Agriculture Reference
In-Depth Information
more risky crops should be shared is debunked by our data. A decade ago this failure to find
support for the risk-incentive trade-off would have been somewhat surprising, but in recent
years the evidence has been accumulating and agency theory has responded as well.
Fourth, we found that farming is dominated by family production when the random
and systematic effects of nature cannot be controlled. Mother Nature not only provides an
opportunity for moral hazard but also limits the possibilities of specialization. Generally
speaking, farm production provides many opportunities for moral hazard and few for
exploiting economies of size. Thus farming is fertile ground for family production. In those
cases where nature's seasons and uncertainty can be controlled, agricultural production
tends to be organized as large-scale corporate firms as in much of the modern economy.
Our findings here offer support for the ideas of many early-twentieth-century agricultural
economists (for example, Brewster 1950, Ellickson and Brewster 1947; Holmes 1928) and
even those of some classical economists (for example, Mill 1965 [1871]).
Beyond agriculture, empirical work in transaction cost economics has been a suc-
cess story, particularly in our understanding of property rights (Alston, Libecap, and
Mueller 2000; Libecap 1989) and modern business organization (Masten 1996; Masten
and Williamson 1999). To this body of literature our study adds a detailed industry-wide
analysis across space and time. Though this study may add more depth than breadth, it
shows how a consistent analytical framework can be used to generate a myriad of testable
hypotheses.
Limits of the Transaction Cost Approach
Again, it is a question of emphasis, but there are three limitations common to this paradigm.
First, a detailed knowledge and data is required for an accurate understanding of contracts
and organization. How can one start with the assertion that contracts are designed to mitigate
transaction costs if one has no idea how these costs arise? Sitting in an office, one can
imagine shirking behavior and outright theft, but often the realistic threats and actual
behaviors are more subtle. Likewise, knowledge about an asset's attributes and their relative
importance often requires some firsthand experience or frontline knowledge. If economists
are really going to understand firm behavior, then they have to become aware of the details
of firms and production.
Second, empirical work in transaction cost economics often, but not always suffers
from data limitations. With transaction cost economics, there are no directly observable
or measurable variables to gauge transaction costs. In farming, there is no input called a
“transaction cost,” nor is there any perfect signal of such a variable. If there was, then the
contracting parties themselves would know what it was, and there would be no transaction
costs. Generally speaking, this problem has contributed to an almost “second-class” stature
for transaction cost economics within the profession. Although using proxy variables is a
