predictions based on changing risk parameters are consistently refuted. In this section we
consider the results in this chapter from the perspective of our transaction cost framework
that assumed risk-neutral contracting parties.
Risk-Neutral Approaches to Share Contracts
In the traditional principal-agent model, risk aversion is a necessary condition for share con-
tracting because the only behavioral margin is the farmer's (unobservable) effort, providing
the well-known trade-off between farmer shirking and risk avoidance. As we have seen, in
chapters 4 and 5, risk aversion need not be required to explain share contracting as long
as other behavioral margins besides farmer effort are considered. With risk neutral parties,
and contrary to prediction 6.2, share contracts can still be optimal when there are additional
incentive problems, such as double-sided moral hazard, multitask agency, or measurement
costs. As argued earlier, for farming there is good reason to believe that the single margin
moral hazard model is an inappropriate model of farmer and landowner incentives. Land,
like farmer effort, is also a variable input that often allows for landowner moral hazard.
Landowners, for instance, may not properly maintain fences or irrigation equipment. At
the same time, farmers can damage the land because lease contracts for land do not, and
cannot, specify all the characteristics of the land. In addition, there are costs of measuring
and dividing shared output and input costs. In part III we consider the organizational issues
of asset ownership and vertical integration, and we see that the number of margins under
which transaction costs of one type or another can arise becomes very large indeed. Ex-
tending the transaction cost framework in these circumstances is straightforward, whereas
extending a model incorporating risk sharing verges on intractable.
In our risk-neutral transaction cost model of share contracting, the trade-offs are distinct
from the risk-sharing versus farmer-shirking trade-off. First, share contracts distribute the
deadweight losses from moral hazard over many margins. Second, share contracts create
incentives for overuse and underreporting of those assets (inputs and the output) that are
shared. As shown in earlier chapters, share contracts are chosen when the costs of dividing
and measuring shared assets are low and the margins for moral hazard are large and many.
On the other hand, fixed payment contracts may emerge when measurement costs are high
and moral hazard margins are small and few.
The Risk-Sharing Evidence Revisited
The most important principal-agent prediction (prediction 6.5)—as output variability in-
creases, sharecropping should be more common—was not supported by the data from
Louisiana, Nebraska, and South Dakota in tables 6.5 and 6.6. Our risk-neutral approach,