Table 5

East versus West

Eastern Half of the CountiesWestern Half of the Counties
Cropland /
(Cropland+
Pasture)
(1)		log(Revenue
from Crops/
Farmland)
(2)		log(Revenue from
Animal Products/
Farmland)
(3)		log(Irrigated
Acreage)
(4)		Cropland /
(Cropland+
Pasture)
(5)		log(Revenue
from Crops/
Farmland)
(6)		log(Revenue from
Animal Products/
Farmland)
(7)		log(Irrigated
Acreage)
(8)
Panel A: First Stage
Proportion in the shelterbelt zone
Each post year*		0.260**
(0.100)		0.137***
(0.044)
Panel B: Second Stage
Proportion protected by shelterbelts*−0.02160.05360.05760.128−0.130−2.011***0.948−12.75
1945-1950(0.0246)(0.275)(0.142)(2.037)(0.131)(0.775)(0.680)(9.207)
1954-1959−0.0921**
(0.0387)		0.442 (0.375)0.617**
(0.269)		1.261
(2.112)		−0.225
(0.163)		−2.686**
(1.181)		2.533**
(1.160)		−9.994*
(5.960)
1964-1969−0.124**
(0.0507)		0.604
(0.388)		0.988**
(0.392)		0.298
(2.355)		−0.310*
(0.179)		−1.947**
(0.948)		2.487*
(1.462)		−10.26*
(5.876)
1978-1982−0.109*
(0.0606)		1.058**
(0.536)		1.764***
(0.574)		−0.134
(1.449)		−0.288
(0.188)		−1.595*
(0.966)		3.614*
(2.168)
1987-1992−0.147**
(0.0706)		1.262**
(0.537)		1.505***
(0.529)		−0.596
(1.517)		−0.358*
(0.211)		−1.838*
(0.991)		3.217
(2.251)		0.347
(2.016)
Year and county dummiesYesYesYesYesYesYesYesYes
Outcome variables in 1910-1930aYesYesYesYesYesYesYesYes
Other control variablesbYesYesYesYesYesYesYesYes
Erosion from the Dust BowlbYesYesYesYesYesYesYesYes
Proportion above the Ogallala AquiferbYesYesYesYesYesYesYesYes
Observations1,2431,2631,2631,1101,1901,3051,3051,035
Number of counties115115115115119119119119
R-squared0.7620.9640.9570.9570.7340.9500.9450.945

  • ↵a Data from 1910 and 1920 are missing in column 1 because the area of cropland and pasture is not reported.

  • ↵b “Other control variables“ include rainfall and temperature for each year and all the other variables listed in Table 1 from 1910–1930; “Erosion from the Dust Bowl” from Hornbeck (2012) include proportions of high- and medium-eroded as in Table 1; ”Proportion above the Ogallala Aquifer” is from Hornbeck and Keskin (2014). All these variables are time invariant and all interacted with year-dummies. Standard errors clustered at county level in parentheses

  • ↵*** p < 0.01

  • ↵** p < 0.05

  • ↵* p < 0.1