Political Business Cycles and Democratization: Evidence From Sub-Saharan Africa

Abstract

The literature on Political Business Cycles (PBCs) has suffered from two limitations, namely a dominant focus on government policies rather than outcomes that could influence voter behaviour, and a lack of attention to the relationship between PBCs and democratization. Using multiple fine-grained data on objective and subjective outcomes we examine the nature of PBCs in Sub-Saharan Africa, a region which has experienced substantial levels of democratization in recent decades. We demonstrate clear evidence for the existence of PBCs in Sub-Saharan Africa and that the nature of the PBC changes with democratization. Specifically, we show that PBCs in non-democracies focus more on the provision of private goods and less on public goods, with this reversing as countries democratize. These findings, which hold across data sources and are robust to various specifications, have important implications for our understanding of the link between elections and development outcomes in Sub-Saharan Africa and beyond.

Keywords

D72 H41 H54 N47 African politics democratization and regime change elections public opinion voting behavior

Introduction

The idea that politicians manipulate the economy to win elections has a long history.¹ The work of Nordhaus (1975) introduced the idea that politicians time fluctuations in macroeconomic outcomes to court voters in what is known as the Political Business Cycle (PBC).² Despite the intuitive appeal of the argument, empirical studies found limited evidence for political cycles in macroeconomic outcomes such as unemployment and inflation (Brender & Drazen, 2005). Acknowledging politicians’ limited ability to control macroeconomic outcomes, this led to a shift in focus instead on policy instruments rather than outcomes (Dubois, 2016). Empirical work directed to this end has proved more fruitful, with various studies finding evidence of political cycles in fiscal and monetary policy instruments.³

Arguably the PBC literature has suffered from two limitations. First, it has focused too much on inputs and government policies, rather than outputs and the actual outcomes of these policies that could affect voters’ behaviour. More specifically, within the PBC literature there has been a relative lack of interest in the relationship between elections and measures of public goods and individual well-being, especially in developing countries. Indeed, while the PBC literature historically arose in the study of American politics, there is strong reason to expect larger and more prominent PBCs in the developing world given the greater political importance of elections in poorer countries (Przeworski, 2004).

A second limitation in the PBC literature has been to assume that PBCs will be identical across countries regardless of their degree of democracy. Indeed, despite growing evidence on how autocrats spend large amounts of energy and resources on winning elections by large margins (Blaydes, 2011; Magaloni, 2008; Simpser, 2013), much of the literature on elections in non-democracies has focused more on the role of sticks than carrots in generating electoral support for incumbent regimes. For instance, much of the recent PBC literature has suggested that PBCs have either a linear (Jäger, 2016; Lokshin and Torre, 2024; Mandon & Cazals, 2019; Philips, 2016) or an inverted U relationship with democracy (Aidt et al., 2020; Eibl & Lynge, 2017; Shmuel, 2020), and do not examine the possibility that different types of regimes will have different types of PBCs.

In this paper we move to correct these two oversights by examining the existence of PBCs in Sub-Saharan Africa using multiple sources of fine-grained data on economic outcomes across different types of political regimes.⁴ We focus on Sub-Saharan Africa because of the availability of multiple types of data over a period of substantial democratization. We first use annual nighttime luminosity data to show that there is more economic activity in election years, but the crude nature of the luminosity data does not allow us to assess the differential nature of PBCs across regimes. As such we turn to three different sources of firm and individual/household survey data - World Bank Enterprise Surveys, Demographic and Health Surveys and Afrobarometer surveys - that allow us to examine how PBCs differ across regimes. Consistent with our theoretical expectations, we show that PBCs in non-democracies effectively consist of vote-buying, with elections associated with increases in subjective and objective individual well-being and economic output, but with lower levels of access to infrastructural public goods such as electricity. However, as countries democratize these outcomes reverse, with elections no longer associated with increasing levels of well-being and investment while they are associated with higher levels of access to public goods. Moreover, we do not find consistent evidence for the existence of “sine wave” PBCs, whereby increased access to goods before elections are countered by lower access to such goods afterwards. Our results thus suggest that democratization in Sub-Saharan Africa has generated more public goods provision during election years, which contributes to scholarship on the positive effects of democratization on human welfare. This paper also adds to the literature on PBCs and suggests that future literature on the topic should both focus more on actual outcomes as well as investigate how democratization can alter the relationship between elections and goods provision.

The rest of the paper is organized as follows. First, we begin with an overview of the last half-century of literature on PBCs and generate our hypotheses. Second, we describe the data sets and their construction, and explain our empirical strategy. Third, we discuss our results. Finally, we conclude with broader thoughts on how our work speaks to broader issues in the study of PBCs and African politics.

Political Manipulations of Economic Policy

Theoretical Arguments

Theoretical models formulated to explain PBCs have differed over their assumptions about politicians and voters. The foundational work in this field started with the simplistic ideas that politicians are identical and opportunistic, and voters are naïve and myopic (Block, 2002). This permitted politicians to be modeled as making macroeconomic policy decisions purely on the basis of electoral concerns. It also allowed models to ignore the possibility that voters might observe and punish such opportunistic behaviour. Later work adapted both assumptions, accommodating ideological/partisan differences in policy concerns on the part of politicians (Alesina, 1987; Hibbs, 1977), and acknowledging rational expectations of voters (Marin et al., 1990; Rogoff, 1990).

In response to the acknowledgment of rational expectations, models overcame the challenges these posed to the PBCs literature by incorporating information asymmetries between voters and politicians. Adverse selection models of this type posited that voters have incomplete information about government competence and are misled by the government’s pre-election expansionary economic activity. Later models relaxed this assumption about information asymmetry but maintained the expectation of PBCs by suggesting a moral hazard problem, whereby incumbents affect an appearance of greater competence prior to elections (Dubois, 2016).

The key take-away from these models is the recognition that where voters are aware of manipulative behaviour by politicians, they will not reward pre-election boons. Adapting the arguments to accommodate this recognition created expectations about the conditional likelihood of PBCs occurring. Some models in this vein have focused on the role of transparency and information. For example, Shi and Svensson (2006) argue that the size of PBCs should depend on the share of informed voters in the electorate. Similarly, Alt and Lassen (2006) suggest that the occurrence of PBCs is conditional on the level of institutional fiscal transparency. A related but slightly different approach has suggested that PBCs should be greater in new democracies, where voters have less experience with pre-electoral fiscal manipulations, and where the media are less experienced at identifying and reporting on such activities (Brender & Drazen, 2005; Mandon & Cazals, 2019).

Relatedly, conditional expectations over the occurrence of PBCs have also emerged from recognition of factors affecting the ability of governments to control fiscal policy instruments (Dubois, 2016). For example, Block (2002) has highlighted the advantage that African incumbents typically face due to the centralization of fiscal policy authority and the lack of independent monetary authorities in many African presidential systems. More generally, PBCs may be more likely in institutional contexts where policymaking power is subject to fewer veto points, such as parliamentary regimes (Chang, 2008; Persson & Tabellini, 2005). And in the same way, while political competition may create the incentives for politicians to generate PBCs, the ability of governments to manipulate fiscal and monetary policy is likely to be limited by a variety of executive constraints present in more strongly democratic states (Shmuel, 2020), and may even lead to a reverse PBC in democracies as investment declines before elections due to uncertainty around future government policies (Canes-Wrone & Park, 2012).

Evidence

While there is limited evidence for political manipulations of macroeconomic outcomes, numerous studies have demonstrated political cycles in the use of fiscal policy instruments.⁵ In line with the conditional expectations highlighted above, Alt and Lassen (2006) find heterogeneity over degrees of fiscal transparency. Moreover, several studies have found stronger evidence for PBCs in less-developed economies, and in newer democracies (Aidt et al., 2020; Brender & Drazen, 2005; Ebeke & Ölcer, 2013; Shi & Svensson, 2006).

In recent work, Veiga et al. (2017) provide evidence that the occurrence of PBCs is conditional on media freedom, which they argue underpins the informational constraint. This fits with evidence of an electoral cycle in social protection spending in authoritarian regimes provided by Han (2022), who argues that opportunistic policy manipulation is facilitated by the information asymmetries present in autocracies. Also focusing on differences in regime types, Shmuel (2020) demonstrates a non-linear relationship between PBCs and the level of democracy. This is argued to follow from the different incentives to engage in political fiscal manipulation faced by autocrats and democrats, as well as the differences in their ability to effectively undertake such manipulation.

Contributions of This Research

While existing work provides compelling support for the existence of PBCs, in particular in lower income countries (Kyriacou & Roca-Sagalés, 2022; Shi & Svensson, 2006), we argue that this work remains somewhat limited by the focus on economic policy instruments rather than outcomes. There is by now good evidence that voters in Africa care about macroeconomic outcomes.⁶ But there is also substantial and compelling evidence that many voters in Africa (as elsewhere) condition their electoral support on their own personal economic situation, and on their access to essential goods and services.⁷ While individual or “pocketbook” economic outcomes may be linked to fiscal policy instruments, if the theoretical argument underpinning PBCs is that economic policies are manipulated to win electoral support, then we should expect to see an impact of the electoral cycle on factors at the individual level that affect such support.

This logic suggests that we should expect to observe higher levels of individual-level measures of well-being, in terms of both pocketbook outcomes as well as access to goods and services, during electoral periods than at other times. However, it seems likely that the nature of these PBCs will vary across different types of political regimes. In particular there are strong theoretical reasons to assume that the level of democracy will influence the nature of PBCs due to incentives on the part of both governments and voters. As regards the former, an older literature assumed that democratic institutions in non-democracies was “mere window dressing” that dictators used to pretend that they had democratic legitimacy (Gandhi & Przweworski, 2006, p. 3), but more recent scholarship has instead argued that many non-democratic regimes actively use democratic institutions as a means to perpetuate their rule. In particular, elections serve multiple functions in non-democratic or hybrid regimes, including by generating information as to the popularity of the regime and allowing governments to divide, demobilize and demoralize opposition movements through large margins of victory (Blaydes, 2011; Gandhi & Lust-Okar, 2009; Magaloni, 2008; Simpser, 2013), which in the long run can perpetuate regime survival (Knutsen et al., 2017).

While much of this literature has focused on the ways in which non-democratic regimes use fraud, oppression, harassment and violence to win elections, such tactics can lead to a public backlash which could be politically costly for the regime while also institutionalizing a security structure that has the potential to overthrow the regime (Haber, 2006). In contrast, however, by increasing public expenditure during election campaigns non-democratic governments can generate more support for the regime at a relatively low cost. It is thus not surprising to find substantial evidence for higher levels of public spending during elections in non-democracies in case studies such as Egypt (Blaydes, 2011), Malaysia (Pepinsky, 2007) and Mexico (Magaloni, 2008) as well as a broader set of non-democracies (Higashijima, 2016; Wright, 2011). Indeed, and as noted above, the strong information asymmetries often present in authoritarian regimes may actually facilitate and encourage the opportunistic manipulation of public spending around electoral cycles (Han, 2022).

We would thus expect increases in public expenditure during election campaigns in both democracies and non-democracies to generate an increase in economic output, consistent with the existence of a PBC. However, it is not clear that regimes of either kind would be incentivized to focus spending on the provision of public goods, for two reasons. First, the complexity of many public good projects can generate challenges in timing their completion to coincide with elections, which in many cases can lead to a lack of investment in long-term infrastructure projects that will not complete until after elections (Bonfiglioli & Gancia, 2013; Thomas & Darsey, 2024). Second, the non-excludable nature of many public goods and services makes it harder to target their benefits towards specific voters. Instead, we would expect governments to focus public expenditure during electoral periods on the allocation of private goods, including doling out public sector jobs and handing out cash and cheap assets to voters that can be both timed and targeted to maximize electoral benefits. Indeed, as noted by Kramon (2018, p. 4), there is evidence that such practices are widespread in contemporary sub-Saharan Africa across both democracies and non-democracies.

However, as countries democratize we should expect to observe a shift in government spending during electoral cycles, with less emphasis on private goods and more focus towards public goods provision. More specifically, we would expect that, as non-elite citizens acquire more political power and the size of the (s)electorate grows, the private provision of goods becomes increasingly inefficient in favor of the economies of scale public goods provision can generate in serving larger numbers of voters (Bueno De Mesquita et al., 2005; Deacon, 2016; Lake & Baum, 2001).⁸ In fact, there is therefore now a substantial amount of evidence linking democratization to a broader and more equitable provision of public goods in health (Kudamatsu, 2012; Wang et al., 2019), infrastructure (Burgess et al., 2015) and education (Harding & Stasavage, 2014).

Moreover, while the aforementioned advantages in terms of timing and targeting generate incentives for democratic rulers to distribute private transfers in election periods, these rulers are likely to be more constrained in doing so than their authoritarian counterparts. Such constraints follow in part from the greater transparency and media freedoms present in democracies (Veiga et al., 2017), as well as from the institutionalization of rules and norms that limit the use of non-programmatic forms of distribution in more democratic contexts (Brun & Diamond, 2014). As a result, the more even electoral playing fields that democrats compete on are likely to constrain the use of clientelist strategies and create incentives for governments to pay the higher costs of constructing electoral majorities through the provision of public goods and services (Harding, 2020b). Nevertheless, to the extent that the provision of public goods and services can be timed to coincide with electoral periods, democratic governments will face incentives to do so.

As with our discussion so far, most of the PBC literature has focused on the actions of governments rather than voters. Yet there is reason to believe that elections can generate incentives for voters as well. More specifically, as countries democratize we would expect elections to be increasingly competitive, thereby generating real uncertainty as to who the winners will be and what government policies they will enact. This uncertainty can cause both voters and international investors to withhold investments and postpone large, irreversible purchases during election years, which can have the effect of reducing economic output and thereby generating negative or reverse PBCs within the private sector (Canes-Wrone & Park, 2012, 2016; Julio & Yook, 2012). This effect is possibly even stronger in presidential systems, where elections are regular and predictable (Bak, 2016). Given the overwhelming preponderance of presidential systems in Africa, we would thus expect to observe the existence of reverse PBCs for private sector economic outcomes in more democratic countries across the continent.

We can thus generate our central hypotheses:

• Access to private goods, measures of individual well-being and private sector economic outputs should be greater in election years in less democratic contexts, and this effect should decline as countries become more democratic.

• Access to public goods should be greater in election years in more democratic contexts, and this effect should decline as countries become less democratic.

Before we turn to our quantitative data, we turn to qualitative evidence from Africa on the existence of PBCs across different political regimes. As with the broader literature, previous scholarship on African PBCs has largely focused on demonstrating evidence for PBCs without examining how PBCs differ across regimes. Thus in their overview of African electoral politics since 1990, Bleck and van de Walle (2018, p. 96) note that governments “increase public sector employment and developmental spending in the run up to the election,” such that “in the run up to the election government ministers tour the countryside and take credit for rural infrastructure and new clinics, schools and roads.” Arguing that African presidents operate in an institutional environment that is particularly conducive to PBCs, Block (2002) provides evidence of electoral interventions across a variety of fiscal and monetary policy instruments in 44 countries in Africa between 1980 and 1995.⁹ Subsequent work updates and corroborates this finding in samples covering more recent time periods (Bobbo & Bate, 2022; Neumann & Ssozi, 2016; Strong, 2023, 2024), while evidence from Marx (2018) finds that there is a PBC in the completion of World-Bank funded projects in Africa - but not in their inception, and only for basic infrastructure and social service projects (and not for agriculture, finance, industry and governance projects). Finally, Higashijima and Kerr (2023) use Afrobarometer data to show that electoral cycles generate higher levels of citizen satisfaction with democracy.

If we turn to case study evidence, however, it becomes clearer how PBCs differ across political regimes in the African context. Here we focus on how PBCs have unfolded in two countries with different levels of democracy. In Ghana President Jerry Rawlings oversaw a period of substantial democratization in the 1990s, with an increase in press freedom and space for opposition parties to function fully. The result has been multiple alterations of power in Ghana since 2000, making it one of the most democratic countries in Africa. However, this period has also seen governments engaged in “uncontrolled spending during election years” comprising an average of 1.1% of GDP, according to Bank of Ghana economists Asiama et al. (2014, p. 5). As such the Ghanaian parliament passed a law in 2018 that capped the annual fiscal deficit at 5% of GDP in each fiscal year to limit election year spending. However, in July 2020 the Finance Minister received parliamentary permission to temporarily suspend the law due to the COVID-19 pandemic, a move which coincidentally took place five months before a national election (Abdul-Gafaru & Sackeyfio, 2022).

Qualitative evidence suggests that successive Ghanaian governments have focused their election year spending on public infrastructure. President Jerry Rawlings arguably timed the completion of electrification projects and roads to occur just before elections in 1992 and 1996, which in the latter case included campaign advertisements with pictures of paved roads and electrical poles (Briggs, 2012, pp. 605–606). In fact, as MacLean et al. (2016, p. 121) point out, the connection between electoral politics and electricity in Ghana goes all the way back to the post-independence period of the 1950s, and has continued to the present day, with problems over load shedding and power cuts arguably responsible for President John Mahama’s loss in his 2016 presidential re-election campaign. Moreover, a focus on electricity in election years has not been limited to one part of the country: “in the run-up to the [2012] elections, electrification projects sped up” in Northern Ghana, one of the poorest regions of the country, with the “months prior to the election in December a time of frenetic infrastructural activity” (Fernández, 2017, p. 76).

In contrast to Ghana, Uganda is one of the less democratic countries in Africa, with no peaceful transitions of executive power since independence. President Yoweri Museveni has been in power since 1986, and has allowed national elections since 1996 and multi-party elections since 2006, which has nonetheless meant severe restrictions on campaign activities by opposition politicians with constant levels of harassment, arrests and threats of violence. Despite the continued electoral dominance of Museveni and his political party, the National Resistance Movement, Museveni’s government has nonetheless focused a great deal of attention on winning national elections by large margins. Thus, for example, his government not only passed an unusually large budget in the year of the 2011 Presidential election but also passed an additional Supplementary Budget a month before the election and withdrew some USD 740 million from the Bank of Uganda during the campaign (Conroy-Krutz & Logan, 2012, p. 633). Instead of spending this money on infrastructure Museveni instead “handed out wads of cash to various associations, fundraisings, churches and mosques throughout the election period” (Vokes & Wilkins, 2016, p. 589).

The high levels of spending in the 2011 campaign generated an increase in inflation after the polls, which led to a more decentralized form of campaign spending in the next election in 2016, when much of the money spent in the campaign was instead raised by MP candidates themselves, especially in the form of loans and remortgages which left many candidates in debt after the election (Vokes & Wilkins, 2016, p. 590).¹⁰ Here again election year spending in Uganda focused on putting cash in the hands of voters, with one study finding that 40% of respondents acknowledged receiving cash for their votes, 83% of which came from incumbent politicians (Blattman et al., 2019, p. 7). Other reports from the 2016 election record the distribution of hoes to voters during campaign rallies, as well as the formation of a new voluntary group of civilians recruited before the elections called Crime Preventers, whose membership numbers ran as high as 11 million people (or over 28% of the total population of Uganda) and whose official task was to report local electoral crimes but who unofficially acted as NRM cadres and in turn received cash payments, bicycles and access to a government-run income scheme (Abrahamsen & Bareebe, 2016). In contrast, however, electrification remains relatively untouched by electoral politics in Uganda and is not a major focus of discussion during election campaigns (MacLean et al., 2016).

Data & Empirical Strategy

In our analysis we used panel datasets from four sources, namely nighttime luminosity measures, World Bank Enterprise Surveys, Demographic and Health Surveys and Afrobarometer survey data. Here we describe each data set in turn. In the first case we have annual country-district-year observations for all countries in Africa between 1992 and 2013, which is drawn from Hodler and Raschky (2014) as originally compiled from the US National Oceanic and Atmospheric Administration (NOAA). Luminosity data has been shown to be a useful measures of GDP and economic activity more broadly, especially in countries where fine-grained data on GDP and economic growth is unavailable or unreliable, making it particularly useful in our analysis (Chen & Nordhaus, 2011; Henderson et al., 2012). However, more recent evidence suggests that satellite data is not very accurate in reflecting economic activity in rural areas, in part because it does not capture agricultural growth very well (Gibson et al., 2021; Keola et al., 2015). Indeed, Gibson et al. (2021, p. 11) conclude their recent review of different types of remote sensing data with the suggestion that researchers interested in well-being among rural populations in developing countries “persevere with their traditional survey-based methods of measuring living standards.”

The second data set used in our paper is the World Bank Enterprise Survey (WBES), which has surveyed firms across the developing world since the early 2000s. The surveys consist of a variety of questions about the firms’ economic activities in the fiscal year prior to the survey, including questions on annual sales, purchases and wage costs. As with the other three survey data sets in our analysis, only recently has the WBES produced enough rounds to use it to examine temporal variation in outcomes, as also seen in two recent papers on the existence of ethno-regional favouritism (Asatryan et al., 2022; Osei-Tutu & Weill, 2023). While the number of surveys in the WBES data set is thus relatively low - with 66 surveys spread across 27 countries - the advantage of the WBES data set is that it is a panel at the firm level.

Our third source, the Demographic and Health Surveys (DHS), consists of survey data for up to eight rounds dating back to the late 1980s across the developing world. As per its name the DHS focuses on demographic and health issues, with its main target being women of reproductive age (aged 15–49); it consists of several data sets per survey, with our focus on household-level data on access to public goods. While the DHS does not consist of longitudinal data, its coverage across Sub-Saharan Africa is substantial, as is its size (with an average of over 10,000 households per country-survey), especially when we supplement the DHS with the UNICEF Multiple Indicator Cluster Surveys (MICS) household data which uses a similar structure with nearly identical questions. As such we take district (or ADM2) level averages from each survey to generate a country-district-survey panel data set by collapsing the data using the provided survey weights, which allows for variation in sampling across rural and urban areas and non-responses.

Our fourth and last data source, the Afrobarometer surveys, consists of up to eight rounds per country of individual-level survey data dating from the late 1990s to the present. The Afrobarometer regularly asks questions about individual well-being and access to public services as well as political attitudes, which has led it to avoid conducting surveys during elections.¹¹ The Afrobarometer surveys are smaller than the DHS - with a maximum of 2400 respondents - and include adult women and men of all ages. As with the DHS we collapse the Afrobarometer data down to the country-district-year in order to generate a panel data set, again by using provided survey weights.

For the DHS and Afrobarometer data we chose to focus on country-district-surveys rather than country-region-surveys or country-surveys as our units of observation as the surveys often take months to complete, which leads to in-survey variation in the length of time to an election, as well as uneven geographical coverage across time due to fieldwork constraints, such as the decision not to survey districts in northern Mali in 2012/2013 due to regional conflict. For most countries we used ADM2 or the second-highest administrative (aka district) level data, with the caveat that in some cases the number of observations per district per survey was too small to generate an accurate panel (such as in Nigeria, with its 774 Local Government Areas). In the two cases of Kenya and Madagascar the whole system of local administration was overhauled during our period of study, which led us to choose units that were comparable in observations per unit to other countries in the data set despite the fact that the units were at one point the highest rather than the second-highest level administrative units.¹² Our ultimate goal in each case was to use the level of administrative unit that was most comparable to units in other countries as regards the number of individual survey observations per unit per round, as can be seen in appendix tables A1 and A2 in the Supplemental Material. As we note below, our results are robust to using ADM1 data instead, with the caveat that there is more variation in coverage within each region across surveys.

Our election variable across all four data sets is a dummy variable that takes the value of 1 when the data pertains to a year in which the country in question had a national election (for the luminosity and WBES data) or the survey was held up to twelve months before a national election (for the Afrobarometer and DHS data), and 0 otherwise. (For the sake of simplicity we henceforth call both measures the election year variable.) Our use of an election-year dummy is limited by the structure of the data in the case of the luminosity and WBES data; for the Afrobarometer and DHS we use the twelve months before the election as our basic measure not only as it is comparable in length of time to the luminosity measure as well as in frequency but also because it matches the length of time of an annual budget and it strikes the right balance between proximity to the election and frequency.¹³ We do, however, include robustness measures using different lengths of time.

In our basic specifications we also include a dummy variable for the year after an election year for the luminosity data, the year of the survey for the WBES data (i.e., the year after the previous fiscal year) and the twelve months after an election for the Afrobarometer and DHS data for two reasons: first, we wish to test the existence of a “sine wave” PBC, whereby increased levels of well-being before an election are countered by decreased levels after an election,¹⁴ and second, we wish to control for the possibility that governments are calling elections in boom periods, which would be consistent with a positive sign on both the pre-election and post-election dummy variables. We code an election as a national election if the election is for either the President and/or Parliament; municipal or local elections that are held across the country thus do not form part of our data set.

Our measure of democracy in all four data sets is the the V-Dem polyarchy democracy measure (at the country-survey level), which is useful for our purposes in several ways. First, it captures a broad understanding of democracy inasmuch as it is based on both weighted and multiplicative averages of measures of freedom of association, clean elections, freedom of expression, elected officials and suffrage, all of which are generated by expert coding. Second, it is a fine-grained measure on an interval scale, which, unlike with binary measures of democracy or even the Polity IV’s 21 point scale, allows for annual variation, which is an important point especially considering the temporal proximity of many of the Afrobarometer surveys. Third, it extends well back in time and up to the present, thus making its coverage broader than both the Polity IV index (which ends in 2018) and the Freedom House Aggregate Scores (which we use as a robustness measure below).

In our specification we interact the democracy measure with the election year variable to capture the changing relationship between democracy and elections over time.¹⁵ We also include both round/year and district fixed effects as well as a country-specific time trend to control for long-term trends in welfare, except in the case of the WBES data set due to the low number of observations per country. For the Afrobarometer and DHS data sets we add a control for proportion urban (at the country-district-survey level), and in all cases we weight all observations equally by country-round.¹⁶ A full list of all surveys included in our analyses can be found in appendix Tables A1 through A3 and the summary statistics are given in Table A4.

Results

We use these four data sets to test our hypotheses in multiple ways. First, we use the luminosity data to check for the existence of PBCs in Africa as regards economic output. We then examine the relationship between elections, democracy and private good provision. We begin with WBES firm-level data on sales, wages, number of employees, wages per worker and asset and equipment purchases (all logged), all in the previous fiscal year. We then turn to the Afrobarometer for both objective and subjective outcomes: we measure the former through an asset index (from an arithmetic mean of respondent answers on radio, television and car ownership), and the latter through questions on respondents’ self-assessed current living conditions, living conditions relative to others and their current assessment of the country’s economy, whether the country is headed in the right direction, their current living conditions compared to 12 months ago and the country’s current conditions compared to 12 months ago.

We then turn to our examination of the relationship between elections, democracy and public goods. In this case our focus is not on the existence of public goods per se but rather on respondents’ access to such goods, on the assumption that only the latter should generate electoral support for the regime. Our measures of public goods provision here are the provision of and access to both piped water and electricity, with a particular focus on the latter for several reasons. First, electricity is an important public good that has significant impacts on broad measures of economic and human development (Cole et al., 2018; Lipscomb et al., 2013; Rud, 2012). Second, there is evidence from multiple countries like India (Baskaran et al., 2015; Min, 2015) and South Africa (Kroth et al., 2016) - as well as the Ghanaian case discussed above - that electricity provision is in part determined by electoral considerations. Indeed, within Africa there is evidence for a robust correlation between democratization and electricity provision (Ahlborg et al., 2015; Trotter, 2016). Third, while we have data on water provision/access from two of our three datasets, we have data from all three data sources on access to electricity. We should note here that we do not use nighttime luminosity as a measure of electricity, as it is now well-established that luminosity best represents a measure of economic activity rather than access to electricity (Chen & Nordhaus, 2011; Henderson et al., 2012). Indeed, as noted by Kroth et al. (2016) in the South African context, luminosity is often generated by factors unrelated to household access to electricity, in particular street lights and industrial estates, which is the reason why they, like us, focus their attention on individual-level data on electrification.

In our analysis we begin with WBES firm-level data on the cost of electricity (logged), a dummy variable measuring whether the firm had any electricity outages, and the average number of outages per month, all in the previous fiscal year. We then turn to DHS data on household access to electricity and piped water, and then examine Afrobarometer data on respondents’ subjective assessment of how the government is doing in the provision of electricity and piped water.

Our first results can be viewed in Table 1, in which we present uninteracted estimates of the effect of an election year dummy on the various outcomes described above, first for the luminosity data and then for the WBES and Afrobarometer data. We start with the luminosity results in the first row, where the election year dummy is positive and statistically significant and indicates that there is around 2.7% more luminosity in years with national elections. We then present the results of the other thirteen outcomes from the WBES and Afrobarometer data. The coefficients on the WBES pre- and post-election dummies are mostly negative, which are consistent with elections taking place at low points in the business cycle; however, the Afrobarometer coefficients are more suggestive of a PBC, with positive coefficients on the pre-election variable for all outcomes and no consistent relationship for the post-election variable.

Table 1.

Electoral Cycles in Economic Output and Private Goods

Panel A:	Lights Data	WBES Data					DHS Data
	(1)	(2)	(3)	(4)	(5)	(6)	(7)
	Nighttime light	Sales	Wages	Number of employees	Wages per worker	Assets purchases	Assets
Election year	0.0268^*	−0.9452^*	−0.5992^**	0.0128	−0.5852^**	−1.9336^**	0.0010
Election year	(0.0105)	(0.1964)	(0.2020)	(0.0445)	(0.2011)	(0.3373)	(0.0015)
Post election	0.0023	−1.2196^**	−1.1449^**	0.1085^**	−1.2035^**	−0.9601^**	0.0025
Post election	(0.0144)	(0.1506)	(0.1366)	(0.0420)	(0.1549)	(0.3647)	(0.0016)
Democracy	0.3328^**	14.2677^**	7.4550^**	−0.5505	7.0517^**	1.7744	−0.0154
Democracy	(0.0716)	(1.7722)	(1.6247)	(0.5150)	(1.6900)	(2.8486)	(0.0126)
Observations	85,470	6,718	5,669	7,121	5,569	2,698	6,471
Districts/Firms	3,643	3,441	3,203	3,445	3,199	2,048	1,663

Panel B:	Afrobarometer Data
	(8)	(9)	(10)	(11)	(12)	(13)	(14)
	Assets	Living conditions	Relative conditions	National economy	Direction of economy	Past conditions	Past country
Election year	0.0014	0.0320^**	0.0409^**	0.0119^†	0.0226^**	0.0489^**	0.0438^**
Election year	(0.0038)	(0.0061)	(0.0069)	(0.0061)	(0.0080)	(0.0111)	(0.0064)
Democracy	0.1940^**	0.2879^**	0.3728^**	0.1523^**	0.1137	0.5228^**	0.3688^**
Democracy	(0.0311)	(0.0503)	(0.0522)	(0.0490)	(0.0716)	(0.1198)	(0.0524)
Post election	0.0116^**	0.0140^*	−0.0023	0.0036	0.0865^**	−0.0109	0.0503^**
Post election	(0.0042)	(0.0067)	(0.0079)	(0.0062)	(0.0107)	(0.0117)	(0.0077)
Observations	7,160	7,588	6,159	7,588	5,456	3,429	7,537
Districts	1,771	1,774	1,743	1,774	1,760	1,441	1,774

Note. OLS models with unit (district/firm) and survey round fixed effects and country-specific time trends (for columns 1 and 7–14). Robust standard errors clustered by district/firm in parentheses. †p < .1, *p < .05, **p < .01.

We now turn to the interacted results for economic output and private goods in Table 2. The results in column 1 do not show any effect of democratization on luminosity measures, which could be the result of the fact that democratization does not have an effect on the nature of PBCs or, alternatively, that the luminosity measure cannot distinguish between different types of PBC outcomes. The results listed in columns 2–14 are less ambiguous and are consistent with our theoretical predictions: election years are correlated with higher levels of economic output as well as subjective and objective measures of individual welfare, with the coefficient on the interaction effect indicating that these effects decline with democracy.¹⁷ The lack of a relationship between elections and the number of employees in column 4 explains why both overall wages and wages per worker are higher in election years in non-democratic countries. The effects can be viewed for each outcome in Figures 1 and 2 at different levels of democracy: collectively they show higher levels of economic output and welfare in non-democracies, while in more democratic countries the effect either reverses or loses statistical significance. Importantly, as demonstrated by Figure A1 in the appendix which shows the distribution of the V-Dem democracy variable within the different datasets, there is common support of the moderating variable (democracy) in the available data (Hainmueller et al., 2019).¹⁸

Table 2.

Electoral Cycles in Economic Output/Private Goods – Conditional Effects

Panel A:	Lights Data	WBES Data					DHS Data
	(1)	(2)	(3)	(4)	(5)	(6)	(7)
	Nighttime light	Sales	Wages	Number of employees	Wages per worker	Assets purchases	Assets
Election year	0.0101	3.9539^**	3.7587^**	−0.0253	3.2708^**	3.9950^*	0.0133^**
Election year	(0.0286)	(0.5339)	(0.4840)	(0.1930)	(0.6431)	(1.7112)	(0.0050)
Post election	−0.0007	−0.8252^**	−0.7977^**	0.1054^*	−0.8949^**	−0.2365	0.0029^†
Post election	(0.0143)	(0.1496)	(0.1325)	(0.0450)	(0.1499)	(0.3947)	(0.0016)
Democracy	0.3095^**	11.8349^**	4.3450^**	−0.5347	4.3253^**	4.1016	−0.0092
Democracy	(0.0768)	(1.6046)	(1.4847)	(0.5155)	(1.6676)	(3.0796)	(0.0129)
Election year ×	0.0377	−8.8494^**	−7.9340^**	0.0683	−7.0235^**	−9.8232^**	−0.0258^**
Democracy	(0.0685)	(1.0459)	(1.0586)	(0.3297)	(1.3057)	(2.8832)	(0.0098)
Observations	85,470	6,718	5,669	7,121	5,569	2,698	6,471
Districts/Firms	3,643	3,441	3,203	3,445	3,199	2,048	1,663

Panel B:	Afrobarometer Data
	(8)	(9)	(10)	(11)	(12)	(13)	(14)
	Assets	Living conditions	Relative conditions	National economy	Direction of economy	Past conditions	Past country
Election year	0.0298^*	0.1331^**	0.0938^**	0.0631^**	0.1045^**	0.1336^**	0.1309^**
Election year	(0.0117)	(0.0193)	(0.0209)	(0.0183)	(0.0244)	(0.0351)	(0.0216)
Post election	0.0127^**	0.0174^**	−0.0015	0.0053	0.0904^**	−0.0080	0.0529^**
Post election	(0.0043)	(0.0067)	(0.0079)	(0.0063)	(0.0107)	(0.0114)	(0.0077)
Democracy	0.2302^**	0.4079^**	0.4336^**	0.2130^**	0.2039^**	0.5980^**	0.4711^**
Democracy	(0.0350)	(0.0547)	(0.0585)	(0.0531)	(0.0754)	(0.1232)	(0.0595)
Election year ×	−0.0513^*	−0.1853^**	−0.0945^**	−0.0938^**	−0.1511^**	−0.1677^**	−0.1597^**
Democracy	(0.0203)	(0.0327)	(0.0352)	(0.0313)	(0.0425)	(0.0619)	(0.0376)
Observations	7,160	7,588	6,159	7,588	5,456	3,429	7,537
Districts	1,771	1,774	1,743	1,774	1,760	1,441	1,774

Note. OLS models with unit (district/firm) and survey round fixed effects and country-specific time trends (for columns 1 and 7-14). Robust standard errors clustered by district/firm in parentheses. †p < .1, *p < .05, **p < .01.

Figure 1.

Electoral Cycles in Economic Output – Conditional Effects.

Figure 2.

Electoral Cycles in Private Goods – Conditional Effects.

We now turn to the results for public goods provision, first with the uninteracted effects in Table 3 and then interacted effects in Table 4. Here again our results are consistent with our theoretical predictions, with increasing levels of democracy associated with a decline in electricity outages and prices with the WBES data, an increase in access to electricity and water with the DHS data, and increases in government provision of electricity and water with the Afrobarometer data. The interaction effects can be viewed at different levels of democracy in Figure 3. Our results thus suggest that democratization leads to PBCs that focus more on the provision of public goods during election campaigns.

Table 3.

Electoral Cycles in Public Goods

	WBES Data			DHS Data		Afrobarometer Data
	(1)	(2)	(3)	(4)	(5)	(6)	(7)
	Price of electricity	Outage dummy	Number of outages	Access to electricity	Access to water	Gov. Electricity	Gov. Water
Election year	−0.9878^**	−0.0098	0.4398^**	−0.0127^**	−0.0099^†	−0.0090	0.0183^**
Election year	(0.2070)	(0.0268)	(0.0933)	(0.0047)	(0.0055)	(0.0070)	(0.0069)
Post election	−1.6361^**	−0.0538^*	−0.1301	0.0048	0.0054	0.0229^**	0.0120
Post election	(0.1565)	(0.0247)	(0.1072)	(0.0049)	(0.0062)	(0.0086)	(0.0075)
Democracy	11.3871^**	0.8790^**	−11.2512^**	0.0617^†	−0.0606	−0.0089	0.1557^**
Democracy	(1.9196)	(0.3351)	(1.2583)	(0.0363)	(0.0440)	(0.0515)	(0.0524)
Observations	5,686	7,049	5,421	6,356	6,485	6,351	7,550
Firms/Districts	3,261	3,444	3,238	1,662	1,663	1,767	1,774

Note. OLS models with unit (firm/district) and survey round fixed effects and country-specific time trends (for columns 4-7). Robust standard errors clustered by district/firm in parentheses. †p < .1, *p < .05, **p < .01.

Table 4.

Electoral Cycles in Public Goods – Conditional Effects

	WBES Data			DHS Data		Afrobarometer Data
	(1)	(2)	(3)	(4)	(5)	(6)	(7)
	Price of electricity	Outage dummy	Number of outages	Access to electricity	Access to water	Gov. Electricity	Gov. Water
Election year	4.4531^**	0.2910^*	0.9019^*	−0.0416^*	−0.0681^**	−0.1178^**	−0.0280
Election year	(0.7030)	(0.1252)	(0.4012)	(0.0176)	(0.0203)	(0.0212)	(0.0217)
Post election	−1.0580^**	−0.0354	−0.0909	0.0039	0.0034	0.0168^†	0.0103
Post election	(0.1818)	(0.0266)	(0.1051)	(0.0049)	(0.0062)	(0.0086)	(0.0076)
Democracy	3.7851	0.5521^†	−11.3893^**	0.0496	−0.0889^†	−0.1452^*	0.1010^†
Democracy	(2.4461)	(0.3315)	(1.2650)	(0.0375)	(0.0459)	(0.0584)	(0.0564)
Election year ×	−9.5971^**	−0.5234^*	−0.8512	0.0606^†	0.1211^**	0.2015^**	0.0847^*
Democracy	(1.3399)	(0.2144)	(0.6958)	(0.0358)	(0.0395)	(0.0372)	(0.0403)
Observations	5,686	7,049	5,421	6,356	6,485	6,351	7,550
Firms/Districts	3,261	3,444	3,238	1662	1663	1767	1774

Figure 3.

Electoral Cycles in Public Goods – Conditional Effects.

We conducted a number of additional robustness tests. First, we generated region-level rather than district-level panel data for the Afrobarometer and DHS data, which meant a smaller number of observations and a potentially cruder and less accurate measure of goods provision, which are nonetheless useful as it captures some more DHS and Afrobarometer surveys which have not been geocoded. As indicated in Tables A6 and A7, the results are largely identical. Second, we turn to an alternative measure of democracy, namely the Freedom House Aggregate (FHA) score, which, like the polyarchy index, similarly captures a broad measure of democracy that is based on both political rights and civil liberties. The FHA is also a fine-grained measure, with a 100-point scale, and its coverage extends from 2005 to 2022, which is ideal for the WBES and Afrobarometer data but means a loss of 30% of the DHS observations.¹⁹ The results, which can viewed in Tables A8 and A9, are almost identical to our main results. Third, we also changed our twelve-month pre- and post-election windows for the Afrobarometer and DHS data sets to fifteen months, with no changes in our results (see Tables A10 and A11).

In other robustness tests we added additional controls to our Afrobarometer analysis, namely a set of individual controls for age and gender and then a district-round level control for the proportion of the respondents who identify with the President’s ethnic group, with no changes in any case. We then broke down the election year variable into parliamentary and presidential elections separately, with the caveats that some countries that do not have presidential elections like Botswana and Ethiopia drop out of the data set in the latter case, and that most countries in the data set hold parliamentary and presidential elections simultaneously. Here our results are largely consistent with our main set of results, indicating that both parliamentary and presidential elections are important in generating incentives for governments to provide more goods in election years. Finally, we interacted a variety of other factors that could play intermediary roles in the provision of private and public goods, namely GDP per capita, district-level urbanization, the district-level proportion of the respondents who identified with ethnic group(s) identified by the Ethnic Power Relations dataset as a Senior Partner in government, and the adoption of a balanced budget rule (Strong, 2023); in all cases the interaction terms were inconsistent across the various outcomes while our main results held.²⁰

Conclusion

In this paper we used multiple types of data to investigate political business cycles across different types of political regimes in Sub-Saharan Africa. We found clear evidence that PBCs exist across multiple regimes but that the nature of PBCs change with democratization, such that less democratic regimes focus more on the provision of private goods that stimulate the economy during election years, while more democratic regimes are focused instead on the provision of public goods such as electricity. Our work represents the first attempt to demonstrate the changing nature of PBCs across different political regimes, whether in Sub-Saharan Africa or elsewhere.

Our results have multiple implications within political science. First and foremost, our results suggest that democratization has a profound effect on the nature of PBCs in Sub-Saharan Africa by providing incentives for governments to focus more on the provision of public rather than private goods. While it is easy to remain cynical about the effects of election campaigns in developing countries as regards a focus on short-term benefits with longer-term costs such as sub-national governmental unit creation or deforestation (Green, 2010; Sanford, 2023), our analysis suggests that democratization can instead spur governments to provide goods that have both short and long term benefits to governments and citizens. Our results thus add to recent literature that provides some optimism to those who remain concerned that elections and democratization in Africa and elsewhere fail to generate incentives for public goods provision (Harding, 2020; Kao & Rakner, 2022).

Second, our results add to the literature on the relationship between democracy and goods provision. As noted above, a substantial body of literature has already focused on how and why democratization can lead to better public goods provision (Deacon, 2016; Harding & Stasavage, 2014; Kudamatsu, 2012; Lake & Baum, 2001), but, with the exception of Min (2015)’s work on electricity, did not focus attention on the timing of these interventions. In other words, there remains much more scope to examine not just the ‘what’ and ‘where’ of the relationship between democracy and the provision of private and public goods, but also the ‘when’ as regards their timing in relation to elections.

A third implication of our work is that electoral politics in Africa is perhaps more similar to other parts of the world than it is often described in the literature. A strong focus in recent years on how ethnic favouritism determines the location of public spending in Africa (Burgess et al., 2015; Franck & Rainer, 2012) has arguably led to a neglect of more basic concerns about the timing of private and public goods provision. Here we show that PBCs not only exist in Africa but follow a logic that is unrelated to ethnicity, and that may be applicable across other parts of the world.

There are at least two avenues of future research which could build upon our results. First, it would be interesting to see if our results hold for local elections, for which there is substantial evidence from single case studies from the developed and developing world on the existence of PBCs (Aidt et al., 2011; Foremny & Riedel, 2014; Labonne, 2016; Veiga & Veiga, 2007). Compiling a data set on local elections in Africa is challenging, not least because local governments differ widely in the powers they hold over the provision of public and private goods, but such an exercise has potential to enhance our understanding of how different types of elections have differential effects on goods provision.

A second such avenue is by using similar types of survey data from other parts of the world that cover periods of substantial variation in the level of democracy to see if our results hold elsewhere. Other PBC literature which has used survey data includes Bhattacharjee (2022), who finds evidence for a decline in neonatal mortality in India in the year before state legislative assembly elections, while Aidt et al. (2020) examine the relationship between elections and food consumption in Armenia. A useful next step would be to extend studies such as these to cover cases that offer variation in regime type. The Latinobarometer and Asiabarometer are two of many potential data sources that could be used to this effect, allowing for a broader understanding of the relationship between PBCs and democratization across the developing world.

Supplemental Material

Supplemental Material - Political Business Cycles and Democratization: Evidence From Sub-Saharan Africa

Supplemental Material for Political Business Cycles and Democratization: Evidence From Sub-Saharan Africa by Elliott Green and Robin Harding in Comparative Political Studies

Footnotes

Acknowledgments

We thank seminar participants at the LSE and the University of Oxford and panel attendees of the Annual Meeting of the European Political Science Association in Glasgow for comments and suggestions. All errors remain our own.

ORCID iDs

Elliott Green

Robin Harding

Funding

The authors received no financial support for the research, authorship, and/or publication of this article.

Declaration of Conflicting Interests

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

Data Availability Statement

Replication materials and code can be found at .

Supplemental Material

Supplemental material for this article is available online.

Notes

Author Biographies

Elliott Green is a Professor of Development Studies in the Department of International Development at the London School of Economics and Political Science. His first book Industrialization and Assimilation: Explaining Ethnic Change in the Modern World was published with Cambridge University Press in 2022, and he is currently completing a second book manuscript on the politics of ethnic identity in modern Africa. His recent work has been published in journals such as the British Journal of Political Science, Comparative Political Studies and World Development, among others. He holds degrees from Princeton University (AB) and the LSE (MSc, PhD).

Robin Harding is a Professor of Politics in the Department of Politics and International Relations at the University of Oxford, and the Gillian Peele Fellow in Politics at Lady Margaret Hall. He holds a PhD from New York University. His research focuses on the political economy of development, with a particular interest in the relationship between political institutions and public policy outcomes, in areas such as education, healthcare, and the environment.

References

Aaskoven

Lassen

D. D.

(2017). Political budget cycles. Oxford Research Encyclopedia: Politics.

Abdul-Gafaru

Sackeyfio

(2022). Introduction: The uncertainties of Ghana’s 2020 elections. African Affairs, 121(484), e25–e53. https://doi.org/10.1093/afraf/adaa028

Abrahamsen

Bareebe

(2016). Uganda’s 2016 elections: Not even faking it anymore. African Affairs, 115(461), 751–765. https://doi.org/10.1093/afraf/adw056

Ahlborg

Boräng

Jagers

S. C.

Söderholm

(2015). Provision of electricity to African households: The importance of democracy and institutional quality. Energy Policy, 87, 125–135. https://doi.org/10.1016/j.enpol.2015.09.002

Aidt

Asartyan

Badalyan

Heinemann

(2020). Vote buying or (political) business (cycles) as usual? The Review of Economics and Statistics, 102(3), 409–425. https://doi.org/10.1162/rest_a_00820

Aidt

T. S.

Veiga

F. J.

Veiga

L. G.

(2011). Election results and opportunistic policies: A new test of the rational political business cycle model. Public Choice, 148(1-2), 21–44. https://doi.org/10.1007/s11127-010-9644-3

Alesina

(1987). Macroeconomic policy in a two-party system as a repeated game. Quarterly Journal of Economics, 102(3), 651–678. https://doi.org/10.2307/1884222

Alt

J. E.

Lassen

D. D.

(2006). Transparency, political polarization, and political budget cycles in OECD countries. American Journal of Political Science, 50(3), 530–550. https://doi.org/10.1111/j.1540-5907.2006.00200.x

Asatryan

Baskaran

Birkholz

Gomtsyan

(2022). Favoritism and firms: Micro evidence and macro implications. CESifo. CESifo Working Paper 9797.

10.

Asatryan

Castellón

Stratmann

(2018). Balanced budget rules and fiscal outcomes: Evidence from historical constitutions. Journal of Public Economics, 167, 105–119. https://doi.org/10.1016/j.jpubeco.2018.09.001

11.

Asiama

Akosah

Owusu-Afriyie

(2014). An assessment of fiscal sustainability in Ghana. Bank of Ghana. Bank of Ghana Working Paper WP/BOG-2014/09.

12.

Bak

(2016). Political investment cycles in democracies and autocracies. International Interactions, 42(5), 797–819. https://doi.org/10.1080/03050629.2016.1173547

13.

Baskaran

Min

Uppal

(2015). Election cycles and electricity provision: Evidence from a quasi-experiment with Indian special elections. Journal of Public Economics, 126, 64–73. https://doi.org/10.1016/j.jpubeco.2015.03.011

14.

Bhattacharjee

(2022). The timing of elections and neonatal mortality: Evidence from India. The World Bank Economic Review, 36(4), 972–998. https://doi.org/10.1093/wber/lhac020

15.

Blattman

Larreguy

Marx

Reid

O. R.

(2019). Eat widely, vote wisely? Lessons from a campaign against vote buying in Uganda. National Bureau of Economic Research. NBER Working Paper 26293.

16.

Blaydes

(2011). Elections and distributive politics in Mubarak’s Egypt. Cambridge University Press.

17.

Bleck

van de Walle

(2018). Electoral politics in Africa since 1990: Continuity in change. Cambridge University Press.

18.

Block

S. A.

(2002). Political business cycles, democratization, and economic reform: The case of Africa. Journal of Development Economics, 67(1), 205–228. https://doi.org/10.1016/s0304-3878(01)00184-5

19.

Bobbo

Bate

A. A. G. M.

(2022). Political budget cycles in Sub-Saharan Africa: New empirical evidence. Journal of Public Finance and Public Choice, 37(1), 149–169. https://doi.org/10.1332/251569121x16397576745954

20.

Bonfiglioli

Gancia

(2013). Uncertainty, electoral incentives and political myopia. Economic Journal, 123(568), 373–400. https://doi.org/10.1111/ecoj.12029

21.

Bratton

Bhavnani

Chen

T.-H.

(2012). Voting intentions in Africa: Ethnic, economic or partisan? Commonwealth & Comparative Politics, 50(1), 27–52. https://doi.org/10.1080/14662043.2012.642121

22.

Brender

Drazen

(2005). Political budget cycles in new versus established democracies. Journal of Monetary Economics, 52(7), 1271–1295. https://doi.org/10.1016/j.jmoneco.2005.04.004

23.

Briggs

R. C.

(2012). Electrifying the base? Aid and incumbent advantage in Ghana. The Journal of Modern African Studies, 50(4), 603–624. https://doi.org/10.1017/s0022278x12000365

24.

Brun

D. A.

Diamond

(2014). Clientelism, social policy, and the quality of democracy. JHU Press.

25.

Bueno De Mesquita

Smith

Siverson

R. M.

Morrow

J. D.

(2005). The logic of political survival. MIT Press.

26.

Burgess

Jedwab

Miguel

Morjaria

Padró i Miquel

(2015). The value of democracy: Evidence from road building in Kenya. The American Economic Review, 105(6), 1817–1851. https://doi.org/10.1257/aer.20131031

27.

Canes-Wrone

Park

J.-K.

(2012). Electoral business cycles in oecd countries. American Political Science Review, 106(1), 103–122. https://doi.org/10.1017/s0003055411000529

28.

Canes-Wrone

Park

J.-K.

(2016). Elections, uncertainty and irreversible investment. British Journal of Political Science, 4(1), 83–106. https://doi.org/10.1017/S000712341200049X

29.

Chang

E. C.

(2008). Electoral incentives and budgetary spending: Rethinking the role of political institutions. The Journal of Politics, 70(4), 1086–1097. https://doi.org/10.1017/s0022381608081073

30.

Chen

Nordhaus

W. D.

(2011). Using luminosity data as a proxy for economic statistics. Proceedings of the National Academy of Sciences, 108(21), 8589–8594. https://doi.org/10.1073/pnas.1017031108

31.

Cole

M. A.

Elliott

R. J.

Occhiali

Strobl

(2018). Power outages and firm performance in Sub-Saharan Africa. Journal of Development Economics, 134, 150–159. https://doi.org/10.1016/j.jdeveco.2018.05.003

32.

Conroy-Krutz

Logan

(2012). Museveni and the 2011 Ugandan election: Did the money matter? The Journal of Modern African Studies, 50(4), 625–655. https://doi.org/10.1017/s0022278x12000377

33.

Deacon

R. T.

(2016). Public good provision under dictatorship and democracy. Public Choice, 139(1-2), 241–262. https://doi.org/10.1007/s11127-008-9391-x

34.

Dubois

(2016). Political business cycles 40 years after nordhaus. Public Choice, 166(1), 235–259. https://doi.org/10.1007/s11127-016-0313-z

35.

Ebeke

M. C.

Ölcer

(2013). Fiscal policy over the election cycle in low-income countries. International Monetary Fund.

36.

Eibl

Lynge

(2017). Constraints, competition, and competitiveness: Explaining the non-linear effect of democratization on political budget cycles. European Political Science Review, 9(4), 629–656. https://doi.org/10.1017/s1755773916000205

37.

Fernández

I. C.

(2017). Kilowatts, megawatts and power: Electric territorialities of the state in the peripheries of Ghana and Tanzania . PhD Dissertation . Centre of African Studies, University of Edinburgh.

38.

Foremny

Riedel

(2014). Business taxes and the electoral cycle. Journal of Public Economics, 115, 48–61. https://doi.org/10.1016/j.jpubeco.2014.04.005

39.

Franck

Rainer

(2012). Does the leader’s ethnicity matter? Ethnic favoritism, education, and health in Sub-Saharan Africa. American Political Science Review, 106(2), 294–325. https://doi.org/10.1017/s0003055412000172

40.

Gandhi

Lust-Okar

(2009). Elections under authoritarianism. Annual Review of Political Science, 12(1), 403–422. https://doi.org/10.1146/annurev.polisci.11.060106.095434

41.

Gandhi

Przweworski

(2006). Cooperation, cooptation and rebellion under dictatorships. Economics & Politics, 18(1), 1–26. https://doi.org/10.1111/j.1468-0343.2006.00160.x

42.

Gibson

Olivia

Boe-Gibson

(2021). Which night lights data should we use in economics, and where? Journal of Development Economics, 149, Article 102602. https://doi.org/10.1016/j.jdeveco.2020.102602

43.

Green

(2010). Patronage, district creation, and reform in Uganda. Studies in Comparative International Development, 45(1), 83–103. https://doi.org/10.1007/s12116-009-9058-8

44.

Green

Harding

(2026). Replication data for: Political business cycles and democratization: Evidence from Sub-Saharan Africa. https://doi.org/10.7910/DVN/FLHHSF.

45.

Haber

(2006). Authoritarian government. In The Oxford handbook of political economy (pp. 693–707). Oxford University Press.

46.

Hainmueller

Mummolo

(2019). How much should we trust estimates from multiplicative interaction models? Simple tools to improve empirical practice. Political Analysis, 27(2), 163–192. https://doi.org/10.1017/pan.2018.46

47.

Han

(2022). Political budgetary cycles in autocratic redistribution. Comparative Political Studies, 55(5), 727–756. https://doi.org/10.1177/00104140211036038

48.

Harding

(2015). Attribution and accountability: Voting for roads in Ghana. World Politics, 67(4), 656–689. https://doi.org/10.1017/s0043887115000209

49.

Harding

(2020a). Rural democracy: Elections and development in Africa. Oxford University Press.

50.

Harding

(2020b). Who is democracy good for? Elections, rural bias, and health and education outcomes in Sub-Saharan Africa. The Journal of Politics, 82(1), 241–254. https://doi.org/10.1086/705745

51.

Harding

Stasavage

(2014). What democracy does (and doesn’t do) for basic services: School fees, school inputs, and African elections. The Journal of Politics, 76(1), 229–245. https://doi.org/10.1017/s0022381613001254

52.

Henderson

J. V.

Storeygard

Weil

D. N.

(2012). Measuring economic growth from outer space. The American Economic Review, 102(2), 994–1028. https://doi.org/10.1257/aer.102.2.994

53.

Hibbs

D. A.

(1977). Political parties and macroeconomic policy. American Political Science Review, 71(4), 1467–1487. https://doi.org/10.2307/1961490

54.

Higashijima

(2016). Political business cycles in dictatorships. Waseda University. WIAS Discussion Paper 2016-002.

55.

Higashijima

Kerr

(2023). When does the honeymoon end? Electoral cycles of satisfaction with democracy in Africa. Political Psychology, 44(4), 709–728. https://doi.org/10.1111/pops.12878

56.

Hodler

Raschky

P. A.

(2014). Regional favoritism. Quarterly Journal of Economics, 129(2), 995–1033. https://doi.org/10.1093/qje/qju004

57.

Jäger

(2016). The role of regime type in the political economy of foreign reserve accumulation. European Journal of Political Economy, 44, 79–96. https://doi.org/10.1016/j.ejpoleco.2016.05.001

58.

Julio

Yook

(2012). Political uncertainty and corporate investment cycles. The Journal of Finance, 67(1), 45–83. https://doi.org/10.1111/j.1540-6261.2011.01707.x

59.

Kao Kristen

L. E.

Rakner

(2022). Vote-buying, anti-corruption campaigns, and identity in African elections. World Development, 160, Article 106064. https://doi.org/10.1016/j.worlddev.2022.106064

60.

Keola

Andersson

Hall

(2015). Monitoring economic development from space: Using nighttime light and land cover data to measure economic growth. World Development, 66, 322–334. https://doi.org/10.1016/j.worlddev.2014.08.017

61.

Knutsen

C. H.

Nygård

Wig

(2017). Autocratic elections: Stabilizing tool or force for change? World Politics, 69(1), 98–143. https://doi.org/10.1017/s0043887116000149

62.

Kramon

(2018). Money for votes: The causes and consequences of electoral clientelism in Africa. Cambridge University Press.

63.

Kroth

Larcinese

Wehner

(2016). A better life for all? Democratization and electrification in post-apartheid South Africa. The Journal of Politics, 78(3), 774–791. https://doi.org/10.1086/685451

64.

Kudamatsu

(2012). Has democratization reduced infant mortality in Sub-Saharan Africa? Evidence from micro data. Journal of the European Economic Association, 10(6), 1294–1317. https://doi.org/10.1111/j.1542-4774.2012.01092.x

65.

Kyriacou

A. P. T. O.

Roca-Sagalés

(2022). Conditional political budget cycles: The role of time preference. Economics & Politics, 34(1), 67–91. https://doi.org/10.1111/ecpo.12187

66.

Labonne

(2016). Local political business cycles: Evidence from Philippine municipalities. Journal of Development Economics, 121, 56–62. https://doi.org/10.1016/j.jdeveco.2016.03.004

67.

Lake

D. A.

Baum

M. A.

(2001). The invisible hand of democracy: Political control and the provision of public services. Comparative Political Studies, 34(6), 587–621. https://doi.org/10.1177/0010414001034006001

68.

Lindbeck

(1976). Stabilization policy in open economies with endogenous politicians. The American Economic Review, 66(2), 1–19. https://www.jstor.org/stable/1817192.

69.

Lipscomb

Mobarak

A. M.

Barilam

(2013). Development effects of electrification: Evidence from the topographic placement of hydropower plants in Brazil. American Economic Journal: Applied Economics, 5(2), 200–231. https://doi.org/10.1257/app.5.2.200

70.

Lokshin Michael

A. R.-F.

Torre

(2024). Electoral cycles and public spending during the pandemic. Review of Development Economics, 28(3), 1077–1107. https://doi.org/10.1111/rode.13094

71.

Long

J. D.

Gibson

C. C.

(2015). Evaluating the roles of ethnicity and performance in African elections: Evidence from an exit poll in Kenya. Political Research Quarterly, 68(4), 830–842. https://doi.org/10.1177/1065912915608946

72.

Lynge

i Coma

F. M.

(2022). The effect of economic downturns on voter turnout in Africa. Electoral Studies, 76, Article 102456. https://doi.org/10.1016/j.electstud.2022.102456

73.

MacLean

L. M.

Gore

Brass

J. N.

Baldwin

(2016). Expectations of power: The politics of state-building and access to electricity provision in Ghana and Uganda. Journal of African Political Economy and Development, 1(1), 103–134. https://hdl.handle.net/10520/EJC-c3d04fbe8.

74.

Magaloni

(2008). Voting for autocracy: Hegemonic party survival and its demise in Mexico. Cambridge University Press.

75.

Mandon

Cazals

(2019). Political budget cycles: Manipulation by leaders versus manipulation by researchers? Evidence from a meta-regression analysis. Journal of Economic Surveys, 33(1), 274–308. https://doi.org/10.1111/joes.12263

76.

Marin

Persson

Tabellini

G. E.

(1990) Macroeconomic Policy, Credibility and Politics (Vol. 38). Taylor & Francis.

77.

Marx

(2018). Elections as incentives: Project completion and visibility in African politics. Department of Economics, Boston University.

78.

Min

(2015). Power and the vote: Elections and electricity in the developing world. Cambridge University Press.

79.

Mosley

Chiripanhura

(2016). The African political business cycle: Varieties of experience. Journal of Development Studies, 52(7), 917–932. https://doi.org/10.1080/00220388.2015.1113263

80.

Neumann

Ssozi

(2016). Political influence on fiscal and monetary policy in Sub-Saharan Africa. Journal of African Economies, 25(1), 55–109. https://doi.org/10.1093/jae/ejv017

81.

Nordhaus

W. D.

(1975). The political business cycle. The Review of Economic Studies, 42(2), 169–190. https://doi.org/10.2307/2296528

82.

Osei-Tutu

Weill

(2023). Regional favoritism in access to credit: Just believe it. BOFIT Discussion Paper 1/2023. University of Stasbourg.

83.

Pepinsky

(2007). Autocracy, elections, and fiscal policy: Evidence from Malaysia. Studies in Comparative International Development, 42(1-2), 136–163. https://doi.org/10.1007/s12116-007-9006-4

84.

Persson

Tabellini

(2005). The economic effects of constitutions. MIT press.

85.

Philips

A. Q.

(2016). Seeing the forest through the trees: A meta-analysis of political budget cycles. Public Choice, 168(3-4), 313–341. https://doi.org/10.1007/s11127-016-0364-1

86.

Posner

D. N.

Simon

D. J.

(2002). Economic conditions and incumbent support in Africa’s new democracies: Evidence from Zambia. Comparative Political Studies, 35(3), 313–336. https://doi.org/10.1177/0010414002035003003

87.

Przeworski

(2004). Institutions matter? Government and Opposition, 39(4), 527–540. https://doi.org/10.1111/j.1477-7053.2004.00134.x

88.

Rhee

(2021). Economic perception to political performance evaluation: Establishing precursors to economic voting in Africa. Political Research Quarterly, 74(1), 131–147. https://doi.org/10.1177/1065912919888017

89.

Rogoff

(1990). Equilibrium political budget cycles. The American Economic Review, 80(1), 21–36. https://www.jstor.org/stable/2006731.

90.

Rud

J. P.

(2012). Electricity provision and industrial development: Evidence from India. Journal of Development Economics, 97(2), 352–367. https://doi.org/10.1016/j.jdeveco.2011.06.010

91.

Sanford

(2023). Democratization, elections, and public goods: The evidence from deforestation. American Journal of Political Science, 67(3), 748–763. https://doi.org/10.1111/ajps.12662

92.

Shi

Svensson

(2006). Political budget cycles: Do they differ across countries and why? Journal of Public Economics, 90(8-9), 1367–1389. https://doi.org/10.1016/j.jpubeco.2005.09.009

93.

Shmuel

(2020). The political budget cycle across varying degrees of democracy. Electoral Studies, 68, Article 102218. https://doi.org/10.1016/j.electstud.2020.102218

94.

Simpser

(2013). Why governments and parties manipulate elections: Theory, practice and implications. Cambridge University Press.

95.

Strong

C. O.

(2023). Tying one’s hand: The effect of fiscal rules on the political business cycle in Africa. Journal of African Economies, 132(4), 438–467. https://doi.org/10.1093/jae/ejac014

96.

Strong

C. O.

(2024). Conditional political cycles in Africa: Myth or reality? Journal of African Economiess, 33(5), 489–514. forthcoming. https://doi.org/10.1093/jae/ejad019

97.

Thomas

Darsey

(2024). How electoral cycles shape the implementation of public works programs: Evidence from India. Political Research Quarterly, 77(1), 106–120. https://doi.org/10.1177/10659129231195079

98.

Trotter

P. A.

(2016). Rural electrification, electrification inequality and democratic institutions in Sub-Saharan Africa. Energy for Sustainable Development, 34, 111–129. https://doi.org/10.1016/j.esd.2016.07.008

99.

Tufte

E. R.

(1978). Political control of the economy. Princeton University Press.

100.

Veiga

F. J.

Veiga

L. G.

(2007). Political business cycles at the municipal level. Public Choice, 131(1-2), 45–64. https://doi.org/10.1007/s11127-006-9104-2

101.

Veiga

F. J.

Veiga

L. G.

Morozumi

(2017). Political budget cycles and media freedom. Electoral Studies, 45, 88–99. https://doi.org/10.1016/j.electstud.2016.11.008

102.

Vogt

Bormann

N.-C.

Rüegger

Cederman

L.-E.

Hunziker

Girardin

(2015). Integrating data on ethnicity, geography, and conflict: The ethnic power relations data set family. Journal of Conflict Resolution, 59(7), 1327–1342. https://doi.org/10.1177/0022002715591215

103.

Vokes

Wilkins

(2016). Party, patronage and coercion in the NRM’s 2016 re-election in Uganda: Imposed or embedded? Journal of Eastern African Studies, 10(4), 581–600. https://doi.org/10.1080/17531055.2016.1279853

104.

Wang

Y.-T.

Mechkova

Andersson

(2019). Does democracy enhance health? New empirical evidence 1900–2012. Political Research Quarterly, 72(3), 554–569. https://doi.org/10.1177/1065912918798506

105.

Wright

(2011). Electoral spending cycles in dictatorships. Department of Political Science, Pennyslvania State University. Unpublished paper.

106.

Youde

(2005). Economics and government popularity in Ghana. Electoral Studies, 24(1), 1–16. https://doi.org/10.1016/j.electstud.2004.02.003

Supplementary Material

Please find the following supplemental material available below.

For Open Access articles published under a Creative Commons License, all supplemental material carries the same license as the article it is associated with.

For non-Open Access articles published, all supplemental material carries a non-exclusive license, and permission requests for re-use of supplemental material or any part of supplemental material shall be sent directly to the copyright owner as specified in the copyright notice associated with the article.

1.28 MB

0.00 MB