Presidents' Politicized Funding Priorities

Thursday, February 19, 2015
Thomas Stratmann

The Congressional Budget Office recently projected that discretionary outlays by the federal government will exceed $12.5 trillion over the next decade. With such large sums at stake, should we grant greater control over the bureaucracy to presidents?

For many scholars and political observers, the answer is clearly "yes." Legislators have a powerful incentive to win their next election, and the most influential members can use the power of the purse to secure their political futures. Over the long term, this can lead to a seemingly unfair distribution of federal funding across the nation. Presidents, so the thinking goes, will limit Congress's inefficient spending due to their nationwide constituency.

Our new research, however, demonstrates that presidents actually target politically important constituencies with federal funds just as much as Congress does. In 2009 and 2010, congressional districts within strongly Democratic states received approximately 50 percent more federal project-grant funding than comparable districts within strongly Republican states.

Although Congress authorizes the overall amount of project-grant funding, federal agencies exercise discretion as to the particular distribution. Presidents then exert influence over the federal agencies through their political appointees. This funding advantage represents $26 million more per district, or $37.50 more per capita, per year on average.

While these sums seemingly pale in comparison with the trillions noted above, imagine the difference $100 million could make for your district over the next four years. Would those changes influence your vote in an election? The evidence suggests that most, if not all, presidential administrations believe it might.

Our own research covers only two calendar years due to data limitations, but it is consistent with the results of several other studies that used longer time horizons as well as different combinations of presidential and congressional control. The presidential incentives driving our main results appear persistent regardless of which administration or party is in office.

This research, while significant in its own right, highlights important principles for a broader array of decisions from mandating vaccinations to education reform. In response to an undesirable outcome, such as a seemingly unfair distribution of federal funding, our initial inclination is often to make stark personnel changes in the decisionmaking body. Such action generally presumes the undesirable outcome was due to individual incompetency or greed. In contrast, our research demonstrates that institutional incentives can largely determine outcomes regardless of which individual or individuals are in charge. We should therefore seek to understand existing institutional incentives better before we make dramatic changes.

Over the coming decade, presidents and Congresses will influence the distribution of at least $12.5 trillion in federal funds. Our results show that presidents target politically important constituencies to influence voter behavior in presidential and congressional elections. Hence, granting greater control over the bureaucracy to presidents will not necessarily result in a more efficient distribution of federal funds. Perhaps scholars and pundits should reconsider whether granting presidents greater control over the bureaucracy is truly preferable.

Presidential Particularism: Distributing Funds between Alternative Objectives and Strategies

February 19, 2015

I. Introduction

The term “congressional particularism” refers to legislators’ deliberate allocation of federal funds to politically influential constituencies in order to achieve certain objectives, usually reelection. Early work on distributive politics suggested that this congressional particularism could be balanced by “presidential universalism.” The widely held belief among scholars of distributive politics was that, because a president’s constituency encompasses the entire nation, presidents would pursue a distribution of federal funds that is relatively independent of their constituents’ political characteristics (Kiewiet and McCubbins 1988; Fitts and Inman 1992; Kagan 2001; Lizzeri and Persico 2001). Recent scholarship, however, theoretically (Fleck 1999; McCarty 2000) and empirically (Shor 2004; Larcinese, Rizzo, and Testa 2006; Bertelli and Grose 2009; Berry, Burden, and Howell 2010; Berry and Gersen 2010; Kriner and Reeves 2012; Albouy 2013; Kriner and Reeves 2014) demonstrates that presidents, like legislators, do influence the distribution of federal funds to target politically influential constituencies. In other words, presidents are particularistic too.

Building on this literature, we attempt to determine whether presidents target specific constituencies to influence their own reelection chances, the reelection chances of their copartisan representatives, or legislators’ votes on important legislation. We also test alternative hypotheses about which specific constituencies presidents will target to achieve the aforementioned objectives. In terms of influencing electoral votes, either presidential or congressional, presidents must choose between targeting core voters (Cox and McCubbins 1986) and swing voters (Lindbeck and Weibull 1987). In terms of influencing legislators’ votes, a president’s optimal strategy for ensuring passage of preferred legislation may include targeting a bill’s current supporters or its moderate opposition voters, or simply refraining from providing any distributive benefits (Groseclose and Snyder 1996). Using a new database that tracks the monthly distribution of project-grant awards, we analyze federal spending obligations during 2009 and 2010 to test which theory or theories best fit our data for each presidential objective. As we explain in section IV, federal project grants are highly discretionary and therefore highly susceptible to political influence.

Our results show that the president attempts to influence presidential and congressional election votes, but not legislative votes, through preferential distribution of federal funds. More specifically, administrative agencies award a disproportionately higher share of federal funds to districts within core Democratic states and core Democratic congressional districts, those generally represented by a member of the president’s party. That share does not, however, expand further during months immediately preceding a congressional election, when constituents’ votes may be relatively more susceptible to political influence. Lastly, shares of federal funding to districts represented by partisan legislators of the president’s party and moderate legislators of both parties did not increase in months immediately before House votes on the Dodd-Frank Wall Street Reform and Consumer Protection Act (Dodd-Frank) or the Patient Protection and Affordable Care Act (PPACA).

Consistently with previous literature (Berry, Burden, and Howell 2010; Berry and Gersen 2010; Kriner and Reeves 2012; Albouy 2013; Kriner and Reeves 2014), we document that districts represented by a member of the president’s party are favored in the budgetary process. Separately, we find that members of committees with substantial influence over budgetary decisions, i.e., Appropriations and Ways and Means, are unable to command a greater share of project-grant funding for their constituencies. These results offer further support for the idea that Congress, at least recently, lacks influence over the distribution of certain federal funds, and that, at least with respect to these funds, presidential particularism is more important than congressional particularism.

Beyond these findings, we consider whether certain executive agencies are more politically motivated than others based on their individual distributions of federal funds to politically influential congressional districts. Our analysis includes four individual executive agencies that account for a substantial portion of project-grant funding and that previous literature identifies as being relatively more amenable to politically targeting federal funding (Berry, Burden, and Howell 2010). We find that project-grant awards by the departments of Health and Human Services, Education, and Agriculture disproportionately favor Democratic districts, in particular core and partisan Democratic districts, which is consistent with relatively high political motivation. Project grant awards by the Department of Transportation, in contrast, demonstrate minimal party favoritism, hence indicating relatively little political motivation. This result for the Department of Health and Human Services contradicts expectations based on a previous finding of relatively low politicization within that department (Berry and Gersen 2010).

In the next section we describe various theories and corresponding evidence regarding distributive politics. Although that literature is notably Congress-centric, our focus is on recent efforts to place the president more at the center in distributive politics. Section III describes alternative hypotheses regarding which political jurisdictions presidents optimally target to achieve different objectives. Section IV describes our methodology and data. Section V presents our main results. Section VI discusses robustness checks and extensions of our results, and section VII concludes.

Continue reading