Agriculture Reference
In-Depth Information
Prediction 7.3 implies a positive sign for the estimated coefficient of NEW FARMER, since
a higher share means the contract has higher power. Only the sample of nonfamily members
provides an estimated coefficients that is positive and statistically significant; the rest are
either negative or statistically insignificant. Prediction 7.4 implies a negative sign for the
estimated coefficient of NEW LANDOWNER. Even though all four estimated coefficients
are negative, none are close to being statistically significant. As with the regression estimates
from table 7.3, there is no consistent support for the ratchet effect within share contracts.
Predictions 7.3 and 7.4 can also be tested by estimating the probability that a farmer's
cropshare changed in the recent past. Table 7.6 shows the estimated coefficients from a logit
regression in which the dependent variable is one if the farmer's share increased within the
last five years. These estimates do show some support for the presence of ratchet effects.
First, the estimated equation shows that NEW FARMER is positively related to rising
farmer shares, consistent with prediction 7.3. Second, the estimated equation also show
that NEW LANDOWNER is positively related to rising farmer shares. Although this result
is statistically insignificant, it is inconsistent with prediction 7.4. 22
Economic Significance and Other Variables
Two separate tests of ratchet effects have been conducted. The first examined whether
contracts would be switched from low-powered share contracts to high-powered cash rent
contracts, and the other looked at whether shares would be increased when a new farmer
started farming the land. In both cases we found limited evidence consistent with ratchet
effects. Furthermore, the magnitude of the estimated coefficients for the two variables NEW
LANDOWNER and NEW FARMER were dwarfed by the size of the estimated coefficients
of the other variables in the regression, suggesting that, unlike our results in chapters 3-
5, ratchet effects have little economic significance. Table 7.3 shows that several nonratchet
variables have much larger coefficients. In particular, the coefficient on the dummy variable
ROW CROP is almost eleven times larger than the coefficients for the binary ratchet
variables. From table 7.3, even in the full sample case where output shares do rise with
the presence of new farmers, the effects of NEW FARMER and NEW LANDOWNER are
not as large as those associated with the choice of crop. 23 Thus, in addition to the lack
of statistical significance, the small coefficient size also suggests that these two variables
(NEW FARMER, NEW LANDOWNER) have a small economic impact and that the ratchet
effect is unimportant in the context of this contracting example.
Although the two variables for new farmers and landowners offer the most direct test for
the presence of ratchet effects, the positive estimated coefficient on the YEARS DURA-
TION variable in table 7.3 could be interpreted as consistent with the ratchet effect. This
positive relationship shows that long-term contracting parties are more likely to choose share
Search WWH ::




Custom Search