Brian F. Crisp Research Projects

Brian F. Crisp

Research Projects

My individual research projects all relate to a general interest in formal democratic institutions. I have devoted some time to studying the origins of institutional designs and how they undergo change. More typically, in my work I assume institutions are exogenous and attempt to discern how they act as incentive structures for a host of actors – voters, party leaders, candidates, legislators, presidents, bureaucrats etc. Bringing institutions and actors together in models of the policy-making process requires the use of key concepts such as “constituencies” and “representation.” Understanding how politicians conceive of “representation” is necessary in order to offer explanations for government decisions, policy outcomes, and government performance.

Research Interests

Working Papers

Ongoing Research Themes

Completed papers (marginally) suitable for public consumption include the following:

Potentially Dominant Presidents:

The Combination of Decree Authority and a Strong Veto in Theory and Practice

with Kristin Kanthak, Jong Hee Park, and Oleh Protsyk

Several constitutions endow presidents with decree authority – the ability to make law without congressional consent. Presidents with constitutionally allocated decree authority have the ability to simply redefine the policy status quo. On the other hand, after the president chooses a policy by decree, if they are so inclined and capable of collective action, a majority of legislators has the authority to pass countervailing legislation. The existence of a strong veto may give opposition legislators pause even where they are in the majority – but not a supermajority. We develop a formal model of decree usage in presidential systems and then empirically test the implications of that model with data from six separation of power systems from Asia, Europe, and the Americas. We find that presidents are more aggressive and confrontational in their use of decree authority than a simple model of interbranch relations would suggest. The collective action issues confronting partisan blocs and multiparty coalitions in the legislature are apparently too great for legislators to check and balance effectively a president with decree powers.

Vote-Seeking Incentives and Investment Environments: Credit Claiming and Protectionism

with Nathan M. Jensen, Guillermo Rosas, and Thomas Zeitzoff

Incentives to cultivate a personal reputation require that legislators generate policy outcomes for which they can claim credit. We show that these incentives can even make themselves felt in international agreements – a domain typically considered to be within the purview of the executive branch. The United States offers a boilerplate bilateral investment treaty to all prospective signatories, but countries with electoral systems which encourage personal vote seeking are more likely to negotiate exceptions to the standard agreement. Legislators benefit by being able to claim credit for having protected their constituents from the competition an unrestricted agreement would entail.

Using Cosponsorship to Estimate Ideal Points

with Scott W. Desposato and Matthew Kearney

Ideal point estimates based on roll-call votes have provided leverage for a wide variety of theory testing efforts. However, in many legislatures, roll-call procedures are not used at all or are used on only a biased sample of votes. What is more, they may reflect unmeasured levels of agenda control or other forms of leadership influence. One recently proposed alternative is cosponsorship data. Conceptually similar to rollcall votes, cosponsorship data are appealing for a variety of reasons, including the fact that they are usually more readily available. However, the data generating process for cosponsorship is undertheorized and little-studied. In this paper, we examine the properties of ideal point estimates from cosponsorship data. We find problems with the use of cosponsorship data in all but very exceptional circumstances.

Electoral Mandates, Political Institutions, and Policy Change
with Santiago Olivella

Elections give voters an opportunity to select policymakers with preferences most closely approximating their own. Campaigns reveal the policy positions politicians will take once in office, and voters use this information to make a vote choice. Through elections politicians are granted a mandate to pursue their preferred policies as revealed during the campaign. If, at a subsequent election, voters abandon previously supported politicians, they are sending a mandate for change. The ability to implement a mandate is in part a function of the institutional requirements for enacting new policies. We provide simple formal models of the budget-making process as it varies across presidential systems. We use these models to design simulations which produce predictions about where we should expect more or less expenditure and revenue stability – given some level of changing preferences among pivotal politicians. We capture the level of change in politicians preferences with data on electoral volatility, measured at individual, party, and ideological bloc levels. We test our predictions about how mandates are translated by budget-making institutions into budget stability with data from Latin America.

Vote Seeking in a Flexible List System: Electoral Incentives and Legislative Politics in Slovakia

with Mike Malecki

Variations in electoral systems alter the incentives legislative candidates face for striking a balance between enhancing their own personal reputations or enhancing the shared reputations of their parties when seeking to further their careers. In a flexible list system, legislative candidates must calculate how many seats their party’s reputation is likely to garner and whether their personal reputation is sufficient to rank them among the group of copartisans elected to parliament. Due to barriers which diminish the prospects of a party’s list being reordered, most existing work assumes that personal reputations and preference votes are unimportant in flexible list systems. We test whether variations in legislators’ vote seeking attributes and activities influence the preference votes they receive when seeking reelection. In terms of attributes, we examine several basic biographical characteristics which may serve to enhance a candidate’s personal reputation, including age, education level, gender, and former occupation. Personal vote seeking actions include voting against the majority of one’s party, initiating private-member bills, and proposing amendments to pieces of legislation. We reason that these actions, not shared with all their copartisans, are opportunities for members of parliament to enhance their own individual reputations. We test our hypotheses with data from Slovakia for the period 2002–2006.

Electoral Efficiency and Electoral Reform

with Yael Shomer

Electoral rules are widely recognized as incentive structures influencing the behavior of voters, individual politicians, and political parties. What explains the decision to reform these formal institutions? Shugart has hypothesized that electoral systems which are "extreme" in intraparty and/or interparty terms will be prone to provoke reform. We use a cross-national, time-serial dataset to explore the incidence of electoral reform. We seek to determine whether extremism predicts reform and the extent to which the specific reforms undertaken moderate the system on interparty and intraparty dimensions.

Some of the papers listed above are subsets of the major research threads outlined briefly below:

Pivotal Politics in Latin America’s Presidential Systems

(with Santiago Olivella)

Krehbiel provided a parsimonious and compelling model of the policy-making process in the U.S. version of presidentialism. However, to be applicable to other separation of power systems, several innovations are necessary. First, in many comparative contexts we lack the roll call vote results necessary to identify the precise ideal points of pivotal actors. We know that elections give voters an opportunity to select policymakers with preferences most closely approximating their own. If, at a subsequent election, voters abandon previously supported politicians, they are sending a mandate for change. We focus on these changing mandates as a means of capturing changes in the relative positions of pivotal politicians over time.

Likewise, we then extrapolate from the typical focus on predicting policy outcomes in terms of precise points in a policy space and instead focus on trying to measure the stability of policy – such as changes in government spending patterns. The ability to implement a mandate for change is in part a function of the institutional requirements for enacting new policies. Amendment and veto powers (as well as reversion points) vary across the institutional designs of Latin American democracies. We provide simple formal models of the budget-making process as it varies across presidential systems. We use these models to design simulations which produce predictions about where we should expect more or less expenditure and revenue stability – given some level of changing preferences among pivotal politicians.

We capture the level of change in politicians preferences with data on electoral volatility, measured at individual, party, and ideological bloc levels. We test our predictions about how mandates are translated by budget-making institutions into budget stability with revenue and expenditure data from presidential systems in Latin America.

The Sample Properties of Roll Call Vote Results Across Legislatures

(with Matt Gabel and Cliff Carrubba)

One of the fundamental issues in representative democracies is how and whether voters can hold politicians accountable for their legislative behavior. Roll call votes, the public recording of how individual members of a legislature voted, is one important device by which interest groups and the public at large evaluate the performance of their representatives and hold them to account.

Surprisingly, despite its normative significance, we know very little about the institution of roll call votes; we know little about the conditions under which votes are roll called, who is responsible for deciding whether a vote is roll called, and what reasons they have for requesting a roll call. In a broad range of systems roll call votes are only a subset of all legislative votes cast. The sample properties of roll-call votes are critical to determining the quality of the inferences voters (and academics) can draw from observed legislative behavior. If these samples of votes are biased, a widely used and fundamental tool of accountability in democratic systems actually may be undermining, rather than enhancing, accountability.

These sample properties are also of enormous importance to academic studies of roll call votes. An individual legislator’s votes are analyzed relative to votes by all other members in order to construct that legislator’s “ideal point” (via NOMINATE, ideal, and other estimators). We put ideal points in order on a single dimension to identify pivotal members. We look at the distribution of those points to gauge the ideological polarization of members or of members grouped into parties. Measures of “likeness” are used to characterize parties and the party system. We interpret high levels of likeness to constitute cohesiveness, parties made up of homogeneous members, or to constitute discipline, the ability of leaders to impose uniformity on a diverse rank-and-file . Roll call votes allow us to characterize the policy space in terms of dimensionality.

The use of roll-call vote analyses to address these questions is ballooning, with scholars increasingly adopting them to study national, subnational (state and local), and international assemblies. Our concern about the potential sampling bias in roll-call votes is based on both empirical evidence and theory. Thus, in this project we will further develop our theoretical models of roll call usage and collect empirical data on the sample properties of roll call votes across legislatures in Latin America and Europe.

Carving Out Constituencies

In all multimember districts competing incumbents must “carve out” constituencies. Nowhere is this process more apparent than when an entire legislative chamber is elected in a single, nationwide, at-large district. Constituents can be geographically concentrated or dispersed. They can be motivated by ethnicity, religion, or ideology. Is it possible for parties and legislators to have constituencies defined in widely disparate terms? If so, how do they design cameral rules to accommodate their diverse electoral incentives? How do differences in constituencies affect legislator behavior in terms of bill sponsorship, committee assignments, and roll call votes. I have collected data relevant to this topic for Colombia, Slovakia, the Philippines, and the Netherlands. For example, . . . . .

József Kvarda received 20,330 preference votes in Slovakia’s 2002 parliamentary elections. In Slovakia’s flexible list system, these votes took him from 26th on the party’s list to 4th. The party only elected 20 members. So, his preference votes made the difference between being elected or not. Nearly sixty percent of his preference votes came from a single region, Dunajská Streda (in dark blue). Kvarda was a candidate for the SMK – the Hungarian Coalition Party. Dunajská Streda borders Hungary and has a large ethnic Hungarian population.

Dusan Caplovic received 42,489 preference votes. No single territory contributed more than 7% of his total, and Dunajská Streda contributed almost none (now in white). Caplovic is a member of the Smer – Direction-Third Way – which, after the 2006 elections, formed the governing coalition along with nationalist parties associated with anti-Hungarian sentiments.