Enhancing Data For Complex Agricultural Establishments

The increasing organizational complexity of farming establishments offers opportunities and challenges for improving the accuracy of statistical estimates and policy data sets. These complexities can affect data collection, accuracy of estimates, use of data (for example, in policy analysis), disclosure and dissemination of estimates. To address these issues, an international expert working group meeting, Enhancing Data Collection For Complex Agricultural Establishments, was convened June 26-28, 2011, at Niagra-on-the-Lake, Ontario, Canada.

This workshop was an opportunity for discussion among experts from statistical and economic policy institutes and universities engaged in the development and use of economic statistics for agriculture and rural development.  The goal was to improve statistical estimates and data bases for policy purposes.  A conceptual paper was prepared by the Planning Committee as guidance for papers, presentations and discussion.

Financial support for the meeting was provided by the Economic Research Service and the National Agricultural Statistics Service of USDA, Agriculture and Agri-Food Canada, and Statistics Canada.  Farm Foundation, NFP co-sponsored and co-organized the meeting.

In addition to representatives from the institutional-sponsor organizations, workshop participants represented: the U.S. Census Bureau, Eurostat, the Organization for Economic Cooperation and Development, Food and Agriculture Organization of the United Nations, the Institute for Ministry of Economic Affairs, Agriculture and Innovation of the Netherlands, and the Instituto Brasileirode Geografia e Estatistica of Brazil.  Participants also included farmers and individuals from universities and agribusiness.

Members of the Workshop Planning Committee were: Mary Ahearn, USDA Economic Research Service (ERS) (co-leader); Kevin Barnes of USDA’s National Agricultural Statistics Service (NASS) (co-leader);  David Culver of Agriculture and Agri-Food Canada (AAFC) (co-leader), Sheldon Jones of Farm Foundation, NFP (co-leader); Jeffrey Smith of Statistics Canada (co-leader); Koen Boone of LEI Wageningen UR, The Netherlands;  Flavio Bolliger of the Instituto Brasileirode Geografia e Estatistica; Bill Iwig, Ashley Leduc of AAFC; Jaki McCarthy of NASS; Jim MacDonald of ERS; Joe Parsons of NASS; Krijn Poppe of the Ministry of Econoimic Affairs, Agriculture and Innovation, The Netherlands; and Daniela Ravindra, Statistics Canada.

Workshop organizers and participants expect discussions to be ongoing.
As available, here are the presentations from the workshop:

A Common Understanding of the Challenges
Welcome and Introduction
Cameron Short, Acting Director General, Research and Analysis, Agriculture and Agri-Food Canada (AAFC)

Goals of the Meeting: Challenges and Solutions
Mary Ahearn, Economic Research Service (ERS), USDA

Data Requirements for Policy and Private Decision-Making
Moderator, Fay Abizadeh, AAFC

U.S. Perspective: Daniel Sumner, University of California, Davis
Abstract: Agricultural establishments are and have long been complex in many dimensions. Relevant data and analysis must develop in ways consistent with complexities for the analysis to be relevant to current and future agricultural issues. Increasingly economists are turning to specialized sources collected to help determine important parameters. The presentation reinforces the main points through a series of examples of quite different data collection efforts: animal welfare; hedonic pricing and identification of willingness to pay for product attributes; supply elasticities for corn and soybeans incorporating rotations and spatial heterogeneity. Often data analysts are driven by supply chain, environmental and other issues which means linking farm data up the marketing chain and back down to resource use.

The Netherlands Perspective: Krijn Poppe, Ministry of Economic Affairs, Agriculture and Innovation and LEI Wageningen, UR, The Netherlands
Abstract: Policy research seldom needs yearly census data; yearly income and other data from a panel is enough in The Netherlands. The presentation provides numerous suggestions for the building of data sets including: Using econometrics as a substitute for data gathering, collaborating with industry (and their datasets), the use of IT to get electronic data, the use of standard definitions (in the 90% of the cases where this is possible), and develop those standards where needed. The presentation emphasized the importance of showing distributions of data and clearly defining the farm unit and how it is integrated into the food chain.

Developing Country Perspective: Naman Keita, Food and Agriculture Organization, United Nations
Abstract: The recent food crisis and food market volatility have lead to a renewed recognition of the critical importance of the agriculture sector as source of economic growth, food security and poverty reduction and improvement of the livelihood of a large proportion of the population in many developing countries. For policies to be effective, they need to be grounded in factual evidence about the sector and make systematic and rigorous use of statistics. However, in many developing countries, the agriculture sector is very complex and evolving rapidly with the simultaneous presence and inter-linkages of several farming systems. This paper reviews the special challenges to data collection in developing countries where there is a wide range of farms, from subsistence family farms to large, modern, market-oriented and highly mechanized systems.

Canadian Perspective: Dave Freshwater, University of Kentucky, and Dave Culver, Farm Performance and Structure, AAFC  Paper   Powerpoint
Abstract: Agricultural policy has been unusual in that it has specified the ongoing existence of a desired production unit, the family farm, as a policy objective. But, despite decades of policy intervention, the majority of Canadian farms no longer meet the common definition of a family farm. Yet, for the most part, the data collected on farming seems trapped in the use of the older and simpler concept of the family farm. The paper reviews these issues, emphasizing different challenges depending on the farm size. Both large and small farms have complex organizational structures. Managers of data systems cannot simply focus on doing a better job of understanding how large farms behave, although they certainly must do this, if only because large farms account for the majority of commodity production. Small farms, while less significant for the production of commodities, play an important role in resource use and in generating political support for agriculture.

Discussant: Catherine Moreddu, Organisation for Economic Cooperation and Development (OECD)
Abstract: The discussion comments focused on four main points: 1) changing priorities in data demand; 2)cost of information; 3) increasing complexity; and 4) distributional issues. As agri-environmental and rural development policies gained in importance in recent decades, more complex information has been needed to evaluate them. These new types of information are at once more local, complex, multidisciplinary and integrated. Recent price movements, however, have prompted renewed interest for market information to analyse price formation and transmission along the food chain, and to identify the causes and consequences of price variability in agriculture. We are all convinced of the need to look at the distribution of variables, but it is not easy to convey distributional information in a printed graph. The comments also ask: Is the household still the unit of analysis for complex farms?

Current and Imminent Data Collection Challenges Faced by Statistical Agencies
Moderator, Kevin Barnes, National Agricultural Statistics Service (NASS), USDA

Canada:   Jeffrey Smith, Agriculture Division, Statistics Canada   Paper   Presentation
Abstract: The paper presents the challenges facing Statistics Canada in two parts: those challenges faced generally when collecting data in business surveys; and those faced more specifically in collecting data from agricultural operations (surveys of agricultural operations are classified as business surveys at Statistics Canada). Increasing the use of administrative data already provided by agricultural operations appears to be a direction that must be pursued to substantially reduce the amount of surveying. While reduction of survey response burden was acknowledged as a good and worthy goal, it appeared that among the large and/or more complex operations, there is an equally strong (or even stronger) desire on their part for the data collector to “get it right”, that is, to take the time with them to properly understand the operation so that the data which are collected are correct, meaningful and useful. Greater use of technology (e.g., to allow sharing of information, to allow pre-filling of information and only getting updates on changes, etc.) and innovative methods on the part of the statistical agencies would be welcomed by respondents.

Brazil: Flavio Bolliger, Instituto Brasileirode Geografia e Estatistica (IBGE)  Paper   PowerPoint
Abstract: The types of complex establishments that raise major challenges for data collection and making records compatible with the information required in Brazil are related to (a) corporations operating in more than one activity and (b) those with a large number of physical operating units. Special attention should also be paid to cases of large corporate agricultural establishments or absentee individual producers, for which the relevant information should be collected from different places in far-off urban centers and even in a different federative state. The sugar-ethanol sector is the largest example of complex establishments in agricultural statistics in Brazil and so the paper uses this supply chain as an example of the challenges. In the Agricultural Census the information from the sugar-ethanol sector had a special collection procedure. The plants have several agricultural establishments spread over several sectors, making it unfeasible to collect at the census level, causing impacts on planning the sample surveys. In the industrial surveys, this agro-industry still has characteristics that result or may result in overestimating intermediary consumption and underestimating the added value, consequently underestimating the sector’s participation in the GDP.

Eurostat’s Farm, Agro-environment, and Rural Development Statistics: Marcel Ernens, Head of Unit
Abstract: There is a recognized need for modernization of EU Agricultural Statistics, in part because of the new demands for data on rural development, agro-environmental indicators, and food safety statistics. The presentation reviewed the plans to modernize statistics relating to agriculture in the major areas of (1) primary statistics, like crop and livestock surveys, (2) derived statistics, like economic accounts, and (3) the broad category of related indicators on land use and cover, food safety, rural development, and agri-environmental indicators.

DG Agri’s Farm Accountancy Data Network: Thierry Vard, Microeconomic Analysis of EU Agricultural Holdings
Abstract: The Farm Accountancy Data Network (FADN) was established in Europe in 1965 with an objective to measure farm income and conduct a business analysis on agricultural holdings. This paper describes the characteristics of the FADN and the challenges faced by data collection in the FADN as agricultural establishments become more complex. There remains a focus on income distribution and an increasing interest in measuring competitiveness (e.g., costs of production and productivity), making the representativeness of FADN an issue since FADN is voluntary.

United States: Joe Parsons,  NASS, USDA
Abstract: In the U.S., data collection from large, complex farms are handled on an individual basis by the decentralized NASS field offices. In order to gain a better understanding of the approaches that these offices employed and found to be most successful, NASS conducted a survey of local offices to collect views about (1) the challenges and (2) their responses to those challenges. (See also Jaki McCarthy’s presentation, based on data from this survey.) This presentation identified a great deal of variation among the local offices in both regards.

Group Discussion: Commonalities and Uniqueness Across Countries: Francine Lavoie, Chief, Enterprise Portfolio Management Program, Enterprise Statistics Division, Statistics Canada

Aspects of Complexity: Current and Emerging Issues in Agribusiness
Moderator, Sheldon Jones, Farm Foundation, NFP

Session abstract: This session featured the unique insights of three agribusiness representatives: George Muehlbach, John Deere; Tom Wegner, Land O’Lakes (presentation posted); and Dewayne Goldmon, Monsanto. Each presentation emphasized the importance of government data on agriculture to agribusiness, e.g., in helping them price their products, and also recommended that government data collection agencies consider utilizing administrative data available from industry sources as a means of reducing burden on farmers. There was also agreement that burden could be reduced by government agencies sharing data amongst themselves.

George Muehlbach, Deere and Company

Tom Wegner, Land O’Lakes

Dewayne Goldmon, Monsanto

Aspects of Complexity: Producers Voices on Complexity and Government Data Collection Needs/Approaches
, Moderator, Kevin Barnes, NASS

Session abstract: This session featured the unique insights of four farmers, each with a different type of business complexity: Kevin A. Green of Greenview Farms, DeWitt, Iowa, a cash grain producer  (see presentation); Craig Yunker, of CY Farms & Batavia Turf, Elba, NY, a vegetable, mixed grains, and sod producer  (see presentation); Beth Kennett,  Liberty Hill Farm, Rochester, VT, dairy producer and agritourism ; and Stewart Skinner of Stonaleen Farms, Ontario, Canada, a hog producer with direct marketing, The farmers shared details about how their farms are organized, emphasizing the challenges that they would face in responding to standard questionnaires, as they attempt to cooperate with government data collection agencies. For example, Green’s operation makes use of multiple legal entities that separate asset ownership from production activities. Similarly, Yunker’s operation is highly complex, in part because of its multiple activities along multiple supply chains. Kennett and Skinner have smaller operations and yet their operations are complex in that they are engaged in niche direct marketing activities and nontraditional enterprises, like recreational services.

Kevin A. Green, Greenview Farms, DeWitt, Iowa

Craig Yunker, CY Farms & Batavia Turf, Elba, NY

Beth Kennett, Liberty Hill Farm, Rochester, VT

Stewart Skinner, Stonaleen Farms, Ontario, CA

Current Best Practices from Statistical Agencies
Moderator, Dave Culver, AAFC

NASS/USDA’s Practices Across State Offices: Jaki McCarthy, NASS, USDA
Abstract: In the United States, data collection from large complex farms are handled on an individual basis by the decentralized NASS field offices. In order to gain a better understanding of the approaches that these offices employed and found to be most successful, NASS conducted a survey of local offices to collect views about (1) the challenges and (2) their responses to those challenges. (See also Joe Parsons presentation, based on data from this survey.) This presentation identified a great deal of variation among the local offices, as reflected in field office responses. The presentation drew on the social psychology literature to consider factors that affected response rates, i.e., persuasion, influence, and cooperation.

Administrative Data for U.S. Agricultural Estimates: Ginger Harris, NASS, USDA
Abstract: This presentation reviewed the current sources of administrative data used by NASS for enhancing and updating the list frame of farmers, including lists of establishments reporting farm income (from IRS, the U.S. Tax Agency) and death lists (from Social Security Administration). NASS has found that aggregate administrative data is most helpful for aggregate estimates (such as, corn acres planted in a certain state and/or county) and most are compiled by summing field level data so are not impacted by the complexity of the organization structure. Problems arise in use when either the definition or the boundaries of the entities do not match between the operational survey program and the administrative data source. The paper drew on the lessons learned from a project using administrative data on government payments. Currently, there is a department initiative to standardize data collection, so farmers only need to report once to USDA for program administration. However, it remains unclear at this point if this includes NASS as a statistical agency, or just the program administration agencies.

Canada’s Enterprise Portfolio Management and Large Agricultural Operations Statistics (LAOS) Unit Approach:  Francine Lavoie, Enterprise Statistics Division, and Steven Danford, Census of Agriculture, Agriculture Division, Statistics Canada
Abstract: This presentation described the Statistics Canada system currently used to collect data from complex farm operations carried out by the Large Agriculture Operation Statistics (LAOS) Unit. Through this program, the data collection for a farmer with a complex operation is coordinated by Statistics Canada’s Agriculture Division so that the same information required on multiple surveys is collected only once. LAOS helps maintain a balance between the need for data and response burden. There was a great deal of interest in the challenges and opportunities of this system for application in other countries. LAOS is complementary to the Statistics Canada Enterprise Portfolio Management (EPM) program, whereby data collection from large and complex (nonfarm) businesses is coordinated through a single point of contact to manage respondent burden, timing and coherence of the collected data.

Lessons from the U.S. Census Bureau’s Efforts on the Large Company Contact Program:  Bob Marske,  U.S. Bureau of the Census  Paper   PowerPoint
Abstract: The Economic Census imposes a significant burden on the Nation’s largest companies, which have hundreds of locations and must complete a form for each of them. These large companies, which comprise at least 35% of the U.S. payroll, are critical to the Census. Because large companies are critical to published data, the U.S. Census Bureau has developed an account manager (AM) program. This program appoints a single contact to help each large company meet its filing requirement. The goal of the program is to help companies understand the census, facilitate use of electronic reporting tools, and accelerate response. While the AM program has been in existence for the past three censuses, the 2007 Economic Census was the most effective in terms of timeliness and unit response. This paper presents background on the AM program, new strategies used for 2007, and plans for the next census.

oncepts to Guide Us Moving Forward
Moderator, Mary Ahearn, ERS, USDA

Translation of Economic Theory into Practical Guidance for Data Collection from Farm Firms and Households: Defining the Target Population and Identifying Relevant Data Items:  Jean-Paul Chavas, University of Wisconsin
Abstract: An integrated analysis of data availability, choices, survey cost, and statistical methods can improve the flexibility, precision and usefulness of data collection and analysis. This paper presents a brief overview of these issues, and reviews what statistical theory and econometrics offer as guidance in the process of collecting and analyzing data. The paper also considers the optimality of data collection and analysis, recognizing the fact that investigators typically have incomplete experimental control. The paper includes a discussion of the definition of a “farm” for data collection purpose, both from a statistical viewpoint and an economic viewpoint. Arguments are presented stressing the role of microeconomic dynamics, and the need for better panel data to help us assess the role of managerial ability and its effects on economic adjustments to changing market conditions and technology. Unfortunately, panel data are rather rare. Yet, having access to panel data can be quite valuable. At this point, some important issues remain poorly understood in large part due to a lack of annual panel data. This includes the role of managerial skills in technology adoption, the role of human capital in market dynamics, and the role of education in the current obesity epidemic.

A Business School Perspective on Agriculture, David Sparling,  University of Western Ontario
Abstract: This presentation emphasized the difference in perspective between a business school and an economics perspective. In a business school perspective, profitability is key. It takes a first person, rather than a third person perspective. Therefore, the focus is on individual decision making and the importance of “owning” your problem. This is consistent with the farmer’s perspective. Both net income to the business units and capital appreciation are key. The paper advocates for a flexible approach to data collection. Because of the focus on individual decisions, a business school perspective finds great value in case studies.

Accounting Practices of Farming Organizations: How They Organize Business Information: Cathy Parciak, Quality Professional Accounting Corporation
Abstract: It is importantto recognize that farmers will organize their data in ways that meet their needs. This includes their needs to provide data to financial institutions, tax officials, owners, Canada’s Agristability/Agriinvest, and other government programs. It is also important to keep in mind that a goal of the farmer is to collect data that will lead to an outcome of producers paying their lowest legal taxes, e.g., many report income on a cash basis. It is very common in complex operations for the assets to be owned by multiple entities, however, they do not get separated out for financial statement purposes. Assets usually include equipment required for custom work, drying facilities, trucks, storage facilities, and even nonfarm work. It is common for farms to buy the output of other producers. For example, there are very few large horticulture producers that do not buy other farmers products to ensure that they have enough diversified produce to sell at Food Terminals, etc. The speaker used case examples of farming organizations to explain the complexity of the organizations for data collection and their motivations for their organizational form, e.g., minimizing taxes.

Innovative Examples and their Potential for the Next Decade
Moderator, Jim MacDonald, Branch Chief, ERS, USDA

The Role of Empirical Research for the Study of Complex Forms of Governance in Agroindustrial Systems: Lessons from Brazil:  Maria Sylvia M. Saes, Guilherme Fowler A. Monteiro, Silvia M. Q. Caleman and Decio Zylbersztajn, PENSA-USP
Abstract: This presentation discusses the role of empirical research in understanding the complex forms of governance in agribusiness. The authors argue that there are three fundamental levels of analysis: (i) the basic structure of the market, (ii) the formal contractual arrangements that govern relations within the agroindustrial system, and (iii) the transactional dimensions governed by non-contractual means. The authors take account of the impact of the concentration in the segments of the supply chains and business strategies. The article concludes by suggesting some indicators which could be collected by statistical agencies to improve understanding of the complex relationships among agribusiness segments.

Innovations Using Farm Records Systems: Allen Featherstone, Kansas State University
Abstract: The paper begins by providing a background of the Kansas Farm Management Association and an overview of reasons leading to additional organizational complexity are discussed. The presentation examines two cases of the organizational structures of actual farms in the Kansas Farm Management Association to understand the depth of the complexity and how that complexity may be accounted for to fully understand the implications for data collection. Finally, the paper provides conclusions regarding managing the complexity associated with data acquisition and performance measurement to adequately capture the entire farming structure. To truly have an understanding of the farm management decision making process, the collection of data on the multiple entities, or super farm, of an economic agent is necessary.

New Trends and Challenges: Integrating Variables and Observations from Multiple Instruments and Experiences in the Exchange of Farmer’s Financial Data among Accounting Offices, Fiscal Authorities, and Research and Statistical Agencies, Koen Boone, LEI Wageningen UR, and Krijn Poppe, Ministry of EL&I and LEI Wageningen, UR, The Netherlands
Abstract: The presentation argues that there is great potential for efficiency gains in data collection. The LEI at Wageningen utilizes many approaches to efficient data collection of complex organizations. Examples included exchanges of data with 5 accounting offices and other private sector firms in the building of the Dutch FADN. For farms for which there is full detail, data are collected from invoices for inputs, bank transactions, and through the linking with other data bases, based on the approval of participating farmers. The presentation also emphasized the importance of harmonizing of definitions, including basic definitions like the farm definition, in order to capture the potential of efficiencies in data collection from multiple sources. There is a large cost in coordinating across all parties to achieve these efficiency gains.

Future Agenda: Where do we go from here?
ynthia Clark, NASS, USDA
Cameron Short, Agriculture and Agri-Food Canada
Jeffrey Smith, Agriculture Division, Statistics Canada
Mary Bohman, ERS, USDA
Abstract: The workshop closed with remarks from leaders participating from the four North American statistical and economic government institutions. A common theme was their interest in fostering the solid programs of each of their agencies, concerns about tight budgets, and their appreciation of the efforts of the workshop participants and their expectations that collaboration will continue in the future.

Cameron Short raised numerous questions about how policy analysis should be conducted with the increasingly complex farm organizations. This is an important question to answer as solutions to the challenges of data collection are addressed because it should be the major driver of solutions. He also mentioned the importance of looking at distributions, rather than simple averages.

Mary Bohman indicated that we must change our approaches to data collection continually as the structure of agriculture changes. Although ERS and USDA believe the whole range of data collection activities have value, her comments were focused on the ARMS. The core use of ARMS is income, but the breadth of the information lends itself to many uses. She indicated that ERS is engaged in a complementary activity focused on reviewing alternative approaches to developing financial statements which also must deal with the complex nature of many farm organizations.

Jeffrey Smith emphasized the importance of customized collections, such as with LAOS, and expects to expand that effort.  The shift to customized collections shifts some burden from the respondent to statisticians in government agencies.  He emphasized that administrative data are fragmented (held by many different players in different jurisdictions) and we can not necessarily assume they are always coherent and ready to use for statistical purposes. There will likely be an increased reliance on administrative data as surveys are reduced in order to address respondent burden issues. There has been a recent initiative to expand access to data by researchers, although business micro data are limited at the present time.

Cynthia Clark reported she had gained insights from the workshop that she believes have implications for many of NASS’ programs. NASS currently is researching respondent burden and nonresponse bias, which is highly relevant to the complex farm issue. A major practice to reduce burden is sample manipulation across surveys by allowing field office leaders to not collect data from some organizations for some surveys. NASS currently does not have guidelines for how field offices should handle data collection for complexfarms, and they are allowed to use their discretion to identify samples as office holds for a variety of reasons. She believes that NASS should be writing procedures on how best to handle these decisions. Procedures should be informed by research, for example, a study by business schools could be beneficial and a better understanding of the value of administrative data, e.g., cooperate with EPA in California.


Share This Project