Shale-boom counties grew jobs and income far faster than the nation — but the boom's benefits are basin-specific and fading fastest in gas country

Summary

Across 12 leading shale-boom counties — four each in the Bakken (ND), Eagle Ford (TX), Permian (TX), and Marcellus (PA/WV) plays — employment grew an average of +87.6% and real per-capita income grew +66.5% from just before the boom to the most recent data (2024), versus +14.3% employment growth and +12.4% real wage growth nationally over the same span. But the gains are highly uneven: the Bakken counties in North Dakota show extraordinary, population-transforming growth (McKenzie County employment +305%, population +156%), while the Marcellus gas counties in Pennsylvania and West Virginia show only modest gains and in one case (Susquehanna County, PA) essentially flat employment. This matches published research: shale employment and income gains are real but temporary in most basins, and the Appalachian gas counties in particular have seen initial gains erode or reverse.

Which counties and why

These are widely-cited top production/employment-growth counties in each of the four dominant U.S. shale plays (Hart Energy, Dallas Fed, Minneapolis Fed): Bakken — McKenzie, Mountrail, Williams (ND); Eagle Ford — Karnes, La Salle, DeWitt (TX); Permian — Midland, Martin, Reeves (TX); Marcellus — Bradford, Susquehanna (PA), Doddridge (WV). Each play's shale-drilling boom (horizontal drilling + high-volume hydraulic fracturing) began ramping roughly 2006-2011 depending on the basin.

Data-coverage limitation — read before the numbers

This corpus's county-level BLS QCEW employment/wage table (econ.county_wages) and BEA regional income table (econ.regional_income) both begin in 2007-2008 — essentially at or just before the Bakken and Marcellus booms started, not a full decade earlier. A genuine decade-before comparison for those two plays would need 1996-1998 data, which is not loaded in this warehouse and was not separately sourced from BLS's historical QCEW archive for this answer given the scope. The one measure here that IS a true decade-before comparison is Census decennial population (2000 vs. 2020). Readers should treat the employment/wage/income figures as "near the start of the boom vs. now" rather than literally "a decade before the boom vs. now," especially for Bakken and Marcellus. For Eagle Ford and Permian, whose shale booms accelerated closer to 2010-2011, the 2007/2008 baseline is a closer — though still imperfect — stand-in for the pre-boom decade.

County-by-county detail

Source: econ.county_wages (BLS QCEW, own_code='0', industry_code='10', 2007 vs 2024), census.decennial_population (2000 vs 2020), econ.regional_income (BEA CAINC1 per-capita personal income, line_code='3', 2008 vs 2024, deflated to 2024$ via CPI-U). National comparators from econ.state_wages summed nationally.

By basin: a genuinely different story in each play

Bakken (ND) — the biggest transformation of any play here. McKenzie County's population grew 156% (2000-2020) and its employment more than quadrupled (2007-2024); Williams County (Williston) went from a stagnant farm-trade county of ~20,000 to a regional oil-services hub of ~41,000. Real wages and incomes rose sharply too, though Williams County's real per-capita income is actually flat-to-slightly-down since 2008 because its 2008 baseline was already elevated by an early spike. This basin, more than the others, converted a genuine multi-decade population and employment trajectory.

Permian (TX) — the strongest and most durable income gains of the four plays: Martin and Reeves counties saw real per-capita income roughly 2.6x since 2008, and employment growth (81-104%) outpaced every basin but the Bakken. Population growth was more moderate (10-47%) — Permian growth has concentrated more in payroll/income than in raw population inflow, consistent with a basin still in a later, more capital- and machinery-intensive phase of development.

Eagle Ford (TX) — a mixed picture. Karnes County shows the strongest income gain in this basin (+134% real per-capita income) alongside falling population (-4.8%) — consistent with published findings that the Eagle Ford's economic upside showed up disproportionately in wages/royalties per resident rather than in durable population growth. DeWitt County's employment gain (+3.9%) is barely distinguishable from noise and trails the national average.

Marcellus (PA/WV) — the weakest and most consistent with the peer-reviewed finding that Appalachian shale-gas gains were "modest and temporary." Bradford and Susquehanna counties in Pennsylvania show essentially flat employment (+2.6%, -0.2%) and both have lost population since 2000 (-4.5%, -9.0%) despite hosting some of the most active Marcellus gas drilling in the country. Doddridge County, WV shows a stronger employment gain (+32.9%) and income gain (+79.4%) than its Pennsylvania peers, but even there population growth was minimal (+5.5%). This lines up with the research literature: gas-focused plays with lower per-well labor intensity and a workforce drawn heavily from outside the county produced a much smaller and less durable local economic footprint than the oil-focused Bakken and Permian.

What the published research says

This pattern matches peer-reviewed and Federal Reserve research on shale-boom counties. A Minneapolis Fed comparison found the Bakken's average weekly wage rose 88% from 2008 through the early boom years versus just 20% in the Eagle Ford, matching the outsized Bakken gains found here. National studies of the fracking boom (e.g. Bartik et al., and Marcellus-specific work by Weber and colleagues) find local employment and wage gains of roughly 7-11% in the first several years after a county's boom began, but that these gains typically fade after about 4 years, and Appalachian-region studies going further out find shale gas counties' initial 1.6% employment gain reversing into a net loss over the longer run — consistent with Bradford and Susquehanna counties' near-zero net employment change here despite roughly 15+ years of Marcellus drilling. In short: oil-heavy Bakken/Permian counties converted their booms into much larger and more durable local economies; gas-heavy Marcellus counties in Appalachia mostly have not.

Methodology and caveats

• County selection: 12 counties, chosen as widely-cited leading production/employment-growth counties in each of the four dominant U.S. shale plays, not a data-driven ranking by this corpus (no single "biggest shale boom" ranking table exists in this warehouse) — see the data-coverage note above.
• "Now" vintage: 2024 for employment/wages/income (most recent complete QCEW/BEA year); 2020 for population (most recent decennial census).
• "Before" vintage: 2007 for employment/wages, 2008 for per-capita income (earliest available in this corpus, not a literal decade before each play's boom start — see caveat above), 2000 for population (a genuine two decades of context, spanning the true pre-boom baseline).
• Real dollars: wages and income are CPI-U deflated to 2024 dollars via this corpus's adjust_inflation tool; population and employment are raw counts/levels, not inflation-relevant.
• Small-n caveat: 12 counties is a small, non-random sample chosen for prominence, not a representative sample of all shale-affected counties; basin averages (3-4 counties each) are illustrative, not statistically robust estimates.

What This Report Does Not Answer

• Compare their economies now vs. a decade before the boom started: County-level BLS QCEW and BEA data in this corpus only reach back to 2007/2008, which for the Bakken and Marcellus plays (booms ramping 2006-2008) is close to the boom's start rather than a full decade before it. Used 2007/2008 as the earliest available 'before' baseline for employment/wages/income, explicitly flagged as a limitation, and supplemented with a genuine 2000-vs-2020 population comparison, which does span the real pre-boom decade.

Every query behind this report

11 warehouse calls ran in this session, in order. Each is reproducible against the same snapshot.

SELECT county_fips, year, annual_avg_emplvl, total_annual_wages, annual_avg_wkly_wage, annual_avg_estabs
FROM econ.county_wages
WHERE county_fips IN ('38053','38061','38105','48255','48283','48123','48329','48317','48389','42015','42115','54017')
AND own_code='0' AND industry_code='10'
AND year IN ('2007','2024')
ORDER BY county_fips, year

SELECT county_fips, "year", total_population FROM census.decennial_population
WHERE county_fips IN ('38053','38061','38105','48255','48283','48123','48329','48317','48389','42015','42115','54017')
AND "year" IN ('2000','2020')
ORDER BY county_fips, "year"

SELECT geo_fips, time_period, data_value FROM econ.regional_income
WHERE geo_fips IN ('38053','38061','38105','48255','48283','48123','48329','48317','48389','42015','42115','54017')
AND cl_unit = 'Per capita personal income' AND tablename='CAINC1' AND time_period IN ('2008','2024')
ORDER BY geo_fips, time_period

SELECT DISTINCT cl_unit, tablename, line_code FROM econ.regional_income WHERE geo_fips='48329' AND time_period='2020' AND cl_unit ILIKE '%per capita%' LIMIT 20

SELECT DISTINCT geo_fips, geo_name FROM econ.regional_income WHERE geo_fips_set='COUNTY' AND geo_name ILIKE '%Midland%TX%' FETCH FIRST 10 ROWS ONLY

SELECT DISTINCT cl_unit, tablename, line_code FROM econ.regional_income WHERE geo_fips='48329' AND time_period='2020' FETCH FIRST 30 ROWS ONLY

SELECT DISTINCT cl_unit, line_code FROM econ.regional_income WHERE geo_fips='48329' AND time_period='2020' AND tablename='CAINC1' FETCH FIRST 20 ROWS ONLY

SELECT geo_fips, geo_name, time_period, data_value AS per_capita_income
FROM econ.regional_income
WHERE geo_fips IN ('38053','38061','38105','48255','48283','48123','48329','48317','48389','42015','42115','54017')
AND tablename='CAINC1' AND line_code='3' AND time_period IN ('2008','2024')
ORDER BY geo_fips, time_period

SELECT county_fips, year, annual_avg_wkly_wage AS wage FROM econ.county_wages WHERE county_fips IN ('38053','38061','38105','48255','48283','48123','48329','48317','48389','42015','42115','54017') AND own_code='0' AND industry_code='10' AND year IN ('2007','2024')

SELECT geo_fips, time_period AS "year", data_value AS pci FROM econ.regional_income WHERE geo_fips IN ('38053','38061','38105','48255','48283','48123','48329','48317','48389','42015','42115','54017') AND tablename='CAINC1' AND line_code='3' AND time_period IN ('2008','2024')

SELECT "year", SUM(annual_avg_emplvl) AS us_emp, SUM(total_annual_wages) AS us_wages FROM econ.state_wages WHERE own_code='0' AND industry_code='10' AND "year" IN ('2007','2024') GROUP BY "year"