The Project Twins
In April 2014, during the second wave of the opioid crisis sweeping the United States, the commissioner of the Food and Drug Administration appeared at a summit on prescription-drug abuse and gave a speech in which she sought to justify the agency’s controversial decision to approve Zohydro ER, a powerful opioid. On the one hand, drug overdose deaths were devastating families and communities, said then-commissioner Margaret Hamburg, but on the other hand, many Americans were suffering and needed painkillers. FDA drug-approval decisions balanced the risks of opioid “misuse, abuse, addiction, and overdose” with the impulse to help “the estimated 100 million Americans living with severe chronic pain or coping with pain at the end of life.”
Hamburg’s use of “100 million” was revisited this year by the New York Times. Two months before her speech, a scientist whose research had led to the figure clarified that only a quarter of the adult population at most was “substantially impaired by chronic pain,” and only 10–15 percent were disabled by it. While still a lot of people, it’s far fewer than 100 million. That estimate had been disseminated in a 2011 report put out by the Institute of Medicine (now called the National Academy of Medicine), which is part of the National Academies of Sciences, Engineering, and Medicine, a respected trio of nongovernmental organizations often sought out by policy makers for independent opinions on matters that are sometimes complex and controversial.
But it seemed the Institute of Medicine was doing little to communicate the clarification. The National Institutes of Health cited the 100 million figure again that September in the first line of a report on opioids, which said one-third of the US population was affected. The American Society of Anesthesiologists is just one of many organizations quoting it on their website today.
What the Times added to the conversation this spring was that, as the figure was introduced and spread, members of the Sackler family—owners of Purdue Pharma, the now-excoriated and bankrupt maker of OxyContin—donated a total of $19 million to the National Academies between 2000 and 2021. What’s more, two of the experts the Institute of Medicine selected to compile the 2011 report on pain were leaders of organizations that had reportedly received grants of nearly $1 million each from Purdue Pharma.
A National Academies spokesperson declined to comment, referring Chicago Booth Review to a statement the academies issued this past April that said the Sackler funds were disclosed, “never used to support an advisory study” on chronic pain or opioids, and frozen in 2019 “as the scale of the opioid epidemic in this country became clear.”
It’s hard to prove influence peddling definitively, and even harder to measure its extent, although such behavior permeates many industries. Harvard’s Susan Crawford wrote a book a decade ago on how telecommunications groups used various means, including charitable giving, to get monopolistic mergers approved in the US. Sunlight didn’t necessarily lead to reform. “I’m afraid I’ve moved away from this policy area. Nothing was changing,” Crawford says.
But data-mining techniques and economic modeling are offering something new to the conversation. “Political charitable giving” in the US represents 2.5 times what companies spend on their investments in political action committees—a third of all federal lobbying dollars, write Chicago Booth’s Marianne Bertrand, University of California at Berkeley’s Matilde Bombardini, Boston University’s Raymond Fisman, and UC Berkeley’s Francesco Trebbi. They’re among the researchers probing how donations might influence policy, providing a growing understanding of the impact of corporate philanthropy on rulemaking and regulatory systems as a whole.
Companies give philanthropically for a number of reasons, including tax incentives and public relations. When a grocery store funds a local community garden, for example, it might lose some sales as potential customers grow rather than buy their own produce—but it can gain valuable goodwill from neighbors and elected officials.
Some gifts might be motivated by policy priorities. In 2015, press reports revealed that the Coca-Cola Company was funding a nonprofit that was downplaying the role of diet in the obesity crisis. (The nonprofit then shut down.) In 2019, the Energy and Policy Institute, funded by charities supporting environmental conservation and climate action, published 10 case studies of utilities that had used philanthropy to “manipulate politics, policies, and regulation.”
But it can be hard to recognize if and when companies are using philanthropy to gain influence. Starting next year, a new anti-money-laundering law will require most public and private companies to disclose their beneficial owners, but charities and nonprofits will be exempt. Bertrand, Bombardini, Fisman, University of Western Ontario’s Brad Hackinen, and Trebbi describe charitable grants as “virtually undetectable by private citizens and civil servants without access to detailed tax returns information.” This opacity, they argue, adds to the appeal for companies of using nonprofits as lobbying allies: “Independent arms-length organizations may extend the credibility of the positions held by special interests,” they write.
As new rules are formulated by federal agencies, companies may look to nonprofits to comment—especially after making a donation. Here’s one way philanthropy can end up shaping policy—and some numbers from the research that demonstrate this system in action over about two decades:
The researchers analyzed logs from the website regulation.gov, which publishes public comments on federal agencies’ rules, alongside tax returns filed by companies’ private charitable foundations. These returns are available to the public, but can be difficult to parse, as a corporation’s foundation or foundations will not necessarily be linked by name to the company, and because, while the returns include appendices listing all grantees who receive payments of $5,000 or more in a tax year, these do not include the nonprofits’ unique identifying numbers on file with the Internal Revenue Service.
The researchers reduced their sample of foundations to those that shared a name with a Fortune 500 company or a company in the S&P 500 in the period 1995 to 2016; this narrowed 120,000 of the biggest active foundations in the US to 629. To confirm which nonprofits were receiving grants, they cross-referenced the foundations’ tax forms with a database of all nonprofits in the US, compiled through these organizations’ applications to the IRS for tax-exempt status.
Looking at data from 2003 to 2016, Bertrand, Bombardini, Fisman, Hackinen, and Trebbi find that receiving a foundation grant of at least $5,000 increased a nonprofit’s chances of commenting within the next year on a rule that had also been commented on by the corporate donor. The effect was large, even when controlling for the likelihood that companies and the foundations they funded had similar areas of interest. Receiving a recent donation increased by 76 percent the likelihood of a nonprofit commenting on the same rule as the company that originated the grant.
The researchers also used natural language processing to analyze the content of the comments, finding that nonprofits and the companies supporting them tended to use more similar language than did the commenters not linked by donations. The companies’ and their recipients’ comments aligned most in the year after a donation.
As an example, after mortgage-backed securities precipitated the Great Recession, Bank of America (along with many other financial institutions) lobbied against a proposed credit-risk retention rule. In a July 2011 comment to the Office of the Comptroller of the Currency, the bank warned that the rule “goes well beyond the regulatory goal of ‘skin in the game’ already met by the risk retention requirement. Instead, it mandates transaction structures that would make private issuance uneconomical for the vast majority of residential mortgage pools.”
In August, it found support from the Greenlining Institute, a nonprofit that describes itself as working “towards a future where communities of color can build wealth, live in healthy places filled with economic opportunity, and are ready to meet the challenges posed by climate change.” The nonprofit wrote that “the proposed rule reaches far past the threshold of safety and instead effectively excludes many deserving families from the opportunity to become homeowners.”
Bank of America gave the Greenlining Institute $150,000 in 2010, according to the research.
When asked to respond to the similarity, Greenlining Institute CEO and president Debra Gore-Mann issued a statement to CBR explaining that when regulatory comments submitted by banks are similar to Greenlining recommendations, the nonprofit considers that evidence of its successful advocacy work.
“The Greenlining Institute engages with financial institutions in coalition with community organizations to advocate for investments, programs, and initiatives that benefit formerly redlined communities of color. During these engagements, we offer guidance, rooted in our policy advocacy work, on how to best repair past harms and serve these communities going forward,” reads the statement.
“Greenlining’s regulatory comments explicitly aim to create economic opportunities for communities of color and close the racial wealth gap, and we encourage the banks we engage with to do the same. At no point in our organization’s history have our funding relationships influenced how we advocate for communities of color.”
Companies spend almost three times the amount on politically motivated charitable giving than they spend trying to influence politicians through PACs.
How Corporations Use Charitable Giving to Wield Political Influence
An expert panel discusses whether lobbying leads to economically inefficient policies and what should be done to prevent regulatory capture.
The Big Question: Is Lobbying Bad for the Economy?Bank of America did not respond to a request for comment.
A revised version of the rule was eventually adopted in 2015. The revisions, according to the agencies responsible, “retained much of the structure of the original proposal, but with more flexibility in how risk retention could be held and with a broader definition of QRM.” (A qualified residential mortgage, or QRM, does not trigger risk retention if it makes up 100 percent of the assets in the securitization.)
Regulators hoped that these changes would “protect investors and enhance financial stability, in part by limiting credit risk, while also preserving access to affordable credit and facilitating compliance.”
They also noted that more than 10,500 individuals and groups commented on the original proposal, and more than 250 on its revision, “including nearly 150 unique comment letters.”
It’s possible that the ties between various members of a coalition do not matter. If the hypothetical Nuts & Bolts Inc. is going to spend $1 million on lobbying to lower import taxes on steel, does it matter whether that money is spent producing official comments and TV ads (entirely transparent, assuming financial support is stated), funneled through a handful of super PACs (less transparent), or invested in persuading nonprofits to join the cause (largely opaque, as only certain types of donations need to be reported)?
Evidence suggests it does. University of Nebraska’s Geoff Lorenz examined the factors that influence whether a US congressional standing committee—such as the House Oversight Committee or Senate Finance Committee—considers a bill, one of the first hurdles proposed legislation must clear on the way to becoming law, and a moment that attracts considerable lobbying pressure.
He analyzed 13,000 organizations’ positions on 4,700 bills introduced in Congress between 2005 and 2014. The biggest factor for a bill to be considered by a committee was not the dollar amount put into a lobbying push nor the number of groups with an opinion on the bill, he finds—it was the diversity of voices.
Lorenz measured diversity by examining the range of interests—such as the industries, unions, issues, or social groups—represented by the organizations that submitted an opinion and subtracting the diversity on one side of an issue from that on the opposing side. Say five consumer goods companies opposed a proposed sugary drinks tax. They would collectively earn a diversity score of 1, as they each represented the same type of organization. Now say the organizations supporting the tax included two doctors’ associations, nine school districts, and one obesity-awareness group. Their diversity score would be 3. With 1 diversity point against the bill and 3 for it, the “net diversity” score would be 2, in support.
The higher the net diversity of a bill’s supporters, Lorenz finds, the better the chance the bill would be considered at the committee level. Between late 2005 and early 2016, fewer than one in three bills were considered, even when sponsored by a member of the party in majority; but when support for a bill went from below average net diversity to above average (as measured by standard deviation), the chances of that bill reaching committee consideration grew by 9 percentage points.
Similarly, Bertrand and her colleagues find that co-commenting nonprofits tend to be effective in promoting companies’ regulatory agendas. The researchers scrutinized federal agencies’ choice of words in documents that were published when a rule was finalized—not the rule itself, which tended to be written in highly codified, formulaic language, but the discussion around the rule. Here, they observed that when companies and grantees co-commented on a rule, the relevant agency’s language in the discussion was more likely to mimic the company’s language in its original comment than if the rule had been commented on by the company but did not receive co-comments from a nonprofit recently funded by that company. The agency was also more likely to mention the particular company’s name, and to do so more often.
“At the very least, it appears that the firm is able to obtain more attention from the regulator in finalizing the rule,” the researchers write.
Lorenz proposes that the effects he observes stem from committee members, and committee chairpersons in particular, using lobbying as a proxy to measure the degree of broader support a bill might get in Congress. This, he writes, could be interpreted as a sign of lobbying being a positive force in lawmaking: “Even though individual groups pursue narrow, parochial concerns, the systemic impact of lobbying includes a ‘bias’ of congressional agendas toward legislation favored by broad elements of the American polity, economy, and society.”
Regulators could apply appropriate discounts to the opinions of potentially compromised parties.
But what if lawmakers are getting false signals? In Congress, this might lead to a stalemate further on in the legislative process: a bill may be advanced because the committee chairperson thought it would receive broad support, when, in fact, it is more divisive than the lobbying suggested.
At the federal agency level, where changes to laws and regulations are not put to a vote but decided by bureaucrats trying to find solutions that satisfy broad segments of the population, false signals can be even more effective for the companies working hard to create them. Academics and practitioners agree that regulators risk capture by the industries they oversee, whether through the prospect of revolving-door jobs or in more subtle ways, such as overlapping social networks with the industries they regulate. “The key question is why agency policies generally ended up favoring the financial sector, with the outcomes we know too well,” writes former University of Connecticut law professor James Kwak, in his 2014 book chapter “Cultural Capture and the Financial Crisis,” about regulatory changes that helped precipitate the financial crisis. “In other words, what mechanisms of influence enabled regulated industry to get its way.”
Bertrand and her colleagues suggest a plausible pathway: reviewing the comments of nonprofits on proposed rules, regulators may think they are getting a read on a broad and disparate swath of American public opinion—when, in fact, they could be operating in an echo chamber of corporate America’s making.
If companies are lobbying by nonprofit proxy, it’s not only effective; it can also offer immediate financial benefits. Companies can rack up a range of tax breaks by giving money to charity—savings they would not realize if they instead put the money into explicitly political donations. For example, giving to their private foundations may allow companies to reduce their taxable income by as much as 10 percent.
It is unclear how much in tax dollars the US loses out on because of all of the deductions available. University of Texas’s Lisa De Simone notes that in 2019, active corporations gave $23 billion to charity, per IRS data. Assuming an average tax rate of 21 percent, that equates to almost $5 billion in allowable deductions. But this figure is a conservative estimate of revenues not collected, she says. There’s a cap on how much corporations can deduct by donating to charity, though various work-arounds exist that allow them to surpass this cap.
Given the tax incentives, it’s not surprising that corporations would spend on lobbying via philanthropy rather than only traditional channels. Bertrand and her colleagues find that co-commenting on federal rules by companies and their nonprofit beneficiaries is a relatively widespread practice: one in 10 of the average company’s comments recorded on regulation.gov shared a co-comment with a nonprofit that had received a grant from its foundation in the previous year. Giving via foundations, they note, is only one path by which companies can support nonprofits (and thus possibly encourage co-commenting). They can donate directly to charities, and their executives can make personal contributions, neither of which require public disclosure.
Nor is co-commenting the only way companies make their philanthropic spending work double time as an investment in political influence. In their 2020 study, Bertrand, Bombardini, Fisman, and Trebbi find that companies’ localized charitable giving fluctuated with the presence or absence of a politician in that district who could wield influence in Washington—following the same pattern as their contributions to politicians via PACs. (For more, read “How corporations use charitable giving to wield political influence.”)
A benign explanation for the co-commenting trend is that foundations’ donations free up resources at grantee institutions, such that the nonprofits have the time and expertise to share publicly, and officially, opinions that they already held—and may in fact have been what drew the companies’ or foundations’ interest in the first place.
That could be the case, yet the researchers find that the similarity between comments by grant-receiving nonprofits and those of the funding company grew after a donation not just relative to other nongrantee commenters, but in comparison with the nonprofit’s co-commentary when it had not recently received a grant.
The system depends on lobbyists offering competing perspectives to less-informed policy makers, who then use these opinions—choosing between them or finding a balance—to create rules and laws that they believe will best serve the public good. Analyzing these dynamics 30 years ago, Northwestern’s David Austen-Smith wrote, “the extent to which any information offered to alter [policy makers’] beliefs is effective depends on the credibility of the lobbyist to the legislator in question.” Quid pro quo arrangements will affect that credibility—but if policy makers are not aware that they exist, a theoretically efficient system can run off kilter.
This sort of damage could be ameliorated through disclosure, argue Bertrand, Bombardini, Fisman, Hackinen, and Trebbi: if regulators understood the links between companies and the nonprofits they support—and if doing so did not require navigating and cross-referencing multiple data sets and building custom machine-learning algorithms, as the researchers undertook—regulators could apply appropriate discounts to the opinions of potentially compromised parties, and take at face value the opinions of other groups.
Companies may well resist greater transparency. The researchers cite an internal email written by a Monsanto executive in 2010 about plans to fund a grantee that would advocate for genetically modified technologies and counter any scientific concerns or proposed additional regulation. “The key,” the executive wrote, “will be keeping Monsanto in the background so as not to harm the credibility of the information.”
Among the biotech products in Monsanto’s portfolio were those modified to be resistant to glyphosate, the key ingredient in its controversial herbicide Roundup. Chemical company Bayer bought Monsanto in 2018, and a spokesperson issued a statement saying that any interactions the company had with regulators regarding glyphosate “have been professional, appropriate, and grounded in the extensive body of science that supports both its safety and non-carcinogenicity.”
As for philanthropists themselves, they might also benefit from opacity, judging by the bipartisan resistance to other proposed reforms, such as no longer allowing foundations to count family members’ salaries and travel expenses toward their minimum annual charitable distributions.
And nonprofits will have their opinions on transparency around funding too. But given donors’ thoughts on the matter alongside the empirical evidence of influence peddling by philanthropic proxy, policy makers would be forgiven for taking at least some of these with a grain of salt.
Combining two transit policies generates far better outcomes than either policy on its own.
Charge Drivers to Improve Public Transit
In Brazil, ESG policies boosted wages but also expanded income gaps.
Skilled Workers Prefer Environmentally Conscious Companies
Turkey’s experience suggests ‘place-based policies’ can have unintended effects.
Does Targeting Aid Help Poorer Regions?Your Privacy
We want to demonstrate our commitment to your privacy. Please review Chicago Booth's privacy notice, which provides information explaining how and why we collect particular information when you visit our website.