CHAPTER 1
Representation, Spending, and the Personal Vote
THIS IS A BOOK ABOUT HOW POLITICAL REPRESENTATION OCCURS on government spending decisionsâone of the most consequential powers of government. The Constitution empowers Congress to âpay the Debts and provide for the common Defence and general welfare of the United States.â Federal spending has a pervasive influenceâimpacting nearly every aspect of American life. How Congress allocates money affects the quality of infrastructure in American cities, the availability of health care in rural towns, and the provision of affordable housing across the country. Spending helps guard against harmâhelping local governments prepare for natural disasters, protect against crime and fire, and deter terror attacks. It also sustains a powerful military, a network of federal law enforcement officials, and the flow of commerce and citizens across international borders. Government spending buoys local economies and even supports universities with funding for research.
Political representation in Congress is, in large part, about how elected officials decide how to spend federal money. While a large literature analyzes how district expenditures affect support for congressional incumbents it remains unclear how constituents hold legislators accountable for expendituresâhow constituents attribute spending to legislators, how constituents evaluate those expenditures, and how constituents reward or punish legislators for spending on projects.1 One reason for this lack of clarity is that constituents are unlikely to learn about the projects on their own. Constituentsâ inability to track spending is not an indictment of their democratic competence. Instead, it reflects the many activities representatives perform and the subtle ways that federal expenditures occur. Constituents lack the time, capacity, and incentives to carefully track what their representatives do in Congress to direct spending to the district. Even when spending reaches the district, it is difficult for constituents to attribute that spending to their representative. Projects in the district often do not have an obvious connection to the federal government.2 And even if constituents do recognize that a project in the district comes from the federal government, they may fail to link the project to their representatives.3
Constituentsâ inattention to spending creates a problem for representatives. Legislators want to use spending to bolster their standing in the district, but inattentive constituents are unlikely to learn about expenditures on their own. Political scientists have long argued that legislators use federal expenditures to cultivate support with their constituents and build a personal voteâsupport based on neither partisan affiliation nor ideological agreement.4 For spending to have a direct effect on constituent support, constituents must at least know that the spending has occurred in the district. But legislators also want constituents to attribute the spending to the representative and to view the spending as beneficial to the district.
Representatives solve this problem with communicationâturning the problem of constituent inattention into an opportunity to receive credit for much more than just spending as it occurs in the district. Legislators use credit claiming messagesâstatements intended to âgenerate a beliefâ that a representative is responsible for spending in the districtâas part of a broad marketing campaign to ensure that constituents learn about spending projects in the district and attribute responsibility to their representatives.5 The goal is to create an impression of influence over expendituresâa reputation of being effective at delivering money to the district. To do this, members of Congress issue press releases announcing new projects, send newsletters to describe work done in Washington to secure expenditures, make appearances at groundbreaking ceremonies as projects begin, and cut ribbons when projects are finished.
We show that this marketing effort is effective. Using a large collection of political texts, a series of experiments, and extensive case studies, we demonstrate that legislatorsâ credit claiming affects constituent credit allocation and leads to a personal vote. To demonstrate how legislators claim credit for spending we analyze a large new collection of House press releases. We show that legislators strategically vary their association with spending, depending on their incentives to cultivate a personal vote. Representatives who need the support of independents and opposing partisans to win elections engage in higher rates of credit claiming than legislators who can rely on the support of copartisans to win election. When claiming credit for spending, legislators lay claim to a broad set of activities and grants. Representatives do claim credit for expenditures as they occur in the district, but members of Congress also claim credit for expenditures they had only an indirect role in securing and expenditures that are unlikely to reach the district soon. And they tend to claim credit for relatively small projects in their district.
Constituents are responsive to legislatorsâ credit claiming efforts. Using a series of experiments, we show that constituents evaluate the actions that legislators report performing and are responsive to who receives the funding. The result is that constituents reward legislators for work throughout the expenditure processâeven if the expenditures have yet to be secured, will be delivered only to the district in the distant future, and even if constituents recognize the project has only a small chance of actually occurring. But constituents are largely unresponsive to the amount of money legislators claim credit for securing. Even large increases in the size of expenditures cause only slight increases in support for legislators.
The credit claiming, credit allocation process we characterize helps explain the institutions that disburse federal funds. We show legislators value the mere opportunity to announce expendituresâeven if they had only an indirect role in securing the expenditure. Bureaucrats at competitive grant programs recognize this and create opportunities for legislators to announce expenditures. Legislators take advantage of the opportunities, using subtle language to imply that they are responsible for expendituresâeven though they never literally claim credit for the project. Constituents allocate credit in response to the messages, inferring that legislators are responsible for the spending. And, in turn, representatives support the grant programs when their budgets are threatened.
We also show that the value of credit claiming is contingent on what other political actors say about spending. We show that after the election of Barack Obama congressional Republicans decreased their credit claiming rate and instead criticized government expenditures. The criticism undermines other legislatorsâ credit claiming effortsâit dampens constituentsâ response to messages, creates opposition to spending programs, and even affects constituentsâ attitudes about prior credit claiming efforts.
We demonstrate how representation occurs around federal spending. It occurs through a dynamic process, with legislators anticipating how constituents will react to particular kinds of messages, constituents rewarding legislators for their credit claiming statements, and other actors attempting to affect how legislators cultivate this support.6 Because legislators are entrepreneurial and anticipate constituentsâ reactions, constituents are able to exercise indirect control over their legislator, making it possible for legislators to be responsive to their constituents and creating conditions for democratic accountability.7 The form of this control, however, is distinct from the usual notions of control and accountability in ideological representation.8 A large literature seeks to measure how well legislators align with constituentsâ stated political preferences. This literature provides insights into how well legislators adopt constituent preferences, but is a less useful framework for studying representation around spending. It is less useful because constituents are unlikely to have the strong preferences over spending necessary to be located in ideological space. Rather than reacting to clearly articulated preferences, legislators anticipate constituentsâ reactions to credit claiming messages. And because legislators want to cultivate support, they will attempt to deliver and claim credit for projects popular with constituents. The result of the process is that constituents can have their underlying preferences enacted in the expenditure process.9 But the process also creates new challenges for assessing the quality of representation and reveals new risks in representation. The risks arise because legislators may fool constituents, which may make it difficult for constituents to hold legislators accountable for actual spending that occurs in the district. And if we prioritize the reasoned exchange of ideas, we may criticize legislators for deceiving constituents.10 But the potential benefits may outweigh the risks. Legislatorsâ deceptions of constituents can lead to more efficient spending decisions and create incentives for legislators to work harder throughout the appropriations process.
The findings in this book provide an expansive characterization of how legislators claim credit for spending and how this affects constituent credit allocation. To do this, we make use of new data, introduce new statistical techniques, and deploy new experimental designs. To measure how legislators claim credit for spending, we use a new collection of nearly 170,000 House press releasesâevery press release from each House office from 2005 to 2010. To measure the content of the press releases, we use text as data methods, providing efficient means for identifying press releases that claim credit for spending. To uncover the effects of the credit claiming statements we introduce new experimental designs that enable us to isolate how features of legislatorsâ credit claiming messages affect constituent credit allocation. We embed the experiments in surveys but also use more realistic settings to replicate how constituents may actually encounter credit claiming messages.
When legislators engage in credit claiming they cultivate an impression of influence over expenditures and build a personal vote with constituents. To illustrate how this process works, and how legislators use credit claiming as part of a broader rhetorical strategy, we examine how one representative, Stephanie Herseth-Sandlin, used credit claiming to bolster support in South Dakotaâand how this credit claiming became a liability when she was attacked by an antispending Republican.
1.1 CREATING A PERSONAL VOTE WITH CREDIT CLAIMING
In 2002, Stephanie Herseth-Sandlinâa Democrat from South Dakotaânarrowly lost election to the stateâs lone seat in the House of Representatives to Bill Jankalow, who was serving as governor. But Herseth-Sandlin would soon have an opportunity to claim the seatâJankalow was forced to resign after a vehicular manslaughter conviction. Herseth-Sandlin ran and won in the June 2004 special election over Larry Diedrich, securing 51% of the vote, and beat Diedrich again in the fall, expanding her support to 53% of the vote. By winning the November election, Herseth-Sandlin would join the 109th Congress as South Dakotaâs lone representative in the House, equipped with the power of incumbency and a full term in office to expand her electoral base.
To use the office to build support, however, Herseth-Sandlin would need to be responsive to her constituentsâand in particular to moderates who supported the Republican Party in national elections. While South Dakota voters tend to elect both Democrats and Republicans to Congress, the state is solidly Republican in presidential elections. Recent elections have seen dismal returns for Democratic presidential candidatesâJohn Kerry carried only 39.1% of the two-party vote in 2004 and Barack Obama won only 45.9% of the vote in 2008 and 39% in 2012. The recent results are in line with a long historical trend: since 1932 only two Democratic presi...