Agriculture Reference
In-Depth Information
33. See Eswaran and Kotwal (1985).
34. Later in chapter 6, in the context of testing for evidence of risk sharing in contract choice, we use data that focus
on farmer wealth rather than just acres owned. The results there are consistent with those here that there is some
modest evidence that capital constraints have an impact on contract choice, even though this effect is small when
compared to the effect of the other transaction cost variables. In chapter 8 we explicitly model capital constraints
in the context of asset ownership. For the decision to own or lease an asset, the effect of capital constraints is much
larger.
35. Sotomayer, Ellinger, and Barry (2000) find support for this prediction in their Illinois contract data, using net
worth to measure financial constraint. One difficulty with our data and the data used by Sotomayer, Ellinger, and
Barry is that there is no information on the presence of multiple payments in cash rent contracts that would reduce
attraction of cash rent contracts in avoiding capital constraints.
36. One bit of evidence in favor of family acting as a trust or information variable is that absentee landowners are
more likely to contract with family members.
37. Hoffman (1984), in an examination of 83 contracts from Lyonnais, found that landowners with vines “found
sharecropping preferable to renting” (316). He reports that “highly commercial crops were often raised by croppers:
not just wine in the Lyonnais, but also mulberry trees for the silk trade farther south” (319). In a discussion of
Hoffman's earlier work, Sella (1982) states: “Sharecropping was best suited to labor intensive farming involving a
variety of arable and tree crops leases involving fixed rent, on the other hand, were preferred where either livestock
or monoculture predominated” (162).
38. See Carmona and Simpson (1999) for a detailed discussion of historical share contracts for grapes in Spain.
Chapter 5: Sharing Inputs and Outputs
1. The analysis in this chapter draws on Allen and Lueck (1993a).
2. In particular, we assume that the partial derivatives of the production function have the signs
h
i
>
0,
h
ii
<
0,
and
h
ij
=
i
j
=
e
l
k
—implying that the inputs are independent.
3. As noted in chapter 4, we maximize expected profits so the optimality conditions do not contain information
about
θ
or its distribution.
4. It may seem possible that
l
s
<l
∗
obtains, but this cannot occur because it would imply such a small output share
(s)
to the farmer that the input distortions would not be minimized. See figure 5.1.
5. The terms
e
s
(h
e
−
w) >
0,
l
s
(h
l
−
r)<
0, and
k
s
(h
k
−
c)
≥
0 are the marginal distortions resulting from using
e
s
,
l
s
, and
k
s
rather than
e
∗
,
l
∗
, and
k
∗
. The positively signed partial derivatives
e
s
,
l
s
, and
k
s
indicate the effect of a
change in the output share on the farmer's input use. Equation (5.3) requires that the marginal cost of an increase in
the farmer's share of the crop—shown by the positive terms
e
s
(h
e
−
w)
and
k
s
(h
k
−
c)
—equal the marginal gain
from the increased share, as shown by the negative term
l
s
(h
l
−
r)
. Equation (5.4) has an analogous interpretation
for changes in the input shares.
6. The assumption of lump-sum input measurement costs seems reasonable for a single farmland contract. In
particular, there is no reason to think these costs would depend on
0 for
,
,
, and
.
7. This solution was first offered by Heady (1947). Heady did not, however, consider input enforcement costs, and
as a result he did not derive the second optimal sharing rule. Berry (1962) and Schickele (1941) also discuss the
effects of sharing input costs.
8. Note that if
s
or
q
l
s
<l
∗
, this condition cannot be satisfied.
9. This prediction is consistent with a pure neoclassical model. If the farmer's contribution is higher, his return
should be too.
10. Braverman and Stiglitz (1986) compare cost-sharing contracts with those “fixed input contracts” that specify
input levels in the contracts. In the data we examine, this type of contract is not observed.
11. We do not report the huge
χ
2
statistics from the cross-tabs of input and output shares.
