Progress Report of NASS, USA

Stan Hoge & Bill Iwig

NASS, USA

1. Status and Update on the NASS Farm Register

Primary activities for calendar year 2004 included the evaluation of the National Agricultural Statistics Services vision and goals related to the Census of Agriculture and Survey programs and a refocus on list building and maintenance.

Results from the 2002 Census of Agriculture showed a decrease in farm number coverage continuing the downward trend of last four Censuses. Over time the number of small farms where coverage is the poorest has been has increasing. The number of medium to large farming operations where coverage is the best has decreased due the economies of scale that have required commercial production operations to consolidate or go out of business. The net effect is a decrease in the overall coverage of the total number of farms (see Table 1 for details). This is further emphasized by the fact that Land in Farm Coverage has remained consistent at around 97.6 percent.

Table 1: NASS Coverage History

1/ Based on the number of farm estimates from the last four Censuses of Agriculture and assuming the 2002 percent Not on Mail List.

The challenge for NASS in the 2007 Census of Agriculture is to provide a high quality Farm Register that is high in coverage, low in the number of non-farm records and possesses accurate control data, all relative to costs. The relative costs in time, money and resources associated with building, maintaining and collecting data for small farms are high and not efficient. Yet small farms are important to the census for demographic purposes. Potential small farms typically are not commercial production operations but tend to be widespread and numerous. The analysis of records not on the census mail list indicated that missing farms were primarily these small operations with cattle, equine or other hay. Many of these agricultural places only qualify as a farm periodically and are frequently alternating between being in scope or out of scope based on the Census definition of having the equivalent of a $1,000 in value of sales. This ultimately means that the NASS Farm Register identifies a large volume of these records that have a low potential of meeting the farm definition and does not meet the challenge of keeping costs down. NASS plans to balance the cost of maintaining low potential farms by limiting the volume of these low potential farms on the 2007 Census Mail List and using coverage adjustments for missing operations on the Census.

To meet the challenge for the 2007 Census of Agriculture NASS has taken several steps including the establishment of the Farm Register Council to monitor policies and procedures, using the Five Year Farm Register Plan (mentioned at the 16th and 17th Roundtables) as a guide, continuing to focus on major list building efforts and aggressively identifying new list sources both at the national and state levels. Farm Register plans have been modified to work with coverage adjustment plans for the 2007 Census. Changes will include list-building efforts that focus on the large production operations, improving the coverage of the demographically unique farms, especially minority owned operations and farms with specialty crops. More general list building efforts will continue to be important in providing an adequate volume of potential farms for adequate coverage while being cost effective. Other major activities for 2004 included populating the Farm register with the control data from the 2002 Census of Agriculture and the establishment of the Farm Register Council. Work on the specifications to transfer the 2007 Census information to the Farm Register began in fall of 2003 and was ready to be populated before the annual population classification and sample selection in the spring of 2004. Final counts included the updating of 14 million data values for 356 control variables and 1.49 million Farm Register rows. This information was limited to actual reported information and does not include any imputation. Spring of 2004 also marked a return to scheduled list building and Farm Register maintenance. These efforts went very well but were still modified to increase the number of medium and low potential farms since the match rates in record linkage were higher than expected which was not that surprising considering the recent Census list building effort. New records from this process totaled approximately 113,000 at the time of this report.

Table 2: NASS 2004 Farm Register Summary Counts.

2. Issues and Problems Resolved During the Past Year

Resolved issues in 2004 are measured by progress made in implementing the Five Year Farm Register Plan:

1).    Establishment of a Farm Register Council. The Farm Register Council met for the first time in May of 2004 to provide guidance and monitor cost and benefits of the various Farm Register Programs. Results of the first meeting were a set of recommendations on Farm Register program and policies. Based on these recommendations work groups have been established to provide analysis and results on each issue for review by the FRC. Immediate action was taken to establish the initial National Criteria Survey and a program change modifying the instructions maintaining large cross state operations.

2).    Implementation of a National Agricultural Identification Survey. A National Agricultural Identification Survey (AIS) is to be conducted in late 2004 to identify the farm status of newly added potential farm records. The AIS will be mailed from and processed through the National Processing Center (NPC) in Jeffersonville, Indiana. It will be used to test the procedures and prepare for a large volume of potential farm records in 2006-2007.

3).    Standardization of Active Status codes. Active Status codes on the NASS Farm Register were standardized in 2004 primarily to enhance the interaction with the NASS record linkage system. This process included the phasing out of outdated, redundant and vague codes. Other changes included the conversion of State use active status codes to National standard and updating definitions.

4).    Implementation of a purge schedule for all States. Prior to taking over the responsibilities for the Census of Agriculture NASS regularly purged records from the Farm Register as part of list maintenance. This was suspended during the period right after the 1997 Census of Agriculture until the completion of the 2002 Census of Agriculture, which was the first Census of Agriculture, conducted completely by NASS. This allowed NASS the time to access the quality of the combined Farm Register after the 1997 Census Mail List was merged on to the Farm Register in 1998. A scheduled purge will return after the 2005 spring classify and sample select cycle. Benefits include a reduction in the maintenance of inactive records, increasing the effectiveness of record linkage since there fewer possible matches to review and it cleans up the Farm Register in preparation for building the next Census Mail List.

5).    Extending the use of Administrative Data. In the spring of 2004 NASS and the Farm Services Agency (FSA) signed a new Memorandum of Understanding. FSA is responsible administration of farm commodity programs; farm ownership, operating and emergency loans; conservation and environmental programs; emergency and disaster assistance; domestic and international food assistance and international export credit programs. The new MOU allows NASS access FSA information from the various FSA administrated programs. This agreement has already produced a new individual query and general extract tools that accesses the FSA Compliance Acres information. Compliance acres are reported to FSA as part of the enrollment process when a farmer signs-up for various FSA programs. This data contains value information on farm operation including names and addresses, cropland acres by commodity and other administrative information.

3.    Issues and Problems Unresolved during the Past Year.

Items that need to be addressed in the next year are as follows.

1).    Integrate the Census and Survey programs. Current census procedures for the 2002 Census allowed for minimal review of name and address changes and addition of new records to the census database in order to facilitate the timely processing of over 2 million records. These changes required a through review by the State Statistical Offices prior to updating the Farm Register. As expected, the process for reconciling the differences between databases required extra resources and was a challenge to update the Farm register correctly. Census and Farm register business rules need to be reviewed and procedures need to be in place by the next census to better facilitate the transfer of Census information to the Farm Register.

2).    Find other ways to use administrative data. NASS has been actively exploring different sources and ways to use administrative data. One example in development is the use of an additional Sampling Frame based on the FSA Certified Land Units to compliment the NASS Farm Register and Area Frame. One of the expected benefits of this approach would be to reduce respondent burden by reducing sample sizes, as more FSA information is available.

3).    Standards for State Criteria Plans. State Statistical Offices have traditionally been solely responsible for conducting surveys to identify the farm status of new potential farms records with limited or no involvement from Headquarters. With the prospect of regularly scheduled National Agricultural Identification Surveys the roles of the SSO, NPC and Headquarters need to be reevaluated and defined. Specifically the SSO will need clear direction on the priorities which potential farm records that will be their responsibility and how data collection and surveys plans are to be coordinated into the SSO schedule.

4. Future Plans

1).    Improve the accuracy of inactive codes on the List Frame. Current Survey procedures at NASS allow for the inactivation landlords, retired farmers and others who are not actively involved in agricultural production but who may still have the resources to be involved in some sort of reduced agricultural activity that would qualify as a farm for the Census. These agricultural places are important to the Census for demographic purposes and farm number coverage. Active status codes need to be reviewed so that out of business on the Survey program is not the same as out of agriculture.

2).    Actively reach out to diverse minority and specialty groups involved in agriculture. NASS has been working with various minority and specialty farm groups and other agencies to identify ways to encourage participation and promote the Census of Agriculture within the various communities. Several projects are in development. One test being conducted is in Louisiana where NASS is enlisting the help of Black Churches in a list building initiative. Efforts are also continuing to actively network with Native American and other minority groups.

3).    Monitor the progress of the National Animal Identification Program. NASS is interested in the recent developments for a National Animal Identification Program. In December of 2003, USDA announced that it would expedite implementation of the program, in the wake of the discovery of a cow with BSE (mad cow disease) in Washington State. The Animal and Plant Health Inspection Service (APHIS) has been funded to develop a national ID system that will trace back diseased or possibly diseased animals. Early pilot programs started this summer are testing identification systems. NASS would use the information for list building and administrative purposes.

4).    Develop a record tracking system. A record tracking system that would keep track of list source information and available criteria information of potential farms would be used to identify records with a higher likelihood of being in scope on the Census. This information would be used in analysis to increase the efficiencies of the census mail list through modeling and analysis.