Do the Largest Firms Grow the Fastest? The Case of U.S. Dairies

all farms for which the owner checked farming as his/her main occupation and for which
at least 50 percent of all agricultural income (exclusive of government payments) in 1992
came from the sale of milk and dairy products. About half of all farms reporting milk
cows in the 1992 Census are included in our sample.

The sample of new entrants in 1997 and 2002 meet the same criteria for the census of
entry. New farm entrants in 1997 constitute a new cohort and are followed through the
2002 census. New farm entrants in 2002 are included as another new cohort.

Dairy farms in the 1992 Census of Agriculture are ranked based on the value of
agricultural sales excluding government payments. They are then partitioned into ten
non-overlapping cohorts based on size. They have equal numbers in each cohort. Farms
in each initial cohort are followed through the 1997 and 2002 censuses. Because some
firms exit the industry between successive censuses, the longitudinal data files for the
incumbent sample form an unbalanced panel.

By using the Census Farm Number (CFN) and Personal Operation Identification
System (POIDS) codes, these data permit us to track most individual farms through
subsequent censuses based on the legal entity for tax purposes. However, because the
administration and conduct of agricultural census was moved from the Bureau of the
Census to the USDA National Agricultural Statistics Service, the tracking was not
perfect. In the 1992 and 1997 censuses, farms were identified by the Bureau of the
Census using the CFN which used land to recognize farms over time. For these censuses,
each individual “farm” operator received one form for his/her farm, even though the farm
might have included several pieces of lands or separate farming businesses. For the 2002
census, farms were identified by the National Agricultural Statistics Service using the

<<   6   7   8   9   10   11   12   >>

More intriguing information