA 1979). Coal is bulky and explosive growth in oil and gas production expensive to transport, which historically after 2005. The low price of natural gas gave Appalachian coal mines an advantage motivated rapid and widespread substitution due to proximity to major eastern cities for coal. During the 1990–2008 period, coal and manufacturers. But improved rail consistently accounted for 48 to 53 percent of transportation from PRB mines beginning in U.S. electricity generation, but fell to 37 percent the 1970s gave further impetus to the region’s in 2012 and 24 percent in 2019. By contrast, 8 Although Appalachian coal has a higher heating value than PBR coal due to its lower moisture content of only about 7%, it also has  about one-and-a-half-times more ash (the toxic residue that remains after burning coal), about two- to five-times more sulphur, which causes acid rain and other air pollutants, and about ten-times more chlorine, which is associated with corroding power plant boilers (Hatt and Mann, 2015; National Academies of Sciences, Engineering, and Medicine, 2007). 9  Source: Carrol, Chris, Wyoming’s Coal Resources Summary Report, 2015. Available at: http://sales.wsgs.wyo.gov/wyomings-coal- resources-summary-report-2015/ (downloaded April 28, 2020). 10 The Wyoming State Geological Survey describes the production-cost advantages of the PRB coal as follows: “In the Powder River  Basin coal field—the most prolific in the world—coal is mined from two major coal seams, the Anderson and Canyon coals. This coal occurs in the Paleocene-age (65 to 55 million years ago) Tongue River Member of the Fort Union Formation. The mineable subbituminous coal seams in the Fort Union Formation are 60 to 80 feet thick, with a moisture content between 20 and 30 percent, and contain less than 6 percent ash and 0.5 percent sulfur. Powder River Basin (PRB) coal also includes beds in the Eocene-age Wasatch Formation, where exploration drilling has encountered coal seams greater than 200 feet thick.” (Source: Wyoming State Geological Survey. https://www.wsgs.wyo.gov/energy/coal-production-mining (downloaded April 30, 2020). 11 Wyoming State Geological Survey. https://www.wsgs.wyo.gov/energy/coal-production-mining. (downloaded April 30, 2020).  Socioeconomic Transition in the Appalachia Coal Region Some Factors of Success 28 natural gas’s share of electricity production 2011 and 2017, the capacity of U.S. coal-fired rose from 12 percent in 1990 to 30 percent in power plants fell nearly 20 percent, and more 2012, and 37 percent in 2019 (EIA 2013, 2020). closures are expected.12 Renewable energy is Appalachia itself has become a major natural also forecast to surpass coal as a source of U.S. gas supplier as northern Appalachian Marcellus electric generation in 2021.13 Shale has become a key source of supply. Expected climate-change regulations also have These combined effects—falling consumer increased demand for gas (and renewables) demand in the last decade and the 50- relative to coal, as CO2 emissions from coal year redistribution of U.S. coal production are about double those of natural gas. By 2019, west—have had devastating impacts on Appalachia’s share of U.S. coal production Appalachian coal country, leading to mine declined to about 26 percent (EIA 2020). closures and layoffs. Scores of U.S. coal companies have filed for bankruptcy in recent Finally, large numbers of older 1950 to 1980- years, including giants Peabody Energy, Cloud era coal power plants have been retired Peak Energy, Arch Coal, Murray Energy, and since the late 2000s recession. Between Alpha Natural Resources.14 12  See EIA (2019) Today in Energy. “U.S. coal plant retirements linked to plants with higher operating costs.” December 3, 2019 (downloaded from https://www.eia.gov/todayinenergy/detail.php?id=42155#. on May 3, 2020). 13 See EIA (2020) Today in Energy “EIA expects U.S. electricity generation from renewables to soon surpass nuclear and coal.”  (downloaded from https://www.eia.gov/todayinenergy/detail.php?id=42655#. on May 3, 2020). 14 See https://en.wikipedia.org/wiki/Coal_mining_in_the_United_States (downloaded April 30, 2020).  29 There is no relief in sight for coal’s near and coal mining moved, coal employment rapidly long-term outlook. During the COVID-19 declined, mainly driven by labor-saving pandemic-induced recession, electricity demand productivity growth (Appendix B Figure 2 has fallen, with coal bearing most of the burden presents annual coal employment for 1929 to (Wade et al. 2020). Coal’s share of U.S. electricity 2019).16 generation declined by over 5 percentage points between February and April, 2020 (Wade et Measured as tons of mined coal per miner, al. 2020). Natural gas and some renewables coal sector productivity rose drastically over represent less expensive sources of electricity time (Figure 3.4). Beginning in the late 1960s, generation than coal. The EIA forecasts a post- steady economic growth and rising energy COVID decline in U.S. coal production of 22 prices, due to supply constraints for oil and percent for 2020. There was a remarkable 41 natural gas, led to resurging coal demand. percent annual drop in U.S. production in the The ensuing search for new coal led to falling week ending May 9, 2020, with Appalachia taking productivity. The result was an 83 percent a disproportionate hit.15 According to the EIA’s increase in mining employment between pre-COVID medium to long-term forecasts, 1968 and 1982. The 1970s coal boom was the coal’s share of electricity generation was already only sustained one since World War I. The expected to decline to 17 percent by 2025 and 1970s-to-early 1980s energy boom temporarily 13 percent by 2050, with renewables becoming interrupted rising mining productivity and increasingly competitive (EIA 2020). declining coal employment. Environmental regulations provided additional headwinds after the passage of the Clean Air Act of 1990. Long-term U.S. Coal-Mining U.S. coal employment declined 68 percent between 1982 and 2000, while productivity Employment Trends and increased 440 percent over the 1982-2020 period, reflecting 8.6 percent annual growth. Productivity Growth Labor productivity in Wyoming coal mines U.S coal mining employment in the last century is 8.4 times greater than in Appalachia. reflects rising concentration in Appalachia Wyoming’s largest coal mine, the Peabody until the mid-20th century, with dissipation Energy North Antelope Rochelle Complex, thereafter as coal production migrated illustrates the higher productivity of western west. Within Appalachia, the last century saw strip mines. The enterprise produced spatial redistribution of coal activity from more than 98 million tons in 2018, three Pennsylvania and northern Appalachia toward million tons more than produced in all of central Appalachia centered on the region where West Virginia.17 Remarkably, only 5,558 Kentucky, Virginia, and West Virginia meet. As workers, barely one-tenth of total U.S. coal 15 Sources: https://www.eia.gov/outlooks/steo/. (Downloaded May 1, 2020) and EIA, Weekly Coal Production (downloaded from:  https://www.eia.gov/coal/production/weekly/. on May 19, 2020). 16  Constructed using an updated version of Betz et al.’s (2015) data. 17 Ibid. and EIA 2019 Annual Coal Report, Available at: https://www.eia.gov/coal/annual/ (downloaded April 30, 2020).  Socioeconomic Transition in the Appalachia Coal Region Some Factors of Success 30 Figure 3.4 U.S. Coal Labor Productivity (tons/miner): 1919-2017 800,000 600,000 400,000 200,000 – 29 39 82 54 63 72 17 12 02 19 92 20 20 19 19 19 19 19 19 19 20 19 Source: U.S. Census Bureau, Census of Mineral Industries. Various Years. employment,18 mined Wyoming’s 304.2 million electric generation to natural gas had begun. tons of 2018 coal production. By contrast, Coal production fell from just over one billion Appalachia coal mining employed 30,620 in tons in 2012 to about 681 million tons in 2019 2018. (Census of Mineral Industries 2017; EIA 2020), while U.S. coal employment and coal-mining Coal experienced a mini-resurgence between support industry employment fell 41 percent 2001 and 2012, but the trend reversed to about 56,000 workers (equivalent to a mere around 2012. The temporary resurgence 0.0037% of total nonfarm U.S. employment in was due to a shift in electricity generation 2019). This compares to 777,000 coal-mining and expanding global demand. During this jobs in 1919 (2 percent of total U.S. employment period, coal mining and coal mining support or 2.8 percent of U.S. nonfarm employment; U.S. jobs increased by nearly one-fourth, although Census Bureau 1975 and 1999, U.S. Census Bureau coal’s diminished importance in the labor Statistical Abstract, Table 1432).19 From a peak market meant that these gains translated of 863,000 coal mining jobs in 1923 (U.S. Census into about 18,000 new coal mining jobs, far Bureau 1975), coal employment contracted by 94 below employment levels seen in 1982. But by percent over the subsequent century. 2012, the increasing shift from coal-powered 18 These Wyoming/Appalachian productivity calculations use data from EIA 2019 Annual Coal Report, Available at: https://www.eia.  gov/coal/annual/ (downloaded April 30, 2020). 19 The source of 1919 U.S. nonfarm employment is Ghanbari, L. and M. McCaul. 2016. “Current Employment Statistics survey: 100  years of employment, hours, and earnings.” Monthly Labor Review. August 2016, https://www.bls.gov/opub/mlr/2016/article/ current-employment-statistics-survey-100-years-of-employment-hours-and-earnings.htm. (downloaded May 2, 2020). 31 Figure 3.5 Coal Employment by State (selected states) 400,000 300,000 200,000 100,000 – 29 39 82 54 63 72 12 17 92 02 19 20 20 19 19 19 19 19 19 19 20 19 PA Co l Emplo m nt WV Co l Emplo m nt OH Co l Emplo m nt VA Co l Emplo m nt KY Co l Emplo m nt AL Co l Emplo m nt Oth r App l chi n Co l Emplo m nt Note 1. Data sources: U.S. Census of Mineral Industries and U.S. Economic Census, various years. For 2017, U.S. Bureau of Labor Statistics, Quarterly Census of Employment and Wages. (https://www.bls.gov/cew/). Note 2. 2002 MD employment and 1992, 2002 2002 TN employment are disclosed only for a discrete range and are imputed. 2017 OH employment is not disclosed and 2018 data is used. 2017 TN employment is not disclosed, so 2014 data is used. Source: 1) U.S. Census Bureau, Census of Mineral Industries. Various Years. Note 3. Kentucky Appalachian coal would be overstated due to the inclusion of western Kentucky coal fields. In 1919, for example, western coal fields account for 27.6% of wage earners and 37.7% of total Kentucky coal employment in 2017 (1919 Census of Mineral Industries and EIA, 2020, Table 18). Coal within Appalachia: and flows. The northern Appalachian states of Pennsylvania and Ohio had no periods of Time and Spatial Trends sustained coal employment increase from 1919 to 2017 (1963-1982 was one of stagnation), The Appalachia coal story reflects general whereas the coal sector in central Appalachian aggressive decline, along with spatial states experienced mini-booms with realignment from north to central employment increasing during 1963-1982 and Appalachia. Pennsylvania’s 300,000 coal 2000-2010. Yet, the alternating bust-periods miners in 1919 was nearly 3.5 times the number vastly overwhelmed any positive employment of coal miners in West Virginia. The decline effects. The central Appalachia coal industry in north Appalachia began immediately after ultimately followed northern Appalachia, but World War I, while the central Appalachian with a lag of 20-30 years. By the 2012-2017 coal industry experienced more volatile ebbs period, central and northern Appalachian coal Socioeconomic Transition in the Appalachia Coal Region Some Factors of Success 32 employment each tracked downward. All five footprint of once-dominant coal is striking, Appalachian states experienced a long-term especially when viewed against service sector collapse of coal employment: over the last jobs. For comparison, Walmart employs 1.5 century, West Virginia lost 85 percent of coal million U.S. workers, and retailer Bed Bath & jobs, compared to 78 percent for Virginia, 85 Beyond employs about 65,000.22 percent for Kentucky, and 98 percent in both Ohio and Pennsylvania (Figure 3.5).20 At the county level, both the degree of local economic dependence on coal and For the three central Appalachian states, the spatial distribution of coal jobs have and especially West Virginia, coal mining changed substantially since 1950, illustrated played critical roles in their respective by comparing Figures 3.6 and 3.7.23 Figure 3.624 state economies through much of the 20th shows that the top one-sixth of ARC counties century. State-level employment data was not had mining employment shares over 40 percent available until the Census Bureau’s 1930 Census in 1950, and the top one-third had mining25 of Unemployment. Prior to that, the federal shares greater than 22 percent. Counties with government only collected employment data the highest shares were generally located along in selected industries. In 1930, PA, WV, and the spine of the Appalachian Mountains from KY coal mining respectively represented 7.3, Pennsylvania to Birmingham, AL, with higher 22.7, and 9.7 percent of their state’s nonfarm densities running from Pennsylvania through employment.21 By 2017, coal job shares had West Virginia and Kentucky. In the buffer fallen markedly to 1.9 percent in WV, 0.3 percent counties outside the ARC boundary, mining in KY, and 0.1 percent in VA (EIA 2019 Annual shares of the labor force are very low, with Coal Report; Bureau of Labor Statistics state the exception of the coal region near where employment data). By 2017, coal employment in Kentucky, Illinois, and Indiana meet to form a the three most coal-intensive ARC states was key “interior” coal region (Appendix B Figures only 14,000 in WV, 6,500 in KY, and 5,400 in PA. 3 to 9 reproduce state maps from the 1954 And between January 2017 and April 2020, total Census of Mineral Industries showing the high- U.S. direct coal employment declined further, coal intensity local economies in Alabama, from 51,000 to 43,800 (BLS, CES). The shrinking Kentucky, Ohio, Pennsylvania, Virginia, and 20 Beginning in the 1920s through the early 1960s, key reasons for declining coal demand were the growing use of electricity, oil,  natural gas and other fuels; substitution of diesel locomotives for steam locomotives; and the general rise of trucking over rail. See http://explorepahistory.com/story.php?storyId=1-9-B&chapter=0 and http://explorepahistory.com/story.php?storyId=1-9- 18&chapter=0. (downloaded May 1, 2020). 21  The 1930 employment-share data was collected from Volume of 1 of the 1930 Census of Unemployment that was collected in conjunction with the 1930 Census of Population. Using Tables 6 and 7 for each state, coal mining employment and total nonfarm employment were calculated by taking gainfully employed (approximately the labor force) minus “Class A” and “Class B” unemployed. 22  Sources: https://www.cnbc.com/2019/03/28/bed-bath-beyond-lays-off-nearly-150-of-its-65000-employees.html (accessed April 30, 2020) and https://corporate.walmart.com/newsroom/company-facts (accessed on April 30, 2020). 23 Note that both Figures 3.6 and 3.7 include an additional 100-mile buffer of counties outside of Appalachia to show the contrast  between the ARC region with nearby counties. Both figures are also divided into six equal-sized mining/coal employment share categories for ease of comparison, in which each category has an equal number of counties. 24  Figure 3.6 shows the 1950 residential share of the civilian labor-force working in mining (both coal and non-coal), which for most of Appalachia is almost exclusively coal mining (especially in 1950). 25 Combined coal and non-coal mining. 33 Figure 3.6 1950 County Mining Employment Shares for ARC and Buffer Counties (% of labor force) Source: 1950 Place-of-residence mining employment share of the labor force is derived from the 1950 Census of Population. The black line is the current boundary of the ARC. Outside the ARC region is a 100-mile buffer of counties. West Virginia).26 The combined mapping of (authors’ estimates). Figure 3.727 shows a these states (Appendix B, Figure 10) illustrates significantly truncated concentration of coal the sheer number and spread of counties with counties in 2016; only five counties in the ARC at least 1,097 coal employees in 1950. region had a coal employment share above 20 percent. Coal mining jobs account for at By 2016, 110 counties in Appalachia out of a least 17.4 percent of total employment in the total 420 ARC counties still had measurable top one-sixth mining-dependent counties,28 coal employment, but at very low levels, and in the top one-third counties, mining averaging 3.7 percent of nonfarm employment, accounts for between 11 and 17 percent of compared to 0.98 percent across 420 counties nonfarm employment. For the 15 most coal- (2014 data), and 0.23 percent nationwide intensive county economies,29 the coal- 26 For example, in 1954, the share of mining employment accounted for by coal in Kentucky, Pennsylvania, and West Virginia were  respectively: 81%, 82%, and 88%. 27 Figure 3.7 shows the (place-of-work) 2016 employed coal mining share of the civilian labor force. 28 Most mining in these states is coal mining. 29 Nine of the top-fifteen most coal-intensive counties are in WV, one is in PA, one is in OH, one is in KY, one is in VA, and one is in MS. Nine more counties had coal-mining nonfarm employment shares above 5 percent (4 in WV, 2 in OH, and one each in KY, MS, and VA). Socioeconomic Transition in the Appalachia Coal Region Some Factors of Success 34 Figure 3.7 2016 County Coal Employment Shares for ARC and Buffer Counties (% of labor force) Source: 2016 Place-of-work mining employment data is from the U.S. Bureau of the Census 2016 County Business Patterns as provided Upjohn Institute of Employment Research, who use a peer-reviewed linear programming model to estimate values for the cases when the federal government suppressed the information to maintain confidentiality. The county’s civilian labor force is from the 2016 American Community Survey five-year averages. The black line is the current boundary of the ARC. Outside the ARC region is the 100-mile buffer of counties. mining share of nonfarm employment For example, Raleigh County, WV had over ranges from 8.2 percent in Fayette County 1,200 miners in 2016, but its relatively large in southern WV to 33.7 percent in nearby population means that its coal employment Boone County, WV. The second highest coal share is below 5 percent. Kanawa County, WV mining employment share is 24 percent in (home of WV’s capital Charleston) and Jefferson Buchanan County in far southwest Virginia, County, AL (home to Birmingham) also have the only Virginia county directly bordering over 1,000 coal miners but very low coal both Kentucky and West Virginia. Overall, employment shares. 55 of the 420 ARC counties have coal mining employment shares over 5 percent. Today, West Virginia has the largest concentration of counties in which coal mining remains an economic driver. Thirteen WV counties have a coal mining share above 5 percent. There are a few other cases where coal mining remains important even though the share is below 5 percent. 35 SECTION 4 Identifying Successful Post-transition Counties Socioeconomic Transition in the Appalachia Coal Region Some Factors of Success 36 Our sample includes the 420 ARC counties Selection Criteria and plus counties located within 100 miles of the ARC region, yielding a total sample of 1,070 Methodology counties.30 Including buffer counties provides more variation in outcomes and explanatory We use the county level as our unit of analysis variables, and helps ensure we are not only to examine the community impact of transition considering a selected group of counties that away from coal mining, for the following reasons: are lagging by definition (ARC was formed to address lagging development). a. This is the unit ARC uses to classify the region’s communities. We define a county as coal-dependent if its mining employment share of the labor b. The most complete local-level secondary data force31 exceeded a certain threshold in 1950 are county level. or in a subsequent period. To determine the threshold for coal-dependence in 1950, we c. Counties are best suited to analyze changes consider the distribution of mining-intensity over time because their boundaries are across the 1,070 sample of ARC and buffer much less likely to shift than boundaries of counties and find that 99 ARC counties had municipalities or other places. at least an eight percent mining-share, with an average mining share of 23.7 percent. The d. Counties are the unit of analysis most often ARC average 1950 mining share of the labor used in rural U.S. research (Isserman 2007; force was 6.75 percent, and the average for Partridge and Rickman 2006). counties within 100 miles of the ARC was 0.8 percent. Given the decline in coal mining’s We first define the criteria by which we importance over time, we lower the coal- measure transition, and then the criteria that dependent threshold to four percent for 1980 define “success”. A county that has transitioned and thereafter.32 An additional 123 candidate must meet the following two conditions: counties met the 4 percent mining share threshold in 1980, bringing the total number of a. The county was dependent on coal mining “successful transition” candidate counties to sometime in the past. 222 (original 99 plus 123). b. It is no longer dependent on coal mining today Population growth is our main metric (either ceased or became a “trivial” part of the for success. There are many aspects to local economy). economic development that might reflect successful post-shock recovery and transition, 30 We do not extend the buffer farther, say to include the whole country, because that would introduce significant heterogeneity into  the model. 31 For data reasons, we use mining share of the labor force rather than coal mining share. We address any potential bias at the county  level below. Moreover, because coal was and still remains the dominant ARC mining industry, “mining-intensive” typically means “coal-intensive”. 32 This threshold may seem low but recall that we use the mining employment as a share of the civilian labor force (which includes  agriculture, proprietors, unemployed), not nonfarm employment, and this yields smaller share values. Between 1950 to 1980, U.S. coal mining employment fell 46 percent and the civilian labor force rose 72 percent. 37 Box 4.1 The case for using population growth to measure “success” Population growth is the most common metric for assessing regional success in U.S. studies. Betz et al. (2015) found that population growth highly negatively correlates with intensity of local coal mining employment, and they found a statistically significant causal link between coal mining employment and population decline. Other variables associated with economic prosperity —per-capita income, poverty rates, and growth in median household income—positively correlate with population growth. Moreover, population growth directly captures the movement of people in and out of these regions based on economic as well as socioeconomic factors. Population data is also readily available at the necessary disaggregated levels. The standard model of American urban and regional economics is the Spatial Equilibrium Model (SEM), which posits that people weigh the trade-off between income and quality-of-life, and locate in a place that gives the most satisfaction (Faggian et al. 2012; Partridge 2010). Given this trade-off, low median household income is not necessarily a sign of failure in settings where quality of life is high. The SEM assumes that households are freely mobile in the medium term; our chosen period of measure is 10 years. If, however, low-income households face information constraints or cannot afford to move, the SEM model is less applicable. But evidence on the outflow of people from low-income coal communities suggests that people do respond to local economic/quality-of-life conditions (Partridge 2010). Population growth is therefore a good measure of community success in countries with relatively high medium to long-term geographical mobility, such as the U.S. (Partridge 2010). People “vote with their feet” and move to places they expect will provide more satisfaction (or utility) according to their personal preferences (Faggian et al. 2012). To take a classic American example, since the 1950s there has been a mass net-outmigration from the (relatively) wealthy Northeast and manufacturing-intensive Midwest toward relatively lower-income but warmer “Sunbelt” states (Partridge 2010). This suggests that people are trading off lower incomes to have a higher quality-of-life, or alternatively, they would require a compensating [income] differential to induce them to live in the cold Northeast and Midwest.33 including measures of population well-being, growth as a metric for “success” also enables average income level, non-income quality statistically robust estimates. Box 4.1 presents of life, availability of work, dynamism and key examples from urban and regional diversification of the local economy, and the economics literature and labor mobility and local economy’s integration into the larger demography studies illustrating the use of national economy or national or global supply population growth as the main variable to chains. The complexity of these aspects makes measure development. them difficult to analyze, however, especially at the county-level (Appendix C summarizes We use a simple regression model to identify various methodologies used to understand the counties that over or underperformed economic impact of coal mining and their pros in terms of population growth given the and cons). Because of our community-level county’s initial endowments (see Box 4.2 for focus and our objective to differentiate coal description of the regression equation). We transition experiences between counties, we test multiple time periods for our analysis, select a relatively simple quantitative measure arriving at our base model which considers for which county data are available for our population growth over the1950 to 2018 long period of interest. Using population period, as well as the sub-periods 1980 to 2018 33 If economic actors are making rational decisions consistent with an underlying utility function, then the “transitivity property”  must hold across all potential migration options—for example., if migrants prefer Georgia to North Carolina and North Carolina to New York, then they must also prefer Georgia to New York. Faggian et al.’s (2012) literature review describes empirical studies that appraise whether the transitivity property holds in U.S. migration decisions; studies find that the transitivity property holds in almost all cases when assessing the migration choice set across all 50 U.S. states, meaning that net-migration reflects a rationale aggregation of people’s preferences and utility functions. Socioeconomic Transition in the Appalachia Coal Region Some Factors of Success 38 Box 4.2 Methodology for identifying Appalachian counties relatively successful at transitioning from coal mining Our dependent variable is the percent change in population between period t and period t+j, where j depends on the time period being considered. We estimate the following regression model: Δpopulationt, t+j = a0 + a1 Mining Employment Sharet + a 2 (Mining Employment Sharet )2 + a3 Populationt + Metropolitan1973+ NonAppalachian + Region where NonAppalachian is a dummy variable equal to 1 if the county is in the 100-mile buffer zone outside of the ARC’s region. Region is a vector of two region variables, ARC North and ARC South (with reference category ARC Central). Population is the log of the initial-period population, and Mining Employment Share is the initial-period share of the county’s civilian labor force employed in mining. It is important to note that the mining variable includes all mining products because the exact number of coal miners is not reported for many counties due to federal confidentiality requirements—though this does not end up being a constraint for subsequent steps. The mining employment share proxies for the degree of dependence of the local county on the mining sector35 (Appendix D provides additional details of the regression specification). and 2000 to 2018.34 We ranked the regression growers for the 1980-2018 model; this resulted model residuals (error terms) from highest to in an additional 41 potential candidate lowest for the 99 “mining-intensive county” counties for successful transition. observations. Using population growth as the dependent variable, positive residuals are The regression results confirm our viewed as a “success” because that county hypothesis of a negative correlation between over-performed in terms of population growth population growth and mining intensity given initial conditions. For each period (that in the local labor market (that is, the mining is, 1950-2018, 1980-2018, and 2000-2018), we employment share takes a negative coefficient considered the top one-third of the mining- value (Appendix D Table 1).36 This is consistent intensive counties in terms of their residuals, with the fact that the average population with the largest positive residuals as potential growth for the entire sample of 1,070 counties candidates for successful transition. This is greater than the average for coal-dependent resulted in 33 cases for the 1950-2018 model. counties, many of which experienced And we repeated this residual ranking for the population loss. ARC counties significantly 123 counties that were “mining-intensive” lagged national average population growth, in 1980, identifying the top one-third fastest and mining-dependent counties pulled 34 We also examined population growth models, decade by decade, starting with 1950 to 1960 and finishing with 2010 to 2018, but  found these redundant with the longer time-periods. 35 We caution that these regressions are descriptive, not causal.  36  We anticipate that, given the standard environmental, rent-seeking, resource curse and crowding-out/Dutch disease concerns about natural-resource dependent economies, the coal mining share coefficient will be negative. Other reasons to expect that the coal mining coefficient will be negative is that coal mining has been negatively linked to small business start-ups and entrepreneurship (Betz et al. 2015). Note that mining may have a nonlinear effect on local economic outcomes, which is why a mining-share quadratic term is included—in which we expect that the negative (linear) effects of mining-dependence generally begins to level off at high levels. In our modelling, we virtually always find that the quadratic mining-share-regression coefficient was the opposite sign of the linear term (which would then be positive) and statistically significant. 39 Figure 4.1 ARC % Population Growth over 1950-2018 for Mining-Intensive Counties Notes: The non-cross-hashed counties are counties with at least 8% of employed residents working in mining in 1950. The cross-hashed counties represent 123 additional counties selected by having at least a 4% mining share in 1980. Figure shows the union of these two cases, or 222 counties. The black line is the current boundary of the ARC. Outside the ARC region is the 100-mile buffer of counties. down the ARC average.37 For the 222 counties a. been within the top-third for population identified as mining-intensive in 1950 growth in either the 1950-2018, 1980-2018, (8 percent threshold) or 1980 (4 percent or 2000-2018 periods after adjusting for threshold), their average population growth conditions in the regression model; between 1950 and 2018 was only 27 percent, and 95 of these 222 counties lost population b. population growth above the ARC average (Figure 4.1). (metropolitan38 or non-metropolitan comparative averages) in at least one of the To qualify as a “successful transition” following periods: 1950-2018, 1980-2018, county, we set the following criteria. The 2000-2018, 1980-1990, 1990-2000, 2000- county must have: 2010; 37 U.S. growth for all counties averaged 116 percent from 1950 to 2018, compared to 94.5 percent across 1,070 counties in our sample,  70 percent for the ARC counties, and 44.9 percent for the non-metropolitan ARC counties. The averages obscure some extremes. For example, McDowell County WV had the highest average 1950 mining share of 62 percent and a population loss of 82 percent between 1950-2018. In 1980, the highest mining share of the labor force was in Martin County, KY at 46 percent, and it experienced a corresponding 19 percent 1980-2018 population loss. In 2000, the highest mining shares of the labor force totaled 16 percent in Campbell County, TN, Martin County, KY, Mingo County, WV, and Wyoming County, WV, with corresponding 2000-2018 population losses of 1 percent, 10 percent, 16 percent, and 19 percent. 38  We use the 1973 metropolitan definition, given it is in the middle of sample period. Socioeconomic Transition in the Appalachia Coal Region Some Factors of Success 40 c. eliminated, or nearly eliminated, coal 1. Sequatchie County, in S.E. Tennessee employment;39 (selected from the 1980-2018 list). d. not substituted other fossil fuel extraction for 2. Laurel County, in S.E. Kentucky (selected coal; and from the 1950-2018 list). e. coal as the dominant mining activity.40 3. Athens County, in S.E. Ohio (selected from the 1950-2018 list). The pool of potential “successful” transition candidates we identified is small. Among the 4. Noble County, in S.E. Ohio (selected from 222 mining-intensive counties, only 28 non- the 1980-2018 list). metropolitan and 4 metropolitan counties grew faster than their corresponding ARC average Figure 4.2 maps the four successful transition population growth. Moreover, population 42 counties. The geographical distribution of growth is spatially persistent; if anything, the low, median, and high-performing counties is relative performance of the most coal-intensive quite dispersed across the ARC, suggesting no areas in central Appalachia lagged further systematic geographic role in their selection. behind after 1980, and especially after 2000 Sequatchie County has fully transitioned away (Appendix E Figures 1 and 2 present population from coal, and Athens County also appears growth maps for the 1980-to-2018 and 2000-to- to have fully transitioned.43 The last year that 2018 periods). coal employees appeared in Athens County was 2014. Laurel County’s coal industry Based on these selection filters, four peaked in the early 1980s and mostly closed “successful” transition counties emerge. Only by the early 1990s, but there remains a small two counties remained after subjecting the top operation with about 10 employees. Since the 33 fastest growing, mining-dependent counties late 1980s, Noble County also has had small in 1950 to the additional filters. And for the 41 coal operations that employ 40-50 workers fastest growing, mining-dependent counties (but no longer coal dependent). Noble County is in 1980, only two met all the selection criteria the weakest case among the “successful” coal- (despite testing for other potential qualifying transition counties. counties over a range of time periods). In order of successful growth, the four “successful” coal transition counties are: 39  We set an upper bound on the coal employment share of the labor force at less than 2 percent to signal that a county is no longer coal-dependent. In a relatively populated Appalachian mining county, that would represent a single typical-sized mine, and for a small ARC county, that would represent one small coal mine. There are no publicly-provided county data for coal employment with less than three or four coal mining companies because of confidentiality reasons. Yet, even in those cases, the Census of Mineral Industries reports a coal employment range, and other data sources report specific values. Although we cannot identify coal mining shares for the entire sample, we are able to for the success cases, enabling us to verify whether they exceeded the 2 percent coal mining employment share threshold. 40 We impose a threshold that at least two-thirds of the county’s mining employment is in coal mining (relying on the 1954 Census of  Mining Industries). 41 The mapping is split into six equal-sized groups of 37 or 38 counties each.  42 The 1950-2018 ARC non-metropolitan county population growth averaged 45 percent, while ARC metropolitan population growth  averaged 177 percent (bottom of Appendix E Table 1) 41 Figure 4.2 Successful Transition and Low, Median, and High Relative Performing Counties41 Notes: The color-coding corresponds to Appendix E, Tables 2-6. Yellow denotes the four successful- transition coal mining counties, blue denotes the relatively worst-performing mining-intensive counties, green denotes the median-performing mining-intensive counties, and orange denotes the higher-performing mining-intensive counties. Robustness Check with Other To do this, we compare our four “successful transition” counties to other mining-intensive Measures of Success counties with respect to key indicators associated with well-being: population growth We test the robustness of our selection rate, mining and coal employment shares, methodology by comparing our results to median household income, and poverty. With other variables commonly used to measure only four successful coal transition counties, successful post-coal economic development. it is impractical to conduct a standard analysis 43  The sources for this coal mining information on the four success stories are: (1) 1954, 1977, 1982, 1987, 1992, 1997, 2002, 2007, 2012, 2017 Census of Mineral Industries, (2) Kentucky Coal Facts, various years but especially the 16th edition in 2017, (available at: https://eec.ky.gov/Energy/News- Publications/Pages/Coal-Facts.aspx. downloaded May 5, 2020). (3) 1993, 2000, 2017, and 2018 Ohio Mineral Reports. (available at: http://geosurvey.ohiodnr.gov/news-events/recent-news-events- and-features-archive/post/2018-mineral-industries-report. Downloaded May 5, 2020). (4) Mining Health Safety Administration CY 2009-CY 2015 Coal Mining Employment by state and county. (downloaded from: https:// www.msha.gov/sites/default/files/Data_Reports/Charts/Coal_Employment_by_State_and_County_CY09to15.pdf. Accessed May 7, 2020.) Socioeconomic Transition in the Appalachia Coal Region Some Factors of Success 42 of statistical correlations; we therefore use economic diversification as the county descriptive analysis to compare outcomes of became more integrated with Chattanooga. our four counties to the top-10, median-10, Like Laurel County, Sequatchie County and bottom-10 ranked performers (based on experienced relatively rapid growth while population growth) within the 99 mining- still coal dependent, and transitioned away dependent ARC counties44 (results in Appendix from coal by the early 1990s before large- E Tables 1-4). scale coal production ceased. Of the four coal-transition success stories, • Athens County’s average population only Sequatchie County, TN is a metropolitan growth rate was slightly below the ARC’s county, part of the Chattanooga metropolitan 1950-2018 non-metropolitan average, but area since 1973. Thus, urban-led growth is exceeded the ARC average in four of seven not the reason for the other three counties’ decades. Between 2000-2010, its 3.8 percent successful transition. Sequatchie and Laurel population growth was between the U.S. Counties both experienced relatively rapid nonmetropolitan average of 3.5 percent and population growth. Laurel County’s growth the ARC’s average of 4.5 percent. While not roughly tripled the nonmetropolitan average booming, Athens County has experienced ARC population growth rate between 1950- steady growth. It had a moderate-sized coal 2018, and exceeded the ARC average during industry in the 1950s, but its coal industry every period. Sequatchie County slightly was very small by the early 1960s and trailed the average ARC metropolitan rate, disappeared after 2009. but exceeded it in the 1970s. By the 1990s, its population was growing faster than the ARC • Noble County generally underperformed average (see Appendix E for further detail on ARC non-metropolitan population growth robustness analysis of “successful” counties). rates from 1950 to 2018, but greatly over- performed in the 1990s. From 2000 to 2010, It is notable that these successful counties Noble County’s 4.2 percent population transitioned at different times: growth was between the U.S. and ARC nonmetropolitan averages. Noble County • Laurel County was an intermittent coal still has a small coal mining sector producer, with heavy production during employing about 40-60 workers since the 1940-1960 and 1970-1988, after which its mid to late 1970s (Census of Mineral Industries coal production was small. The county and Ohio Mineral Report). Noble County transitioned from coal dependence around also has had a small oil and gas industry 1990.45 during high-drilling periods. Having more intensive-coal mining neighbors such as • Sequatchie County’s coal dependence ended Belmont County, Ohio means that some in the mid-1980s as the share of employees Noble County residents commute elsewhere in mining declined. This coincided with to work in mining, mainly in oil and gas but the 1980s coal market bust, and increasing also coal. This explains why Noble county’s 44  We use the 1950-2018 model’s 99 regression residuals to determine the included counties. We select this model because it indicates long-run success and we have already shown strong spatial persistence in population growth over the 1950-2018. 45 Source: op. cit. Kentucky Coal Fact Book, 2016.  43 place-of-residence mining share of the tend to have had higher mining shares than labor force is higher than anticipated given the four successful counties (except in the its small local mining/coal employment 1980s), although the gaps became small levels. Therefore, while we include it as a beginning in the 1990s. Not surprisingly, “successful” case of transitioning from coal, sector diversification away from mining it is a marginal example. After the energy is associated with economic prosperity— bust in the early 1980s, Noble County was although the top-10 performers are skewed less exposed to the energy industry, and because 5 of the 10 counties became coal in particular. Thus, it appears to have metropolitan by the end of the sample period. transitioned from coal about a decade before Until 2000, median-performing, mining- the relative prosperity of the 1990s and early intensive counties had higher mining-shares 2000s. Nonetheless, Noble County did not than either the relatively successful transition fare well after the 2008 Great Recession. counties or the top-10 performing mining- intensive counties, after which their mining Taken together, our results indicate shares converged. The bottom-10 performing that after eliminating over 98 percent mining-intensive counties exhibit a surprising of mining-intensive county candidates, pattern; their mining-shares converged to the four selected counties are relatively near the four successful transition counties successful. The four successful transition by 1970, and afterwards maintained mining counties experienced much faster average shares either equal to or below the four population growth than many counties. Yet, successful transition counties. One possible the successful counties modestly trailed the explanation is that the underperformers top-10 performing mining-intensive counties include metropolitan counties that had both between 1950-2018. This is partly explained coal and complementary manufacturing in the by outliers Shelby County, AL, part of the mid-20th century—for example, integrated Birmingham metropolitan area, and Pickens steel mills. After these manufacturing County, GA, which has experienced rapid industries failed in the face of foreign amenity-led growth. The four successful competition and technological change (for counties grew faster than the 10 median- example, shifting from integrated steel to performing counties that continued to be mini-mills), the economic drivers of these mining-dependent and especially the 10 worst small industrial cities dissipated, leading to -performing counties, which experienced depopulation and declining coal demand. declining populations between 1950 to 2018. Since 1990, the four successful transition counties have experienced slightly faster average growth than the top-10 cohort, suggesting that the benefits of transitioning away from coal are increasingly paying off. The four successful transition counties had above average mining shares compared to the ARC region until 1990, after which their shares fell (Appendix E Table 2). The top-10 fastest growing, mining-intensive counties Socioeconomic Transition in the Appalachia Coal Region Some Factors of Success 44 SECTION 5 Factors Underlying Successful Coal Transition Why were the four counties successful? In this section, we explore factors most likely to have contributed to their success, based on variables that social scientists typically associate with local economic development. Some of these factors appear to correlate with successful economic transition, but others play little role in the Appalachian context. Our results are based on descriptive quantitative data for the 99 mining-intensive Appalachian counties, using the same comparison categories as the prior section—that is, top-10, median-10, and bottom-10 performers in terms of population growth. 45 Human Capital to West Virginia University in Morgantown, significantly increases that category’s average The four successful coal mining transition education. High state investment in university counties had average educational attainment settings is hard to replicate. Based on these levels within ARC but below national relatively mixed results, with the exception of averages. Further, this performance is skewed Athens County, we conclude that educational upward by the relatively strong performance attainment in itself does not explain why the of Athens County, home to Ohio University. counties had relative success in transitioning Table 5 in Appendix E shows the average away from coal. share of the over-25 population with at least a Bachelor’s degree (CG), some college including an Associate’s degree (SC), and high school Sectoral Composition and graduates (HS) for 1950, 1980, 2000, and 2018. Besides Athens, tertiary educational Diversification attainment is generally low in the other three successful transition counties, especially The underlying structure of local economic the college graduate shares, which are well production affects the types of jobs available below the US average (Figure 5.1). The top- and productivity levels. By comparing the 10 performing mining-intensive counties 1950 pre-transition economic structure of also have a skewed educational profile. For mining counties across the ARC region, we example, Monongalia County, WVA, home may detect variations that have helped some Figure 5.1 Share of College Graduates (% of over-25 population) 35 30 25 20 15 10 5 – 50 80 18 00 20 19 19 20 Av US L ur l Top 10 p rformin ARC non-m tro Ath ns M di n 10 p rformin ARC m tro S qu tchi Bottom 10 p rformin Nobl Source: Author calculations using U.S. Census Bureau’s Decennial Census of Population (various years). Socioeconomic Transition in the Appalachia Coal Region Some Factors of Success 46 counties diversify. Appendix E Table 6 reports thereafter. Whereas the four successful the employment shares for the agriculture transition counties had much lower sector, manufacturing, and self-employment in manufacturing shares in 1950 (averaged 11 1980, 2000, and 2018 to show patterns that might percent) compared to other coal-intensive explain the relative successful performance of counties, manufacturing sectors in our the four coal-transition counties. “successful” counties increased rapidly after 1950, peaking in 1980-2000. Meanwhile, the Farming. Farming can have mixed economic other three mining-intensive categories—and effects. On the positive side, agriculture in fact much of the U.S.—were experiencing cannot be offshored and can be a source of deindustrialization. The exception was Athens new entrepreneurship. However, a degree of County, OH, where economic transformation import substitution together with labor-saving centered around Ohio University. After productivity increases have led to fewer farm 2000, however, manufacturing intensity in jobs. Farming’s share of total U.S. employment the four successful transition counties fell was 41 percent in 1900, 21.5 percent in 1930, 4 sharply, in line with regional and national percent in 1970, 1.9 percent in 2000, and just over trends. Moreover, the future prospects of 1 percent in 2019 (U.S. BLS; Dimitri et al. 2005). “reshoring” manufacturing from abroad Heavy reliance on farming is unlikely to offset faces stiff headwinds, given established and declining coal mining employment. The four highly efficient global supply chains that rely successful transition counties all had relatively on low-wage countries for labor-intensive large average shares of agriculture employment production. Even the COVID-19 pandemic and in 1950, averaging 32 percent, larger than attendant pressure on global trade is unlikely the 1950 ARC non-metro farm share, and to significantly realign manufacturing toward considerably greater than the other comparison the U.S. Relying on new manufacturing to counties. Consistent with the national trend, replace coal-mining going forward is risky, average farm employment shares fell in all four and additional costs are associated with performance categories after 1950, becoming Appalachia due its remoteness from suppliers relatively insignificant today. and customers. Manufacturing. A large manufacturing Self-Employment. Within highly developed sector typically raises average productivity economies like the U.S., higher self- and wages compared to resource or service- employment46 intensity tends to be associated based economies. Manufacturing has long with faster economic growth, larger multiplier been considered key for rural development, effects through locally-based supply chains, especially during the mid-20th century. U.S. and profits retained locally, all of which support Bureau of Economic Analysis (BEA) data local growth (Stephens and Partridge 2011; indicates that beginning in the mid-1970s, Stephens et al. 2013; Tsvetkova et al. 2019). High rural-based manufacturing grew relatively levels of local entrepreneurship can help make faster than in metropolitan areas, and had a local communities more resilient to economic higher share of manufacturing employment shocks. Stephens and Partridge (2011) find 46  Self-employed workers are defined as those who own their own firm as either a pass-through (income) business, S corporation, or partnership. These firms can employ other workers. A large share of firms start as self-employed, but as they grow, there are legal advantages to become a C corporation—e.g., they can be publicly traded on financial markets. 47 that greater shares of self-employment support significantly from national patterns (Appendix growth in the ARC region. The four successful E Table 7 reports population shares for those transition counties exhibit quite high shares of under age 18 and over 65). The four successful self-employment, averaging 33 percent in 1950 coal-mining transition counties exhibit similar and sustained average self-employment rates patterns to the rest of the ARC region, suggesting throughout the adjustment period relative to the that demographic-age structures do not explain top, median, and bottom-performing, mining- the relative success of these counties. intensive counties (note that Athens County is an outlier). While based on a small sample size, these results point to potentially using local assets and Geographic Characteristics and capacity to promote small-business development to counteract declining coal employment. Agglomeration Economies We look at several geographic characteristics Population Demographics that commonly affect local economic growth: extent of urbanization, proximity to metropolitan Population age structure affects current areas, agglomeration economies, natural and future economic growth. In particular, amenities, and topography. We also look at the “dependency ratio”—that is, the share of sociodemographic factors such as inequality, children and senior citizens in a household— unemployment, and race (summarized in reflects the population share not contributing Appendix E Table 8). to economic production and at least partially dependent on local government for education Urbanization. The degree of urbanization is or services, including healthcare. A larger positively correlated with economic growth share of children can support future growth rates, but this is not systematically borne out once they enter the local labor force, whereas in the data for Appalachia. Based on the 2013 a larger share of seniors may increase the local USDA rural-urban codes (RUC) methodology,48 fiscal burden.47 The age structure and trends which assigns a value of “1” to the most urban, in Appalachian coal counties do not vary and a value of “9” to the most rural, we find that 47 There can be positive returns to large retired populations, e.g., in destination retirement markets such as Arizona, the far south  Atlantic states and Gulf Coast states, where local economies can be boosted by attracting sufficient numbers of (wealthy) retirees. In Appalachia coal-country, however, we do not observe a pattern of senior citizen wealth inflow that offsets the fiscal burden, not least because degraded environmental conditions do not attract in-migration. 47 The 2013 RUC codes are defined below with more details from the USDA, available at: https://www.ers.usda.gov/data-products/  rural-urban-continuum-codes.aspx, downloaded on July 17, 2020. Metropolitan Counties Code Description 1 Counties in metro areas of 1 million population or more 2 Counties in metro areas of 250,000 to 1 million population 3 Counties in metro areas of fewer than 250,000 population Nonmetropolitan Counties 4 Urban population of 20,000 or more, adjacent to a metro area 5 Urban population of 20,000 or more, not adjacent to a metro area 6 Urban population of 2,500 to 19,999, adjacent to a metro area 7 Urban population of 2,500 to 19,999, not adjacent to a metro area 8 Completely rural or less than 2,500 urban population, adjacent to a metro area 9 Completely rural or less than 2,500 urban population, not adjacent to a metro area. Socioeconomic Transition in the Appalachia Coal Region Some Factors of Success 48 Figure 5.2 Distance to Metropolitan Areas (miles) ARC m tro 75 Bottom 10 p rformin 60 ARC non-m tro 45 30 15 M di n 10 p rformin L ur l 0 Top 10 p rformin S qu tchi Nobl Ath ns DistMSA Av US Source: Author calculations using U.S. Census Bureau’s Decennial Census of Population (various years). Appalachia’s non-metropolitan counties are area.49 While closer than the average U.S. county more rural than the national average, and the and the ARC non-metropolitan county average, four successful counties are even more rural three of the successful transition counties than other mining-intensive counties. This are actually farther away on average than the result suggests that lack of urbanization did top-10 performing mining-intensive counties not impede their relative success. (Figure 5.2; details in Appendix E Table 8). This implies that urban access is unlikely to have Proximity to metropolitan areas. Rural played a tangible role in the transition of the communities close to metropolitan areas four successful counties. grow faster due to greater access to urban labor markets for commuters, urban services Agglomeration. Lack of agglomeration for households, and urban markets for firms economies represents an impediment to (Partridge and Rickman 2008; Partridge and economic development in lagging U.S. Olfert 2011). To assess how much of the relative regions. Agglomeration effects include cost- “success” of the four transition counties is saving efficiencies from economies of scale due to proximity to urban areas, we consider and network effects that arise from urban the distance from the population-weighted agglomeration, namely when firms are centroid of the county to the population- located near to each other. Many rural under- weighted centroid of the nearest metropolitan developed regions lack the population size 49 The source of the DISTMSA variable is author calculations using the STATA statistical software and U.S. Census geocoding for  county centroids. 49 to sustain a critical mass of producer and tourism, may affect their transition success. The consumer services to support growth. To county case studies presented in section 6 below understand whether urbanization supported examine this aspect in more detail. our successful transition counties, we consider their population levels (Appendix E Table 9). Topography. Topography matters for economic The four successful coal transition counties development. Mountains can attract migrants are all sparsely populated, with a 2018 average and tourists, for example, but can also impede population of 38,800. The median-performing access and complicate construction of buildings counties are also small with an average 2018 and physical infrastructure. Unlike the Western population of about 34,800. Surprisingly, U.S. mountains with their typical wide valleys, the top-performing and bottom-performing Appalachia’s mountain terrain tends to lack mining-intensive counties are more populated, level land. Using the USDA’s TOPO51 index with 2018 populations averaging over 90,000. scale, which ranges from 1 (flat) to 21 (most This pattern seems to contradict the tenet mountainous), the ARC region ranks over 15 on that agglomeration economies are a major the scale, compared to a national average of determinant of local economic growth. The 9. But the four successful transition counties finding is nevertheless encouraging because have an average topographical score of 17.7. it suggests that relatively unpopulated coal- This is similar to other mining-intensive mining communities can transition without counties, suggesting that topography was not a devastating economic consequences. contributing factor to their relative success. Natural Amenities. The U.S. Department of Inequality. Household income inequality, as Agriculture (USDA) Economic Research Service measured by the Gini coefficient (where “zero” calculates the level of natural amenities at the reflects perfect equality and “1” is perfect county level based on the presence of “warm inequality) is very similar across the entire winter, winter sun, temperate summer, low ARC region and national sample. This suggests summer humidity, topographic variation, and that differences in inequality and its associated water area.” This amenity index uses a one-to- 50 negative social effects are not a determinant seven scale, seven indicating the most amenities. factor for successful transition from coal. Partridge (2010) reviews the literature to show that a nice climate and pleasant landscape Unemployment. Unemployment rates (2018 attracts migration, with (net) moves toward data) are slightly high across the ARC region, warm winters, mountains, lakes, and oceans. but consistent with the trend of higher For the four successful transition counties, their unemployment rates in less-populated counties. average level of natural amenities measures There are otherwise no discernable differences 3.5, which is quite similar to Appalachian between the four successful transition counties comparator counties and to the entire U.S. and other mining-intensive counties. Despite the implied small variation in natural attractiveness, the degree to which counties Race. Finally, the minority share of the local leverage their natural amenities, such as for population can be an important socioeconomic 50 The source of the AMENTIY measure and details in its construction are available from the USDA at: [https://www.ers.usda.gov/  data-products/natural-amenities-scale/] downloaded on July 17, 2020. 51  See the source of the AMENITY measure for details on the TOPO index. Socioeconomic Transition in the Appalachia Coal Region Some Factors of Success 50 indicator commonly correlated with lower in close-knit communities, these residents may human capital and lower incomes. Rural ARC lack “bridging” social capital that connects counties are heavily white, with no clear them with outside communities and their differences across the mining-intensive innovative ideas. An illustration of this is counties. This suggests that race does not when residents in a small town are skeptical of explain different long-term economic effects outsiders and do not integrate them. in the four successful transition counties. Social capital cannot be directly measured, so we use proxies. For example, researchers Social Capital often use the proportion of the population that engages in various local associations, Social capital, defined as “norms of trust, votes, or voluntarily responds to the Census. reciprocity, and networks of relationships The common theme of these proxies is that existing within local communities” ,52 tends people engage in behavior that may have little to lubricate economic activity and the individual benefit but can advance community ability to collectively organize to improve interests or serve a public good (Rupasingha et socioeconomic well-being (Audia and al. 2006). Social capital measures considered Teckchandani 2010). Social capital makes here are drawn from Rupasingha et al. (2006).53 basic market transactions easier as more trust Appendix E Table 10 presents the following implies less need to draw up complex contracts county-level social capital proxy values: or to monitor the behavior of economic parties. Social capital also facilitates common • Total number of religious, civic, business, community efforts through enhanced political, professional, labor, and cooperation and trust that all parties are recreational establishments per 1,000 working for the same goal. Researchers have residents (ASSN2014). pointed to the importance of social capital in increasing community resiliency in the wake • Voter turnout in the 2012 general election of disasters and downturns (Aldrich and Meyer (PVOTE2012). 2015). In terms of economic development, social capital increases the cooperation • 2010 Census household response rate necessary to support and promote new local (RESPN2010). investment and the likelihood of institutional support for winners compensating losers • Number of non-profit organizations (Reese and Rosenfeld 2002; Woolcock 1998). including those with an international approach (NCSS2014). Yet social capital may have unintended or counterproductive consequences (Putman • A composite Social Capital Index (SK2014).54 2000; Woolcock 1998). For example, while high social capital may bind local residents together 52 Putman 2000.  53 Social capital data is from https://aese.psu.edu/nercrd/community/social-capital-resources.  54  he Social Capital Index SK2014 is standardized across all US counties, such that it has a mean of zero and standard deviation of 1. T Negative numbers reflect below-average social capital and positive numbers reflect above-average social capital. 51 While social capital may change over time, it government spending and fiscal pressures tends to be persistent at the local level. If we (such as large fiscal deficits) are often used believe that higher social capital positively as indicators of capacity (Sharp and Moody correlates with local economic development, as 1991). Public spending can create economic the literature posits, then we might expect these multiplier effects, reduce economic instability, factors to facilitate transition from coal mining. improve infrastructure, and produce goods and services for local populations (Allard Strikingly, throughout the ARC region, in both 2017). Fiscal pressure, by contrast, can deter mining- and non-mining-intensive counties, government investment in activities needed to overall social capital is below the US average. create growth and local well-being (Johnson et This may help explain why ARC counties lagged al. 2015). even prior to coal transition. Lower social capital may also impede transition given the lack of Overall, local government capacity in bridging capital to connect with outsiders. The the four successful transition counties four successful transition counties exhibit social appears broadly similar to the ARC region capital levels broadly in line with their ARC (Appendix E Table 11). On average the four neighbors. With the exception of Noble County, successful transition counties raise about half they underperform the US average across the of their revenue locally, similar to the top-10 range of proxy indicators. As measured by performing mining intensive counties. With traditional social capital proxies, therefore, we the exception of Athens County, they tend find no strong link between social capital and to experience slightly less fiscal pressure the successful transition of coal-dependent compared to other counties.55 counties. Health Outcomes Government Capacity The four successful counties compare well Economic development literature posits a to national and regional health outcomes, link between local government capacity and except in deaths by drug overdose. Health economic well-being, although there are few outcomes tend to be better in communities quantitative studies. Lobao et al. (2014) find with more social capital and fiscal capacity, that U.S. county governments with greater and thus indirectly correlate with economic institutional capacity tend to be more active growth. Moreover, the availability of local in social and business policy formulation economic opportunities may further reduce and provide more public services. Local fiscal “deaths of despair”, such as drug overdoses resource autonomy is expected to make local and suicides. We therefore consider the governments more accountable to civil society mortality rate as a proxy for the effectiveness and more motivated to improve services (Pöschl of local social capital and government capacity and Weingast 2015). Localities able to raise to support better quality of life. greater revenues tend to face lower barriers to improving conditions (Johnson et al. 1995). Local 55  U.S. counties are required to balance their budgets annually so a close association between revenues and expenditures is to be expected. Socioeconomic Transition in the Appalachia Coal Region Some Factors of Success 52 Figure 5.3 Drug Overdoses per 100,000 Population 25 20 1980 2014 US Av 2014 15 (p r 100,000 popul tion) Source: Author calculations using DRUG OVERDOSES 10 data from the Institute for Health Metrics and Evaluation (IHME). 5 – o l o ns hi bl tr tr in in in ur tc No m m m h m m At or or or L n- qu C rf rf rf AR no S p p p C 10 10 10 AR n p om To di tt Bo M Mortality: National life-expectancy in 1980 Opioid Epidemic: The Appalachia region averaged 73.8 years, slightly higher than the has been severely affected by the opioid ARC region. By 2014, longevity had increased epidemic, and includes some of the most to 77.8 years nationally, consistent with the afflicted counties nationwide. By 2014, ARC top-performing coal-intensive counties and metropolitan and non-metropolitan counties four success stories, but higher than the saw drug overdose death rates of 15 and 17 ARC non-metropolitan average of 76.0 years per 100,000 population, respectively, a sharp (Appendix E Table 12). With respect to other increase from 0.7 in 1980 (Figure 5.3). Overdose mortality indicators—such as death by suicide, deaths were particularly high in the 10 middle- alcohol poisoning, or violent acts—the four performing counties, where the rate averaged successful transition counties performed 31/100,000.56 The four successful transition better than U.S. and ARC region averages, with counties had slightly lower rates (averaging one key exception: drug overdoses. It is notable 13.8/100,000), although still nearly 40 percent that in all other “deaths of despair” categories, above the U.S. average. Laurel County, KY, the successful transition counties—and in fact had an alarmingly high overdose rate of much of the Appalachia region—performed 20/100,000, twice the national average. Nobel better than the national average, a surprising County, OH, by contrast, posted a much lower outcome given their remote rural locations. overdose rate, long life expectancy (81 years), and above average performance in other health-related indicators. 56 This group includes Floyd County, KY (47.7/100,000), Leslie County, KY (44/100,000), and Whitley County, KY (35.4/100,000). 53 SECTION 6 A Closer Look at Successful Transition: County Case-studies In this section, we present case-study analysis of the four successful transition counties in Appalachia, and one case-study from outside the region. We describe the types of opportunities and resources available within each community that likely contributed to its relative success in transitioning from mining- dependence. It is notable that all four counties increased their coal activity during the mini-coal boom in the late 1970s and early 1980s before ultimately reversing course. The factors discussed here do not explain the decline of coal activity, but rather provide insight into the post-coal economic recovery. We focus particularly on the role of distance and road infrastructure, local industries beyond mining, institutional capacity, and social capital networks. For comparison, we also present a case study of successful transition by a resource-dependent county outside the ARC region to explore common factors. We end with a discussion of the cross-cutting economic benefits of the Appalachian Development Highway System. Socioeconomic Transition in the Appalachia Coal Region Some Factors of Success 54 Information about the counties comes from The story of Athens County’s success primary sources through key informant appears to be heavily tied to the presence of interviews,57 quantitative population a public university within the county. Athens and survey data58, and secondary data County is home to Ohio University, a public, sources such as official government websites, four-year college; Hocking College a public, published reports, and media articles. Given two-year college; and an Ohio Technical the current COVID-19 pandemic, locating and Center. Ohio University’s rising enrollment contacting individuals proved challenging, more than doubled from about 8,000 students especially in Sequatchie, as its website is in 1960 to over 19,000 by 1970. In the course limited and out of date. We asked informants a of our case-study work, we interviewed Dr. series of questions about factors affecting their Jason Jolley, Professor of Rural Economic county that could help explain the transition Development and Director of the Masters of (for example, distance/infrastructure Public Administration at Ohio University. His issues, local economic structure, population summary of Athens County’s transition from characteristics, institutions/government, and coal employment is as follows: social capital networks such as collaboration within the county and with outside entities). “Brick-making from clay and coal mining were two major employers in Athens County historically. However, after World War II Athens County, Ohio there was growth in the city of Athens [the county seat], as the GI benefits led to influx Athens County, Ohio, the first community of enrollment at Ohio University (OU) that that successfully transitioned from coal then lead to the campus boom and growth. mining, began its transition in the late This is also about when Athens County saw 1950s and early 1960s, when the coal mining its decline in coal employment, and again share of the labor force fell below 2 percent, with the Vietnam War. Those are the two and ultimately fell to zero. Its population of periods of growth in the county, and that 54,889 persons in 1970 grew to 65,936 by 2018, was largely driven by enrollment at the during which time mean household income university. Simultaneously, technological rose from $56,889 to $76,771 (constant $2018). advances in the 1950s and 60s opened up Despite these income gains, the household new areas of coal mining, that required less poverty rate increased from 13 percent in 1970 employment in coal. So, the story is that to 17 percent in 2018. As of 2017, the county’s the University (OU) drove the economic main sectors of employment were educational diversification, and then the technological services (28 percent of employed residents), changes made it easier to get to coal, healthcare and social assistance (12.9 percent), allowing for decline in coal employment and retail trade (11.4 percent) (Data USA 2020). (Jolley 2020).” 57  We selected informants through a search of county government websites, federal/state/county-funded Extension Service offices (agents who provide development assistance to counties), and library and media outlets. We also asked each key informant to recommend other knowledgeable people. The key informants include county commissioners, mayors, Extension agents, local business people, Chamber of Commerce, directors of public-private economic development associations, local librarians, and researchers. 58  Decennial U.S. Census, the American Community Survey (ACS), and other secondary sources provide county-level data on socioeconomic well-being and other conditions over the past half century (1950-2018). 55 The nexus of the university, the ARC’s politically liberal than other counties within regional economic development district Ohio’s Appalachian region. Dr. Jolley linked (Buckeye Hills Regional Council), and this political liberalism with greater focus small businesses and nonprofits within on promoting environmentally sustainable Athens county reflects a high degree of practices and concern about the negative administrative institutional capacity to impacts of coal mining. For example, in pursue economic development strategies partnership with Ohio University, Athens and diversification (Jolley 2020). These County received an Environmental Protection development partnerships have benefited Agency (EPA) water stream restoration grant Athens County: Mike Jacoby (President of to clean county streams degraded from coal the Appalachian Partnership for Economic waste. Growth) noted that a number of new companies “grew out of the university— Location and natural amenities have also Global Cooling, Quidel (formerly Diagnostics played a role in Athens’ success. Athens Hybrids), and RXQ Compounding—… [these] has had easy access to its nearest major came out of the university business incubator metropolitan area, Columbus via U.S Route or because of the university talent.” Further, 33, which was constructed in 1938 and runs cross-county partnerships resulted in a through the county. Between Athens’ County retraining program administered by the seat (also named Athens) and the Columbus Hocking-Athens-Perry Community Action I-270 beltway, US 33 is mostly four lanes, much Program (HAPCAP), which helped miners of limited access. The county is also home to and their spouses obtain decent-paying jobs some recreation and tourism amenities such elsewhere after the last coal mine in the area as Wayne National Forest. closed in 2002 (Harris 2020). Athens has been active in securing ARC funding, and developed innovative ARC projects, such as those that Noble County, Ohio provide philanthropic support and programs for youth, some of which use local campus Noble County, Ohio, began its transition space (ARC 1999). Recent ARC Power Grant before 1980, when the coal mining share of funding data shows that the county received the labor force fell below 2 percent. Nobel 7 grants out of a total 23 received by Ohio County’s population was 10,428 in 1970 and counties (ARC 2020). increased to 14,443 by 2018, during which mean household income rose from $51,231 in Ohio University also contributes 1970 to $73,906 in 2018 (constant $2018), and significantly to social capital in Athens the household poverty rate fell sharply from County. Dr. Jolley noted the strong base in 20 percent in 1970 to 9 percent in 2018. The the county/city of Athens because of Ohio principle sectors of employment in 2017 were University, which motivates the community’s healthcare and social assistance (16.9 percent), greater willingness to invest in the arts, manufacturing (14.1 percent), and retail trade community events, and local culture. (12.8 percent) (Data USA 2020). Additionally, there is greater support for and investment in local businesses and the local Noble County benefitted significantly from food system. Finally, due to the University, the building of a state prison in the area. the population in Athens County is more We interviewed the retired Community Socioeconomic Transition in the Appalachia Coal Region Some Factors of Success 56 Development Director of Noble County One example is the Mahle Engine Plant – Extension, Edwin Lloyd, who served the still operational when Noble County began county during the transition period. When transitioning away from coal employment. asked about what contributed to the county’s According to Gwynn Stewart, current move away from coal, former Director Lloyd development director of County Extension, said: the plant provided some good local jobs. In terms of current economic development “A brief sketch of Noble county first includes opportunities, Director Lloyd noted that the surprise that Noble has consistently although Noble County, along with other grown, because it is still one of the smallest counties in the region, have benefited from counties in the state in terms of population. manufacturing plants since the decline of coal However, as to how the county has achieved mining, many factories have since closed and that – in the 1960s the interstate was built moved abroad, where production and labor through the county (I-77) and it created costs are lower. The community has not in fact an opportunity for the county to develop entirely abandoned coal; B&N Coal still has a at least a little as a bedroom community small operation in Noble County today. for other counties in the region with more employment opportunities, and as a Geography has had some positive impact on result most of the residents work outside diversification and growth but also creates of the county. The main real change that challenges. Interstate-77 passes through happened was when the state prison (the the county, allowing residents to commute Noble County Correctional Institution) was elsewhere. Wayne National Forest is also opened in the county in 1996. The prison located in parts of Nobel County, attracting employs 396 people, which contributed to tourism and recreational activities. But lack of Noble County’s growth during the period.59 infrastructure limits the ability to recruit and The facility has a population of just under expand industries within the county. Former 2,500 inmates, and they would also be Community Development Extension Director included in the population of Noble County Lloyd noted that water and sewer systems are under the US Census “group quarters.” As limited and would likely be unable to support to how the prison came to be located in the large-scale development, and cellphone and community – some groups, such as county broadband services remain limited due to the commissioners and state legislatures, county’s mountainous terrain. worked together. The county bought the land and then gave it to the state to build the Noble County’s record on social capital prison. However, to some degree it was luck networks and collaboration is mixed. The that the prison was located in Noble County, Noble County Chamber of Commerce website but it remains a major employer in the indicates that its development efforts began county (Lloyd 2020).” formally in 1971, and that collective efforts have been ongoing. As noted, county groups Manufacturing provided another pathway had worked with the State Legislature to bring for the county’s economic diversification. a prison to the county. Nevertheless, there are 59 Prison data from the Ohio Department of Rehabilitation and Correction, [downloaded from: https://drc.ohio.gov/nci. on May 19,  2020]. 57 many challenges to the community’s economic and while its population has grown as mining wellbeing, according to Director Lloyd, such as declined, it continues to struggle with small size the lack of non-governmental organizations and limited infrastructure. and agencies that promote economic development. For example, he stated regional partners like United Way were not present, Laurel County, Kentucky and basic social services, such as a homeless shelter or emergency medical services, Laurel County, Kentucky, made its post-coal are limited or nonexistent. An initiative to transition between the late 1980s and the 1990s support small business development through (after increasing its coal dependence during loans was abandoned because the program the late 1970s/early 1980s coal boom). It’s 1970 was undersubscribed. Noble County has population of 27,386 more than doubled by 2018 successfully collaborated with other counties to 60,180. Mean household rose from $40,193 in to obtain ARC funding for “innovative” grants 1970 to $60,981 in 2018 (constant $2018), while projects (ARC 1999), and had received two ARC household poverty rates fell from 34 percent POWER grants by 2020. to 19 percent in the same period. By 2017, the main sectors of employment were retail trade In short, Noble County’s economic (15.4 percent), manufacturing (13.9 percent), and development limitations persist. While the healthcare and social assistance (12.9 percent) county benefitted from securing a state prison, (Data USA 2020). Socioeconomic Transition in the Appalachia Coal Region Some Factors of Success 58 The primary driver behind the growth of The county has also benefited from tourism- Laurel County and its capital (City of London) driven growth, notably as the birthplace of into a regional hub was the construction Kentucky Fried Chicken ((Colonel) Harland of two major roads, according to the City Sanders once lived in Laurel County). The of London’s Tourism Office. Interstate 75 (I- county takes pride in its “annual world chicken 75), which runs north/south, opened in 1969, festival”, which is a tourist draw. Daniel Boone and Highway 80, which runs east/west and National Forest also spans parts of Laurel is known as Hal Rogers Parkway (formerly County, offering tourism amenities. Daniel Boone Parkway), opened in 1971. Other communities east of Laurel County did not have The role of networks and collaboration the advantage of the new interstate, whereas appear to have facilitated Laurel County’s most of the counties along that corridor have turnaround. There is evidence of longstanding seen progressive growth. Other infrastructure collaboration between the county and its investments also helped facilitate economic county seat, London. In 1971, the London-Laurel development. The regional London-Corbin County Industrial Development Authority was Airport was established in the early 1970’s; formed to “secure and develop industrial parks whereas it has no scheduled airline passenger and provide new locations for companies like flights, it is among the five busiest airports in Laurel Grocery, Walmart Distribution Center, Kentucky. and Aisin Automotive Casting” (Sentinel Echo 2019). London’s mayor Troy Rudder notes: Investments made to pipe water from Laurel Lake to London meant the city had “The county’s growth is largely due to the the necessary water for factories, according cooperation of local leaders who work to a respondent from the Laurel County together to continuously improve the quality Historical Society. In 1960, Congress authorized of life for residents…the biggest part in our construction of a dam on the Laurel River. The success is that the city and the county [work] U.S. Army Corps of Engineers built the high closely together. You don’t have that in most Laurel River Dam between 1964 and 1974, cities and counties. We’ve built such a close and hydropower production began in 1977, relationship between the city and county and resulting in safe drinking water, recreational that has helped us bring businesses here and opportunities, and low-cost hydroelectric power improve the lives of everyone who lives and which ultimately helped crowd out coal and works here.” (Johnson 2020) coal-powered electricity (Laurel County 2020). Overall, Laurel County’s economic transition The county’s location and infrastructure have and resilience appears to have been driven contributed to a relatively diverse economy. by its location and infrastructure capacity. Many durable and nondurable manufacturing The presence of an interstate highway likely plants operate in London, KY, such as Aisin—a facilitated greater industrial development by Japanese producer of automobile components reducing transportation costs and increasing and systems—Bimbo Bakeries USA, Flowers access to Laurel Country by nearby commuters. Foods, and others. In 2015, commuters from This was likely aided by effective collaboration other counties represented about 55 percent of between city and county government the county’s workforce, and the county hosts officials to leverage economic development seven industrial parks. opportunities. 59 Sequatchie County, Tennessee of becoming a bedroom community to Chattanooga, offering very cost-effective Sequatchie County, Tennessee—the fourth, real estate to retirees, workers, and others and final, Appalachian successful case- on incredible sites with vistas and views study—transitioned from coal mining that are breathtaking. They worked with largely in the 1980s. Sequatchie County’s very private developers to purchase mountain-top small population of 6,331 in 1970 more than properties formerly owned by coal mine and doubled by 2018, reaching 14,730. During the timber companies to develop into residential same period, mean household income rose properties. Both the city and county worked from $42,550 to $64,711 (constant $2018), and together to focus on providing critical the household poverty rate declined from 25 infrastructure such as water, sewer, and percent to 14 percent. As of 2017, its major roads to these sites; and as such folks have sectors of employment were manufacturing continued to buy and develop residential (15.7 percent), retail trade (12.6 percent), and properties, retail has followed the growth; healthcare and social assistance (11.6 percent) and the commute from Chattanooga to (Data USA 2020). Sequatchie County takes no more than 30 minutes. I live on the north end of Hamilton Location played a key role in Sequatchie County and can be to Dunlap as quick as I County’s economic development. Part of can be in downtown Chattanooga.  They have the Chattanooga, Tennessee Metropolitan worked diligently to keep the county beautiful Statistical Area since 1973, the county’s and scenic – it provides a quaint, small town proximity undoubtedly contributed to life with amenities very close to a mid-sized its growth. Two major highways, U.S. urban area. Sequatchie County has continued Route-127 and Tennessee State Route-111, to be our fastest growing county for the intersect in Dunlap, the county seat. For our past two census counts due to the influx of case-study work, we contacted Southeast residents. The city and county have several Tennessee Development District Director Beth industries, but primarily the focus has been Jones, who has 38 years of experience working on quality of life!” in the county. She characterized Sequatchie County’s transition from coal employment as The county’s location provides recreational follows: and scenic opportunities. It is part of the Cumberland Plateau, sitting more than “Sequatchie County’s success in transitioning 1000 feet above the Tennessee River Valley away from coal has to do with its close (Sequatchie County TN 2020). proximity to Chattanooga and the construction of a major Appalachian Sequatchie County is home to Chattanooga Development Highway System (ADHS) State Technical Community College, providing Corridor, U.S. Highway-111 (known as a local source of workforce development. Corridor J), which constructed a major divided two-lane highway from the north Overall, Sequatchie County’s path of end of Hamilton County (Chattanooga), successful economic transition was providing easy and quick access across facilitated by its proximity to Chattanooga, the mountain into the Sequatchie enabling county residents to access a much Valley. Sequatchie County adopted a strategy larger and diverse job market. Socioeconomic Transition in the Appalachia Coal Region Some Factors of Success 60 Ouray County, Colorado industry that continued to support the town into the 2000s. Today, Ouray is known as Our final case study considers the non- ‘The Switzerland of America’ and is one of Appalachian Ouray County, Colorado, which Colorado’s most popular tourist destinations. transitioned from silver and gold mining Many of Ouray’s remaining historic employment, and is relatively prosperous buildings were built during the 1880s today. We include Ouray County as an example and 1890s. Ouray, Colorado is one of the of a hard-rock mining experience with similar West’s best preserved mining towns from characteristics to coal mining, but being in the the 1800s. It also happens to be one of the Western United States, it has a different set of most spectacularly beautiful destinations structural and cultural characteristics than anywhere in America.” (Ouray, Colorado Appalachia. 2020) Ouray County’s transition away from mining An important aspect of Ouray’s tourism occurred from 1970 to 1980. Ouray County’s success was the early development 1950 population of 2,103 fell to 1,537 in 1970. of highways through its surrounding Yet, by 2018, its population had tripled to 4,722. mountains. For example, the stretch of The 1970 mean household income of $49,029 U.S. Highway-550 from Ouray to Silverton increased to $87,907 in 2018 (constant $2018), is known as the “Million Dollar Highway”. and household poverty rates fell from 11 percent The highway began as the Otto Mears toll in 1970 to 7 percent in 2018. The main sectors road built in 1883 to link Ouray with the Red of employment in 2017 were retail trade (22 Mountain Mining District (Ouray, Colorado percent), accommodation and food service 2020). U.S.550 was part of the original 1926 (21 percent), and professional, scientific and federal highway system, and today the entire technical services (12.6 percent) (Data USA route is part of the San Juan Skyway Scenic 2020). Byway (Ouray County, Colorado 2020). Natural amenities have played a big role in Ouray County’s economic resilience appears Ouray County. Ouray County has a rugged to stem from its location, natural amenities mountain topography and many national/state and accessibility that make it a desirable parks and trails. The county hosts the largest tourist destination. The presence of a federal ice climbing event in North America. Additional highway likely eased its transition to a tourism industries include mountain biking, tourism-based economy. Similar to the four hiking, trail running, and off-roading in ARC case studies, Ouray County shows that four-wheel drive expeditions (Ouray County, it is not necessary to have a large population Colorado 2020). The towns of Ridgway and in order to thrive after transitioning from Ouray in Ouray County have been the locations reliance on a mining economy. for numerous films, including True Grit and How the West Was Won. According to the Western Mining History Society: “The closure of [mines in the early 1990s] spelled the end of Ouray’s mining economy, but fortunately tourism become the major 61 Figure 6.1 Map of Planned Appalachian Development Highway System in 2017 ADHS Miles Open to Traffic ADHS Miles Not Open to Traffic Interstate Highway System Source: Appalachian Regional Commission, Map of the Appalachian Development Highway System as of February 25, 2018,  https://www.arc.gov/ adhs Regional Economic Benefits final completion dates are uncertain (ARC 2020). The purpose of the ADHS is to generate of the Appalachian Development economic development in isolated areas and, by integrating the Appalachian region to the Highway System national transportation system, provide access to regional, national, and global markets. The Appalachian Development Highway System (ADHS) has been a significant source The ADHS plan has slightly expanded over of economic benefit to counties across the time. Figure 6.1 shows the system as it exists Appalachia region. The first highway system and is envisioned today. As of 2016, ARC has authorized by Congress for the express purpose received over $34 billion in federal expenditures, of stimulating ARC economic development, with the majority going towards funding the the ADHS is a network of 31 distinct highway construction of ADHS corridors. Congress corridors totaling 3,090 miles that connects appropriated $100 million in FY 2020 for the ARC region with the Interstate Highway ongoing ADHS construction (ARC 2020). System (Appalachian Regional Commission 2019; Cambridge Systems, Inc. 2008). The Since the beginning of ADHS’s construction corridors developed by the ADHS consist of a in 1965, it has led to improvements in terms mixture of state, U.S., and interstate routes. of travel efficiency and direct and indirect As of September 2019, the ADHS is 85 percent economic impacts. Travel efficiency benefits complete, with another 5 percent open to traffic have been immense, largely in the form of travel but planned enhancements not completed. time and reliability. In their analysis of travel Only 340 miles are unconstructed, though time saved, Economic Development Research Socioeconomic Transition in the Appalachia Coal Region Some Factors of Success 62 Group, Inc. estimates that “the ADHS now saves Sequatchie County, TN are connected to the 231 million hours per year of travel each year, ADHS through corridor J. Although the ADHS representing 632,000 hours per day, compared does not intersect with Noble County, OH, it to what would have been without the ADHS. In benefits from being crossed by U.S. Interstate 77. addition to the hours of time saved by increased travel speeds and shorter routes on the ADHS, it • Athens County has benefited from its is estimated that 129 million hours of ‘reliability proximity to Corridor D (U.S. Highway 50), time’ are saved annually, representing the major east-west route in southern Ohio another 360,000 hours saved daily” (Economic (Appendix F, Figure 1). While already linked to Development Research Group, Inc. 2017, p. 5). Columbus, OH through U.S. Highway 33, the introduction of Corridor D facilitated shorter Besides saving travel costs, the ADHS travel times to other key metropolitan areas has substantially helped Appalachian such as Cincinnati, OH and Parkersburg, WV.60 communities access goods and labor markets. The role of transport networks in facilitating • Laurel County has benefited from Corridor the movement of goods and people is widely J, which intersects the county and whose recognized as a key contributor to economic northern terminus is Interstate 75 near development and growth. The ADHS has London, KY, the county seat (Appendix promoted labor force accessibility in areas F Figure 2). This is noteworthy because remote from major population centers, as well as Interstate 75 is a critical transport route in economically distressed counties (Economic for U.S. manufacturing supply chains, Development Research Group, Inc. 2017). especially for the auto industry. Completed Investments in the ADHS have enabled the ARC in 1984, Corridor J runs through Kentucky region to attract businesses that spur economic counties Laurel, Pulaski, Wayne, Clinton, growth; added business sales are estimated and Cumberland. It is estimated to have at over $24 billion annually, and added gross generated 3,785 additional jobs, $140 million regional product is an estimated $11 billion per in additional income, and a 5% drop in annual year ($2015) as a result of the ADHS (Economic vehicle-miles-traveled, and a 29% cut in Development Research Group, Inc. 2017). Not drive time (Econ Works 2020; Wilber Smith all counties benefited equally from the ADHS, Associates 1998). which suggests that connectivity infrastructure addresses only part of the underlying challenges • Sequatchie County is located near the facing the Appalachia region. Corridor J terminus in Chattanooga (see Appendix F Figure 3). The section of Corridor J The ADHS corridors have benefitted our between Dunlap (Sequatchie’s county seat) and case-study communities. Among our four US 27, built between 1988 and 1994 (Tennessee Appalachian communities, three counties— State Route 111, 2020), provides an additional Athens, OH; Laurel, KY; and Sequatchie, commuting route into Chattanooga, whose TN—are located on sections of an ADHS metropolitan area had a 2019 population corridor. Athens, OH is connected to the ADHS of 563,000 (compared to 183,000 within through corridor D, while both Laurel, KY and Chattanooga’s city limits; U.S. Census Bureau). 60 According to some of the region’s residents, “With the highway in place, people in Appalachian Ohio can work the next county over,  or drive to Athens or Cincinnati for work or recreation – time-consuming or impossible before the highway was built” (Fugleberg 2016). 63 SECTION 7 Conclusions and Policy Implications The aim of this study is (a) to identify Appalachian counties that have been relatively successful at transitioning from coal-based economies, and (b) to identify factors that promoted this transition. The rise and decline of U.S. coal production over the last century had the effect of building mining-dependent economies across the remote counties of Appalachia through coal-based employment, and then exposing these communities to severe negative shocks as demand for coal dried up. Many of these relatively small and remote communities were already lagging in terms of socio-economic development; the sharp reduction in coal demand over the past several decades created further dislocation that undermined their economic sustainability. The magnitude of the impact across Appalachian counties varies, as does the degree of economic resilience and capacity to transition to alternative industries. Some counties transitioned early, while others delayed the difficult and costly adjustment needed to diversify their local economies. Socioeconomic Transition in the Appalachia Coal Region Some Factors of Success 64 To understand these differences, we • In terms of economic well-being, the four draw on evidence from the traditional counties’ level of success is modest. While economics literature, as well as broader the four counties have grown in population research disciplines related to economic and diversified production, and most have geography, demography, sociology, and experienced significant poverty reduction, community capital to identify the main average household incomes are relatively low variables associated with economic and poverty rates exceed both the national transition and local economic development and ARC averages. Median household most relevant to the Appalachian income in the four successful coal-transition experience. We use these variables to counties is below the top-10 fastest growing develop a measurement strategy to mining counties, as well as the 10 slowest compare transition performance across growing mining counties. Moreover, the the 420 ARC counties and select the best highest-performing and poorest-performing performers. mining-intensive counties had lower initial poverty rates than the four successful Our quantitative analysis identifies counties, and still have lower poverty rates four Appalachian counties that have today. been most successful in growing their communities while eliminating • Severe structural impediments across dependence on coal jobs: Athens and Noble Appalachia constrain growth. Being Counties in Ohio; Laurel County, Kentucky; small and remote, most ARC counties have and Sequatchie County in Tennessee. We limited access to labor markets that offer complement our quantitative findings with more and diverse job opportunities. ARC qualitative case-study research on each counties have low levels of physical capital— “successful” Appalachian county to try especially infrastructure—and human to understand the factors associated with capital, with lower educational attainment successful transition. and lower quality education and health services. High transport costs, both for The following main conclusions emerge bringing in intermediate and final goods from our combined quantitative and and for exporting outputs, reduce firms’ qualitative analysis: competitiveness and raise the cost of living for residents. • Very few counties manage a positive transition. Out of 222 ARC counties • Non-structural impediments reinforce with a high level of coal dependence at poor economic outcomes and reduce some point after 1950, only four counties local economic resilience. Historically managed to transition out of coal coal-dependent communities exhibit less and remain economically viable with economic diversification; more uniform, sustained population growth. This result low-skill job profiles; and local labor is surprising in light of global trends market wage distortions (driven by high- in coal’s declining competitiveness, paying mining jobs). The modest share of especially Appalachian coal, and sharply manufacturing activity—which tends to lower market demand. serve larger, non-local markets and raises average productivity and wages compared 65 to resource or service-based sectors—also which provides ongoing economic increases the economy’s vulnerability to stimulus. It also benefited from good coal price shocks, thereby locking-in path- highway connections. dependence. The patterns of boom and bust cycles and low levels of entrepreneurship ° Laurel County’s relative success at reflect limited flexibility to adapt to diversifying from coal is largely due to changing market realities. major infrastructure improvements. These included construction of two • Institutional capacity and social capital highways, the regional London-Corbin matter. Local government capacity and Airport, a major dam, and investments social capital are generally low across the in piped water and hydroelectric power. ARC region compared to national averages, These investments connected the county as measured by common proxy measures. to key manufacturing supply chains, and This may explain why ARC counties were helped it exploit some unique tourism lagging even prior to coal transition. The opportunities. case studies highlight examples where local government capacity to design, finance, and ° Sequatchie County found alternative implement economic development initiatives post-coal economic stimulus through and attract external financing—especially in its proximity to Chattanooga, TN and collaboration with local civil society agencies investments in highways to access this and regional government authorities— large metropolitan market offering diverse seems to help sustain the transition impetus. job opportunities. • Transition paths in our four “successful” Are there any common features or lessons Appalachian counties reflect idiosyncratic, that can help policymakers and local unique features that drove their success: community leaders promote transition toward a sustainable post-coal economy? We ° Athens County’s economic development do not offer specific policy recommendations, has centered around its large public because this would require a comprehensive Ohio University, which created high- assessment of county, regional, state, and skilled teaching, administrative, and federal policies in the ARC region and how these health services jobs and stimulated local have affected outcomes. Detailed research on businesses to provide goods and services the incentive framework that steers producers’ to the large student population. The and workers’ decisions, and the fiscal stance university has also fostered innovative and public investment decisions by particular graduate or affiliated start-ups, generated counties or states, is beyond the scope of our social capital by coalescing community research. Nor do we examine regional and members’ interests around enhancing national coal-specific policies and incentives the university community, and has raised relevant to the local policy context. institutional capacity within the county. Our analysis points to some broad ° Noble County, despite its small approaches policy makers should consider population, was able to attract a large to address common economic development public investment to build a state prison, impediments in Appalachia: Socioeconomic Transition in the Appalachia Coal Region Some Factors of Success 66 • Enhance connectivity: For remote that the ARC region is filled with distressed communities, connectivity to larger counties that also received significant markets is essential for growth, although government investments, beginning with infrastructure investments alone are the Tennessee Valley Authority (TVA) in the not enough to guarantee successful 1930s, and later through ARC and POWER community transition from coal dependence. programs and widespread ADHS road Connectivity is key to achieve scale investments throughout the region. economies and competitive margins for locally produced goods and services, as • Invest in human capital: Enhancing human well as for matching excess labor supply capital through investments in education with jobs in nearby markets. Even when and health services quality and/or access remoteness represents an asset by virtue of will improve residents’ well-being, even natural amenities, deriving economic value in the absence of a market return on their from these amenities—such as through improved human capital. Similar to physical recreational services and tourism—requires infrastructure, increasing human capital non-residents to be able to access these makes workers more productive and allows communities. Infrastructure that supports them to compete for better, higher-skilled this connectivity—whether roads or internet jobs and generate more added value. But digital connectivity—is necessary but not in the absence of employer labor demand, sufficient to support diversified economic human capital investments will generate development. This is illustrated by the fact negligible returns, and may raise job 67 seekers’ expectations for skilled jobs that for providing public goods consistent with are not available locally. Lessons from economic development, and they must successful investments in education and design and enforce the right mix of rules training require aligning the curricula with and regulations. Governments can also identified private sector needs or focusing on play an active role in stimulating private developing entrepreneurial capacity for new sector activity, such as through investment business start-up. incentives, public-private partnerships (PPP), local-content procurement thresholds, • Seek economic diversification to ease and strategic planning to promote priority “boom and bust” cycles: The long- sectors, to name a few options. Coordinating term economic health of coal-dependent with local stakeholders and regional communities requires that government agencies and collaborating with neighboring authorities and civil society make concerted communities or state authorities to design efforts to facilitate new economic activities complementary approaches can maximize and move beyond the coal “boom-bust” cycle. economic impacts and potentially leverage Resource-dependent economies are subject more financial and/or institutional support. to volatility due to market fluctuations in This approach appears to have been effective the value, and therefore profitability, of their in successful transition counties, which are resource extraction and production activities. part of ARC-established, multi-county local We illustrated how volatile coal “boom- economic development areas. bust” cycles in Appalachia affected county- level mining employment. Dependence on • Coordinate economic development commodity-based sectors is associated strategies: Ultimately, counties that with long-term economic and social costs, are economically constrained need a such as reduced economic activity and coordinated set of economic development entrepreneurship, slower average economic strategies and approaches to (a) foster larger growth, and weak institutional capacity economic agglomeration and/or linkages to to meet local social needs (Betz et al. 2015, larger regional/national/global markets, and Billings and Blee 2000, Duncan 2014). The (b) exploit natural amenities in a sustainable promise of a future new “boom” cycle delays manner that can attract demand for local difficult decisions, especially those that services. Some communities may not have entail short-term costs, such as job losses. the minimum endowments to remain, or Attracting investments in new firms that become, economically viable in the absence serve a local or regional market niche, or that of coal. Policymakers at the state and reach more distant markets seeking their national levels will need to weigh the fiscal products or services—which in turn depend cost of propping up very small communities on local inputs—can stimulate local job against the opportunity cost of not investing creation and value addition. in communities with greater economic potential. This will require a hard look at • Build local institutional capacity: communities’ long-term sustainability, and Diversifying sectors of economic activity consideration of alternative approaches, requires effective local institutional such as facilitating out-migration through capacity to develop an enabling business mobility grants or other incentives. environment. Governments are responsible Socioeconomic Transition in the Appalachia Coal Region Some Factors of Success 68 References Aldrich, Daniel P. and Michelle A. Meyer. 2015. Black, D., McKinnish, T., Sanders, S., 2005. The “Social Capital and Community Resilience.” American economic impact of the coal boom and bust. The Behavioral Scientist 59 (2):254-269. Economic Journal 155, 449–476. Allard, Scott W. 2017. Places in Need: The Changing Blank, Rebecca. 2005. “Poverty, Policy, and Place: Geography of Poverty. New York: Russell Sage How Poverty and Policies to Alleviate Poverty Are Foundation. Shaped by Local Characteristics.” International Regional Science Review 28(4):441–64. Appalachian Citizens’ Law Center, Appalachian Voices, Coalfield Development Corporation, Rural Brown, David L. and Kai A. Schafft. 2011. Rural Action, and Downstream Strategies. 2019. “A New People and Communities in the 21st Century. Malden, Horizon: Innovative Reclamation for a Just Transition.” MA: Polity. Reclaiming Appalachian Coalition. Cambridge Systematics, Inc., Economic Appalachian Regional Commission (ARC). 1999. Ideas Development Research Group, Inc., and HDR that Work http://www.arc.gov/images/newsroom/ Decision Economics. 2008. “Economic Impact publications/Ideas_That_Work_1999.pdf Study of Completing the Appalachian Development Highway System.” Final Report. Washington, D.C.: Appalachian Regional Commission (ARC). 2020. Appalachian Regional Commission. Power Award Summaries by State. https://www. arc.gov/images/grantsandfunding/POWER2020/ Carley, Sanya, Tom P. Evans, and David M. ARCPOWERAwardSummariesbyStateFebruary2020.pdf Konisky. 2018. “Adaptation, Culture, and the Energy Transition in American Coal Country.” Audia, Pino. G. and Atul Teckchandani. 2010. The Effect Energy Research & Social Science 37 (March): of Connected and Isolated Voluntary Associations on 133–39. Economic Activity in the United States, 1984-2000. Social Science Research, 39 (6) 1153-1163. Census History Staff - US Census Bureau. Various Years. “Metropolitan Areas - History - U.S. Besser, Terry L., Nicholas Recker, and Kerry Agnitsch. Census Bureau.” Various Years. https://www. 2008. “The Impact of Economic Shocks on Quality of census.gov/history/www/programs/geography/ Life and Social Capital in Small Towns.” Rural Sociology metropolitan_areas.html. 73 (4): 580–604. Centers for Disease Control, The National Betz, Michael R, Mark D. Partridge, Michael Farren, Institutes for Occupational Safety and Health. and Linda Lobao. 2015.“Coal Mining, Economic 2018. “Prevalence of Black Lung Continues to Development, and the Natural Resources Curse.” Energy Increase among U.S. Coal Miners.” Available Economics 50:105-116. at: https://www.cdc.gov/niosh/updates/upd- Billings, Dwight B. and Kathleen M. Blee. 2000. The 07-20-18.html#:~:text=A%20Steady%20 Road to Poverty: The Making of. Wealth and Hardship Increase&text=Recent%20NIOSH%20 in Appalachia. New York: Cambridge University Press. research%20indicates%20an,CWHSP%20in%20 the%20late%201990s. Downloaded July 24, 2020. Bishop, Bill. 2017. “Federal Spending in Rural Lags Cities.” The Daily Yonder. https://www.dailyyonder.com/ Cook, Ak. 1995. “Increasing Poverty in Timber- federal-spending-rural-lags-cities/2012/03/08/ Dependent Areas in Western Washington.” Society & Natural Resources 8 (2): 97–109. 69 Cotter, David A. 2002. “Poor People in Poor Places: Economic Development Research Group, Inc. Local Opportunity Structure and Household Poverty.” 2017. “Appalachian Development Highway System Rural Sociology 67(2):534–55. Economic Analysis Study: Synthesis of Findings to Date.” Technical Report. Boston, MA: Appalachian Cowan, Tadlock. 2012. Military Base Closures: Regional Commission. https://www.arc.gov/assets/ Socioeconomic Impacts. Congressional Research Service: research_reports/ADHSEconomicAnalysisStudy 1-9. SynthesisofFindingstoDate.pdf. Daniels, S. E., C. L. Gobeli, and A. J. Findley. 2000. “Econ Works.” 2020. https://planningtools. “Reemployment Programs for Dislocated Timber transportation.org/13/econworks.html. Workers: Lessons from Oregon.” Society & Natural Resources 13 (2): 135–50. (EIA) U.S. Energy Information Agency, 1999. Annual energy outlook 2000with projections to 2020. Available Data USA. 2020. https://datausa.io/profile/geo. at: http://www.eia.gov/forecasts/aeo/archive.cfm Accessed April 29, 2020. (downloaded January 15, 2014). Deaton, J.B., Niman, E., 2012. An empirical examination (EIA) U.S. Energy Information Agency, 2005. Annual of the relationship between mining employment and energy outlook 2005. Available at: http://www.eia.gov/ poverty in the Appalachian region. Appl. Econ. 44, forecasts/aeo/archive.cfm (downloaded January 15, 303–312. 2014). Dimitri, C., A. Effland, and N. Conklin. 2005. “The 20 th (EIA) U.S. Energy Information Agency, 2013. Annual Century Transformation of U.S. Agriculture and Farm energy outlook 2014 early release. Available at: Policy.” U.S. Department of Agriculture, Economic (downloaded January 15, 2014). Research Service, [Available at: https://www.ers. usda.gov/publications/pub-details/?pubid=44198. (EIA) U.S. Energy Information Agency, 2014. U.S. Downloaded July 16, 2020.] Coal supply, consumption, and inventories. EIA short-term energy forecasts (January 7, 2014) Doan, D. Paddock, B., and Dyer J. (undated mimeo). Available at: http://www.eia. gov/forecasts/steo/ “Grain Transportation Policy and Transformation tables/?tableNumber=18#. (downloaded January 15, in Western Canadian Agriculture.” Agriculture 2014). and Agri-food Canada. (Available at: https://www. semanticscholar.org/paper/The-reform-of-grain- (EIA) U.S. Energy Information Agency. Annual Coal transportation-policy-and-in-Doan-Paddock/ Report. Various years. https://www.eia.gov/coal/ b7c0b228313a692fa439c0723a93ec88c418229a?p2df. annual/pdf/acr.pdf downloaded May 25, 2020). (EIA) U.S. Energy Information Agency. 2019. U.S. Douglas, Stratford and Anne Walker. 2017. "Coal Coal Supply, Consumption, and Inventories. EIA Short- Mining and the Resource Curse in the Eastern United Term Energy Outlook. https://www.eia.gov/outlooks/ States.” Journal of Regional Science 57(4): 568-590. steo/data/browser/#/?v=18&f=A&s=&start=1998& end=2020&map=&ctype=linechart&maptype=0&id Duncan, Cynthia. 2014 Worlds Apart: Poverty and Politics =&linechart=CLPRPUS_TON in Rural America. Yale University Press. Faggian, A., M.R. Olfert, and M.D. Partridge. 2012. “Inferring Regional Well-being from Individual Revealed Preferences: The ‘voting with your feet’ approach.” Cambridge Journal of Regions, Economy and Society. 5: 163-180. Socioeconomic Transition in the Appalachia Coal Region Some Factors of Success 70 Farren, Michael D., and Mark D. Partridge. 2015. Hatt, Rod and Charles Mann 2015. “Operational “Reevaluation of the Impact of Coal Mining on the Considerations When Burning Higher-Chlorine Coal.” Virginia State Budge.” 2015: Appalachian Research Power. January 31, 2015. (downloaded from https:// Initiative for Environmental Science. https://kb.osu. www.powermag.com/operational-considerations- edu/handle/1811/81431. when-burning-higher-chlorine-coal/ on May 19, 2020). Freudenburg, William R. 1992. “Addictive Economies: Extractive Industries and Vulnerable Localities in a IMPLAN. 2011. IMPLAN Economic Impact Analyses Changing World.” Rural Sociology 57(3):305–32. Software (version Industry Code 21 - Mining Coal). Freudenburg, William R., and Lisa J. Wilson. Institute for Health Metrics and Evaluation. n.d. 2002. “Mining the Data: Analyzing the Economic “United States Mortality Rates by County 1980-2014.” Implications of Mining for Nonmetropolitan Regions.” Accessed August 17, 2020. http://ghdx.healthdata.org/ Sociological Inquiry 72 (4): 549–75. record/ihme-data/united-states-mortality-rates- county-1980-2014. Fugleberg, Jeremy. 2016. “Ohio 32: A Road of Unintended Consequences.” Cincinnati Enquirer, 2016. Isserman, Andrew . 2007. “Getting State Rural https://www.cincinnati.com/story/news/2016/10/15/ Policy Right: Definitions, Growth, and Program ohio-32-road-unintended-consequences/91496144/. Eligibility.” Journal of Regional Analysis and Policy 37:73–79. Glasmeier, Amy K. 2002. “One Nation Pulling Apart: The Basis of Persistent Poverty in the USA.” Progress in Isserman, Andrew, Edward Feser, and Drake Warren. Human Geography 26(2):155–73. 2009. “Why Some Rural Communities Prosper While Others Do Not.” International Regional Science Review Graff, Michelle, Sanya Carley, and David M. Konisky. 32 (3): 300-342. (An earlier more detail version was 2018. “Stakeholder Perceptions of the United States published in the 2007 Report to the Office of Rural Energy Transition: Local-Level Dynamics and Development, USDA http://www.jj0955.com/PdfFiles/ Community Responses to National Politics and Policy.” IssermanFeserWarren_070523_RuralProsperity.pdf) Energy Research & Social Science 43 (September): 144–57. Johnson, Kenneth, M. Johnson, John P. Pelissero, David B. Holian and Michael T. Maly. 1995. “Local Haggerty, Julia H., Mark N. Haggerty, Kelli Roemer, and Government Fiscal Burden in Non-Metropolitan Jackson Rose. 2018. “Planning for the Local Impacts of America.” Rural Sociology 60 (3):381-398. Coal Facility Closure: Emerging Strategies in the U.S. West.” Resources Policy 57 (August): 69–80. Johnson, Nita. 2020. “Laurel County Leaders Proud of New Business Development In 2019.” Corbin Times- Haggerty, Julia. 2014. “Long-Term Effects of Income Tribune. January 6, 2020. https://www.thetimestribune. Specialization in Oil and Gas Extraction: The U.S. West, com/news/local_news/laurel-county-leaders-proud- 1980-2011.” Energy Economics 45: 186–95. of-new-business-development-in/article_7a8b0534- Haggerty, Mark. 2019. “Communities at Risk from 1c32-505b-8690-ebaa919dd957.html Closing Coal Plants.” Working Paper. Bozeman, MT: Jolley, Jason. 2020. Athens County Transition from Coal Headwaters Economics. Employment. Phone Interview. 71 Jolley, Jason, Christelle Khalaf, Gilbert Michaud, Lobao, Linda, and Gregory Hooks. 2003. “Public and Austin Sandler. 2019. “The economic, fiscal, and Employment, Welfare Transfers, and Economic Well- workforce impacts of coal-fired power plant closures Being across Local Populations: Does a Lean and Mean in Appalachia Ohio.” Regional Science Policy Practice, 11: Government Benefit the Masses?” Social Forces 82 (2): 403-422. 519–56. Jones, Beth. 2020. Sequatchie County Transition from Lobao, Linda and Paige Kelly. 2020. “Local Coal Employment. Email Interview. Governments across the Rural-Urban Continuum: Findings from a Recent National County Government Kelsey, Timothy, Mark Partridge, Nancy White. 2016. Study.” State and Local Government Review “Unconventional Gas and Oil Development in the (forthcoming). United States: Economic Experience and Policy Issues.” 2016. Applied Economic Perspectives and Policy 38 (2): Lobao, Linda, Wilner P. Jeanty, Mark Partridge, and 191-214. David Kraybill. 2012. “Poverty and Place across the United States: Do County Governments Matter to the “Kentucky Coal Facts - Kentucky Energy and Distribution of Economic Disparities?” International Environment Cabinet.” 2016. https://eec.ky.gov/ Regional Science Review 35 (2): 158-187. Energy/News-Publications/Pages/Coal-Facts.aspx. Lobao, Linda, Mark Partridge, Minyu Zhou, and Krugman, Paul, 1991. "Increasing Returns and Michael Betz. 2016. “Poverty, Place and Coal Economic Geography." Journal of Political Economy, Employment in a New Economic Era.” Rural Sociology University of Chicago Press, vol. 99(3), pages 483-499. 81 (3):343-386. “Laurel County, Kentucky.” 2020. In Wikipedia. London Laurel Economic Development Authority. https://en.wikipedia.org/w/index.php?title=Laurel_ 2020. https://www.lookatlondonky.com/ Accessed County,_Kentucky&oldid=950105254 April 29, 2020. Lazarus, W.F., D.E. Platas, and G.W. Morse 2002. Moller, Stephanie, Arthur S. Alderson, and Francois “IMPLAN’s Weakest Link: Production Functions or Nielsen. 2009. “Changing Patterns of Income Regional Purchase Coefficients?” Journal of Regional Inequality in U.S. Counties.” American Journal of Analysis and Policy. 32 (1): 33-49. Sociology 114:1037–1101. Lichter, Daniel T. and Lisa A. Cimbaluk. 2012. “Family Moretti, Enrico. 2012. The New Geography of Jobs. Change and Poverty in Appalachia.” Pp. 81-105 in Boston: Houghton Mifflin Harcourt. Appalachian Legacy: Economic Opportunity after the War on Poverty edited by James P. Zilliak. Washington, D.C.: Morris, Adele Noah Kaufman and Siddhi Doshi. 2019. Brookings. The Risk of Fiscal Collapse in Coal Reliant Communities. Center on Global Energy Policy, Columbia University. Lloyd, Edwin Michael. 2020. Noble County Transition New York https://www.energypolicy.columbia.edu/ from Coal Employment. Phone Interview. sites/default/files/file-uploads/RiskofFiscalCollapse Lobao, Linda. 2004. “Continuity and Change in Place inCoalReliantCommunities-CGEP_Report_080619.pdf Stratification: Spatial Inequality and Middle-Range National Academy of Sciences, Engineering and Territorial Units.” Rural Sociology 69(1):10–30. Medicine. 2007. Coal, Research and Development to Support National Energy Policy. National Academy Press: Washington, DC. (downloaded from: https://www.nap. edu/catalog/11977/coal-research-and-development- to-support-national-energy-policy on May 19, 2020). Socioeconomic Transition in the Appalachia Coal Region Some Factors of Success 72 Noble County Chamber of Commerce. 2020. https:// Pöschl, Caroline and Barry R. Weingast. 2015. “A www.noblecountychamber.com/ Accessed April 29, Fiscal Interest Approach to Subnational Government 2020. Incentives.” Pp. 163- 178 in Is Decentralization Good for Development? Perspectives from Academics and Policy Ohio Development Services Agency. “County Trends.” Makers, edited by Jean-Paul Faguet and Caroline 2020. https://www.development.ohio.gov/reports/ Pöschl. Oxford, UK: Oxford University Press. reports_countytrends_map.htm. Accessed April 29, 2020. Potera, Carol. 2019. “Black Lung Disease Resurges in Appalachian Coal Miners.” American Journal of Nursing. Ohio Department of Natural Resources. n.d. “Ohio (119): 14. Mineral Reports.” Accessed August 19, 2020. https:// ohiodnr.gov/wps/portal/gov/odnr/discover-and- Putnam, Robert D. 2010. Bowling Alone: The Collapse learn/safety-conservation/about-ODNR/geologic- and Revival of American Community. New York: Simon & survey. Schuster. “Ouray, Colorado.” 2020. Western Mining History Reese, Laura. A. and Raymond A. Rosenfeld. 2002. The Society. 2020. https://westernmininghistory.com/ Civic Culture of Local Economic Development. Thousand towns/colorado/ouray/. Oaks: Sage Publications. “Ouray County, Colorado.” 2020. Wikipedia. April 10, Rupasingha, A., Goetz, S. J., & Freshwater, D. (2006, 2020. https://en.wikipedia.org/wiki/Ouray_County,_ with updates). The Production of Social Capital in U.S. Colorado. Counties.”  Journal of Socio-Economics, 35, 83–101. doi:10.1016/j.socec.2005.11.001. https://aese.psu.edu/ Partridge, Mark D. 2010. "The Dueling Models: NEG nercrd/community/social-capital-resources vs Amenity Migration in Explaining U.S. Engines of Growth." Papers in Regional Science. 89: 513-536. Sentinel Echo. 2019. “Industrial Authority now has a new name”, Sentinel Echo December 30, 2019, London, Partridge, Mark D. and Dan S. Rickman. 2006. The KY. Geography of American Poverty: Is There a Need for Place- Based Policies? Kalmazoo, MI: Upjohn. “Sequatchie County TN.” 2020. https:// sequatchiecountytn.gov. Accessed April 29, 2020. Partridge, Mark D. and Dan S. Rickman. 2008. “Distance from Urban Agglomeration Economies and Snyder, Brian F. 2018. “Vulnerability to Rural Poverty.” Journal of Regional Science 48: 285–310. Decarbonization in Hydrocarbon-Intensive Counties in the United States: A Just Transition to Avoid Post- Partridge, Mark D. and M. Rose Olfert. 2011. “The Industrial Decay.” Energy Research & Social Science Winners' Choice: Sustainable Economic Strategies for 42 (August): 34–43. Successful 21st Century Regions.” Applied Economic Perspectives and Policy. (33): 143-178. Sharp, Elaine B. and Steven Maynard Moody. 1991. “Theories of the Local Welfare Role.” American Journal Partridge, Mark D., Michael R. Betz, and Linda of Political Science 35(4):934-950. Lobao. 2013. “Natural Resource Curse and Poverty in Appalachian America.” American Journal of Agricultural Stedman, Richard C., Mike N. Partiquin, and John Economics 95 (2): 449–56. R. Parkins. 2012. “Dependence, Diversity, and the Well-being of Rural Community: Building on the Freudenburg Legacy.” Journal of Environmental Studies and Sciences 2:28–38. 73 Stephens, Heather and Mark D. Partridge. 2011. “Do US Census Bureau. Decennial Census of Population. small businesses matter for economic growth in Washington, DC: Government Printing Office. Various Appalachia?” Growth and Change. 42: 431-465. Years. Stephens, Heather, Mark D. Partridge, and Alessandra US Census Bureau. County Business Patterns. Faggian. 2013. “Innovation, entrepreneurship and Washington, DC. 2016. economic growth in lagging regions.” Journal of US Census Bureau. 2018. United States Census of Mineral Regional Science. 53: 778-812. Industries, 1954, Vol. 2: Area Statistics, Industry Statistics Stewart, Gwynn. 2020. Noble County Transition from for Geographic Divisions, States, and Counties. Coal Employment. Email Interview. US Census Bureau. 2012 Census of Governments – “Tennessee State Route System.” 2020. Organization. Washington, DC: Government Printing In Wikipedia. https://en.wikipedia.org/w/ Office. 2012. index.php?title=Tennessee_State_Route_ US Department of Labor. “Local Area Unemployment System&oldid=958475326. Statistics.” Various Dates. https://www.bls.gov/lau/. Tsvetkova, Alexandra and Mark D. Partridge. 2016. USDA Economic Research Service. “Natural Amenities “Economics of Modern Energy Boomtowns: Do Oil Scale.” https://www.ers.usda.gov/data-products/ and Gas Shocks Differ from Shocks in the Rest of the natural-amenities-scale/. Economy?” Energy Economics 59: 81-95. U.S. Department of Energy. Energy Information Agency Tsvetkova, Alexandra, Mark D. Partridge, and Michael (EIA). Annual Energy Outlook. Various Years. (Available Betz. 2019. “Self-Employment effects on regional at https://www.eia.gov/outlooks/aeo/. Downloaded growth: a bigger bang for a buck?” Small Business May 6, 2020). Economics 52: 27-45. Van der Ploeg, Frederick., 2011. Natural Resources: U.S. Department of Agriculture, Economic Research Curse or Blessing? Journal of Economic Literature 49 (2): Service. 2019. Rural America at a glance 2019 Edition. 366–420. Washington, D.C. https://www.ers.usda.gov/webdocs/ publications/95341/eib-212.pdf?v=5832 Voss, Paul R., David D. Long, Roger B. Hammer, and Samantha Friedman. 2006. “County Child Poverty U.S. Census Bureau, Census of Mineral Industries. Rates in the U.S.: A Spatial Regression Approach.” Washington, DC: Government Printing Office. Various Population Research and Policy Review 25:369–91. years. Wade W., C. Martin, and M. Carr. April 29, 2020. “In U.S. Census Bureau, 1975. Historical Statistics of the Global Electricity Slump, Coal Is the Big Loser.” United States from Colonial Times to 1970. Washington, Bloomberg News. https://www.bloomberg.com/news/ DC: Government Printing Office. articles/2020-04-27/in-global-coronavirus-energy- U.S. Census Bureau, Statistical Abstract of the United slump-coal-is-the-biggest-loser. (Downloaded April States. Washington, DC: Government Printing Office. 30, 2020). Various Years. Weber, Bruce, Leif Jensen, Kathleen Miller, Jane US Census Bureau. Various Years. “American Mosely, and Monica Fisher. 2005. “A Critical Review Community Survey (ACS).” The United States Census of Rural Poverty Literature: Is There Truly a Rural Bureau. Various Years. https://www.census.gov/ Effect?” International Regional Science Review programs-surveys/acs. 28(4):381–414. Socioeconomic Transition in the Appalachia Coal Region Some Factors of Success 74 Wilbur Smith Associates. 1998. “Appalachian Development Highways Economic Impact Studies - Appalachian Regional Commission.” Washington, D.C.: Appalachian Regional Commission. https://www.arc. gov/research/researchreportdetails.asp?REPORT_ ID=68. Woolcock, Michael. 1998. Social Capital and Economic Development: Toward a theoretical Synthesis and Policy Framework. Theory and Society 27(2): 151-208. 75 APPENDIX A Targeted Literature Review and Findings In this Appendix, we present the results from the targeted literature review. We explain the methodology of selecting and classifying the studies that are used to inform section 2 of this report, which addresses the factors analysts have identified as barriers or facilitators to improved community well-being in light of resource and other transitions. Socioeconomic Transition in the Appalachia Coal Region Some Factors of Success 76 Methodology: Selection of Studies parsimonious conclusions for researchers and policy-makers, we excluded literature where We used two databases to procure the that nation was not the centerpiece. Thus, necessary literature. The first database was we excluded articles on the just-transition in “Web of Knowledge.” Within this database, we European Union countries and Australia as well key-termed searched for the following: “just as articles on the impacts of coal mining and transition”; “transition from coal”; “mono- resource extraction in developing countries industry economies”; “military base closures”; within the Global South. Second, we excluded “rural economic resiliency”; and “social articles that were reviews of the literature, impacts of resource extraction.” To develop specifically if those reviews did not focus on the the pool of literature for this report, we social impacts of transitions from coal or coal constrained the surveyed literature to those employment, but rather on the available data articles written between 1992 to the present and statistical modeling techniques available and focusing on communities or impacts for assessing the effects of a transition. Third, within the United States. Results were further we excluded review articles that were mainly constrained to the following social science theoretical or conceptual. These pieces were disciplines: economics, geography, public more abstract discussions about what could or administration, sociology, and general social should be done to assist in the just transition, sciences. Next, search results were ordered by rather than providing concrete information Times Cited to ensure that the most referenced about what is being done currently. Finally, we and consequential works were reviewed. We excluded articles that focused on the successful then proceeded to review additional citations endurance and resilience of industries within from the articles we reviewed. The second communities rather than their decline and database was the Appalachian Regional communities’ responses to such declines. As a Commission’s (ARC) digital archive of reports. result, the included literature summarized in Within the research reports archive, we this report represent the major research on the searched for using the before mentioned key- topic, but due to the selection criteria and the terms. We then selected reports that focused inherent limitations of research reviews, they on assessing previous policies or programs are not exhaustive of past work. aimed at assisting Appalachian communities through the transition away from coal. Classification of Studies The resulting literature summarized in this report, therefore, mainly comprises published In Appendix A Table 1 we classify these studies refereed articles from academic journals as by four general criteria: 1) the methodology well as reports from governmental and non- employed; 2) the thematic focus of the study; governmental agencies. Our findings are based 3) the impact or outcome variable(s) examined; on 37 articles we found most relevant to the and 4) what factors matter in terms of the research questions this report aims to address. opportunities for community revitalization and However, we also reviewed approximately the barriers to community improvement. 30 additional articles that we deemed not to include for several key reasons. First, because In classifying the studies by methodology, our focus is on the just-transition in the we found several major types of research United States and because we aim to draw designs. The most commonly employed 77 research design were quantitative empirical future implementation and evaluating policies analyses. These studies mainly were previously applied. regional impact assessments typically involving regional econometric and/or The third class of studies is based on the input-output statistical models using either impact or outcome variable(s) examined secondary or primary survey data to assess by the reviewed literature. The most the community impact of industries’ decline common outcomes explored were related to or change.  The unit of analysis in these communities’ socioeconomic well-being and studies is the community and most of these included indicators such as poverty rates, studies are concerned with the impacts of employment rates, median household income, industry change across counties. Second, etc. A second type of impacts were those on the studies reviewed included qualitative individuals’ attitudes and perceptions, such empirical analyses research designs. These as acceptance of coal employment decline, studies included case-studies focused on perceptions of just transitions, and views of single or a small number of communities community change in light of shocks. The or projects, and/or interviews with key final type of outcomes explored were the community informants, and/or focus extent to which policy interventions and groups among community residents. Next, implementations accomplished their intended additional studies reviewed included policy goals. evaluations which aimed to assess policies already implemented in attempts to The final class considers the factors that revitalize communities experiencing losses mattered to the outcome variable(s) and in coal employment. The remaining studies include either opportunities or barriers. included systematic literature reviews of past Opportunities include those factors found studies to assess the current state of scholarly to have positive impacts and/or improve knowledge about the impacts of various community well-being; barriers include those industries on community well-being. Finally, factors found to have negative impacts and/ studies are classified by scale of their unit of or impede community well-being. The factors analysis, and include: cities; counties; states; that mattered, as both opportunities and and regions. barriers, were further parceled into categories of main factors (e.g., the focus of the study) To further classify the studies reviewed, and other factors (e.g., geographical location, we included three major categories for the local economic structure). thematic focus of studies. First are studies focused on the impact of employment in various industries (e.g., coal mining, oil and gas mining) on community well-being. The next classification of studies comprises event- based studies on the impact of temporal changes in the global or local economy (e.g., various shocks and boom and bust cycles) on community well-being. The last major category of studies includes those focused on policy implementation – both planning for Socioeconomic Transition in the Appalachia Coal Region Some Factors of Success 78 Overview of Table 1 Sections term decline; and perceptions of who is responsible for paying for policies to assist • Citation in the transition away from coal. ° Abbreviated Citation: First Author, Date, ° Policy Evaluations: Extent to which policy/ Title programs accomplished intended goals • Methodology • What Mattered? ° Research Design Format: Quantitative ° Opportunities: Positive impacts and/or empirical analysis (primary survey data factors that improve community well- or secondary data); qualitative empirical being analysis (focus groups, case-studies, interviews); policy evaluations; systematic ° Barriers: Negative impacts and/or factors literature reviews. that impede community wellbeing ° Scale of Study: City; county; state; region – Main Factor: See focus of studies. • Focus of Studies – Other Factors: ° Employment by Industries: Coal mining; • Geographical Location: Rural-urban oil and gas mining; general mining; location; proximity from metropolitan restoration economy; manufacturing areas ° Event: Temporal changes in global/local • Local Economic Structure: Local economy (e.g., economic shocks; boom and industrial composition; unionization; bust cycles); military base closures; just entrepreneurialism rates transition • Demographic Composition: Age ° Policy: Assessing policy intervention(s); composition; educational attainment; proposed policies expected effects racial/ethnic composition; population size/density; • Community Impact Indicators • Institutional Factors: Governmental or ° Socio-Economic Well-Being: Poverty; non-governmental amenities/services/ employment; income; social capital; policies; government capacity (fiscal population size/density; educational and/or administrative); communities’ attainment; crime rates; affordable/safe social capital; intergovernmental housing collaboration/coordination ° Individuals’ Attitudes: Acceptance of coal • Other: Temporal changes in global/ employment’s decline; perceptions of local economy (e.g., boom and bust transition away from coal; perceptions of cycles, economic restructuring who/what is responsible for coal’s long- 79 Appendix A Table 1: Targeted Literature Review, Table of Findings Citation Appalachian Citizens’ Law Center et al. 2019. “A New Horizon: Innovative Reclamation for a Just Transition.” Methodology 20 case-study communities in Virginia, West Virginia, Kentucky, and Ohio Policy evaluation Focus of Studies Policy: assessing the impacts of Abandoned Mine Lands recovery projects Community Impact or Socio-economic well-being: development of economic opportunities in three different sectors: Outcome Variable(s) 1. Recreation and ecotourism 2. Solid waste, recycling and sustainable materials management 3. Technology (Renewable energy) What Mattered? Barriers • A glut of vacant, condemned properties (brownfields) that hamper redevelopment of the properties for new economic development opportunities and growth. • Lack of private investment capital and return on investment potential – less certain returns hamper development in these areas. • Unknown extent of contamination and cleanup costs are potentially prohibitive without major assistance from public or philanthropic funding sources. • Many low-income communities, like coalfield municipalities, suffer from a lack of municipal and civil society capacity to lead extensive multi-year, multi-stakeholder redevelopment processes. • The lack of legal and development expertise hinders planning and momentum at the local level and requires obtaining outside support. Opportunities • Projects start with robust and locally-grounded planning process for project concept development. • Providing essential support to local stakeholders to plan for the regions’ economic futures. • Need for restoration economy to accelerate the adoption of innovative approaches to land restoration that contribute meaningfully to the regions’ economic rebirth. Citation BenDor, Todd K. et al. 2015. “Defining and Evaluating the Ecological Restoration Economy: Defining and Evaluating the Restoration Economy.” Methodology 14 case-studies of restoration projects at national, state, and county levels Systematic literature review Focus of Studies Employment by industry: restoration economy Community Impact or Socio-economic well-being: employment multiplier effects of the restoration economy. Outcome Variable(s) What Mattered? Opportunities • Restoration investments appear to have particularly localized benefits which can be attributed to the tendency of projects to employ local labor and materials. • Studies show a range of 6.8-39.7 jobs per $1 Million invested for the restoration economy • Studies show range of employment multipliers of 1.97-3.8 for the restoration economy Socioeconomic Transition in the Appalachia Coal Region Some Factors of Success 80 Citation Besser, Terry L. et al. 2008. “The Impact of Economic Shocks on Quality of Life and Social Capital in Small Towns.” Methodology 99 small towns in Iowa Longitudinal survey data gathered from residents (1994 & 2004) and telephone interviews with key informants (2004) Quantitative empirical analysis of primary data & qualitative empirical analysis of primary data Focus of Studies Event: community economic shocks (e.g., loss of major employer, opening of new prison, boom and bust from energy development, natural disasters) Community Impact or Socio-economic well-being: Outcome Variable(s) Quality of Life (QoL) • Factor scale composed of three items measuring residents’ overall satisfaction with government services, non-governmental services, and the community in general. Social Capital • Perceptions of Within-Community Social Capital • Perceptions of Between-Communities Social Capital What Mattered? Opportunities • Gaining employment (+) * on QoL & both forms of between group social capital • Non-governmental amenities (+) ** on QoL & Structural between-group social capital • Net sum of community shocks are (+) * on QoL, subjective between-group social capital, & structural within- group social capital Barriers • Losing employment (-) * on subjective within-group social capital The significance of the net sum of shock significance suggests that the cumulative strength and degree of the shock are more important than the kind of shock in predicting changes in QoL and social capital in a community. 81 Citation Betz, Michael et al. 2015. “Coal mining, economic development, and the natural resource curse.” Methodology Continental U.S. counties & separate counties within the Appalachian Regional Commission (ARC) borders for two time periods: 1990-2000 and 2000-2010. Quantitative empirical analysis of secondary data Focus of Studies Employment by industry: total county coal mining employment share and the impact on community well- being indicators Community Impact or Socio-economic well-being: percent changes for the decadal models (1990-2000 and 2000-2010): Outcome Variable(s) 1. Per capita income 2. Wage and salary income 3. Median household income 4. Rental and investment income 5. Population 6. Accommodation employment 7. Retail employment 8. Level-measure of poverty rate 9. Employment/population ratio 10. Disability/employment ratio 11. Proprietors’ share of total employment. What Mattered? Opportunities • Coal employment is generally associated with more positive (or less negative) effects in the post-2000 boom period relative to 1990s; these results hold across US and ARC counties • Coal mining appears to have benefits to lower and middle-income households for the U.S. as a whole • 1960 poverty rates were (+) * on change in socio-economic well-being Barriers • ARC counties tended to fare worse on economic indicators relative to other U.S. counties • Coal employment appears to not have same benefits to lower and middle-income households in ARC counties • Coal employment (-) * on changes in population and measures of entrepreneurship as reflected by self- employment measures; these relationships are particularly strong in ARC counties • Oil and gas employment generally had negative impacts on socioeconomic wellbeing, particularly during the 90-00 period. Socioeconomic Transition in the Appalachia Coal Region Some Factors of Success 82 Citation Black, Dan, et al. 2005. “The Economic Impact of the Coal Boom and Bust.” Methodology Counties in Kentucky, Ohio, Pennsylvania, and West Virginia Boom-bust cycle in 1970-1980s: Boom, 1970-7 Peak, 1978-82 Bust, 1983-9 Quantitative empirical analysis of secondary data Focus of Studies Employment by industry & event: coal mining, employment, and earnings per mining worker and the impact on community well-being across boom-bust cycles Community Impact or Socio-economic well-being: Outcome Variable(s) 1. Employment Multipliers: jobs created in traded or local sector per job created in mining sector 2. Wage growth by sector, 1970-80 & 1980-90 • Mining • Non-Mining • Construction • Retail Trade • Services • Manufacturing 3. Population Growth by Gender, 1970-80 & 1980-90 4. Change in Poverty, 1970-80 & 1980-90 Testing is for differences in average annual changes in the logarithm of employment, earnings, and earnings per worker for the non-mining sector between treatment and non-treatment counties. Treatment counties are those that produce at least 10% of their total earnings from the coal industry in 1969. What Mattered? Opportunities: • Coal boom spurred economic growth in the non-mining sectors of coal dependent counties. • Employment grew during the boom in construction and service sectors • Coal boom and bust did generate modest employment spillover into local sectors. • Coal boom decreases number of families in poverty. Barriers: • Employment declined during the bust for all three non-mining local sectors (construction, retail, and services). • Earnings per worker also decrease in all three local sectors during the bust • The multiplier effect of the bust is almost twice the size of the multiplier effect of the boom (for every 10 jobs lost in the coal sector, 3.5 jobs were estimated to be lost in construction, retail, and service sectors) • Bust produces wage declines in all sectors. • Bust produces population declines for both men and women ages 10-39. • Bust produces increases in the number of families in poverty. 83 Citation Carley, Sanya, et al. 2018. “Adaptation, Culture, and the Energy Transition in American Coal Country.” Methodology Focus-groups with individuals that reside or work within the range of Appalachia identified (25 individuals; conducted in 2 WV cities) + 23 interviews with experts Qualitative empirical analysis Focus of Studies Event: coal mining communities’ perceptions and acceptance of transition away from coal. Community Impact or Individuals’ attitudes: residents of coal communities’ acceptance of – Outcome Variable(s) 1. The energy transition 2. The implications of the transition for their personal circumstances 3. How they fared as the transition has evolved in their own community What Mattered? Opportunities: • Respondents perceived that new professional opportunities encouraged acceptance of the transition • Perception that emphasis on training or community college programs to prepare community members for transition Barriers: • Perceived that lower levels of education across coal miners and their families make them more susceptible to shock and limits their potential adaptability • Community cultural identity of mining • Lack of alternative economic opportunities within coal communities • Promise of return to coal jobs (e.g., Trump administration_ is damaging to community and individuals’ efforts to adapt and accept changes. Socioeconomic Transition in the Appalachia Coal Region Some Factors of Success 84 Citation Chamberlin, Molly et al. 2019. “Success Factors, Challenges, and Early Impacts of the POWER Initiative: An Implementation Evaluation.” Methodology Evaluation and report on the qualitative impact of investments made to date with POWER (Partnerships for Opportunity and Workforce and Economic Revitalization) Initiative Sample of 88 grantees were selected for document review and experiences with implementation. Policy evaluation Focus of Studies Policy: POWER initiative – an evaluation of 88 projects and whether they met their shared objectives of projects. Community Impact or Socio-economic well-being: Outcome Variable(s) 1. Economic Diversification 2. Job Creation 3. Capital Investment 4. Workforce Development 5. Reemployment What Mattered? Opportunities: • Success factors related to target population recruitment and engagement of community residents: • Strategic recruitment; Use of multiple types of media; Tailoring programming; Demonstrating progress to participants • Success factors of projects related to organizational capacity: • Internal resources; Organizational experience and reputation; Promoting a creative and nimble organizational culture • Success factors related to partnerships and collaborations: • Building of community self-determination; Building a pipeline of future leaders; Increasing grantee presence and connection with communities Barriers: • Common challenges for the recruitment of residents/workers as well as retention and continued engagement in programs included: • Participant reluctance; Competition among organizations; Access to funding and capital • Common challenges related to organizational capacity included: • Project or time management; Adequate staffing; Financial management; Grants management • Common challenges related to partnerships and collaborations: Limited pool of community resources; Social and environmental barriers; Community infrastructure 85 Citation Cook, Ak. 1995. “Increasing Poverty in Timber-Dependent Areas in Western Washington.” Methodology Counties in Western Washington: 8 timber-dependent non-metropolitan counties 4 nonmetropolitan counties that are less timber-dependent 7 metropolitan counties Quantitative empirical analysis of secondary data Focus of Studies Employment by industry: timber-dependent employment (forestry/fisheries and furniture, lumber and wood products). Community Impact or Socio-economic well-being: poverty rates Outcome Variable(s) What Mattered? Barriers: • Percent change in labor force participation by both men and women were lower in timber-dependent counties • Wage and salary incomes were lower in timber dependent counties • Greater number of female-headed households in timber-dependent counties ➞ higher poverty rates. Citation Cowan, Tadlock. 2012. Military Base Closures: Socioeconomic Impacts. Methodology Communities/localities with military bases Systematic literature review Focus of Studies Event: military base closures (BRAC) Community Impact or Socio-economic well-being: employment multipliers & income multipliers Outcome Variable(s) Input-Output analyses What Mattered? Barriers: • Rural location – harder to recover from loss of base • Loss of population potentially leading to loss of local government revenues Opportunities • Employment multipliers were less than one ~ losses were associated with military transfers out of the area • Per capita income was little affected by the closures • Widespread community commitment to a sound plan for base reuse has been shown to be crucial to positioning communities to life without a military base. Socioeconomic Transition in the Appalachia Coal Region Some Factors of Success 86 Citation Daniels, S.E., et al. 2000. “Reemployment Programs for Dislocated Timber Workers: Lessons from Oregon.” Methodology Evaluation of policy implemented in 2 counties in Western Oregon through interviews with displaced timber workers 1. Linn County – 15% employment tied to timber; six communities classified as timber dependent 2. Benton County – more diversified economy based on timber and wood products Assessment of 2 worker training programs Qualitative empirical analysis & policy evaluation Focus of Studies Policy: reemployment programs for Displaced Timber Workers Community Impact or Individuals’ attitudes: Perceived Success of and Satisfaction with reemployment as the result of two Outcome Variable(s) programs: 1. State initiated career planning workshop – Choices and Options (C&O) – 2-week workshop Designed and implemented by three agencies: local Job Training Partnership Act (JRPA) agency, community college, and employment department. 2. Federal initiative – Jobs in the Woods (watershed restoration projects) – provided means to receive classroom and field work retraining in new occupation – ecosystem restoration – while they worked in the woods and earned a wage. Designed by State Community Economic Revitalization Team; implemented by steering committee composed of representatives from the state economic development agency, higher education, federal and state land management agencies, local JTPA agencies, and labor. What Mattered? Opportunities • Both programs produced modest, positive effects on the displaced workers’ job satisfaction Barriers • Challenges to starting small business in nascent industry (ecosystem restoration) Citation Deaton, James B., et al. 2012. “An Empirical Examination of the Relationship between Mining Employment and Poverty in the Appalachian Region.” Methodology 399 Counties in Appalachia (ARC) Quantitative empirical analysis of secondary data Focus of Studies Employment by industry: mining employment Community Impact or Socio-economic well-being: poverty rate Outcome Variable(s) Using panel data decomposition models the effects of an increase in a sector’s employment share to identify an immediate and lag effect. What Mattered? Opportunities • Immediate effect of increased mining employment is (-) * on poverty rates • Share white population (-) ** • Share high school graduates (-) ** Barriers: • Higher share of mining employment, at the expense of manufacturing, agriculture, or service sectors, (+) * on long-term poverty rate • Share of dependent population (+) ** • Unemployment rate (+) ** 87 Citation Douglas, Stratford and Anne Walker. 2017. “Coal Mining and the Resource Curse in the Eastern United States.” Methodology 409 Appalachian Counties (1970-2010): Compare coal counties to coal-free counties Quantitative empirical analysis of secondary data Focus of Studies Employment by industry: resource-sector dependence as measured by employment in coal mining Community Impact or Socio-economic well-being: long-run (1970-2010) per capita personal income growth Outcome Variable(s) What Mattered? Barriers: • A one standard deviation increase in resource dependence is associated with 0.5-1 percentage point (-) in annual long-run growth rate of per capita personal income • A standard deviation in resource dependence is associated with a 0.2 percentage point (-) in annual short- run growth of per capita personal income • Disincentives to education explain about 15 percent of the apparent resource curse of resource-dependence Opportunities • Metropolitan counties had (+) ** on growth in per capita personal income Citation Freudenburg, William R., and Lisa J. Wilson. 2002. “Mining the Data: Analyzing the Economic Implications of Mining for Nonmetropolitan Regions.” Methodology Meta-analysis of all quantitative studies on mining (301) comparing nonmetropolitan mining regions compared against other nonmetro regions and/or their own experiences over time. Systematic literature review Focus of Studies Employment by industry: mining and mining dependence Community Impact or Socio-economic well-being: Outcome Variable(s) 1. Income 2. Poverty 3. Unemployment 4. Overall Findings What Mattered? Barriers • Mining has predominantly adverse for unemployment, poverty, and overall findings • Findings of favorable economic conditions in mining regions have become relatively rare since 1982, making up only about 20 percent of the available findings that come from 1983 and thereafter. Meanwhile, adverse findings make up nearly three times that number (57.3%) for the same era. Opportunities • Income has greater favorable outcomes among the reviewed studies (47.5%); while only 33.9% are adverse and 18.6% were neutral. • Mining in western region had more favorable outcomes (52.1%) than in South, Great Lakes, or Other regions Socioeconomic Transition in the Appalachia Coal Region Some Factors of Success 88 Citation Graff, Michelle, et al. 2018. “Stakeholder Perceptions of the United States Energy Transition: Local-Level Dynamics and Community Responses to National Politics and Policy.” Methodology Interviews and surveys with community energy stakeholders (individual sand organizations working on the energy transition – both non-profit, private, and public sector) on perceptions in three frontline communities: 1. Detroit, MI 2. St. Louis County, MO 3. Appalachia Coal Country (Counties in KY and WV) Qualitative empirical analysis Focus of Studies Employment by industry: sectors related to energy transition: 1. Automobiles 2. Coal company headquarters and coal railroad hub 3. Coal mining Community Impact or Policy evaluations: Outcome Variable(s) 1. Community Activities related to Transition 2. Actors and institutions across communities leading (or not) on energy policy issues What Mattered? Opportunities: • Solar farms, panels and policies offering new economic opportunities • Perception that “bottom-up” approaches are more successful than “bottom-down” efforts Barriers • All localities: · Lack of mobility of local residents · Necessity of moving for economic opportunities but lack of affordable housing and place attachment impeding ability to move • Appalachia · Job loss – layoffs and business closures in other local industries · Concern about wider adverse labor market effects – schools and retail closing due to families migrating out of communities · Vulnerable populations – low income and minority groups & individuals just entering the labor force (age 18-25); those in mid- to late-stage career (age 50+), and women. • Detroit and St. Louis · Concern about potential for increasing utility prices · Lack of understanding about the potential benefits and opportunities inherent in transition · Vulnerable populations – low income individuals and communities of color particularly vulnerable 89 Citation Greenberg, Pierce. 2018. “Coal Waste, Socioeconomic Change, and Environmental Inequality in Appalachia: Implications for a Just Transition in Coal Country.” Methodology Neighborhoods (Census tracts) in Central and North Central Appalachian Region between 1990-2000 Quantitative empirical analysis of secondary data Focus of Studies Employment by industry: Coal waste impoundments (negative environmental externality of mining) Community Impact or Socio-economic well-being: change in poverty rate from 1990-2000 Outcome Variable(s) What Mattered? Barriers • Proximity to impoundments (+) * with poverty rate change • Higher levels of mining employment have (+) * poverty changes over time Opportunities • Past poverty rate (1990) was (-) * to poverty change over time. Citation Haggerty, Julia H., et al. 2018. “Planning for the Local Impacts of Coal Facility Closure: Emerging Strategies in the U.S. West.” Methodology Characterization and assessment of strategies of local governments policies surrounding the closing of coal-fired plants in the U.S. West. Systematic literature review Focus of Studies Employment by industry: coal-fired power plants Community Impact or Policy evaluations: four key transition planning criteria for coal plant retirements Outcome Variable(s) 1. Importance of replacing and stabilizing revenue streams 2. The necessity to plan, fund, and execute complete environmental remediation 3. The risk of focusing on economic development strategies that are inappropriate to local context 4. The association of willingness to change and positive outlook with community resilience during transitions What Mattered? Barriers • Remoteness • Lack of alternative economic opportunities • Low education of populations • Uncertainty surrounding policies and government funding for rural services • Degraded environments associated with coal mining and coal burning can hamper long-term growth • Many communities do not have adequate health services, universities or training centers • Resistance to change – tend to blame closure on restrictive environmental regulations, while ignoring role of markets, particularly price competition with natural gas Opportunities • Promoting agro-tourism and outdoor recreation • Existing transmission infrastructure as competitive advantage for renewable development Socioeconomic Transition in the Appalachia Coal Region Some Factors of Success 90 Citation Haggerty, Julia. 2014. “Long-Term Effects of Income Specialization in Oil and Gas Extraction: The U.S. West, 1980-2011.” Methodology Counties within six major oil and gas producing states in the US West: Colorado, Montana, New Mexico, North Dakota, Utah and Wyoming between 1980-2011 Quantitative empirical analysis from secondary data Focus of Studies Employment by industry & event: oil and gas specialization Community Impact or Socio-economic well-being for 1980-2011 (effect of duration of boom): Outcome Variable(s) 1. Per capita income* 2. Average earnings per job 3. Total employment 4. Total income 5. Per capita investment income 6. Unemployment rate 7. Percent of adults with college education* 8. Percent of individuals in poverty 9. Percent of renter occupied units with gross rent >35% of household income 10. Gini coefficient 11. Violent and property crimes per 1000 people* *only three found to be statistically associated with the duration of oil and gas development. What Mattered? Opportunities • Shorter duration of boom associated with (+) * per capita income growth (relative to longer duration boom) Barriers • Per capita income (-) * over the period if county participated in 1980-82 boom • The longer a county has specialized in oil and gas, the higher the county’s crime rates • Longer specialization in oil and gas associated with fewer adults with a college degree 91 Citation Haggerty, Mark. 2019. “Communities at Risk from Closing Coal Plants.” Methodology Report summarizing characteristics of communities most at risk from coal plant closures Technical report based on Coal Transition Solutions Forum which brought together 17 diverse participants to share their expertise in local and state government, academia, consulting and policy work. Policy evaluation Focus of Studies Employment by industry: coal plants Community Impact or Policy evaluations: vulnerabilities and barriers to transition Outcome Variable(s) What Mattered? Barriers • Isolation from major population centers and markets • Limited institutional and leadership capacity (e.g., little or no planning staff); lack capacity to apply for federal and state assistance. • Weak ties to state or regional actors, thereby limiting access to regional and state support for transition. • Fragmented ownership of plants and mines · Owners will vary in response to different incentives depending upon their ownership structure and their customer base. • Workers and families are “stuck-in-place” due to mortgages or other factors. • Have older/aging workforce that are highly skilled, yet overly adapted. • Long-term dependence on coal-revenue · Delays acceptance of transition • Lacks adequate fiscal autonomy – state restrictions on local budgeting authority. • Few funds or none for transition planning and implementation • Lack adequate information to assess fiscal risks and the limitations imposed by state and federal fiscal policies • Federal assistance is over-prescribed and poorly targeted – limiting local autonomy and flexibility to use funds to meet locally defined needs. Opportunities (Potential) • Reallocate coal revenue to transition needs and priorities · Reform current tax policies that mandate that coal revenue be used for tax relief. Coal revenue would be used for transition purposes including permanent savings and long-term investments. • Broaden the tax base to replace coal revenue • Ensure resource taxes address transition needs and priorities • Fund effective federal grant and loan programs for coal communities • Secure one-time transition payments from facility owners • Increase local fiscal autonomy, including allowing for local savings authority Socioeconomic Transition in the Appalachia Coal Region Some Factors of Success 92 Citation Haller, Melissa, et al. 2017. “The End of the Nuclear Era: Nuclear Decommissioning and Its Economic Impacts on U.S. Counties.” Methodology County level study of 24 nuclear reactors that have undergone decommissioning from 1975-2014. Quantitative empirical analysis of secondary data Focus of Studies Event: nuclear decommissioning Community Impact or Socio-economic well-being: Outcome Variable(s) 1. Employment 2. Income 3. Population Method • Difference-in-difference regression • Propensity score matching What Mattered? Opportunities • Decommissioning (+) * in employment and per capita income over time Barrier • Decommissioning has a non-significant effect (neither positive or negative) on population growth over time Citation Hooker, Mark A. and Michael M. Knetter. 2001. “Measuring the Economic Effects of Military Base Closures.” Methodology 57 counties with military bases experiencing closures Quantitative empirical analysis of secondary data Focus of Studies Event: Military base closures Community Impact or Socio-economic well-being: job loss multipliers – how many fewer jobs a closure county had than would be Outcome Variable(s) expected if it grew at counterfactual rates What Mattered? Opportunities • Military base closures had (+) job multiplier effects, likely due to: • Counties receiving technical and financial aid in their reuse (recover) of base property efforts • The sample for this study is not random – counties might have been more-adaptable than average military base closure counties. Citation Hultquist, Andy and Tricia L. Petras. 2012. “An Examination of the Local Economic Impacts of Military Base Closures.” Methodology 510 individual counties containing one or more military bases active from 1977 to 2005 + all neighboring counties containing one or more military bases = 1,721 counties observed Quantitative empirical analysis of secondary data Focus of Studies Event: Military base closures Community Impact or Socio-economic well-being: Total employment Outcome Variable(s) What Mattered? Opportunities • Places with higher military base personal (own county) prior to the closure of a base experienced (+) * change in county employment 93 Citation Isserman, Andrew M., et al. 2007. “Why Some Rural Communities Prosper While Others Do Not.” Methodology All rural counties in the continental U.S. Quantitative empirical analysis of secondary data Focus of Studies Employment by industry: Attributes of Prosperous rural counties Community Impact or Socio-economic well-being: Outcome Variable(s) 1. Poverty rate 2. Unemployment rate 3. High school dropout rate 4. Housing problem rate (% of households with at least one housing condition) Findings based on t-tests** What Mattered? Opportunities • Geographic Indicators: · Weak findings for impacts of distance from MSA and location variables. This is encouraging factors like temperature, distances to cities, and employment in the nearby region are beyond the control of local rural development actions. • Economic Indicators · More vigorous private sector · More active and prosperous farm sector · More diversified economies (Herfindahl index) · Lower income inequality • Human and Social Capital Indicators · Higher educational levels on average · Educational attainment among 25-34 year olds is higher · More creative class as percent of occupations · Greater number of associational activity based establishments per capita · More adherent to civically engaged religions · Higher proprietor income per capita (~ of entrepreneurialism) • Demographic Indicators: · Slower growth (2% compared to 7% for less prosperous counties between 1990 and 2000) · Larger elderly population Barriers • Economic Indicators · Greater number of government, public sector jobs · Greater number of resource based, value added manufacturing jobs • Demographic Indicators · Greater share demographic changes · more foreign born populations · More foreign-born who arrived in the past decade · More recent in-migration · Greater racial heterogeneity · Rural counties with minority concentrations Socioeconomic Transition in the Appalachia Coal Region Some Factors of Success 94 Citation Jolley, Jason, et al. 2019. “The economic, fiscal, and workforce impacts of coal-fired power plant closures in Appalachia Ohio.” Methodology Adams county, OH – two Dayton power & light coal-fired power plants (in Appalachia Region) Quantitative empirical analysis of secondary data Focus of Studies Employment by industry: coal-power plants Community Impact or Socio-economic well-being: Outcome Variable(s) 1. Employment Impacts – direct, indirect, and induced employment 2. Fiscal Impacts – losses due to lower tangible personal property tax 3. Workforce Impacts – viability of retraining employees in comparable wage jobs What Mattered? Barriers • Supports finding of limited fiscal resilience of Appalachian communities to exogenous shocks • State fiscal policies limit the ability of localities to recoup lost taxes as a result of plant closures • Access to information for displaced workers – communities have low digital literacy which complicates both occupational transitions as well as the search for job openings • Transportation to technical centers or community colleges where displaced workers can bridge skill Citation Kelsey, Timothy, et al. 2016. “Unconventional Gas and Oil Development in the United States: Economic Experience and Policy Issues.” Methodology Examine the economic experience of past energy booms and of the current unconventional gas and oil development era Survey of key economic issues that tend to arise with energy (oil and gas) development. Policy evaluation Focus of Studies Event: gas and oil development Community Impact or Policy evaluations: recommendations and implementation Outcome Variable(s) What Mattered? Barriers • Taxation limited by state government policies • Use of Revenue – local governments have little discretion over use of revenue • Locus of Decision-Making vis-à-vis federal, state, and local governments (preemption of local authority) • Lack of distributional equity over economic benefits (among residents/workers) • High wages in in resource sector for less-skilled workers reduces the incentive for further education and training for less-skilled workers in a vulnerable position Opportunities (Recommended) • Use any financial proceeds from activity to fund long-run investments that will strengthen the community so it will be better able to adapt when drilling ends • Do not make long-run financial commitments that may burden the community for a long time • Strive to maintain diversified economy • Protect important environmental and community assets and amenities so they are not harmed during the boom 95 Citation Kirschner, Annabel R. 2010. “Understanding Poverty and Unemployment on the Olympic Peninsula after the Spotted Owl.” Methodology Sub-county areas formed from census tracts in the Olympic Peninsula of Washington State Quantitative empirical analysis of secondary data Focus of Studies Employment by industry: timber and timber dependency Community Impact or Socio-economic well-being: Outcome Variable(s) 1. Poverty (2000) 2. Unemployment (2000) What Mattered? Barriers • Past poverty (1990) (+) * on current poverty rates • Greater racial/ethnic minority population size (+) * on unemployment rates Opportunities • Greater change in the share population with bachelor’s degree (-) * poverty rates Socioeconomic Transition in the Appalachia Coal Region Some Factors of Success 96 Citation Lobao, Linda, et al. 2016. “Poverty, Place, and Coal Employment across Appalachia and the United States in the New Economic Era.” Methodology Counties across the US as well as specifically in Appalachian region (ARC) from 1990 to 2010. Quantitative empirical analysis of secondary data Focus of Studies Employment by industry: coal mining employment and change in coal mining employment Community Impact or Socio-economic well-being: Outcome Variable(s) 1. Poverty 2. Household income 3. Unemployment What Mattered? Opportunities • Poverty · Coal employment (-) * from 2000-2010 for US and ARC · Share oil and gas employment (-) * 2000-2010 for all US counties · Growth in coal employment (-) * for ARC in both periods • Median Household Income · Share coal employment (+) * in US and ARC for 2000-10 period · Oil and gas mining employment (+) * for US in 00-10’ • Unemployment · Coal employment (-) * for US and ARC in 00-10’ period · Gas and oil employment (-) * for US in 00-10’ period · Change in other mining employment for US and ARC in 90-00’ and 00-10’ · Change in employment for coal (-) * for US and ARC in 00-10’ · Change in employment for oil and gas (-) * for US and ARC in 00-10’ Barriers • Poverty · Higher coal employment (+) * 1990-2000 for US • Median Household Income · Share coal employment (-) * in US and ARC for 90-00’ period · Oil and gas employment (-) * in US for 90-00’ period · Growth in oil and gas (-) * for 90-00’ period in US • Unemployment · Gas and oil employment (+) * for ARC in 90-00’ period 97 Citation Mayer, Adam. 2018. “A Just Transition for Coal Miners? Community Identity and Support from Local Policy Actors.” Methodology County and city policy actors in Colorado and Utah Quantitative empirical analysis of survey data Focus of Studies Employment by industry: coal mining Community Impact or Individuals’ attitudes: policy actors’ perceptions of just transition – Outcome Variable(s) 1. What accounts for coal’s long-term poor fortunes? 2. Who is responsible for paying for policies to assist in the transition away from coal? What Mattered? Opportunities • A multi-causal explanation – both competition from oil and gas as well as regulations to blame – for decline in coal employment (+) * programs to provide education and training to displaced miners Barriers • Community identity tied to extractive industries (-) * to allocating special funds for relocation to new areas to find jobs • Regulations explanation – environmental regulations as cause of coal miner displacement – (-) * that communities will support policies to ensure miners pensions are fully funded Citation Partridge, Mark D. et al. 2013. “Natural Resource Curse and Poverty in Appalachia America.” Methodology Counties in Appalachia region (ARC) over the 1990-2010 period. Quantitative empirical analysis of secondary data Focus of Studies Employment by industry: coal mining employment & mountain top mining Community Impact or Socio-economic well-being:poverty rates (2000) Outcome Variable(s) What Mattered? Barriers • Past poverty (1990) (+) * for both US and ARC • Share of coal employment (+) * for ARC counties • Growth in oil and gas employment (+) * for US • Distance to higher-tiered metropolitan areas (+) * Opportunities • Educational attainment – high school graduates (-) * Socioeconomic Transition in the Appalachia Coal Region Some Factors of Success 98 Citation Partridge, Mark D. and M. Rose Olfert. 2011. “The Winners’ Choice: Sustainable Economic Strategies for Successful 21st-Century Regions.” Methodology Functional Economic Areas for Atlanta, GA; Columbus, OH; Des Moines, IA; and Minneapolis-St. Paul, MN MSAs from 1950-2009 Quantitative empirical analysis of secondary data & systematic literature review Focus of Studies Employment by industry: general indicators of growth Community Impact or Socio-economic well-being: population growth (size and density) as well as urbanization intensity Outcome Variable(s) What Mattered? Opportunities • Expanding regional boundaries as more rural areas functionally become tied to urban centers leads to greater within-region cohesiveness • Build from within through retention, expansion, and supporting local entrepreneurship; will likely lead to a more diverse economy than attracting one or two large outside firms/industries. • Regions with an attractive quality of life for high-skilled workers will have advantage in growth Citation Pollin, Robert, and Brian Callaci. 2019. “The Economics of Just Transition: A Framework for Supporting Fossil Fuel–Dependent Workers and Communities in the United States.” Methodology National overview of potential framework for transitioning communities and workers away from fossil fuel-based employment. Systematic literature review Focus of Studies Employment by industry: fossil-fuel dependent employment Community Impact or Policy evaluations: transitioning workers away from coal employment Outcome Variable(s) What Mattered? Opportunities • Clean energy investments will produce jobs for electricians, steel workers, machinist, engineers, truck drivers, research scientists, lawyers, accountants, and administrative assistants. • Attritions by retirement – about 85 percent of the necessary job retrenchments can be managed through attritions by retirement when current employed fossil fuel workers reach age 65. Barriers • In a 20-year transition period where the coal industry contracts by 60%, 600 younger (<45) workers will need to be laid off 99 Citation Poppert, Patrick E. and Henry W. Herzog Jr. 2003. “Force Reduction, base closure, and the indirect effects of military installation on local employment growth.” Methodology 3092 counties in the United States Quantitative empirical analysis of secondary data Focus of Studies Event: military base closures Community Impact or Socio-economic well-being: employment in private, nonfarm sectors Outcome Variable(s) 1. Employment by industry: hydrocarbon dependent occupations What Mattered? Opportunities • Military base closures show (+) * on employment two years after base closure • Transfers for educational assistance (+) * on employment rates • Population-employment ratio (+) * effect on employment rates • Population density (+) * on employment rates • Other non-manufacturing, non-services sector employment (+) * on employment • State and Local government employment (+) * on employment • Agricultural sector employment (+) * on employment Barriers • Past employment loss (-) * employment • Coefficient of specialization (-) * on employment Citation Snyder, Brian F. 2018. “Vulnerability to Decarbonization in Hydrocarbon-Intensive Counties in the United States: A Just Transition to Avoid Post-Industrial Decay.” Methodology Community level overview of the development of a metric to assess vulnerabilities/resiliencies to decarbonization of the economy to develop an index of vulnerability Quantitative empirical analysis of secondary data Focus of Studies Employment by industry: hydrocarbon dependent occupations Community Impact or Socio-economic well-being: vulnerability Index – to assess areas vulnerable to socioeconomic declines due Outcome Variable(s) to decarbonization What Mattered? Barriers • Geographic isolation • High-levels of preexisting socioeconomic disadvantage • Large levels of hydrocarbon employment • Low educational levels may negatively impact communities’ ability to adapt • Role and power of labor (workers) is more restricted in the US and is, therefore, unlikely to play a prominent role in the transition. Policy will therefore rely on federal, state, and local governments and policy-makers. Socioeconomic Transition in the Appalachia Coal Region Some Factors of Success 100 Citation Sorenson, David and Peter Stenberg. 2015. “The Effect of Military Base Closures on Rural County Economies: An Evaluation of the 1988-1995 Rounds of Cuts.” Methodology Six counties that meet the criteria of 1. Experiencing a major loss of jobs from military base and 2. Being located in a non-metro area (91-93) and counties active as control group (matching) Quantitative empirical analysis of secondary data Focus of Studies Event: military base closures Community Impact or Socio-economic well-being: Outcome Variable(s) 1. Employment rates 2. Population size Used matching techniques based on sectoral composition of local economy, spatial setting, income levels and sources, and prior growth What Mattered? Barriers • Counties with military base closures (-) * employment levels relative to those without • Counties with military base closures experienced (-) * population growth compared to those without Citation Tsvetkova, Alexandra and Mark D. Partridge. 2016. “Economics of Modern Energy Boomtowns: Do Oil and Gas Shocks Differ from Shocks in the Rest of the Economy?” Methodology Counties Quantitative empirical analysis of secondary data Focus of Studies Employment by industry: oil and gas Community Impact or Socio-economic well-being: Outcome Variable(s) 1. Employment across 14 sectors 2. Population size What Mattered? Opportunities • Oil and gas employment has (+) * spillover into other industry sectors • Metropolitan areas – no link between total employment and energy sector employment Barriers • Overall, given that job effects of energy booms are relatively modest in magnitude compared to the scale of the rest of the economy, local economies would be better off if they were to experience broad-based growth rather than energy booms both in terms of multiplier effects and enhanced economic diversity. 101 Citation Tsvetkova, Alexandra, et al. 2019. “Self-Employment effects on regional growth: a bigger bang for a buck?” Methodology Counties (2001-2013) Quantitative empirical analysis of secondary data Focus of Studies Employment by industry: self-employment vs. wage/salary employment Community Impact or Socio-economic well-being: employment growth rates Outcome Variable(s) What Mattered? Opportunities • Estimated benefits for self-employment on employment growth rates (+) * are substantially larger than identical effects of paid employment Socioeconomic Transition in the Appalachia Coal Region Some Factors of Success 102 Citation Weinstein, Amanda L., et al. 2018. “Follow the money: aggregate, sectoral and spatial effects of an energy boom on local earnings.” Methodology Counties by metro and nonmetro status that experience energy expansion compared to other areas Quantitative empirical analysis of secondary data Focus of Studies Employment by industry: oil and gas Community Impact or Socio-economic well-being: Outcome Variable(s) 1. Change in total earnings growth 2. Change in Earnings per worker growth 3. Change in employment growth What Mattered? Opportunities • Total earnings • Employment · Nonmetro · Nonmetro · Energy growth (+) * · Energy growth in county (+) ** · Change in energy performance in bordering · Change in energy performance in bordering counties (+) * counties (+) ** · Industry mix metric (+) · Industry mix (+) ** · Metro · Metro · Energy growth (+) * · Energy growth in county (+) ** · Change in energy performance in boarding · Change in energy performance in bordering counties (+) * counties (+) ** · Industry mix metric (+) · Industry mix (+) ** • Earnings per worker · Nonmetro · Energy growth (+) * · Change in energy performance in boarding counties (+) * · Industry mix metric (+) · Metro · Energy growth in county (+) · Industry mix (+) Barriers • Total earnings • Employment · Nonmetro · Nonmetro · Mining employment share (-) * · Mining employment (-) * • Earnings per worker · Nonmetro · Mining employment share (-) * 103 Citation Weber, Jeremy G. 2012. “The Effects of a Natural Gas Boom on Employment and Income in Colorado, Texas, and Wyoming.” Methodology Counties in Colorado, Texas and Wyoming through boom and bust cycles of natural gas extraction Quantitative empirical analysis of secondary data Focus of Studies Employment by industry & event: natural gas Community Impact or Socio-economic well-being: Outcome Variable(s) 1. Employment 2. Wage and salary income 3. Median household income 4. Poverty What Mattered? Opportunities • Employment · Boom (+) * · Wyoming (+) * relative to Colorado • Wage and Salary Income · Boom (+) * · Mining share of earnings (+) * · Wyoming (+) relative to Colorado • Median Household Income · Population density (+) * · Ag share of earnings (+) * · Mining share of earnings (+) · Population density of contiguous counties (+) • Poverty · Population density (-) * · Manufacturing share of earnings (-) * · Wyoming compared to Colorado (-) * Socioeconomic Transition in the Appalachia Coal Region Some Factors of Success 104 List of Studies Reviewed Appalachian Citizens’ Law Center, Appalachian Voices, Daniels, S. E., C. L. Gobeli, and A. J. Findley. 2000. Coalfield Development Corporation, Rural Action, “Reemployment Programs for Dislocated Timber and Downstream Strategies. 2019. “A New Horizon: Workers: Lessons from Oregon.” Society & Natural Innovative Reclamation for a Just Transition.” Resources 13 (2): 135–50. Reclaiming Appalachian Coalition. Deaton, B. James, and Ekaterina Niman. 2012. “An BenDor, Todd K., Avery Livengood, T. William Lester, Empirical Examination of the Relationship between Adam Davis, and Logan Yonavjak. 2015. “Defining Mining Employment and Poverty in the Appalachian and Evaluating the Ecological Restoration Economy: Region.” Applied Economics 44 (3): 303–12. Defining and Evaluating the Restoration Economy.” Douglas, Stratford, and Anne Walker. 2017. “Coal Restoration Ecology 23 (3): 209–19. Mining and the Resource Curse in the Eastern United Besser, Terry L., Nicholas Recker, and Kerry Agnitsch. States.” Journal of Regional Science 57 (4): 568–90. 2008. “The Impact of Economic Shocks on Quality of Freudenburg, William R., and Lisa J. Wilson. Life and Social Capital in Small Towns.” Rural Sociology 2002. “Mining the Data: Analyzing the Economic 73 (4): 580–604. Implications of Mining for Nonmetropolitan Betz, Michael, Mark Partridge, Micheal Farren, Regions.” Sociological Inquiry 72 (4): 549–75. and Linda Lobao. 2015. “Coal mining, economic Graff, Michelle, Sanya Carley, and David M. Konisky. development, and the natural resources curse.” Energy 2018. “Stakeholder Perceptions of the United Economics 50: 105-116. States Energy Transition: Local-Level Dynamics Black, Dan, Terra McKinnish, and Seth Sanders. 2005. and Community Responses to National Politics “The Economic Impact of the Coal Boom and Bust.” and Policy.” Energy Research & Social Science 43 The Economic Journal 115 (April): 449-476. (September): 144–57. Carley, Sanya, Tom P. Evans, and David M. Konisky. Greenberg, Pierce. 2018. “Coal Waste, 2018. “Adaptation, Culture, and the Energy Transition Socioeconomic Change, and Environmental in American Coal Country.” Energy Research & Social Inequality in Appalachia: Implications for a Just Science 37 (March): 133–39. Transition in Coal Country.” Society & Natural Resources 31 (9): 995–1011. Chamberlin, Molly, Nicole Dunn, Abigail Kelly- Smith and Dorinda Byers. 2019. “Success Factors, Haggerty, Julia H., Mark N. Haggerty, Kelli Challenges, and Early Impacts of the POWER Initiative: Roemer, and Jackson Rose. 2018. “Planning for the An Implementation Evaluation.” Appalachian Regional Local Impacts of Coal Facility Closure: Emerging Commission: 1-89. Strategies in the U.S. West.” Resources Policy 57 (August): 69–80. Cook, Ak. 1995. “Increasing Poverty in Timber- Dependent Areas in Western Washington.” Society & Haggerty, Julia. 2014. “Long-Term Effects of Income Natural Resources 8 (2): 97–109. Specialization in Oil and Gas Extraction: The U.S. West, 1980-2011.” Energy Economics 45: 186–95. Cowan, Tadlock. 2012. Military Base Closures: Socioeconomic Impacts. Congressional Research Service: Haggerty, Mark. 2019. “Communities at Risk from 1-9. Closing Coal Plants.” Working Paper. Bozeman, MT: Headwaters Economics. 105 Haller, Melissa, Michael Haines, and Daisaku Partridge, Mark D., Michael R. Betz, and Linda Yamamoto. 2017. “The End of the Nuclear Era: Nuclear Lobao. 2013. “Natural Resource Curse and Poverty Decommissioning and Its Economic Impacts on U.S. in Appalachian America.” American Journal of Counties.” Growth and Change 48 (4): 640–60. Agricultural Economics 95 (2): 449–56. Hooker, Mark A. and Michael M. Knetter. 2001. Partridge, Mark D. and M. Rose Olfert. 2011. “The “Measuring the Economic Effects of Military Base Winners’ Choice: Sustainable Economic Strategies Closures.” Economic Inquiry 39(4): 583-598. for Successful 21st-Century Regions.” Applied Economic Perspectives and Policy 33(2): 143-178. Hultquist, Andy and Tricia L. Petras. 2012. “An Examination of the Local Economic Impacts of Military Pollin, Robert, and Brian Callaci. 2019. “The Base Closures.” Economic Development Quarterly 26(2): Economics of Just Transition: A Framework for 151-161. Supporting Fossil Fuel–Dependent Workers and Communities in the United States.” Labor Studies Isserman, Andrew M., Edward Feser, and Drake Warren. Journal 44 (2): 93–138. 2007. “Why Some Rural Communities Prosper While Others Do Not.” USDA Rural Development. AG RBCS RBS- Poppert, Patrick E. and Henry W. Herzog Jr. 2003. 02-12. IL: University of Illinois at Urbana-Champaign. “Force Reduction, base closure, and the indirect effects of military installation on local employment Jolley, Jason, Christelle Khalaf, Gilbert Michaud, growth.” Journal of Regional Science 43: 459-481. and Austin Sandler. 2019. “The economic, fiscal, and workforce impacts of coal-fired power plant closures Snyder, Brian F. 2018. “Vulnerability to in Appalachia Ohio.” Regional Science Policy Practice, 11: Decarbonization in Hydrocarbon-Intensive 403-422. Counties in the United States: A Just Transition to Avoid Post-Industrial Decay.” Energy Research & Kelsey, Timothy, Mark Partridge, Nancy White. 2016. Social Science 42 (August): 34–43. “Unconventional Gas and Oil Development in the United States: Economic Experience and Policy Issues.” 2016. Sorenson, David and Peter Stenberg. 2015. “The Applied Economic Perspectives and Policy 38 (2): 191-214. Effect of Military Base Closures on Rural County Economies: An Evaluation of the 1988-1995 Rounds Kirschner, Annabel R. 2010. “Understanding Poverty of Cuts.” International Advances in Economic Research and Unemployment on the Olympic Peninsula after the 21: 167-187. Spotted Owl.” Social Science Journal 47 (2): 344–58. Tsvetkova, Alexandra, Mark D. Partridge, and Lobao, Linda, Minyu Zhou, Mark Partridge, and Michael Michael Betz. 2019. “Self-Employment effects on Betz. 2016. “Poverty, Place, and Coal Employment across regional growth: a bigger bang for a buck?” Small Appalachia and the United States in a New Economic Business Economics 52: 27-45. Era.” Rural Sociology 81 (3): 343–86. Weinstein, Amanda L., Mark D. Partridge, Alexandra Mayer, Adam. 2018. “A Just Transition for Coal Miners? Tsvetkova. 2018. “Follow the money: aggregate, Community Identity and Support from Local Policy sectoral and spatial effects of an energy boom on Actors.” Environmental Innovation and Societal Transitions local earnings.” Resources Policy 55:196-209. 28 (September): 1–13. Weber, Jeremy G. 2012. “The Effects of a Natural Partridge, Mark D. (2010) “The Dueling Models: NEG vs Gas Boom on Employment and Income in Colorado, Amenity Migration in Explaining U.S. Engines of Growth” Texas, and Wyoming.” Energy Economics 34 (5): Papers in Regional Science. 89: 513-536. 1580–88. Socioeconomic Transition in the Appalachia Coal Region Some Factors of Success 106 APPENDIX B Appalachian and U.S. Coal Industry Figures 107 Appendix B Figure 1: Coal Fields of the United States Source: Source: East, J.A., 2013, Coal fields of the conterminous United States—National Coal Resource Assessment updated version: U.S. Geological Survey Open-File Report 2012–1205, one sheet, scale 1:5,000,000, available at http://pubs.usgs.gov/of/2012/1205/. Socioeconomic Transition in the Appalachia Coal Region Some Factors of Success 108 Appendix B Figure 2: Annual U.S. Coal Mining Employment using BLS/BEA Data 700 525 COAL MINING EMPLOYEES (Thous nds) 350 175 – 29 39 49 59 79 89 99 19 09 69 20 19 19 19 19 19 19 19 19 20 Co l minin (BEA) Co l minin (BLS) Support ctiviti s for co l minin (BLS) Sources: 1. BEA NIPA Tables 6.4A-D; 2. BLS Quarterly Census of Employment and Wages 109 Appendix B Figure 3: 1954 Alabama Mining Industries by Employment Source: Reproduced from 1954 Census of Manufacturing, Vol. 2. Each symbol represents 100 mining employees, e.g., Black circles denote 100 coal-mining jobs. Note the concentration of coal and metal mining in and near Birmingham in Jefferson County. Socioeconomic Transition in the Appalachia Coal Region Some Factors of Success 110 Appendix B Figure 4: 1954 Kentucky Mining Industries by Employment Source: Reproduced from 1954 Census of Manufacturing, Vol. 2. Each symbol represents 100 mining employees, e.g., Black circles denote 100 coal-mining jobs. 111 Appendix B Figure 5: 1954 Ohio Mining Industries by Employment Source: Reproduced from 1954 Census of Manufacturing, Vol. 2. Each symbol represents 100 mining employees, e.g., Black circles denote 100 coal-mining jobs. Socioeconomic Transition in the Appalachia Coal Region Some Factors of Success 112 Appendix B Figure 6: 1954 Pennsylvania Mining Industries by Employment Source: Reproduced from 1954 Census of Manufacturing, Vol. 2. Each symbol represents 100 mining employees, e.g., Bla Source: Reproduced from 1954 Census of Manufacturing, Vol. 2. Each symbol represents 100 mining employees, e.g., Black circles denote 100 coal-mining jobs. ack circles denote 100 coal-mining jobs. 113 Appendix B Figure 7: 1954 Tennessee Mining Industries by Employment Source: Reproduced from 1954 Census of Manufacturing, Vol. 2. Each symbol represents 100 mining employees, e.g., Black circles denote 100 coal-mining jobs. Socioeconomic Transition in the Appalachia Coal Region Some Factors of Success 114 Appendix B Figure 8: 1954 Virginia Mining Industries by Employment Source: Reproduced from 1954 Census of Manufacturing, Vol. 2. Each symbol represents 100 mining employees, e.g., Black circles denote 100 coal-mining jobs 115 Appendix B Figure 9: 1954 West Virginia Mining Industries by Employment Source: Reproduced from 1954 Census of Manufacturing, Vol. 2. Each symbol represents 100 mining employees, e.g., Black circles denote 100 coal-mining jobs. Socioeconomic Transition in the Appalachia Coal Region Some Factors of Success 116 Appendix B Figure 10: 1950 Mining Workers by Place of Residence, ARC Counties Source: 1950 Census of Population, U.S. Census Bureau 117 APPENDIX C Methodologies Used in the Literature to Understand the Economic Impact of Coal Mining Socioeconomic Transition in the Appalachia Coal Region Some Factors of Success 118 Basic Input-Output Effects. When assessing corresponding multipliers for other economic the expansion or contraction of an industry outcomes such as output and value added. such as coal mining, one applied model is an “input-out” (IO) model, commonly used IO model estimates can be problematic, by practitioners for estimating economic however. The question of what happens to total development impacts. IO models capture three local employment is related to total multiplier types of effects: direct effects, indirect effects, effects based on the direct contribution of and induced effects. For example, if the coal the new coal industry jobs. However, there industry increases employment by 100 workers are caveats regarding IO estimates that are and output by 3 million short tons, the direct generally believed to lead to overestimates effects are simply 100 jobs and 3 million tons. of the economic impacts. First and foremost is direct reverse causality in which people, There are then indirect economic effects including workers, move into a local area from the supply-chain input purchases of for (say) quality-of-life or in response to the the coal industry, including corresponding creation of jobs. Reverse causality leads to responses as input suppliers make their own multiplier estimates being overstated as new purchases of inputs, and so on. And there residents create added demand for local firms. are induced economic effects from increased Second, there are crowding-out effects. For spending by newly employed workers in example, if there is a positive expansion of coal (direct) and coal-supply chain (indirect) the local coal industry, this will bid up wages activities, such as for groceries, household and land prices as the local economy expands, goods, leisure and essential services. The total which leads to other local firms hiring fewer effects are the sum of the direct, indirect, and workers or perhaps going out of business due induced effects. Although IO models are not to higher land and labor costs. Crowding- necessary for estimating multipliers, they out effects are reinforced if the expansion have the advantages of being well known by of economic activity in turn increases local policymakers and enabling a decomposition of housing costs, causing some local residents responses outside of the direct industry being to relocate, which further leads to offsetting considered. reductions in local economic activity. The total economic multiplier can be There are other offsetting effects in IO calculated as the (direct effect + indirect effect estimates. Betz et al. (2015) found that local + induced effect)/(direct effect), and reflects self-employment was negatively impacted by the total number of (say) new jobs supported greater coal-mining intensity, which reduces when the industry directly creates (say) 100 entrepreneurship and long-run economic jobs. For example, a multiplier of 2.2 means growth. Likewise, they found that greater that if the coal industry directly creates 100 coal intensity negatively impacted long- new jobs, there are an additional 120 jobs run population growth, all else equal. The created by indirect and induced effects for a population results are consistent with people total of 220 supported jobs (i.e., 2.2 × 100). IO not wanting to be near coal mining, perhaps results should use the words jobs supported due to environmental reasons. Overall, these rather than jobs created because it captures effects offset any positive gross economic the gross effect plus other (usually) offsetting impacts of coal mining, compounding the negative employment effects. There are also overestimation of economic impacts by IO 119 models. production intensities (location quotients), econometric estimation, or surveys, but The presence of agglomeration economies regardless of the approach, all are prone to can lead IO models to underestimate multiplier measurement error. Commercial IO models can effects. For example, positive feedback loops produce literally hundreds of direct, indirect, between customers and suppliers can create and induced effects by industry, but such positive linkages that increase the size of estimates can give inexperienced users a false economic impacts within an agglomeration. sense of precision in the results because unlike The best examples are found in the New statistical analysis, there is no indication of the Economic Geography (NEG) literature (e.g., degree of precision, e.g., standard errors. Krugman 1991). Yet, even in NEG models, congestion effects eventually limit the size of a Statistical Approaches. There are also city as it becomes crowded, polluted, imposes statistical or econometric estimates of the longer commutes, etc. multiplier-impacts of the energy industry, which is a more recent innovation. Statistical Other basic shortcomings of IO models methods have the advantage of taking into are unavoidable. For one, the underlying account all of the net positive and negative coefficients such as IO technical coefficients offsetting effects and can be casual estimates that show how much input from industry if done correctly. Studies typically find energy- i is used in the production of industry j. In sector and average-industry multiplier effects the case described below, these production- to range between about 1.3 and 2 for county- technology coefficients for each local area level economic impacts (e.g., Tsvetkova and are estimated using national IO coefficients Partridge 2016), magnitudes that are similar or with some relatively minor adjustments—i.e., slightly smaller than IO estimates. For larger assuming the average national technology geographical areas such as metropolitan areas, applies to every county despite the vintage states, or the entire nation, multipliers tend of the capital investments. Lazarus et al. to increase as economic spillovers generate (2002) find evidence that errors in measuring added economic activity through commuting or local production technology is a key reason purchases of inputs. for commercial IO models such as IMPLAN to inaccurately estimate economic impacts. Coal Industry Supply-Chain in the Broader Another method is to survey local businesses U.S. Economy. In the case of the coal industry, regarding their production technologies and one key input is heavy coal-mining equipment, purchasing decisions. While survey approaches which tends to be manufactured throughout capture local production idiosyncrasies, they the world. Such heavy-mining equipment still involve measurement error and are costly. is not typically available in geographically remote, sparsely populated areas such as There are a host of other coefficients and central Appalachia or the PRB (e.g., Caterpillar’s parameters that must be estimated in an IO main manufacturing site is in Peoria, IL). model. Foremost is the regional purchase In the U.S., the big heavy equipment that coefficient (RPC), which estimates how much extracts and transports coal around a mining input purchases by a particular industry site is manufactured by the (relatively small) can be provided locally. The RPC can be Construction Machinery Manufacturing estimated multiple ways such as using relative industry (NAICS 333120) for which coal mining Socioeconomic Transition in the Appalachia Coal Region Some Factors of Success 120 is just a sliver of its output, generating only plants are not located at the coal mine itself, modest indirect employment effects. The other though there are exceptions such as at Colstrip, main coal mining equipment manufacturing Montana. As described below, coal-burning industry is Mining Machinery and Equipment power plants have been under pressure for the Manufacturing (NAICS 333131), which includes last decade due to low natural gas prices and machinery for all mining except oil and gas. the falling cost of alternative electricity sources This industry only employed 10,000 U.S. such as solar and wind. Further weighing workers in 2019. There are other industries down coal is the large number of coal-burning that supply inputs for coal mining, but they all power plants built 40 or more years ago, which employ small numbers across the U.S. Thus, are highly inefficient and environmentally the relative impact of coal mining in terms problematic, leading to wide-scale closures of U.S. employment up the supply chain is of older coal power plants. This engendered limited, given that there are approximately 160 a precipitous decline in fossil-fuel power- million Americans in the labor force. plant employment from 137,000 in 2009 to 82,000 a decade later. Much of these job losses Three large buyers of coal are the foundry reflect simple efficiency gains associated with industry for iron and steel (NAICS 33111), Other reallocation to modern technology that is also Petroleum and Coal Products Manufacturing less labor-intensive. Some job losses may stem (NAICS 32419), and fossil fuels electric power from the fact that alternative energy generation plants (NAICS 221112). The iron and steel may take place in locations far from traditional industries have increasingly dispersed across power plants. Nonetheless, reallocating away the country with the rise of mini-mills in from coal-electricity power generation to other the last 40-50 years. Even so, foundries only sources may have a small net effect on U.S. employed 64,000 in the U.S. in 2019 and it is employment because the jobs lost in the coal- unclear how, if at all, it would be affected if electricity supply chains would be offset to its coal inputs were ever cut off. Likewise, some degree by employment gains in renewable Other Petroleum and Coal Products are also energy or natural gas-electricity supply chains. dispersed throughout the country. Examples are products produced in coke ovens such Case Study of Coal-mining Multipliers in as asphalt and roofing materials. It is also Virginia. To give a sense of the size of localized uncertain how these products would be impacts on coal communities from changes in substituted if coal was no longer available. coal mining employment, we examine the case For example, other materials may replace of far southeast Virginia coal country. Virginia asphalt, whereas shingles can use a host of has 7 counties that produce coal, three of which substitutes such as wood or petrochemical – Buchanan, Wise, and Dickenson counties – plastics. The employment effects are further account for well over 80% of the state’s total limited because coal is not necessarily the coal production. The other four counties are Lee, main feedstock for some firms in this industry Russell, Taxewell, and Scott Counties (Farren including those who produce lubricants. Even and Partridge, 2015). All are part of the ARC. so, this industry only employed about 16,000 in 2019. Farren and Partridge (2015) analyzed the impact of the coal industry on Virginia’s budget Power plants serve as the main downstream for state and for local “coal” counties. To do buyer of coal. Generally, coal-burning power this, they estimated coal-mining’s impact on 121 Virginia’s economy, both in coal country and region is associated with supporting an output the state. As noted above, as the geographical increase of $1.29 million and $1.62 million area expands, the multiplier increases because for the 3-county region and for the state, of spillovers such as from coal-mining input respectively. Note, the results likely overstate purchases in (say) Richmond, VA. However, IO multipliers and estimates have statistical error. models do not pick up displacement effects. For example, if there is a new factory that opens up in southeastern Virginia coal country to supply the coal industry, it may put out of business a competing factory outside of coal country that also supplies the coal industry. Farren and Partridge (2015) (herewith FP) used proprietary IO software IMPLAN to estimate these effects. IMPLAN is a popular, relatively low-cost software, based on IO models and can be purchased for states, metropolitan areas, and for multiple or single counties. As described above, IMPLAN like other commercial models lack any notion of “uncertainty” of the estimates (IMPLAN could incorporate a Monte Carlo framework to bootstrap standard errors). FP considered coal-mining effects for three regions: (1) all of Virginia; (2) the broader “7-county” coal region; and (3) the narrower “3-county region” most heavily coal-dependent. Specifically, FP assessed the impact of a $1 million increase in coal production in 2011 on the three Virginia study regions. The expected pattern emerges, namely that multipliers are larger when expanding from the 3-county core coal mining region to the 7-county region, and then to the entire state. For example, the employment multiplier is 2.60 for Virginia and 1.74 for the core 3-county region, suggesting that 100 new coal mining jobs is respectively associated with a total increase of 260 and 174 jobs supported by coal in Virginia and in the core 3 coal counties. For output multipliers, FP estimate that a $1 million increase in coal-mining output in the core 3-county coal Socioeconomic Transition in the Appalachia Coal Region Some Factors of Success 122 APPENDIX D Regression Model Description and Results 123 The following briefly describes the regression ARC regions and Appendix D Figure 2 is a map of model we employed. Our dependent variable the ARC designation of distressed to competitive is the percent change in population between counties to show that coal country has some of period t and period t+1. The sample is the 420 the weakest economies in Appalachia. ARC counties and other counties within 100 miles from the ARC region, yielding a sample The initial log population is a measure of of 1070 counties. Adding the additional buffer agglomeration economies that might support counties provides more variation in outcomes faster local economic growth. Likewise, the and in the explanatory variables and helps indicator for being part of a metropolitan area in ensure we are not estimating an equation on a 1973 is another measure of both agglomeration selected group of counties that are lagging by economies and more generally for rural areas, definition—i.e., the ARC was set up to address whether the county has commuting opportunities lagging development. Thus, we estimate the for their workforce. following very simple regression model: Before describing additional steps in identifying Δpopulationt, t+j = a0 + a1 Mining Employment Sharet + successful transition counties, we briefly discuss a2 (Mining Employment Sharet )2 + a3 Populationt + Metropolitan1973+ the empirical results. Appendix D, Table 1 reports NonAppalachian + Region the descriptive statistics in column (1) and then regression results for 1950-2018, 1980-2018, and Where NonAppalachian is an indicator equaling 2000-2018 in columns (2)-(5). The results for 1 if the county is in the 100 mile buffer zone these models end up being the most important in outside of the ARC’s region. We did not extend our analysis. the buffer farther, say to include the whole country, because we thought that would The regression results are generally expected introduce significant heterogeneity into the and consistent across models. For example, model. Region is a vector of two indicator the mining share coefficient is negative and variables for ARC regions. Population is the log statistically significant in all three cases, while of the initial-period population and Mining the squared mining share coefficient is positive Share of the Labor Force is the share of the and statistically significant in two out of three county’s civilian labor force employed in mining. cases. For the 1950-2018 model, taking the Thus, the mining share variable proxies for the derivative of mining share and its square yields importance or dependence the local county has a positive marginal relationship on population on the mining sector. We caution that these growth with a share > 42 percent, and in the regressions are descriptive, not causal. 2000-2018 model, the marginal association turns positive after the mining share of the labor force The rationale for including the non-ARC surpasses 12 percent (the 1980 mining share indicator is that we anticipate that these squared coefficient was statistically insignificant counties will grow faster than ARC counties, in the 1980-18 model). Not surprisingly given because if not, they likely would have been the rise of the Sunbelt, the ARC South indicator included in the ARC in the first place. The North is highly positive in both periods. Likewise, the and South ARC dummies are the specific ARC MSA indicator shows that metropolitan counties region. The more economically disadvantaged were associated with faster growth. central ARC region is the omitted group. Appendix D Figure 1 provides a map of these Socioeconomic Transition in the Appalachia Coal Region Some Factors of Success 124 Appendix D Figure 1: ARC Defined Regions Source: https://www.arc.gov/assets/maps/related/Subregions_2009_Map.png. (Downloaded May 4, 2020). The regression model merges north and northcentral and southcentral with south for the regression analysis. 125 Appendix D Figure 2: ARC Distress Indicators of Economic Well-Being Source: https://www.arc.gov/research/MapsofAppalachia.asp?MAP_ID=149. (Downloaded May 4, 2020). ARC county rankings from distressed, or least well off, to competitive, or best well off counties. The measures are a weighted average of the county’s unemployment rate, official federal poverty rate, and per-capita market personal income. For more details, follow the ARC link above. Socioeconomic Transition in the Appalachia Coal Region Some Factors of Success 126 App ndix D T bl 1: R r ssion Estim t s for S l ct d Tim P riods (1) Summ r St ts R r ssion r sults 94.46 %Popul tion Growth 1950 to 2018 (215.02) (2) (3) (4) 30.54 %Popul tion Growth 1980 to 2018 Popul tion Growth Popul tion Growth Popul tion Growth (63.53) 1950 to 2018 1980 to 2018 2000 to 2018 6.67 %Popul tion Growth 2000 to 2018 (17.69) %Minin EmpSh r 1950 3.40 -5.34*** (8.33) (-4.18) Minin Sqr1950 79.17 0.064*** (367.51) (2.82) Lo Pop1950 67,616 -0.000181*** (189,522) (-5.20) %Minin EmpSh r 1980 6.97 -3.22*** (6.35) (-4.85) Minin Sqr1980 88.9 0.0260 (196.72) (1.55) Lo Pop1980 88,434 -4.52 -05*** (20,2230 (-5.094) %Minin EmpSh r 2000 0.78 -3.58*** (1.89) (-7.60) Minin Sqr2000 4.19 0.148*** (22.83) (4.57) Lo Pop2000 101,195 7.67 -08 (216,298) (0.0333) M tro1973 0.27 197.4*** 32.27*** 10.11*** (0.45) (7.857) (4.511) (6.290) NonApp l chi n 0.61 10.17 5.40 0.097 (0.49) (0.66) (1.35) (0.092) ARC North 0.16 -2.53 -5.20 -3.19** (0.36) (-0.16) (-1.24) (-2.33) ARC South 0.71 68.57*** 32.33*** 8.10*** (0.45) (4.67) (7.35) (6.21) Const nt 9.75 20.36*** 0.76 (0.60) (3.54) (0.55) R-squ r d 0.18 0.16 0.15 Obs rv tions 1,070 1,070 1,070 Std. dev. in parentheses and Robust t-statistics are in parentheses. The descriptive statistics are in levels, not in logs in the cases that the table states log in the variable names, though the regression results report the results using log as described in the empirical implementation section. *** p<0.01, **p<0.05, *p<0.1 127 APPENDIX E Further Details on “Successful” County Selection Robustness Analysis and Data Socioeconomic Transition in the Appalachia Coal Region Some Factors of Success 128 We test the robustness of our selection U.S. average is exceeded by the ARC average. In methodology by comparing our results to that case, the cell is grey cross-hatched when other variables commonly used to measure county population growth exceeds the U.S. successful post-coal economic development. average but trails the ARC average. To do this, we compare our four “successful transition” counties to other mining-intensive Appendix E Table 2 presents the mining share counties with respect to four key indicators of the labor force for each decade spanning associated with well-being: population growth 1950 to 2010, along with the coal-mining rate, mining and coal employment shares, nonfarm employment share in 2016. Recall median household income, and poverty. that the denominator is the civilian labor force. The mining share is based on the county We use descriptive analysis to compare of residence. Conversely, the coal-mining outcomes of our four counties to the top-10, share is a place-of-work measure based on the median-10, and bottom-10 ranked performers actual county where the worker is employed. (based on population growth) within the 99 The difference between the mining share and mining-dependent ARC counties. Results coal-mining share, besides broader industry are reported below, with the following classifications, is commuting patterns. By color coding (which corresponds to colors considering the county of residence in the in Figure 4.2): yellow denotes the four mining share variable, we account for counties successful-transition coal mining counties, that are mining dependent even if workers orange denotes the top-ten fastest growing are employed elsewhere. Yet, to determine mining-intensive counties (excluding the whether a local coal industry exists in the four successful counties), green denotes the county, we use place-of-work data from the 10 median-performing counties in terms of Census of Mineral Industries. population growth, and blue denotes the 10 worst-performing mining-intensive counties. Median household income is the third category of comparison. As shown in Appendix E Appendix E Table 1 reports the four groupings’ Table 3, median household income in the population growth over various sub-periods. four successful coal-transition counties is The groupings are highlighted in different below levels in the top-10 performing mining colors for ease of comparison. At the bottom counties, but no clear time trend is observed. of each group is that group’s average and Because the top-10 performers tend to be standard deviation. At the very bottom is the more populated, a better comparison is to variable’s average and standard deviation for: the median-performing counties given their all U.S. metropolitan counties (1973 definition), similar populations. In 1960, successful all U.S. nonmetropolitan counties, all ARC transition counties had a median household metropolitan counties (1973 definitions), and income 32% greater than median-performing all ARC nonmetropolitan counties (Table 3 counties, and this margin falls to 23.3% in excludes U.S. averages). Table 2 also shows the 1990, and stabilizes at 22.8% higher in 2018. periods for which the successful transition While the four successful-transition coal counties exceed their respective metro/ counties had higher median-household nonmetro ARC average population growth income over the entire period, the median- rates. These cases are shaded grey. The 2000- performing counties made relative gains over 2010 nonmetro case is the rare one when the time. As noted above, median-performing 129 counties were losing population unlike metropolitan county poverty rate averaged relatively successful coal-transition counties; 12.7 percent (vs. 30.4 percent in 1960) and the this income pattern may therefore reflect nonmetro county poverty rate averaged 19.1 the effects of falling labor supply in median- percent (vs. 45.9 percent in 1960). The overall performing counties rather than a relative 1980 U.S. person rate equaled 13 percent. After improvement in local well-being that attracts 1980, however, progress stagnated as overall migrants. The four successful transition U.S. income inequality began to rise, and counties made impressive gains in median the ARC region actually lost ground. All four household income compared to the bottom-10 comparison groups experienced increased performers, which were hemorrhaging poverty rates between 1980 and 2018. And the population. Falling relative income in relatively weakest-performing counties turned conjunction with falling population implies in the worst performance between 1960 and that labor demand decreased more than labor 2018. supply. One explanation could be that more deindustrialization took place in the poorest- Considered together, this comparison exercise performing mining-intensive locales. As noted suggests that our selection criteria and in the SEM discussion in Appendix C, these methodology yield results that are relatively patterns are unsurprising because average consistent with standard “success” measures. income shows no clear trend due to regional compensating differentials. We also examine performance with respect to alternative socioeconomic variables including Poverty is the final category of comparison. population size, human capital, local economic Appendix E Table 4 shows poverty rates from structure, population age structure, social 1960 to 2018, revealing very high rates, and capital, government capacity and other socio- some surprising patterns. In terms of their economic characteristics. Results are reported levels, median-performing counties had in Appendix E Tables 5-12 below. remarkably high average poverty rates in 1960, averaging 20 percentage points above the average rate in the four successful-transition counties, which at 44.1%, was already alarmingly high. The highest-performing and poorest-performing mining-intensive counties also had high initial poverty rates, respectively averaging 35.6% and 29.8%, both lower than the successful-transition counties. These figures reflect the significant poverty across Appalachia. For comparison, the overall U.S. poverty rate was only 22.2% in 1960. It is therefore not surprising that in 1965, Congress enacted the ARC to address the region’s severe and persistent poverty. Between 1960 and 1980, Appalachia made remarkable progress. In 1980, the average ARC Socioeconomic Transition in the Appalachia Coal Region Some Factors of Success 130 Appendix E Figure 1: 1980-2018: ARC %Population Growth in 1950 and 1980 Mining-Intensive Counties 131 Appendix E Figure 2: 2000-2018: ARC %Population Growth in 1950 and 1980 Mining-Intensive Counties Socioeconomic Transition in the Appalachia Coal Region Some Factors of Success 132 App ndix E T bl 1: Succ ssful nd S l ct d Counti s’ % Popul tion Growth R t 1950-2018 MSA MSA %∆p %∆p %∆p %∆p %∆p %∆p %∆p %∆p Count St t FIPS 73 13 5018 5060 6070 7080 8090 9000 0010 1018 Succ ssful Co l Tr nsition Counti s LAUREL K ntuck 21125 0 0 133.3 -3.5 10.0 42.3 11.4 21.4 11.7 2.2 SEQUATCHIE T nn ss 47153 1 1 159.1 4.0 7.0 35.9 3.0 28.3 24.7 3.9 ATHENS Ohio 39009 0 0 43.8 2.5 16.8 2.8 5.6 4.5 3.8 2.1 NOBLE Ohio 39121 0 0 22.9 -6.5 -5.0 8.5 0.2 24.0 4.2 -1.4 Av r 89.8 -0.9 7.2 22.4 5.1 19.5 11.1 1.7 Std D v 66.5 5.0 9.1 19.7 4.8 10.4 9.8 2.2 T n B st P rformin Minin -Int nsiv Counti s GREENE P nns lv ni 42059 0 0 -18.2 -13.2 -8.5 12.2 -2.3 2.8 -5.1 -3.8 PERRY Ohio 39127 0 1 24.1 -3.9 -1.5 13.1 1.7 8.0 5.9 -0.3 FAYETTE P nns lv ni 42051 0 1 -30.3 -10.8 -8.7 3.1 -8.8 2.3 -7.9 -3.4 PICKENS G or i 13227 0 1 248.2 0.5 8.1 21.1 23.9 59.3 28.1 4.7 SHELBY Al b m 1117 1 1 595.8 5.8 18.4 74.3 49.9 44.2 36.3 8.2 MONONGALIA W st Vir ini 54061 0 1 73.1 -8.5 14.6 17.8 0.6 8.4 17.8 9.2 CUMBERLAND T nn ss 47035 0 0 210.6 1.4 8.4 38.3 21.1 34.7 20.0 4.4 CLEARFIELD P nns lv ni 42033 0 0 -6.7 -5.1 -8.5 12.0 -6.6 6.8 -2.1 -1.8 INDIANA P nns lv ni 42063 0 0 11.2 -2.3 5.4 16.1 -2.5 -0.4 -1.1 -3.3 SCHUYLKILL P nns lv ni 42107 0 0 -28.4 -13.7 -7.5 0.3 -5.0 -1.5 -1.6 -3.0 Av r 89.8 -0.9 7.2 22.4 5.1 19.5 11.1 1.7 Std D v 66.5 5.0 9.1 19.7 4.8 10.4 9.8 2.2 T n M di n P rformin Minin -Int nsiv Counti s LEE Vir ini 51105 0 0 -33.2 -28.5 -21.3 27.7 -5.6 -3.7 8.2 -5.5 LESLIE K ntuck 21131 0 0 -32.6 -29.6 6.2 28.0 -8.3 -9.1 -8.7 -7.5 ST CLAIR Al b m 1115 1 1 227.1 -4.9 10.1 47.4 21.4 29.5 29.0 4.5 FLOYD K ntuck 21071 0 0 -31.0 -22.2 -13.8 35.9 -10.6 -2.6 -7.0 -6.4 FAYETTE W st Vir ini 54019 0 1 -46.5 -25.1 -20.1 17.3 -17.1 -0.8 -3.3 -4.1 CLAY K ntuck 21051 0 0 -10.8 -10.2 -10.9 23.1 -4.4 12.9 -11.1 -5.6 KNOX K ntuck 21121 0 0 3.5 -16.9 -6.2 27.6 -1.9 7.1 0.2 -1.3 LINCOLN W st Vir ini 54043 0 1 -6.2 -9.8 -6.7 25.2 -9.7 3.4 -1.8 -2.9 WHITLEY K ntuck 21235 0 0 13.0 -19.2 -6.5 38.3 -0.2 7.6 -0.3 0.9 GREENBRIER W st Vir ini 54025 0 0 -10.0 -12.3 -6.8 17.4 -7.9 -0.7 3.2 -0.6 Av r 7.3 -17.9 -7.6 28.8 -4.4 4.4 0.8 -2.8 Std D v 79.4 8.5 10.0 9.4 10.3 10.9 11.4 3.7 T n Worst P rformin Minin -Int nsiv Counti s MARION T nn ss 47115 1 1 38.5 2.5 -2.2 18.7 1.8 11.7 1.5 0.8 KANAWHA W st Vir ini 54039 1 1 -22.5 5.5 -9.3 0.8 -10.3 -3.6 -3.7 -3.6 CARBON P nns lv ni 42025 1 1 11.1 -8.1 -4.4 5.4 6.7 3.4 10.7 -1.8 LACKAWANNA P nns lv ni 42069 1 1 -17.8 -8.9 -0.2 -2.6 -3.9 -2.6 0.5 -1.4 SOMERSET P nns lv ni 42111 1 0 -8.4 -5.3 -1.8 6.8 -3.7 2.3 -3.0 -3.5 BELMONT Ohio 39013 1 1 -22.0 -4.4 -3.5 2.0 -13.9 -1.2 -0.1 -2.4 WALKER Al b m 1127 1 1 1.1 -15.0 3.8 22.1 -1.4 4.5 -5.2 -3.8 CAMBRIA P nns lv ni 42021 1 1 -35.8 -3.0 -8.1 -1.9 -11.0 -6.4 -6.1 -6.1 JEFFERSON Ohio 39081 1 1 -30.7 2.8 -3.0 -4.8 -12.3 -8.0 -6.2 -3.5 WAYNE W st Vir ini 54099 1 1 5.2 0.7 -3.6 22.5 -9.5 3.0 -1.5 -3.7 Av r -8.1 -3.3 -3.2 6.9 -5.8 0.3 -1.3 -2.9 Std D v 22.5 6.3 3.7 10.4 6.8 5.9 5.0 1.8 (continued) 133 MSA MSA %∆p %∆p %∆p %∆p %∆p %∆p %∆p %∆p Count St t FIPS 73 13 5018 5060 6070 7080 8090 9000 0010 1018 Av r US M tropolit n (1973) 301.0 32.4 25.0 23.3 13.1 16.5 11.9 5.4 Std D v 465.3 65.9 82.2 27.5 18.4 17.2 14.9 7.8 Av r US Nonm tropolit n (1973) 72.6 0.4 1.1 15.6 2.0 9.6 3.5 -0.2 Std D v 313.6 22.6 17.0 23.8 15.6 15.4 11.8 7.1 Av r ARC M tropolit n (1973) 176.5 9.8 9.3 18.5 6.6 12.7 8.2 1.0 Std D v (B s d on ll ARC m tro counti s) 438.7 15.7 14.9 23.3 20.5 21.7 15.9 7.2 Av r ARC Nonm tropolit n (1973) 44.9 -4.9 0.8 17.5 1.5 10.5 4.5 -1.1 Std D v (B s d on ll ARC nonm tro counti s) 91.0 11.8 11.1 10.5 11.3 13.0 9.9 5.2 Sources: The 1973 and 2013 MSA categories are from the U.S. Census Bureau Historical MSA Classifications. Population data is from U.S. Census Bureau, Population Estimates. Notes: 1. FIPS denotes the county code. MSA73 and MSA 2013 denote whether a county was considered a "metropolitan statistical area" based on the 1973 and 2013 MSA categories from the U.S. Census Bureau Historical MSA Classifications, respectively. 2. The individual categories are the four relatively successful coal-transition counties. The higher-, median-, and low-performing mining counties are from the regression ranking of the 99 residuals from the 1950- 2018 population growth models described in the text. The high-performing are the 10 highest residual cases, net of the successful coal-transition counties, representing “over-performing” mining-intensive counties. Median- and Low-performing mining counties are the middle 10 counties from the 99 ranked residuals and the lowest-10 residuals, reflecting median performing counties and underperforming mining counties. 3. The grey shading in the successful county grouping means it grew faster than the county’s respective ARC average growth rate for metropolitan or nonmetro counties for the period. Grey cross-hatching means that the county’s growth rate was between the U.S. and the ARC growth rate in period when the ARC’s average rate exceeded the U.S. average. Socio conomic Tr nsition in th App l chi Co l R ion Som F ctors of Succ ss 134 App ndix E T bl 2: Succ ssful nd S l ct d Counti s’ Minin /Co l Emplo m nt Sh r 1950-2016 MSA MSA MSH MSH MSH MSH MSH MSH MSH MSH COAL Count St t FIPS 73 13 50 60 70 80 90 00 10 15 16 Succ ssful Co l Tr nsition Counti s LAUREL K ntuck 21125 0 0 5.8 1.8 0.9 9.9 2.8 0.8 3.2 0.8 0.8 SEQUATCHIE T nn ss 47153 1 1 15.8 5.4 6.3 7.8 1.5 0.5 0.0 0.0 0.0 ATHENS Ohio 39009 0 0 11.6 2.4 0.7 5.0 1.7 0.6 0.5 0.0 0.0 NOBLE Ohio 39121 0 0 9.9 5.2 9.9 17.7 7.8 2.9 1.3 0.5 0.7 Av r 10.7 3.7 4.4 10.1 3.4 1.2 1.2 0.3 0.4 Std D v 4.1 1.9 4.5 5.4 3.0 1.1 1.4 0.4 0.4 T n B st P rformin Minin -Int nsiv Counti s GREENE P nns lv ni 42059 0 0 37.5 46.0 17.2 22.9 9.6 6.1 18.2 6.3 6.6 PERRY Ohio 39127 0 1 18.2 8.1 6.8 12.6 5.1 1.6 0.5 1.9 5.0 FAYETTE P nns lv ni 42051 0 1 26.9 4.1 7.7 9.2 2.9 2.1 0.3 0.0 0.0 PICKENS G or i 13227 0 1 11.3 0.0 3.0 6.5 1.6 0.7 0.0 0.0 0.0 SHELBY Al b m 1117 1 1 10.5 3.0 1.5 2.7 0.6 0.5 0.2 0.1 0.2 MONONGALIA W st Vir ini 54061 0 1 25.4 14.0 8.7 9.7 5.5 1.9 0.8 1.0 0.9 CUMBERLAND T nn ss 47035 0 0 11.1 4.8 2.5 5.2 1.4 1.1 0.0 0.0 0.0 CLEARFIELD P nns lv ni 42033 0 0 21.4 10.2 6.5 10.1 5.3 1.7 1.1 1.2 1.4 INDIANA P nns lv ni 42063 0 0 27.6 13.3 7.7 16.1 8.1 2.5 1.3 0.7 0.9 SCHUYLKILL P nns lv ni 42107 0 0 25.8 11.1 3.7 4.4 1.7 0.9 0.6 0.5 0.7 Av r e 21.6 11.5 6.5 10.0 4.2 1.9 2.3 1.2 1.6 Std D v 8.8 13.0 4.5 6.1 3.1 1.6 5.6 1.9 2.3 T n M di n P rformin Minin -Int nsiv Counti s LEE Vir ini 51105 0 0 29.7 6.5 10.0 20.6 9.7 5.3 0.0 0.0 0.1 LESLIE K ntuck 21131 0 0 33.4 75.4 25.8 32.4 29.3 13.2 2.0 1.9 3.3 ST CLAIR Al b m 1115 1 1 11.2 0.0 0.3 2.6 0.3 0.1 0.0 0.0 0.0 FLOYD K ntuck 21071 0 0 49.4 35.6 18.3 24.9 15.6 8.5 2.2 3.6 3.8 FAYETTE W st Vir ini 54019 0 1 49.5 23.1 20.3 18.3 7.0 4.1 1.8 3.8 5.6 CLAY K ntuck 21051 0 0 22.4 26.6 14.1 23.1 12.7 4.3 0.9 0.0 0.1 KNOX K ntuck 21121 0 0 18.5 2.6 3.1 10.4 5.7 1.4 0.9 0.4 0.5 LINCOLN W st Vir ini 54043 0 1 22.0 6.6 8.2 13.2 8.3 4.2 0.4 0.0 0.0 WHITLEY K ntuck 21235 0 0 13.8 0.0 3.2 9.5 4.5 1.4 0.7 1.1 1.2 GREENBRIER W st Vir ini 54025 0 0 19.2 8.3 7.8 9.7 3.9 1.3 0.0 1.6 1.6 Av r 26.9 18.5 11.1 16.5 9.7 4.4 0.9 1.2 1.6 Std D v 13.6 23.4 8.3 9.0 8.2 4.0 0.8 1.5 2.0 T n Worst P rformin Minin -Int nsiv Counti s MARION T nn ss 47115 1 1 13.6 8.4 2.8 4.9 1.4 0.3 0.0 0.0 0.0 KANAWHA W st Vir ini 54039 1 1 10.4 5.2 4.4 5.0 2.0 1.3 1.3 1.1 0.9 CARBON P nns lv ni 42025 1 1 18.4 2.6 1.5 2.7 0.9 0.5 0.0 0.0 0.0 LACKAWANNA P nns lv ni 42069 1 1 12.4 4.0 0.7 1.0 0.2 0.1 0.0 0.0 0.0 SOMERSET P nns lv ni 42111 1 0 26.6 6.9 4.7 12.0 3.2 1.6 2.8 1.1 1.5 BELMONT Ohio 39013 1 1 17.2 6.7 10.9 14.4 4.6 2.6 2.5 1.7 2.1 WALKER Al b m 1127 1 1 24.1 8.8 7.8 17.0 8.5 3.6 2.4 1.2 1.7 CAMBRIA P nns lv ni 42021 1 1 19.9 9.9 5.8 8.4 3.6 0.8 0.1 0.1 0.1 JEFFERSON Ohio 39081 1 1 8.5 2.8 3.0 3.8 1.5 0.7 0.3 0.4 0.6 WAYNE W st Vir ini 54099 1 1 8.2 2.7 1.4 4.7 3.0 1.7 4.5 2.7 4.5 Av r 15.9 5.8 4.3 7.4 2.9 1.3 1.4 0.8 1.1 Std D v 6.4 2.7 3.2 5.4 2.4 1.1 1.6 0.9 1.4 135 MSA MSA MSH MSH MSH MSH MSH MSH MSH MSH COAL Count St t FIPS 73 13 50 60 70 80 90 00 10 15 16 Av r ARC M tropolit n (1973) 4.2 1.5 1.3 3.8 0.9 0.5 1.5 1.5 0.5 Sdt D v (B s d on ll ARC m tro counti s) 6.3 2.4 2.0 3.2 1.3 0.8 1.4 1.3 2.0 Av r ARC Nonm tropolit n (1973) 7.4 4.5 3.7 9.1 2.8 1.5 4.0 3.5 1.1 Std D v (B s d on ll ARC nonm tro counti s) 12.9 10.6 7.3 8.0 5.4 2.8 3.9 2.7 3.8 Sources: The 1973 and 2013 MSA categories are from the U.S. Census Bureau Historical MSA Classifications. 1950-2010 mining employment is derived from U.S. Census Bureau, decennial censuses. 2016 mining employment are from the five-year county average in the American Community Survey. The 2016 coal mining share is derived using coal employment data from the U.S. Census Bureau, County Business Patterns along with estimates from The Upjohn Institute of Employment Research (see the text for details). The denominator is the civilian labor force from the U.S. Census Bureau decennial census, except for 2010 and 2016, which use the five year county average from the American Community Survey. Notes: 1. FIPS denotes the county code. MSA73 and MSA 2013 denote whether a county was considered a "metropolitan statistical area" based on the 1973 and 2013 MSA categories from the U.S. Census Bureau Historical MSA Classifications, respectively. 2. The individual categories are the four relatively successful coal-transition counties. The high-, median-, and low-performing mining counties are from the regression ranking of the 99 residuals from the 1950-2018 population growth models shown in Table 1. The high-performing mining- intensive counties are the 10 highest residual cases, net of the successful coal-transition counties, representing “over-performing” mining-intensive counties. Median- and Low-performing mining counties are the middle 10 counties from the 99 ranked residuals and lowest-10 residuals, reflecting median performing counties and underperforming mining counties. Socio conomic Tr nsition in th App l chi Co l R ion Som F ctors of Succ ss 136 App ndix E T bl 3: Succ ssful nd S l ct d Counti s’ Nomin l M di n Hous hold Incom 1950-2018 Count St t MI1950 MI1970 MI1980 MI1990 MI2000 MI2010 MI2018 Succ ssful Co l Tr nsition Counti s LAUREL K ntuck 1,260 6,088 11,961 18,584 27,015 36,835 39230 SEQUATCHIE T nn ss 1,307 6,445 10,972 19,223 30,959 33,181 51,750 ATHENS Ohio 2,112 8,617 11,839 19,169 27,322 33,836 37,778 NOBLE Ohio 1,841 7,760 14,442 21,617 32,940 39,544 47,456 Av r 1,630 7,228 12,304 19,648 29,559 35,849 44,054 Std D v 416 1,172 1,492 1,344 2,879 2,932 6,670 T n B st P rformin Minin -Int nsiv Counti s Av r 2,258 7,957 14,329 23,314 34,283 45,228 52,484 Std D v 562 760 1,878 5,064 8,374 9,258 10,092 T n M di n P rformin Minin -Int nsiv Counti s Av r 1,687 5,762 10,983 15,933 23,099 31,887 35,870 Std D v 473 1,102 2,027 3,491 5,908 7,771 8,137 T n Worst P rformin Minin -Int nsiv Counti s Av r 2,543 8,618 15,205 22,078 31,282 42,076 46,488 Std D v 656 1,148 1,762 2,030 2,475 4,052 4,748 Av r US counti s 3,300* 8,605 14,313 23,979 34,832 44,973 50,792 Sdt D v - 1,936 3,413 6,612 9,457 12,525 14,456 Av r ARC M tropolit n (1973) 2,368 8,912 15,648 25,727 36,988 46,238 52,066 Std D v (B s d on ll ARC m tro counti s) 723 1,371 2,342 4,882 7,711 9,083 10,044 Av r ARC Nonm tropolit n (1973) 1,710 7,149 12,471 20,353 29,989 37,332 42,667 Std D v (B s d on ll ARC nonm tro counti s) 659 1,405 2,211 4,099 5,567 6,953 7,886 *the national average—i.e., not the national average across counties Sources: The 1973 and 2013 MSA categories are from the U.S. Census Bureau Historical MSA Classifications. 1950-2000 Median Household Income is from the U.S. Census Bureau, Decennial Census. 2010 and 2018 are from the American Community Survey five-year county estimates. Notes: In 1950, average family income is used due to data availability. The individual categories are the four relatively successful coal-transition counties from the text. The high-, median-, and low-performing mining-intensive counties are from the regression ranking of the 99 residuals from the 1950-2018 population growth models described above. The high-performing are the 10 highest residual cases, net of the successful coal-transition counties, representing “over- performing” mining-intensive counties. Median- and Low-performing mining counties are the middle 10 counties from the 99 ranked residuals and lowest-10 residuals 137 Appendix E Table 4: Successful and Selected Counties’ Household Poverty Rate 1950-2018 County State PV1960 PV1970 PV1980 PV1990 PV2000 PV2010 PV2018 Successful Coal Transition Counties LAUREL Kentucky 59.3 39.1 21.1 24.8 21.3 20.4 23.7 SEQUATCHIE Tennessee 51.8 30.0 22.4 22.9 16.5 19.3 16.6 ATHENS Ohio 32.4 29.1 21.6 28.7 27.4 32.2 30.6 NOBLE Ohio 32.8 27.9 13.0 16.4 11.4 14.1 15.4 Average 44.1 31.5 19.5 23.2 19.1 21.5 21.6 Std Dev 13.6 5.1 4.4 5.1 6.8 7.7 7.1 Ten Best Performing Mining-Intensive Counties Average 35.6 24.9 14.4 16.5 13.8 15.5 15.0 Std Dev 12.9 3.6 3.4 4.5 4.9 4.2 4.0 Ten Median Performing Mining-Intensive Counties Average 61.1 43.5 26.0 29.8 26.8 25.1 26.5 Std Dev 13.6 11.1 9.3 8.5 8.2 6.7 8.2 Ten Worst Performing Mining-Intensive Counties Average 29.8 21.0 12.3 15.7 13.9 15.6 16.4 Std Dev 11.3 5.8 4.1 3.8 2.9 3.1 3.2 Average US counties 34.3 24.1 15.8 16.7 15.1 17.1 16.4 Sdt Dev 16.5 9.8 7.3 8.0 8.9 8.2 8.2 Average ARC Metropolitan (1973) 30.4 20.6 12.7 13.6 12.1 14.9 14.5 Std Dev (Based on all ARC metro counties)  11.7 6.0 4.1 4.1 3.4 3.4 3.6 Average ARC Nonmetropolitan (1973) 45.9 30.4 19.1 20.3 17.3 20.1 19.1 Std Dev (Based on all ARC nonmetro counties)  15.5 10.3 7.4 8.1 6.6 5.7 5.7 Sources: The 1973 and 2013 MSA categories are from the U.S. Census Bureau Historical MSA Classifications. The 1960 poverty rate is from the U.S. Department of Agriculture, Economic Research Service special tabulation. 1970, 1980, 1990, and 2000 poverty rates are from U.S. Census Bureau, Decennial Censuses. 2010 and 2018 are from the American Community Survey five-year county estimates. Notes: The individual categories are the four successful coal-transition counties described above. The high-, median-, and low-performing mining counties are from the regression ranking of the 99 residuals from the 1950-2018 population growth models shown in Table 1. The high-performing are the 10 highest residual cases, net of the successful coal-transition counties, representing relatively “over-performing” mining-intensive counties. Median- and Low-performing mining counties are the middle 10 counties from the 99 ranked residuals and the lowest-10 residuals, reflecting median performing counties and underperforming mining counties. Socioeconomic Transition in the Appalachia Coal Region Some Factors of Success 138 App ndix E T bl 5: Succ ssful nd S l ct d Counti s’ Hum n C pit l Sh r (% of popul tion > 25 rs old) CG CG CG CG SC SC SC SC HS HS HS HS Count St t 1950 1980 2000 2018 1950 1980 2000 2018 1950 1980 2000 2018 Succ ssful Co l Tr nsition Counti s LAUREL K ntuck 2.2 6.7 10.6 13.4 4.9 8.5 18.4 27.1 6.3 27.3 34.8 40.6 SEQUATCHIE T nn ss 1.5 5.7 10.2 14.6 2.1 5.0 18.0 28.8 8.3 28.5 38.5 39.2 ATHENS Ohio 6.5 20.3 25.7 29.7 7.8 11.9 23.1 25.0 17.1 35.7 34.2 34.9 NOBLE Ohio 2.6 5.9 8.1 10.3 4.2 7.1 22.7 25.8 18.7 50.3 47.8 49.1 Av r 3.2 9.7 13.7 17.0 4.8 8.1 20.5 26.7 12.6 35.5 38.9 40.9 Std D v 1.9 6.2 7.0 7.5 2.0 2.5 2.4 1.4 5.4 9.2 5.4 5.2 T n B st P rformin Minin -Int nsiv Counti s Av r e 3.6 10.5 16.8 22.5 4.4 8.7 19.6 25.6 14.1 37.0 41.4 39.8 Std D v 1.7 5.5 9.3 10.1 1.3 2.2 3.5 3.4 4.7 7.9 9.4 9.6 T n M di n P rformin Minin -Int nsiv Counti s Av r 2.1 7.2 9.7 13.4 3.8 7.4 17.6 25.2 7.3 25.4 33.3 38.2 Std D v 0.9 1.5 2.5 3.9 1.2 1.8 3.9 3.7 3.0 6.1 4.1 3.8 T n Worst P rformin Minin -Int nsiv Counti s Av r 3.4 8.3 12.9 17.6 4.3 9.1 21.2 27.7 16.2 39.5 42.3 41.8 Std D v 1.3 2.7 3.8 5.1 1.2 1.9 2.2 3.3 4.6 5.9 5.7 4.1 Av r US counti s 4.2 11.5 16.5 21.6 6.5 13.1 26.0 30.5 16.0 34.7 34.4 34.2 Sdt D v 2.4 5.5 7.7 9.4 2.8 4.6 5.8 5.4 7.1 7.3 6.7 7.2 Av r ARC M tropolit n 3.8 10.5 17.1 23.3 4.8 11.0 24.3 29.1 14.3 36.3 36.7 35.6 (1973) Std D v 1.9 3.9 6.7 8.3 1.7 2.8 3.3 2.7 6.7 6.6 7.0 7.0 (B s d on ARC m tro counti s) Av r ARC Nonm tropolit n 3.0 8.2 11.9 16.5 4.4 8.6 20.0 26.5 10.5 30.9 37.5 39.4 Std D v (B s d on ARC nonm tro 1.8 3.8 5.2 6.5 1.6 2.5 4.0 4.0 5.8 7.6 6.2 6.5 counti s) Sources: Educational Attainment is Population data is from U.S. Census Bureau, Decennial Census of Population. Notes: 1. CG: Four-year College Graduate; SC: Some college and/or Associate Degree; HS: High School Graduates share; measured as a share of the population 25 years and above. 2. The individual categories are the four relatively successful coal-transition counties from the text. The high-, median-, and low-performing mining-intensive counties are from the regression ranking of the 99 residuals from the 1950-2018 population growth models described above. The high-performing mining-intensive counties are the 10 highest residual cases, net of the successful coal transition counties, representing relatively “over-performing” mining-intensive counties. Median- and Low-performing mining counties are the middle 10 counties from the 99 ranked residuals and the lowest-10 residuals. 139 App ndix E T bl 6: Succ ssful nd S l ct d Counti s’ Economic Structur (% of Tot l Emplo m nt) AG AG AG AG MF MF MF MF SE SE SE SE Count St t 1950 1980 2000 2018 1950 1980 2000 2018 1950 1980 2000 2018 Succ ssful Co l Tr nsition Counti s LAUREL K ntuck 42.1 20.1 8.8 5.3 6.0 25.0 17.0 10.8 39.9 23.4 18.3 18.4 SEQUATCHIE T nn ss 28.9 17.2 11.0 7.8 16.8 20.6 25.5 8.9 29.0 25.8 27.2 36.1 ATHENS Ohio 13.8 6.6 5.0 4.4 12.3 6.6 4.6 2.2 19.6 16.3 18.6 20.1 NOBLE Ohio 44.6 36.6 25.9 23.3 10.1 16.7 24.3 4.5 42.3 30.6 27.7 35.5 Av r 32.3 20.1 12.7 10.2 11.3 17.2 17.9 6.6 32.7 24.0 23.0 27.5 Std D v 12.3 10.8 7.9 7.7 3.9 6.8 8.3 3.4 9.1 5.2 4.5 8.3 T n B st P rformin Minin -Int nsiv Counti s Av r e 13.0 8.8 6.1 4.5 23.0 18.8 12.4 7.5 17.1 18.3 21.7 24.9 Std D v 9.7 5.7 4.3 3.6 9.6 8.6 5.6 4.1 6.7 5.5 5.7 7.6 T n M di n P rformin Minin -Int nsiv Counti s Av r 25.7 15.8 9.3 7.1 28.0 9.7 7.6 4.9 24.0 22.5 22.8 25.5 Std D v 12.3 14.6 8.4 7.0 14.4 6.3 4.1 4.3 8.9 6.3 7.7 4.2 T n Worst P rformin Minin -Int nsiv Counti s Av r 8.2 5.0 3.7 3.0 17.0 19.8 11.7 7.9 13.1 14.5 19.2 21.1 Std D v 6.7 3.9 2.6 2.0 6.9 7.5 4.3 4.6 4.9 3.3 6.0 4.3 Av r US counti s 12.1* 23.9 16.3 13.0 23.9* 15.6 12.0 8.4 30.9 25.0 26.3 29.3 Sdt D v - 20.1 16.0 13.1 - 11.6 9.6 7.6 13.2 11.3 11.2 10.0 Av r ARC M tropolit n 18.4 12.1 6.8 4.7 4.4 23.6 15.2 9.7 20.0 18.9 21.2 25.0 (1973) Std D v 16.9 15.6 8.7 6.2 6.8 9.1 7.0 5.5 11.9 9.8 9.7 8.8 (B s d on ARC m tro counti s) Av r ARC Nonm tropolit n 32.9 21.9 14.1 11.1 7.7 23.6 17.4 11.2 30.6 24.2 25.8 28.4 Std D v (B s d on ARC nonm tro 19.8 19.3 12.5 9.8 13.5 13.2 11.1 8.8 14.2 10.6 10.2 8.9 counti s) *the national average—i.e., not the national average across counties. Sources: Industry composition and self-employment data is from U.S. Census Bureau, Decennial Census of Population. Notes: 1. AG: Agricultural (including fishing, hunting) employment share; MF: Manufacture employment share; SE: Self- employment share, all measured by place of residence. 2. The individual categories are the four relatively successful coal-transition counties from the text. The high-, median-, and low-performing mining-intensive counties are from the regression ranking of the 99 residuals from the 1950-2018 population growth models described above. The high-performing mining-intensive counties are the 10 highest residual cases, net of the successful coal transition counties, representing relatively “over-performing” mining-intensive counties. Median- and Low-performing mining counties are the middle 10 counties from the 99 ranked residuals and the lowest-10 residual. Socio conomic Tr nsition in th App l chi Co l R ion Som F ctors of Succ ss 140 App ndix E T bl 7: Succ ssful nd S l ct d Counti s’ A Structur U18 U18 U18 U18 065 065 065 065 Count St t 1950 1980 2000 2018 1950 1980 2000 2018 Succ ssful Co l Tr nsition Counti s LAUREL K ntuck 46.7 32.4 25.4 23.2 7.2 10.8 11.5 15.7 SEQUATCHIE T nn ss 47.7 31.7 24.6 21.3 6.6 10.7 12.3 19.5 ATHENS Ohio 33.9 23.3 18.3 14.9 11.1 9.6 9.3 12.1 NOBLE Ohio 35.3 30.4 22.6 18.4 14.1 14.9 13.1 25.9 Av r 40.9 29.5 22.7 19.5 9.8 11.5 11.5 18.3 Std D v 6.4 3.6 2.7 3.2 3.0 2.0 1.4 5.1 T n B st P rformin Minin -Int nsiv Counti s Av r e 38.7 28.2 22.7 19.8 8.2 12.4 15.0 18.8 Std D v 4.4 2.9 2.7 2.3 1.4 2.4 3.8 4.6 T n M di n P rformin Minin -Int nsiv Counti s Av r 48.1 32.3 24.1 21.8 5.8 11.4 13.4 17.6 Std D v 4.2 2.6 1.6 1.6 1.1 2.1 2.2 2.2 T n Worst P rformin Minin -Int nsiv Counti s Av r 37.6 27.6 22.2 20.1 7.7 13.2 17.2 19.9 Std D v 5.3 2.0 0.9 1.1 1.5 1.7 2.2 0.9 Av r US counti s 38.1 29.5 25.6 22.3 8.7 13.2 14.6 18.4 Sdt D v 5.2 3.7 3.4 3.5 2.6 4.2 4.2 4.5 Av r ARC M tropolit n (1973) 38.5 28.6 23.8 21.2 7.5 11.3 14.2 18.2 Std D v (B s d on ARC m tro counti s) 4.7 2.4 2.1 2.2 1.8 2.2 3.3 2.9 Av r ARC Nonm tropolit n 41.8 29.5 23.8 20.8 8.0 12.6 14.6 19.3 Std D v (B s d on ARC nonm tro counti s) 5.1 2.8 2.3 2.6 2.2 2.2 2.6 3.5 Sources: Age structure is from U.S. Census Bureau, Decennial Census of Population. Notes: 1. U18: Population share younger than 18 (in the case of 1950, younger than 20 and we use U20); O65: population share older than 65; measured as a share of the county’s population. 2. The individual categories are the four relatively successful coal-transition counties from the text. The high-, median-, and low-performing mining-intensive counties are from the regression ranking of the 99 residuals from the 1950-2018 population growth models described above. The high-performing mining-intensive counties are the 10 highest residual cases, net of the successful coal transition counties, representing relatively “over-performing” mining-intensive counties. Median- and Low-performing mining counties are the middle 10 counties from the 99 ranked residuals and the lowest-10 residuals. 141 App ndix E T bl 8: Succ ssful nd S l ct d Counti s’ Dist nc to N r st MSA, Am nit L v l, GINI ind x, Un mplo m nt R t nd R ci l Composition Count St t DISTMSA AMENITY RUCODE TOPO GINI UNEMP WHITE Succ ssful Co l Tr nsition Counti s LAUREL K ntuck 85.5 4.0 7 19 0.47 8.1 97.0 SEQUATCHIE T nn ss 16.4 3.0 6 16 0.40 5.6 97.1 ATHENS Ohio 20.7 3.0 4 18 0.51 7.7 90.6 NOBLE Ohio 30.7 4.0 8 18 0.45 4.8 92.5 Av r 38.3 3.5 6.3 17.8 0.46 6.5 94.3 Std D v 27.7 0.5 1.5 1.1 0.04 1.4 2.8 T n B st P rformin Minin -Int nsiv Counti s Av r e 24.1 3.9 4.4 17.7 0.45 6.3 92.36 Std D v 17.1 0.3 2.2 1.7 0.03 1.1 4.70 T n M di n P rformin Minin -Int nsiv Counti s Av r 54.5 3.2 7.1 18.7 0.47 8.4 94.9 Std D v 25.4 0.4 2.0 1.7 0.03 2.0 3.2 T n Worst P rformin Minin -Int nsiv Counti s Av r 7.7 3.5 2.8 17 0.45 6.6 93.2 Std D v 8.1 0.5 1.2 1.8 0.02 1.2 2.6 Av r US counti s 48.7 3.5 5.5 8.9 0.45 6.1 82.6 Sdt D v 50.3 1 2.7 6.6 0.04 3.6 17.1 Av r ARC M tropolit n (1973) 9.0 3.6 2.4 15.2 0.45 5.6 88.2 Std D v (B s d on ARC m tro counti s) 8.4 0.5 1.6 4.6 0.03 1.4 11.1 Av r ARC Nonm tropolit n 39.6 3.5 6.5 15.5 0.45 6.9 90.2 Std D v (B s d on ARC nonm tro counti s) 26.9 0.6 2.1 5.0 0.03 2.4 12.1 Sources: DISTMSA is from author calculations; AMENITY RUCODE, TOPO are from USDA Economic Research Service; GINI, UNEMP, and WHITE are from the U.S. Department of Labor for the Unemployment Rates and U.S. Census Bureau, American Community Survey for the GINI and percent of the population that is White. Notes: 1. DISTMA denotes distance in miles from the population weighted centroid of the county to the population-weighted centroid of the nearest metropolitan area. AMENITY index is a one to seven scale reflecting the county’s level of natural amenities (seven is highest amenities). The amenity measure is calculated by the U.S. Department of Agriculture (USDA). RUC denotes the 2013 USDA rural-urban codes (RUC), expressed on a 1 to 9 scale, 1 being the most urban, 9 being the most rural. TOPO is the topography index expressed on a scale of 1 (flat) to 21 (most mountainous) used to calculate the USDA AMENITY measure. See the text for more details of variable definitions. 2. The individual categories are the four relatively successful coal-transition counties from the text. The high-, median-, and low-performing mining-intensive counties are from the regression ranking of the 99 residuals from the 1950-2018 population growth models described above. The high-performing mining-intensive counties are the 10 highest residual cases, net of the successful coal transition counties, representing relatively “over-performing” mining-intensive counties. Median- and Low- performing mining counties are the middle 10 counties from the 99 ranked residuals and the lowest-10 residuals. Socio conomic Tr nsition in th App l chi Co l R ion Som F ctors of Succ ss 142 App ndix E T bl 9: Succ ssful nd S l ct d Counti s’ Popul tion 1950-2018 Pop Pop Pop Pop Pop Pop Pop Pop Count St t MSA73 MSA13 1950 1960 1970 1980 1990 2000 2010 2018 Succ ssful Co l Tr nsition Counti s LAUREL K ntuck 0 0 25,797 24,901 27,386 38,982 43,438 52,715 58,891 60,180 SEQUATCHIE T nn ss 1 1 5,685 5,915 6,331 8,605 8,863 11,370 14,173 14,730 ATHENS Ohio 0 0 45,839 46,998 54,889 56,399 59,549 62,223 64,592 65,936 NOBLE Ohio 0 0 11,750 10,982 10,428 11,310 11,336 14,058 14,643 14,443 Av r 22,268 22,199 24,759 28,824 30,797 35,092 38,075 38,822 Std D v 17,829 18,378 22,058 22,943 24,808 26,152 27,428 28,084 T n B st P rformin Minin -Int nsiv Counti s Av r e 74,682 68,234 66,445 74,906 76,117 84,166 90,711 92,092 Std D v 68,156 58,958 53,088 51,844 47,397 48,998 54,874 57,600 T n M di n P rformin Minin -Int nsiv Counti s Av r 36,150 29,206 26,244 33,640 32,051 33,953 35,258 34,757 Std D v 19,368 14,056 10,734 13,035 12,196 14,962 19,738 20,967 T n Worst P rformin Minin -Int nsiv Counti s Av r 115,316 111,837 106,853 109,034 101,029 99,030 96,997 93,957 Std D v 86,670 85,502 79,816 76,205 69,421 65,495 63,732 61,928 Av r US M tropolit n (1973) 157,499 198,717 231,722 258,450 287,394 325,204 357,344 381,580 Std D v 350,511 419,829 466,725 481,156 536,714 590,823 626,439 664,334 Av r US Nonm tropolit n (1973) 204,867 21,048 21,887 26,683 28,373 31,982 34,818 35,911 Std D v 18,127 19,952 22,032 44,037 52,317 61,409 70,074 76,911 Av r ARC M tropolit n (1973) 104,267 113,650 118,781 127,917 130,733 141,819 153,334 158,528 Std D v (B s d on ll ARC m tro counti s) 186,759 201,229 199,853 187,368 177,661 180,127 187,392 196,526 Av r ARC Nonm tropolit n (1973) 29,273 28,342 28,698 32,893 33,192 36,204 38,194 38,182 Std D v (B s d on ll ARC nonm tro counti s) 26,055 25,839 26,083 28,280 28,218 30,192 32,543 33,435 Sources: The 1973 and 2013 MSA categories are from the U.S. Census Bureau Historical MSA Classifications. Population data is from U.S. Census Bureau, Population Estimates. Notes: 1. MSA73 and MSA 2013 denote whether a county was considered a "metropolitan statistical area" based on the 1973 and 2013 MSA categories from the U.S. Census Bureau Historical MSA Classifications, respectively. 2. The individual categories are the four relatively successful coal-transition counties from the text. The high-, median-, and low-performing mining-intensive counties are from the regression ranking of the 99 residuals from the 1950-2018 population growth models described above. 2. The high-performing mining-intensive counties are the 10 highest residual cases, net of the successful coal transition counties, representing relatively “over-performing” mining-intensive counties. Median- and Low-performing mining counties are the middle 10 counties from the 99 ranked residuals and the lowest-10 residuals. 143 App ndix E T bl 10: Succ ssful nd S l ct d Counti s’ Soci l C pit l M sur s SK ASSN PVOTE RESPN NCCS Count St t 2014 2014 2012 2010 2014 Succ ssful Co l Tr nsition Counti s LAUREL K ntuck -1.8 0.7 0.5 0.8 154 SEQUATCHIE T nn ss -1.5 0.5 0.6 0.8 50 ATHENS Ohio -0.5 1.0 0.6 0.7 446 NOBLE Ohio -0.1 1.5 0.7 0.7 70 Av r -1.0 0.9 0.6 0.7 180 Std D v 0.7 0.4 0.0 0.0 158.5 T n B st P rformin Minin -Int nsiv Counti s Av r e -0.5 1.3 0.6 0.8 400.3 Std D v 0.4 0.3 0.1 0.0 202.5 T n M di n P rformin Minin -Int nsiv Counti s Av r -1.5 0.9 0.5 0.7 119.0 Std D v 0.6 0.3 0.1 0.1 74.3 T n Worst P rformin Minin -Int nsiv Counti s Av r -0.3 1.5 0.6 0.7 476.0 Std D v 0.8 0.4 0.1 0.1 372.8 Av r US Counti s 0.0 1.4 0.7 0.7 463.2 Std D v 1.3 0.7 0.1 0.1 1,399 Av r ARC M tropolit n (1973) -0.5 1.2 0.6 0.8 724.5 Std D v (B s d on ll ARC m tro counti s) 0.6 0.4 0.1 0.1 1,218 Av r ARC Nonm tropolit n (1973) -0.7 1.2 0.6 0.7 159.5 Std D v (B s d on ll ARC nonm tro counti s) 0.8 0.4 0.1 0.1 155.9 Source: Data from Rupasingha et al. (2006) Notes: 1. SK2014 is an overall measure of Social capital index; ASSN2014 is the total number of religious, civic, business, political, professional, labor, and recreational establishment divided by population per 1,000; PVOTE2012: Voter turnout in the 2012 election; RESPN2010 is the 2010 Census response rate; NCCS2014 is the number of non- profit organizations without including those with an international approach. 2. The individual categories are the four relatively successful coal-transition counties from the text. The high-, median-, and low-performing mining-intensive counties are from the regression ranking of the 99 residuals from the 1950-2018 population growth models described above. The high-performing mining-intensive counties are the 10 highest residual cases, net of the successful coal transition counties, representing relatively “over-performing” mining-intensive counties. Median- and Low-performing mining counties are the middle 10 counties from the 99 ranked residuals and the lowest-10 residuals. Socio conomic Tr nsition in th App l chi Co l R ion Som F ctors of Succ ss 144 App ndix E T bl 11: Succ ssful nd S l ct d Counti s’ Gov rnm nt C p cit Count St t Fisc lAuto R v nu P rC pit Exp ndP rC pit Fisc lStr ss Succ ssful Co l Tr nsition Counti s LAUREL K ntuck 0.5 2,332.0 2,050.8 1.1 SEQUATCHIE T nn ss 0.4 2,487.0 1,901.0 1.3 ATHENS Ohio 0.5 3,599.6 3,639.5 1.0 NOBLE Ohio 0.5 3,371.3 2,513.1 1.3 Av r 0.5 2,947.5 2526.1 1.2 Std D v 0.0 546.8 681.2 0.1 T n B st P rformin Minin -Int nsiv Counti s Av r e 0.5 3,105.8 2,874.8 1.1 Std D v 0.1 559.1 487.0 0.1 T n M di n P rformin Minin -Int nsiv Counti s Av r 0.4 2,735.2 2,669.2 1.0 Std D v 0.1 390.9 468.7 0.1 T n Worst P rformin Minin -Int nsiv Counti s Av r 0.6 3,488.3 3,155.8 1.1 Std D v 0.1 1,128.5 1,104.3 0.1 Av r US Counti s 0.6 4,649.0 4,108.0 1.1 Std D v 1.4 2,817.9 2,328.6 0.3 Av r ARC M tropolit n (1973) 0.6 3,707.7 3,292.4 1.1 Std D v (B s d on ll ARC m tro counti s) 0.1 1,306.9 1,043.4 0.2 Av r ARC Nonm tropolit n (1973) 0.5 3,364.8 3,012.7 1.1 Std D v (B s d on ll ARC nonm tro counti s) 0.1 1,075.4 986.5 0.2 Source: U.S. Census Bureau, 2012 Census of Local Governments. Notes: 1. The various FY 2012 measures for the local county fiscal variables are: FiscalAuto: Own-Source Revenue/ State & Federal Revenue ($); RevenuePerCapita: Local Revenue per Capita ($); ExpendPerCapita: General Expenditures per Capita ($); FiscalStress: Revenue/Expenditures ($). 2. The individual categories are the four relatively successful coal-transition counties from the text. The high-, median-, and low-performing mining-intensive counties are from the regression ranking of the 99 residuals from the 1950-2018 population growth models described above. The high-performing mining-intensive counties are the 10 highest residual cases, net of the successful coal transition counties, representing relatively “over-performing” mining-intensive counties. Median- and Low-performing mining counties are the middle 10 counties from the 99 ranked residuals and the lowest-10 residuals. 145 App ndix E T bl 12: Succ ssful nd S l ct d Counti s’ Mort lit R t s LE LE AL AL DR DR SU SU MU MU Count St t 1980 2014 1980 2014 1980 2014 1980 2014 1980 2014 Succ ssful Co l Tr nsition Counti s LAUREL K ntuck 73.8 75.5 1.3 1.1 0.6 20.2 13.9 17.7 11.1 5.8 SEQUATCHIE T nn ss 73.2 75.9 2.5 3.0 0.8 14.4 13.2 18.7 11.8 5.9 ATHENS Ohio 73.4 76.9 1.4 2.4 0.5 14.0 14.2 13.7 3.2 3.0 NOBLE Ohio 74.8 81.1 0.9 2.0 0.3 6.6 12.8 13.1 3.2 3.5 Av r 73.8 77.3 1.5 2.1 0.5 13.8 13.5 15.8 7.3 4.6 Std D v 0.6 2.2 0.6 0.7 0.2 4.8 0.5 2.4 4.1 1.3 T n B st P rformin Minin -Int nsiv Counti s Av r e 73.4 77.9 1.4 1.9 0.6 13.3 14.9 16.6 5.2 3.7 Std D v 0.9 1.0 0.6 0.4 0.3 3.7 2.0 2.6 2.6 1.1 T n M di n P rformin Minin -Int nsiv Counti s Av r 72.3 73.5 2.2 2.3 0.7 30.6 15.4 20.3 13.1 7.2 Std D v 0.4 1.6 0.5 0.8 0.2 9.8 2.3 2.2 5.2 1.4 T n Worst P rformin Minin -Int nsiv Counti s Av r 73.0 76.3 1.4 2.1 0.7 19.7 13.6 18.5 6.2 5.4 Std D v 0.8 2.0 0.8 0.9 0.2 7.0 2.0 4.2 3.8 2.7 Av r US counti s 73.8 77.8 2.9 3.1 0.7 9.9 14.9 17.1 7.7 5.0 Sdt D v 1.8 2.4 2.8 2.7 0.5 6.0 3.5 5.7 5.5 3.5 Av r ARC M tropolit n (1973) 73.5 77.2 2.1 2.3 0.7 15.0 14.2 16.5 7.0 5.1 Std D v 0.8 1.6 1.4 0.9 0.3 6.0 2.3 3.4 3.8 2.3 (B s d on ARC m tro counti s) Av r ARC Nonm tropolit n 73.2 76.0 2.4 2.5 0.7 16.9 15.1 18.6 8.3 5.7 Std D v 1.2 2.2 1.2 0.9 0.2 9.7 2.4 3.9 4.3 2.7 (B s d on ARC nonm tro counti s) Source: Institute for Health Metrics and Evaluation (IHME) as provided by Global Health Data Exchange, available at: http://ghdx.healthdata.org/record/ihmedata/united-states-mortality-rates-county-1980-2014. Notes: 1. The variables are measured in 2014. LE: Life expectancy at birth; AL: Age-standardized mortality rate for both sexes combined (deaths per 100,000 population) from Alcohol use disorders; DR: Age-standardized mortality rate for both sexes combined (deaths per 100,000 population) from Drug-use disorders; SU: Age-standardized mortality rates for both sexes combined (deaths per 100,000 population) from Self-harm; MU: Age-standardized mortality rate for both sexes combined (deaths per 100,000 population) from Interpersonal violence. 2. The individual categories are the four relatively successful coal-transition counties from the text. The high-, median-, and low-performing mining-intensive counties are from the regression ranking of the 99 residuals from the 1950-2018 population growth models described above. The high-performing mining-intensive counties are the 10 highest residual cases, net of the successful coal transition counties, representing relatively “over-performing” mining-intensive counties. Median- and Low-performing mining counties are the middle 10 counties from the 99 ranked residuals and the lowest-10 residuals. Socio conomic Tr nsition in th App l chi Co l R ion Som F ctors of Succ ss 146 APPENDIX F Appalachian Development Highway System Maps Socioeconomic Transition in the Appalachia Coal Region Some Factors of Success 148 Appendix F Figure 1: Appalachian Development Highway System Corridors in Ohio 149 Appendix F Figure 2: Appalachian Development Highway System in Kentucky Socioeconomic Transition in the Appalachia Coal Region Some Factors of Success 150 Appendix F Figure 3: Appalachian Development Highway System in Tennessee 151 Supporting Transition in Coal Regions A Compendium of the World Bank’s Experience