A Local Islamist Advantage? Toward a Theory of National and Local Service Provision in Algeria

Lindsay J. Benstead, Portland State University

In this memo, I take a preliminary step toward filling this gap by leveraging original data from a survey of 780 Algerian citizens I conducted in 2006 with Ellen Lust. This unique dataset allows me to test whether the service provision practices of Islamist parties differ from non-Islamist members and across levels of government. Here, the emphasis in on linkages between elected legislators and citizens, through which citizens asks for help with personal and community problems, or express opinions.

I find that Islamist parties do indeed govern differently; more often than other parties, they receive requests from citizens with whom they have no personal connection and citizens who are more likely to be marginalized from patronage networks with political elites – operationalized here as women (Benstead 2016). Moreover, I find that the extent to which Islamists serve these citizens increases at the local level.

These findings have several implications for existing literature, and for addressing why Islamist movements and parties are often electorally successful. First, they extend existing work on Islamist parties’ strategy to reach citizens outside established patronage networks by suggesting that these efforts are more effective at the local than the national level. Second, they highlight variation in governance patterns across two Islamist parties – one allied with the government and the other outside the governing coalition – and suggest that cooperative relationships with the regime may be needed to reach disengaged citizens with services. Third, they underscore the need to examine why local politics differ from the regional and national levels in authoritarian regimes.

Along with other contributions in this volume, the findings suggest that authoritarian states’ reliance on patron-client relationships leaves them vulnerable to challenges by opposition movements when they are able to mobilize support among marginalized groups.

Theoretical Framework

Extant literature suggests that Islamist parties govern differently from other parties, bypassing existing patronage networks through institutionalization of service provision and direct contact with citizens. This strategy impacts both their electoral success (Masoud 2014), as well as the extent to which citizens who are not tapped into existing state-society patronage networks, including women, benefit from opportunities to request clientelistic services (Blaydes and Tartouty 2009; Meyersson 2014; Blaydes 2014; Abdel-Samad and Benstead 2016; Benstead 2016).

Extending this literature, I investigate several hypotheses:

At all levels of elected office (national, provincial, and municipal councils), Islamist members will reach out to citizens who are outside existing patronage networks more than non-Islamist officials, such that:

H1: Citizens who approach Islamist members are more likely than those who contact non-Islamist members for help to say they have no personal connection with the legislator.

H2: Female citizens are more likely to approach Islamist members than non-Islamist members.

 Islamist parties’ strategy to institutionalize service provision, engage in redistributive politics, and establish direct contact with citizens in their district (Cammett and Jones Luong 2014; Arat 2005; Abdel-Samad and Benstead 2016) will also impact their governance practices differently at the local than at the national level. Specifically, the extent to which Islamists’ serve citizens with whom they have no previous connection will be greater at the local level. This “local advantage” results from greater social embeddedness – that is, the greater ease maintaining contact with citizens through offices and social and religious institutions at the local than the national constituency.

H3: The gap between Islamist and non-Islamist member in governance practices – serving citizens with whom legislators have no prior connection and women – will be greater at the local than at the national level.

Legislative Elections in Algeria

Algeria holds elections by universal suffrage to five year terms for its National Popular Assembly (L’Assemblée Populaire Nationale, APN), which had 389 seats as of the 2002-2007 mandate covered by the survey, as well as to its 48 regional assemblies (Popular Assembly of the Wilaya/L’Assemblée Populaire du Wilaya, APW) and 1541 municipal assemblies (Popular Assembly of the Commune/L’Assemblée Populaire Communale, APCs). The upper house of parliament is comprised of 48 seats appointed by the president and 144 seats indirectly elected by the APWs. Elections are held by closed-list proportional representation in multi-member districts. For elections to the APN, the electoral districts are the country’s 48 provinces (wilayat; Image 1). During the 2002-2007 APN mandate, districts had between 4 and 32 seats and there were 8 overseas seats.

Image 1. Algeria’s 48 wilayat


The Constituent Survey was conducted in eight wilayat: (38) Tissemsilt, (16) Algiers, (42) Tipaza, (19) Sétif, (23) Annaba, (5) Batna, (39) El-Oued, and (1) Adrar (Wikipedia 2017).

Semi-competitive elections and executive dominance

Algeria has a single party dominant regime in which military and intelligence officials and civilian elites, including the President and the cabinet, compete to shape major political and economic policy (Entelis 2011). While President Bouteflika succeeded in consolidating unprecedented power behind the presidency, powerful, unelected military elites and intelligence services – called “le pouvoir” or “les décideurs” – still wield influence behind the state’s civilian political institutions. These individuals are at the pinnacle of Algeria’s patronage system, an informal system of largely personalistic relationships through which rents from the state’s vast oil reserves are distributed (Volpi 2013).

Algeria’s political system is highly centralized, with electoral structuring, institutional design, and clientelism ensuring regime continuation through executive dominance at the national, regional, and local levels. Since 1997, national, regional, and local levels have been structured to ensure that the two regime parties – National Liberation Front (FLN) and National Rally for Democracy (RND) – make up a majority in all legislative bodies, thereby guaranteeing that the executive branch wields law-making power. This executive dominance extends to the regional and local levels.

Without autonomy, national, regional, and local assemblies potentially offer their members opportunities to benefit personally, or to utilize state patronage networks to help voters with personal problems – and enhance future personal or party electoral prospects. To a large extent, this requires maintaining cooperative relationships with regime patrons and gaining party representation in the cabinet and ministries, who can help legislators resolve citizen requests.

Data and methods

To investigate governance practices across Islamist and no-Islamist parties and levels of government, I use a household survey of 780 Algerians conducted face-to-face by a local team from August-September 2006. The response rate was 73 percent. At the time of the survey, members of the APN, APWs, and APCs had been in office since 2002. Since 2002, national/regional/local elections were held in 2007, 2012, and 2017. No significant changes to the electoral code or prerogatives of the assemblies occurred during this time.

The survey was conducted among a nationally-representative sample in 8 wilayat and 36 communes selected through multi-stage probabilistic sampling with the electoral district (wilayat) as Primary Sampling Units (PSUs) and communes as the Secondary Sampling Units (SSUs). The eight communal seats were self-representing.


To measure linkages, respondents were asked if they had a parliamentarian during the current mandate for help with a personal or community problem, or to express an opinion. This includes both successful as well as unsuccessful requests. Respondents were then asked a series of questions about the most recent time, including the member’s party. To measure personalistic and rationalized linkage types, respondents were asked why they contacted the parliamentarian. Response choices included “just because the deputy was elected” (“rationalized” ties/no connection), or “member of your tribe,” “friend,” “another connection such as business,” and “a member of the party you belong to” (personal ties). Respondents could answer more than one type (Benstead 2017). This battery was repeated for the APW and APC. (See Appendix for questions wording).

Islamist and non-Islamist parties

When the survey was conducted in 2006, the parties held a similar proportion of seats at the national, regional, and local levels. In the 389-seat APN elected May 30, 2002, pro-regime parties, the FLN and the RND has 199 and 47 seats, respectively. Two Islamist parties, el-Islah and MSP, held 43 and 38 seats. The socialist Workers’ Party (PT) had 21 seats, while independents and smaller parties held 41 seats. The Constitutional Rally for Democracy (RCD) and Democratic and Social Movement (MDS) boycotted national, regional, and local elections.

At the time of the survey, one Islamist party, the Movement for Society and Peace (MSP, formerly HAMAS), was in a coalition with the governing FLN and RND parties and had representation in the cabinet – while a second, Islah, did not join the coalition and was weaken by regime interference and had split it into two groups within the parliament (Benstead 2008).

These same parties were similarly represented in the APW and APC after October 10, 2002 elections. Partial elections were later held for 60 communes in the Kabylie in 2003 due to electoral violence which closed the polls in 2002.

The FLN, which was the only party fielding candidates in all districts in 2007 (Bustos, 2012), won 2.5 million votes, RND (1.2 million), Islah (960,000), MSP (600,000), PT (500,000), and 19 more parties receiving 270,000 or fewer votes each (Hamidouche 2012). APW results were similar to those of the APCs.

Member-citizen linkages 

I present bivariate analyses using two-tailed Chi-squared tests of independence to investigate member-citizen linkage across parties and levels of government. I find that, while relatively fewer citizens have contacted a member of the APN or an APW for help with a problem or to express an opinion, many more contacted a local councillor. About 11 percent had contacted an APN or APW member and 25 percent of citizens have asked for help from a local councillor (Figure 1).

Figure 1. Percentage of citizens who have contacted a member, by wilaya and level

Differences significant across wilayat for APWs and APCs (p<.001). N=780

Contact rates for APN members shown in light grey in Figure 1 vary across districts. The rate is highest in Batna, where 19 percent have asked for help from a parliamentarian and lowest in El Oued, where only 2 percent have asked for help. Batna is part of the BTS triangle (Batna-Tébessa-Souk Ahras) – a region with dense state-society patronage networks, due to the large number of public officials and military personnel originating from the region since the 1980s (Hachemaoui 2013, p. 21). Contact rates for deputies from the APW (shown in black in Figure 1) also vary across wilayat, and the difference is statistically significant (p<.001).

When it comes to asking for help from APW members, Batna also stands out as the wilayat with the highest rates, suggesting similarities in electoral politics across the national and provincial levels. There, 21 percent have asked for help in Batna, while only 2 percent in El Oued have. That Batna has the highest rates for both the APN as well as the APW suggests that similar political mechanisms underlie national and regional electoral politics and shed light on electoral politics in a highly centralized authoritarian state. APN and APW elections are likely more tightly controlled since the incumbents must ensure a majority for the regime parties, the FLN and RND in the APN and the APWs, which indirectly elects 2/3 of seats in the Senate, a key institution guaranteeing regime stability.

There are also important differences in the type of member-citizen linkages across the national and local levels. Many more citizens have contacted members of APCs for help, with rates varying significantly across provinces (p<.001) and municipalities (p<05). While Batna has the highest rates of legislator-citizen linkages for the APN and APW, its rate for APC members is 19 percent, below 25 percent average for local councils. El Oued has among the lowest rates of legislator-citizen linkages for all three legislative levels. There, only 7 percent of citizens have contacted a member of a local council for help with a problem. The highest levels of local service provision are in Tissemsilt, where 39 percent have asked for help, and lowest in Tipaza, where 7 percent have done so. As noted, this suggests that the mechanisms underlying local political recruitment differ from the regional and national levels in ways that are unexplored in the literature.

There are other differences across the levels as well. Citizens infrequently contact a deputy at the APN without having a connection, but rationalized linkages (“just because they were elected”) are more common for the APW and APC. While only 24 percent of attempts to contact APN members were rationalized, the proportion increases to 28 percent for APW and 57 percent for APC members (Figure 2).

Figure 2. Percentage of citizens who contacted official with whom they have no personal connection, by party and level

N=30 (APN)/N=44 (APW)/N=69 (APC). Where the figure indicates “0%”, this indicates that 0% of the reported contacts for that party were rationalized (i.e., 100% were personalized). No citizens reported contacting local councillors (APC members) from Islah or PT, while no citizens reported contacting parliamentarians (APN members) from MSP or PT. 

Do Islamists reach more citizens with whom they have no connection?

To what extent do Islamists reach those who are marginalized from existing patron-client relations, and do they do so more than do non-Islamists? As shown in Figure 2, I find that Islamist parties do govern differently; they serve more citizens with whom they have no connection, and the gap increases at the local level. When considering the FLN, 14 percent of contacts at the national level, 39 percent at the regional level, and 54 percent at the local level were characterized by “no connection.” For the RND, 0 percent at the national level, 17 percent at the regional level, and 28 percent at the local level were rationalized.

In contrast, at the national level, there was no data recording Islamist contacts, yet 100 percent of contacts at regional level and 80 percent at the local level involving the MSP took place involved a citizen who had no connection with the legislator, in partial support of H1 and H3.

At the same time, there are differences across the two Islamists parties in this respect. All of the reported contacts of Islah deputies were via a personal connection. One possible reason for this difference is that the MSP, due to its coalition with the government, has better connections within the state and greater freedom to implement an Islamist mandate by reaching out more broadly and directly within their constituencies (Abdel-Samad and Benstead 2016). Islah may simply have had fewer resources and connections to effectively reach out to citizens. Islamist parties, like other movements and parties, must adapt their strategies, based on the constraints and opportunities they encounter (see Margot Dazey’s memo in this volume). 

Do Islamist parties serve more women?

As shown by Figure 3, both Islamist parties are more likely than non-Islamist parties to contact women, in support of H2. Due to the small number of respondents, differences are not statistically significant, but they are large. At the APN, 67 percent of those who contacted Islah were women, compared to 25 percent who contacted deputies of the FLN and 33 percent who contact the RND.

Figure 3. Percentage of female citizens contacting member, by party and government level

N=30 (APN)/N=44 (APW)/N=69 (APC). No citizens reported contacting local councillors (APC members) from Islah or PT.

Women were more likely to contact Islamist legislator than those from other parties at the regional level as well, in support for H2. 100 percent of those who reported contacting Islah for a services were women. Local officials from the Islamist MSP party, like those from the RND, serve 50 percent women, in partial support for H2.

Conclusions and implications

Three conclusions emerge. First, according to citizen reports, Islamist members serve more female citizens than do non-Islamist members. Second, more than non-Islamists, Islamists are in contact with citizens who report having no personal connection with them. And, this gap in rationalized connections across Islamist and non-Islamist parties becomes larger at the local level – evidence for a “local advantage.” At the same time, these relationships were more pronounced for the MSP, an Islamist party in coalition with the government, than Islah, which did not join the coalition. This difference across Islamist parties should be examined in future research.

These findings have several implications for existing literature and speak to the themes of this volume, which emphasize the ways that weak institutions and reliance on patron-client relations in authoritarian regimes create opportunities for opposition parties to reach the disengaged. When Islamist movements mobilize populations outside existing state patronage structures, they are more likely to succeed electorally (Masoud 2014). Following this same logic, rebel groups also take advantage of these gaps in state capacity and legitimacy to build a nascent state structure and control territory, as IS has in Iraq and Syria (see Mara Revkin’s piece in this volume).

First, Islamists’ greater use of rationalized connections, particularly at the local level, indicates that they may utilize social and political institutions relatively more effectively at the local than the national level. This suggests that the extent to which rationalized linkages will be dominant in a given case depends not only on institutional settings and rents, but also on the presence of Islamist members and the level of government (Benstead 2017). Further research should explore this “local advantage” with additional qualitative and quantitative data.

Second, at all levels, Islamist legislators reach out to those who are marginalized from patronage networks and women more than do non-Islamists. For literature on Islamist parties and gender and governance, this finding supports existing literature by showing that Islamist parties bypass existing patronage networks and serve citizens broadly. This also has a positive impact on women’s access to services not only at the national level, but also at the local level.

There are broader implications as well for theories of authoritarian politics. Local linkages differ from national-level linkages, suggesting that different mechanisms drive service provision at the local and national levels. For instance, in Algeria, the types of parliamentarian-citizen connections – largely personalistic, non-tribal connections – that dominate at the national level are less common at the regional level, and even less common at the local level. Local councillors from all parties are more likely than those at the national level to be in contact with citizens with whom they have no direct connections. Existing theories of electoral and legislative politics in authoritarian regimes should be extended to explain this variation across levels of government and its implications for authoritarian regime persistence and Islamist success.

These preliminary findings shed light on the ways that local and Islamist politics – and the intersection of the two – differ from traditional understandings of national level legislative politics. Because of the small N in this survey (and the infrequency of the event being measured), more survey data are needed to measure member-citizen linkages that questions asking about state-citizen linkages and include the party and use a large N sampling approach that is representative at the communal level, similar to the methodology employed by the Local Governance Performance Index (LGPI) conducted the Program on Governance and Local Development (GLD 2017).

Future work should examine differences in Islamists strategies across urban and rural areas. This could better inform efforts to understand why the PJD has been unsuccessful in rural areas in national elections in Morocco (see Yasmina Abouzzohour’s memo) and why the AKP party has had varying levels of success across localities in municipal elections in Turkey (see the contribution of Melissa Marschall, Marwa Shalaby, and Saadet Konak Unal; Evren Celik Wiltse). Following the approach of Brooke and Ketchley (in this volume) and Margot Dazey writes (in this volume), it is critical to interrogate the relationship between the local and national officials in Islamist movements. Research should explore how networks between local and national officials affect the successfulness of efforts to provide services to citizens. These and other extensions will help to better explain how and why local Islamist and other movements reach citizens who are outside established patronage networks and fill spaces left by authoritarian states.


Question wording.

Local Councils (APCs)

Have you or any member of your household living here tried either successfully or unsuccessfully to contact an elected official of your local council to seek help with a personal problem, to seek help with a social or economic problem your community is facing, or to express an opinion? No(=1), Yes(=1), Yes we tried, but could not contact(=1).

(Now I am going to ask just about the most recent request you made to an elected official in your local council). Which party this member represent?

Which of the following best describes why you tried to contact this official specifically? This official was a member of your family or tribe? A member of the party in which you are/were a militant? A friend or someone you knew through a previous connection (e.g. business, association, etc)? Or was it simply because this official was elected to your local council and you had no previous connection with him/her? [Check all that apply]


Regional Councils (APWs)

As above (“your regional council“)

National Popular Assembly (APN)

As above (“the Chamber of Representatives“)



Abdel Samad, Mounah, and Lindsay J. Benstead. 2016. “Why Does Electing Women and Islamist Parties Reduce the Gender Gap in Service Provision?” Paper presented at the “After the Uprisings: Public Opinion, Gender, and Conflict in the Middle East” workshop, Kansas State University, May 5.

Arat, Yesim. 2005. Rethinking Islam and Liberal Democracy: Islamist Women in Turkish  Politics. Albany: State University of New York Press.

Benstead, Lindsay J. 2008. “Does Casework Build Support for a Strong Parliament? Legislative Representation and Public Opinion in Morocco and Algeria.” (PhD diss., University of Michigan, 2008). Retrieved from ProQuest Dissertations Publishing, (Accession No. 20083343007).

———. 2016. “Why Quotas Are Needed to Improve Women’s Access to Services in Clientelistic Regimes.” Governance 29 (2): 185-205. doi:10.1111/gove.12162. First published online 2015. doi: 10.1080/21565503.2015.1084338.

———. 2017. “Explaining Parliamentarian-Citizen Linkage Type and Satisfaction across Authoritarian regimes.” Unpublished paper.

Blaydes, Lisa. 2014. “How Does Islamist Local Governance Affect the Lives of Women?” Governance 27 (3): 489-509.

Blaydes, Lisa, and Safinaz El Tarouty. 2009. “Women’s Electoral Participation in Egypt: The Implications of Gender for Voter Recruitment and Mobilization.” The Middle East Journal 63 (3): 364–380.

Bustos, Rafael. 2012. “Local Elections in Algeria Against a Backdrop of Disinterest and Party Crises.”Observatorio Politico y Electoral del Mundo Arabe y Musulman. Last modified December 12.


Cammett, Melanie, and Pauline Jones Luong. 2014. “Is There an Islamist Political Advantage?” Annual Review of Political Science 17: 187-206.

Entelis, John P. 2011. “Algeria, Revolutionary In Name Only.” Foreign Policy, September 7. http://foreignpolicy.com/2011/09/07/algeria-revolutionary-in-name-only/

Gandhi, Jennifer, and Adam Przeworski. 2006. “Cooperation, Cooptation, and Rebellion under Dictatorships.” Economics & Politics 18 (1): 1-26.

———. 2007. “Authoritarian Institutions and the Survival of Autocrats.” Comparative Political Studies 40 (1): 1279-1301.

Hamidouche, Younes. 2002. “Seuls trois partis approchent ou dépassent le cap du million de Voix. Nouvelle carte politique, mêmes réflexes électoraux.” La Tribune, October 13. http://www.algeria-watch.org/farticle/communales_2002/million_voix.htm

Jamal, Amaney A. and Ellen Lust-Okar. 2002. “Reassessing the Influence of Regime Type on Electoral Law Formation.” Comparative Political Studies 35(3): 337-366.

Lust-Okar, Ellen. 2006. “Elections under Authoritarianism: Preliminary Lessons from Jordan.” Democratization 13 (3): 456-471

———. 2009. “Competitive Clientelism in the Middle East.” Journal of Democracy 20 (3): 122-135.

Malesky, Edmund, and Paul Schuler. 2010. “Nodding and Needling: Analyzing Delegate Responsiveness in an Authoritarian Parliament.” American Political Science Review 104 (3): 485-502.

Masoud, Tarek. 2014. Counting Islam: Religion, Class, and Election is in Egypt. Cambridge, MA: Cambridge University Press.

Meyersson, Erik. 2014. “Islamic Rule and the Empowerment of the Poor and Pious.” Econometrica 82 (1): 229-269.

Program on Governance and Local Development (GLD). Accessed July 1, 2017. http://gld.gu.se

Roberts, Hugh. 2003. The Battlefield: Algeria, 1988-2002: Studies in a Broken Polity. London: Verso.

Volpi, Frédéric. 2013. “Algeria Versus the Arab Spring.” Journal of Democracy 24 (3): 104-11.  http://www.journalofdemocracy.org/article/algeria-versus-arab-spring

Wikipedia. 2017. “List of Algerian Provinces by Population.” Last modified May 17. https://en.wikipedia.org/wiki/List_of_Algerian_provinces_by_population