In nearly all developed health systems, including the US Medicare program, governments set regulated prices for health care providers. There are inherent trade-offs between relying on markets versus governments to allocate resources and set providers’ reimbursements (Acemoglu and Finkelstein, 2008, Chandra et al., 2016, Clemens and Gottlieb, 2017). At present, there is a well-developed literature on the imperfections of health care markets in general (e.g., Arrow, 1963) and on the challenges with relying on bargaining between insurers and providers to set prices in particular (e.g., Cooper et al., 2019). However, much less is known about the political challenges and trade-offs inherent in regulating providers’ reimbursements and, more generally, the influence of political dynamics on health care spending.

In the United States, Congress and the executive branch have significant capacity to influence health spending, particularly via reforms of the $800 billion Medicare program. Congress votes on the laws that define the scope and structure of the Medicare program, including those that determine how the program reimburses providers. The Department of Health and Human Services (HHS) and the Centers for Medicare and Medicaid Services (CMS), both of which are directed by political appointees, implement the laws passed by Congress, including those that dictate how providers get paid. However, despite the large role of Congress and the executive branch in shaping US health policy, outside of a handful of studies (e.g., Leider et al., 2015, Vladeck, 1999, McKay, 2019), there is scant empirical evidence on political economy issues in the US health sector, and political scientists have even asserted that variation in Medicare spending is unlikely to be explained by political dynamics (Berry et al., 2010).

In this paper, we analyze the interplay between congressional politics, the actions of the executive branch, and hospitals’ regulated Medicare fee-for-service payments. Our focus is on examining whether Medicare payments are malleable and can be influenced by political dynamics and on testing for the presence of a feedback loop that allows legislative leaders to work with the executive branch to raise Medicare payments in an effort to pass wider pieces of social policy legislation. To do so, we study events leading up to and following the passage of the Medicare Modernization Act (MMA) of 2003, which was one of the largest expansions of social insurance in US history and gave government-funded prescription drug coverage to seniors via the introduction of the Medicare Part D program.

The MMA was a signature policy of the George W. Bush administration. The law was politically fraught and was passed in the US House of Representatives by a one-vote margin. We focus on the impact of a provision inserted into the MMA of 2003 – Section 508 – which created a process through which hospitals could apply to receive an increase in their regulated Medicare payments. Of note, the rules governing which hospitals received increases in their reimbursements from Section 508 were written by the executive branch after the MMA was passed, allowing officials to steer funds to specific hospitals and regions (thus benefiting specific politicians).

In the canonical analysis of legislator behavior, Mayhew (1974) argued that the primary goal of members of Congress is to be re-elected. This pressure to be reelected drives members of Congress to pass legislation with direct benefits to their constituents, like grants to universities or the building of bridges, for which the legislators can directly claim credit (Mayhew, 1974, Weingast et al., 1981, Rocca and Gordon, 2013). The desire to be reelected also dissuades members of Congress from devoting time to forming the coalitions necessary to pass sweeping laws, such as expansions of the Medicare program, where credit claiming can be difficult (Evans, 2004). To encourage members of Congress to join coalitions, legislative leaders in the US House of Representatives often include provisions with targeted benefits to get reluctant members to vote for sweeping laws where credit-claiming is difficult (Evans 2004). These provisions, such as the building of a train station in a district, are often referred to as “pork-barrel projects”, “sweeteners”, “earmarks”, or “distributive policies”.

This paper tests whether the Section 508 program in particular and Medicare payment increases in general can serve as appealing legislative sweeteners. Our analysis proceeds in three stages. First, we identify which hospitals received a Section 508 waiver, quantify the extent to which Section 508 waivers raised their Medicare payments, and test for a link between a hospital’s receipt of a waiver and the vote its representative in Congress made on the MMA. To do this, we identify, via a Freedom of Information Act (FOIA) request, the hospitals that applied for and were granted a Section 508 waiver (hereinafter “Section 508 hospitals”). This FOIA also allows us to identify the 376 hospitals that applied for a waiver and had their applications rejected. These rejected hospitals form our control group. We illustrate that hospitals that received a Section 508 waiver – 120 in total – saw their Medicare prospective payment system (PPS) payments increase by approximately 7%. We then calculate what hospitals would have made in inpatient Medicare revenue with and without the Section 508 payment increases, ignoring any increases in activity that the rise in payments could precipitate.

We show that hospitals that received a Section 508 waiver were more likely to be represented by a member of Congress who voted in favor of the MMA. To bolster this analysis, we also interview one of the key political architects of the MMA, Tommy G. Thompson, who was the secretary of Health and Human Services (HHS) in the George W. Bush administration from 2001 to 2005.1 Thompson served as secretary of HHS when the MMA was being developed, during the passage of the MMA into law, and as the MMA was implemented. In our interview with him, Secretary Thompson confirmed that the Section 508 program was added to the MMA in an effort to secure votes for the passage of the law. His statements also dovetail with our empirical analysis and anecdotal press reporting at the time that suggested the MMA included a series of smaller provisions designed to win legislators’ support (e.g., Lee, 2003). We also explored, with Secretary Thompson, the mechanism through which politically appointed officials at HHS and CMS could influence Medicare payments.

Second, we show that the Section 508 payment increases led to an 11% increase in Medicare spending at affected hospitals. These spending increases in spending were driven by the mechanical payment increases generated by the Section 508 program and hospitals’ behavioral to the payment increases (e.g., hospitals responded to their payment increases by raising their activity). Hospitals that experienced larger increases in their Medicare payments responded with larger increases in their inpatient activity, highlighting how the increase in Medicare payments induced movement along the supply curve. Ultimately, we show that the Section 508 program did not simply lead to a reallocation of spending across hospitals. Rather, we illustrate that the Section 508 program led to an average increase in Medicare spending of approximately $10 million annually in congressional districts with Section 508 hospitals.

Notably, the Section 508 program led to large increases in health spending in a modest number of congressional districts. From 2005 to 2010, the 20% of districts that received the largest gains from the Section 508 program got hundreds of millions in additional Medicare payments. Measured in federal dollars per capita, this implies the Section 508 program led some districts to get increases in federal spending that were on the same order of magnitude as gains generated by the American Recovery and Reinvestment Act (Boone et al., 2014). Outside the health care sector, there is substantial evidence that elected officials often allocate state and federal resources in an effort to garner electoral support (Bickers and Stein, 1996, Evans, 2004, Levitt and Snyder, 1997, Wright, 1974, Clemens and Veuger, 2021, Berry et al., 2010). As Clemens and Veuger (2021) note from their analysis of the distribution of COVID-19 aid, political considerations can shift the allocation of distributive policies by billions of dollars. Our findings echo that broader literature and underscore why adjustments to the Medicare fee schedule, which allocates billions of dollars in Medicare funding annually, can lead to substantial local benefits that are appealing to lawmakers, who can use them to claim credit on wider pieces of health care legislation.

Finally, we show that the Section 508 program precipitated large increases in campaign contributions for members of Congress representing recipient hospitals. The original language on Section 508 in the MMA allocated $900 million to fund the program and scheduled the program to expire three years after it was introduced (i.e., it included a sunset provision). As a result, the hospitals and wider constituencies that were aware of and benefited from the waivers had considerable interest in seeing the program extended by Congress beyond its slated expiration. Indeed, their shared financial interest in the program’s continuation was so great that Section 508 hospitals formed the Section 508 Coalition, which, according to data obtained from the Center for Responsive Politics, spent millions lobbying to preserve the Section 508 program (The Center for Responsive Politics, 2019b). Past work highlights how interest groups like the Section 508 Coalition can make legislators’ actions more visible to voters and precipitate increases in campaign contributions that serve as ex post rewards for past actions (Stratmann, 2005, Fox and Rothenberg, 2011).

We find that, on average, a member of Congress with a Section 508 recipient hospital in her district received an increase in overall campaign contributions of approximately $191,000 after the Section 508 program was extended. Likewise, we observe suggestive evidence that there were increases in campaign contributions to members representing Section 508 hospitals from individuals working in the hospital sector immediately before the vote to reauthorize the Section 508 program and again after the law was passed. We do not find corresponding increases in campaign contributions from individuals working in sectors that were plausibly unaffected by changes in hospital Medicare reimbursements (e.g., donors working in the agriculture industry or the oil and gas industry). We also observe that the larger the gains a district received from the Section 508 program, the larger the increases in campaign contributions the members representing those districts received. We find, for example, that for each additional $1000 in inpatient spending generated by the Section 508 program, members of Congress received an additional $61 in overall campaign contributions. These findings help underscore why inserting provisions like Section 508 into law can win the votes and support of members of Congress.

Our work contributes to and intersects with three literatures. First, this study highlights the impact of raising Medicare reimbursement rates on hospital volume. There is a growing literature that assesses the extent to which increasing provider payments raises the volume of care provided. Consistent with Clemens and Gottlieb, 2014, Gross et al., 2021, Hackman and Pohl, 2018, Dillender et al., 2023, and Cabral et al. (2021), we find that when hospitals’ Medicare rates are increased, it raises the quantity of care hospitals provide. These findings are vital for policymakers because they imply that the impact of increasing Medicare rates on spending will be a function of both the mechanical effect of the payment increases and the behavioral response (e.g., volume increases) those payment increases induce from providers.

Second, this work adds to the literature on political logrolling (trading favors for votes), the allocation of pork, and work that has assessed the influence of politics on the allocation of federal funds. Our discussion with Secretary Thompson suggests that the Section 508 program was used to win votes for the MMA, and that Medicare payments can be malleable and influenced by political dynamics. Likewise, our work highlights that, given the scale of the benefits they create, Medicare payment increases can be an appealing tool for steering local benefits and thus should be expected to be used for logrolling and the transfer of political rewards. Our findings on the allocation of the Section 508 waivers are consistent with predictions by Evans (2004) that distributive policies, like increases in regulated hospital payments, can be used to entice and reward legislators for “yea” votes on legislation where credit-claiming will be difficult. Whereas previous work, such as Berry et al. (2010), assumes that formula-based payment policies in the Medicare program are unlikely to be gamed, our interview with Secretary Thompson illustrates that the HSS can use the Medicare PPS program to steer federal funding. This is consistent with others’ findings that the executive branch (e.g., the program administrators) can steer benefits to political allies and that federal aid from nationwide programs can be steered toward congressional districts in the majority (Kasdin and Lin, 2015, Clemens and Veuger, 2021). For example, Gordon (2011) finds that during the Bush administration, under pressure from the political staff at the White House, the General Services Administration, a US federal agency that allocates government contracts, steered a disproportionate share of government contracts to congressional districts with vulnerable Republican legislators. Thus, our work underscores the potential for regulated provider fee schedules, like the Medicare PPS, to be used as a tool to steer political benefits.

Third, we add to the literature analyzing the link between legislators’ actions and campaign contributions. Past work has argued that campaign contributions can serve as both ex ante inducements for legislators to act or ex post rewards for past action (Grossman and Helpman, 1994, Grossman and Helpman, 1996, Chamon and Kaplan, 2013, Rocca and Gordon, 2013, Powell, 2014). Interest groups, such as political action committees, it is argued, can both sway legislators and mobilize the public to reward legislators ex post for their actions (Stratmann, 2005, Fox and Rothenberg, 2011). Our findings are consistent with work by Rocca and Gordon (2013) that has found, via analysis of earmarks for the defense industry, that donations tend to be ex post rewards for previous votes. Our work is also consistent with research suggesting that lobbying can be particularly effective when it targets specific, less salient provisions within a law (e.g., hospital payment increases in the MMA) as opposed to lobbying over whole pieces of legislation (McKay, 2019, Kang, 2015).

Collectively, the main contributions of this work are to show a link between domestic politics and health spending in the United States, and to illustrate how regulated provider payments can be used as a tool to steer benefits to specific geographic areas. Past work by Brown (1985) and Vladeck (1999) suggested Medicare policy can be meaningfully swayed by political concerns; Leider et al. (2015) has found that health care facilities represented by representatives or senators serving on appropriation committees receive more in earmarks. We highlight how the Bush administration and legislative leaders influenced the Medicare fee schedule and how changes in the fee schedule raised health spending. Our work is consistent with predictions that, when government policy transfers large amounts of capital from one group to another via regulatory interventions, the policy process grows susceptible to outside influence (Acemoglu and Verdier, 2000, Chan and Dickstein, 2016).

Going forward, this paper is structured as follows. In Section 2, we provide background information on the Medicare program and the Medicare Modernization Act of 2003. In Section 3, we describe the Section 508 program, identify recipient hospitals, quantify the payment increases they received via Section 508 waivers, and test for an association between hospitals’ receipt of Section 508 waivers and votes by the members of Congress who represent them. Section 4 describes the impact of the Section 508 program on hospital behavior and health spending. In Section 5, we explore whether members of the House of Representatives were rewarded for reauthorizing the Section 508 program. We conclude in Section 6.