Strategic States: The Congressional Roots of Federal Grant Applications

Abstract

While a large body of research explores the federal-level influences over distributive politics decisions, very little attention has been given to the active role state and local governments play in the geographic distribution of federal funds. Before presidents, legislators, and agency leaders can influence the selection of federal grants, state and local governments must expend time and resources to submit grant proposals. We focus on grant applications as our unit of analysis and advance a theory that congressional representation influences the grant application behavior of state and local governments. We analyze US Department of Transportation grant applications and awards from 2009 to 2022 and find evidence that congressional representation meaningfully influences state-level grant application behavior. States apply more aggressively for federal transportation grants when represented by senators in the Senate majority party, and states apply more efficiently for grants when represented by a senator holding an advantageous committee leadership post.

Introduction

The state of Maine received two $20+ million-dollar federal transportation grant awards in 2020, continuing the state’s run of success with the Department of Transportation’s Better Utilizing Investments to Leverage Development (BUILD) grant program. Receipt of those two major awards is even more impressive when we consider that only six of the 655 BUILD grant proposals in 2020 came from Maine. The state similarly won two BUILD awards with only three proposals in 2019 and three BUILD awards with three proposals in 2018. Maine’s success at securing BUILD awards is often attributed to Senator Susan Collins, who is the highest ranking Republican on the Appropriations subcommittee that oversees the BUILD grant program. Senator Collins’ office notes in a press release that, “[s]ince 2009, when Senator Collins became a member of the Appropriations Committee, she has secured more than $721 million for competitive transportation grants for Maine.” ¹ This narrative fits a common theme in the congressional distributive politics literature – that positioning on jurisdictionally relevant committees, as well as being in the majority party and the president’s party, provides legislators with increased access to federal pork (Shepsle and Weingast 1987; Bickers and Stein 2000; Berry, Burden, and Howell 2010; Curry and Donnelly 2021).

However, Maine’s success at pursuing BUILD grant funds is not solely attributable to Senator Collins’ influence over award selection. Local and state government officials in Maine must submit shovel-ready project proposals to have a chance at receiving BUILD grants, and grant proposals require significant investments of time and money. Having a senator in Collins’ position likely increases state and local officials’ knowledge of the grant program and confidence that certain proposals have a high enough chance of success to be worth the opportunity costs. Further, insider access to information on the BUILD program likely prevents Maine officials from wasting time and money on proposals that are unlikely to succeed, as such information decreases uncertainty about each potential proposal’s probability of success. Thus, in addition to potential influence over the DOT’s grant selection process, Senator Collins’ position on the Appropriations committee likely influences Maine officials’ grant proposal decisions.

A large literature examines the federal-level forces shaping the geographic distribution of public resources, but most research ignores the active role state and local governments play in the distribution of federal funds. Grants require applications, and resource-strapped governments face opportunity costs in applying for competitive federal grants. Findings from public administration show that factors such as local government capacity influence the distribution of public resources (Collins and Gerber 2006; Hall 2008; Nelles 2013; Lowe, Reckhow and Gainsborough 2016; Lowe and Sciara 2018). Further, state and local governments do not decide whether to apply for grants in a vacuum. They likely recognize that their chances at acquiring federal funds vary based on the national political landscape and their representation in Congress.

We argue that factors arising from congressional representation – having legislators in the majority party, in the president’s party, and well-positioned on jurisdictionally relevant committees – influence grant application behavior by providing information to potential applicants. This information offers insights into the probability of success before the grants are distributed, thereby influencing application behavior. Majority party membership and presidential co-partisanship generally increase legislators’ access to federal funding, and potential grant applicants likely use information on congressional power dynamics to ascertain whether they have favorable representation in Congress. We hypothesize that favorable representation in Congress, indicated by majority party status and presidential co-partisanship, results in more aggressive grant application behavior. Alternatively, membership on relevant congressional committees provides legislators with more specific information about grant programs, which they can pass on to potential grant applicants in their constituencies. We argue that favorable committee positioning generates useful information for potential grant applicants, leading to more efficient application behavior.

Using a comprehensive dataset of Department of Transportation BUILD grant applications and awards from 2009 to 2022, we find that state application behavior is responsive to congressional representation. States apply for additional BUILD grants and higher levels of funding when they have more senators in the Senate majority. Additionally, we find that states apply more efficiently for grants when represented by a senator who holds an advantageous committee leadership post. Specifically, states submit fewer unsuccessful BUILD grant applications and receive a greater return on proposed funding when represented by a senator in a leadership position on the Appropriations committee. BUILD grants represent only one case study of bureaucratic spending, but our findings suggest that grant application behavior is an important element of distributive politics that is responsive to congressional representation and we hope that this work will stimulate more research into these earlier phases of the process.

Distributive politics literature

Allocating and distributing public resources is a core function of government, and scholars of American political institutions have long sought to uncover the influences over the geographic distribution of federal funds. The federal bureaucracy is the primary institution through which funds are distributed, but other powerful actors – namely, members of Congress and the president – have a stake in influencing federal spending decisions (Arnold 1980). Consequently, existing distributive politics research focuses on the congressional and presidential determinants of bureaucratic resource allocation.

Canonical accounts of congressional behavior clarify legislators’ desire to influence the distribution of public resources. Reelection concerns shape congressional action, and federal projects offer opportunities for members of Congress to build a personal vote (Mayhew 1974; Cain, Ferejohn, and Fiorina 1987; Arnold 1990; Stein and Bickers 1997). Grimmer, Messing, and Westwood (2012) show members of Congress can generate constituent support by claiming credit for federal grants, even though members are not directly responsible for such projects. In addition to building a personal vote, pork helps legislators pursue reelection by lowering the likelihood of facing a quality challenger and increasing campaign contributions (Bickers and Stein 1996; Rocca and Gordon 2013; Lazarus, Glas, and Barbieri 2012). Members of Congress, in constant pursuit of reelection, are therefore incentivized to pursue pork benefits. Congressional committees provide an institutional framework for distributive politics, allowing legislators to secure high-priority and district-specific spending projects for constituents (Shepsle 1979; Weingast and Marshall 1988; Shepsle and Weingast 1987). However, the electoral benefits to be gleaned from pork are conditional on party, constituency characteristics, and the type of pork (Crespin and Finocchiaro 2013; Lazarus and Reilly 2010).

Historically, members of Congress have utilized two venues for procuring electorally valuable pork projects – congressionally mandated projects (earmarks) and influence over bureaucratic policymaking. ² Earmarks have always comprised an exceptionally small percentage of federal spending, but they offer legislators the most direct and traceable form of distributive spending (Lee 2003). The 2011 earmark moratorium eliminated individual members’ ability to directly secure federal funding for their constituents and address district needs (Gordon 2018; McLaughlin 2023). In the absence of earmarks, exerting influence over federal agencies represents the only viable strategy for legislators to secure federal projects. Congressional earmarking has returned under a new program title – Community Funding Projects – although legislators are still incentivized to influence bureaucratic spending decisions. Due to the budgetary and oversight powers Congress maintains over the bureaucracy, bureaucratic spending is amenable to congressional influence (Arnold 1980). Consequentially, members of Congress actively seek to influence bureaucratic distributive policy.

Members of Congress often bypass legislation and influence policy by back-channeling with the bureaucracy (Ritchie 2023), though legislators’ capacity to individually solicit federal projects from bureaucratic agencies is limited. In one of the few studies of direct pleas to agencies by members of Congress over the allocation of resources, Mills, Kalaf-Hughes, and MacDonald (2016) find congressional letter-marking to have little sway with the Federal Aviation Administration (FAA) on air traffic control tower closure decisions. Two aspects of this study are worth highlighting. First, members of Congress do, in fact, try to individually influence bureaucratic distributive policy – members wrote over 100 letters to the Secretary of Transportation pleading for protection of traffic control towers in their districts. Second, legislators’ power over bureaucratic decision-making is limited. FAA decision-making tracked more closely with agency priorities and benefit–cost ratios than letter-marking pressures. A second study of letter-marking, conducted by Neiheisel and Brady (2017) on Department of Labor stimulus funding, similarly finds the practice to have nuanced and conditional effects. “Even though certain members of Congress may be able to guide the process from time to time, it appears that the bureaucracy is very much in the driver’s seat when it comes to the distribution of stimulus dollars” (Neiheisel and Brady 2017: 5). Members of Congress stand to gain from securing and protecting federally funded projects, but their ability to do so via direct contact with agencies appears to be rather limited.

Agencies may ignore individual legislator demands when they conflict with agency goals and priorities, but the geographic distribution of federal resources is generally responsive to congressional influence. Agency leaders recognize the higher powers shaping their budget and exerting oversight, and the geographic distribution of federal resources tends to tilt in favor of partisan power in Congress. Albouy (2013) uses panel data to show that states receive more federal funds when they have senators in the majority party. Having two senators in the majority party equates to a 2% boost in government grants to a state as compared to having two senators in the minority party (Albouy 2013). Similarly, Curry and Donnelly (2021) find that state congressional delegations secure more federal funding as their share of members in the majority party increases. Previous research on earmarks also reveals the partisan distribution of public resources. Balla et al. (2002) and Clemens, Crespin, and Finocchiaro (2015b) find legislators in the majority party are awarded more money for earmarked projects. Additionally, electorally vulnerable members of Congress receive additional earmarks, but only when they are in the majority party (Lazarus 2009).

Placement on jurisdictionally relevant committees may also increases legislators’ ability to procure public resources for their districts. Membership on both the House and Senate Appropriations Committees offers additional access to earmarks (Lazarus and Steigerwalt 2009). Further, Clemens, Crespin, and Finocchiaro (2015b) find relevant House Appropriations subcommittee membership is associated with an increase in the number of earmarks members receive. Conversely, research on aggregate federal funding distribution finds little support for the claim that committee membership yields greater access to federal pork. Multiple studies using total federal outlay data find no significant relationship between federal funding and membership on either the Appropriations or Ways and Means Committees (Dynes and Huber 2015; Kriner and Reeves 2015; Berry and Fowler 2016; Christenson, Kriner, and Reeves 2017). However, Hammond and Rosenstiel (2020) make a compelling case that data on total federal outlays offers a problematic measure of budgetary influence. Using a dataset of military construction appropriations, Hammond and Rosenstiel uncover a substantial effect of an Appropriations subcommittee seat on district funding.

Studies specifically focused on transportation policy reveal a similarly nuanced relationship between committee membership and pork benefits. At the project level, Evans (2004) finds a positive association between membership on the Public Works Committee and transportation project distribution, and Lee (2003) shows membership on the Transportation and Infrastructure Committee yielded additional earmarks in the 1998 surface transportation re-authorization law. However, committee membership did not have a significant impact on state-level funding amounts in Lee’s 2003 analysis, and Lee (2000) finds no relationship between senator membership on the Environment and Public Works Committee and additional surface transportation funding. Taken together, membership on jurisdictionally relevant committees appears to help legislators procure individual projects but has little impact on general funding levels.

Presidents also exert influence over the geographic distribution of federal funds to further their political goals. Hudak’s (2014) analysis of federal grants from 1996 to 2011 reveals substantial presidential influence over distributive policy outcomes. Hudak shows presidents work to organize and take advantage of a “federal grants process that offers presidents and their appointees numerous opportunities to influence the distribution of funds and thereby a variety of outcomes”(Hudak 2014: 7). Because they represent the entire country and hold broad authority over federal agencies, presidents pursue multiple particularistic goals. Presidents direct additional federal funds toward core constituencies, districts with competitive elections, and districts of copartisan members of Congress (Berry, Burden, and Howell 2010; Kriner and Reeves 2015; Dynes and Huber 2015; Christenson, Kriner, and Reeves 2017). Presidential power over the distribution of federal funds is not a modern development – presidents’ wielded this power in the late nineteenth century (Rogowski 2016). In sum, bureaucratic leaders are generally in charge of distributive policy decisions, but members of Congress and the president exert influence over the geographic distribution of public resources.

As the preceding review demonstrates, extant distributive politics scholarship offers a comprehensive account of the federal-level forces that influence final spending decisions. However, research solely focused on federal-level influences overlooks important developments that occur earlier in the distributive politics process. Clemens, Crespin, and Finocchiaro (2015a) highlight this issue in their examination of spatial heterogeneity in demand for pork. They find incentives for pork vary across time and space, suggesting one-size-fits-all models of distributive politics fail to capture significant geographic variation. A vein of research in public administration addresses this topic from a different angle by examining how variation in the capacity of local governments and civic communities influences federal grant distribution (Collins and Gerber 2006; Hall 2008; Nelles 2013; Lowe, Reckhow, and Gainsborough 2016; Lowe and Sciara 2018). For instance, Lowe, Reckhow, and Gainsborough (2016) show that civic capacity positively predicts whether metropolitan statistical areas receive transportation grant awards. ³ Increased capacity arms potential applicants with the resources required for grant success. We aim to further clarify this early stage of the distributive politics process by examining the ways in which federal grant application behavior is responsive to congressional representation.

Theory

A substantial portion of the distributive politics process occurs before agency leaders, the president, and members of Congress squabble over final spending decisions. Federal grants are a central component of distributive politics, and grants require actors outside of the federal government – predominantly state and local governments – to prepare and submit proposals. State and local governments play an active but relatively under-examined role in the distribution of federal resources.

Our theory of the relationship between congressional representation and grant application behavior centers on the concept of information provision. We argue that state and local governments receive information from Congress regarding grant opportunities and apply for grants accordingly. Information provision ranges from vague indications of an encouraging grant environment to specific information about which grant proposals are likely to succeed. After interpreting information from Congress, state and local governments apply for federal grants based, in part, on their perceived chances of success.

Indirect information: partisan power dynamics

Indirect information concerning grants comes from partisan power dynamics in Congress. Extant studies repeatedly reveal two paths through which congressional partisan power dynamics influence the geographic distribution of funds by federal agencies. First, majority party representation in Congress increases legislators’ access to federal funds. In addition to congressionally directed spending through earmarks (Balla et al. 2002; Clemens, Crespin, and Finocchiaro 2015b; Lazarus 2009), majority party status provides increased access to federal grants (Albouy 2013; Curry and Donnelly 2021). Potential grant applicants face a more favorable application environment when represented by legislators in the majority party.

Second, the executive branch shows favoritism to presidential allies in Congress when distributing federal funds. Presidential co-partisanship increases legislators’ access to federal grants (Berry, Burden, and Howell 2010; Hudak 2014; Kriner and Reeves 2015; Dynes and Huber 2015; Christenson, Kriner, and Reeves 2017). As such, potential grant applicants face a more favorable application environment when represented by legislators in the presidents’ party.

Majority party membership and presidential co-partisanship generally increase legislators’ access to federal resources, and potential grant applicants likely recognize such dynamics as material information on the grant environment they currently face – all else equal, an applicant has a better chance of success when it has more legislators in the majority party and in alignment with the president. Applying for federal grants is not a costless activity, and potential applicants must weigh the chances of grant success with the resources required for grant submission. Applications for BUILD grants, a program we describe in more detail below, require up to 25-page narratives, budgets, cost–benefit analysis statements and calculations, community involvement plans, potential matching funds, and evidence of project readiness.

Consequently, we theorize that state and local governments apply more aggressively for grants when they are represented by majority party and presidential co-partisan legislators. Our theory is based on two mechanisms regarding potential grant applicants. First, we assume that grant applicants are sensitive to the opportunity costs of applying for federal grants. Federal grant applications demand time and resources, and potential applicants weigh the cost of applying against the probably of success. Second, we assume grant applicants understand the federal grant process to be political. As discussed above, research clearly demonstrates that congressional representation affects the distribution of federal grants. We assume that potential grant applicants understand the political nature of federal grants and aim to take advantage of favorable representation. ⁴

Direct information: legislator committee positioning

Alternatively, potential applicants likely receive more direct information about grant programs when represented by legislators on relevant congressional committees. Appropriations subcommittee membership offers legislators increased access to distributive benefits (Evans 2004; Clemens, Crespin, and Finocchiaro 2015b), and membership also increases access to information about grant opportunities. Thus, if a potential applicant has representation on the relevant committees, the uncertainty around grant applications should decrease. If legislators have enough committee-based access to influence resource allocation, they likely have the ability to transmit detailed information to potential applicants in their district on grant opportunities. This information advantage aligns with Krehbiel’s conception of congressional committees (Krehbiel 1992). The committee system incentivizes legislators to develop expertise in a specific policy domain, providing them with an information advantage that can be converted into policy benefits.

We expect that the effect of legislator committee positioning meaningfully differs from that of the partisan power dynamics discussed above. Whereas being favored by partisan power dynamics offers indirect information on a favorable grant environment, legislators with committee-based informational access to a grant program can pass along specific information about which grant proposals are likely to succeed and fail. Well-positioned legislators, and their staffs, might also be able to help grant applicants refine applications using knowledge gained from prior years.

Direct, committee-based information about grant programs should not lead to more applications. Instead, applicants with an information advantage should apply more efficiently for grants, meaning they only submit proposals for projects with a high probability of success and shelve lower probability proposals. Advantageous committee positioning allows members of Congress to pass along efficiency-boosting information to potential applicants in their constituency, resulting in a more targeted use of application resources. Further, the staffing advantages afforded to committee leaders provide greater capacity for correspondence with potential grant applicants, enhancing the efficiency of grant applications.

Therefore, we theorize that state and local governments apply more efficiently for grants when represented by legislators with advantageous committee positioning. Here, the mechanism behind our theory involves three steps. First, legislators with advantageous committee positioning gain access to useful information on grant programs. Second, legislators transmit this information to potential grant applicants in their constituencies, thereby reducing uncertainty on the likelihood of application success. This includes information that increases the likelihood that promising proposals succeed, as well as information on which proposals are unlikely to succeed. Third, grant applicants use this information to apply more efficiently for federal grants, improving promising proposals and shelving less promising proposals. ⁵

Together, our theory posits that congressional representation accounts for important variation in grant application patterns. States receive information from Congress concerning grant opportunities and submit proposals accordingly. We expect vague and indirect information stemming from partisan power dynamics to predict variation in application aggressiveness and specific and direct information stemming from committee positioning to predict variation in application efficiency. The empirical analysis that follows uses grant application and awards data for a long-running and competitive transportation grant program to clarify the effects of legislator representation factors on grant application behavior.

Department of Transportation BUILD Grant Program

The Better Utilizing Investments to Leverage Development (BUILD) grant program provides competitive federal grant awards for surface transportation projects. The BUILD program – previously referred to as TIGER grants – began in 2009 as part of the American Recovery and Reinvestment Act and is funded on a year-to-year basis by the annual transportation appropriations bill (Peterman 2019). ⁶ In 2020, 70 BUILD grants were awarded across 44 states, totaling just over $1 billion (US DOT 2020). While only comprising a small percentage of overall transportation spending – the DOT’s 2020 gross discretionary funding resources totaled $87.2 billion (Davis 2019) – the BUILD program represents a consequential component of surface transportation policy. The program was again re-branded as Rebuilding American Infrastructure with Sustainability and Equity (RAISE) grants in 2021, but the structure of the program remains the same. In 2022, the most recent year of awards, 166 RAISE grants were awarded across the US.

BUILD grants are an ideal case study for distributive policy analysis due to the high demand for surface transportation grants, significant DOT discretion over award selection, and continuity of the program over multiple presidential administrations and partisan control scenarios in Congress. The need for surface transportation funding is ubiquitous in America, and the BUILD grant program reflects universal demand for such projects. Between 2009 and 2020 the grant program “provided a combined $9 billion to 678 projects in all 50 states, the District of Columbia, Puerto Rico, Guam, the Virgin Islands” (US DOT 2020). However, the magnitude of awards pales in comparison to the magnitude of applications. Over the same period, the DOT received “more than 9,700 applications requesting more than $175 billion for transportation projects across the country” (US DOT 2020). High demand for BUILD grants makes sense, as they offer one of the only venues through which state and local governments can secure federal transportation funds outside of the annual transportation funding formula. Universal demand and intense competition for BUILD grants make the program an ideal distributive politics case study. Broad geographic demand means the results of our analysis are more generalizable since grant distribution is not merely a function of geographic variation in the condition of transportation infrastructure.

Another notable feature of the BUILD program is the substantial flexibility in who can apply for funding and what types of projects fit under the program umbrella. “BUILD can provide capital funding directly to any public entity, including municipalities, counties, port authorities, tribal governments, MPOs, or others in contrast to traditional Federal programs which provide funding to very specific groups of applicants” (US DOT 2020). Additionally, BUILD grants are awarded for a diverse array of project types, including road, rail, transit, and port projects. A diverse pool of applicants and project types means that a strictly merit-based evaluation and selection system is unlikely – apples to oranges project comparisons likely leave the DOT decision-making process open to other influences, such as congressional pressure.

DOT leadership’s discretion over BUILD grant selection stems from vague selection criteria, a surplus of project applications, and a diverse array of incomparable project-types. Further, the DOT has come under fire for not closely following the existing, albeit limited, selection criteria: “The U.S. Government Accountability Office (GAO) has reported that, while DOT has selection criteria for the TIGER grant program, it has sometimes awarded grants to lower-ranked projects while bypassing higher-ranked projects without explaining why it did so, raising questions about the integrity of the selection process” (Peterman 2019: 1). A separate study found cost–benefit analyses played an insignificant role in the final selection process, whereas projects flagged by a small team of senior policy members had a significantly higher chance of success (Homan, Adams, and Marach 2014).

The DOT’s discretion over project selection indicates the process is amenable to outside pressure. In turn, Congress has shown a keen interest in monitoring and influencing the distribution of BUILD grants. When the Trump administration began favoring rural projects for BUILD grants, Congress responded by placing an upper limit on the share of funding to rural areas (Peterman 2019). Additionally, Congress rejected a DOT initiative to favor applications that provided non-federal sources of revenue and instructed the agency not to factor federal share of project cost into award decisions. Members of Congress pay attention to BUILD grant distribution and deem the program important enough to put a thumb on the scale regarding the geographic distribution of grant funding.

Application aggressiveness hypotheses

Our theory on application aggressiveness argues that state and local governments apply more aggressively for federal grants when they are represented by majority party and presidential co-partisan legislators. The BUILD program offers an ideal case study for testing whether application behavior is responsive to these partisan power dynamics. Members of Congress have shown a keen interest in the geographic distribution of BUILD grants and demonstrated their oversight authority by mandating the proportion of urban and rural grants to be awarded. Further, the BUILD program’s future continuously rests in the hands of congressional appropriators, as the program requires yearly funding. Consequentially, the DOT operates under the influence of congressional pressure, and potential grant applicants plausibly understand that their chances of success vary based on their representation in Congress.

We focus on Senate representation and aggregated applications at the state-level, as it is common for state government agencies to apply for large grants that span multiple congressional districts under this program. US House members plausibly play a role in the grant process, but the sprawling nature of most BUILD grant applications means applications do not always map onto individual US House districts. Further, because senators represent entire states rather than US House districts, they are more likely involve themselves in state-level funding programs (Lee 2003).

Our first set of hypotheses predicts that state-level BUILD grant application behavior is responsive to partisan power dynamics in the Senate. Potential grant applicants recognize that having senators in the majority party and aligned with the president increases their chances of success and apply for federal grants accordingly. In other words, congressional representation provides applicants with indirect information on the likelihood of successful grant proposals. Therefore, we expect states to apply more aggressively for BUILD grants when they have more senators in the majority party and the president’s party:

Hypothesis 1: States with more senators in the majority party apply more aggressively for BUILD grants.

Hypothesis 2: States with more senators in the president’s party apply more aggressively for BUILD grants.

Application efficiency hypotheses

Our theory on application efficiency argues that state and local governments apply more efficiently for grants when represented by legislators with advantageous committee positioning. The BUILD grant program is funded on a yearly basis through appropriations bills, and the Senate Transportation, Housing and Urban Development, and Related Agencies subcommittee of the Appropriations Committee (hereafter referred to as the Transportation subcommittee) has oversight authority over the program. Subcommittee membership likely leads to increased access to information about the BUILD grant program, and we expect states represented by Transportation subcommittee members to apply more efficiently for BUILD grants.

However, recent research suggests appropriations power may be concentrated among subcommittee chairs and ranking members. Berry and Fowler (2016) find that leaders of Appropriations subcommittees maintain influence over resource allocation, whereas subcommittee membership carries little weight. The subcommittee leadership advantage likely extends to our topic of inquiry, and we argue that leadership atop the full Appropriations Committee offers a similar informational advantage. Committee leadership positions provide senators with increased influence over spending programs, access to information, and staffing capacity, all of which help senators relay efficiency-boosting information to grant applicants in their home state. Consequentially, we expect that states represented by leaders of the Transportation subcommittee and Appropriations Committee apply more efficiently for BUILD grants.

Hypothesis 3: States with a senator on the Transportation subcommittee apply more efficiently for BUILD grants.

Hypothesis 4: States with a senator holding a leadership position on the Transportation subcommittee or Appropriations Committee apply more efficiently for BUILD grants.

Developing testable hypotheses for our theory of grant application efficiency requires choosing the most relevant committee positioning for the grant program under inquiry. ⁷ As noted above, our choice of the Transportation subcommittee is based on the fact that the BUILD grant program is funded on a yearly basis through appropriations bills, and the Transportation subcommittee has immediate oversight authority over the program. However, in addition to the yearly appropriations process, funding for the BUILD program is authorized through multi-year surface transportation bills. Through this process the Senate Commerce, Science and Transportation Committee also has oversight authority over the BUILD program. We include Senate Commerce Committee positioning in the analysis, as such positioning potentially leads to a similar efficiency boosting effect that we hypothesize for the Transportation subcommittee. However, the Transportation subcommittee deals with the BUILD program on a more regular and granular basis than the Commerce committee. As such, our hypotheses focus on Appropriations positioning.

Additionally, our hypotheses focus on US Senate representation. House committee positioning – through the House Appropriation Committee and the Transportation and Infrastructure Committee – plausibly influences grant application behavior, but the data on BUILD grant applications is not well-structured to test this proposition. House members have geographically-limited interests when it comes to distributive politics (Lee 2003), and many BUILD grant applications occur at the state level with projects spanning multiple House districts. Further, because our data is aggregated at the state level, House member influence over grant applications is harder to detect. Per our theory, applicants in the House district represented by the chair of the Appropriations Committee plausibly apply more efficiently for grants. However, the grant application data is not associated with specific House districts, and House-level influences are substantially harder to detect at the state level. As such, we include variables for House committee positioning in our analysis, but our hypotheses remain focused on the Senate.

Modeling grant application aggressiveness

Hypotheses 1 and 2 predict senatorial representation dynamics influence state grant application aggressiveness, and we test these claims by examining a comprehensive dataset of BUILD grant applications from 2009 to 2022. We obtained BUILD grant application data from the DOT. BUILD applications were coded by the DOT at the individual project level, and we recode the data into a yearly count of applications, dollar sum of applications, and dollar sum of applications per capita for each state. The number of applications indicates aggressiveness, as proposing additional grants both incurs greater opportunity costs and offers a higher ceiling of success. The dollar sum of applications offers a similar measure but highlights a different dimension of application aggressiveness. Larger projects require more extensive applications, increasing the opportunity costs, but increase the payoff if the grant is successful. In short, strategic and rational potential applicants who believe they face an encouraging grant environment will maximize both the amount and size of grant proposals. Finally, to account for the fact that states vary drastically in population and transportation infrastructure needs, we include a dollar sum of applications per capita measure of application aggressiveness.

The distributions of our three measures of grant application aggressiveness have substantial positive skews (see Supplementary Figure A1), so we take the natural log of each state’s yearly count, dollar sum, and dollar sum per capita of BUILD grant applications. The dependent variables used to test Hypotheses 1 and 2 are approximately normally distributed after the logarithmic transformation, as shown in Figure 1. ⁸

Figure 1. Distribution of BUILD Application Dependent Variables.

The unifying theoretical claim behind Hypotheses 1 and 2 is that states apply more aggressively for BUILD grants under advantageous Senate representation conditions. Specifically, states apply more aggressively for BUILD grants when represented by senators in the majority party (Hypothesis 1) and the president’s party (Hypothesis 2). As such, the independent variables of interest are each state’s number of majority party senators and presidential co-partisan senators. If Hypotheses 1 is accurate, we would expect to observe that a state’s number of majority party senators positively predicts application aggressiveness. Likewise, if Hypotheses 2 is accurate, we would expect to observe that a state’s number of presidential co-partisan senators positively predicts application aggressiveness.

To test Hypotheses 1 and 2, we use time-series cross-sectional (TSCS) analyses to explore variation in state-level transportation grant behavior from 2009 to 2022. We estimate three TSCS linear regressions – one for the count of state applications per year, one for the dollar sum of state applications per year, and one for the dollar sum of state applications per capita per year – predicting application behavior. ⁹ State fixed effects are included in each model to account for unique state characteristics unrelated to Senate delegation partisanship, and year fixed effects are included to control for year-to-year changes in the BUILD grant program. In sum, we employ the fixed effects specifications to focus the analysis on variation in the independent variables of interest rather than other, and potentially unobserved, differences across states and years. ¹⁰

We include a host of covariates in the regressions to account for time-variant factors that plausibly affect state application behavior. While committee positioning is not central to our theory on application aggressiveness, it remains possible that committee-based representation impacts state application behavior. As such, we include the following Senate committee positioning variables, which capture the relevant Appropriations subcommittee (Transportation subcommittee) and authorizing committee for surface transportation legislation (Commerce, Science, and Transportation), in the models: Appropriations subcommittee membership, Commerce membership, Appropriations leadership, Appropriations subcommittee leadership, and Commerce leadership.

Our hypotheses focus on US Senate representation, per our reasoning in the hypotheses section above, but US House representation may still play a role in application behavior. We include variables in the models for the percent of US House delegation members in the majority party, the percent of US House delegation members in the president’s party, and US House committee positioning: percent of US House delegation members on the relevant Appropriations subcommittee, percent of US House delegation members on the Transportation and Infrastructure committee, Appropriations leadership, Appropriations subcommittee leadership, and Transportation and Infrastructure leadership.

A state’s changing level of demand for funding may also impact application decisions. Thus, we also include measures of state population, vehicle miles traveled on state roads, and state contributions to the Highway Trust Fund from taxes on gas and diesel to control for state-level variation in demand for transportation projects. All three demand variables are highly skewed, so we use the natural log of each measure.

Modeling grant application efficiency

Hypotheses 3 and 4 claim that states apply more efficiently for BUILD grants when represented by senators on the Transportation subcommittee and leading either the Transportation subcommittee or the full Appropriations committee. As such, the committee membership and leadership variables are the key independent variables of interest. As noted above, we conceptualize application efficiency as devoting time and effort to proposals with a high likelihood of success and shelving proposals with a low likelihood of success. We operationalize this concept in two ways.

We first operationalize the concept of efficiency by calculating each states’ yearly number of failed BUILD proposals. Failed proposals indicate inefficiency because they represent a waste of time and resources. With perfect information and strategic application behavior, a state would not submit any failed applications. Failed applications equate to wasted time and resources on grant applications, so a count of failed applications offers a useful measure of inefficiency. The distribution of failed BUILD grant proposals per year is shown in the left panel of Figure 2. Because the dependent variable is a count variable with high dispersion, we use a negative binomial estimator for the regression.

Figure 2. Distribution of BUILD Awards Dependent Variables.

Failed applications offer a useful but incomplete measure of application efficiency. It captures the negative component of efficiency – wasted time and resources – but not the positive component of efficiency. We develop a second measure of grant application efficiency to address this concern, calculating the ratio of aggregated application dollars to aggregated dollars received through grants. For example, a state applying for $100 million across all projects and receiving $50 million in BUILD grant awards would yield an efficiency ratio of 0.5. ¹¹ The ratio of awarded dollars to proposed dollars is shown in the right panel of Figure 2.

Similar to the research design for grant application aggressiveness, we employ TSCS linear regression to estimate the effects of Senate committee positioning on grant application efficiency. In addition to all of the independent variables from the application aggressiveness models, we include a categorical variable that denotes whether a state is competitive in presidential elections. Because additional federal funding tends to flow into presidential swing states (Berry, Burden, and Howell 2010; Hudak 2014), it is important to account for competitiveness in award selection and grant efficiency. We define competitive constituency as a presidential vote share between 45% and 55% in the previous presidential election.

Results

Results from the application aggressiveness models are presented in Table 1. We find support for Hypothesis 1 – the relationship between the number of majority party senators and state BUILD grant application behavior is positive, statistically significant (p-value $ < $ 0.05), and substantively meaningful in all three models. States with more senators in the majority party apply for more BUILD grants and higher levels of BUILD grant funding. ¹²

Table 1. Determinants of BUILD Grant Applications

OLS regression with fixed effects for state and year. Clustered standard errors in parentheses.

* p < 0.05

Results from the count of applications model reveal the propensity of states to apply for more build grants as their number of majority party senators increases. All else equal, a state applies for 10% more grants for every additional majority party senator, meaning a state with two majority party senators applies for 20% more grants than one with zero majority party senators. ¹³ This relationship is further clarified in Figure 3, which displays the number of BUILD grants a state is predicted to apply for based on Senate majority representation, holding all other covariates constant. An average-sized state with zero majority party senators is predicted to apply for approximately nine BUILD grants, whereas an average-sized state with two majority party senators is predicted to apply for approximately 12 BUILD grants. ¹⁴

Figure 3. BUILD Grant Applications by Majority Party Senators.

The dollar sum of application, both total and per capita, models in Table 1 tell a similar story to the application count model. All else equal, a state applies for 10% more total grant funding and funding per capita for every additional majority party senator. Figure 4 visually presents these effects. Holding all covariates constant, Figure 4 reveals a substantively meaningful effect of Senate representation on application behavior. The estimated dollar sum difference in applications between an average state with zero majority party senators and an average state with two majority party senators is over $30 million. In sum, consistent with expectations, state BUILD grant application behavior is responsive to national partisan power dynamics. States with more majority party senators apply for additional BUILD grant funding.

Figure 4. BUILD Grant Application $ by Majority Party Senators.

However, grant application behavior is only partially responsive to Senate representation dynamics. We find no support for Hypothesis 2 – states do not appear to apply for additional BUILD grants or funding when they have additional senators in the president’s party. The coefficients for the Senators same party as president variable are statistically insignificant and in the opposite direction as predicted in all three models. Therefore, we report partial support for the theory that states interpret partisan power dynamics in Congress as actionable information and adjust application behavior accordingly.

Table 2 presents findings from the BUILD grant efficiency models, addressing the question of whether advantageous Senate committee positioning predicts more efficient BUILD grant application behavior. The underlying theory behind Hypotheses 3 and 4 claims advantageous committee positioning allows senators to communicate more detailed information to potential applicants, leading applicants to invest in proposals with a higher likelihood of success and shelve less promising proposals. If the theory is accurate and applicable to the BUILD grant program, we would expect Transportation subcommittee membership, Transportation leadership, and Appropriations leadership to negatively predict failed grant proposals and positively predict efficiency ratio (aggregated dollars awarded/aggregated dollars applied for).

Table 2. Determinants of BUILD Grant Proposal Efficiency

Both models include fixed effects for state and year. Clustered standard errors in parentheses.

* p < 0.05

The results displayed in Table 2 offer mixed support for our grant application efficiency theory. Having a senator on the Transportation subcommittee does not appear to influence state application efficiency, suggesting little support for Hypothesis 3. The coefficient for Senate Appropriations Subcommittee Member is neither statistically nor substantively significant. Hypothesis 4, however, is generally supported by our findings. All four Appropriations committee leadership positions negatively predict failed grant applications, and two of the four coefficients – chair and ranking member of the Appropriations Committee – reach statistical significance (p-value <0.05). ¹⁵ Additionally, we find that states represented by the chair of either the Appropriations Committee or Transportation subcommittee have a significantly higher application efficiency ratio (p-value <0.05). That is, they receive a higher return on the amount they apply for via BUILD proposals. All else equal, a state with a senator who gains an Appropriations Committee leadership position applies more efficiently for BUILD grants. ¹⁶

Considering committee leadership has a more substantial impact on the allocation of resources than committee membership (Berry and Fowler 2016), it makes sense that we observe support for our committee leadership hypothesis but not our committee membership hypothesis. Subcommittee members may not have the level of information access necessary to transmit efficiency-boosting information to potential applicants. Appropriations Committee leaders, however, have greater influence over the grant program, access to information, and staffing capacity than ordinary members. As a result of this information and staffing advantage, their states waste less resources on proposals that stand little chance of success.

Qualitative evidence on theoretical mechanisms

Thus far, we have posited two theoretical expectations on the relationship between congressional representation and federal grant application behavior – one on application aggressiveness and one on application efficiency – and tested the implications of these theories using quantitative analysis of BUILD grant applications. Our results offer some support for each theory. On application aggressiveness, we find that states with more senators in the majority party, but not the president’s party, apply more aggressively for BUILD grants. On application efficiency, we find that committee leadership positioning, but not membership, results in more efficient BUILD grant applications. However, the nature of our quantitative analysis is unable to shed light on the underlying mechanisms proposed by our theories. Therefore, after conducting our quantitative analysis, we also conducted qualitative interviews with officials involved in various stages, both on the congressional side and the applicant side, of the federal grant application process.

Our survey methodology was one of convenience and we do not claim that our respondents are necessarily representative of potential sample populations. We offered all of our subjects anonymity. We initially talked via telephone with a staffer in a senior House appropriator’s office. The staffer then provided us with contact information for a staff member who handles transportation for a senior senator who served on Commerce, Science, and Transportation. This interview was also conducted via telephone. Telephone interviews lasted approximately 30 minutes and while we started with some initial questions, we asked followups when interesting avenues presented themselves. Since we are interested in grant applications, we were also able to interview a city manager for a mid-sized city as well as an official who was successful in this particular grant program. Both of these interviews were conducted via email at the request of the subjects. Interviewees were sent questions, and both subjects provided information they deemed important beyond our initial questions. We did not ask follow-up questions. We made an effort to contact other subjects in the government, but those attempts did not yield additional responses. Since our interviews were conducted after we performed the quantitative analysis, we view them as offering a degree of convergent validity as well as a look at the mechanisms involved in the process rather than a basis for building our theory of grant applications.

The overarching theory we delineate suggests congressional offices play an active role at the beginning of the grant application process, before the grant selection process. The pre-submission involvement of congressional offices in the grant application process was confirmed in our interviews with congressional legislative staffers and a local government official. When asked about their office’s involvement in federal grant applications, one appropriator’s staffer said that their office recently sent out a notice to local officials about grant opportunities for transportation projects. The notice included information about the program, encouraged applications, and invited collaboration between local officials and the congressional office. Additionally, a staffer in a senior senator’s office confirmed their collaboration with local governments and the state government on grant opportunities, emphasizing their office’s technical assistance in ensuring applications meet grant program requirements. On the topic of BUILD grants, the staffer said that “we want to be helpful and make sure applications meet program requirements.”

Our interview with a city manager of a mid-sized city further reinforced our theory that congressional offices play a meaningful role in the grant application process:

We keep open lines of communication with our Federal Reps in both the House and Senate. We keep our long-range planning document in front of their staff members in an effort to trigger outreach when they become aware of funding availability. We also reach out to those staffers when particular issues arise and ask if they are aware of any funding on the horizon we should look for. This open line of communication has proven fruitful over the past several years, whether it’s in pursuit of grant dollars or budget earmarks for specific projects.

Our theory on grant application aggressiveness is based on two mechanisms regarding potential grant applicants, and our interview with an official who spearheaded a successful BUILD grant application lends credibility to these proposed mechanisms. First, we argue that grant applicants are sensitive to the opportunity costs of applying for federal grants. Federal grant applications demand time and resources, and potential applicants weigh the cost of applying against the probably of success. On this point, the public official stated, “Our BUILD grant was the result of a combination of thorough research and covers every detail of the project from start to finish.” Further, the successful application came after three years of failed applications, each time revising the proposal based on DOT recommendations. For this BUILD grant awardee, it was a long, intensive, and costly application process. Second, we assume grant applicants understand the federal grant process to be political. We received the following response from the previous grant winner on this topic: “Despite being told this process was not political, we quickly learned that it is highly political and how important our congressional delegations are to the process.” In this case, the interview evidence suggests that our proposed theoretical model of congressional representation and grant aggressiveness was borne out. Grant applicants are sensitive to opportunity costs and understand the political nature of federal grant programs.

The underlying mechanism behind our theory that advantageous committee positioning leads to more efficient application behavior involves three steps. First, legislators with advantageous committee positioning gain access to useful information on grant programs. The staffer in a senior appropriator’s office confirmed the information advantage offered by committee positioning: “When you are on the committee that is providing the funds to the agency, you know that it is there. That it exists. How much is there. How it works.” Second, legislators transmit this information to potential grant applicants in their constituencies, thereby reducing uncertainty on the likelihood of application success. This includes information that increases the likelihood that promising proposals succeed, as well as information on which proposals are unlikely to succeed. According to the city manager with whom we spoke about congressional offices and federal grants, “those open lines of communication have proven valuable in gaining an understanding of the priorities of the agencies awarding grant dollars.” Third, grant applicants use this information to apply more efficiently for federal grants, improving promising proposals and shelving less promising proposals. The city manager emphasized that the decision to submit a federal grant application is made by evaluating the available information on opportunity costs and proposal success: “The decision to apply or not is driven by how closely our project fits the grant parameters, our ability to meet the grant match requirements, and our staff availability to administer the grant [i.e., do the paperwork]. If we are confident we can be successful, we prepare an application.” Therefore, the informational advantage of committee membership plausibly boosts application efficiency by decreasing uncertainty about the grant program and potential applications.

In sum, while our quantitative analysis reveals meaningful relationships between congressional representation and federal grant application behavior, our qualitative interviews yield insight on how these relationships work in practice. Our theory posits specific mechanisms through which congressional representation influences grant application behavior, and we observe those mechanisms at play in our qualitative interviews. Therefore, we believe that these interviews provide additional credibility for our theoretical argument on the pre-submission involvement of congressional offices in the grant application process.

Conclusion

When deciding whether to apply for federal grants, state and local governments must consider the application costs, the potential benefits of the grant, and the likelihood of grant success. Our analysis of DOT BUILD grant applications from 2009 to 2022 shows that grant application behavior is responsive to congressional representation. Congressional offices play an active role in the grant application process, and application behavior varies predictably based on Senate representation. Recognizing when the door is open for higher grant success, state and local governments apply more aggressively for BUILD grants as the number of majority party senators from their state increases. All else equal, a state applies for roughly 20% more grants and funding when it has two senators in the Senate majority as compared to no senators in the Senate majority.

State BUILD grant application behavior is also responsive to congressional committee representation. We find evidence for the claim that Senators with privileged committee positions are able to provide useful informal to potential applicants in their state about the likelihood of grant success. As a result, states with senators holding leadership positions on the Appropriations Committee apply more efficiently for BUILD grants.

Our analysis is restricted to a single transportation grant program, which raises the question of how generalizable our findings are to other grant programs. On one hand, the BUILD program is relatively unique in its size and scope. The program funds essentially all forms of transportation infrastructure projects and accepts applications from a notably large range of applicant types (US DOT 2020). We argue above that the ubiquity of demand for BUILD grants across geography and time make the program an ideal topic of inquiry from an internal validity standpoint, but these characteristics distinguish the program from other forms of federal grants. Different dynamics might shape the application process for more targeted and specific federal grant programs. For instance, our theory of application aggressiveness and congressional representation likely decreases in explanatory power as demand for the grant program decreases. For grant programs with relatively low competition, applicants are less likely to weigh opportunity costs and strategically apply based on partisan power dynamics.

On the other hand, there is reason to believe our findings are generalizable to other federal programs. The large-scale funding, high-profile nature, and long-term staying power of the BUILD program mean that the program is well known amongst potential grant applicants. As such, BUILD grants offer a relatively conservative testing ground for our theory of congressional offices’ active role at the beginning of the grant application process. Our theory on congressional committees and information provision is more likely to shape the application process of smaller-scale and lesser-known grant programs. Further, our interviews indicate that congressional offices are in constant communication with local and state governments on federal grant opportunities. This continuous collaboration – what the city manager we interviewed described as “open lines of communication” – is a plausibly meaningful component of federal distributive politics that extends beyond the BUILD program. Of course, further research should examine how the role of congressional representation changes according to the characteristics of various grant programs.

In sum, we advance the distributive politics literature by highlighting grant application behavior as a meaningful and variable component of bureaucratic spending. Funding for many federal programs requires the submission of grant applications, and grant application behavior varies among state and local governments. We show that this variation is responsive to congressional representation and believe that future research should continue to probe the connection between grant application behavior and distributive policy outcomes. Distributive politics research should give greater consideration to early-stage factors that influence the allocation of federal benefits before bureaucrats, presidents, and members of Congress are able to exert direct influence over the geographic distribution of public resources. For instance, additional research focusing on the distribution of federal grants should consider application aggressiveness as a key explanatory variable. Such research will present a fuller picture of this realm of policymaking, thereby adding clarity to this important but largely missing link in prior research on distributive politics between federal and state/subnational politics that is embedded in the system of American federalism.