Agriculture Reference
In-Depth Information
Table A.3
Summary statistics for leased farmland Louisiana and British Columbia
Louisiana (1992)
British Columbia (1992)
British Columbia (1992)
Variable
Mean
Standard deviation
Mean
Standard deviation
Mean
Standard deviation
CONTRACT
0.67
0.47
0.29
0.46
0.39
0.49
SHARE
77.4
6.8
75.07
11.1
73
11.4
ACRES
2,989
369
281.8
l00
111.3
l,204
AGE
47
13
47.6
11.4
40.9
12.5
EDUCATION
13.2
2.3
12.5
2.5
11.05
3.14
FARM INCOME
3.06
1.17
2.6
1.2
NA
NA
FULLTIME
0.93
0.26
0.75
0.44
NA
NA
FUTURES
0.78
0.41
0.18
0.39
0.12
0.33
IRRIGATED
26.2
42.01
35
46
0.54
0.49
LlVE-FARM
0.74
0.44
0.86
0.35
NA
NA
MILES
5.36
8.34
3.18
6.03
NA
NA
WIDOW
0.02
0.13
0.01
0.09
NA
NA
BUILDINGS
71,970
71,964
139,650
147,309
NA
NA
EQUIPMENT
142,907
162,513
79,334
91,408
NA
NA
LAND
173,022
442,614
153,770
326,906
NA
NA
WEALTH
243,530
298,190
288,100
305,800
NA
NA
Note:
NA = not available.
results across the six equations. In general, risk sharing implies that crops with CVs greater (less) than the WHEAT
CV should have positive (negative) coefficient estimates. The crop CVs can be ranked by referring to table A.12.
Consider South Dakota where WHEAT has the largest CV, so that the risk-sharing prediction is that the
coefficient estimates for all crop dummies should be negative. As table A.12 shows, none of the estimates are
consistent with this prediction. Only two estimates have a negative sign, and none of the estimates are significantly
different from zero. For Nebraska, WHEAT has the lowest CV (of the dryland crops), so all coefficient estimates
are predicted to be positive. Inspection shows that estimates for OATS and SORGHUM are positive and significant
but that the estimate for CORN is negative and significant and the estimate for BARLEY is insignificantly different
from zero. For British Columbia and Louisiana, the comparisons require separating crops into two categories using
table 6.1: CV greater than WHEAT and CV less than WHEAT. For British Columbia, ALFALFA and HAY are low
CV crops, so their coefficient estimates should be negative; the estimates for all other crops should be positive.
Only the estimate for BARLEY in the 1992 sample is consistent with these predictions. In the 1979 sample only
the ALFALFA estimate is consistent with risk sharing. Most of the parameter estimates are insignificantly different
from zero but many have signs that are the direct opposite of those predicted by risk sharing. Finally, for Louisiana,
CORN and RICE have CVs greater than WHEAT, so their estimated coefficients should be positive and the other
estimates should all be negative. None of the estimates support the risk-sharing thesis for the Louisiana data, and
some estimates (for example, COTTON) directly refute the prediction.
Control Variables Used in Chapter 6
Throughout chapter 6, we reported coefficient estimates only for those variables explicitly cited in the theoretical
predictions. For table 6.5, the Nebraska-South Dakota equations included ACRES, AGE, DENSITY, and FAMILY.
For the 1992 British Columbia-Louisiana equations the controls included ACRES, AGE, EDUCATION, FARM
INCOME, FULLTIME, INSTITUTION, IRRIGATED, LIVE-FARM, and MILES. The 1979 British Columbia
equation included ACRES, AGE, EDUCATION, IRRIGATED, and LAND VALUE.
