How many lobbyists are in Washington? Shadow lobbying and the gray market for policy advocacy

Transcription

1 Int Groups Adv (2017) 6: DOI /s y ORIGINAL ARTICLE How many lobbyists are in Washington? Shadow lobbying and the gray market for policy advocacy Herschel F. Thomas 1 Timothy M. LaPira 2 Published online: 18 September 2017 Ó Macmillan Publishers Ltd 2017 Abstract How many lobbyists are in Washington, and how common is it for them to have worked in the federal government? We assume that high-profile cases like former Senator Tom Daschle the namesake of the so-called Daschle loophole to the Lobbying Disclosure Act (LDA) in the USA are not isolated. In this article, we systematically account for lobbying and policy advocacy in as large an empirical scope as possible to uncover the presence of shadow lobbyists. Using a new data set of professional biographies of both registered lobbyists and unregistered policy advocates, we estimate that there are an equal number of paid professionals in a gray market for lobbying services. We also find that registered lobbyists are more likely to have previously worked in government and are more likely to specialize in legislative advocacy. Since policymaking at the American national level has increasingly shifted to federal agencies and to the states, our results indicate that the LDA and similar lobbying regulations may be becoming increasingly obsolete. The evidence we present indicates a growing divide between transparency laws and recent changes in the marketplace for policy advocacy. Keywords Lobbying Disclosure Act Lobbying Policy advocacy Shadow lobbyists Revolving door Political reform & Herschel F. Thomas herschel.thomas@uta.edu 1 2 University of Texas at Arlington, 601 S. Nedderman Dr., Box 19539, Arlington, TX 76019, USA James Madison University, 91 E. Grace St., MSC 7705, Harrisonburg, VA 22807, USA

2 200 H. F. Thomas, T. M. LaPira Introduction How many lobbyists have previously worked in government? This seemingly simple empirical question about the size and scope of the Washington s revolving door is deceptively challenging to answer. This is due to the lack of knowledge about how many lobbyists are actually active in the nation s capital. Anecdotal evidence suggests that many public policy consultants, strategists and advocates lobbyists by another name and their clients do not disclose because they are not technically required to report their activities under the 1995 Lobbying Disclosure Act (LDA) and the 2007 Honest Leadership and Open Government Act (HLOGA) that revised it. Many of these consultants previously held positions of significant power in the federal government and, presumably, use their professional relationships and political process knowledge to help their clients pursue policy goals. Government transparency advocates and LDA compliance officers colloquially refer to this problem as the Daschle loophole after former Senate Majority Leader Tom Daschle (D-SD), who now earns a living as chairman of the Daschle Group, a subsidiary of major Washington law and lobbying firm Baker, Donelson et al. 1 Until 2016, Daschle never registered as a lobbyist, presumably because he interpreted his work as not fitting the strict statutory definition of lobbying. Yet, most reasonable people would consider his value to be rooted in his extensive policy knowledge, as well as his contacts on Capitol Hill and throughout the federal government. Moreover, it is safe to assume that Daschle is not alone. Former Speaker of the House Newt Gingrich (R-GA) earned millions of dollars offering strategic historical advice to Freddie Mac after he left Congress. His doing so became a contentious topic in the 2012 Republican presidential nomination contest. This research is motivated by the assumption that the Daschle phenomenon extends well beyond these anecdotes. Media attention to well-known politicians like Daschle and Gingrich may be misleading, since we know the overwhelming majority of lobbyists are not former elected officials. While LaPira (2015) normatively assesses shadow lobbying as it threatens the democratic legitimacy of pluralism among organized interests (23), to date there have been no systematic empirical estimates of the true lobbyist population in Washington. In this article, we go beyond the narrow, statutory definition of a lobbyist to include a broad range of policy advocates. Doing so allows us to provide a systematic estimate of the full size and scope of policy advocacy in Washington. Using a new data set of professional biographies, we determine that the size of the gray market for policy advocacy is likely equal to that of lobbying disclosed to the public. Our empirical approach also allows us to compare the characteristics of registered lobbyists versus unregistered policy advocates. We show that those policy advocates who are not registered as lobbyists are less likely to have government experience as staff in Congress and tend to be political process specialists. We interpret these findings as evidence of a legislative lobbying bias in the existing disclosure framework. We conclude that a recent trend in the decline of publicly disclosed lobbying is an artifact of stricter revolving door cooling off periods in the 1 See Accessed May 2017.

3 How many lobbyists are in Washington? Shadow lobbying and law and a shift in policymaking attention away from the legislature. Lobbying has not likely declined; it has simply moved into the shadows. Who is a lobbyist? To achieve our straightforward empirical objective, we are compelled to define lobbyist more broadly than the LDA statutory definition. The LDA defines a lobbyist as a person who (a) contacts more than one government official (b) on behalf of a paying client or employer for (c) 20% or more of their time in a given quarter. In an in-person interview, a prominent Washington political law attorney explained the law this way: What the [LDA] has done is help to define [lobbying] and to place a meaningful threshold that would trigger the burdens of filing reports and registering with the government. So the definition basically says a lobbyist is someone who is compensated to do two things: to have two or more communications with certain people in government and to spend more than 20 percent of his or her time engaged in what the statute describes in further detail as lobbying activity. [ ] [The 20 percent threshold is] determined in two ways. One is in any occupation or industry that has time keeping, and you would see this particularly in government contracting industries, employees of the company have to keep track of their time, describe whether or not they re lobbying or not. This is a corollary of the rules on government contracting where the proceeds, the federal government payments for contracts, cannot be used for lobbying. The more typical tracking method for organizational employers of lobbyists is a good faith period of guesstimates of time spent on that particular activity. So in practice, the law requires registered lobbyists to maintain documentation in case the Government Accountability Office (GAO) an administrative support agency in the legislative branch randomly selects them for annual audits (Luneberg and Susman 2009). But the law allows lobbyists themselves to determine, or guesstimate, whether they are lobbyists. As a result, the statutory definition alone does not offer a good operational definition to guide our observation of the lobbyist population. In this article, however, we follow LaPira (2015) to broadly operationalize shadow lobbyists which we empirically identify as unregistered policy advocates as those who are: paid to challenge or defend the policy status quo, to subsidize policymakers with information, or to closely monitor intricate policy and political developments that are not readily available to the public or those who offer expertise, knowledge, and access in support of these activities yet who do not register as lobbyists (LaPira 2015, 229).

4 202 H. F. Thomas, T. M. LaPira In other words, we consider registered lobbyists to simply fall in a specific category of the general policy advocates population who happen to disclose their activities. We draw a random sample of 367 registered lobbyists, unregistered policy advocates and other professionals from the commercial Lobbyists.info directory of 31,976 political and policy actors. Using this sample, we create an original data set of professional biographies that describes current lobbying and policy advocacy work and previous government employment. Doing so provides sound evidence to generalize about shadow lobbying and the market for policy advocacy in Washington. Why do we care about unregistered policy advocates? From the perspective of government transparency, the public clearly has an interest in knowing who gets hired to influence the public policies that affect them and their tax dollars. While those lobbyists who take advantage of the Daschle loophole are not necessarily in violation of the LDA, it is likely that many are not following the spirit of lobbying transparency. 2 Indeed, uncovering instances of unregistered lobbyists has been the subject of a number of investigative journalist and transparency advocate reports (Revolving Door Working Group 2005; Public Citizen 2005; Project on Government Oversight 2013). Adjustments to the law appear to be increasing unreported lobbying activities, effectively driving policy influence activities underground (Center for Responsive Politics 2013). According to the Center for Responsive Politics data, after increasing an average of 4% annually between 1998 and 2007, the number of registered lobbyists has declined an average of 3% annually between 2008 and These government watchdog and journalistic reports consistently find that LDA loopholes are extremely broad, implying that the regulatory regime governing lobbying disclosure falls well short of the transparency and ethical ideal. The so-called revolving door the movement of individuals from positions in government to the private sector, and especially as lobbyists is the raison d etre for requiring individual lobbyists to disclose their activities. Otherwise, lobbying transparency would be sufficient if only the organized interests seeking to monitor or influence the policy process reported what they were doing in the public sphere. As a result, any changes to norms of lobbying transparency may amplify concerns about the revolving door. Interest group scholarship has long recognized the need to uncover the causes and consequences of the revolving door. Beginning with Gormley (1979), interest group scholars have tested the revolving door hypothesis in a variety of government venues (Quirk 1981; McGuire 2000), over time (Cohen 1986) and in selected policy domains (Salisbury et al. 1989; Hansen 1991). More recently, Baumgartner et al. 2 The purpose of this research is not to identify individual instances of non-compliance, but rather to evaluate the scope of lobbying disclosure avoidance based on an empirically identifiable population. 3 Calculated by authors from data reported at Accessed May Actual data may have been slightly revised after we calculated these results.

5 How many lobbyists are in Washington? Shadow lobbying and 203 (2009) found that hiring covered official lobbyists to be the only extragovernmental variable that systematically predicted policy success. And several studies have uncovered data on individual lobbyists from public biographies and government payroll records to see what effect their work has had on a variety of outcomes such as revenue and issue advocacy activity (Vidal et al. 2012; Bertrand et al. 2014). And, LaPira and Thomas (2017) find that revolving door lobbyists have worked mostly in Congress, tend to work as hired-gun lobbyists, represent a wide variety of economic interests and generate more lobbying revenue based on the depth of their experience on Capitol Hill. The incidence of revolving door lobbying has increased over time. Parker (2008) shows that, since 2000, 53% of departing members of Congress have moved to K Street. Drutman and Furnas (2014) document an explosion in the number of active lobbyists reporting any previous government experience since Lazarus et al. (2016) offer further evidence of a steady and remarkable increase in lobbying among former lawmakers. Finally, Cain and Drutman (2014) reveal how new limitations on legislative staff changed the incentive structure to move on to postgovernment employment as lobbyists. Though the literature has begun to empirically examine the revolving door phenomenon more rigorously, interest group scholars still lack both good theories and data on individual lobbyists likelihood of disclosing their activities. That is, we know that some lobbyists have gone through the revolving door and some have not, and that some lobbyists meet the strict LDA statutory definition of lobbyist and some do not. But we do not know how many fit into these categories, or why. The shortcomings of lobbying disclosure laws and the presence of lobbyists working in the shadows inhibit our ability to know who is hiring lobbyists with previous government experience, and so we have no way of knowing which interests benefit from their advocacy and influence. Methods and data In this article, we simply seek to account for the population of policy advocates which includes both LDA-registered lobbyists and unregistered policy advocates, as well as other professionals such as executive branch legislative liaisons to provide some empirical estimate of those working on policy advocacy in Washington. Our undertaking is both Herculean and Sisyphean, as the true population of policy advocates in Washington is in the tens or perhaps hundreds (Thurber 2011) of thousands. And their professional backgrounds are often impossible to fully reconstruct based on publicly available information alone. Nevertheless, we believe our sample of policy advocates represents a defensible, comprehensive and replicable estimate of the size of the policy advocacy market in Washington. We obtained the full database of all 31,976 individuals listed in the Lobbyists.info directory on September 14, 2012, to draw our sample. This database is the continuously updated, online version of what used to be annually published in print as Washington Representatives. The volume has served as the sampling frame for several large-scale interest group studies and has been the main source for

6 204 H. F. Thomas, T. M. LaPira identifying organizations in Washington for decades, even preceding the Lobbying Disclosure Act of 1995 (see Schlozman and Tierney 1986; Schlozman et al. 2012). Among interest group scholars, it is viewed as the nearest equivalent to an exhaustive listing of the organizations active in national politics (Schlozman 1984, 1019) and also includes those individuals that organizations employ to represent their interests in Washington. 4 We then randomly ordered individuals within this data set and supervised student assistants as they conducted extensive searches for professional biographies using a variety of electronically available sources. Coders used proprietary search engines such as CQ Press s First Street, Leadership Directories, Legistorm Pro and the Lexis Nexis Martindale Hubbell directories, as well as publicly available sources such as the OpenSecrets.org Lobbying and Revolving Door databases, lobbying firm and organizational websites and the LinkedIn social network. From these sources, coders recorded each lobbyist s employment status, advocacy specialization and previous government employment, as well as partisan affiliation, policy substance expertise and demographic information. Given the time-intensive nature of our search procedure, we were able to collect this information for 367 living individuals (1.1%). Our dataset includes independently verified information about each individual s policy advocacy employment and previous government employment as it was recorded between September, 2012, and April, While this search process is as thorough as possible, some limitations are unavoidable. First, though the sampling frame of more than 30,000 individuals is by far the most complete source of political and policy professionals available on the market, we recognize that many actors who engage in policy advocacy activities may not be listed in Lobbyists.info directory for a variety of reasons. For instance, the chief executives of corporations and organizations who most certainly visit Washington and meet with policymakers in so-called fly-ins are not likely listed because they are primarily managers. Likewise, regulatory and appellate lawyers may not think of themselves as policy advocates, though we could easily include attorneys who prepare notice-and-comment letters and amicus briefs as interest group representatives according to our operational definition (Yackee 2006; Box- Steffensmeier and Christenson 2014; Haeder and Webb Yackee 2015). Similarly, public relations and grassroots lobbying consulting firms that do not primarily engage in direct contact lobbying, but instead use social media, and telephone 4 In September, 2012, the authors met with the publisher of the directory information, Columbia Books, and learned that there are three main avenues in which individuals are entered into the database. First, organizations such as lobbying firms, associations and companies voluntarily provide information (including information about their staff) for inclusion in the directory either as an advertising mechanism or upon request from the publisher. Such requests are routinely sent by the publisher to update existing entries or to new organizations that have purchased access. Second, the publisher retrieves information from public disclosure documents required by the Lobbying Disclosure Act of And, third, the publisher actively searches for those working in political advocacy in Washington. While Schlozman et al. (2012) identify some error as a function of the type of activity an organization engages in as well as temporal variation, they note that the publisher views the database as a snapshot of those engaged in Washington politics.

7 How many lobbyists are in Washington? Shadow lobbying and 205 contact lists and other outside lobbying tactics therefore avoiding both LDA and common definitions of lobbyists are likely excluded. Accordingly, the population estimates and policy advocacy activities that we report are, if anything, a conservative estimation of the full population of lobbyists and unregistered policy advocates in Washington. Registered lobbyists are a matter of public record, which can be verified. Annual Government Accountability Office (GAO) audits show that the overwhelming majority of filers are able to supply supporting documentation for data in their reports, so we trust the data are as internally reliable and valid as can be reasonably expected (GAO 2016). But those advocates who choose not to register under the LDA, and who either have not been identified by Columbia Books or who choose not to be listed in the Lobbyists.info directory, remain an empirical unknown. Our discussion proceeds as if our sampling frame is a genuine census of lobbyists and policy advocates, with the caveat that it is very likely an underestimation of unregistered policy advocates and supporting personnel. There are most certainly other highly paid agents in Washington who sell their political process and policy knowledge. For example, Senator Charles Grassley (R- IA) and Representative Louise Slaughter (D-NY) introduced legislation in the 114th Congress to require so-called political intelligence consultants to register like lobbyists do. This newly emerging niche industry collects information on possible or actual statutory or regulatory actions that may impact companies market capitalization. They sell this information to institutional investors such as hedge fund managers, who then use it to valuate companies and strategically take long or short positions on equity contracts that may be affected by forthcoming government activity. Additionally, as we show, many people listed in Lobbyists.info are primarily engaged in political and campaign consulting activities such as political action committee (PAC) managers whom we do not operationalize as policy advocates. We concede that this distinction may be becoming less, not more, clear in the aftermath of Citizens United. So-called social welfare organizations like Republican Party-affiliated Crossroads Grassroots Policy Solutions (GPS) and Democratic Party-allied Organizing for Action (OFA) blur the line between policy advocacy and campaign political activity. We make what we believe to be the conservatively cautious choice to not count those engaged primarily in electoral activities as policy advocates. The gray market for policy advocacy Our novel sample enables us to not only describe the proportion of unregistered advocates engaged in the policy process for the first time, but also to make inferences about the type of professionals present in the full population. Table 1 reports the number and percentage of sampled actors whom we have verified as employed as political and policy professionals in 2012 and Only 13 (3.5%) of the 367 sampled individuals are currently employed by the federal government as legislative liaisons or similar job titles. These executive branch government

8 206 H. F. Thomas, T. M. LaPira Table 1 Population estimates of Lobbyists.info sample, 2012 Sample Population estimates Count Percentage Low Mean High Current federal employee Electoral politics professional Unregistered policy advocate ,079.0 Inactive LDA lobbyist Active LDA lobbyist , , ,477.0 Total , , ,113.9 n = 367 individuals who have been verified as employed as political and policy professionals in 2012 and The sample proportion is 1.1% of 31,976 individuals listed in Lobbyists.info directory as of September 14, The margin of error =±5.1% at the 95% confidence level. The actual census of LDA lobbyists in 2012 was 12,389 (Center for Responsive Politics 2013) employees are not lobbyists in the private sector. We exclude them from subsequent analyses. The second category in Table 1 includes 69 (18.8%) people involved primarily in electoral politics. We likewise exclude them from subsequent descriptions of the policy advocacy population. The remaining three categories include people currently employed as policy advocates. Those who have never registered to lobby but who are primarily engaged in policy consulting, strategy and other advocacy-related activities including the Gingrich s and Daschle s in Washington represent more than one-quarter of our sample with 97 individuals. The next category includes 52 individuals who are currently inactive federally registered lobbyists. According to the Center for Responsive Politics Lobbying Database, inactive lobbyists did not file LDA reports with the Senate in the 2012 calendar year, though there is a record of them filing reports at some point between 1998 and Finally, the most common type of advocate is that of federally registered lobbyists who reported activity as required by the LDA in Formally registered lobbyists represent more than one-third of our sample. Though our random sample of actors is just 1.1% of the 31,976 people listed in Lobbyists.info, we can make a simple empirical estimate about how many of each type of actor exists in the population. To do this, we begin with the sample percentage of individuals within each included category and then calculate the corresponding number of equivalent persons in our sampling frame of 31,967. We also subtract and add 5.1% from this value to generate the low and high bounds of the estimate and account for sampling error. As shown in Table 1, we calculate that there are between 6824 and 10,079 unregistered policy advocates. In other words, we estimate that there are 8451 under the radar advocates on K Street (margin of error =±5.1%). That is, if we were to repeatedly draw samples of 367 individuals from the Lobbyists.info list, we can say with confidence that 95% of the time the number of unregistered advocates would 5 We do not distinguish between those who took the formal steps with the Secretary of the Senate to deregister and those who deactivated by simply ceasing to file quarterly lobbying disclosure reports.

9 How many lobbyists are in Washington? Shadow lobbying and 207 Table 2 Political economy of policy advocacy in Washington, 2012 Population Expenditures Estimate Percentage Unreported policy advocacy 12, $3,485,628,942 Reported LDA lobbying 11, $3,181,513,666 Total 24, $6,667,142,607 These estimates are based on n = 285 professional policy advocates. Current federal employees and election-oriented political professionals are excluded. The unreported policy advocacy category combines both unregistered advocates and inactive LDA advocates. Unweighted lobbying expenditures per lobbyist is based on the census of LDA lobbyists reporting activity in 2012 and the sum of LDA-reported expenditures, $3,326,389, ,389 lobbyists = $268,495 (Center for Responsive Politics 2013) fall within the range of roughly 6800 and 10,000 people, and most frequently would be about 8500 people. We also calculate that there are approximately 11,849 active federally registered lobbyists in the population, with a 95% confidence interval of 10,221 13,477. According to the Center for Responsive Politics (2013), the actual census of LDA lobbyists in 2013 was 12,389, which is well within our population estimate. In Table 2, we collapse the unregistered advocate and inactive lobbyist categories to get a combined measure of unregistered policy advocates in When we restrict the sample to unregistered advocates alone, a nominal majority (52%) do not disclose activities according to LDA guidelines. That is, in 2012, most policy advocacy in Washington was in the shadows. These estimates are subject to two caveats. First, low-end estimates of the unregistered policy advocates could put this estimate at roughly 45% of the population relative to the mean of registered lobbyists, though it could be as high as 58%. And second, as mentioned, these individuals are likely underestimated by an unknown factor. Estimates based on the Lobbyists.info directory sample fall well short of Thurber s (2011) estimate of 100,000 advocates and support personnel who spend $9 billion per year. If we make the admittedly generous assumptions that (a) our estimates are indeed accurate, and that (b) unregistered advocates will on average generate the same amount of expenditures as LDA-registered advocates, then we can predict the amount of money spent on both reported and unreported policy advocacy. We know that roughly $3.3 billion was spent on LDA lobbying in If clients spent the same amount per policy advocate as LDA clients report spending on LDA lobbyists or an unweighted mean of $268,495 then the total amount spent on policy advocacy in 2012 would be about $6.7 billion at the 95% confidence interval. According to Bureau of Economic Analysis estimates, the Washington metropolitan area s regional gross domestic product was $435 billion in 2012, making policy advocacy at least 1.5% of the local economy (Bureau of Economic 6 The means of calculating LDA expenditures are subject to controversy as well. Clients choose one of three ways to account for expenditures, so they may under- or overestimate spending. We consider this problem to be a matter of measurement error that is distributed randomly across individual lobbyists, since this accounting decision is made by the organization-registrant that lists them in disclosure reports.

10 208 H. F. Thomas, T. M. LaPira Analysis 2016). That is, the size of the marketplace for policy advocacy in Washington alone excluding similar advocacy at the state and local level or at the supranational institution level is a roughly $7 billion industry, half of which is unregulated and unaccounted for. But, consider these findings in context of a National Retail Federation report: American consumers spent $18 billion on Valentine s Day consumption in 2012 (National Retail Foundation 2012). So, we are cautious to draw sweeping conclusions about what the raw amount of policy advocacy spending alone means for political influence and policy outcomes (Ansolabehere et al. 2003). The true size and shape of Washington s revolving door These findings provide a general overview of the population of unregistered policy advocates and the gray market for their advocacy activity. Here we describe the government experience of each of those individuals in our sample to determine whether, and to what extent, unregistered policy advocates spin through the revolving door. In Table 3, we report the number of policy advocates previously held government jobs (if any). Though a majority of advocates are conventional in the sense that we found no professional history of federal employment (164, 57.5%), more than 2 in 5 actors in our sample held at least one government position. Of the 285 sampled advocates, about 15% held two or more federal government jobs. In sum, our search process reveals that 121 (42.5%) policy advocates in our sample went through the so-called revolving door. If we distinguish each policy advocate by LDA status and whether or not they previously worked for the federal government, then we can examine how the Daschle Loophole contributes to the size and shape of the revolving door. Figure 1 plots the numbers of unregistered policy advocates and registered LDA lobbyists by revolving door status. The differences between the categories are easily visible. Those who go through the revolving door are more likely to have registered to lobby in 2012 (v 2 = 28.5, p \ 0.001). Nearly one-quarter of unregistered policy advocates and approximately 60% of LDA-registered lobbyists hail from the halls of government. Though the majority of registered lobbyists previously worked inside government, a non-trivial portion of unregistered policy advocates has also gone through the revolving door. So, though the Daschle exemption is real, we can infer that previous government Table 3 Number of policy advocates previous federal government jobs n = 285 policy advocates. Current federal employees and election-oriented political professionals are excluded Count Percentage None One Two Three or more Total

11 How many lobbyists are in Washington? Shadow lobbying and 209 Fig. 1 Previous government experience and type of policy advocate. Note n = 285 professional policy advocates. Pearson v 2 (1) = 28.5, p \ employees are actually more likely to disclose their lobbying than their counterparts without government experience. Of the subsample of 135 policy advocates that we verified as having revolved from federal employment to policy advocacy, where did they previously work in government? We aggregate revolving door advocates by place of employment. The unit of analysis is the advocate-job, as individual actors may have held jobs in more than one location. Exact positions range from the mid-level agency bureaucrats and legislative assistants to members of Congress, chiefs of staff and cabinet officers. The personal, committee, and leadership offices on Capitol Hill account for 94 advocate jobs, agencies in the federal bureaucracy account for 30, and just 13 were traced to the White House. Positions on Capitol Hill make up the vast majority of those held by advocates in our sample. Not only is Congress the keystone of the Washington establishment, it is the stepping stone to K Street. Figure 2 compares the percentages of registered lobbyists and unregistered advocates by institution. Though we find that 37% of previous Congressional employees are presently acting as unregistered policy advocates, former White House and Bureaucracy employees are much more likely to be unregistered advocates than lobbyists registered under the LDA. From the White House, 66.7% of revolving door policy advocates are unregistered, though the extremely small sample of these former White House personnel calls for caution. Nearly three quarters (74.7%) of those who held positions in the federal bureaucracy are unregistered policy advocates. Moreover, the institutional differences in unregistered advocates are statistically significant for both the Congress and Bureaucracy, though the correlation coefficients are in the opposite direction. Finally, we also investigate the professional characteristics of the policy advocates in our sample. While they may or may not have the same revolving door experiences, do unregistered policy advocates differ systematically in policy specialization? Figure 3 illustrates the percentage of both registered lobbyists and

12 210 H. F. Thomas, T. M. LaPira Fig. 2 Policy advocates and the revolving door. Note n = 135 revolving door policy advocates. Unit of analysis is the advocate-job. Two advocates worked in more than one part of government. Difference between proportion of registered advocates and unregistered lobbyists: Congress, (z =-3.62*); Bureaucracy, (z = 3.49*); and White House, (z = 0.37) Fig. 3 Policy advocacy specialization. Note n = 285 professional policy advocates. Pearson v 2 (3) = 29.4, p \ unreported advocates across four specialization types: legislative affairs, regulatory affairs, political process experts and a category for unknown or other specializations such as government contracts experts or policy researchers. 7 Legislative and regulatory affairs specialists are self-explanatory. Political process specialists are those policy advocates who explicitly described themselves as knowing the innerworkings of the federal government, but did not identify particular institutional expertise. 7 There was too little variation in the current sample in these specialization categories, so we collapsed them into a single, catch-all category.

13 How many lobbyists are in Washington? Shadow lobbying and 211 Of the 285 professional policy advocates in our sample, 174 specialize in legislative affairs issues, further evidence of the legislative-centric nature of lobbying as defined by both the LDA and Lobbyists.Info. LDA-registered lobbyists account for 97 (55.8%) of those individuals, with 77 (44.3%) of them as unregistered advocates. The breakdown across registration status is nearly identical for regulatory affairs specialists: 32 (52.5%) are registered lobbyists and 29 (47.5%) are unregistered policy advocates. Though these results suggest that advocates are similar in their legislative and regulatory expertise, we find that unregistered policy advocates are much more likely to have specialized knowledge of the political process of policymaking. While just six (22.2%) political process specialists are LDA-registered lobbyists, 21 (77.8%) are active unregistered policy advocates. Given the under the radar nature of non-lda-reported policy advocacy it is unsurprising, though telling, that for those advocates who we were unable to determine a particular specialty, 22 (95.7%) were unregistered policy advocates. All told, the archetypal registered lobbyist in Washington has gone through the revolving door, worked in only one government job most likely on Capitol Hill and specializes in legislative lobbying. Alternatively, the typical unregistered policy advocate is less likely to have gone through the revolving door and is more likely to be a general political process expert rather than an institutional specialist. Tom Daschle and Newt Gingrich are the exceptions, not the norm, although their greater influence among former colleagues may make them exceptional indeed, even if they do not trigger the strict statutory definition of lobbyist. Methodological limitations and inferences about the market for advocacy Our estimation of the policy advocacy population and our description of revolving door patterns among policy advocates are novel, yet incomplete. We report findings from a relatively small sample due to our extremely time-intensive data collection process. Along with increasing sample size, future iterations of research along these lines ought to report a variety of descriptive variables such as seniority in government job, party affiliation of previous employer, policy expertise in and out of government and demographic variables that may correlate with theoretically relevant policy advocacy outcomes. Moreover, our methods are limited by information available in the public domain. Though we are confident that our process exhaustively located professional biographical data about each of the cases in our sample, our approach cannot fully explain why a policy advocate chooses to formally report lobbying activity or not. Moreover, unlike our related work that links professional histories with subsequent lobbying activities (LaPira and Thomas 2017), the lack of information about unregistered policy advocates clients and issue activities prevents us from making causal inferences about the revolving door, interest representation and influence with this sample. We simply do not know who shadow lobbyists represent, what policy issues they seek to influence and monitor and whether their activities are the

14 212 H. F. Thomas, T. M. LaPira same or different than those lobbyists who do disclose under the law. These data and transparency problems are precisely what motivates this research. Answers to these questions ought to use alternative research methods such as qualitative in-person interviews and questionnaires to expressly learn more about non-lobbying policy advocacy. With increased empirical evidence revealing the problems associated with political gray markets, policymakers may begin to seriously reconsider lobbying disclosure laws. The public should not be forced to rely on hypothetical assumptions about how many highly paid professionals seek to influence their government on behalf of organized interests. Practical implications for lobbying transparency The findings we report here show that a great deal of policy advocacy exists out of view from public scrutiny. If recent trends continue, then unreported policy advocacy will certainly be more common than publicly disclosed lobbying activities, as we assume it already is. Though our evidence of unreported advocacy only represents a snapshot in time and our forecasts are based on admittedly simple hypothetical assumptions, the clear trend in LDA deregistration or deactivation can be attributed to the 2007 lobbying disclosure reforms and the Obama administration s exclusion of lobbyists from executive branch appointments, among other causes (LaPira 2015). The policy-induced loopholes in lobbying transparency produced during the Obama administration may only be exacerbated by underspecified efforts by the Trump administration to drain the swamp. Post-government cooling off periods seem only to drive lobbyists into the shadows. There is little reason that further limitations will improve what we document here. The actual and future incidents of shadow lobbying may be worse than our assumptions, and empirical observations lead us to conclude whether prospective registered lobbyists are subjected to greater professional disincentives to disclose their advocacy activities. If transparency is the genuine policy objective of lobbying disclosure laws, then the LDA and HLOGA are seriously flawed. Congress ought to reconsider statutory definitions of lobbying and lobbyists to account for a variety of policy advocacy activities and actors that are effectively the same thing. Any efforts by the legislative or executive branches of government that institute new restrictions without first greatly expanding these definitions are not only futile, they make the lobbying disclosure regime even less transparent. This implication is a critical point for political reform efforts outside the USA, as well as the state and local level within the USA, which often assume the LDA disclosure regime to be the gold standard. In addition to reconsidering who and what gets reported, Congress should also take on the burden of public disclosure. Lobbying and policy advocacy is a two-way street. The lobbying and government relations transparency regime have always required lobbyists to self-report their activities; it has never asked members of Congress, White House personnel, policymaking bureaucrats and other relevant government employees to disclose which private organized interests they interact

15 How many lobbyists are in Washington? Shadow lobbying and 213 with. Lobbyists and policy advocates who have a contractual relationship with their clients do not have a responsibility to the public, though the law compels them to shoulder the transparency burden. But in a democracy, elected representatives and appointed officeholders do have a direct and immediate responsibility to their constituents and legislative overseers. We recognize that reelection-motivated members of Congress have little incentive to provide potential challengers with fodder for attacks, but that is a separate issue from thinking through normatively ideal political reforms. The growth in shadow lobbying is an unarguable setback for open government and transparency proponents. And it is a significant impediment for basic research on interest group politics in the USA as well. The 1995 lobbying law has helped propel two decades of interest group scholarship that has taken seriously Baumgartner and Leech s (1998) call for theoretically rich and empirically rigorous scholarship of interest representation. If lobbying disclosure becomes increasingly unreliable and externally invalid based on the unilateral, subjective determination of some organizations and their agents to not report their activities, interest group scholars ability to make inferences about representation, mobilization and influence over time will suffer. The magnitude of unreported policy advocacy that we report here comes at a time when interest group scholars have begun to take advantage of publicly available data to increase our understanding of interest representation, political influence and money in politics more generally. As we show in this article, though, alternative data sources both for our sampling frame and the publicly available raw data sources promise to improve our knowledge of the size, scope and potential biases of the interest group system. We hope this research serves as an impetus for future efforts to better estimate the marketplace for policy influence and the true size of the advocacy community, both in and out of Washington. References Ansolabehere, S., J.M. de Figueiredo, and J.M. Snyder Jr Why is there so little money in U.S. politics? Journal of Economic Perspectives 17 (1): Baumgartner, F.R., and B.L. Leech Basic interests: The importance of groups in politics and in political science. Princeton, NJ: Princeton University Press. Baumgartner, F.R., J.M. Berry, M. Hojnacki, D.C. Kimball, and B.L. Leech Lobbying and policy change: Who wins, who loses, and why. Chicago: University of Chicago Press. Bertrand, M., M. Bombardini, and F. Trebbi Is it whom you know or what you know? an empirical assessment of the lobbying process. American Economic Review 104 (12): Box-Steffensmeier, J.M., and D.P. Christenson The evolution and formation of amicus curiae networks. Social Networks 36 (January): Bureau of Economic Analysis Regional data: GDP & personal income. itable.cfm?reqid=70&step=1&isuri=1&acrdn=2#reqid=70&step=1&isuri=1&7003=900&7004= naics&7005=-1&7001=2900&7002=2&7090=70. Retrieved May Cain, B.E., and L. Drutman Congressional staff and the revolving door: The impact of regulatory change. Election Law Journal: Rules, Politics & Policy 13 (1): Center for Responsive Politics Lobbyists 2012: Out of the game or under the radar?. Washington, DC: Center for Responsive Politics.

16 214 H. F. Thomas, T. M. LaPira Cohen, J.E The dynamics of the revolving door on the FCC. American Journal of Political Science 30 (4): Drutman, L., and A. Furnas Revolvers dollars. Sunlight Foundation. com/blog/series/revolvers-dollars/, Retrieved May Government Accountability Office Observations on lobbyists compliance with disclosure requirements. Retrieved May Gormley, W.T A test of the revolving door hypothesis at the FCC. American Journal of Political Science 23 (4): Haeder, S.F., and S. Webb Yackee Influence and the administrative process: Lobbying the U.S. President s Office of Management and Budget. American Political Science Review 109 (3): Hansen, J.M Gaining access: Congress and the farm lobby, Chicago: University of Chicago Press. LaPira, T. M Lobbying in the shadows: How private interests hide from public scrutiny and why that matters. In Interest group politics, eds. A. J. Cigler, B. A. Loomis and A. J. Nownes, 9th ed., LaPira, T.M., and H.F. Thomas Revolving door lobbying: Public service, private influence, and the unequal representation of interests. Lawrence, KS: University Press of Kansas. Lazarus, J., A. McKay, and L. Herbel Who walks through the revolving door? Examining the lobbying activity of former members of congress. Interest Groups & Advocacy 5 (1): Luneberg, W.V., and T.M. Susman The lobbying manual: A complete guide to federal lobbying law and practice, 4th ed. Chicago: American Bar Association. McGuire, K.T Lobbyists, revolving doors, and the us supreme court. Journal of Law and Politics 16: National Retail Federation Americans to pull out all the stops this valentines day. media/press-releases/americans-pull-out-all-the-stops-this-valentine%c2%92s-day. Accessed May Parker, G.R Capitol investments: The marketability of political skills. Ann Arbor, MI: University of Michigan Press. Project on Government Oversight Dangerous liaisons: Revolving door at SEC creates risk of regulatory capture. Washington, DC: Project on Government Oversight. Public Citizen Congressional revolving doors: The journey from congress to k street, Congress Watch (July). Washington, DC: Public Citizen. Quirk, P Industry influence in federal regulatory agencies. Princeton, NJ: Princeton University Press. Revolving Door Working Group A matter of trust: How the revolving door undermines public confidence in government and what to do about it. Washington, DC: The Revolving Door Working Group. Salisbury, R.H., P. Johnson, J.P. Heinz, E.O. Laumann, and R.L. Nelson Who you know versus what you know: The uses of government experience for Washington lobbyists. American Journal of Political Science 33 (1): Schlozman, K.Lehman What accent the heavenly chorus? Political equality and the American pressure system. The Journal of Politics 46 (4): Schlozman, K.Lehman, and J.T. Tierney Organized interests and American democracy. New York: Harper & Row. Schlozman, K., S.Verba Lehman, and H.E. Brady The unheavenly chorus: Unequal political voice and the broken promise of American democracy. Princeton, NJ: Princeton University Press. Thurber, J Changing the way Washington works? Assessing President Obama s battle with lobbyists. Presidential Studies Quarterly 41 (2): Vidal, Jordi Blanes I., M. Draca, and C. Fons-Rosen Revolving door lobbyists. American Economic Review 102 (7): Yackee, S. Webb Sweet-talking the fourth branch: Assessing the influence of interest group comments on federal agency rulemaking. Journal of Public Administration Research and Theory 16: