Funding Faction or Buying Silence? Grants, Contracts, and Interest Group Lobbying Behavior

By law, federal funds may not be used for lobbying, but many corporations and nonprofit organizations that receive government grants and contracts also engage in lobbying activities using private funds. The legal tightrope that this creates has been the focus of numerous attempts at policy changes during the past decade. Some of these changes have simply sought better disclosure by recipients of federal monies, but others have been aimed specifically at nonprofit organizations and have attempted to prevent them from lobbying the government on any topic if they receive federal funds.

These efforts include the unsuccessful Istook amendment and its successors, which focus on lobbying activity by 501(c)(3) charitable organizations; the SimpsonCraig amendment banning all lobbying activity by 501(c)(4) social welfare organizations receiving federal funds; and the "Truth-in-Testimony" rule, under which witnesses appearing before House committees must disclose federal grants and contracts (Gelbart 1997; Weinstein and Stofferahn 1996). Nonprofit advocates say they have noted a rise in such efforts, citing letters from administration officials discouraging advocacy by Head Start programs, proposed provisions in the Individuals with Disabilities Act of 2003 and the Federal Housing Finance Reform Act of 2005, and increased tax audits of politically active nonprofits (Bass, Guinane, and Turner 2003).

The proposals have been attacked as partisan attempts to "defund the left" by liberal organizations, while conservative organizations have lauded such proposals as necessary to stop "the funding of faction" through taxpayer-subsidized lobbying activities (Bennett and DiLorenzo 1985; Wittmann and Griffin 1995). Both liberals and conservatives have criticized the proposals because they address only the lobbying of nonprofits while allowing defense contractors and other recipients of federal contracts to lobby as much as they care to (Gallagher 1995).

The rationale behind these efforts to limit lobbying stems from the idea that group resources are fungible, and therefore that even those organizations that abide by the letter of the law may be using government funds to indirectly subsidize their advocacy efforts (Wittmann and Griffin 1995). These policy proposals in effect assume that lobbying by grant- and contract-receiving interest groups is not proportional to lobbying by nonfederally funded groups, and that government monies are allowing federally funded organizations to lobby more than they otherwise would (and thereby use government funds to seek yet more government spending). Opponents of lobbying limit focus on the free-speech issues that such limitations raise. Lobbying is clearly a form of political speech, and while we certainly would not want taxpayers paying for interest groups' political activities, neither would we want the government to prevent the normal level of speech that such a group might engage in.

This suggests an empirical question, one that has not previously been addressed directly: Is the level of lobbying engaged in by organizations that receive government funds different from the level for similar organizations that do not receive government funds? That is, is there any empirical evidence that government spending is not only driving an increase in lobbying but funding an increase in lobbying?

Existing research casts doubt on this assumption. Some studies suggest that organizations that are highly reliant on government funding, far from lobbying more energetically, may tend to muffle their advocacy attempts to avoid alienating their budget sources (see Beck 1970; Gais and Walker 1991; Smith and Lipsky 1993, 57-38). Other scholars have found that the majority of nonprofits with charitable status incorrectly believe that they are prohibited by law from lobbying and therefore are very unlikely to do anything that might be construed as political (Berry and Arons 2003). While government attention to a policy area tends to increase the level of lobbying in it, government spending alone has no independent effect on the amount of lobbying in a policy area (Leech et al. 2005). And although new government programs often create constituencies that then become willing to fight to preserve those programs (Campbell 2003), this phenomenon occurs whether the program in question is delivered by a nonprofit or by a government agency.

Berry and Arons (2003, 87-90) have found that the charitable organizations that lobby the most are the least likely to say that they are highly motivated to seek government grants and contracts. In a multivariate analysis, the desire for government funding was not a significant predictor of the amount of lobbying the organization engaged in. Berry and Arons did not, however, directly compare actual recipients of federal funds to nonrecipients, and their sampling frame is all 501(c)(3) organizations, regardless of where in the country they are based and regardless of whether or not they are politically active. This sample was appropriate for their research questions, but makes their data less appropriate for assessing whether congressional concerns about nonprofits that are politically active are well founded or not. The analysis in this paper, by focusing on a sample of politically active nonprofits, attempts to address these remaining questions.

The questions surrounding advocacy and nonprofits are important because government spending on services delivered by nonprofits has more than tripled over the past two decades. Government grants and contracts to nonprofits totaled more than $200 billion in 1998 compared to about $60 billion in 1982 (Independent sector 2001).1 The majority of these grants are aimed at providing health and social services and reflect a growing tendency since the late 1970s for both the state and federal governments to contract with the nonprofit sector to provide such services. Smith and Lipsky (1993) note that most of these grants and contracts are not gifts that nonprofits spend as they wish, but rather come with explicit requirements for how they are spent. Governments "contract for whole programs, and even create providers where they otherwise do not exist" (1993, 9).

This article uses data from a survey of more than 700 politically active interest organizations with offices in Washington, DC, to measure the lobbying behaviors of federal contractors and grantees and compare them with interest groups that do not receive such funding. I consider not only the nonprofit entities on which the congressional policies have been focused, but also businesses to assess whether federal grants and contracts also spur these for-profit entities to lobby more extensively. I begin with a discussion of theories of group motivations in lobbying, followed by a description of the data and an analysis of the organizations' lobbying behaviors.

Why Do Groups Lobby?

Lurking unstated behind congressional attempts to limit lobbying is a theory of interest groups as rent-seekers, entities that are motivated primarily by benefits for the group or its leaders at the expense of the public (see, e.g., Buchanan, Tollison, and Tullock 1980; Tollison and McCormick 1981; Tullock 1967). While pluralist and neopluralist perspectives on interest groups emphasize the representational roles interest groups can fill, rent seeking and related "transactions" theories of lobbying emphasize the wasteful, antidemocratic properties of "special interests" (see Baumgartner and Leech 1998; Lowery 2005; Lowery and Brasher 2003; Mansbridge 1992; McFarland 2004 for discussions of these various perspectives). In the context of politics, rent-seeking usually refers to efforts by industries to use the coercive power of the government to seek profits over and above what they could achieve in the unregulated marketplace. The benefits the businesses seek from government may be in the form of direct payments and subsidies, or in the form of regulations such as tariffs and licensing that limit the entry of competitors into the market. They may also be in the form of rules that make the product the business produces more desirable-for example, regulations classifying sports utility vehicles (SUVs) as light trucks rather than passenger vehicles made it possible for SUVs to be produced more cheaply because fewer safety features were required.

Rent-seeking theory, developed as it was by economists, deals primarily with lobbying by economic interests rather than by groups interested in social policy. By extension, however, the theory might predict that groups interested in social goods have parallel motivations. While the leaders of businesses seek to make their enterprises successful and to increase profits, the leaders of these ideologically based groups seek to make their enterprises successful and have greater effects on society. These nonprofit groups may, for example, seek to provide more food for the poor, greater restrictions on abortion, or more stringent testing of school children. To achieve these goals, they have two basic options. The first is to turn exclusively to the private sector: raising money from private sources, launching their own anti-abortion educational campaigns, or themselves recruiting teachers and principals who also favor testing. This would allow their groups to survive as well as further their goals for society. The second option is to make use of the power of government to tax, ban, or regulate. Harnessing this governmental power could make the work that the groups advocate occur much more quickly. And so, nonprofits may choose to spend at least some of the time that they might be fundraising, educating, and organizing, and instead spend it lobbying. Pluralist and postpluralist approaches here would emphasize the representational role that this lobbying can serve and the important information that can be transferred from individuals and organizations to elected and appointed officials through the lobbying process. For rent-seeking theorists, however, rent-seeking is always a negative sum game, as it involves the use of government authority to transfer benefits to the rent-seeker: "Resources devoted to efforts to curry the queen's favor might be used to produce valued goods and services elsewhere in the economy, whereas nothing of net value is produced by rent-seeking" (Buchanan 1980, 8).2

More General Theories Of Lobbying

The decisions groups make about when and where and whether to lobby are complex. The theory of rent seeking, taken alone, predicts that organizations will lobby whenever they can successfully extract benefits from the government. In addition, both economists and political scientists are well aware that problems of collective action affect which organizations will form and decide to lobby in the first place, with larger organizations and organizations that were formed primarily to provide private goods to their members more likely to take the lead (Olson 1971). Since this article considers only existing organizations, the issue of group formation is less central. But even once the collective action dilemma has been overcome, not all groups lobby with equal intensity.

Once a group has formed and has decided to petition the government in some way, many other factors affect decisions about what type of advocacy and how much advocacy will take place. Lowery (2005) points out that a full theory of why groups lobby must begin with the organization's need to survive as an organization. Lobbying decisions are based not only on what will maximize profits and/or ideological goals, but also on which decisions allow for organizational maintenance. These decisions take place in a particular political venue and particular issue context, and these in turn affect the lobbying decisions that are made (Baumgartner and Leech 1998).

The level of lobbying by an organization on a given issue is therefore expected to vary not only by the organization's degree of interest in the issue, but also by external, issue-based variables and variables related to organizational maintenance. .Issues in which the government is active, that attract many other interest groups, that are publicly salient, and feature partisan conflict will attract more lobbying than smaller, less conflictual issues will. The level of lobbying is also expected to be affected by internal, group-level variables related to organizational maintenance, including tax status and source of income. I will consider each of these variables in turn.

Issue Context

It is reasonable to expect that the actions of government should affect the lobbying behavior of organizations. Certainly that assumption lies behind attempts to curtail lobbying by recipients of federal funds. Government affects the fortunes and futures of organizations not simply through contracts and grants, however, but also through policies and regulations. In their study of interest group registrations to lobby before the federal government, Leech et al. (2005) found that the number of congressional committee hearings in an issue area was a better predictor of the amount of lobbying than the size of the federal budget for that issue area. Government activity certainly drives lobbying activity, but the work of Leech et al. suggests that government activity is more important than government spending per se. Gray et al. (2005), and Lowery et al. (2004) also find support at the state level for the idea that government activity leads to greater lobbying activity in a policy area.

Many interest-group scholars have noted the importance of other external political and issue-related factors in influencing levels and types of lobbying, although few have had the data to directly test these theories (see, e.g., Baumgartner and Leech 1998; Berry 1977; Schlozman and Tierney 1986 for general discussions; for partial tests see Hojnacki 1997; Hojnacki and Kimball 1998 and 1999). These theories suggest that organizations that are involved in issues that are highly conflictual may lobby more than organizations that are involved in issues that are fairly consensual, with little partisan conflict or conflict among interest groups, so these factors should be controlled for in the analysis (but see McKay and Yackee 2004 who find that conflict does not lead to more lobbying in agency rule making). Kollman (1998) has found that outside lobbying in particular increases when there is public support for an issue, and certainly lobbying of all types increases when an issue is salient to the general public, as this salience increases the chance that government will act on the issue. For the same reason, lobbying would also be expected to increase when organizations have allies within the government who would be sympathetic to the organizations' efforts-no group likes to waste its time, and groups with many governmental allies are likely to believe that their chances of success are better than groups with fewer such allies.

Organizational Maintenance

Both the rent-seeking perspective and the pluralist and postpluralist perspectives lead to the expectation that we should find more lobbying where there is more government activity. Whether it is to profit from direct grants and new regulations, or to speak out on issues that affect their organizations, interest groups are more likely to lobby on issues and areas of the economy in which government has been most involved. Factors related to organizational maintenance may temper this tendency, however.

Gais and Walker (1991) have shown that among the not-for-profit associations Walker studied, the factor source of income affected lobbying behavior. Reliance on private and governmental patrons for organizational funds, for example, tended to moderate a group's lobbying tactics. This would work in opposition to any tendency for government payments to increase lobbying levels. Membership-based groups that are highly reliant on dues, however, are more likely to turn to more public types of lobbying to demonstrate to their widely flung membership how active the group is.

In addition, government regulations may directly restrict some types of lobbying. Berry and Arons (2003) describe 501(c)(3) organizations as the group type with the strongest disincentive to lobby because of government regulation limiting their actions. Having 501(c)(3) tax-exempt status is advantageous for an organization because its supporters then can legally deduct their contributions as charitable donations, unlike contributions to most other types of nonprofits. Such tax benefits come at a price in terms of advocacy, however. Organizations with 501(c)(3) status may not participate in any campaign activity for or against a particular candidate, and they may lobby only if their lobbying activities do not make up a "substantial part" of their activities.3 These restrictions and the incentives to retain advantageous tax status lead to an expectation that 501(c)(3) organizations will lobby less than other types of organizations. This should especially be the case for advocacy activities that are limited or banned for charitable organizations.

In addition to external contextual variables and variables affecting organizational maintenance, the level of a group's resources is expected to affect the amount of lobbying it does, with organizations with larger budgets and/or larger lobbying staffs able to lobby more intensely. The number of lobbyists an organization employs is of course to some extent endogenous to the amount of lobbying it is engaged inthat is, over the long term an organization that wants to lobby should attempt to garner additional resources and to hire more lobbyists. Nonetheless, going from wanting more lobbyists to finding the money to pay more lobbyists is not an easy transition, and wealthy organizations with many resources are expected to have more lobbyists than less well-to-do organizations. In particular, businesses and trade associations that are active in Washington tend to have larger lobbying staffs than nonprofit organizations, so it is important to control for that tendency.

Data and Methods

Data presented here comes from a mail survey conducted by the author of more than 700 interest groups. The organizations were randomly chosen from among all organizations that maintained their own offices in the Washington, DC area and were listed in the Washington Representatives directory of interest organizations.4 There are more than 700,000 organizations in the nation with 501 (c) (3) status, and an additional 140,000 organizations with 501(c)(4) status (Independent sector 2001). Less than 1 percent is active enough politically to appear in my sampling frame. Thus, the question being addressed by this article is whether among politically active groups in Washington the receipt of federal funding affects the type or amount of lobbying. Berry and Arons' work (2003) already suggests that among a random sample of all 501(c)(3) groups, recipients of federal funds are less likely to lobby. But because nonparticipants in the political process dominate their results, it remains to be seen whether a sample of groups that are active politically show the same pattern.

The organizations in the sample range from a corporation with more than 40 full-time government representatives to a small citizen group that relied solely on volunteer labor, including the use of unpaid college interns as lobbyists. The respondents were asked to provide background information about their group and then answer questions about their advocacy efforts on the most recent policy issue in which they had been involved, including questions about 30 tactics of advocacy. As the respondents were selected randomly, selecting the most recent issue of any type in which the organization was involved creates a random snapshot of advocacy efforts on all issues current at the time of the survey. Respondents also were asked questions about their usual levels of lobbying and about their membership and finances.

Measuring Lobbying

Definitions of lobbying vary substantially from regulation to regulation and law to law, so it makes sense to measure lobbying, the dependent variable, in a variety of ways. The description of lobbying in the Individuals With Disabilities Act was quite broad ("federal relations"), so a measure of lobbying that is as inclusive as possible would best match its intent. Under tax law, which limits lobbying by 501(c)(3) organizations, "lobbying expenditures" refers only to expenditures aimed at influencing legislation, excluding agency lobbying or litigation but including many grassroots efforts. Congressional opponents of nonprofit lobbying are in general most concerned with congressional lobbying because this sort of lobbying is seen as leading to political pressure and rent-seeking activities, which in turn are feared to be driving additional government spending. On the other hand, lobbying of agencies should be expected and allowed by recipients of federal funds, since it is the agencies that distribute and administer these grants and contracts. These additional contacts may in fact be required by the government rather than sought by the interest groups. The four dependent variables used in the analyses of all of the groups are thus measured as follows:

Lobbying Index. This is the broadest measure of lobbying, built on the reported frequencies of 30 separate lobbying tactics, which range from contacting a member of Congress, to writing a research report, to engaging in protest.5 Each tactic is measured on a scale ranging from 1 (never) to 5 (daily), and the resulting additive index ranges from 30 to 150, with a mean of 75.43 and a standard deviation of 19.47. Cronbach's alpha, a measure of index reliability, was 0.94.

Regulated Lobbying Index. This index taps the lobbying behaviors that are limited or banned under tax law for 501(c)(3) organizations-lobbying aimed at affecting legislation (limited), grassroots efforts aimed at influencing legislation (limited), and involvement in elections (banned). There are eight items comprising the index6 ranging from 8 to 40, with a mean of 21.17 and a standard deviation of 6.86. Cronbach's alpha was 0.87.

Agency Lobbying Index. This index comprises five lobbying tactics aimed at federal agencies: testifying at agency hearings, directly contacting agency officials, submitting written comments or reports, responding to requests for information or briefing agency officials, and working with people within an agency who agree with this organization's policy position. The index ranges from 8 to 40, with a mean of 13.59 and a standard deviation of 5.23. Cronbach's alpha was 0.88.

Lobbying Frequency. The indices above are based on reported lobbying on a single issue. To address concerns that the most recent issue may not be related to the reason the group received government funds, a variable that measures average lobbying frequency is also included. This variable, based on a question that asked organizations to estimate how often they usually lobbied, ranges from 1 to 5, with a mean of 3.93 and a standard deviation of 1.6.7

Measuring Funding

The independent variable of central concern in these analyses is reliance on federal funding. This variable is based on a survey question that asked interest-group leaders to assess the percentage of total group income that came from eight different sources, including dues, fund-raisers, state and local government grants and contracts, and federal government grants and contracts.8 Reliance on federal government funds is a dichotomous measure that takes a value of 1 if the organization received any income from a federal grant or contract and a value of 0 otherwise.9

Lobbying and Federal Funds

State and federal governments often rely on nonprofits to provide social services, and support these services through contracts or grants. According to the Independent Sector, a research and advocacy organization for nonprofits, charitable and social welfare organizations, on average, receive about a third of their revenues (31.3 percent) from government sources (Independent Sector 2001). Table 1 shows that among politically active organizations, the levels are much lower. Politically active organizations with 501(c)(3) tax status overall receive about 11 percent of their income from the federal government.10 More than 60 percent of these organizations, however, receive no federal funding at all. For the subset that does receive federal funds, 29 percent of the budget came from federal funds.

For tax-exempt organizations that do not have status as charities, the percentages are much lower. Mean reliance on federal funds for this category of groups is only 3 percent, and only 40 of the 293 organizations in this category received any federal funds at all. Even for those 40 groups, mean reliance on federal funds was only about 19 percent-much lower than the national average for all nonprofits. Of the three tax-status categories, business had the highest mean reliance on federal funds. For businesses in the sample taken as a whole, federal funds comprised 13 percent of the total estimated budgets. About a third of the businesses (54 out of 147 total) reported receiving some government contracts, and relied on these contracts for 34 percent of their total budgets.

These data indicate that overall reliance on government funds is lower among politically active groups than among nonprofits as a whole, as expected. However it remains to be seen whether politically active organizations that do receive government grants and contracts actually lobby more intensely.

Comparison of Lobbying Levels

The stated rationale behind limiting lobbying by nonprofits that receive federal funds is to avoid taxpayer subsidy of partisan politics. If recipients of federal funds are having their advocacy efforts subsidized, then we would expect to see higher mean levels of lobbying among recipients of federal grants and contracts than among nonrecipients. This section of the article tests that expectation.

First, I look at mean levels of lobbying for groups with all types of tax status combined-501(c)(3), other tax-exempt organizations, and businesses. Table 2 compares the mean levels of lobbying between organizations that received federal funds and organizations that did not receive federal funds. The data show that for the broad index of lobbying, regulated lobbying, and overall lobbying frequency, there is no statistically significant difference between groups that received federal funds and those that did not. For agency lobbying, however, the difference is statistically different, although the substantive difference is slight (a score of 13.28 on the index versus a score of 14.44).

While these simple means tests match my predictions, it is possible that while there is no overall difference in the lobbying levels of recipients and nonrecipients of federal funds, there may be effects within the different types of tax status. That is, we know from previous studies that 501(e)(3) organizations tend to lobby less than other organizations, most likely because of the restrictions put on them through the tax code. But it is possible that 501(c)(3) organizations that receive federal funds lobby more frequently than 501(c)(3) organizations that do not, and the same pattern could hold for either businesses or tax-exempt groups that lack charitable status. If one of these group types but not the others lobbied more frequently in the presence of federal funds, that could explain the lack of difference seen in the means from Table 2. The next analysis addresses this concern.

Table 3 compares the means of groups with three different types of tax status depending on whether the organizations did or did not receive income from a federal grant or contract. Because I am considering so many means, the possibility of committing a Type I error increases, and I turn to analysis of variance (ANOVA) to test for these differences. The ANOVA results presented in Table 3 indicate that there were significant differences in intensity of lobbying across the different categories, but do not allow us to determine whether these differences are attributable to tax status (i.e., businesses lobby more than charities do) or to differences in funding sources (i.e., whether organizations with federal funds lobby more).11 Therefore post hoc Scheffe and Tamhane tests of these differences were conducted to assess which of the means were different.12 The results of these tests are indicated in Table 3 by the letters in parentheses after each mean. Within each column, means that share a letter are not statistically different from each other.

The analysis shows that the greatest differences in lobbying levels come not from whether or not an organization receives federal funds, but from its tax status. Across the different measures of lobbying, businesses and tax-exempt organizations without 501(c)(3) status tend to lobby the most, with 501(c)(3) charitable organizations trailing behind. For the purposes of the research question posed in this article, however, the most relevant differences are those that occur within each tax status. For the most part, such differences do not exist. The lobbying measures in the first three columns show no differences between charitable organizations that do receive federal funds and those that do not, no differences between other tax exempt organizations that do and do not receive federal funds, and no differences between businesses that do and do not receive federal funding. These three columns, measuring overall levels of lobbying and levels of Internal Revenue Service-restricted types of lobbying, show no support for the idea that organizations that receive federal grants and contracts lobby more than organizations that do not receive such funds.

The final column, measuring the frequency of agency lobbying, shows a somewhat different story. Here the Scheffe test indicates that 501(c)(3) organizations that receive federal funds contact government agencies more frequently than 501(c)(3) organizations without such funds. Although the mean score for the other tax-exempt organizations with federal funds is the highest among the groups, we cannot say with certainty that it is statistically different from equivalent groups without federal funds because the internal variance of this category is great and there are so few groups in this category. Businesses also show no significant variation between those with funds and those without on this index. Although the mean amount of agency lobbying is slightly higher for businesses with federal funding, the difference is not statistically significant. These results support the contention by previous scholars that the receipt of federal funds is not a positive correlate of lobbying intensity, except in the case of agency lobbying. The positive relationship between federal funding and agency lobbying makes sense given the need for recipients of grants and contracts to contact agency officials and file reports, and the resulting increased probability that such organizations would develop allies within the agency and be asked to testify at agency hearings.

A Model of Lobbying

Three of the lobbying frequency indices assess organizational lobbying on a single issue. Issues vary in terms of salience, conflict, and the number of venues available for interest-group access, and so a final test of the findings of this article should take the form of a multivariate model that can control for these other factors. Groups vary in ways other than tax status, and groups with greater resources will lobby more frequently than groups with fewer resources. It is possible that once these factors have been held constant, federal funding may affect lobbying levels. The model that results from the theory outlined above is:

E |Lobbying levels^sub i^| = ß^sub 0^ + ß^sub 1^Lobbying staff^sub i^ + ß^sub 2^Government attention^sub i^ + ß^sub 3^Public salience^sub i^ + ß^sub 4^Party conflict^sub i^ + ß^sub 5^Opposing groups^sub i^ + ß^sub 6^Institutional allies^sub i^ + ß^sub 7^Tax status^sub i^ + ß^sub 8^Dues^sub i^ + ß^sub 9^Federal funds^sub i^ + e^sub i^

where

Lobbying staff = an ordinal measure of the number of government relations staff members employed by the respondent's group.13 Expected to be positively related to lobbying levels.

Government attention = a count measure of the number of congressional committees and agencies considering the issue of interest to the respondent. The higher the number, the greater the number of venues in which the interest group might lobby on this issue.14 (For the agency lobbying analysis, only agency venues are included.) Expected to be positively related to lobbying levels.

Public salience = the degree to which a respondent thought public opinion was important to the outcome of the policy issue in question.15 Expected to be positively related to lobbying levels except for agency lobbying, where no relationship is predicted.

Party conflict = the degree to which the respondent thought Republicans and Democrats differed on the policy issue, measured on a 5-point scale.16 Expected to be positively related to lobbying levels.

Opposing groups = a dichotomous variable indicating whether the interest group faced opposition by one or more interest groups on this policy issue.17 Expected to be positively related to lobbying levels.

Institutional Allies = an ordinal measure indicating the breadth of support that the interest group has among government actors, such as allied committees, agencies, and other officials (i.e., White House, courts). The variable is at its maximum (3) when at least one of each type of ally is present.18 (For the agency lobbying analysis, only agency allies are considered.) Expected to be positively related to lobbying levels.

Tax status = a dichotomous variable indicating whether the interest group has 501(c)(3) tax status.19 Expected to be negatively related to lobbying levels, with the strongest effects seen in the regulated index of lobbying.

Dues = the percentage of an organization's operating expenses that are derived from dues.20 Expected to be positively related to lobbying levels.

Federal funds = a dichotomous variable indicating whether the interest group received any income from federal grants or contracts during the past year. The nature of this relationship is disputed.

The results of the regression analyses on the four measures of lobbying are shown in Table 4. Across all of the models, organizations with more staff lobby more, and this measure of resources consistently has the greatest impact of any of the explanatory variables. Across all of the indices related to a particular issue, government attention and institutional allies are among the important predictors of lobbying. And for all of the models except one, organizations with 501(c)(3) tax status lobby less than other types of organizations, and federal grants and contracts are not a significant predictor of interest-group lobbying.

Let us first consider the lobbying frequency analysis, which employs a simpler model than the other three because the dependent variable reflects lobbying in general, rather than on a particular issue. The issue-based variables such as conflict and allies are therefore dropped in this analysis, and we consider only lobbying staff, tax status, reliance on dues, and federal funding. This stripped-down model supports the conclusion that while controlling for group resources, federal funds still have no impact on the level of lobbying by a group. Charitable tax status as a 501(c)(3) organization, however, is inversely related to lobbying levels, as expected. Charitable status depresses the levels of lobbying in this model and has an effect size second only to the number of lobbying staff. The coefficient for reliance on dues, predicted to be positive, is not statistically significant.

In the analysis for the broad-based lobbying index based on 30 tactics and the lobbying index based on eight tactics regulated by tax law, we can begin to see the effect of the issue-based factors. Government attention, public salience, party conflict, opposition interest groups, and institutional allies all increase the level of lobbying in both models, as predicted. Tax status as a 501(c)(3) organization has a strong negative effect on lobbying levels in both models, although the impact is somewhat greater for the regulated lobbying index, as predicted, as these are the lobbying tactics that are mentioned in the tax code as limited or prohibited for 501(c)(3) organizations. As for other organizationally based variables, the size of lobbying staff and reliance on dues are, as expected, positively related to lobbying. Most importantly, the receipt of federal funds is not a significant factor in determining how frequently an organization lobbies, and even the sign of the coefficient is negative.

The analysis of agency lobbying, however, shows a different pattern, just as it did in the means analysis. Public salience and party conflict were not contributing factors to how much an organization lobbied agencies. This, of course, makes sense if we consider the more technical nature of bureaucratic decision making and the fact that agency officials are not elected. The presence of interest-group conflict also was not related to levels of lobbying, paralleling the findings of McKay and Yackee (2004). The coefficient for the 501(c)(3) tax status variable no longer is statistically significant. Lobbying of agencies is not limited under the Internal Revenue Code and thus there is no reason to expect 501(c)(3) status to depress lobbying in this category.21 And finally, the pattern that we saw in the difference of means tests holds here as well: Organizations that receive federal funds are more likely to engage in agency lobbying than organizations that do not receive such funds, although substantively this is the least important of the statistically significant variables.

Funding and Faction

The debate between those who would restrict lobbying by nonprofits and those who would free it comes down to a question of whether the government is funding faction or buying silence. There is no support in this analysis for the idea that federally funded organizations lobby significantly more than similar nonfederally funded organizations, and it is tax status, not the lack of direct federal subsidies, that seems to do the best job of quieting advocacy efforts. This holds true even when controlling for common influences on lobbying behavior, such as the size of the government relations staff, the salience of the issue, and the level of political conflict.

The finding that government grants and contracts are not a prime motivator of interest-group lobbying may be surprising to some legislators, but not to most scholars of interest groups. Even those taking a rent-seeking perspective argue that direct payments are not the most likely candidates for attracting rent-seeking interest groups. Parker (1996) argues that contracts and direct payments are less likely to be sought by organized interests than are rules and regulations that limit competition and entry to the market. "The reason is simple: unless the list of beneficiaries can be limited by some acceptable devise, the amount of the direct subsidy and industry or group obtains will be quickly dissipated among a growing number of rivals" (17). Government attention of any kind certainly attracts groups to an issue or issue area, but grants are less likely than legislation and regulation to generate this interest-group attention.

To the extent that federal funding does increase lobbying, it does so in agencybased lobbying, not in the legislative arena where most proposed restrictions on lobbying have been aimed. The area of greatest concern to most critics of nonprofit lobbying has been legislative lobbying because of fears that federally supported organizations may use their increased budgetary strength to intensify rent-seeking efforts with those who hold the purse strings. The analysis here indicates that levels of such legislative lobbying are no higher among recipients of federal funds than among nonrecipients.

Should organizations that receive federal funds be lobbying at all? Clearly it is reasonable for taxpayers to wish to avoid funding advocacy campaigns, especially advocacy campaigns that they disagree with. Existing federal law, however, already prohibits the use of federal funds for lobbying, and despite the theoretical fungibility of organizational budgets, there is no indication based on lobbying levels that groups are subsidizing their legislative lobbying with federal funds. In fact, the opposite seems to be true, at least in the case of organizations with charitable tax status, which lobby significantly less than other types of groups.

The charities, other tax-exempt organizations, and businesses discussed in this article are among the most active lobbying organizations in the country. They all are known to have engaged in some type of policy-related activity in the past and maintain their own offices in or near Washington, DC. Lobbying among all recipients of federal grants and contracts nationwide is much lower than the levels shown here. Berry and Arons (2003) have shown that when politically active and politically inactive 501(c)(3) organizations are compared, it is the politically inactive that are more motivated to seek federal funds. This analysis adds to their findings, showing that among politically active organizations, federal funding does not correlate with increased lobbying.

Government affects the lives and fortunes of interest groups and individuals in myriad ways. Direct subsidies are only a small part of what government does. Interest groups lobby to try to affect the decisions that in turn affect them. Sometimes these efforts may be related to government grants or contracts, but more often it will be a regulation or legislation that affects how businesses operate, how the rights of some constituencies are interpreted, or what behaviors are legal and illegal. If we consider the many areas in which government is involved and the many ways that interests are affected, there is no logical reason to expect only direct recipients of government funds to be highly motivated to lobby or for their lobbying to exceed that of others.