Natural Resource Dependence and Rural American Economic Prosperity From 2000 to 2015

Abstract

Research suggests dependence on natural resource development leads to decreases in per capita income, increases in inequality, and elevated poverty. Natural resource development generally takes two forms—extractive (e.g., oil and gas, mining, timber) and nonextractive (e.g., tourism, recreation, real estate). However, research has rarely examined both in tandem. Drawing on the concept of dependence (i.e., overspecialization), the author tests the hypothesis that increasing levels of both forms of development were associated with diminishing returns to economic prosperity—operationalized as per capita income, inequality, and poverty—in rural America over the period of 2000 to 2015. Extractive development exhibited the expected relationship in remote rural counties for all outcomes, while nonextractive development had a generally negative relationship with per capita income, a positive relationship with poverty, and no relationship with inequality. Support for the overall hypothesis was limited due to the returns of nonextractive development being more negative than expected.

Keywords

natural resource dependence resource curse poverty dependency uneven development rural tourism

Natural resource dependence—meaning the overspecialization of a local economy in the sector of natural resources—is commonly believed to lead to anemic economic growth, increased poverty, and higher levels of inequality (Freudenburg, 1992; Havranek et al., 2016; Krannich et al., 2014; Perdue & Pavela, 2012). Previous longitudinal studies of the relationship between natural resource development and economic production have found both negative (Douglas & Walker, 2017; James & Aadland, 2011; Papyrakis & Gerlagh, 2004, 2007) and positive (S. M. Brooks & Kurtz, 2016; J. P. Brown, 2014; Deller & Schreiber, 2012; Havranek et al., 2016) impacts. While negative economic impacts, the so-called “resource curse,” have received significant attention internationally (Papyrakis, 2017), tests of these relationships within the United States have been more limited.

Within the United States, extractive natural resource development, with perhaps the exception of the recent fracking boom, has been on the decline since the 1920s, with extraction’s labor market share dropping by 70% from 1920 to 1990 (Freudenburg & Gramling, 1994), and remaining generally flat since 2000 (Thiede & Slack, 2017). Although extraction now plays a smaller role in the national economy of the United States, nonextractive forms of natural resource development, such as tourism, outdoor recreation, real estate, and amenity development, have continued to grow (English et al., 2000; Green, 2017; Thiede & Slack, 2017). For illustration (I use nonmetropolitan U.S. counties from 2000 to 2015), the average share of county employment in extractive natural resource development was 1.53% in 2000 and 1.90% in 2015. Meanwhile, the average share of county employment in nonextractive development was 5.35% in 2000 and 5.81% in 2015. Highlighting that, on average, nonextractive development represented at least three times as much of rural employment as extraction from the beginning to the end of the 15-year period.

Rural America finds itself in an era where sustainable economic development is especially needed. The time period of 2000 to 2015, the focus of this study, was one of dramatic economic change for both urban and rural America. The period began with the 2001 recession following the dot-com bubble and was immediately followed by the “jobless recovery,” where even though the overall economy recovered, job growth was historically weak (Aaronson et al., 2004). Shortly after this came the Great Recession of 2007 to 2009, which, up to that point in time, represented the most significant financial contraction since the Great Depression. The recovery from the Great Recession was slow for rural America, and poverty, labor force nonparticipation, and underemployment remained elevated through 2015 (D. L. Brown & Schafft, 2019).

In tandem with these broader economic upheavals, rural America experienced continued industrial restructuring, originating in the 1970s, from 2000 to 2015 (D. L. Brown & Schafft, 2019; Thiede & Slack, 2017). While extractive and nonextractive development increased their employment share in rural America from 2000 to 2015, agriculture continued consolidating in many regions and manufacturing declined significantly. Given the stability and growth of extractive and nonextractive natural resource development over this otherwise tumultuous period, rigorous theoretical and empirical analyses of their relationship with economic prosperity is needed to inform future economic development in rural America. In this article, I address this need using spatial fixed effects models to evaluate the county-level relationship between both extractive and nonextractive natural resource development, and three metrics of economic prosperity—per capita income, income inequality, and poverty—in the rural counties of the contiguous United States from 2000 to 2015.

Theoretical Framework

This analysis tests hypotheses suggested by the theoretical framework of dual dependency articulated by Mueller (2020). This integrative theoretical framework of rural natural resource dependence argues the negative impacts of extractive and nonextractive development on economic prosperity should be similar due to the location of the rural resource-rich community in the global capitalist system. In doing so, it draws on four main research traditions: (a) the resource curse literature from resource economics, (b) the resource dependence literature from rural sociology, (c) the literature on the impacts of rural tourism, (d) and the literature on spatially uneven development, dependency theory, and world systems.

The Resource Curse

The resource curse, also known as the Dutch disease, is the unexpected finding that areas with rich natural resources have lower and slower economic growth than neoclassical economic theory would suggest (James & Aadland, 2011; Sachs & Warner, 1995). This body of work is exclusively focused on natural resource extraction, whether that be in the form of mining, oil and gas, or timber. Although far from agreed on, numerous researchers have found an inverse relationship between natural resource abundance and economic growth, especially when measured at the national level (Douglas & Walker, 2017; Gylfason, 2001; Havranek et al., 2016; Papyrakis & Gerlagh, 2004, 2007; Sachs & Warner, 1995). The pathways through which the resource curse is expected to impact economic growth include the way investment in natural resource extraction crowds out other sectors (Sachs & Warner, 1995; Tsvetkova & Partridge, 2017), the fact that natural resource regions do not incur the same learning-by-doing effects as other economies (Matsuyama, 1992), the low returns to education in a natural resource-dominant economy (Gylfason, 2001; Sachs & Warner, 1995), the booms and busts of the extractive sectors (Frankel, 2012), and the rent-seeking and poor institutions’ hallmark of the industry (Frankel, 2012; Gylfason, 2001; Havranek et al., 2016).

Natural Resource Dependence

While economists have generally relied on the resource curse framework, natural resource dependence is the framework commonly used in sociology. Generally using measures like poverty and unemployment as outcome variables, natural resource sociologists have consistently demonstrated the negative impact that dependence on natural resource extraction has on U.S. rural communities (Freudenburg & Wilson, 2002; Humphrey et al., 1993; Krannich et al., 2014; Peluso et al., 1994; Perdue & Pavela, 2012). Where the resource curse literature has focused on natural resource abundance, rural sociology has generally focused on dependence. The term dependence intimates a threshold at which an economy becomes overspecialized in the natural resource sector, meaning that small degrees of specialization are not expected to have the same adverse effects as large degrees of specialization.

Many reasons for the negative outcomes of resource dependence have been discussed in the literature (see Freudenburg, 1992; Humphrey et al., 1993; Krannich et al., 2014; Mueller, 2020; Peluso et al., 1994 for a full review). Those relevant to the theoretical framework of this article include the underinvestment in human capital at the individual and community levels that occurs in resource dependent communities (Humphrey et al., 1993); and the role of power, domination, and natural resource bureaucracy in rural communities where decisions regarding how a local resource base is used, and who profits from that use, are made by those who control the resource base at a detriment to local residents (Humphrey et al., 1993; West, 1994). Additional works include theories of internal colonialism (Peluso et al., 1994), dependency (Humphrey et al., 1993), and world systems (Bunker, 2003) that are central to the framework of dual dependency presented by Mueller (2020) and have argued that rural resource dependent communities in both the United States and abroad exist at the periphery of a global economic system where core, or urban, communities underdevelop the rural, or peripheral, communities to facilitate cheap and easy access to goods and natural resources (Bunker, 2003; Franke, 1969/2007; Humphrey et al., 1993; Wallerstein, 1979/2015).

Impacts of Rural Tourism

Notably absent from the perspectives of either the resource curse or natural resource dependence is the nonextractive development of natural resources (e.g., tourism, recreation, real estate, retirement, amenity development). However, as stated earlier, this form of development represented a larger share, on average, of rural employment from 2000 to 2015. Although it should be clear that these two activities are not mutually exclusive—many regions have active extractive and nonextractive sectors—this form of development does represent a competing, and sometimes alternative, use of the same resource base as extraction. Thus, a deeper theoretical and empirical understanding of the differences and similarities between the two is needed.

At present, research on the impacts of rural tourism in the United States remains contested. One body of largely econometric research suggested that rural tourism and amenity development is a possible vehicle for economic growth in rural America (Deller et al., 2001, 2008), or at least not a net harm (Deller, 2010). Other research has suggested that once spatial clustering or cost of living is controlled for, the impact of rural tourism on economic growth is insignificant (Hunter et al., 2005; Kim et al., 2005). Beyond just the economic impacts, a growing body of research has demonstrated the negative and divisive impacts—such as conflict between long-time residents and new arrivals, as well as between seasonal and long-term residents—that the change wrought by rural amenity development can have on rural communities (Armstrong & Stedman, 2013; Sherman, 2018; M. D. Smith & Krannich, 2000; Ulrich-Schad, 2018; Ulrich-Schad & Qin, 2018).

Dual Dependency

At present, a theoretical integration of the subnational impacts of extractive and nonextractive development on U.S. rural communities and regions does not exist. Therefore, Mueller (2020), integrating the theoretical perspectives of the resource curse, natural resource dependence, and critical perspectives of capitalism, developed and presented the integrative theoretical framework of dual dependency. The framework posits that rural natural resource dependence, meaning the overspecialization of the rural economy in the natural resource sector, is characterized by the formation of a dual dependency—first on the national and global capitalist economy, and second on the resource rich local environment. This theoretical framework distinguishes between natural resource dependence—meaning the general state of overspecialization in the natural resource sectors—and dual dependency, which refers to the two dependency relationships that develop as a result of overspecialization in natural resource sectors.

Foundational to the first form of dependency is the argument that U.S. capitalism represents a microcosm of free markets, wherein its federalist orientation resembles some degree of a macroeconomy, but the constitutional ban on the regulation of interstate commerce by states greatly amplifies the ability of capital to flow across both political boundaries and distance. Thus, the uneven development brought on by capitalism (D. Harvey, 1982/2018), neoliberalism (D. Harvey, 2006), and the rise of the transnational capitalist class (Sklair, 2002) is allowed to accelerate, and rural areas have limited power or legal authority to determine their own fate. Within this system, rural areas in the United States are geographically and structurally positioned in a peripheral position relative to their urban, or core, counterparts (Wallerstein, 1979/2015). This placement fosters exploitation of the periphery, by the core, for their labor and natural resources. This results in the extra local interests of the core fostering underdevelopment in rural areas and directing economic growth in a direction outside of local interest.

This exploitative structure, which exists for all rural America, is exacerbated in the case of natural resources due to the contradiction between the spatially fixed nature of natural resources and the spatially unbound nature of capital. Capital fixed in space is subject to a falling rate of profit due to improving labor conditions, agglomeration effects, and local crises of overaccumulation (D. Harvey, 1982/2018; N. Smith, 1984). Thus, to maintain a high rate of profit capital must constantly be in motion. Other industrial sectors keep capital in motion by regularly moving their investment to a more profitable location as the rate of profit falls (N. Smith, 1984, 2011). This results in the seesaw patterns of development discussed by N. Smith (1984), where capital leaves one place due to falling rates of profit and does not return until the region degrades to a certain level of underdevelopment. However, capitalists are unable to move the natural resource base in the same way they might move a factory. This results in a contradiction between the need for capital to flow freely across space and the static nature of natural resources.

This contradiction creates additional friction for the movement of capital invested in natural resources and results in investments remaining in place longer than they would in other industrial sectors. The strategies required to keep the rate of profit high while remaining fixed in these peripheral regions are what create the negative impacts of natural resource dependence. Ultimately, unlike common perspectives of uneven development—which argue capitalists seek out poor regions for high rates of profit (D. Harvey, 2006; N. Smith, 1984)—in the case of natural resources, capitalists “double-down” on the location, and therefore pursue especially perverse strategies to keep profits high. Importantly, natural resource interests could address the falling rate of profit by fully disinvesting in the region or pursuing technological or regional substitutes to keep profits high (Bunker, 2003). The dual dependency framework, however, posits that in the United States it is far easier for capital to pursue the strategies listed above and foster the negative local outcomes’ hallmark of natural resource dependence.

Under this framework, high levels of both extractive and nonextractive forms of natural resource development are expected to foster similarly negative outcomes, thus highlighting the importance of considering both types of resource use in tandem. These mutually negative outcomes are due to both forms of development sharing the contradiction between generating profit from a static resource base and the need for capital to be in motion to keep profits high. This alignment creates a mutual interest between the sectors to keep the rural area generally the same, thus limiting diverse and broad-based local economic growth while using the resource base in exclusively one manner. Both extractive and nonextractive interests have invested in the rural area, at least in part, due to its current level of development. For example, nonextractive development in the form of rural tourism often requires an area to have a specific character. Thus, diverse and broad-based development can run counter to the preservation of the characteristics that make the location ideal for nonextractive development in the first place. In line with this, both extractive and nonextractive interests will pursue strategies antithetical to diverse and broad-based economic growth to maintain the local characteristics most favorable to their bottom line (Billings & Blee, 2000; Duncan, 2014; Sherman, 2018; Ulrich-Schad, 2018). While the actual use of the resource base may differ, the incentives to exploit and underdevelop the rural area are aligned between extractive and nonextractive interests.

Importantly, this framework does not argue all levels of development in natural resources will result in negative outcomes. The framework of dual dependency draws a distinction between natural resource specialization—where positive economic outcomes occur due to the benefits of regional economic specialization (Kemeny & Storper, 2015)—and natural resource dependence—where negative economic outcomes occur due to overspecialization and the formation of the dual dependencies discussed earlier. Where this exact level of overspecialization occurs will likely vary due to unique county-level characteristics such as local history and political economy. However, we should be able to expect, on average, that the negative impacts will begin to accrue as specialization in the sector increases. This means that the relationships between natural resource development and economic indicators are expected to be nonlinear.

This nonlinear view more appropriately grounds the testing of resource dependence within its conceptual footing. What is meant by this is that if theory says natural resource dependence is when overspecialization and negative outcomes occur, it does not make sense to assess the impacts of resource dependence on economic outcomes; this would be tautological. Rather, we should look across the range of development and see when or if this dependence develops. In the following analysis, I examine a core tenet of the framework of dual dependency by testing the global hypothesis:

Hypothesis 1: The negative county-level relationship between high levels of natural resource development and economic prosperity, as measured by per capita income, local income inequality, and poverty, as well as the formation of natural resource dependence, will be similar for both extractive and nonextractive natural resource development.

Although the ultimate cause of natural resource dependence suggested by (Mueller, 2020) is the contradiction between static natural resources and free-flowing capital, there are several proximate causal mechanisms expected to stem from this contradiction. Mueller (2020) outlined how the formation of dual dependency is expected to negatively impact three indicators of local economic prosperity: per capita income, local income inequality, and poverty. While the analysis conducted in this article only tests the direct relationships between natural resource development and economic outcomes—not testing any of these specific mechanisms—they must be discussed to ensure hypotheses regarding direct relationships are properly justified, formulated, and tested.

Beyond the ultimate cause of the spatial contradiction, seven proximate causal mechanisms were identified as reasons why an area with high levels of natural resource development should expect lower economic prosperity: (1) underinvestment in human capital; (2) the crowding out of growth in other sectors; (3) the undiverse and oligopsonistic nature of rural natural resource-dependent labor markets, wherein the rural economy, already prone to high levels of market concentration, becomes even more concentrated due to the low level of market diversity in the region; (4) the quality of the jobs in natural resources, which tend to be precarious and low paying in the nonextractive sector (Green, 2017) and highly subject to the ebbs and flows of global commodity markets in the extractive sector (Freudenburg, 1992); (5) the actions of local elites, wherein the power elite in a community will capture what economic growth does exist, resulting in limited overall economic gain; (6) rural gentrification due to natural amenity migrants; (7) and the gendered nature of employment within the natural resource sectors, where extractive labor is masculinized and nonextractive labor is feminized (Sherman, 2009), resulting in limited employment opportunities for half of the population when a community is dependent on just one form of natural resource development.

These seven mechanisms are expected to relate, or not relate, to each of the three indicators of economic prosperity in different ways. High levels of natural resource development are expected to negatively impact per capita income due to the underinvestment in human capital, the crowding out of other sectors, the undiverse and oligopsonistic labor market, the quality of jobs, and the gendered nature of employment. Heightened inequality in the face of high levels of natural resource development is expected due to the undiverse and oligopsonistic labor market, the action of local elites, and rural gentrification effects. Finally, poverty is expected to increase at high levels of natural resource development due to the underinvestment in human capital, the crowding out of other sectors, the quality of the jobs, the actions of local elites, and the gendered nature of employment. As stated, these mechanisms and their negative effects are not expected to develop at low levels of specialization but will instead emerge as specialization in the natural resource sectors increases. Drawing from these mechanisms, the outcome-specific hypothesis is stated below:

Hypothesis 2: The county-level share of employment in extractive and nonextractive natural resource industries had a nonlinear relationship with per capita income, poverty, and inequality in U.S. rural counties from 2000 to 2015, with higher levels of specialization in either form of development resulting in diminishing, and ultimately negative, returns.

Method

Data

The data used for this analysis represent all counties in the contiguous United States for the years of 2000, 2010, and 2015—the longest possible time period of analysis that can be conducted while using unsuppressed industry employment data—and come from five sources: the Decennial U.S. Census, the American Community Survey (ACS),¹ the Bureau of Economic Analysis (BEA) Local Area Personal Income and Employment data,² Wholedata: Unsuppressed County Business Patterns data from the W.E. Upjohn Institute for Employment Research,³ and the U.S. Department of Agriculture Economic Research Service Rural–Urban Continuum Codes (RUCC; Economic Research Service, 2020). Demographic characteristics came from both the Decennial Census as well as the ACS 5-year estimates for 2008 to 2012 and 2013 to 2017 and were extracted through the National Historic Geographic Information System hosted by the Integrated Public Use Microdata Series (Manson et al., 2017). BEA data were used to calculate both per capita income to residents and overall employment totals. Wholedata was used to calculate industry-specific employment totals. The RUCC codes, discussed in greater detail in the subsection titled Rural Indicator, were used to determine metropolitan, nonmetropolitan metropolitan-adjacent, and nonmetropolitan remote status. To ensure consistent geographic units, counties that changed boundaries during the study period were collapsed into larger time-consistent geographic areas in both the data set and within a county-level shapefile from Integrated Public Use Microdata Series (Manson et al., 2017). A description of all county boundary adjustments is provided in the appendix, available in the online supplemental material.

Dependent Variables

Per capita income is a common outcome variable when analyzing natural resource development and is used as one of the indicators of economic prosperity in this analysis (Havranek et al., 2016; James & Aadland, 2011; Sachs & Warner, 1995). The hypotheses under examination pertain to impacts to local residents; therefore, instead of using pure per capita income generated within a county, I use the “earnings to residents” values from the BEA (2019). All income was adjusted for inflation and put into 2017 dollars using the Consumer Price Index inflation calculator provided by the Bureau of Labor Statistics. Per capita income was created by dividing the BEA reported income to residents by the BEA reported total population for that county for that year and is reported in thousands of dollars. Income inequality was operationalized in this study as the local Gini Index, which is the conventional Gini coefficient scaled by multiplying it by 100. This means that 0 represents perfect equality and 100 represents perfect inequality.⁴ The poverty rate used in this analysis is the portion of persons for whom poverty was determined in a county, using the official U.S. poverty measure. The poverty rate was created by dividing the number of persons for whom poverty was determined, as reported by National Historical Geographic Information System, by the total county population and then multiplied by 100.

Independent Variables of Interest

Extractive and nonextractive natural resource related development was operationalized as the share of total employees within a county working in either extractive or nonextractive industries. Employment share is used as an indicator for the relative level of natural resource development within a given county. Unlike outcome variables, these measures allow for workers to move across county boundaries for work. This allowance means these measures provide a measure of the level of development in each sector within a county, regardless of where those employees live.

The industries classified as extractive industries included forestry and logging (North American Industry Classification System [NAICS] = 113); fishing, hunting, and trapping (NAICS = 114); support activities for forestry (NAICS = 1153); and mining, quarrying, and oil and gas (NAICS = 21). Only those industries directly involved in the extraction of resources were included in this definition, meaning no processing, manufacturing, or energy production was included. This decision was made to ensure conceptual clarity in the industries considered. Extractive employment share was calculated by dividing the number of employees in these extractive industries by the total number of employees within a county and multiplying by 100. Nonextractive natural resource-related development included accommodation and food services (NAICS = 72); arts, entertainment, and recreation (NAICS = 71); real estate and rental and leasing (NAICS = 531); and scenic sightseeing and transportation (NAICS = 487). Nonextractive employment share was calculated by dividing the number of employees in these nonextractive industries by the total number of employees within a county and multiplying by 100.

Time-Variant Control Variables

Although the use of unit and period fixed effects in this study accounts for unobserved county-level and year-specific heterogeneity, relevant variables that are time variant within units are not controlled for by this model specification. To ensure precise estimates, I adjust for the four confounding variables suggested by the literature to have a causal relationship with both the independent and dependent variables within rural America: population, portion of the population over 65, portion of the population that is Black, and portion of the population that is Latinx (James & Aadland, 2011; Lobao et al., 2016).

The rationale for this set of confounders is as follows: counties increasing in overall population are likely to have lower levels of extractive development and higher levels of economic prosperity (Johnson & Lichter, 2019); counties with an increasing proportion of elderly residents can confound the relationships of interest in two ways—population aging can lead to decreased economic growth resulting in lower natural resource development and lower aggregate economic prosperity (Thiede et al., 2017), and elderly migration into retirement communities can increase nonextractive development and aggregate economic welfare for at least a portion of the population (Poudyal et al., 2008); counties with increasing Black population proportions have historically faced higher levels of resource extraction and environmental injustice, as well as lower economic prosperity due to the deeply embedded racism in American society (Brulle & Pellow, 2006; Duncan, 2014; M. H. Harvey, 2017); and counties with increasing Latinx populations, due to the linked migration networks between natural amenity migrants and Latinx immigrants (Nelson et al., 2009), are more likely to have higher levels of nonextractive development, as well as higher levels of inequality and poverty (Lichter, 2012; Monnat & Chandler, 2017).

Notably absent from this list of controls, but which have been included in other studies on this topic, are the dimensions of education, share of employment in manufacturing, cost of living, and unemployment. These variables were excluded to avoid overadjusting the model by conditioning on, or controlling for, downstream variables (Schisterman et al., 2009). What this means is that it is not appropriate or necessary to control for variables that theory suggests are post “treatment.” As outlined in the section on Dual Dependency, the pathways along which natural resource dependence is expected to influence economic prosperity include the underinvestment in human capital (e.g., education), the crowding out of manufacturing, rural gentrification effects (increases in cost of living), and increase in overall unemployment. Therefore, including these variables would only serve to suppress the true impact of natural resource development while also potentially opening the model to additional omitted confounding variables (Schisterman et al., 2009).

That said, additional attention is warranted related to manufacturing employment share, unemployment, and cost of living. It is possible that the exit of manufacturing in a location could lead to higher unemployment and decreased economic prosperity while raising the employment share of natural resources by default due to the changing denominator in employment share. Furthermore, a secular increase in cost of living could occur and reduce the profitability of natural resource development while changing economic welfare of residents. From this, it could be argued that these three variables represent time-varying confounders. However, within the framework of dual dependency guiding this study, these variables are expected to function as downstream mechanisms. Thus, the role of manufacturing, cost of living, and unemployment could be argued to play a dual role as a possible confounder and as a downstream mechanism; this distinction would occur at a temporal level of precision unavailable in these data. As such, they are not included in the primary models, but are included, along with education, in a robustness check in the section on Robustness Checks.

Rural Indicator

The focus of this analysis is on rural outcomes. As such, the relationship between natural resource development and outcomes is isolated using an interaction technique described below. Given that rurality is not simply a dichotomous trait but a continuum, relationships were estimated for two kinds of rural counties, those which are adjacent to metropolitan areas and those that are not. I divide these counties using the RUCC produced by the U.S. Department of Agriculture Economic Research Service, which divide the dichotomous Office of Management and Budget metropolitan/nonmetropolitan county classification into nine levels based on both population size and adjacency (Economic Research Service, 2020). I combine all nonmetropolitan counties that are adjacent to metropolitan areas into one group (RUCC Codes 4, 6, and 8), and all nonmetropolitan counties not adjacent to metropolitan areas into another (RUCC Codes 5, 7, and 9). I use a fixed definition anchored in the year 2000. For the sake of clarity, I refer to these as nonmetropolitan metro-adjacent and nonmetropolitan remote counties.

Analytic Approach

The use of a spatial econometric model was necessary due to the well-documented spatial clustering of social phenomena across the United States (M. M. Brooks, 2019; Lobao et al., 2007; Thiede et al., 2018), as well as the ease at which changes in one county can influence its neighbors due to permeable boundaries (Chi & Zhu, 2019; Leicht & Jenkins, 2007). If an aspatial model were used, the assumptions of independence between units built into that model would be violated, producing incorrect estimates of model parameters (Chi & Zhu, 2019). Due to the complexity of the models and the use of polynomial terms, I use a spatial lag of X (SLX) model, as opposed to the more complicated model like the Spatial Durbin Model, to ensure interpretable results (Gibbons & Overman, 2012; Vega & Elhorst, 2015).

The hypothesized nonlinear effects were estimated by including both the first and second order (i.e., linear and quadratic) terms of employment shares in the models. A quadratic nonlinear relationship was specified due to theory and prior work suggesting an expected threshold where we see diminishing, and possibly negative, returns from increased specialization of either extractive or nonextractive development (Freudenburg & Wilson, 2002; Stedman et al., 2005).

The desired results for this model were direct effect estimates for the two kinds of rural counties; however, the use of spatial regression means that the exclusion of urban neighbors from the overall model would have been inappropriate. Therefore, I estimated the models using a simple spatial regime approach where a three-level categorical indicator, where the reference group was nonmetropolitan remote, was interacted with both extractive and nonextractive employment shares.

Ultimately, an SLX fixed-effects model with both unit and period fixed effects was used to control for time-invariant unobserved county-level heterogeneity, as well as any secular trends occurring over the study period, while also controlling for spatial dependence and spillovers. The full SLX model is presented in Equation 1, where y_it is the outcome variable of interest; metro represents the three-level metropolitan-status indicator; ex_it represents the share of local employment in the sector of extractive natural resource development; nx_it represents the share of local employment in nonextractive natural resource development; β₁ and β₃ capture the first-order term of the within-county direct effect of extractive and nonextract employment share on y_it, respectively; β₂ and β₄ capture the second order, quadratic within-county direct effect of extractive and nonextract employment share on y_it, respectively; W is a row-standardized first-order queen’s contiguity matrix; θ represents the spatially lagged coefficient for its corresponding β; X is a vector of time-variant control variables; u_it represents the error in the model that has the components µ_i and c_t representing the unit and period fixed effects; and ε_it represents stochastic error. The interpretation of this model is straightforward, as β represents the average direct effect of a change in X_it on y_it and θ represents the average indirect effect of a change in average neighboring levels of X_it on y_it (Golgher & Voss, 2016; Vega & Elhorst, 2015). All models were estimated using cluster robust standard errors at the county level due to the repeated observation of counties over the study period (Angrist & Pischke, 2008; Cameron & Miller, 2015).

\begin{array}{l} y_{i t} = m e t r o * [β_{1} e x_{i t} + β_{2} e x_{i t}^{2}] \\ + m e t r o * [β_{3} n x_{i t} + β_{4} n x^{2}] i t + W θ_{1} e x_{i t} \\ + W θ_{2} e x_{i t}^{2} + W θ_{3} n x_{i t} + W θ_{4} n x_{i t}^{2} + β_{7} X_{i t} \\ + W θ_{7} X_{i t} + u_{i t} u_{i t} = μ_{i} + c_{t} + ε_{i t} \end{array}

(1)

Evaluation of Hypotheses

The global hypothesis regarding similar relationships between extractive and nonextractive development and economic prosperity (Hypothesis 1) was assessed by considering all results for all three outcomes variables in tandem and is discussed at the end of the results. The outcome-specific hypothesis (Hypothesis 2) was evaluated using marginal estimates of changes in the direct effects for each outcome across a constrained range of each form of natural resource employment share. Marginal effects and predicted means were estimated while holding all other model variables constant at their means. Thus, by calculating and plotting the nonlinear predicted marginal means and marginal effects across a constrained range of each form of development, I assessed the shape and significance of the nonlinear relationships between extractive and nonextractive employment shares and per capita income, inequality, and poverty. Given the considerable ambiguity regarding what it means to have a “significant nonlinear effect,” I do not focus on one omnibus test of effects and instead assess the strength and significance of the direct effects as they vary across the selected range of employment share. All tests of statistical significance were evaluated at p < .05. Finally, given the complexity of the models due to the use of both interactions, polynomials, and direct and indirect effects, I only present model results visually in the main body of this manuscript. Full tables are available in the appendix, available in the online supplemental material.

Results

Summary statistics for all variables included in the primary models by metropolitan status are presented in Table 1. After county boundary adjustments, there were a total of 3,073 counties, with 961 classified as nonmetropolitan remote, 1,049 classified as nonmetropolitan metro-adjacent, and 1,063 counties classified as metropolitan. All counties had three observations (i.e., 2000, 2010, and 2015) and there were no missing data.

Table 1.

Summary Statistics for Model Variables by Metropolitan and Nonmetropolitan Across All Years.

	M	SD	Min	Max
Nonmetropolitan nonadjacent: N/county-years = 961/2883
Per capita income to residents (thousands)	21.47	8.29	0.99	115.68
Local Gini coefficient	43.32	3.78	30.75	80.67
Poverty rate	15.85	6.93	0.00	56.69
Extractive employment share	1.80	3.35	0.00	55.33
Nonextractive employment share	5.46	4.40	0.00	54.97
Population (thousands)	17.17	17.71	0.07	179.90
Percentage over 65	18.16	4.68	3.00	39.16
Percentage of White	86.93	16.05	2.92	100.00
Percentage of Black	6.07	13.81	0.00	86.49
Percentage of Hispanic	7.57	13.66	0.00	95.68
Nonmetropolitan adjacent: N/county-years = 1049/3147
Per capita income to residents (thousands)	20.21	5.33	7.70	84.36
Local Gini coefficient	43.05	3.50	32.91	70.33
Poverty rate	16.16	6.03	2.91	50.45
Extractive employment share	1.60	2.88	0.00	36.47
Nonextractive employment share	5.79	4.03	0.00	72.27
Population (thousands)	31.59	27.42	0.69	215.19
Percentage over 65	16.90	3.87	5.21	54.18
Percentage of White	83.93	16.80	9.34	99.52
Percentage of Black	9.50	15.48	0.00	86.92
Percentage of Hispanic	7.64	13.64	0.00	99.18
Metropolitan: N/county-years = 1063/3189
Per capita income to residents (thousands)	26.98	8.67	10.70	95.34
Local Gini coefficient	41.95	3.21	32.44	56.96
Poverty rate	12.90	5.21	2.11	38.59
Extractive employment share	1.09	1.84	0.00	44.45
Nonextractive employment share	7.00	4.22	0.00	90.13
Population (thousands)	237.54	507.87	1.54	10123.25
Percentage over 65	13.94	3.64	1.80	38.30
Percentage of White	81.28	15.26	17.10	99.43
Percentage of Black	10.89	13.69	0.00	81.48
Percentage of Hispanic	8.47	12.11	0.02	95.74

Model Results

Results for both nonmetropolitan adjacent and nonadjacent counties are presented in Figures 1 through 3. Each figure presents plots for both types of nonmetropolitan counties’ predicted marginal means and direct effects for a constrained range of extractive and nonextractive employment share. Marginal predictions and effects of employment share were estimated for every integer from 0 to 30 for all relationships. The range of 0 to 30 was selected as it contains most of the observed values for each form of development. Each figure first presents the marginal predicted means for the relevant combination of employment share variable and outcome variable on the left, and then presents the corresponding marginal direct effects on the right.

Figure 1.

Predicted means and marginal direct effects for per capita income to residents across constrained ranged of natural resource employment share.

Figure 2.

Predicted means and marginal direct effects for Gini index across constrained ranged of natural resource employment share.

Figure 3.

Predicted means and marginal direct effects for poverty rate across constrained ranged of natural resource employment share.

Per Capita Income

In the case of per capita income to residents, extractive development had the expected relationship for nonmetropolitan remote counties only (Figure 1). In nonmetropolitan remote counties, increases in extractive employment share corresponded to larger increases in per capita income at lower absolute levels of employment share than at higher levels. At high levels of extractive employment share the effect tapered and ultimately became nonsignificant. In nonmetropolitan metro-adjacent counties the relationship was not statistically distinguishable from zero across its range. Taken together, these models show support for the expected nonlinear relationship in nonmetropolitan remote counties for extraction, but not in nonmetropolitan metro-adjacent counties. Although we see the expected nonlinear relationships, the effects do not become negative across the range of predicted values; instead, they simply diminish to zero.

The relationship for nonextractive development also varied between remote and metro-adjacent counties. In nonmetropolitan remote counties, low levels of nonextractive employment share were associated with decreases in per capita income. As the absolute level of employment share grew, this negative effect tapered toward zero. In metropolitan-adjacent counties the effect was far more consistent: Increases in nonextractive employment share were associated with decreases in per capita income across the range of predicted values. Although there does appear to be a nonlinear effect, these relationships do not support the hypothesis as the effect either becomes less negative (nonmetro remote) or was a strictly negative relationship (nonmetro metro-adjacent).

Taking all of this in tandem, the hypothesis regarding per capita income (Hypothesis 2) is partially supported for extractive development and not supported for nonextractive development. The hypothesized negative returns were present for nonextractive development, but counter to the hypothesis, they were present across all predicted levels of employment share.

Inequality

In the model of inequality, we again see the expected relationship for extractive development in nonmetropolitan remote areas—wherein there are decreases in inequality at low levels of employment share and the effect tapers at high levels (Figure 2). The relationship is more nuanced for nonmetropolitan metro-adjacent counties. The relationship between extractive employment share and Gini Index is not significantly distinguishable from zero until a relatively high level of extractive employment share, where we then see a positive relationship with increases in extractive employment share corresponding to increases in inequality. Although there is nuance in these findings, they broadly support the outcome-specific hypothesis. High levels of extractive employment share were associated with negligible to negative returns for the reduction of inequality.

In the case of nonextractive development, the relationship between employment share and Gini Index is not statistically distinguishable from zero at any point along the range of predicted values. Thus, the results do not support the hypothesis. Ultimately, the hypothesis regarding inequality (Hypothesis 2) is supported for extractive development and unsupported in the case of nonextractive development.

Poverty

When examining the results for poverty, we again see the expected relationship for extraction in nonmetropolitan remote counties. Increases in extractive employment share corresponded to decreases in poverty at low absolute levels of employment share, and this effect tapered to zero as the absolute level of employment share grew. In nonmetropolitan metro-adjacent areas, the relationship between extractive employment share and poverty was not statistically distinguishable from zero at any predicted value. Thus, the hypothesis of nonlinearity and diminishing returns is supported for nonmetropolitan remote counties, but not metro-adjacent counties. Once again, there were no negative returns at high levels of extractive employment share in remote counties, poverty did not increase, but the effect did diminish to zero.

The results for nonextractive development in some ways mirror those for extraction in the case of poverty. In nonmetropolitan remote counties, there was no statistically significant effect across the range of predicted values, leaving the hypothesis unsupported. However, in nonmetropolitan metro-adjacent counties, there was a significant effect. At low levels, increases in nonextractive employment share resulted in increased poverty in metro-adjacent areas. As the absolute level of nonextractive employment share increased, this effect tapered. While nonlinear, this does not support the study hypotheses. Instead of the relationship becoming more severe across the range of development, the relationship tapered—suggesting a ceiling where further increases in nonextractive development did not continue to correspond with increases in poverty.

Similar to per capita income, the hypothesis regarding poverty (Hypothesis 2) is partially supported for extraction and not supported for nonextractive development. Although there is a significant nonlinear effect for nonextractive development in nonmetropolitan metro-adjacent counties, the relationship gets better, not worse, as absolute levels of employment share rise.

Evaluation of the Global Hypotheses

To assess support for the global hypothesis proposed by the theoretical framework of dual dependency—that the negative county-level relationship between high levels of natural resource development and economic prosperity, as measured by per capita income, local income inequality, and poverty, and the formation of natural resource dependence will be similar for both extractive and nonextractive natural resource development—requires assessing all three models in tandem. High levels of specialization in both forms of development were associated with adverse outcomes for per capita income and poverty; however, the similarities end there.

Extractive development demonstrated the expected curve of diminishing returns for all three outcome variables in nonmetropolitan remote counties, but in metro-adjacent counties the hypothesis was only supported for inequality. Furthermore, only in the case of inequality were actual negative effects observed, as opposed to simply diminishing returns. This indicated significant support for the outcome-specific hypothesis in remote counties and partial support in metropolitan-adjacent counties. Nonextractive development did not support the outcome-specific hypothesis in any model. The relationships between nonextractive employment share and per capita income and poverty were far more severe than expected and the relationships did not take the expected shape. Additionally, the lack of a significant relationship between nonextractive development and inequality did not support the global hypothesis of similar outcomes and highlights the way these forms of development relate to economic prosperity in different ways.

Taking these findings together, there is little support for the global hypothesis as written. For extraction, I find the expected curves, but the relationships do not become negative except in the case of inequality. For nonextractive development I do find negative outcomes to per capita income and poverty at high levels of development, but they are less severe than they would be at lower levels.

Robustness Checks

To assess model robustness to variables that could be viewed as either downstream mechanisms or time-varying confounders, models were also estimated with the variables of percent without a high school diploma, percent with at least a college degree, median owner-occupied housing value, unemployment rate, and manufacturing employment share included. The findings were consistent between the primary models and the models with these variables included. Figures and tables for this robustness check are included in the appendix, available in the online supplemental material. Additionally, models were estimated using a row-standardized rooks first-order contiguity matrix to ensure estimate stability. The conclusions suggested by models were equivalent with the models using the preselected queen’s contiguity matrix and results are provided in the appendix, available in the online supplemental material.

Discussion

This study presents an investigation into the relationship between both extractive and nonextractive development and rural economic prosperity in the United States from 2000 to 2015. This period was one of significant economic upheaval and industrial restructuring in rural America. As both extractive and nonextractive natural resource development showed considerable stability and growth—as measured by employment share—during this otherwise difficult period, it was important to understand how these forms of development related to rural economic outcomes. While I do find support for the hypothesized nonlinear relationships of overspecialization for extraction in nonmetropolitan remote counties, the results were not as evident in those rural areas adjacent to metropolitan centers. Furthermore, nonextractive development did not conform to the theoretical hypotheses, with the relationships being less straightforward than hypothesized under the framework of dual dependency articulated by Mueller (2020).

Nonextractive forms of natural resource development have been suggested as a boon for struggling rural economies (Deller, 2010; Deller et al., 2008). However, these results call into question how effective these forms of development really are for rural America. These results provide support for the qualitative findings of both Sherman (2018) and Ulrich-Schad (2018), showing negative outcomes in the face of increased nonextractive development in rural areas, as well as the quantitative findings of Deller (2010) showing a limited impact of nonextractive development on poverty reduction. When controlling for relevant county characteristics, the share of employment in nonextractive natural resource industries such as tourism, recreation, accommodations, and real estate, generally decreased per capita income to residents either had no effect or raised poverty and had no relationship with inequality. The reasons for this finding remains an empirical question and may stem from the generally low quality of jobs in the sector. Regardless of reason, when considering these findings, it appears that rural natural amenity development may not only be a poor panacea for rural America but may be a direction to be seriously avoided.

While the effects found for extractive forms of development were not as severe as those found for nonextractive development, extractive natural resource employment was not strictly positive for rural economic prosperity. These results support the notion that natural resource dependency develops at high levels of extraction where we see diminishing returns to per capita income, inequality, and poverty. However, given the lack of actual negative returns—generally the effects simply tapered toward zero—it appears remote rural counties are better off with some extraction than none. Thus, while these results both support prior literature demonstrating the resource curse within the United States (James & Aadland, 2011; Perdue & Pavela, 2012) and highlight the possible negative socioeconomic outcomes of natural resource dependence, the use of nonlinear effects highlights the complexity of these relationships and the positive absolute outcomes extraction can have for remote rural counties.

I find the relationship between natural resource development and economic prosperity varies between nonmetropolitan remote and nonmetropolitan metro-adjacent counties. The framework of dual dependency argues one of the primary reasons overspecialization develops is the peripheral nature of rural communities. That I only find consistent support for study hypotheses for extraction in the remote rural counties supports this element of the framework. The fact that nonextractive development did not show a clear trend between the two types of rural counties highlights the work left to do before we will fully understand the trajectory of nonextractive natural resource development in the United States. Although Mueller (2020) may have been correct in the prediction that both forms of development ultimately result in negative dependent relationships for rural communities, it does not appear as simple as assuming a similar trajectory between extractive and nonextractive natural resource development.

Limitations

There remain two limitations to acknowledge. The use of the county as the spatial unit, although common and ideal for subnational research (Hooks et al., 2004; Lobao & Kraybill, 2005), poses limitations for this study due to the variability in size and structure of counties throughout the United States. Counties have a large degree of internal heterogeneity. Treating them as a single unit of analysis necessarily removes nuance from findings and assumes impacts will be shared equally. For example, within a county that appears dependent on extraction you may have communities that are, in fact, dominated by the sector, and other communities with little or nothing to do with extractive activities. Future work using restricted data will be needed to understand these effects at a community level.

Finally, I elected to group industrial activities into either extractive or nonextractive categories. This approach, while necessary for the analysis at hand, is rather coarse when considering the intricacies of context and the varying trajectories of the component sectors over the time period. It will be crucial for future work to explore each kind of extractive or nonextractive development separately, while also looking at varying contextual and regional effects if we are to fully understand the benefits and drawbacks of these forms of economic development.

Supplemental Material

sj-pdf-1-edq-10.1177_0891242420984512 – Supplemental material for Natural Resource Dependence and Rural American Economic Prosperity From 2000 to 2015

Supplemental material, sj-pdf-1-edq-10.1177_0891242420984512 for Natural Resource Dependence and Rural American Economic Prosperity From 2000 to 2015 by J. Tom Mueller in Economic Development Quarterly

Footnotes

Acknowledgements

I would like to thank Ann R. Tickamyer, Brian Thiede, Kathryn Brasier, and Alan R. Graefe for their feedback and guidance on this article.

Declaration of Conflicting Interests

The author declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Funding

The author disclosed receipt of the following financial support for the research, authorship, and/or publication of this article: This material is based on work supported by the National Science Foundation under Grant No. 1903924: Effects of Natural Resource Dependence. Any opinions, findings, and conclusions or recommendations expressed in this material are those of the author and do not necessarily reflect the views of the National Science Foundation.

ORCID iD

J. Tom Mueller

Supplemental Material

Supplemental material for this article is available online.

Notes

Author Biography

J. Tom Mueller is an assistant professor of sociology at Utah State University. His work focuses on spatial inequality, well-being, and the environment, with a particular emphasis on rural regions of the United States.

References

Aaronson

Rissman

E. R.

Sullivan

D. G.

(2004). Assessing the jobless recovery. Economic Perspectives, 28(2), 2-21. https://www.chicagofed.org/publications/economic-perspectives/2004/2qtr-2004-part-1

Angrist

J. D.

Pischke

J.-S.

(2008). Mostly harmless econometrics: An empiricist’s companion. Princeton University Press. https://doi.org/10.2307/j.ctvcm4j72

Armstrong

Stedman

R. C.

(2013). Culture clash and second home ownership in the U.S. Northern Forest. Rural Sociology, 78(3), 318-345. https://doi.org/10.1111/ruso.12010

Billings

D. B.

Blee

K. M.

(2000). The road to poverty: The making of wealth and hardship in Appalachia. Cambridge University Press.

Brooks

M. M.

(2019). The advantages of comparative LISA techniques in spatial inequality research: Evidence from poverty change in the United States. Spatial Demography, 7(2-3), 167-193. https://doi.org/10.1007/s40980-019-00052-4

Brooks

S. M.

Kurtz

M. J.

(2016). Oil and democracy: Endogenous natural resources and the political “resource curse.” International Organization, 70(2), 279-311. https://doi.org/10.1017/S0020818316000072

Brown

D. L.

Schafft

K. A.

(2019). Rural people and communities in the 21st century: Resilience and transformation (2nd ed.). Polity Press.

Brown

J. P.

(2014). Production of natural gas from shale in local economies: A resource blessing or curse? Economic Review: Federal Reserve Bank of Kansas City. https://kansascityfed.org/publicat/econrev/pdf/14q1Brown.pdf

Brulle

R. J.

Pellow

D. N.

(2006). Environmental justice: Human health and environmental inequalities. Annual Review of Public Health, 27, 103-124. https://doi.org/10.1146/annurev.publhealth.27.021405.102124

10.

Bunker

S. G.

(2003). Matter, space, energy, and political economy: The Amazon in the world-system. Journal of World-Systems Research, 9(2), 219-258. https://doi.org/10.5195/jwsr.2003.241

11.

Bureau of Economic Analysis. (2019). Local area personal income and employment methodology. https://www.bea.gov/resources/methodologies/local-area-personal-income-employment

12.

Cameron

A. C.

Miller

D. L.

(2015). A practitioner’s guide to cluster-robust inference. Journal of Human Resources, 50(2), 317-372. https://doi.org/10.3368/jhr.50.2.317

13.

Chi

Zhu

(2019). Spatial regression models for the social sciences. Sage.

14.

Deller

(2010). Rural poverty, tourism and spatial heterogeneity. Annals of Tourism Research, 37(1), 180-205. https://doi.org/10.1016/j.annals.2009.09.001

15.

Deller

S. C.

Lledo

Marcouiller

D. W.

(2008). Modeling regional economic growth with a focus on amenities. Review of Urban & Regional Development Studies, 20(1), 1-21. https://doi.org/10.1111/j.1467-940X.2008.00139.x

16.

Deller

S. C.

Schreiber

(2012). Mining and community economic growth. Review of Regional Studies, 42(2), 121-141. https://www.researchgate.net/publication/285787676_Mining_and_Community_Economic_Growth

17.

Deller

S. C.

Tsai

T.-H.

Marcouiller

D. W.

English

D. B.

(2001). The role of amenities and quality of life in rural economic growth. American Journal of Agricultural Economics, 83(2), 352-365. https://doi.org/10.1111/0002-9092.00161

18.

Douglas

Walker

(2017). Coal mining and the resource curse in the eastern United States. Journal of Regional Science, 57(4), 568-590. https://doi.org/10.1111/jors.12310

19.

Duncan

C. M.

(2014). Worlds apart: Poverty and politics in rural America (2nd ed.). Yale University Press.

20.

Economic Research Service. (2020). Documentation of rural urban continuum codes (Technical report). U. S. Department of Agriculture.

21.

English

D. B.

Marcouiller

D. W.

Cordell

H. K.

(2000). Tourism dependence in rural America: Estimates and effects. Society & Natural Resources, 13(3), 185-202. https://doi.org/10.1080/089419200279054

22.

Franke

A. G.

(2007). The development of underdevelopment. In Roberts

J. T.

Hite

A. B.

(Ed.), The globalization and development reader: Perspectives on development and global change (pp. 76-85). Basil Blackwell. (Original work published 1969)

23.

Frankel

J. A.

(2012). The natural resource curse: A survey of diagnoses and some prescriptions (HKS Faculty Research Working Paper Series RWP12-014). https://dash.harvard.edu/bitstream/handle/1/8694932/RWP12-014_Frankel.pdf

24.

Freudenburg

W. R.

(1992). Addictive economies: Extractive industries and vulnerable localities in a changing world economy. Rural Sociology, 57(3), 305-332. https://doi.org/10.1111/j.1549-0831.1992.tb00467.x

25.

Freudenburg

W. R.

Gramling

(1994). Natural resources and rural poverty: A closer look. Society & Natural Resources, 7(1), 5-22. https://doi.org/10.1080/08941929409380841

26.

Freudenburg

W. R.

Wilson

L. J.

(2002). Mining the data: Analyzing the economic implications of mining for nonmetropolitan regions. Sociological Inquiry, 72(4), 549-575. https://doi.org/10.1111/1475-682X.00034

27.

Gibbons

Overman

H. G.

(2012). Mostly pointless spatial econometrics? Journal of Regional Science, 52(2), 172-191. https://doi.org/10.1111/j.1467-9787.2012.00760.x

28.

Golgher

A. B.

Voss

P. R.

(2016). How to interpret the coefficients of spatial models: Spillovers, direct and indirect effects. Spatial Demography, 4(3), 175-205. https://doi.org/10.1007/s40980-015-0016-y

29.

Green

(2017). The opportunities and limits of economic growth. In Tickamyer

A. R.

Sherman

Warlick

(Eds.), Rural poverty in the United States (pp. 416-438). Columbia University Press. https://doi.org/10.7312/tick17222-026

30.

Gylfason

(2001). Natural resources, education, and economic development. European Economic Review, 45(4-6), 847-859. https://doi.org/10.1016/S0014-2921(01)00127-1

31.

Harvey

(2006). Spaces of global capitalism. Verso.

32.

Harvey

(2018). The limits to capital. Verso books. (Original work published 1982)

33.

Harvey

M. H.

(2017). Racial inequalities and poverty in rural America. In Tickamyer

A. R.

Sherman

Warlick

(Eds.), Rural poverty in the United States (pp. 141-167). Columbia University Press. https://doi.org/10.7312/tick17222-009

34.

Havranek

Horvath

Zeynalov

(2016). Natural resources and economic growth: A meta-analysis. World Development, 88(December), 134-151. https://doi.org/10.1016/j.worlddev.2016.07.016

35.

Hooks

Mosher

Rotolo

Lobao

(2004). The prison industry: Carceral expansion and employment in US counties, 1969–1994. Social Science Quarterly, 85(1), 37-57. https://doi.org/10.1111/j.0038-4941.2004.08501004.x

36.

Humphrey

C. R.

Berardi

Carroll

M. S.

Fairfax

Fortmann

Geisler

Johnson

T. G.

Kusel

Lee

R. G.

Macinko

Peluso

N. L.

Schulman

M. D.

(1993). Theories in the study of natural resource-dependent communities and persistent rural poverty in the United States. In Summers

(Ed.), Persistent poverty in rural America (pp. 136-172). Westview Press.

37.

Hunter

L. M.

Boardman

J. D.

Onge

J. M. S.

(2005). The association between natural amenities, rural population growth, and long-term residents’ economic well-being. Rural Sociology, 70(4), 452-469. https://doi.org/10.1526/003601105775012714

38.

Isserman

A. M.

Westervelt

(2006). 1.5 Million missing numbers: Overcoming employment suppression in county business patterns data. International Regional Science Review, 29(3), 311-335. https://doi.org/10.1177/0160017606290359

39.

James

Aadland

(2011). The curse of natural resources: An empirical investigation of us counties. Resource and Energy Economics, 33(2), 440-453. https://doi.org/10.1016/j.reseneeco.2010.05.006

40.

Johnson

K. M.

Lichter

D. T.

(2019). Rural depopulation: Growth and decline processes over the past century. Rural Sociology, 84(1), 3-27. https://doi.org/10.1111/ruso.12266

41.

Kemeny

Storper

(2015). Is specialization good for regional economic development? Regional Studies, 49(6), 1003-1018. https://doi.org/10.1080/00343404.2014.899691

42.

Kim

K.-k.

Marcouiller

D. W.

Deller

S. C.

(2005). Natural amenities and rural development: Understanding spatial and distributional attributes. Growth and Change, 36(2), 273-297. https://doi.org/10.1111/j.1468-2257.2005.00277.x

43.

Krannich

R. S.

Gentry

Luloff

Robertson

(2014). Resource dependency in rural America: Continuities and change. In Bailey

Jensen

Ransom

(Eds.), Rural America in a globalizing world (pp. 208-225). West Virginia University Press.

44.

Leicht

K. T.

Jenkins

J. C.

(2007). New and unexplored opportunities: Developing a spatial perspective for political sociology. In Lobao

L. M.

Hooks

Tickamyer

A. R.

(Eds.), The sociology of spatial inequality (pp. 63-84). State University of New York Press.

45.

Lichter

D. T.

(2012). Immigration and the new racial diversity in rural America. Rural Sociology, 77(1), 3-35. https://doi.org/10.1111/j.1549-0831.2012.00070.x

46.

Lobao

Kraybill

D. S.

(2005). The emerging roles of county governments in metropolitan and nonmetropolitan areas: Findings from a national survey. Economic Development Quarterly, 19(3), 245-259. https://doi.org/10.1177/0891242405276514

47.

Lobao

Zhou

Partridge

Betz

(2016). Poverty, place, and coal employment across Appalachia and the United States in a new economic era. Rural Sociology, 81(3), 343-386. https://doi.org/10.1111/ruso.12098

48.

Lobao

L. M.

Hooks

Tickamyer

A. R.

(2007). The sociology of spatial inequality. SUNY Press.

49.

Manson

Schroeder

Van Riper

Ruggles

(2017). IPUMS: National Historical Geographic Information System: Version 12.0 [Data set]. University of Minnesota. https://doi.org/10.18128/D050.V12.0

50.

Matsuyama

(1992). Agricultural productivity, comparative advantage, and economic growth. Journal of Economic Theory, 58(2), 317-334. https://doi.org/10.1016/0022-0531(92)90057-O

51.

Monnat

S. M.

Chandler

R. F.

(2017). Immigration trends and immigrant poverty in rural America. In Tickamyer

A. R.

Sherman

Warlick

(Eds.), Rural poverty in the United States (pp. 167-201). Columbia University Press.

52.

Mueller

J. T.

(2020). The effect of natural resource dependence on economic prosperity [Doctoral dissertation, The Pennsylvania State University]. https://etda.libraries.psu.edu/catalog/17542jfm5607

53.

Nelson

P. B.

Lee

A. W.

Nelson

(2009). Linking baby boomer and Hispanic migration streams into rural America: A multi-scaled approach. Population, Space and Place, 15(3), 277-293. https://doi.org/10.1002/psp.520

54.

Papyrakis

(2017). The resource curse-what have we learned from two decades of intensive research: Introduction to the special issue. Journal of Development Studies, 53(2), 175-185. https://doi.org/10.1080/00220388.2016.1160070

55.

Papyrakis

Gerlagh

(2004). The resource curse hypothesis and its transmission channels. Journal of Comparative Economics, 32(1), 181-193. https://doi.org/10.1016/j.jce.2003.11.002

56.

Papyrakis

Gerlagh

(2007). Resource abundance and economic growth in the United States. European Economic Review, 51(4), 1011-1039. https://doi.org/10.1016/j.euroecorev.2006.04.001

57.

Peluso

N. L.

Humphrey

C. R.

Fortmann

L. P.

(1994). The rock, the beach, and the tidal pool: People and poverty in natural resource-dependent areas. Society & Natural Resources, 7(1), 23-38. https://doi.org/10.1080/08941929409380842

58.

Perdue

R. T.

Pavela

(2012). Addictive economies and coal dependency: Methods of extraction and socioeconomic outcomes in West Virginia, 1997-2009. Organization & Environment, 25(4), 368-384. https://doi.org/10.1177/1086026612464767

59.

Poudyal

N. C.

Hodges

D. G.

Cordell

H. K.

(2008). The role of natural resource amenities in attracting retirees: Implications for economic growth policy. Ecological Economics, 68(1-2), 240-248. https://doi.org/10.1016/j.ecolecon.2008.02.022

60.

Sachs

J. D.

Warner

A. M.

(1995). Natural resource abundance and economic growth (NBER Working Paper 5398). https://doi.org/10.3386/w5398

61.

Schisterman

E. F.

Cole

S. R.

Platt

R. W.

(2009). Overadjustment bias and unnecessary adjustment in epidemiologic studies. Epidemiology, 20(4), 488-495. https://doi.org/10.1097/EDE.0b013e3181a819a1

62.

Sherman

(2009). Those who work, those who don’t: Poverty, morality, and family in rural America. University of Minnesota Press.

63.

Sherman

(2018). Not allowed to inherit my kingdom: Amenity development and social inequality in the rural West. Rural Sociology, 83(1), 174-207. https://doi.org/10.1111/ruso.12168

64.

Sklair

(2002). Democracy and the transnational capitalist class. Annals of the American Academy of Political and Social Science, 581(1), 144-157. https://doi.org/10.1177/000271620258100113

65.

Smith

M. D.

Krannich

R. S.

(2000). “Culture clash” revisited: Newcomer and longer-term residents’ attitudes toward land use, development, and environmental issues in rural communities in the Rocky Mountain West. Rural Sociology, 65(3), 396-421. https://doi.org/10.1111/j.1549-0831.2000.tb00036.x

66.

Smith

(1984). Uneven development: Nature, capital, and the production of space. Basil Blackwell.

67.

Smith

(2011). Uneven development redux. New Political Economy, 16(2), 261-265. https://doi.org/10.1080/13563467.2011.542804

68.

Stedman

R. C.

Parkins

J. R.

Beckley

T. M.

(2005). Forest dependence and community well-being in rural Canada: Variation by forest sector and region. Canadian Journal of Forest Research, 35(1), 215-220. https://doi.org/10.1139/x04-140

69.

Thiede

Kim

Valasik

(2018). The spatial concentration of America’s rural poor population: A postrecession update. Rural Sociology, 83(1), 109-144. https://doi.org/10.1111/ruso.12166

70.

Thiede

Slack

(2017). The old versus the new economies and their impacts. In Tickamyer

A. R.

Sherman

Warlick

(Eds.), Rural poverty in the United States (pp. 231-256). Columbia University Press. https://doi.org/10.7312/tick17222-015

71.

Thiede

B. C.

Brown

D. L.

Sanders

S. R.

Glasgow

Kulcsar

L. J.

(2017). A demographic deficit? Local population aging and access to services in rural America, 1990–2010. Rural Sociology, 82(1), 44-74. https://doi.org/10.1111/ruso.12117

72.

Tsvetkova

Partridge

(2017). The shale revolution and entrepreneurship: An assessment of the relationship between energy sector expansion and small business entrepreneurship in US counties. Energy, 141(December), 423-434. https://doi.org/10.1016/j.energy.2017.09.101

73.

Ulrich-Schad

J. D.

(2018). “We didn’t move here to move to Aspen”: Community making and community development in an emerging rural amenity destination. Journal of Rural and Community Development, 13(4), 43-46.

74.

Ulrich-Schad

J. D.

Qin

(2018). Culture clash? Predictors of views on amenity-led development and community involvement in rural recreation counties. Rural Sociology, 83(1), 81-108. https://doi.org/10.1111/ruso.12165

75.

U.S. Census Bureau. (2014). American Community Survey design and methodology. https://www.census.gov/history/pdf/acsdesign-methodology2014.pdf

76.

Vega

S. H.

Elhorst

J. P.

(2015). The SLX model. Journal of Regional Science, 55(3), 339-363. https://doi.org/10.1111/jors.12188

77.

Von Hippel

P. T.

Scarpino

S. V.

Holas

(2016). Robust estimation of inequality from binned incomes. Sociological Methodology, 46(1), 212-251. https://doi.org/10.1177/0081175015599807

78.

Wallerstein

(2015). The rise and future demise of the world capitalist system: Concepts for comparative analysis. In Roberts

J. T.

Hite

A. B.

Chorev

(Eds.), The globalization and development reader: Perspectives on development and global change (2nd ed., pp. 126-146). John Wiley. (Original work published 1979)

79.

West

P. C.

(1994). Natural resources and the persistence of rural poverty in America: A Weberian perspective on the role of power, domination, and natural resource bureaucracy. Society & Natural Resources, 7(5), 415-427. https://doi.org/10.1080/08941929409380878

Supplementary Material

Please find the following supplemental material available below.

For Open Access articles published under a Creative Commons License, all supplemental material carries the same license as the article it is associated with.

For non-Open Access articles published, all supplemental material carries a non-exclusive license, and permission requests for re-use of supplemental material or any part of supplemental material shall be sent directly to the copyright owner as specified in the copyright notice associated with the article.

0.00 MB

1.33 MB