Think Tank Soldiers in the Battle for Tax Reform

Abstract Contemporary scholars argue that high wealth donors in the U.S. influence political decision making through generous funding of nonprofit organizations like think tanks. In response to that potential influence, some endorse curbs on implicit subsidies that favor higher-income donors more than lower-income donors. To highlight the debate, this study selects a particular topic – tax policy – that generates highly partisan viewpoints and political agendas. The article first models predicted partisan operational differences, based on donors’ ideological differences. The study then explores the financial, staffing, and board resources of think tanks and associated advocacy organizations. The data were collected in the year immediately prior to the passage of the 2017 U.S. Tax Cuts and Jobs Act, providing a snapshot view of contrasting operations of left-, centrist, and right-leaning tax policy think tanks. Given the notably more generous resources utilized by right-leaning tax policy organizations, it is possible that donor wealth differences enabled right-leaning nonprofits to contribute their influence to get the historic tax reform package passed. However, the successful passage of the Tax Cuts and Jobs Act could have resulted as well from the more targeted messaging and narrative framing employed by right-leaning think tanks and advocacy organizations.


Introduction
In theory, great wealth does not directly lead to great influence in a democracy. A high net worth individual might gain more access to elected representatives, but those representatives may be voted out of office if they consistently ignore the wishes of the voting majority. Recent studies question these assumptions. Page, Bartels, and Seawright (2013) show marked differences between the policy preferences of the general public versus the wealthy, with top 1 % individuals favoring lower taxation on corporations and high-income individuals, plus even sharper preferences for tax reduction among the top 0.1 %. Cook, Page, and Moskowitz (2014) also show much higher political participation among the wealthy, compared to the general public. That is, very high net worth individuals have significantly more conservative tax policy preferences than the preferences shown by the general public, and are far more active politically (contributing, contacting legislators, etc.) than the non-wealthy. To illustrate, Confessore, Cohen, and Yourish (2015) report that approximately half of the contributed funds in the 2016 U.S. presidential campaign came from 158 families contributing over $250,000 each and 200 families contributing between $100,000 and $250,000 each. Some evidence supports the concern that concentrated wealth is effective in influencing policy change: Elected politicians appear to be supporting legislation that is more in sync with high-wealth individuals' preferences than with the broader public (Gilens and Page 2014).
Bypassing the constraints inherent in a representative democracy can be a laborious task, requiring persuasion of both elected officials and the public. Public intellectuals (or more recently, "thought leaders" -see Drezner 2017) can research issues and disseminate solutions for legislators to consider. Convincing the public is more than just an intellectual exercise, however. The nonprofit think tank and policy advocacy industry provides a leverage point for individuals to employ in persuading the general public and influencing policy formulation and implementation. This circuitous route to political influence would probably not be cause for alarm if the funding for such organizations were broad-based. As wealth distribution is increasingly concentrated (Atkinson 2015), however, scholars are beginning to criticize the U.S. regulatory framework that allows philanthropic funding of policy advocacy and even provides a greater tax incentive to do so, the wealthier the donor (Reich 2018). Callahan (2017b. p. 35) cautions against the "absolutist embrace of philanthropic freedom" and argues for limits to be placed on deductibility of policy-related charitable donations, given that Renee Irvin is the corresponding author.
© 2020 Irvin and Sokolowski, published by De Gruyter. This work is licensed under the Creative Commons Attribution 4.0 Public License. "(p)hilanthropists are buying ever more influence over public policy, wielding power that's not available to most citizens. Anyone worried about how economic inequality is translating into political inequality should worry about this flow of money." Callahan, it should be noted, co-founded Demos, a progressive think tank. Others argue that right-leaning philanthropy is necessary to counter the strength of the left's pull on academia and other agenda-setting in politics. Husock (2017) characterizes the think tank ideology battle as "a war in which the right is vastly outgunned financially" by the left.
In order to respond to calls to reform philanthropy to policy advocacy organizations, an important exercise is to first understand the mechanism of nonprofit sector advocacy; its financial and staffing resources, its research and advocacy activities, and its persuasive communications. This article provides insight on a highly partisan think tank sub-sector -organizations actively engaged in tax policy research and commentary -on the eve of historic federal tax reform. Passage of the Tax Cuts and Jobs Act of 2017 ushered in a new era of markedly lower taxes for high wealth households and large corporations in the U.S.
How did a tax cut for high-wealth citizens and corporations come to be considered as potential legislation, despite increasing concerns about economic inequality? One answer could be the advocacy done by partisan think tanks and associated lobbying organizations, funded by ideologically-motivated foundations and philanthropists. Miller (2006), for example, describes the Olin Foundation's decades of funding conservative writers, think tanks, and campus newspapers to advance public preferences for free markets. Mosley and Galaskiewicz (2010, 185) report, "Strategic giving to a network of increasingly influential conservative think tanks like the Heritage Foundation and the American Enterprise Institute has been a key part of this effort … In contrast, centrist and progressive foundations may have had a more mixed orientation, playing roles associated with both philanthropy and charity, by attempting to balance traditional notions of providing for those in need while working to ensure that new service models and policies were reflective of their ideological preferences." Think tanks' long-term role in generating new ideas, combined with the immediacy of promoting policy to elected officials, the media, and the public, suggest an influence on policy that goes beyond think tanks' reputation for ivory tower intellectualism. A robust literature describes the development of the think tank field over time. McGann's book (McGann 2016) and annual report (McGann 2017) catalogue and track the evolving global think tank industry. Providing an important discussion of the think tank industry as an interstitial field -"neither the privileged sanctuary for intellectuals … nor the mercenary 'lobbying firm in disguise'" - Medvetz (2012, p. 41) traces the origins of the U.S. think tank industry from its intellectual beginnings to the competitive and politically potent role in U.S. politics today. Drezner (2017) tracks the growth of the industry as well, noting that increasing competition among think tanks has rewarded those who adopt a stridently ideological tone. Selee (2013) and Struyk (2015) provide practical management advice to founders, funders, and other leaders of think tanks.
Although the think tank literature provides an excellent overview of the field, with particularly good coverage of large, foreign policy think tanks, this article focuses on the specific subset of U.S. think tanks engaged in tax policy issues just prior to the passage of the Tax Cuts and Jobs Act of 2017. Along with 501(c)(3) think tanks, we have included advocacy organizations such as 501(c)(4)s, as well as 501(c)(6)s that are actively involved in disseminating research and other persuasive communications in order to educate the public; i. e. to sway public opinion on the topic of taxation. We have also included both national and state-level think tanks in our analysis. Regional players inform and influence state and local policy making agendas, and efforts to convince state or local-level voters of fiscal issues may also yield benefits at the national level, as voters convey those same opinions when voting for candidates for national offices.
Most importantly, unlike the more broadly-focused literature on think tanks, we analyze operations and resources of the think tanks based on their partisan slant. Determining how they differ allows exploration of two questions: First, what structures and operations might have been successful in influencing the tax reform process? Second, given the observed operations of highly partisan policy advocacy organizations, what policy reforms might help to reduce the influence that high net worth funders have on political processes?

What are U.S. Think Tanks Allowed to Do?
By projecting an identity of independent scholarly centers of expertise, think tanks strive to be sources of "thought leadership," or culturally accepted, trusted sources of wisdom and expertise for a specific topic area. Research and discussion points by think tanks and associated advocacy organizations are circulated by the media and organizations with similar ideological viewpoints. For example, Fox News (McCaughey 2017), a conservative-leaning news source, reported "(t)he nonpartisan Tax Foundation was estimating that the Republican tax plan would lift wages nearly 8 % and produce 1.7 million new jobs over the decade." This support was voiced while political planning was underway to pass the historic tax reform package.
The full scope of work of think tanks is broad and varied; some are university-affiliated research centers, some are funded by coalitions of firms to promote policy benefitting their industry; some provide research primarily at a government's or group of donors' request; and others are fully independent research nonprofits. A common thread among various definitions is that think tanks conduct studies related to government actions or public policy and promote the fruits of that research to specific publics that often include policy makers, the media, and the public (Ahmad 2008;McGann 2007;Medvetz 2012;Pautz 2011;Rich 2001;Weaver 1989).
Most think tanks in the U.S. are nonprofit 501(c)(3) organizations, which face some restrictions on how they seek to influence the political process. Donations to 501(c)(3)s are tax-deductible. Nonprofits may engage in direct and grassroots lobbying as long as the lobbying is not a substantial amount of the organization's activity (see Prentice (2018a) for a detailed description of lobbying regulations for U.S. nonprofits). Direct lobbying is defined by the U.S. Internal Revenue Service as attempting to influence specific legislation by communicating with a voting politician or government staff member, and grassroots lobbying is defined as attempting to influence legislation by communicating with the general public to support or oppose specific legislation.
Although lobbying is only one advocacy tactic (Prentice 2018b), other forms of advocacy can achieve nearlobbying outcomes. Furthermore, by establishing sister 501(c)(4)s, the non-tax-exempt "social welfare" nonprofits that may participate in political campaigning, 501(c)(3)s, can engage in substantial lobbying and political activity (Simon, Dale, and Chisolm 2006). Finally, if the organization wants to fund more direct political activity, they can establish an affiliated 527 organization to do so. Table 1 summarizes Internal Revenue Service structural and regulatory boundaries of 501(c)(3)s, 501(c)(4)s, and 527s. Nonprofit think tanks conduct or promote research that identifies preferred policy actions, gather and package information from a variety of sources to provide to policy makers and the public, and they help shape the narratives adopted by left and right to sell the preferred policy. The following quote, though exhibiting the hyperbole that often accompanies a fundraising pitch, illustrates how a state-level think tank in North Carolina touts its role in the policy arena: The Civitas Institute is North Carolina's most broadly supported conservative policy organization. We're not just a "think tank." We aggressively work to diminish the influence of the Left in our state and effectively communicate the benefits of conservative policies to the legislature and the citizens of North Carolina. Legislators rely on Civitas' expertise when making critical policy decisions. Our research and education has helped drive the most significant reforms of the past decade (Civitas 2017, www.nccivitas.org/donate/).
Out of the 198 organizations in our study, only 16 appeared to operate solely as centrist nonpartisan research think tanks carrying out original research on tax policy. The rest advance a sharply delineated partisan agenda. Interestingly, 79 % of the clearly left-leaning and 74 % of the clearly right-leaning think tanks claimed on their website to be "bi-partisan" or "nonpartisan" (see Judis 2017 for a caustic complaint about think tank partisanship). Even those that did not self-identify as nonpartisan were often labeled "nonpartisan" by media sources, as illustrated by the Fox News quote on page 5. Independence -or the appearance of it -is clearly a treasured image among tax policy think tanks in the U.S., as an independent mien enhances the think tank's appeal to both the public (Jacobsen 2019) and the media; both seeking credible information. U.S. think tanks boast of being independent from (i. e. not funded by) government and political parties, unlike think tanks in other countries . In the tax policy environment, however, financial independence from government and political parties may mean little if the think tank still operates at the behest of key major donors (see Froelich 1999).

Donor Incentives
Using data from the 2010 Survey of Economically Successful Americans, Page, Bartels, and Seawright (2013) analyzed the differing policy preferences of the both top one percent and one-tenth of one-percent of wealthholders in the U.S., revealing that high net worth individuals hold views that are more economically conservative than the general public. Note that these preferences differ from the merely well-off upper twenty or ten percent of the population. Page, Bartels, and Seawright (2013) showed, in particular applicability to our study, that high net worth individuals favor the Earned Income Tax Credit in much lower numbers (13 %) than the general public (49 %), favor a wealth tax at a much lower rate (17 %) than the general public (52 %), and also favor lower capital gains and estate taxes. Given these figures, it is expected that high net worth donors would prefer to fund conservative think tanks promoting reforms that reduce tax rates in the top tax brackets.
Middle and low-income populations, although implied by the above results to be likely donors to liberal causes such as left-leaning think tanks, face an incentive structure that is not as straightforward as with high net worth individuals. Middle and low income individuals may favor reductions in taxes for their income group, yet may also favor the services that the taxes pay for. When it comes to taxing the very rich, however, the public's incentives are sharper. Several recent polls (Reuters/Ipso 2017; Gallup 2017; Pew Research Center 2017) showed majority support across the general public for increasing taxes on the wealthy. Aside from taxing the wealthy, the general public's ambivalence toward taxes (paying taxes versus receiving tax-funded services) may incentivize them to pay more attention to service provision, rather than tax policy. That is, with less disposable income at hand, low-and middle-income populations may be indifferent to donating to upstream policy-related nonprofits, and instead may support downstream to the local food bank, homeless shelter, and other visible causes.
In addition to reflecting the differing donor incentives outlined above, the hypotheses introduced below reflect the situational characteristics of the tax policy environment in 2016. Right-leaning advocacy was proactively focused on advancing reforms that would lower taxes, while the left-leaning stance was reactive, as there was little proactive policy advocacy at the time to increase corporate income taxes or taxes on high-income or high-wealth individuals.

Hypothesis 1 (Financial Resources):
Right-leaning think tanks will receive more total donations than left-leaning think tanks receive. We expect total donations to right-leaning think tanks to predominate because high net worth donors are highly invested in the potential policy gains of a right-leaning tax policy think tanks, whereas low-wealth donors will be indifferent to or unaware of the tax policy agenda.

Hypothesis 2 (Financial Resources):
Right-leaning think tanks will have boards composed of individuals with greater wealth, compared to left-leaning think tank boards. This is the obverse view of representativeness explored in the nonprofit governance literature, where representativeness usually refers to participation of community stakeholders and marginalized groups in organizations serving them (Miller 1999;Gazley, Chang, and Bingham 2010;Weisinger, Borges-Méndez, and Milofsky 2016). In this study, representativeness refers to the think tank's involvement of elites in policy advisory activities that are more likely to yield benefits for high net worth individuals.

Hypothesis 3 (Governance):
Left-leaning think tanks will have larger boards and more volunteers, as they attempt to gain more resources by building strong networks. Schmid, Bar, and Narel (2008) show that the larger number of volunteers with an organization, the greater the organization's policy influence. Left-leaning tax policy think tanks are expected to have public sentiment (and therefore volunteers) on their side. Page, Seawright, and Lacombe (2018) suggest an additional motive for right-leaning boards to be comparatively small: Since they will be promoting policies that are not likely to be popular with the public, a certain amount of discretion -for example, keeping board deliberations within a small circle -would be advised in strategizing how to appeal to a wary public. Hypothesis 3 is tentative, however, because a concentrated policy push -a "movement" -typically involves legions of volunteers. In this case, the push for tax reform was on the right, so for the purposes of movement-building, we expect to see more volunteers in right-leaning think tanks.

Hypothesis 4 (Spin):
Left-leaning think tanks will emphasize a broader range of issues, in order to appeal to more constituents and widen their resource networks. Right-leaning think tanks, in contrast, will focus simply on reducing the scope of government and the reach of taxation.

Hypothesis 5 (Advocacy):
Left-leaning think tanks will spend less on advocacy and will not be affiliated with as many 501(c)(4) and 527 political advocacy and campaigning organizations, due to a comparatively strong existing base of public support for progressive tax policy. In addition, Lu's (2018) meta-analysis suggests a more aggressive policy engagement role for right-leaning tax policy think tanks, who faced a negative policy environment.

Hypothesis 6 (Social Media):
Left-leaning think tanks and affiliated advocacy organizations will show a broader base of support on social media platforms like Facebook and Twitter. Think tanks that want to extend their role past research and into advocacy utilize social media as a low-cost method of informing and mobilizing networks of supporters (Guo and Saxton 2014;Xu and Saxton 2019;Lilleker and Koc-Michalska 2017). Because of the differing levels of public support, we expect left-leaning think tanks to have an easier time attracting a larger social media network of followers. However, a countering viewpoint would be to assume that right-leaning think tanks and affiliated advocacy organizations will make a concerted social media effort to identify supporters, convey information to them, and mobilize them to reach a broader base of public support.

Data Sources
To collect names of US think tanks involved in tax policy research or advocacy, we gathered initial lists from: the right-leaning State Policy Network and Atlas Network, the left-leaning State Priorities Partnership and Economic Analysis and Research Network, University of Chicago Harris School of Public Policy's Think Tanks Guide, and McGann's (2016) report. From a review of foundations self-identifying as being involved in grant making for tax policy issues, we added think tanks that appeared as grantees on the foundations' 990PF tax forms. From donor advised funds such as Donors Trust, Donors Capital, Democracy Alliance, New Venture Fund, and Tides, we gleaned names of think tank grantees. Finally, we identified many additional think tanks from Schedule I of other think tanks' tax forms, as there is considerable sub-granting activity in the sector. In total, we reviewed websites and U.S. Internal Revenue Service Form 990 tax return data for 541 organizations for possible inclusion in the study.
All think tank websites in our sample were reviewed to determine if they were actively engaged in conducting or disseminating research on tax policy and/or promoting partisan viewpoints on tax issues in 2017. The threshold for inclusion in the study required at least four observations of tax policy-related activity in 2017 documented on their website, including press releases, blog posts, research reports, position statements, convenings and conferences, podcasts, and testimony on pending legislation. We included nonprofits that were operating foundations supporting 501(c)(4)s. (See Bryan and Isett (2018) for a discussion of foundations involved in policy reform.) Finally, selected 501(c)(6)s (chambers of commerce) and one 501(c)(5) (union) were included because they were actively engaged in tax policy issues.
Organizations focused on media (watchdog groups, etc.), building leadership skills, community organizing, voter mobilizing, local government, and international issues were excluded. University-based think tanks were excluded because their financial data were not reported separately from their host university. The exception is the right-leaning Mercatus Center, which is located at George Mason University but is an independent 501(c)(3). Note that if academic think tanks more often lean left, the absence of left-leaning academic think tanks in our study introduces bias to our findings. On the other hand, Drezner (2017, p. 75) discusses the "Standard Indictment" that academicians are of little consequence in a partisan policy battle, because academicians rarely or ineptly communicate directly to the public.
We also did not include several huge 501(c)(3) donor-advised funds (listed above), which serve as donor portals to fund think tanks. Donors Trust and Donors Capital, for example, direct donors to selected organizations that are active in conservative or libertarian tax policy research and advocacy. These donor-advised funds attracted contributions totaling more than $500 million in 2016, which suggests the tax benefits accruing to the original donors are not trivial. Because the funds function as pass-through organizations, they act as "donors" to think tanks, and are thus not included in the dataset. We did, however, review their activity to find names of potential tax policy organizations to include in the dataset. Donor-advised funds allow donors to make donations to nonprofits anonymously.
We did not include organizations for which we could not find a Form 990 tax return. This could be because the organization was too new, had dissolved, or was a 527 political action committee (although we noted the existence of the 527 organizations).
Partisan slant and form of organization. As we reviewed the 198 remaining organizations' websites to determine if they were actively researching or promoting viewpoints on tax issues, we noted the political lean of the organizations as indicated by their arguments vis-à-vis the Tax Cuts and Jobs Act, whether they were members in partisan networks such as the State Policy Network and the State Priorities Partnership, and their website verbiage on the topic of other taxation issues. The political lean of the think tank was almost always obvious, as the left-leaning organizations advocated for policies assisting working families, retention of income assis-tance programs, and clearly opposed the Tax Cuts and Jobs Act. Right-leaning organizations expressed strong support for the Tax Cuts and Jobs Act, other forms of lowering taxes, and reducing the size of government.
Readers may differ with the assigned classification of certain organization's political lean. We have included a data appendix so that others can reconfigure the data based on their own determination of political lean. McGann (2017) reports with 2014 data that 39 % of think tanks affiliated with a party report as Independents, 32 % as Democrats, and 23 % as Republicans. In contrast, we labeled just 16 of 198 think tanks in the tax policy arena as "centrist." Five of the 16 centrist organizations have a narrow focus on reducing the federal debt. Table 2 shows the breakdown of the 198 organizations in terms of legal type, and whether the organization operates primarily at the state or national level. Right-leaning think tanks involved with tax policy issues were more involved at the national level but showed a similar rate of affiliation with 501(c)(4) organizations to carry out advocacy and lobbying projects. However, right-leaning think tanks and advocacy groups were more likely to have an affiliated political action committee (527) or business league/chamber of commerce (501(c)(6)). Seven right-leaning think tanks in our dataset had affiliated 527 organizations and three left-wing think tanks had an affiliated 527 organization. Four right-leaning think tanks had sister 501(c)(6) chambers of commerce, but there were no affiliated left-leaning 501(c) (6)  One left-leaning union 501(c)(5) is included in the dataset because they discussed tax policy often enough to meet our threshold minimum. Centrist think tanks did not have affiliated 527s, chambers of commerce, or unions. Because many of the organizations in the dataset were linked with one another, Table 2 shows the net number of organization groups in the dataset.

Think Tank Financial Resources
Hypothesis 1: Right-leaning think tanks will receive more total donations than left-leaning think tanks receive.
Total revenue. From Lakoff's lament (2002) that conservatives fund their think tanks much more generously than the left, to the State Policy Network's complaint that "the Left is way ahead in working with millennials" (Custer 2016, p. 6), both sides express dismay at the resources and influence that the other has on the policy agenda. Table 3 explores the financial resources of the think tanks and associated advocacy organizations involved in tax issues in the US. Using the most recent form 990 available (usually 2016), we recorded total grants and contributions, total revenue, total assets, total liabilities, and unrestricted net assets. Grants and contributions. The first and obvious result was the astonishing importance of grants and contributions in the think tank revenue stream, averaging 92 % of total revenue. For context, note that grants and contributions for reporting public charities in the U.S. (excluding hospitals and higher education) comprise 39 % of total revenue (Roeger, Blackwood, and Pettijohn 2012). Table 3 also reveals differences in the size of the organizations. The 16 centrist think tanks dwarf their partisan counterparts. The average and median left-leaning think tanks receive fewer grants and contributions (and overall revenues) annually than right-leaning think tanks, which lends support for hypothesis 1.
Net assets. Net assets (total assets minus total liabilities) expresses the net worth of the organization. The median left-or right-leaning think tank has less than a year's worth of total revenue held in net assets, but a few well-endowed organizations skew the total and average net assets sharply higher. Centrist think tanks have notably more net assets (compared to their total revenue) than the left-or right-leaning think tanks. One centrist think tank (the Peter G. Peterson Foundation, a private nonprofit foundation) has $732 million in net assets, which is over 50 times the median net asset size for the centrist organizations. Left-leaning think tanks show the lowest net worth, with a median net assets figure of only $749,367. This result also lends support for hypothesis 1.
Restrictions on assets. We recorded unrestricted net assets to determine how flexible those net assets are. Unrestricted net assets are the bread and butter of nonprofit operations (see Bowman 2011; Calabrese 2013); in essence, they are accumulated funds (savings) that are not restricted by donors to limited purposes. Rightleaning think tanks have fewer restrictions on their assets.
Salaries. Given the differences in resource levels, our data show (Table 4) higher salaries (and benefits) for top executives and per person employed at right-leaning and centrist think tanks. This result also holds regardless of the state-versus national (i. e. smaller versus larger) organizations. For example, when only statelevel think tanks are reviewed, salaries are still sharply higher at right-leaning think tanks. Left-leaning think tanks employ more staff members than right-leaning think tanks, despite having less revenue overall. Public support. The form 990s provide a potential way to measure how broadly-based the think tank's support is via Schedule A part II. The "public support percentage" measures support from donors not exceeding 2 % of the total support in a given year, divided by total support. This ratio is computed over five years of data. A higher percentage indicates that the base of support is broader. Table 5 shows the public support percentages for the think tanks, with somewhat higher values for right-leaning think tanks. This result does not support the hypothesis that left-leaning think tanks receive donations from a broader base of donors. There are three complicating factors in interpreting the public support percentage score. First, the score is calculated only for 501(c)(3) organizations, as the form 990s do not require 501(c)(4) advocacy organizations to report these data. Second, the public support percentage score is not calculated for organizations that are new, because five years of data are averaged into the final score. Thus, a new 501(c)(3) with substantial launch funding from a founding donor would have a very low public support percentage score but will not be required to report it. Third, sub-granting via intermediary organizations (like donor advised funds, private foundations, or larger think tanks) can disguise a huge gift by breaking it into small gifts from several sources. Mayer (2016); Page, Seawright, and Lacombe (2018); and Skocpol and Hertel-Fernandez (2016) describe seeding of conservative think tanks and bundling of donations by networked high net worth donors. If left-leaning think tanks are receiving more of their income from larger gifts, and simultaneously showing weakness in overall resource acquisition, it could be that the left-leaning think tanks have both fewer smaller donors and fewer major donors. The wealth level of the board members, for example, could indicate access to major gifts. Board member connections to extreme wealth are examined in hypothesis 2.

Hypothesis 2 (Financial Resources): Right-leaning think tanks will have boards composed of individuals with greater wealth, compared to left-leaning think tank boards.
Fortunately, the very top U.S. wealth-holders are in the public eye, enabling a check to determine the presence of billionaires serving on the boards. We obtained the Forbes 400 Richest Americans list for 2016, which ranks U.S. citizens with the highest net worth. While the calculation of wealth and the relative ranking of each person on the list is approximate, the list is dependable as a data source for indicating extreme wealth (Peterson-Withorn 2016). Table 6 shows the Forbes 400 2016 list members who were serving on tax policy-focused think tanks in 2016. We were also able to find spouse names for most of the Forbes 400 billionaires, but found no evidence of spouses serving on think tank boards. The data show more billionaires serving on both centrist and rightleaning boards, compared to left-leaning boards. This finding lacks the nuanced results we might find from analysis of the mere 10 % or 1 % (Page, Bartels, and Seawright 2013), yet suggests that the highest net worth individuals are more likely to advance the interests of right-leaning think tanks involved in tax policy issues.

Michael Milken Milken Institute Paul Singer
Manhattan Institute

Think Tank Governance
Hypothesis 3 (Governance): Left-leaning think tanks will have larger boards and more volunteers.
Board size. We recorded the names and roles of all independent voting board members (totaling 2,149) listed on the form 990 for the most recent year available (usually 2016). "Independent" requires the board member to not receive a salary from the organization, not to receive more than $10,000 compensation in contract work provided for the organization, and not to have a family member involved in financial transactions with the organization. The median number of independent voting board members in centrist, left, and right think tanks is 17.5, 12, and 7, respectively (average figures are similar to the median figures). That is, right-leaning boards were significantly smaller, which lends support for hypothesis 3. Struyk (2015) recommends a board size of eight members as ideal, in order to enable more meaningful exchange among members, while Selee (2013) endorses larger boards, in order to garner more resources from board members and their networks (see also Hartarska and Nadolnyak 2012). Among the 2,149 board members in our dataset, some board members served on two or more boards, but no board members served on boards with opposing partisan identities. Independence of board members. Small boards were an interesting feature of the board membership data, revealing less independence of board members on right-leaning boards. The sixteen centrist think tanks all had boards with more than five members. Five of the 81 left-leaning think tanks had four or fewer board members, while 29 of the 101 right-leaning think tanks had four or fewer members (and seven right-leaning think tanks had 0 or 1 independent voting board members). Thus, some right-leaning boards exhibited very little separation between board and staff members; a practice described by researchers (Ostrower 2007; LeClair 2019) as potentially enabling more conflicts of interest.
Volunteer engagement. We had hypothesized that left-leaning think tanks would have comparatively more volunteers (despite tax policy reform efforts on the right constituting a proactive policy push, which might imply a key role for volunteers). Centrist think tanks reported the highest median number of volunteers (16), left-leaning think tanks had 10 median volunteers, and right-leaning think tanks reported the lowest (5). In addition, 20 % of the left-leaning and 23 % of the right-leaning organizations left the volunteer question blank. With incomplete data, no conclusions can be made, but given the smaller size of the left-leaning think tanks, it is noteworthy that they still had more reported volunteers and more board members than right-leaning think tanks. Note also that because of the broader policy agenda at left-leaning think tanks (hypothesis 4), their volunteers could have been attracted to the organizations' other, more appealing policy topics.

Hypothesis 4 (Spin): Left-leaning think tanks will emphasize a broader range of issues.
Spin. Classifying the think tanks into their partisan groupings required a website review of each organization. The website rhetoric by the organizations revealed distinct left-right patterns with remarkable consistency. These differences are summarized in Table 7.

Left-Leaning
Right-Leaning

View on inequality overall
Inequality is a fundamental problem that must be solved.
Inequality is overstated. Liberty is more important.

Approaches to alleviating inequality
Minimum wage laws, EITC enhancement, TANF, funding of social services, education, health care No need to alleviate. People are empowered when they are self-reliant, not receiving assistance. How to grow the economy Assist the low-and middle-class (demand side benefits) Tax cuts for high income earners and corporations, reduce regulations (supply side benefits)

Common phrasing
Democratic, voices for children, common good, shared prosperity, progress, fiscal responsibility Liberty, free-market ideals, individual/personal responsibility, freedom

Policy focus
Wide ranging; no national consensus. State-level think tanks often focus on minimum wage and EITC enhancements.
Focused on reducing entitlement spending, reducing taxes on individuals and corporations, and reducing regulations.
Right-leaning think tanks consistently supported reducing government spending, reducing taxes, and reducing regulations. The left-leaning think tanks did not exhibit any countering focus on taxes, compared to the rightleaning think tanks. That is, although left-leaning think tanks expressed opposition to the Tax Cuts and Jobs Act in Fall 2017, they did not, on the other hand, consistently show support for progressive taxation in general. Tax policy verbiage in left-leaning think tanks was almost never a sole focus, and more often discussed alongside a lengthy menu of other concerns, ranging from climate change, social justice, immigration rights, basic rights, voting rights, health, community development, and so on. This corroborates Mosley and Galaskiewicz (2010) description of the "mixed orientation" of progressive think tanks and suggests support for our hypothesis that left-leaning think tanks will cultivate a broader constituency by appealing to potential supporters with diverse interests. Advocacy expenditures. We also hypothesized that left-leaning think tanks would spend comparatively less on advocacy because the public already is expected to be more supportive of their positions (especially with respect to taxation of high net worth individuals). Measuring lobbying expenditures is problematic, as Prentice (2018a) shows: nonprofits often underreport lobbying activity in their tax returns. Right-leaning think tanks are somewhat more likely to have affiliated 501(c)(4) advocacy organization, so the higher number of leftleaning 501(c)(3) organizations' lobbying could partially be attributed to the right-leaning think tanks' offloading lobbying activity to the sister 501(c)(4) advocacy organizations, which are not required to report lobbying expenditures. Note the very high total revenue for the right-leaning 501(c)(4) advocacy organizations shown in Table 8. The funding totals suggest support for hypothesis 5. However, some differences in the social media data are notable. Left-leaning 501(c)(4)s follow others more on Twitter (median 3,635) compared to right-leaning 501(c)(4)s (median 1,359), suggesting wider networking on the left. The sharpest difference in social media activity among partisan think tank advocacy organizations was in numbers of Facebook followers. Left-leaning 501(c)(4)s had (approximate median) 29,000 followers, while right-leaning 501(c)(4)s had (approximate median) 66,000 followers. The social media results, therefore, failed to support the hypothesis that left-leaning think tanks have a broader base of support on social media platforms.

Summary of Partisan Differences in Tax Reform Advocacy
Resources vs. focus. We found that left-and right-leaning think tanks involved in tax policy issues are similar in number. However, the right-leaning think tanks covering tax policy issues have almost twice the revenue and net assets as the left-leaning think tanks and have more billionaires on their boards. Right-leaning and left-leaning 501(c)(3) think tanks covering tax policy issues both have sister 501(c)(4) organizations to aid in advocacy activity, but right-leaning 501(c)(4) advocacy organizations are funded much more aggressively than their left-leaning counterparts. Left-leaning think tanks exhibit more connection with their public constituency, with larger boards and more reported volunteers. Right-leaning think tanks have significantly smaller boards, fewer independent board members, and better-paid staff. Qualitative differences also emerged from our review of think tank programming. Left-leaning think tanks are more often engaged in a broad array of economic inequality concerns, including environmental justice and social/racial inequality. Tax policy is a small portion of their issue agenda. In contrast, right-leaning think tanks are more consistently focused on upstream, federallevel tax reforms such as reduction or elimination of taxes.
Our qualitative results, therefore, suggested an imbalance of focus, which is more difficult to measure empirically than an imbalance of financial resources: In 2016, prior to the passage of the federal tax reform, leftleaning think tanks included tax issues on their agenda, but spread their advocacy and research across a broad spectrum of issues. Left-leaning think tanks and advocacy organizations are also less prominent nationally, with 30 left-leaning organizations and 48 right-leaning organizations focused at least partially on federal-level tax issues. In other words, the passage of the Tax Cuts and Jobs Act of 2017 could have resulted from a larger financial resource base at right-leaning think tanks, but also could have resulted from the right's national, proactive, and more controlled approach to their narrative interface with the public. Reforming tax policy vis-à-vis philanthropy to policy organizations -such as changing tax deductibility of donations -may not produce a change in tax policy outcomes if left-leaning organizations are reluctant to engage at the federal level and are ineffective at reaching the public with a clearly articulated tax policy message.
Limitations. Information is not available to enable identifying who is donating how much to policy-related nonprofits. Enough workarounds exist (via donor advised funds, for example) to enable a nonprofit think tank to be substantially funded by one or a handful of donors and still pass the "public support" test shown in Table  5. That is, at present we cannot determine if a small number of donors are funding the entire operation.
The process by which the right-leaning think tanks achieved advocacy success -in this case the passage of the 2017 Tax Cuts and Jobs Act -remains an unexamined black box. Therefore, policy prescriptions based on just a partial understanding of the think tank policy process could fall short of achieving an impact on the present U.S. patterns of funding for policy advocacy. Nevertheless, we do know that think tanks are receiving funding from high net worth individuals with a keen interest in policy influence (Mayer 2016), and the more money donated, the more leverage those individuals are likely to have in shaping policy to meet their personal preferences. Pundits denounce the huge influence that billionaires George Soros (on the left) and the Koch brothers (on the right) have in promoting their agendas. Thus, Callahan's (2017a) and Reich's (2018) calls for curbs on the favorable tax incentives for high-wealth donors seem warranted.
Finally, the analysis in this article constitutes a case study of the tax reform environment of 2016 in the U.S., when right-leaning organizations were pushing actively for reform and left-leaning organizations were more reactive. Results are not generalizable to policy environments where proponents and opponents both are pursuing proactive and targeted agendas, where public sentiment is more evenly divided on an issue, or where the issue is not germane to high net worth individuals' preferences versus the preferences of the broader public.

Reforming Policy Philanthropy
Reform efforts in the tax policy philanthropy arena may center on concerns with the balance of resources available to think tanks, fairness in incentivizing donors to donate, and transparency.
Achieving balance and fairness in deductibility. As Reich (2018) and others point out, the tax deductibility of donations to 501(c)(3) organizations coupled with the marginal structure of income tax rates ensures that more rewards accrue to wealthier individuals for donating. To level this incentive, the tax deductibility could be changed to be a tax credit or could be tax deductible regardless of the taxpayer's standard deduction limit. Unfortunately, our results suggest that proposals aimed at encouraging more deductibility of donations for lower and middle-income Americans would not necessarily address the left-right policy philanthropy imbalance, because it is not evident that the broad public has an interest in supporting left-leaning policy think tanks. Indirect evidence such as the higher public support percentage and robust number of Facebook followers for right-leaning 501(c)(4)s both suggest plenty of broad-based support for certain conservative policy organizations. Selee (2013), for example, notes the successful broad-based fundraising achieved by the right-leaning Heritage Foundation and Cato Institute.
Deductibility limits. Disallowing deductibility for policy-related philanthropy, above a certain threshold donation, may not work to discourage philanthropy to policy advocacy either, because donations to 501(c)(4) and 527 organizations are already non-deductible and have experienced a surge in funding this decade.
Structural limits: Structural curbs on high net worth individuals' giving to think tanks and associated advocacy organizations, similar to the maximum allowable donations from individuals to political campaigns (currently $2,800 per candidate per election 1 ), might be more effective in diminishing the policy arena's development into a billionaires' battlefield. Caps on donations from individuals might also help to address imbalance of resources across organizations. This simple concept, however, requires determining where to draw the conceptual line between research -which should presumably be left unregulated -and advocacy, which is at least indirectly engaged in policy promotion. Caps for individuals on donating to 501(c)(4) and 527 organizations might be a defensible policy reform, but caps on 501(c)(3) operations would be difficult to justify.
Public Disclosure: Disclosing donor names for donations over a threshold (e. g. $1000 annually) would allow the public to more readily identify the sources of funding on which think tanks and other advocacy organizations depend. To be effective, the donor disclosure must also include the original donor names from sub-granted funds from other organizations -for example, from donor advised funds. Daylighting the donor names would have the effect of revealing where organizations have broad support and where they are narrowly supported by a few high net worth donors. A think tank's credibility is likely to be undermined if it is revealed to be funded by a few key individuals. Regulatory effort with respect to disclosure has moved in the opposite direction: In 2018, the U.S. Treasury Department and IRS announced they would no longer collect donor information from 501(c)(4) and 501(c)(6) organizations (U.S. Department of the Treasury 2018).
Even though the analysis in this article is essentially a case study, the study provides a service by narrowing the focus: Here is a subsector where a small portion of the population is materially invested in think tank and advocacy organization success, while the broader public is largely indifferent. This is also a subsector where public opinion supported taxing high net worth individuals more, yet the opposite tax reform emerged. We encourage future advocacy research on such topics where certain constituents are highly incentivized to influence policy. The apparent imbalance in resources on each side of the tax policy advocacy side forces readers to grapple with the concerns that Reich (2018) and Callahan (2017a) raise. Although we discourage policy solutions that focus merely on deductibility, policy recommendations for the advocacy arena must pay attention to resource balance as a key structural feature that could result in differential influence. We do not claim that resources directly translate to results, but $51 million speaks to the public and lawmakers in ways that $1 million cannot.
Another contribution of this study that merits a much closer look is the surprisingly small boards found (in some cases, having no independent voting board members) among right-leaning tax policy think tanks. This finding, along with the routing of funding through identity-shielding organizations, provides some support for Page, Seawright, and Lacombe's (2018) characterization of "stealth" advocacy by high-wealth donors.
Regardless of policy prescriptions, differences in focus among the left and right-leaning think tanks can still arise from the ambiguity of financial incentives among low and middle-income voters. That is, without a clear personal benefit to influencing tax policy agendas, low-and middle-income donors may simply cede the floor to individuals highly invested in the political agenda. In an effort to appeal to as many constituents as possible, left-leaning think tanks may sacrifice their clarity of purpose and persuasive appeal, as our results indicate. Left-leaning think tanks hoping to push the tax policy agenda may well be advised to look inward for practices that limit their efficacy.

Notes
1 However, individual giving to multiple candidates, PACs, and political parties has no aggregate annual limit.