Agriculture Reference
In-Depth Information
Table 7.4
Summary statistics for new farmers and landowners: Share contracts
Variables
Only a new farmer
Only a new landowner
Number of contracts (%)
264 (100)
115 (100)
Share of inputs changed (%)
47 (17.8)
18 (15.6)
Number of shared inputs changed (%)
37 (14)
8 (6.9)
Lease changed to cropshare (%)
36 (13.6)
10 (8.6)
Farmer's share dropped (%)
15 (5.7)
10 (8.7)
Farmer's share increased (%)
20 (7.5)
5 (4.3)
the null hypothesis of no ratchet effect. Prediction 7.2 implies that the presence of a new
landowner should increase the probability of a share contract; the estimated coefficient for
NEW LANDOWNER is expected to be positive. In three cases the estimated coefficients
are negative and in all four specifications the standard errors are too large to allow statistical
significance. Overall then, the estimates shown in table 7.2 fail to reject the null hypothesis
of no ratchet effects and suggest that ratchet effects are not present in the choice of land
leasing contract.
Changes in Share Contracts
The second set of empirical tests examines the determinants of particular sharing rules
within cropshare contracts in order to test predictions 7.3 and 7.4. Table 7.4 reports some
summary statistics for cases in which a new landowner or farmer is involved in the cropshare
contract. The data in table 7.4 are generally inconsistent with ratchet effects existing within
the set of all share contracts. When there is a new farmer, both input cost shares and
the number of inputs change, which refutes prediction 7.3 that states they should remain
constant. The output share to the farmer increased more often than it decreased, which is
consistent with prediction 7.3. Furthermore, when there is a new landowner input cost shares
change less often and output shares decrease more often than they increase. However, the
difference in proportions between the two samples is not statistically significant.
20
In order to further test predictions 7.3 and 7.4, we use several samples containing only
share contracts to estimate the following model:
ln(s
i
/(
1
−
s
i
))
=
Z
i
φ
i
+
i
,
(7.10)
th
where
s
i
is the farmer's share of the output for the
i
contract, Z is a row vector of
is an error term.
21
Table 7.5 presents the results from four OLS estimations of equation (7.10). As in
the logit estimates, the variables NEW FARMER and NEW LANDOWNER are used.
explanatory variables,
φ
is a column vector of unknown coefficients, and
