2007 Agricultural Census, and note of caution

When working at USDA I did a lot of research using county-level and micro files of the Agricultural Census. I'm anxious to dig into the latest, but I probably won't have time until summer.

The statistics from the Agricultural Census that I see reported most are (1) the number of farms and (2) average farm size. From Farmpolicy.com, here's the announcement:

A news release issued yesterday by USDA provided a broad overview of the report, which is conducted every five years, and stated that, “The number of farms in the United States has grown 4 percent and the operators of those farms have become more diverse in the past five years, according to results of the 2007 Census of Agriculture released today by the U.S. Department of Agriculture’s National Agricultural Statistics Service (NASS).”
The release added that, “The 2007 Census counted 2,204,792 farms in the United States, a net increase of 75,810 farms. Nearly 300,000 new farms have begun operation since the last census in 2002. Compared to all farms nationwide, these new farms tend to have more diversified production, fewer acres, lower sales and younger operators who also work off-farm.”

These numbers, while factually correct, could be extremely misleading.

You see, both depend importantly on the definition of a farm.

A farm is a farm, you say? No. If you define a farm as anyone with an herb garden or a tomato plant, then the average farm is about like the average household. The actual definition of a farm is not quite that generous. But it's close, and getting more generous with time: it's a place that has sales or potential sales of $1000. This definition has been constant, in nominal dollars, for 35 years.


What's not mentioned explicitly is that USDA has also been working harder and harder to find and count hiding $1000 potential farms. Most of these guys don't know they're farms and so they can be hard to find and difficult to entice to return their census forms. So non-response rates are growing, mostly for tiny farms that probably don't realize they're farms in the first place.

Non-response? No problem. USDA just uses weights to account for the non-response which boosts the officially reported number of farms.

An interesting tidbit: the 2002 census officially counted more farms than the 1997 census, but if you look at official changes in farm numbers, they went down. How could that be? Well, they changed the weights to count more of the micro farms they thought were probably hiding out there and that caused the numbers to go up in 2002. When they count changes they keep weights constant.

All of this may seem trivial, and on some level it is. But farm numbers do matter for allocation of funding from USDA-CSREES across states. It also matters for some research findings. For example, there have been wide reports of farm households deriving an increasing share of their income from off farm. But mostly reverses the truth. Rather, the census is counting more households as farms, and since these "new farms" derive nearly all their income off farm (indeed, probably don't even think of themselves as farms), the aggregate share of off-farm income has been rising.

It's the definition that's been changing much more than the actual disposition of farms.

Before USDA took over and the Census Bureau managed the Ag. Census, they wanted to restrict the definition of a farm to a place with $10,000 in sales or potential sales. This was in the mid-1990s. Since this seemingly innocuous change would have halved the number of farms and had a significant influence on the distribution of federal money to the states, it wasn't going to happen. I'm unsure about the details but I think this is a key reason the Ag. Census moved from the Census Bureau to USDA.

Anyhow, this little article explains how much all of this matters.

The point is that despite the "growing number of farms," agricultural production is actually becoming far more concentrated in fewer much larger farms. So it's important not to miss this part of the announcement:

“The latest census figures show a continuation in the trend towards more small and very large farms and fewer mid-sized operations. Between 2002 and 2007, the number of farms with sales of less than $2,500 increased by 74,000. The number of farms with sales of more than $500,000 grew by 46,000 during the same period,” USDA said.