Joel Barkan’s early research on Kenya (Barkan Reference Barkan1976,1978,Reference Barkan1984; Barkan & Okumu Reference Barkan and Okumu1974,Reference Barkan, Okumu and Lawson1980) revealed an important relationship between constituents and Members of Parliament (MPs), structured by the country’s single-member district (SMD), first-past-the-post (FPTP) electoral system, which provided a unique logic to Kenya’s political system. The Barkan thesis, which he claimed applied to other African countries as well (Barkan Reference Barkan1995), was as follows. First, in large, predominantly agricultural, and newly independent polities, far-flung rural voters living at the periphery might have been expected to have weak attachments to the state. But MPs, Barkan found, played a crucial role in linking them to the center and integrating them into the new political system. Second, in an African version of Richard Fenno’s “home style” (1978), the electoral logic of the single-member constituency presented MPs with a choice between staying in the capital versus traveling to and spending significant amounts of time in the constituency. MPs, he found, devoted most of their time to working with local self-help projects and obtaining state resources for the constituency because this was the most rational strategy to secure reelection. All these factors, finally, conspired to produce an especially weak legislature because few MPs had the time or the inclination to use their institutional position to strengthen it or to check the power of the president.
While subsequent research took him in different directions, Barkan began to turn his attention back to the legislature in the first decade of the twenty-first century. Not coincidentally, this came on the heels of his spending several years in the mid-1990s as the first Regional Democracy and Governance Advisor in U.S. AID’s regional office for East and Southern Africa in Nairobi, an experience that made him increasingly aware of the importance of a strong legislature in overcoming Africa’s authoritarian legacy and fostering democracy. He was asked by the World Bank to lead a study of African legislatures, not so much as an institution of democracy (if only because the Bank was still very hesitant to take on this subject) but as an “institution of horizontal accountability” (Barkan et al. Reference Barkan, Levy and Kpundeh2004). This study was subsequently expanded with additional case studies and resulted in the publication of Legislative Power in Emerging African Democracies (2009), a study of legislatures in Benin, Ghana, Kenya, Nigeria, South Africa, and Uganda. This book produced at least two crucial insights. First, some (but not all) African legislatures were beginning to develop institutional capacities that had a positive impact on democratic development. Second, this development seemed to be linked, not so much to the extent of democratization, but to the presence (or absence) of a small but critical mass of MPs who focused on strengthening the institutional capacity of their legislatures and especially in countering executive dominance. Barkan described this group of MPs as “institutionalists.” However, these insights, derived as they were from country case studies, cried out for wider comparative analysis and systematic testing.
It was during Barkan’s field research on the South African Parliament in 2004 that the three of us first gathered in Cape Town for initial discussions about the possibility of a more systematic comparative study of a larger number of legislators. The result was the African Legislatures Project (ALP). Footnote 1 As a co-principal investigator, Barkan was intimately involved in all aspects of ALP, but his most important contribution was his insistence that we devote significant time to measuring MPs’ orientations toward the legislative roles that correspond to the different functions that legislatures perform. These orientations were collected through structured interviews of representative samples of legislators in each country. Footnote 2 In this article we examine a portion of this data in conjunction with Afrobarometer data on citizens’ views on their national legislatures in order to test, update, and extend some of Barkan’s arguments and early findings about the linkages between African citizens and their MPs. We end by briefly speculating on the implications of our analysis for legislative performance and democratic development in Africa.
Legislative Roles and Legislative Performance
Democratic legislatures perform four core functions: lawmaking, oversight of the executive, representation (which involves, literally, re-presenting constituency opinion within the legislature), and constituency service (i.e., what the MP does back in the district). No legislature—and no legislator—can carry out each of these functions to the same extent, or to the same level of effectiveness. For several reasons, these functions are inherently contradictory and exist in a permanent state of tension. One reason is simply a matter of time and space. Time spent in committee meetings, for instance, may reduce the number of days MPs can spend traveling back and forth to the constituency. Tensions also emanate from the fact that each function has a distinct animating logic and target constituencies with varying size, scope, and interests. Lawmaking and representation deal typically with broad issues affecting large, often national, constituencies; constituency service deals with the specific provision of material benefits to localized constituencies or to specific groups or constituents; and oversight involves the horizontal institutional relationship between the legislature and the executive aimed at holding the executive accountable for its actions. These tensions present legislators with difficult choices in terms of trade-offs between the time, effort, and organizational as well as personal resources they can devote to each.
Evidence
In May 1974 Barkan and John Okumu carried out a representative survey of thirteen rural constituencies in Kenya. After presenting the 3,828 respondents with a list of seven potential activities that MPs could perform, they found that an overwhelming proportion (71%) selected as most important those activities that facilitated linkages between the constituency and the central government. These included “tell[ing] government what people in the district want” (29%), “obtain[ing] projects and benefits for the district” (25%), and visiting the district “frequently” (11%). In contrast, just 5 percent prioritized parliamentary activities such as debating or passing bills. More important, and in contrast to much of the literature of the day (e.g., Lemarchand Reference Lemarchand1972; Lemarchand & Wegg Reference Lemarchand and Legg1972), just 6 percent of all respondents chose “help[ing] constituents with their personal problems (Barkan Reference Barkan1976).
By studying what MPs actually did, Barkan found that while they certainly never missed a chance to help individual constituents by donating food parcels, helping with school fees, or even dispensing out-of pocket cash contributions, they devoted the bulk of their time to organizing community self-help development projects to build water wells, clinics, and schools and obtaining funds for the constituency from central government (Barkan & Okumu Reference Barkan and Okumu1974, Reference Barkan, Okumu and Lawson1980). He also found that the percentage of respondents in each constituency who rated the MP as “very active” on the function they identified as most important was strongly and negatively related to the number of challengers to the incumbent who emerged in the subsequent KANU one-party election, and strongly and positively related to the overall proportion of the vote obtained by the incumbent in that election (Barkan Reference Barkan1976). Thus, MPs’ focus on linkage activities—especially constituency service, rather than institutional activities such as committee work—was a rational resolution to the kinds of role conflicts created by SMDs and one-party elections (Barkan Reference Barkan1976,Reference Barkan, Musolf and Smith1979).
One of our central concerns in the ALP was to test, update, and extend these findings. Our joint exploration of these issues began with a review of responses to an “open-ended” questionnaire fielded by Afrobarometer in 2002–2004 (Round 2), which asked respondents: “In your opinion, what are the most important responsibilities of a Member of Parliament?” Respondents could offer up to three answers. The aggregated responses closely resembled Barkan’s findings of three decades earlier. Across sixteen countries, the most frequent type of response related to constituency service. And as in Barkan’s earlier study, people focused on the delivery of material collective goods, with responses ranging from general requests for “development” assistance (27%), to more specific requests for “infrastructure” development (26%), “employment” opportunities (15%), “education” (13%), “health” services (12%), and agricultural development or assistance to the poor (>10%). Only 8 percent said the MP should provide personal services, or private goods such as “spending his own money” in the constituency, or providing loans or food parcels to individuals (though, perhaps ironically, this figure rose to 27% in Kenya). The second most frequent type of broad response referred to aspects of representation. One quarter (26%) said MPs should “represent the people” and another 13 percent said they should “listen to the people.” Only a small minority referred to issues of oversight (7%), or lawmaking (with 6% indicating the need for MPS to “make the laws” and 4% mentioning their “debating” or “discussing and solving” problems).
Based on both the responses to this open-ended questionnaire and our interest in how MPs negotiate the inherent tension between the four legislative functions, we designed a closed-ended, forced-choice question for Afrobarometer Round 4 (2008–2009), which asked: “Representatives to the National Assembly have different responsibilities. Which of the following do you think is the most important responsibility of your representative to the National Assembly?” When presented with these options, the modal response among the citizens of the seventeen countries studied by ALP shifted from the delivery of material goods to the provision of representation. Just under one-half of all respondents chose “listen to constituents and represent their needs” (45%). The next largest group chose constituency service, defined as “deliver[ing] jobs and development to your constituency” (31%). But as with the Round 2 open-ended question, the realms of lawmaking (“make laws for the good of the country”) and executive oversight (“monitor the President and his government”) were chosen by substantially smaller percentages of respondents—15 percent and 6 percent, respectively (see figure 1).
![](https://static.cambridge.org/binary/version/id/urn:cambridge.org:id:binary-alt:20161220065034-47423-mediumThumb-S0002020616000834_fig1g.jpg?pub-status=live)
Figure 1. Citizen Role Expectations of MPs
Source: Afrobarometer Round 4, n = 20,339.
These data confirmed two of Barkan’s earlier findings: namely that African citizens, in general, expect their MPs to focus more on local issues than on national legislative or institutional issues, and that specifically with respect to local issues, they expect their MPs to focus on the delivery of development for the constituency as a whole rather than on the provision of personal assistance. Yet the modal Afrobarometer respondent of the twenty-first century is significantly more likely than the rural Kenyans of the 1960s to rank representation ahead of constituency services. We found this change surprising, not only from the perspective of Barkan’s earlier Kenya research, but also because the scholarship on African politics focuses almost exclusively on the downward flow of state resources to local constituencies, and in particular, the provision of private goods.
But what about MPs? How do they define their roles? Our ALP survey of legislators across the same seventeen countries asked MPs: “In your opinion, which of these following jobs is the most important part of being an MP?” Because MPs are likely to have a more cognitively sophisticated view of their job than ordinary citizens do, we offered a more fine-grained set of responses. As the results in figure 2 show, African MPs, in general, see constituency service as the most important part of their job (35%). Again, in contrast to the widely held view in Africanist scholarship, MPs view constituency service in terms of delivering collective goods (“bringing development to the constituency” and “solicit[ing] funds for the constituency” [31%]) rather than the provision of personal assistance (“help[ing] with personal problems” [4%]). With respect to the other three roles, 35 percent of African MPs selected lawmaking as the most important part of their job (“debating bills and passing laws” and “making public policy by writing laws”), 19 percent chose representation (“representing constituency views in Parliament), and 10 percent picked oversight (“overseeing the executive”).
![](https://static.cambridge.org/binary/version/id/urn:cambridge.org:id:binary-alt:20161220065034-02287-mediumThumb-S0002020616000834_fig2g.jpg?pub-status=live)
Figure 2. MP Role Orientations
Source: African Legislatures Project MP Survey, n = 838.
Not only do African MPs see constituency service as the most important part of their job, relative to the other three legislative roles, they also see it as the most gratifying. In a follow-up question (not shown), the MPs were asked, “For you personally, which role brings you the most satisfaction?” In the responses to this question, each of the relevant options related to constituency service (delivering development, helping solve personal problems, raising funds) was selected by significantly larger proportions of respondents, while lawmaking and oversight were selected by significantly smaller proportions.
Then, in order to compare MP and citizen responses directly, we collapsed the more specific responses offered to MPs in the ALP survey into the four legislative functions offered in the Afrobarometer question. The results, displayed in figure 3, show that MPs’ role orientations closely resemble mass expectations in two ways. Neither group emphasized oversight (10% of MPs compared to 6% of citizens). And roughly one-third of both MPs (35%) and citizens (31%) pointed to constituency service. But there are also sharp differences. MPs (33%) are far more likely than citizens (15%) to view lawmaking as an important function. Footnote 3 The most significant difference, however, occurs with respect to representation. Citizens (45%) were more than twice as likely as MPs (19%) to select representation as the most important function.
![](https://static.cambridge.org/binary/version/id/urn:cambridge.org:id:binary-alt:20161220065034-19055-mediumThumb-S0002020616000834_fig3g.jpg?pub-status=live)
Figure 3. Citizen Role Expectations and MP Role Orientations Compared
Source: Afrobarometer Round 4, n = 20,339. African Legislatures Project MP Survey, n = 838.
In general, then, MPs do not seem to grasp fully the extent of public demands for representation. This misperception took on even more interesting contours once we disaggregated the data by country (see figures 4 and 5). The cross-national variation in public role expectations suggests that Barkan’s intuition of the important role of the electoral system was prescient, though for the wrong reason. Of the eleven countries where representation is selected by at least 40 percent of citizens, nine elect legislators from SMDs with an FPTP formula. The other two use mixed types of systems that also emphasize constituency effects. Senegal’s mixed parallel system allocates constituency seats by a majoritarian formula, which typically trumps the effects of the national seats allocated by a proportional representation (PR) formula (Mozaffar & Vengroff Reference Mozaffar and Vengroff2002). Mali’s system of multimember districts (MMDs) with majoritarian formulas also heavily reinforces constituency effects (Vengroff Reference Vengroff1993). In contrast, five of the six countries in which less than 40 percent of citizens select representation use PR formulas for allocating all seats, while the sixth (Lesotho) combines SMDs with FPTP with a PR formula for national seat allocation. In these countries, public emphasis on constituency service varies across a much smaller range, from 41 percent in Lesotho to 16 percent in Senegal, and appears to be much more weakly related to the type of electoral system. Thus, where Barkan inferred that SMDs create public demands for bringing goods and services downward from the central government (constituency service), the cross-national data strongly suggest instead that they amplify the desire for MPs to take their interests and opinions upward to central government (representation).
![](https://static.cambridge.org/binary/version/id/urn:cambridge.org:id:binary-alt:20161220065034-57903-mediumThumb-S0002020616000834_fig4g.jpg?pub-status=live)
Figure 4. Citizen Role Expectations of MP Linkage Activities by Country
Source: Afrobarometer Round 4, n = 20,339.
![](https://static.cambridge.org/binary/version/id/urn:cambridge.org:id:binary-alt:20161220065034-00592-mediumThumb-S0002020616000834_fig5g.jpg?pub-status=live)
Figure 5. MP Linkage Role Orientations by Country
Source: African Legislatures Project MP Survey, n = 838.
For MPs, in sharp contrast to the citizens, it is the emphasis on constituency service that varies widely, from 76 percent in Tanzania to 13 percent in Benin, while the variation in representation is relatively smaller, ranging from 46 percent in Namibia to just 2 percent in Mali. Moreover, the variation in MP orientations as “constituency servants” turns sharply along lines of electoral systems: MPs who are elected from single-member districts are far more likely to see themselves as “constituency servants.” In the eight countries where at least 40 percent of MPs see themselves as “constituency servants,” seven use SMDs with an FPTP formula. Of the seven countries where fewer than one-quarter of MPs see themselves this way, six use at least some form of PR formula.
A comparison of figures 4 and 5 suggests that MP misperceptions of citizen preferences are greatest in countries that have single-member districts, and figure 6 demonstrates that this is precisely the case. We constructed a measure of the gap between MP and citizen emphases on representation for each country by subtracting the country percentage of MPs who cite representation from the country percentage for citizens. At one end of the spectrum, citizens of Burkina Faso are actually less likely than MPs to choose representation (a –14 point difference). In Uganda and Tanzania, in contrast, the gap between citizens’ demands for representation and MP role emphasis is +52 and +51 points, respectively. But there is a clear influence of the electoral system at work here. The most pronounced gaps (greater than 30 percentage points) exist in the three Anglophone countries of East Africa, and in Nigeria in West Africa. However, the size of the gap is only slightly less pronounced (10–29 percentage points) in Mali, Ghana, Botswana, Lesotho, and Senegal. All of these countries, with the exception of Mali and Senegal, employ plurality formulas to elect MPs for single-member districts. Footnote 4 As Barkan (Reference Barkan, Musolf and Smith1979, 1980, Reference Barkan1995) argued, and the comparative literature substantiates (see, e.g., Cox Reference Cox1997; Reynolds Reference Reynolds1999), single-member districts create greater incentives than proportional representation systems to cultivate the personal linkages between voters and MPs that underpin a localist emphasis on constituency service and representation. Footnote 5 The smallest gap (less than 10 percentage points) exists in five countries, all of which use proportional representation formulas to elect MPs for large multi-member districts. Footnote 6 These institutional designs, combined with the strong centralized control of party nominations and the legislative agenda exercised by dominant parties, lead to greater emphasis on broad national issues over local interests.
![](https://static.cambridge.org/binary/version/id/urn:cambridge.org:id:binary-alt:20161220065034-29438-mediumThumb-S0002020616000834_fig6g.jpg?pub-status=live)
Figure 6. Citizen–MP Gaps in Emphasis on Representation by Country
The number for each country is the percentage point difference between the percentage of Afrobarometer respondents minus the percentage of MPs from each country who answered that listening to constituents and representing their needs is the most important job for MPs to perform.
Sources: African Legislature Project MP Survey (n = 838) and Afrobarometer, Round 4 (n = 20,339).
Thus MPs elected from single-member districts seem to understand that their citizens want MPs to have a “localist” focus on the constituency itself. However, they misperceive what it is that citizens want them to focus on while they are there. Citizens want MPs to listen, while MPs think they have to deliver welfare. Working through this data, Barkan was delighted by the crude, but apt, analogy between the MP and the amorous male suitor who showers the object of his affection with jewels and expensive clothes, while the woman (i.e., the citizen) secretly confides to her friends that “I just want someone to listen to me.”
Role Orientations and Constituency Development Funds
None of this is to say that Joel Barkan was oblivious to, or resisted, these new facts and the evolving interpretation. On the contrary, he was fascinated by the rich implications of the ALP survey. He rubbed his hands with glee at the prospects of describing these findings and exploring their implications for our understanding of democratization and governance in Africa in the “great book” that would eventually lay out the findings of the ALP project. Tragically, that book did not materialize before he died.
However, in the very last piece we wrote together—for a volume edited by Mark Baskin and Michael Mezey on constituency development funds—we began to tease out these findings and at least one apparent consequence (Barkan & Mattes Reference Barkan, Mattes, Baskin and Mezey2014). Since 2002, constituency development funds (CDFs) have been established in nine African countries, and another two countries have created “approximations” of CDFs in that they address the perceived need by members of the legislature for budgeted funds to spend on the development of the districts they represent. Footnote 7 Thus just under one-quarter of the forty-eight countries in sub-Saharan Africa have adopted some type of CDF.
The absolute size of MP misperceptions of citizens’ expectations of representation is positively and strongly related to the recent growth of these controversial funds. That is, the greater the gap between citizens and MPs, the more likely the country in question has established a constituency development fund (see table 1). CDFs exist in those countries with a very high gap, or level of misperception, about what citizens want from their MPs (i.e., +30 points or more). Indeed, the relationship is striking (Tau b = .830***).
Table 1. CDF Status by Citizen–MP Gaps in Emphasis on Representation
![](https://static.cambridge.org/binary/version/id/urn:cambridge.org:id:binary-alt:20161220065034-78409-mediumThumb-S0002020616000834_tab1.jpg?pub-status=live)
Kendall’s Tau b=.830*** / Pearson’s r=.792***, N=17.
The number for each country is the percentage point difference between the percentage of Afrobarometer respondents minus the percentage of MPs from each country who answered that listening to constituents and representing their needs is the most important job for MPs to perform.
Sources: African Legislature Project MP Survey (n=838) and Afrobarometer, Round 4 (n=20,339).
MPs rightly perceive the need to maintain close contact with their constituents, but wrongly believe that their constituents look to them mainly for “pork.” Instead, African constituents’ primary expectations of their MP is that they regularly visit the district to learn what is on their minds, and to then quite literally “re-present,” or transmit, these views back to the central government via the legislature. In other words, while citizens desire stronger representation of their needs at the center, MPs respond by delivering services and favors at the periphery (i.e., the district), thinking mistakenly that the CDFs are the answer to what the public wants. One indication of the validity of this mismatch argument is the finding that reelection rates in the countries that have established CDFs do not appear to have risen, compared to the period before there were CDFs (see table 1).
Conclusion
Joel Barkan’s pioneering survey of rural Kenyans’ expectations of Members of Parliament showed that most people emphasized MPs linkage role (carrying views upward to the capital, as well as bringing goods down from national government to the constituency) rather than their institutional roles: for instance, writing laws or overseeing the executive. But in contrast to much of the literature of his day (Lemarchand Reference Lemarchand1972; Lemarchand & Wegg Reference Lemarchand and Legg1972; Wanthekon Reference Wantchekon2003; Lindberg Reference Lindberg2003, Reference Lindberg2010), as well as that of several decades since, he argued that what Kenyans’ wanted was not personal or private goods, but collective goods designed to develop the larger constituency. Based on this perception, he concluded that MPs’ emphasis on constituency service and pork-barrel politics was a rational reaction to the need to be reelected.
Afrobarometer and ALP data, collected three decades later, broadly confirm the fundamental insight that both MPs and citizens from countries with electoral systems based on single-member districts with plurality rules converge around a strong “localist” focus on the constituency through the linkage functions of representation and constituency service. However, the Afrobarometer data also demonstrate that the demand for representation outstrips the demand for constituency service, especially in countries that use SMD and plurality formulas. Moreover, the ALP data reveal that while MPs in SMD districts broadly understand voters’ demand that they pay close attention to the constituency, they misinterpret this as a demand for material goods and development and systematically underappreciate the public’s demand for representation. In this sense, MPs’ focus on the downward distribution of state resources is not, as Barkan thought, entirely rational, since it is based on an inaccurate reading of what voters want, and hence what gets them reelected. Finally, while Barkan’s initial argument holds up with respect to countries whose electoral systems are based on SMDs with FPTP formulas, its varied manifestations in countries with other types of electoral systems also suggest that there are much more complex interactions of country-level effects of democratization and institutional choice which demand further investigation.
Future research needs to take the (largely unrequited) demand for representation more seriously. In the very last piece Joel Barkan wrote with us (Barkan & Mattes Reference Barkan, Mattes, Baskin and Mezey2014), we demonstrated that this misperception had at least one major consequence by distinguishing between the countries whose legislatures have created controversial constituency development funds and those that have not. We believe that these misperceptions have other consequences as well, and should be explored in future research. For instance, do MPs’ inaccurate readings of voter expectations prompt them to spend considerable amounts of time in the constituency at the expense of work in plenary sessions and committees? Perhaps if African MPs realized that the most important citizen demand was for representation, they might be better able to square this circle and represent their constituencies by spending more time in the legislature and engaging more deeply in lawmaking and oversight.
Acknowledgments
Fieldwork for the African Legislatures Project was supported by grants from the World Bank, the United Kingdom Department for International Development, the United States Agency for International Development, the Vice Chancellor’s Fund at the University of Cape Town, the Heinrich Böll Stiftung, PACT/Zambia, and the Calouste Gulbenkian Foundation. Other research activities were supported by the Harry Oppenheimer Institute, the Centre for Higher Education Transformation, the South African National Research Foundation, and the Andrew Mellon Foundation. Shaheen Mozaffar thanks the Center for Democratic Governance and Leadership at Bridgewater State University for research support.