Agriculture Reference
In-Depth Information
REGIONAL CV and REGIONAL STD measure the coefficient of variation and standard
deviation, respectively, for each crop for regions within the states. Similarly, COUNTY
CV and COUNTY STD measure exogenous risk at the county or parish (for Louisiana)
level.
We combine the Nebraska and South Dakota samples because they are contiguous states
from south to north and because the contract data for both come from the same survey
during the same year. In both states the eastern reaches are comprised of better soils, greater
precipitation, and a more predictable climate than their western counterparts. It is worth
noting that the far eastern portions of these two states border Iowa and are effectively
part of the Corn Belt, while the far western portions border Wyoming and are effectively
part of the High Plains. The general consequence is that crops in western counties tend to
have lower and more variable yields compared to eastern counties.
25
Louisiana exhibits a
similar variability, although it runs mainly from south to north instead. South Louisiana
tends to have a more stable, subtropical climate that makes crop yields higher and less
variable than those grown in the north. British Columbia, which is larger than all three
states combined, exhibits greater heterogeneity than do any of these three states. In many
cases the heterogeneity is so strong that crops are strictly limited to certain regions.
26
The variation in crop yield CVs (and STDs) across these jurisdictions is substantial.
27
For example, in Nebraska the minimum COUNTY CV for corn is 0.21, the maximum
COUNTY CV is 0.75, and the mean COUNTY CV is 0.28. For South Dakota, the same
measures are 0.10, 0.40, and 0.23, respectively. In Louisiana, the measures are 0.14, 0.76,
and 0.30, respectively. Table A.4 shows these statistics for other crops and for COUNTY
STD. This variability in natural conditions within the jurisdictions we study allows us to
conduct tests of risk sharing under favorable conditions.
The Choice of Contract: Cropshare vs. Cash Rent.
To test prediction 6.5 with crop-
specific contract data we use the following empirical specification, where for any contract
i
and crop
j
the complete model is
C
ij
=
σ
2
ij
j
+
X
j
j
+
ij
i
=
1,
...
,
n
j
;
j
=
1,
...
, 15;
and
(6.9)
1 f
C
ij
>
0
C
ij
=
(6.10)
C
ij
≤
0 f
0,
C
ij
C
ij
where
is an unobserved contract response variable;
is the observed dichotomous
j
choice of farmland contracts for crop
, which is equal to 1 for cropshare contracts and
equal to 0 for cash rent contracts;
n
j
is the number of contracts for a crop-specific sample;
2
ij
σ
is the crop-specific variability of the random input for a given plot of land (as measured
by CV or STD);
j
is the corresponding coefficient for crop
j
;
X
j
is a row vector of control
