- Executive Summary
- Cooperatives in U.S. Economy
- Economic Impacts
- Commercial Sales & Marketing
- Social & Public Services
- Financial Services
- Discussion Papers
- Future Research
- Download Report
University of Wisconsin Center for Cooperatives
Research on the Economic Impact of Cooperatives
The aim of the project was to create a complete census of U.S. cooperative businesses and measure their economic impact on the U.S. economy. The process of creating a census involved three distinct steps:
Most businesses were identified with the help of key contacts in various trade associations, academic partners and collaborators, and primary population discovery conducted by the UWCC using business software. In the next section, we discuss each of these venues for population discovery.
For regulated industries such as credit unions, corporate credit unions, the farm credit system, and federal home loan banks, we used annual reports available at the regulatory Federal agencies' websites. The data for rural electrics comes from NRECA. Agricultural Marketing and Supply Co-ops data come from the USDA 2006 annual survey.
Purchasing cooperative lists were provided by NCBA, and housing cooperative lists were provided by NCB. The EPA provided a list of water mutuals and associations which was supplemented with Guidestar data.
For many sectors we created primary lists with the assistance of undergraduate researchers. Online searches were conducted with key phrases such as "co-op", "cooperative", and "mutual" for each economic sector. Once cooperatives were identified, lists were created and downloaded into a database with appropriate contact information.
Childcare, Healthcare, Mutual Insurance, Transportation, Education, Water and Waste, and Telephones lists were created using Google, Broadlook, Onesource, Dunn, and Guidestar; UWCC purchased the software. Finally, for grocery and worker cooperatives, we used lists maintained by Professor Ann Hoyt and Professor Christina Clamp, respectively.
We used standardized survey instruments and a uniform sampling methodology to minimize measurement error and to yield data that would be comparable across economic sectors. The instruments were also designed to identify businesses and collect firm-level data that can be used for future longitudinal studies of cooperative performance.
Implementing a survey involved numerous separate tasks. These activities included:
The identical survey instrument was used for all economic sectors, except that adjustments were made as needed for inherent structural differences. The core instrument has four sections:
The cooperative business surveys were targeted to a particular set of firms in the following sectors the USDA identified: Commercial Sales and Marketing; Social and Public Services' Financial Services; and Utilities.
Our interest was to collect firm-level data. A firm may have one or many establishments. Financial information for the purposes of this study was collected at the aggregate level, so all reported financial data is consolidated unless otherwise specified.
Our sampling strategy was as follows: If the total number of firms were <400 in a given economic sector, then we interviewed all firms in the list. Our goal was to elicit a 50% survey response rate. The following sectors were surveyed using this approach: Grocery and other consumer retail; Arts and Craft; Education; Healthcare; (not Community Healthcare Centers) Transportation, Bio-fuels; Telephone; and Purchasing and Worker cooperatives.
For economic sectors with >400 firms we selected a stratified random sample of 300 firms. We employed this approach for the following sectors: Mutual Insurance; Water; and Housing Cooperatives. Our sampling unit for stratification was U.S. states. We followed this approach to ensure that the resulting sample represented underlying distribution within each state for a particular economic sector. To preserve the anonymity of firms, we excluded any state that had fewer than 5 firms in a particular economic sector.
Even following this sampling strategy, identifying telephone numbers for cooperatives was sometimes difficult particularly in the case of . housing, and water and waste cooperatives. Most of these cooperatives are small, or without offices, and no one is available during regular business hours. To maximize data points, we redrew our stratified sample from firms with telephone numbers, preserving the population distribution.
We piloted the survey to pretest the questions to minimize question ambiguities, check for clarity and consistency, incorporate input from key participants, and allow survey modification to address sector-specific differences. Finally, piloting enabled better training of enumerators. Our piloting consisted of up to 20 interviews, depending on the number of firms in the sector.
Publicizing a survey increases participation. Because we were surveying multiple sectors simultaneously, we used various mediums to invite participants. To increase participation, we solicited help from trade associations to distribute invitations to member lists, on their websites, and in their newsletters. UWCC also posted an announcement about the survey on its website, mailed invitation letters and e-mails, and often extended direct invitations by telephone.
We intended to create a web form that firms could visit annually to update their profile. Although we followed this approach early in survey implementation, survey responses were not adequate. We therefore hired a staff of 12 students to conduct phone surveys to reach this desired 30% response rate. Calling individual firms and scheduling appointments with the CEO or accountant was more efficient, because this approach gave the respondents time to collect financial information before the phone survey.
Using supplementary data from Guidestar, and Onesource we attained a response rate of 30% for all sectors except housing. We surveyed the following sectors: healthcare; childcare; groceries; purchasing; worker; transportation; education; telephones; water and waste; mutual insurance; farm credit system (only employment information); arts and crafts; housing; and bio-fuels. We contacted each firm at least three times. Specific response rates for each economic sector are provided in the sector analysis section under "population discovery".
Although the data needed for this economic impact analysis was fairly straightforward, the reporting of financial information varies greatly by sector and posed challenges to standardizing data for analysis. This was especially true for defining a patronage refund. Further research needs to carefully document patronage practices across cooperatives.
Once the data was standardized, it was used to create the maps and the IMPLAN analysis.