Skip to main content
. Author manuscript; available in PMC: 2023 Mar 1.
Published in final edited form as: J Health Econ. 2022 Jan 30;82:102595. doi: 10.1016/j.jhealeco.2022.102595

Table 2:

Water Quality Impacts of SGD Chemicals

A. Water Quality Impacts of Drilled Well Bores
Sample SGD Chemicals Non-SGD
Dep. Var. Log Result Detection Std. Result Log Result
Bores, <0.5km 0.0133**
(0.00675)		 0.0587***
(0.0166)		 0.0840
(0.0716)		 0.00245
(0.00417)
Bores, 0.5–1km 0.00907**
(0.00413)		 0.0263***
(0.00965)		 0.0912**
(0.0365)		 −0.00280
(0.00230)
Bores, 1–2km −0.000347
(0.000772)		 2.26e-05
(0.00209)		 0.0273**
(0.0118)		 0.000364
(0.000468)
Bores, 2–3km −0.000300
(0.000618)		 0.00146
(0.00197)		 −0.0135*
(0.00740)		 −0.000255
(0.000449)
Bores, 3–4km 0.000148
(0.000528)		 0.000737
(0.00164)		 0.0103*
(0.00618)		 0.000324
(0.000337)
Bores, 4–5km −0.000247
(0.000427)		 −0.00250*
(0.00149)		 0.00408
(0.00585)		 −0.000857
(0.000537)
Obs. 69,237 69,237 65,573 102,370
B. Heterogeneity & Placebo Tests (SGD Chemicals)
Dep. Var.: Log Result Threat Type: Uphill (A) vs. Downhill Produced (A) vs. Never Produced Future Drilling Permitted Wells
Type A Bores, <1km 0.0101**
(0.00448)		 0.0131*
(0.00730)
Type B Bores, <1km 0.00497
(0.00321)		 0.00555
(0.00402)
Bores, Next 180 days, <1km 0.00587
(0.00393)
Permitted Bores, <1km 0.00642
(0.00499)
Obs. 69,237 69,237 69,237 69,237

Note. Table presents water quality impacts of SGD. Each column is a separate regression. The estimation sample consists of either SGD or non-SGD related water measurements from ground water-based community water systems with any water source within 10 km of any gas well. In panel A, the main regressors of interest are drilled well bores (at various distances to the water source). Across columns, we vary the sample (SGD or non-SGD related) and the dependent variable. Panel B re-estimates the specification in column 1 of panel A, except (1) it does not separate out well bores drilled within 0.5 and 1 kilometer, and (2) either distinguishes well bores within 1 kilometer as being Type A (uphill or producing) or Type B (downhill or never producing), or (3) adds additional variables for which we would expect no impact (future drilling or permitted well bores). Impacts at 1 kilometer for “Future Drilling” and “Permitted Wells” in panel B are not shown. Robust standard errors in parentheses are clustered at the CWS level.

***

p<0.01,

**

p<0.05,

*

p<0.1.