Do Localities Benefit from Natural Resource Extraction?

AuthorDe Silva, Dakshina G.
  1. INTRODUCTION

    With the advances in oil and gas drilling and recovery techniques that have occurred in the last decade, the State of Texas recently experienced another oil and gas boom. Whereas annual crude oil production had been in long term decline in Texas for decades prior to 2010, annual Texas crude oil production nearly tripled between 2009 and 2015, increasing from just below 400 million barrels in 2009 to 1.155 million barrels in 2015. Indeed, according to the U.S. Energy Information Administration, Texas accounted for nearly 40 percent of U.S. crude oil production in February, 2015, or about twice the share it held in February, 2009. This recent explosion in oil and gas production that occurred in Texas is, of course, attributable to the application of horizontal drilling and hydraulic fracturing technologies that have enabled extraction of oil and gas from shale deposits.

    Economic research in the 1990s consistently found evidence that resource dependent economies exhibit slower long-term growth than more diversified economies. This phenomenon came to be called the Natural Resource Curse. While various reasons have been proposed for this resource curse, both theoretical and empirical analyses conclude that natural resource driven economic booms draw resources from non-booming export activities, lead to higher prices of non-tradables, and contribute to greater regional specialization. While most of the research in this area has focused on cross-country comparisons, similar results have been found at both the state and county levels in the United States. Yet, there is little research at the sectoral level to identify the microeconomic dynamics that would be inherent in a process of increasing regional specialization. We address that question.

    In this paper, using both OLS and censored instrumental variable approaches, we investigate the localized economic effects of oil and gas production and revenues among a defined set of non-urban counties in Texas, as explained below. This paper adds to the literature in several dimensions. Our analysis extends the previous research on the question of resource endowment and employment growth by considering inter-industry effects at the county level. That is, we not only look at county overall employment growth, but investigate employment changes in terms of their industrial composition and the likely inter-industry spillovers that a resource boom might engender. In terms of the data, this would imply an increase in the relative size of the mining sector. This is important since one explanation for the resource curse is regional specialization and the re-allocation of labor toward the booming industry. These latter effects will not be evident in broader measures of employment growth, and would be obscured if county-level labor supply is inelastic.

    Previous shorter term analyses of resource booms at the county level, Weber et al (2012), Weber (2014), and Brown (2014) for example, focus on broader measures of employment to analyze the broader impact on economic growth from growth in natural gas production. This paper not only considers a cross-industry view of employment, but also examines possible effects from the rapid increase in petroleum production that occurred more or less concurrently in many counties. We also estimate effects on both median and per capita county income for comparative and interpretive purposes. Our paper further undertakes an analysis of property tax base and public school finance at the school district level, based on host county resource endowment as instrument.

    Using only the State of Texas as the region for analysis, we are able to exploit the controlled comparison presented by the uneven distribution of oil and gas resources at the county level to identify the localized impacts of oil and gas production on our variables of interest, i.e., employment, personal income, and public school finance. This provides an important control in the case of public finance that is not present in cross-state analyses. By using a single state for analysis, we have a consistent means by which to consider changes in property tax bases, rates, and public school finance. Although we are unable to observe directly whether or not the increases in tax capacity result in higher levels of local public goods provision, we consider the question of changes in levels of per-student public education expenditures as a direct measure of investment in human capital and an indirect measure of changes in levels in local public goods. (1) We make no attempt to include the environmental costs of the production activity to the localities in which the activity occurs. The jury is still out on the question of the short and long-term environmental costs and consequences of hydraulic fracturing.

    We find that, at best, direct and indirect employment effects are modest while increases in per capita county personal income can be important. However, given that we also find lesser effects on county median income, we find it likely that gains in personal income have been rather more concentrated at higher income levels. As expected, we find that the value of county property tax bases increases with increases in production levels. Although we find no evidence that school finances were affected by oil and gas revenues over the course of the analysis, school districts appear to benefit from the higher levels of oil and gas activity in the post-2005 period (shale boom) as school tax rates are lower and per pupil expenditures higher in counties with higher levels of oil and gas production. This paper is the first, to our knowledge, to conduct a controlled analysis (single state regime) to investigate the economic effects of oil and gas extraction in relatively small geographies (counties and school districts) and to consider the effects of natural resource extraction on public finances. It is our view that increased resource mobility within small geographies, as opposed to state or national level economies, should accelerate the collateral economic impacts of a sharp expansion in natural resource extraction and facilitate identification of the ingredients that lend themselves to a resource curse over the longer term, if they occur, within a relatively shorter time frame.

    After a brief discussion of the economic context of these research questions, we proceed with the empirical analysis in terms of industry effects, county personal income, and property taxes and school expenditures. We finish with a robustness analysis and discussion of conclusions.

  2. ECONOMIC CONTEXT

    The question of how an endowment of natural resources affects economic growth rates has been extensively studied in the literature. In cross-country comparisons, Sala-i-Martin (1997) finds that primary sector production is negatively correlated with growth and Sachs and Warner (1997) find a negative association between countries' growth rates and their ratios of natural resource exports to GDP. These earlier studies relied on cross-country growth comparisons -assuming convergence in growth rates among regions in the same country- to identify what has come to be called the natural resource curse. Papyrakis and Gerlagh (2007) analyzed growth rates across states in the U.S. and found a significant negative relationship at the state-level between natural resource dependence and income growth. Working at an even finer geographic scale, James and Aadland (2011) draw similar conclusions at the county-level in the United States.

    Observation of the apparent natural resource curse has of course spawned a large literature that seeks to explain it. It is commonly argued that natural-resource dependence creates market and institutional failures that induce slower economic growth (Auty, 1994, Bhattacharyya and Hodler, 2010, Gylfason, 2001, Matsuyama, 1992, and Sachs and Warner 1997). Sachs and Warner (2001) note that resource-abundant economies are often high-price economies and tend to miss out on export led growth, i.e., Dutch Disease. James (2015) concludes that resource sectors have generally tended to grow more slowly than other sectors and, therefore, industrial composition is important to take into account. Papyrakis and Gerlagh (2007) conclude that natural resource abundance decreases investment, schooling, openness and R&D expenditures while increasing corruption, which explain the lower state-level growth rates.

    To the contrary, Weber (2012) looks at natural gas booms at the county-level in three U.S. states, Colorado, Texas, and Wyoming, for the period 1998/99-2007/08 and finds that income and employment exhibit positive, but modest, gains with respect to increases in production of shale gas. He considers the impact of the gas booms only on total county employment, wage and salary effects, and effects on median income. He recognizes that the length of his study period may not capture long-term effects. However, Weber (2014) studies a decade of shale gas production in Texas, Louisiana, Arkansas and Oklahoma counties to look for symptoms within that time frame that might be suggestive of a resource curse in the longer term. Specifically, he looks for increased dependence on the mining sector, higher earnings per job, and declines in the educational attainment of the adult population. He concludes there is little evidence to suggest that gas production creates conditions conducive to a resource curse.

    Oil and gas have notoriously exhibited boom-bust cycles. In fact, natural gas prices had been on an upward trend over the decade of Weber's analysis, but collapsed mid-year 2008. Drilling, in particular, is stimulated by high oil and gas prices. The resulting shifts in the supply curves can lead to steep price declines and bring new field development to a halt with an abrupt drop in employment. Much of the local oil and gas industry employment is associated with drilling and other oil and gas field service activities...

To continue reading

Request your trial

VLEX uses login cookies to provide you with a better browsing experience. If you click on 'Accept' or continue browsing this site we consider that you accept our cookie policy. ACCEPT