Partisanship and political accountability in new democracies: Explaining compliance with formal rules and procedures in Ghana

Abstract

Standard political agency models generally predict an inverse relationship between the degree of partisan allegiance among citizens and political accountability. Does variation in voter attachments to political parties influence the behavior of public officials in new democracies? I take advantage of data from a unique audit of local governments in Ghana—the Functional Organizational Assessment Tool (FOAT)—to examine the impact of partisanship on public officials’ compliance with formal rules and procedures. Because unattached or weakly attached voters are more responsive to the performance of incumbent officials, they are more likely to deter rent seeking, corruption, and other administrative malpractices. Analysis of the baseline FOAT results provides strong support for this idea: compliance with formal rules and procedures is significantly higher in districts where voters evince weak attachments to political parties. This result is robust to controls for alternate explanations and sheds light on the conditions under which politicians would abjure rent seeking and corruption, even in the context of a new democracy where they have so much opportunity to do so.

Keywords

Voter attachment compliance Ghana local government

Introduction

Political parties are an important instrument and institution of democratic politics. That modern representative democracy is not conceivable without political parties is widely accepted among researchers (Ladd, 1992). By articulating and integrating different interests, visions, and opinions, parties mobilize citizen engagement in the political process with the ultimate goal of winning political power and influencing government policy. While popular participation in the political process is essential for democratic development, research suggests that the degree of partisan attachments among citizens has important implications for political accountability and, hence, the proper functioning of democracy and the overall welfare of citizens.

Standard political agency models generally predict an inverse relationship between the degree of partisan allegiances among citizens and political accountability (Besley, 2007; Persson and Tabellini, 2000). When voters grow strongly attached to and prioritize electoral victory of their favorite parties over other considerations, electoral control of politicians suffers. A large body of empirical work in different contexts confirms this inverse relationship. Strong partisanship reduces voter responsiveness to politicians’ performance (Anderson, 2000; Hellwig and Samuels, 2008; Kayser and Wlezien, 2011; Moehler and Lindberg, 2009) and undermines their ability to remove corrupt and nonperforming politicians from office (Eggers, 2014). From the politicians’ point of view, strong partisanship reduces elected officials’ incentives to exert energy toward providing quality basic services to their constituents (Keefer and Khemani, 2009). I extend this logic to address an important but relatively underexplored question in new democracies where formal rules and procedures are less well institutionalized and politicians have significant opportunity to manipulate the weaknesses in the accountability system and extract public resources for private ends. I ask whether variation in voters’ partisan preferences influences the behavior of public officials in these contexts. In other words, does the inverse relationship between partisanship and political accountability apply in developing democracies?

Because the voting decisions of unattached or weakly attached voters are increasingly influenced by the performance of incumbent politicians in new democracies (Fridy, 2007; Lindberg and Morrison, 2008; Youde, 2005), the theory would predict that these voting behaviors are likely to deter official impropriety; incumbent politicians would risk damaging their re-election chances and those of their party if they engage in rent seeking, corruption, and other administrative malpractices in contexts where partisan loyalties are weak. To test this idea empirically, I use a unique and highly comparable baseline dataset from a standardized assessment of the budget implementation activities of local governments in Ghana. This on-going annual assessment program, known as the Functional Organizational Assessment Tool (FOAT), started in 2008 and is designed to gauge the extent of local governments’ compliance with the formal rules and administrative procedures governing budget implementation. These rules and procedures, notably those contained in Ghana’s Financial Administration and Public Procurement Acts (more below), are designed to decrease the prevalence of rent seeking and corruption among public officials. Thus, greater adherence to these rules and procedures means less rent seeking and/or corruption. However, the formal institutional structure of local government in Ghana (see overview below) makes local politicians more accountable to the president and the ruling national party and thus significantly weakens enforcement of and accountability vis-à-vis the rules at the local level. Even despite this single country-wide administrative structure, however, the results of the baseline FOAT assessment reveal wide variation in local governments’ compliance with rules and procedures. I argue that the degree of partisan attachments among voters accounts for this variation.

To measure the extent of voter attachments to political parties in each local government district, I use two indicators based on electoral outcomes in the two general elections preceding the baseline assessment in 2006 (i.e. the 2000 and 2004 general elections). Firstly, taking advantage of Ghana’s concurrent presidential and legislative elections, I create a dichotomous variable that takes a value of 1 if the majority of voters in the district voted for the presidential candidate of one party and the parliamentary candidate of a different party in one or both prior elections, and 0 otherwise. I call this split-ticket voting.¹ Strongly attached voters have an incentive to vote for the candidates of their favorite parties in both the presidential and legislative elections. Thus, if the winning candidates in the (concurrent) presidential and legislative elections in a district are from different parties, that would be indicative of weaker partisan preferences among some voters in the district. Secondly, to capture the effect of the magnitude of nonpartisan voters, I use the Pedersen index of electoral volatility to compute the proportion of voters that switched between the two parties in each district in the two prior elections.² For a stable two-party system with consistently high voter turnout as in Ghana, we would expect this measure to capture the rate of voter switching between the two parties over time.

Consistent with the idea that weak partisanship deters rent seeking and other corrupt practices, I find that compliance with formal budget rules and administrative procedures is significantly higher in districts where voters have weaker attachments to political parties. The results suggest that even in contexts where formal institutional constraints are limited and politicians have the greatest opportunity to manipulate or override the existing rules and procedures for personal gain, weak partisan loyalties can constrain the incentive to exploit those opportunities. This paper contributes to the broader literature about the relationship between partisanship and political accountability by focusing on the effect of partisan loyalties on rent seeking behavior among public officials in new democracies with fledgling institutions. The paper also contributes to the literature on how competitive configurations create incentives for politicians to tie their own hands through administrative reforms (Finkel, 2008; Geddes, 1991; Lehoucq and Molina, 2002). Finally, the evidence presented here is consistent with recent research that shows a rising trend of performance-based voting in Africa (Bratton, 2013; Fridy, 2007; Lindberg and Morrison, 2008; Youde, 2005), a trend that is largely concentrated among unattached or weakly attached voters (Weghorst and Lindberg, 2013).³

Empirical setting

Ghana is a unitary state divided into 10 administrative regions. The principal units of local government are the districts, of which there are presently 216 nested within the 10 regions. This is the level of government at which the national budget is allocated and spent. The political heads of local government are known as District Chief Executives (DCEs)—the equivalent of city mayors elsewhere. The president appoints all the DCEs and they have significant authority over all spending decisions in their districts. The administrative setup of local government means that the president and the governing national party have broad, centralized control over the affairs of all districts in the country.

As the main agents of development at the local level, local governments, specifically the DCEs, control large amounts of financial and other development resources. To check abuse or misappropriation of public resources, Ghana has, since its transition to democracy in 1992, introduced various laws to govern development policy implementation at all levels of government, including, in particular, local governments. These laws, in particular the Financial Administration Act and the Public Procurement Act, both enacted in 2003, are designed to prevent rent seeking and corruption among public officials. In addition to these two pieces of legislation, there are many other formal guidelines and administrative procedures for administering the numerous development resources that flow to the districts. However, the sweeping appointment powers of the president significantly weaken local-level accountability and enforcement of the formal rules and procedures. Thus, local officials have significant opportunity to manipulate or override the existing rules and procedures and extract public resources for political or private ends with little or no legal sanction.

The apparent weak enforcement of the rules and procedures and, hence, widespread rent seeking and corrupt practices at the level of local government partly prompted the introduction of the independent assessment program in 2008. Four of Ghana’s major bilateral donors partnered with the government to introduce the FOAT program to assess all local governments annually on their compliance with budget rules and administrative procedures, to motivate compliance with the established rules procedures, and to eliminate official malfeasance.⁴ The program’s accountability indicators are divided into two groups: minimum conditions (MCs) and performance measures (PMs). The PMs, the focus in this paper, are the dimensions used to assess local governments’ compliance with the existing formal rules and procedures governing implementation of the national budget in nine broad areas.⁵

Although the program was introduced in 2008, the baseline assessment was conducted on the 2006 fiscal year activities. I use the results of the baseline assessment for the analysis in this paper. The baseline results are ideal for investigating the political sources of the variation in compliance with formal rules and procedures for at least two reasons. Firstly, the 2006 baseline assessment was akin to a surprise audit of the behavior of local officials in the budget implementation phase. The 2006 fiscal year activities were executed without any knowledge of a possible assessment in the future and, hence, local officials would have conducted their activities on the basis of business-as-usual. This should be contrasted with the standard annual audits of local governments conducted by the auditor general, which local officials are very much aware of and often take steps to ensure that they get favorable audit opinions. We can imagine that corrupt officials are capable of buying the best audit opinions in the circumstance.⁶ Secondly, the FOAT program came with an incentive package: districts that do well in complying with the rules are rewarded with extra discretionary funding. This monetary incentive, which became apparent only in 2008, now complicates the behavior of local officials and thus makes subsequent FOAT results less ideal for investigating the political sources of official compliance with rules and procedures. In fact, all the districts have become “masters of the game” over time; in recent assessments, the scores on all the FOAT PMs are so high that there is very little cross-district variation to make any meaningful analysis.

Data and measurement

In 2006, the baseline year of the FOAT assessment, Ghana had a total of 138 districts and all were assessed and hence are included in the analysis below.

Dependent variables

Compliance with formal rules and administrative procedures

I operationalize the dependent variable in two ways: first I use the total compliance score on all the eight FOAT PMs used for the baseline assessment. The ninth PM—Environmental Sanitation Management—was introduced in subsequent assessments. The total score on all PMs is 100. Secondly, I use the compliance score on procurement alone. The procurement PM gauges local governments’ compliance with the formal rules governing contracting and contract execution and includes indicators such as procedures for inviting tenders, contract mobilization, tender review process, timeliness of projects execution, and contract retention. The maximum score on this measure is 9. I single out procurement for two reasons: firstly, studies on Ghana’s national budget show that more than half—between 50 and 70 percent—is procurement-related (World Bank, 2003). Since much of government spending goes into procurement, rent seeking and corruption is likely to be concentrated in this area. Secondly, procurement is an activity that is most susceptible to rent seeking behavior and corruption: officials can bend the rules of contracting and pricing of goods and services. If local officials seek to extract private benefits, they would most likely do so through procurement.⁷

Control variables

I control for several socio-economic and demographic characteristics of the districts. Firstly, I control for Education, as well-educated voters may demand more efficient management and use of public resources and may be more likely to punish politicians who behave otherwise. I operationalize this concept using the share of district population that has at least a high school education. Secondly, I control for ethnic diversity, arguably the most important factor in African elections (Posner, 2007). Ghana’s 2000 population and housing census includes a breakdown of the ethnic composition of the population in each district. I use this data to compute an index of ethno-linguistic fractionalization (ELF), which is a decreasing transformation of the Herfindahl concentration index.⁸ Thirdly, I control for bureaucratic capacity, as districts with greater bureaucratic capacity may just do better at following the formal rules and procedures than those with limited capacity. I use a measure of per capita locally generated revenues from the FOAT results to measure bureaucratic capacity. Fourthly, I control for population density (logged) as a proxy for urbanization. Finally, I control for voter turnout in each district in the 2004 presidential elections. Turnout in national elections has been shown elsewhere to influence the behavior of incumbent politicians (Fleck, 1999; Horiuchi and Saito, 2009).

Empirical model

I estimate a simple ordinary least squares (OLS) model that takes the form below and includes region fixed effects in all specifications, since the districts are nested within the 10 regions

\begin{array}{l} C o m p l i a n c e_{i} = b_{0} + b_{1} A t t a c h e d_{i} + b_{2} E t h n i c . F r a c_{i} \\ + b_{3} E d u c a t i o n_{i} + b_{4} B u r e a u . C a p a c i t y_{i} \\ + b_{5} l o g (P o p . D e n s i t y_{i}) + b_{6} T u r n o u t_{i} + ε_{i} \end{array}

Results

Table 1 presents the OLS regression results of the effect of split-ticket voting on local governments’ compliance with formal rules and administrative procedures. The dependent variable in columns 1 and 2 is the total score on all the eight PMs; that in columns 3 and 4 is the compliance score on procurement only. Columns 1 and 3 report estimates of the effect of split-ticket voting on total compliance and procurement compliance scores respectively without any covariates. In both models, the effect is statistically significant. Without accounting for any covariates, the results show that on average, the total compliance score on all PMs is about six points higher in districts that voted for candidates of different parties in at least one of the two prior general elections compared to those that voted straight ticket in those elections. Similarly, split-ticket districts score about one point higher on procurement compliance than straight ticket districts.

Table 1.

Split-ticket voting and local governments’ compliance with formal rules and administrative procedures in Ghana.

VARIABLES	Dependent variable: Total FOAT score (Range: 0–100)		Dependent variable: Procurement score (Range: 0–9)
VARIABLES	Model 1	Model 2	Model 3	Model 4
Split ticket	5.547***	3.413**	0.911***	0.570**
	(1.561)	(1.470)	(0.283)	(0.266)
Ethnic fractionalization		14.503***		2.971***
		(3.453)		(0.624)
Percent over high school		−6.575		−2.335
		(13.993)		(2.529)
Bureaucratic capacity		1.199**		−0.061
		(0.484)		(0.087)
Pop density (logged)		4.858**		1.070***
		(1.864)		(0.337)
Average turnout		−4.878		2.151
		(13.966)		(2.524)
Constant	42.831***	24.778**	5.528***	0.982
	(1.753)	(11.568)	(0.318)	(2.091)
Region fixed effects	Yes	Yes	Yes	Yes
Observations	138	138	138	138
R-squared	0.210	0.381	0.133	0.324

Standard errors in parentheses.

***

p < 0.01; **p < 0.05; *p < 0.1.

FOAT: Functional Organizational Assessment Tool.

In columns 2 and 4, I control for all the relevant covariates as well as region fixed effects. In both models, the effect of split-ticket voting remains positive and statistically significant. Local governments in districts that have experienced split- ticket voting in prior elections are significantly more likely to stick to the formal rules and administrative procedures governing development policy implementation. In these districts, the total score on all eight compliance indicators is about four points higher compared to districts that voted straight ticket in the two prior elections. These effects are quite large given the design of the indicators. Most of the indicators are dichotomous—taking a value of 1 if the district fulfils a particular formal rule or administrative procedure and 0 otherwise. This means that on average, split-ticket districts comply with about four more formal rules and/or administrative procedures than straight ticket district. The results on procurement are similar. Districts with a history of split-ticket voting score, on average, about half a point higher on procurement compliance than those with no such experience. This translates approximately to compliance with one more procurement rule or procedure.

Alternative specifications of main independent variable

Table 2 presents the results using the second operationalization of voter attachments to political parties: the Pedersen index of electoral volatility. Similar to the preceding analysis, I find that higher voter volatility is associated with greater compliance with formal rules and procedures. Based on the results in models 2 and 4, a 10 percentage-point switch between the two parties, which is just around the average, is associated with about a 4-point increase in the total FOAT score (i.e. compliance with roughly four more formal rules and procedures) and about 1 point increase in the procurement compliance score (i.e. compliance with approximately two more rules and procedures on procurement).⁹ Since each rule or procedure is designed to prevent rent seeking behavior and corruption in specific areas, the results suggest that officials in electorally volatile districts are relatively more likely to constrain their own ability to extract private rents.¹⁰

Table 2.

Electoral volatility and local governments’ compliance with formal rules and administrative procedures in Ghana.

VARIABLES	Dependent variable: Total FOAT score (Range: 0–100)		Dependent variable: Procurement score (Range: 0–9)
VARIABLES	Model 1	Model 2	Model 3	Model 4
Index of electoral volatility	0.485**	0.347**	0.105***	0.078**
	(0.189)	(0.173)	(0.035)	(0.032)
Ethnic fractionalization		14.152***		3.213***
		(3.436)		(0.633)
Percent over high school		−10.465		−2.056
		(13.803)		(2.542)
Bureaucratic capacity		1.379***		−0.034
		(0.474)		(0.087)
Pop density (logged)		4.918***		0.996***
		(1.829)		(0.337)
Average turnout		1.403		4.736*
		(14.715)		(2.710)
Constant	41.113***	18.004	5.144***	−1.481
	(1.878)	(12.181)	(0.350)	(2.243)
Region fixed effects	Yes	Yes	Yes	Yes
Observations	130	130	130	130
R-squared	0.216	0.405	0.120	0.347

Standard errors in parentheses.

***

p < 0.01; **p < 0.05; *p < 0.1.

FOAT: Functional Organizational Assessment Tool.

The other important determinants of public officials’ compliance with formal rules and procedures are ethnic diversity and population density. Both variables are statistically significant in all specifications, suggesting that local governments in ethnically heterogeneous and densely populated districts are more likely to stick to the rules and thus relatively less likely to engage in rent seeking and corruption. Firstly, the effect of ethnicity is to be expected given the nature of Ghana’s ethnic geography. The two ethnic groups that identify strongly with, and consistently vote straight ticket for candidates of the two major parties in both the presidential and legislative elections, are concentrated in only two of the 10 regions. Voting behavior in the remaining eight regions, which are relatively more ethnically diverse, is driven by factors other than ethnicity (Fridy, 2007). High ethnic diversity means that no single ethnic group can mobilize and win an election on its own and the performance of incumbents would matter in those settings. Secondly, densely populated areas, which are essentially the urban areas, are home to relatively wealthier and perhaps better-informed voters. Economically empowered voters are more likely to demand efficient management and use of public resources and they may punish incumbent politicians who engage in rent seeking and corruption. This result is consistent with studies that show that politicians are more likely to engage in clientelism and other corrupt electoral practices among the poor (e.g. Magaloni et al., 2007; Wantchekon, 2003) and thus less likely to do so in wealthy areas (Weitz-Shapiro, 2012).

Conclusion

This paper examines the impact of voter attachment to political parties on public officials’ compliance with formal rules and administrative procedures designed to eliminate or reduce opportunities for rent seeking and corruption. The evidence shows that compliance with formal rules and procedures is significantly higher in districts where voters demonstrate weak commitments to political parties. Because unattached or weakly attached voters are more responsive to the performance of incumbent officials, politicians in these districts risk undermining their own political futures if they exploit public resources for immediate personal gain. This association is robust to controls for a range of potential confounding covariates, including factors unique to each region. These findings thus advance our understanding of the conditions under which politicians would play by the rules and potentially limit their ability to extract private rents. Using the case of Ghana, where the formal legal system is relatively weakly institutionalized, has allowed insight in particular into a context where politicians have significant opportunity to engage in corruption, given the relatively low chance of legal sanction. The results suggest that when voters are less committed to political parties, rent seeking behavior and other administrative malpractices are likely to decline.

Footnotes

Declaration of Conflicting Interests

The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Funding

The author(s) disclosed receipt of the following financial support for the research, authorship, and/or publication of this article: UCLA dissertation year fellowship; UCLA International Institute; and UCLA Charles and Sue Young Fellowship.

Supplementary material

The online appendix is available at:

Carnegie Corporation of New York Grant

The open access article processing charge (APC) for this article was waived due to a grant awarded to Research & Politics from Carnegie Corporation of New York under its ‘Bridging the Gap’ initiative.

Notes

References

Anderson

(2000) Economic voting and political context: a comparative perspective. Electoral Studies 19(2–3): 151–170.

Besley

(2007) Principled Agents? The Political Economy of Good Government. Oxford: Oxford University Press.

Bratton

(2013) Voting and Democratic Citizenship in Africa. Boulder, CO: Lynne Rienner Publishers.

Carlson

(2010) Great Expectations: Ethnicity, Performance, and Ugandan Voters. Working Paper, UCLA Department of Political Science, Los Angeles.

Eggers

(2014) Partisanship and electoral accountability: Evidence from the UK expenses scandal. Quarterly Journal of Political Science 9(4): 441–472.

Ferree

(2006) Explaining South Africa’s racial census. Journal of Politics 68(4): 803–815.

Finkel

(2008) Judicial Reform as Political Insurance: Argentina, Peru, and Mexico in the 1990s. 1st ed. Notre Dame, IN: University of Notre Dame Press.

Fleck

(1999) The value of the vote: A model and test of the effects of turnout on distributive policy. Economic Inquiry 37(4): 609–623.

Fridy

(2007) The elephant, umbrella, and quarrelling cocks: Disaggregating partisanship in Ghana’s fourth republic. African Affairs 106(423): 281–305.

10.

Geddes

(1991) A game theoretic model of reform in Latin American democracies. The American Political Science Review 85(2): 371.

11.

Hellwig

Samuels

(2008) Electoral accountability and the variety of democratic regimes. British Journal of Political Science 38(1): 65–90.

12.

Horiuchi

Saito

(2009) Rain, Elections and Money: The Impact of Voter Turnout on Distributive Policy Outcomes in Japan. Canberra, Australia: Australian-Japan Research Centre, Australian National University, Crawford School of Economics and Government.

13.

Ichino

Nathan

(2013) Crossing the line: Local ethnic geography and voting in Ghana. American Political Science Review 107(02): 344–361.

14.

Kanyinga

(1994) Ethnicity, Patronage and Class in a Local Arena: High and Low Politics in Kiambu, Kenya, 1982–92. Johannesburg, South Africa: University of the Witwatersrand.

15.

Kayser

Wlezien

(2011) Performance pressure: Patterns of partisanship and the economic vote. European Journal of Political Research 50(3): 365–394.

16.

Keefer

Khemani

(2009) When do legislators pass on pork? The role of political parties in determining legislator effort. The American Political Science Review 103(1): 99–112.

17.

Ladd

(1992) Political Parties, ‘reform’ and American democracy. In: White

Mileur

(eds) Challenges to Party Government. Carbondale, IL: Southern Illinois University, pp. 22–39.

18.

Lehoucq

Molina

(2002) Stuffing the Ballot Box: Fraud, Electoral Reform, and Democratization in Costa Rica. Cambridge, UK: Cambridge University Press.

19.

Lindberg

Morrison

MKC

(2005) Exploring voter alignments in Africa: Core and swing voters in Ghana. The Journal of Modern African Studies 43(4): 565–586.

20.

Lindberg

Morrison

MKC

(2008) Are African voters really ethnic or clientelistic? Survey evidence from Ghana. Political Science Quarterly 123(1): 95–122.

21.

Magaloni

Diaz-Cayeros

Estévez

(2007) Clientelism and portfolio diversification: A model of electoral investment with applications to Mexico. In: In Patrons, Clients, and Policies: Patterns of Democratic Account- ability and Political Competition. Cambridge University Press. Available from: http://dx.doi.org/10.1017/CBO9780511585869.008 (accessed 25 August 2012).

22.

Moehler

Lindberg

(2009) Narrowing the legitimacy gap: Turnovers as a cause of democratic consolidation. The Journal of Politics 71(04): 1448–1466.

23.

Persson

Tabellini

(2000) Political Economics: Explaining Economic Policy. Cambridge, MA: MIT Press.

24.

Posner

(2005) Institutions and Ethnic Politics in Africa. Cambridge; New York: Cambridge University Press.

25.

Posner

(2007) Regime change and ethnic cleavages in Africa. Comparative Political Studies 40(11): 1302–1327.

26.

Wantchekon

(2003) Clientelism and voting behavior: Evidence from a field experiment in Benin. World Politics 55(03): 399–422.

27.

Weghorst

Lindberg

(2013) What Drives the swing voter in Africa? American Journal of Political Science 57(3): 717–734.

28.

Weitz-Shapiro

(2012) What wins votes: why some politicians opt out of clientelism. American Journal of Political Science 56(3): 568–583.

29.

World Bank (2003) Ghana-Country procurement assessment report (Vol. 4): Annex 8-Public Procurement Act (English). Available from: http://documents.worldbank.org/curated/en/2003/06/3582686/ghana-country-procurement-assessment-report-vol-4-5-annex-8-public-procurement-act (accessed 3 October 2013).

30.

Youde

(2005) Economics and government popularity in Ghana. Electoral Studies 24(1): 1–16.