Skip to content Skip to navigation

Crossing party lines: The effects of information on redistributive politics

Please note that prior to September 2017, the Center on Global Poverty and Development was known as the Stanford Center for International Development (SCID).

A Southern Sudanese woman displays her voter registration card as she waits in a long line to cast her vote.

<p>Voters display registration cards while waiting at the polls</p>

Jun 19 2015

Posted In:

Research Spotlights, SCID News

By Sam Zuckerman

What is the Issue?

In Sub-Saharan Africa, political loyalties sometimes divide along ethnic lines. Voters commonly make choices based on tribal allegiances, which gives party leaders little incentive to field high-quality candidates. The result: too often elected officials are incompetent and unaccountable, and public resources are spent for narrow partisan purposes. One reason voters may keep voting out of tribal loyalty is they may not have the facts needed to make informed choices. Voters may not be able to read and, even if they can, news sources providing information about candidates may not be at hand.

These entrenched political dynamics can lead to poor leadership and ineffective government, hampering economic development. But can improved access to information loosen the grip of tribal politics, opening a path to better governance? In particular, can competent public officials gain a following and attract significant electoral support from outside their ethnic group? In “Crossing Party Lines: The Effects of Information on Redistributive Politics,” Katherine Casey, Assistant Professor of Political Economy in the Stanford Graduate School of Business and SCID Faculty Affiliate, looks at the case of Sierra Leone—a nation with a history of bitter tribal conflict—to investigate whether party loyalties weaken as citizens learn more about the people running for office. Casey also examines whether political parties change how they spend campaign funds and distribute public resources when elections become more competitive. She finds reasons for optimism about African politics, noting that “deeply entrenched allegiances are not in fact immutable.”  


The West African republic of Sierra Leone provides a natural experiment in how access to information affects voter choice. Two ethnically based political parties are dominant. The Sierra Leone Peoples Party (SLPP) gets support from the Mende and other ethnic groups in the south, while the All Peoples Congress (APC) is strong among the Temne and other groups in the north. The country went through many years of one-party rule and civil war, which ended in 2002. Parliamentary races in 2007 were the first free elections after a harrowing conflict. The APC won 36 of 39 seats in the northern regions, and the SLPP and a splinter party took 24 of 25 southern seats, underscoring the strength of tribal ties.

Casey took advantage of several features of Sierra Leone politics to investigate whether different amounts of information influenced which parties voters supported in the 2007 parliamentary races and in 2008 local elections. Specifically, the country’s decentralized government system and its varying levels of radio coverage allowed her to examine three factors that affected how much voters knew about candidates:

  • Local versus national voting. Sierra Leone’s decentralized political system includes 19 local councils, and citizens tend to have more information about local than national politicians. Local officials live in or near the communities they represent, and people often know them personally or by reputation. Surveys in 2007 showed that some 37 percent of respondents could name their local councillor, while only 17 percent could identify their parliamentary representative.
  • Radio ownership. Radio broadcasts are the second most important source of political information in Sierra Leone after word of mouth, but only 48 percent of households own radios.  
  • Community versus national radio. Sierra Leone is divided into regions with no radio coverage, community stations only, or both national and community programming. Community stations cover local elections more heavily, and radio owners in areas that only have community stations may get more information about local politicians than do people in areas where listeners have a choice of programming.


Casey found that voters in local elections, radio owners, and residents of places with only community programming had access to more information about candidates than, respectively, voters in national elections, people without radios, and those who lived in areas with dual radio coverage. Voters in these higher-information categories displayed more political knowledge, as shown by their greater ability to name local politicians in surveys.

These higher-information voters were also more willing to break ethnic party bonds. Casey estimates that 74 percent of voters supported their ethnic party in local elections compared with 85 percent who did so in national elections. In addition, when several offices were contested, voters in local races were more likely to split their tickets. To a lesser degree, radio owners were also more likely to break with their ethnic party, both in areas with local and national programming, and in areas with community stations only.

In addition to analyzing information and party voting, Casey looked at whether candidate quality made a difference in voter willingness to cross ethnic party lines. Casey asked local councillors in several districts to rank their peers on effectiveness according to a 7-point scale. She found that the higher the performance ranking of an incumbent councillor, the more likely that candidate was to win votes from different tribes.

Casey also confirmed a well-known political science hypothesis that parties direct resources to areas that are most competitive, which in Sierra Leone means places where several ethnic group have settled. She estimates that a candidate for national office passes out US $18.30 less in a minimally competitive jurisdiction than in a maximally competitive jurisdiction. Similarly, district councils make an estimated $19,577 less in public investments in minimally competitive districts.

Finally, Casey connects these spending patterns with her findings on information. When greater information makes balloting more competitive, political parties smooth out their spending, allocating resources more evenly across districts. For example, Casey’s data suggests campaign spending responds only half as strongly to underlying ethnic party loyalties in local races, in which voters tend to have more information about candidates, than in national races.

Policy Implications

Casey’s research indicates that greater access to information weakens ethnic party loyalty and fosters more even distribution of campaign spending across jurisdictions. Moreover, voters with better information are more apt to cast ballots outside their ethnic parties to support high-quality candidates. Thus, providing better information about candidates could potentially make governments more effective and more responsive to citizen needs. “Information could break the low-accountability equilibrium in which citizens cast their votes blindly along partisan lines, creating little incentive for political parties to invest in candidate quality or provide resources to areas outside the most tightly contested jurisdictions,” Casey concludes.

Casey suggests that governments take steps to improve the quantity and quality of information available to voters. The example of Sierra Leone indicates that political decentralization offers one strategy for distributing political information more widely. “Decentralization brings government closer to the people. It enhances the amount of information available to citizens,” she notes. In addition, rising literacy makes information more accessible. “With the recent growth in mass media and communications technology across Africa, an optimistic implication of this result is that it may lead to a reduced reliance on ethnic politics in the future,” Casey concludes.