Shortly before announcing his plans to resign from the presidency of the American Enterprise Institute (AEI), Arthur Brooks outlined a framework for evaluating the “impact” of think tanks such as his own in the Harvard Business Review.1 In his article, Brooks identified several metrics for assessing impact, such as the number of op-eds placed in leading newspapers and congressional testimony by think-tank experts.
AEI is hardly alone in turning to these quantifiable measures. The Aspen Institute published a brief note in 2017 observing that organizations as diverse as the center-left Brookings Institution and the libertarian Cato Institute have sought to use similar metrics to demonstrate their value to donors and other stakeholders.2
Although there appears to be a rough consensus around the metrics relevant to these organizations, studies published to date frequently suffer from at least three key limitations. First, evaluations of a given think tank’s influence are typically published by the institution itself and intended as marketing material. Second, most studies focus on the relative position of a think tank versus comparable organizations at a particular moment, rather than on changes in influence over time (whether of an individual institution or of think tanks as a class). Third, studies have almost completely ignored executive compensation, an especially glaring omission since think-tank president/CEO compensation has risen dramatically in recent years without any corresponding increase in impact, even according to the think tanks’ chosen metrics. These trends may indicate that think-tank compensation has become divorced from the stated goals of these institutions and might also point to serious flaws in the current legal regime governing nonprofits more broadly.
The Rapid Rise in Think-Tank CEO Compensation
Compensation of top think-tank executives has risen tremendously, growing faster on average than for-profit executive compensation and dramatically outpacing general economic indicators. Brooks’s compensation, for example, increased from slightly over $500,000 in 2011 to just over $1 million in 2015,3 representing an annualized increase of about 18.5 percent. By comparison, nominal U.S. wage increases never exceeded 2.5 percent during that period, and nominal GDP growth never exceeded 4.4 percent. Inflation averaged just 1.3 percent. Total direct compensation of S&P 500 CEOs, meanwhile, did not rise above 6 percent during those years.4 The rate of increase in Brooks’s compensation outpaced even the S&P 500 index itself. Although markets were in the midst of a postcrisis recovery, the annualized return of the index during this period (including dividend reinvestment) was 15.3 percent.5
Brooks’s pay increases rank among the highest relative to peers, but he is by no means an outlier. Heritage Foundation chief executive Edwin J. Feulner Jr.’s compensation rose from about $1.1 million in 2011 to an astonishing $3.6 million in 2013. His successor, James W. DeMint, received about $900,000 in total compensation in 2014, which jumped to $1.1 million in 2015. At the Council on Foreign Relations, Richard N. Haass’s compensation grew about 11 percent annually, from around $880,000 in 2011 to more than $1.3 million in 2015. Brookings Institution president Strobe Talbott’s compensation grew at an annual rate of about 10 percent, from just over $400,000 in 2011 to slightly over $600,000 in 2015. During the same time, it is worth noting, think-tank policy analysts and scholars, on average, did not experience anywhere near those rates of increase in their compensation.
As indicated by the above, large executive pay increases are common across the political spectrum, though “conservative” or “pro-market” think tanks seem more comfortable with larger compensation packages and salary increases. By contrast, presidents of the Center for American Progress, which self-identifies as progressive, only received pay increases of about 6 percent annually, and began from a much lower base (growing from about $250,000 in 2011 to just over $315,000 in 2015). Ideologically, the libertarian Cato Institute might appear to be an outlier: chief executive compensation only grew at an annual rate of 3.8 percent during the same period, from approximately $450,000 to $530,000. Nevertheless, as was revealed in recent press coverage of a sexual harassment scandal at the think tank, former president Ed Crane continued to receive almost his entire salary for years after becoming “president emeritus,”6 representing substantial compensation that the above numbers do not take into account. Perhaps surprisingly, this libertarian think-tank executive seemed to prefer the equivalent of a hefty defined-benefit pension plan over a higher salary while in office.
Assessing Think Tanks’ Impact
Presumably, the justification for the rapid rise in think-tank executive compensation is that these individuals have dramatically improved their institutions’ ability to “make an impact.” According to their own preferred metrics for measuring impact, however, this does not appear to be the case.
Attempting to quantify this, it may be useful to devise an “impact score” based on the metrics already advertised by a number of think tanks. The score used in this analysis combines op-eds placed in the New York Times and Wall Street Journal, congressional testimonies, h-index scores (a measure of citations in other papers), and best-selling books, ascribing reasonable weightings to each component (best-sellers are presumably more valuable and much harder to produce than one op-ed, for example). If think-tank executive compensation is driven by impact, one would expect to see a strong correlation between an institutional impact score and executive compensation. Using data from a sample of the largest American think tanks, however, yields an r-squared value of only 0.09, suggesting no significant correlation.7
To be sure, such statistics are inevitably an exercise in false precision. Among other problems, what performance metrics to include is highly debatable—other newspapers or magazines, television appearances, internet video views, article clicks, and so on—or how they should be weighted. The point of the above example is simply to show that the performance metrics identified by the think tanks themselves seem to have no significant correlation with executive compensation.
There also appears to be no significant correlation between the change in an organization’s impact over time and growth in executive compensation, much less the dramatic increases in impact that would be required to justify the high compensation growth shown above. On the other hand, there does appear to be some evidence of a change in focus at most major think tanks. Almost across the board, h-index scores (a more academic measure) have noticeably declined, while op-eds (typically written to advance a more immediate political concern) have noticeably increased. Various explanations can be offered, but one interpretation is that think tanks, initially conceived of as “universities without students,” are now mostly functioning as advocacy organizations.
But does this new emphasis on advocacy translate to more influence at the activist or party level? The answer to this question is also highly debatable, but one admittedly anecdotal point is difficult to overlook, especially for “conservative” institutions—the election of President Trump. Although think tanks cannot technically endorse candidates without losing their 501(c)(3) nonprofit tax status, even the most casual observer is aware of their ideological orientations and the political preferences of their leading voices. It seems safe to say that, during the 2016 Republican primary, Trump and his agenda were hardly popular among conservative think-tank analysts or executives, and yet he comfortably won the nomination anyway. Tariffs, a border wall, and an “America first” foreign policy were far from the dominant messages coming out of AEI and Heritage in the years leading up to 2016, yet these were the distinctive aspects of Trump’s winning campaign. In short, whatever all those op-eds accomplished, they appear to have had little impact on public debates over the most visible issues.
Alternative Explanations of
Think-Tank Compensation Policies
The mismatch between think-tank executive compensation and intellectual influence suggests that other factors may be more important in determining the value of these institutions. Donors may in fact care little about an organization’s position in the “marketplace of ideas” and much more about its access to officeholders.
Think tanks are widely known to be breeding grounds for congressional staffers and presidential appointees (despite Trump’s relative unpopularity at think tanks, the current administration is no exception), and they have numerous points of contact with politicians. Thus a more significant measure of impact might be the number of think-tank fellows appointed to executive or legislative roles or the number of nonpublic meetings held with politicians and their staffs. In other words, think tanks today may mainly function as lobbying firms, and many donors may evaluate their impact primarily on that basis.
This was essentially the conclusion of a 2014 paper by Ken Silverstein entitled “Pay to Play Think Tanks: Institutional Corruption and the Industry of Ideas.”8 Silverstein details the many ways in which think-tank agendas (across the ideological spectrum) are increasingly driven by donors, including large corporations and foreign governments. In exchange for larger contributions, many think tanks allow donors to set or veto research topics, sponsor public events under the aegis of a nonprofit organization, enjoy private access to influential politicians or foreign leaders, and so on. Silverstein reports that scholars at the Center for American Progress were told to speak to the development office before publishing on topics that might adversely affect donor interests. Silverstein also describes situations in which think tanks worked to burnish the image of foreign governments and corporations after receiving large donations from them.
Lobbying through think tanks offers a number of advantages. It is, in some ways, less regulated than straightforward lobbying, and contributions are fully tax-deductible, whereas ordinary lobbying expenses are not. Disclosure requirements are often less strict, as well. Think tanks might also offer a unique patina of legitimacy, either as a result of a nonprofit, nonpartisan image or precisely because of a think tank’s ideological reputation. A Republican politician, for example, may be more likely to support a policy if it is endorsed by “leading conservative thinkers” rather than seen as a mere exercise in corporate rent seeking.
Self-serving donor influence on think-tank activities recently flared into a public scandal at New America. In August 2017, the New York Times reported that New America scholar Barry Lynn was forced out of the organization after praising a European antitrust penalty against Google, which was said to be unacceptable to Eric Schmidt, a former Google executive and major donor to New America.9 (Certain elements of this account were disputed by New America and Google.)
The runaway growth of think-tank executive compensation would seem to confirm Silverstein’s thesis. If major donors view think tanks as ersatz lobbying firms, then it is only logical that their executives would be paid more like lobbyists. Moreover, such donors might prefer to reward executives rather than to fund scholarship in order to better ensure institutional loyalty and message discipline, and because the executive can exert the most control over the organization’s interactions with policymakers. In addition, these donors are probably less focused on a think tank’s participation in broad policy debates and more focused on whether the institution’s activities advance their own narrower interests.
Think-Tank Executive Compensation and
The rapid rise in think-tank executive compensation raises important questions about the current regulatory treatment of think tanks and the nonprofit sector more generally. By definition—and by law—a nonprofit is only a nonprofit so long as no part of its net earnings inures to the benefit of any individual. This requirement not only prohibits “shareholders” from having a claim on a nonprofit corporation’s net assets or earnings (nonprofits typically do not have shareholders at all) but also imposes a standard on reasonable compensation for executives and employees of 501(c)(3) organizations. A 501(c)(3) that raises money only to pay its executives absurdly high salaries would almost certainly lose its tax status. Indeed, current law imposes rather stiff penalties on nonprofits (and their board members) that approve inappropriately large officer compensation packages.
At the same time, it is only logical that the CEO of, say, a large nonprofit hospital network with thousands of employees should be paid more than, say, the part-time leader of a small organization dedicated to maintaining a local park. Thus the standard usually employed to determine when nonprofit compensation becomes excessive to the point of inurement is whether a nonprofit’s compensation policies are in line with those of comparable organizations.
In themselves, both of these requirements seem entirely sensible, though it is not too difficult to imagine how they could potentially conflict, or to imagine the difficulty in determining when increasing executive pay represents the de facto inurement of nonprofit earnings to an individual. In reality, the comparability standard can be stretched to render the prohibition on inurement almost meaningless—especially if for-profit corporations offering their executives stock-based compensation can be treated as comparable organizations.
Large Washington think tanks provide an interesting case study that demonstrates just how flexible the regulation (and internal governance) of nonprofit compensation has become. Researchers at Cornell University have determined that for-profit sector CEO compensation typically increases at about 30 percent of the rate of the increase in company revenue. In other words, if a company’s revenue increases by 10 percent, one would expect CEO pay to increase by 3 percent.10 Given the prohibition against inurement at nonprofit organizations, one would further expect that think-tank chief executive compensation growth relative to organizational revenue growth would follow—certainly not exceed—this ratio. In fact, however, from 2011 to 2015, many think-tank executives received pay increases significantly in excess of this for-profit sector compensation-to-size growth ratio. This means that institutional revenue growth actually inures to the benefit of nonprofit think-tank presidents at a higher rate than it does to the average CEO of a for-profit business. Perhaps it is no surprise, then, that many think tanks have significantly ramped up fundraising efforts—and expenses.
Compensation-to-revenue growth ratios address comparability purely in terms of organization size, but it is also worth exploring comparability in terms of organizational purpose. For instance, Harvard University’s faculty outperforms AEI in terms of intellectual impact by virtually any measure, including Arthur Brooks’s preferred metrics discussed above. In addition, Harvard is a much larger and more complex organization with thousands of employees and students. Yet, in 2015, the president of Harvard received compensation of about $816,000—despite several years of healthy increases as well—still significantly lower than the $1 million plus that Brooks received. Thus whatever Brooks’s metrics may be used to evaluate, his personal compensation appears to have been determined by other factors.
Recommendations and Conclusions
The growing separation between think tanks’ stated goals and their executives’ compensation exacerbates the confusion currently surrounding their role in public life. If think tanks are to be treated as public charities whose value lies in their intellectual contributions to society, then nonprofit compensation standards should be enforced with greater vigilance—if not by the IRS, then at least by their donors and directors, who should want to maximize funding for research.
On the other hand, if think-tank executives want to be paid more like lobbyists—and their donors’ main interests are instrumental rather than philanthropic—then their 501(c)(3) status should be revisited. Many of these organizations would arguably be better classified as 501(c)(4) advocacy organizations, for which personal donations are not deductible and any deductions claimed by businesses must exclude the portion spent on lobbying activity. The mere fact that think tanks sponsor some research does not necessarily mean that these institutions merit 501(c)(3) status. Lobbyists commission studies and occasionally write op-eds, too. For many corporate and other large donors, think tanks effectively represent yet another example of regulatory arbitrage.
In any event, the rapid rise of think-tank executive compensation seems out of step with public perceptions of their mission, with the performance metrics they claim to rely on, and perhaps even with nonprofit anti-inurement provisions. Laments over the politicization and corruption of think tanks are nothing new, of course, but the dramatic compensation increases witnessed in recent years should raise new questions for donors, regulators, and the general public about the role of these organizations in American politics.
This article originally appeared in American Affairs Volume II, Number 2 (Summer 2018): 152–61.
2 “Are Think Tanks Effective?,” Aspen Institute (website), February 17, 2017.
3 Compensation figures taken from publicly available nonprofit organization tax filings.
4 Aubrey E. Bout and Brian Wilby, “S&P 500 CEO Compensation Increase Trends,” Harvard Law School Forum on Corporate Governance and Financial Regulation, October 7, 2017.
5 Represents the period from June 30, 2011, to June 30, 2015 (for consistency with annualized compensation growth rates).
6 Daniel Lippman and Maggie Severns, “Former Cato Employees Describe Years of Harassment,” Politico, February 8, 2018.
7 Includes American Enterprise Institute, Brookings Institution, Cato Institute, Center for American Progress, Council on Foreign Relations, Heritage Foundation, and New America Foundation.
8 Ken Silverstein, “Pay to Play Think Tanks: Institutional Corruption and the Industry of Ideas,” June 15, 2014.
9 Kenneth P. Vogel, “Google Critic Ousted from Think Tank Funded by the Tech Giant,” New York Times, August 30, 2017.
10 Kevin F. Hallock, “The Relationship between Company Size and CEO Pay,” Workspan, February 2011.