Using Data from the National Center for Charitable Statistics (NCCS) for Advocacy-Related Quantitative Research by Jeff Krehely Contents I. Lobbying Variables II. Strengths of the Current Data III. Weaknesses of the Current Data IV. Future Data Improvements V. Works Cited The National Center for Charitable Statistics (NCCS) , a project of the Center on Nonprofits and Philanthropy (CNP) at the Urban Institute, is the national repository of data on nonprofit organizations in the United States. NCCS Core Files are produced annually and are compiled from information 501(c)(3) nonprofit organizations report to the IRS, primarily on Form 990 and Form 990's Schedule A. More specifically, the Core Files combine descriptive information from the IRS Business Master Files and financial variables from the IRS Return Transaction Files, after they have been cleaned by NCCS. The 1999 Core File contains 125 variables for the approximately 230,000 501(c)(3) organizations required to file a Form 990 with the IRS The Core Files contain several variables that can be used to analyze the amount of money charities spend on lobbying activities. Lobbying is just one type of advocacy in which these groups may engage; other forms of advocacy include educating the public on a particular issue or cause; providing research or testimony to government officials; litigation; and monitoring the implementation of public policies. Due to current data limitations - outlined below - it is difficult to assess these broader forms of nonprofit advocacy. In short, the lobbying data provide us with an overview of one piece of the advocacy picture. Future data improvements will help overcome this challenge to nonprofit advocacy research. I. Lobbying Variables The lobbying variables are derived from Form 990's Schedule A, which all registered 501(c)(3) organizations with gross receipts greater than $25,000 are required to file annually with the IRS. These variables are derived from responses to the following lines on Schedule A: Part III, Line 1: "During the year, has the organization attempted to influence national, state or local legislation, including any attempt to influence public opinion on a legislative matter or referendum? If 'Yes,' enter the total expenses paid or incurred in connection with the lobbying activities." Part VI-A, Line 36: "Total lobbying expenditures to influence public opinion (grassroots lobbying)." Part VI-A, Line 37: "Total lobbying expenditures to influence a legislative body (direct lobbying)." All organizations must respond to Part III, Line 1. However, Part VI-A Lines 36, 37, and 38 are only answered by organizations that make the 501(h) election to lobby. The Core Files also contain data on "excess" lobbying expenditures for electing organizations. According to the data, in 1998 only 37 organizations exceeded the limits for grassroots lobbying (total excess grassroots spending equaled $456,362), and only 57 exceeded the limits for direct lobbying (total excess direct spending equaled $717,149). Total lobbying limits are currently set at 20 percent of the first $500,000 of exempt purpose expenditures, and then calculated on a sliding scale based on total exempt purpose expenditures, up to a cap of $1 million on total lobbying expenditures. The limit on grassroots lobbying is set at 25 percent of an organization's total lobbying limit (because total lobbying is capped at $1 million, grassroots lobbying is effectively capped at $250,000). II. Strengths of the Current Data The Core Files provide the most comprehensive financial data available on charitable organizations. Directly surveying all 220,000 organizations in the data set to get this information would be costly and tedious. And due to the diverse nature of these organizations, choosing a representative sample to survey would be difficult. In short, these files, and other data sets available through NCCS, are invaluable for quantitative research on the nonprofit sector. Particular to advocacy research, the Core Files are quite useful for studying lobbying activities of 501(c)(3) organizations: These files go back to 1989, so detailed time-series analyses on these variables are possible. It is also possible to focus on a particular category or type of organization and quantitatively analyze how lobbying groups differ financially (if at all) from non-lobbyists in that category or type. By sorting out the electing organizations from the non-electing groups among all groups that lobby, research can also be done to see how non-electors differ financially (if at all) from electors. For people interested in conducting qualitative research on organizations that lobby, the Core Files can be used to provide financial background and trend information for in-depth case studies. With the variety of financial variables available, the data could also be used to write shorter organizational profiles of particular organizations. In general, working with the lobbying variables is a simple way to locate and learn about the politically active 501(c)(3) organizations that are required to file a Form 990 with the IRS. III. Weaknesses of the Current Data The Core Files have several weaknesses that must be considered when conducting research on the advocacy activities of 501(c)(3) groups: First, and most generally, if NCCS and other IRS data are to be used for advocacy-related research, it is important to remember that advocacy and lobbying are not synonymous terms. As mentioned earlier, organizations that advocate can use a variety of methods to influence public opinion or the government - lobbying national and state legislatures is just one of those methods. Overall, the Core Files tell us very little about nonprofit advocacy, broadly defined. Second, the IRS does not collect data on all organizations that are technically 501(c)(3) charities. For example, many religious-related organizations, including churches and associations of churches, are not required to register with the IRS. Organizations with less than $5,000 in annual gross receipts are also not required to register. Among those that must initially register, only those organizations with greater than $25,000 in annual gross receipts must file a Form 990 with the IRS. Mary Anna Culleton Colwell (1997) estimates that the IRS has records of only 10 percent of the "voluntary action field." Her research also shows that many of these unregistered organizations are likely to be involved in advocacy activities. (Considering points one and two together, the NCCS core files understate the frequency and magnitude of the advocacy work of charitable organizations.) The third weakness, which is not particular to lobbying research but all uses of the Core Files, is that the organizations themselves provide this information to the IRS, based on their understanding of the instructions for Form 990 and Schedule A. Because there are over 200,000 organizations in the 1999 Core File, it is inevitable that reporting errors have occurred. As one example, a few organizations enter the same value for total lobbying expenditures and for Line 81A on the Form 990, which captures the amount spent for political purposes (related to elections). This line is on the 990 primarily because other tax-exempt organizations (such as labor unions and trade associations) also use the form and must report this information. The 501(c)(3)s are legally barred from spending money on directly influencing the outcome of an election, so they should not be reporting any information on this line. Fourth, most organizations do not engage in lobbying and therefore do not report any lobbying expenses. It is difficult, then, to do any large-scale analysis of how all lobbying groups differ from all those that do not lobby, because the overwhelming majority of non-lobbyists dilutes the sample to an extreme. Looking at differences between particular types of organizations - for example, comparing environmental groups that lobby to those environmental groups that do not - is one remedy to this problem. Finally, many of the key variables for analysis are aggregate measures of several smaller variables, which limits the scope and depth of any analysis. For example, the contributions revenue variable is actually a total of donations from private individuals and foundations, as well as government grants. Likewise, the total expenses variable is a function of many other subcategories, including administrative and program costs. Being able to analyze specific components of an organization's revenue stream or its expenses would allow more in-depth research on the financial characteristics of groups that lobby, as well as possibly find subtle, hidden relationships between these subcategories and an organization's decision to lobby. For example, preliminary work by Lester Salamon (1995) suggests that there may be a positive relationship between government sources of revenue and an organization's propensity to advocate. This research could be expanded or made more certain if detailed revenue data were available. IV. Future Data Improvements Despite these data weaknesses, there is hope that quantitative advocacy research will improve in the near future. The new "Digitized Data" that NCCS is currently preparing, in cooperation with Philanthropic Research Inc., has the potential to enhance all of the strengths of the current data, while compensating for several of the weaknesses. The Digitized Data, like the Core Files, are based on IRS Form 990 and Schedule A, but include a variable for every entry on the form (which adds up to over 400 variables). The Core Files, by comparison, include only a fraction of all entries. Here are some of the highlights of this new data resource, as it pertains to lobbying and advocacy research: Information on the lobbying activities and techniques of non-electing organizations will be available for analysis. New variables representing the sub-categories for the currently aggregated revenue and expense variables will allow in-depth research on how these particular expenses and revenue streams relate to lobbying activities. Program descriptions, as reported in Part III of Form 990, will be included in the data set. Keyword searches can be performed on these descriptions to identify the organizations that engage in broader advocacy work. Line 80 on Form 990 - which requires organizations to list any separate organizations they are financially and managerially related to - will also be available for searching and analysis. This information will allow researchers know if a 501(c)(3) organization has an affiliated 501(c)(4) social advocacy group through which it might conduct a larger advocacy or lobbying agenda. The Digitized Data are currently being prepared at NCCS. Extensive analyses on several aspects of this data - including the new advocacy and lobbying data - are currently being planned. V. Works Cited Colwell, Mary Anna Culleton. 1997. "The Potential for Bias When Research On Voluntary Associations Is Limited to 501(c)(3) Organizations." Aspen Institute, Washington, DC: The Nonprofit Sector Research Fund Working Paper Series. Autumn. Salamon, Lester M. 1995. "Explaining Nonprofit Advocacy: An Exploratory Analysis." Draft paper prepared for delivery at the Independent Sector Spring Research Forum. Comments and questions may be sent via email to advocacyresearch@ui.urban.org.