Central transfers and incentives to collect local revenue among the Central Region of Ghana’s local government officials: analysing the flypaper effect

Abstract

Local governments that rely heavily on central transfers are likely to face a laxity problem when increasing locally generated revenues. The flypaper effect identifies this overdependence by examining the relationship between local expenditures and revenues (central transfers and locally generated). Local governments in Ghana depend on central transfers and locally generated revenues to finance local development. Central transfers are important because they balance fiscal capacities across local governments and promote equalisation in the provision of public goods. The study used panel data estimation to analyse revenue, expenditure, distance, and population from 2009 to 2020 for 17 Ghanaian local governments in the Central Region. The results show the presence of the flypaper effect, a situation explained by the fiscal interest model to reduce the effectiveness and efficiency of local government officials as they tend to be less responsive to the needs of the local people they serve. Furthermore, the presence of flypaper effects reduces local governments’ incentives to increase the mobilisation of local revenue. This paper recommends policies that strengthen local government financial autonomy and, consequently, economic development.

Introduction

Local governments require a steady and sufficient supply of revenue to achieve their development mandate. Public finance literature has identified two primary sources of revenue for local governments: internally generated revenue and central transfers, known as the District Assembly Common Fund (DACF) in Ghana (Dick-Sagoe, 2013). Internally generated revenue includes both taxable and non-taxable income collected within the jurisdiction of a local government (Dick-Sagoe, 2017), whereas central transfers are monies transferred from the central government to local governments. The rationale for intergovernmental fiscal transfers is rooted in public finance principles that seek to address vertical and horizontal fiscal imbalances. The vertical imbalance refers to disparities in revenue and expenditure responsibilities between central and local governments, whereas the horizontal imbalance addresses discrepancies in revenue and expenditure across different local governments. Therefore, central transfers are expected to promote financial equity and efficiency at the local level.

Theoretically, central transfers have two primary objectives: political and economic (Rao and Singh, 1998; Brollo et al., 2013; Shah, 2007; Singhal, 2008; Oates, 2003; Dick-Sagoe and Tingum, 2021; Dick-Sagoe et al., 2022). Economically, the transfers should mitigate structural inequalities by ensuring equitable resource distribution, thereby reducing revenue and expenditure imbalances across the local governments. In practice, central transfers have been instrumental in addressing both vertical and horizontal imbalances, helping to achieve budgetary parity among local governments and curbing cross-jurisdictional spillovers. However, despite their intended economic benefits, political considerations often overshadow their primary economic objectives, making central transfers a tool for political control rather than a means to enhance fiscal efficiency and decentralisation (Dick-Sagoe et al., 2022).

Politically, central transfers are often leveraged by politicians to influence local governments for their own vested political interests or advantage rather than serving the needs of the people. Instead of enhancing decentralisation and local autonomy, these transfers create socio-economic dependencies, enabling the central government to impose strict conditions that limit local discretion. In federal systems, for instance, central transfers are used to regulate local government spending and enforce national priorities at the local level. This undermines the core principles of decentralisation by restricting local governments’ ability to determine their own developmental agendas.

A significant challenge associated with central transfers is the flypaper effect, which suggests that local government expenditures respond more to increased central transfers than to an equivalent rise in locally generated revenue (Hines and Thaler, 1995). The presence of the flypaper effect implies that local governments become overly reliant on central funding, leading to inefficiencies in public service delivery generally (Weingast, 2013). The flypaper effect is more pronounced in local governments with low tax autonomy, where officials prioritise central transfers over efforts to improve local revenue generation (Ferreira et al., 2019; Rodden, 2003). Consequently, decentralisation, which is supposed to enhance local accountability and efficiency, is undermined when local governments heavily depend on external funding rather than locally generated revenue.

Despite the presence of a large body of studies on flypaper effects (Hines and Thaler, 1995; Bailey and Connolly, 1998), the presence of flypaper effects is considered by public finance researchers as an anomaly because of its inconsistency with the ‘equivalence theorem’ (Bradford and Oates, 1971). Theoretically, decentralisation fosters greater accountability as local governments are directly responsible to their constituents for public service delivery (Ferreira, 2021). According to public management and public choice theory, decentralisation is championed by Tiebout’s (1956) ‘inter-jurisdictional competition,’ Musgrave’s (1959) ‘solution to the assignment problem,’ Oates’ (1972) ‘decentralisation theorem’ and Rubinfeld (1987), all of which promote individual freedom and the right to local autonomy from the central government (Gruening, 2001). This autonomy is crucial for ensuring that local revenue generation efforts align with the specific needs and preferences of local communities (Osei-Owusu, 2022).

Second-generation decentralisation theorists, such as Weingast, argue that reliance on locally generated revenue is essential for providing pro-poor services tailored to local needs. Since local taxes are collected from residents, they foster accountability by ensuring that local government officials remain transparent about how public funds are utilised in order to sustain collection of revenue from them (Dick-Sagoe, 2013). In contrast, when local governments depend on central transfers, they become more accountable to the central government rather than to their constituents (Dick-Sagoe and Tingum, 2021; Kang and Setyawan, 2012; Dick-Sagoe and Tingum, 2021; Poschl and Weingast, 2013; Rao and Singh, 1998); and this weakens local participation in governance and reduces incentives for prudent financial management at the local level.

Another critical aspect of government transfers is the distinction between conditional and unconditional transfers. Although unconditional transfers grant local governments greater flexibility in allocating funds based on their priorities, conditional transfers, such as those in Ghana, impose strict expenditure guidelines that align with central government priorities. Ghana’s District Assemblies Common Fund (DACF), established under Article 252 of the 1992 Constitution, requires that at least 5% of national revenue be allocated to local governments. Although this policy supports local development, it has also been linked to a decline in local revenue generation efforts. Empirical evidence suggests that local governments in Ghana tend to rely more on central transfers than on locally generated revenue, reinforcing the flypaper effect. For example, a study analysing 17 local governments in Ghana’s Central Region between 2008 and 2015 found that central government transfers accounted for a larger share of local government expenditures than locally generated revenue (Dzansi et al., 2019).

On the other hand, conditional transfers, as in Ghana, restrict expenditure at the local level to central government priorities. Central government transfers in Ghana are usually accompanied by strict implementation guidelines that give specific proportions of funds to spend on human capital development, infrastructure, and other areas. Although these are global development priorities, the rates at which local governments need one over the other vary. Therefore, unconditional transfers from the central government violate the purpose of decentralisation.

Between 2013 and 2017, data from the Ministry of Local Government and Rural Development further indicated a decline in local revenue generation, compelling many local governments to heavily depend on central transfers (Delali, 2019). The same study reveals that local governments with higher central transfers make higher expenditures. However, this does not necessarily translate to improved public services or development outcomes. This suggests inefficiencies in resource allocation due to the flypaper effect. Therefore, a study that seeks to investigate the relationship between central transfers and local revenue collection in Ghana’s Central Region is crucial for understanding the fiscal behaviour of local governments. Addressing the flypaper effect can enhance fiscal autonomy, improve public service delivery, and promote sustainable local economic development in Ghana.

The study was therefore motivated by the potential challenges that may arise from the presence of the flypaper effect and its broader implications for decentralisation. Overreliance on central government transfers may lead to moral hazard, weakening local governments’ ability to maintain sound fiscal policies and implement necessary and indigenously conceived structural reforms. This is supported by several empirical works that show that local government dependence reduces incentives for innovation and productivity, thereby limiting economic growth (Dick-Sagoe and Tingum, 2021; Lu and Wu, 2022). Moreover, the flypaper effect can undermine local governments’ independence, making them instruments of central government control rather than agents of local development. More importantly, when local governments depend on central transfers, it weakens their sense of accountability to constituents, who may not have so much vested interest as opposed to when constituents contribute more revenue to local governments.

Using panel data models, this study empirically examines the flypaper effects of central transfers on the local government expenditure of the Central Region of Ghana’s local governments. With the narrative derived from the flypaper effect that ‘money sticks where it hits,’ the study examines whether central transfers significantly result in disproportionately higher local government expenditures compared to locally generated revenues from the Central Region of Ghana.

Furthermore, this study is significant considering the growing dependence of local governments on central transfers, which threatens self-sufficiency and fiscal sustainability. By filling the empirical gap on the flypaper effect in Ghana, this study provides insights into why central transfers from the Central Region of Ghana have not translated into improved service delivery and governance. The findings also offer policy recommendations for strengthening local revenue generation efforts and reducing excessive dependence on central government transfers. Ultimately, this study seeks to promote a decentralised governance framework that enhances fiscal autonomy, accountability, and local economic development in Ghana.

Sources of local government funding in Ghana

Ghana’s current decentralised local government system started in 1998 and was supported by the Provisional National Defence Council’s (PNDC’s) Law 207, which was later given legal backing in the 1992 constitution of the 4th Republic of Ghana, chapter 20, and the Local Government Act of Ghana (Act 462, 1993). However, the Local Government Act, Act 462 of 1993, has been replaced with the Local Government Act, Act 936 of 2016. The constitution provides room for the establishment of metropolitan, municipal, and district assemblies (MMDAs) with administrative and political powers to ensure the total development of their jurisdictions.

Regarding the finances of local governments (Metropolitan, Municipal and District Assemblies, MMDAs), the constitution of Ghana mandates the central government to allocate 5% of the total revenues of Ghana and set aside as to be distributed to the local governments in Ghana based on a parliamentary-approved sharing formula (Osei-Owusu, 2022). For example, in 2022 fiscal year 2022, 5% of Ghana’s total tax revenue, amounting to GH 3,342,970,849.00, was disbursed to all local governments in Ghana according to the Parliament of Ghana’s approved sharing formula (Osei-Owusu, 2022). In addition, local governments in Ghana are constitutionally mandated to collect imposed rates and non-tax and tax revenues. This non-tax revenue includes licence rates, fees, and fines (Dick-Sagoe and Tingum, 2021; Otoo and Danquah, 2021).

Literature review

Theory

The theories that describe the relationship between central transfers and local government expenditures are the public choice theory, fiscal interest model, and flypaper effect model. This section explains these theories.

Public choice theory

The theory of public choice is central to the notion that individuals follow their personal interests and behave according to their desires. Slavinskaitė, Liučvaitienė, and Gedvilaitė (2019) summarised the objective of public choice into two parts: efficiency of public services and strengthening local governments’ responsibility. Promoting individual freedom allows the public choice theory to achieve the following two principles (Slavinskaite, 2017): diversity and equivalence.

Diversity argues that the differences among the jurisdictions of local governments reflect the corresponding differences in the provision of services in each jurisdiction. Hence, the same model in service provision argues that the political objective it seeks to achieve does not address the argument of allocative efficiency. Equivalence argues against the strict ‘one size fits all’ provision of public services under the political objective, where all local governments are given particular instruments to follow in spending resources allocated by the central government. Instead, equivalence argues that each type of service should be allocated, assigned, and implemented at the local level, which is the level that requires the most. By doing so, efficiency is achieved.

According to public-choice scholars, the inefficient use of local governments’ resources and the suppression of the interests of local people result from traditional budgeting by central government officials and executives within bureaucratic organisations. Such bureaucratic organisations (defined as organisations that are partially or entirely inaccessible on markets) have various shortcomings. This is what the political objective brings to decentralised local governments, rendering them unable to perform their real economic objective.

Fiscal interest model

According to Wallis, Sylla, and Legler (1994), local government officials are biassed towards policies that seek to increase revenues. Dependence on locally generated revenues (taxes) generates the following interests for local governments (Weingast, 2013):

1. Local governments that depend on locally generated revenues tend to focus on providing services that satisfy residents’ needs (known as allocative efficiency). By doing so, they maintain or increase residents’ tax compliance. They tend to be more accountable to residents for their actions and financial resources. Greater tax compliance from the local residents results in higher local revenue mobilisation for local governments. Timmons (2005) established a positive statistical relationship between voluntary compliance with local taxes and the provision of local public services valued by citizens. Local governments that depend on local revenue (raised through local taxes) tend to represent their residents’ interests when making decisions and provide local public goods and services that the population needs. This forms the basis of the decentralisation concept.

2. Citizens tend to scrutinise the activities of local government operations if they are made to pay taxes on the local government. Citizens then emphasise their representation in local governments’ expenditure on services that enhance their growth and development (Bird, 2010). As put it, citizens will have no business checking the activities of local governments if all the financial resources for development are transferred from the central government to the local government.

3. It has been observed that the government tends to be more careful in the expenditure of taxable revenues than ‘unearned’ revenues from government transfers. Grants make local governments in developing countries less accountable to local people their fiscal decisions. They then increase spending without increasing taxes, a circumstance that eliminates the incentive to enhance local government operations and create cutting-edge strategies for providing local public goods.

Rodden’s (2003) study revealed a significant difference between local governments and their dependence on central transfers. The observation was that local governments differ significantly with respect to their dependence on central government transfers and those that depend on local revenues. Studies conducted by various researchers indicate that local governments tend to be very accountable and responsive to local needs if they depend solely on local revenues.

Providing more social services, reducing corruption, and swifter provision of local public services are considered important factors to enhance compliance in taxes payments. To increase revenues, local governments pay particular attention to their citizens and provide services that enhance their development—receiving public goods and services whose value corresponds with high levels of voluntary tax compliance (Timmons, 2005). Greater tax compliance is intrinsically linked to higher government revenues. Simply put, local governments that rely on their own income are incentivised to be more responsive to their constituents, offer improved local public goods and services, and are less likely to be corrupt (Ambrosio and Borgignon, 2008; Careaga and Weingast, 2003; Rodden, 2003; Singh and Srinivasan, 2006).

Supporting the explanation of fiscal interest theory, this study uses a case involving property taxes. For example, if a local government depends on property taxes and wants to increase its revenue, it is advisable for that local government to provide local public goods that can improve the value of property within their jurisdiction. Inter-jurisdictional competition from other local governments will serve as an incentive to maximise property to bring in scarce capital and labour to locate and stay in their jurisdiction. On that note, the local government receives revenue to support its budget, which then forces the local government to collect more revenues and provide local public goods in line with the demands of the taxpayers. This is the fundamental argument for the fiscal interest model.

Flypaper effect

The flypaper effect was introduced in 1960 by James Henderson and Edward Gramlich, who were interested in establishing a correction between demographics and economics on the one hand and government budgets on the other. They developed a theoretical model that argued that money is money and that personal income (locally generated revenue) and central transfers (share of government grant) should have the same impact on local government spending (Inman, 2008).

The flypaper effect refers to the greater stimuli effect observed on unconditional grants on the expenditure of local government than increases in locally generated revenue (Aragon, 2009; Shah, 2007). Therefore, rising local government expenditure because of higher central transfers than an equivalent increase in locally generated revenue sources is referred to as the ‘flypaper effect.’ This effect challenges the traditional grant-in-aid theory (Oates, 1999). Prediction from the standard approach states that central transfers to local governments match equivalent increases in local governments’ locally generated revenues (Bradford and Oates, 1971; Gamkhar and Shah, 2007; Aragon, 2009; Kaur et al., 2021). This prediction of the standard approach is termed the ‘veil hypothesis,’ as it posits that central transfers serve as a veil for the central government’s tax refunds (Oates, 1999). Flypaper effects have been developed to explain the evident conflict between empirical observations and theory prediction (Hines and Thaler, 1995). Thus, the flypaper effect is treated as an irregularity inconsistent with economic theory (Inman, 2008).

Although an extensive body of literature has empirically confirmed the effects of flypaper (Witterblad, 2007; Dahlberg et al., 2006; Dick-Sagoe and Tingum, 2021; Kaur, Mohanty, Chakraborty and Rangan, 2021), others, such as Aragon (2009), report that the effect remains paradoxical. Political and bureaucratic factors also account for the flypaper effects as political agents maximise their own budgets to increase their influence over the localities within their jurisdiction (Brollo et al., 2013; Rao and Singh, 1998; Shah, 2007; and Singhal, 2008).

Empirical evidence of flypaper effects in developing countries

A handful of studies exist on the determinants of the flypaper effect, some measuring the effect across authorities and others measuring the effect of transfers on taxes. Empirical reviews of central and intergovernmental transfers to local expenditure are discussed below. These studies have been conducted in South Africa (Amusa et al., 2008), Indonesia (Kang and Setyawan, 2012), Turkey (Sa and Saruç, 2004), Brazil (Ferreira, 2021), Germany, and Ghana (Dick-Sagoe and Tingum, 2021). In the first two cases (South Africa and Indonesia, the authors observed a lack of flypaper effects, whereas the study in Brazil, Germany, Turkey, and Ghana observed the presence of flypaper effects. The details of their study are presented in this section.

Amusa et al. (2008) investigated fiscal illusion in the local sphere by using South African Municipal data to test flypaper effects. The authors argue that there is a gap between the revenue sources of local governments in South Africa and their expenditure responsibilities, which is addressed through intergovernmental transfers. The authors further cautioned that overdependence on intergovernmental transfers by local governments has been the resulting challenge. To achieve this, the authors used revenue and expenditure data from 237 local governments in South Africa from 2005 to 2006. Their study resulted in the lack of a flypaper effect, as the marginal effect of local governments’ own-source revenues on local revenues exceeded that of intergovernmental transfers.

Similarly, Kang and Setyawan (2012) studied the flypaper effect of intergovernmental transfers on Indonesian municipalities from 2001 to 2008. Specifically, the authors were interested in the impact of their own-source revenue and intergovernmental transfers on local expenditures and how the flypaper concept mediates these relationships. Their study was conducted in two parts. The first part used cross-sectional data from 188 local governments from 2006 to 2008. The second part uses panel data from 484 local governments from 2001 to 2008. Their results proved that own-source revenue and intergovernmental transfers (partially and simultaneously) have a significant impact on local expenditure. By comparing the two effects (intergovernmental transfers and own-source revenue) on local revenue, the authors found that the effect of own-source revenue on local expenditure was stronger than that of intergovernmental transfers. This indicates the absence of flypaper effects in the responses of local governments to intergovernmental transfers.

Chakraborty and Jha (2019) employed panel data regression models, including fixed and random effects, to analyse the relationship between central transfers and the revenue efforts of Indian states. This study finds that central transfers have a significant impact on Indian states’ revenue efforts, suggesting that these transfers can influence the fiscal behaviour of state governments. In contrast, Dash and Raja (2022). employs panel data analysis techniques to assess the impact of fiscal autonomy on public expenditure performance, controlling for various state-specific factors, and concluded that higher fiscal autonomy is associated with improved public expenditure performance. This indicates that when states have greater control over their revenues, they tend to manage public funds more efficiently.

In 2004, Sa and Saruç (2004) showed that grants in Turkey caused the flypaper effect, despite a linear budget constraint in Turkey, and that the population and grant increase rate influence the flypaper effect even more. Much later, Baskara (2016) employed an instrumental variable design with panel data to show that the Hessian municipalities use transfers to increase expenditures, which is consistent with the presence of the flypaper effect. They then concluded on the existence of the flypaper effect in German local fiscal equalisation. Recently, Ferreira (2021) observed the presence of flypaper effects, which the author considers to be part of fiscal illusion, in Brazilian local governments from 2005 to 2012. The author explains flypaper effects, which is in line with what the present study considers, as receiving local governments converting intergovernmental transfers to local public expenditures at a significantly higher rate than that for local revenue. Ferreira (2021) conducted three distinctive and complementary studies on flypaper effects. The first study examined flypaper effects on 476 Brazilian local governments, where flypaper effects were present for the 476 local governments from 2005 to 2012.

For the first analysis, the author included variables that influence grants, such as mayor’s coalition with state governor and federal president, party, and political alignment, where the local government tax autonomy index was used as a control variable. Tax autonomy was defined as the percentage of local revenue relative to total local government revenues. The second study analysed 5568 Brazilian local governments from 2006 to 2013 using the tax base elasticity of local governments’ tax rates as a proxy for the marginal cost of public funds. The second study also confirmed the presence of the flypaper effect. Finally, the third study analysed 27 Brazilian states from 1985 to 2010. In addition, 5568 Brazilian local governments were analysed from 2000 to 2018. The result from the third study proved the stimulative effect of intergovernmental transfers on local government expenditures, thus reflecting the presence of flypaper effects.

In a recent study, close to flypaper effects in Ghana, Otoo and Danquah (2021) observed the impact of fiscal decentralisation on the efficiency of local public services and goods in Ghana’s 216 local governments (comprising of Metropolitan, Municipal, and District Assemblies) in 2021. What made their study relevant for this study was the presence of a specific objective that sought to assess the contributions of own-source revenues to local expenditure on projects and services in Ghana. Using parametric and non-parametric frontier methods, the authors resorted to the composite budget of all 216 local governments in Ghana for their study. Based on the two indicators of fiscal decentralisation (central transfers and local revenue), the authors observe that the share of local revenue in total local expenditure (technically called fiscal autonomy) has a positive influence on the efficiency of local service provision. On the other hand, and more relevant for this study, a high proportion of central transfer to local expenditure does not improve the efficiency of local service provision. The findings indicate a need to give high priority to policy initiatives aimed at enhancing local revenue collection. Therefore, the question that arises here is why central transfers, which dominate local revenues and expenditures, do not improve efficient service provision, whereas local revenue does? Does this revelation make the position of researchers in the fiscal interest model relevant?

Despite the relevance of investigating the flypaper effect, the majority of studies on the flypaper effect have been conducted in countries outside Africa. The need to improve the development of local government units in developing countries, such as African countries, cannot be overemphasised. However, few studies have been conducted in South Africa and Ghana. The study in Ghana by Dick-Sagoe and Tingum (2021) used panel data for 17 local governments from 2008 to 2015 with the aid of the Generalised Least Squares (GLS) to examine the fiscal behaviour of local government officials when presented with intergovernmental fiscal transfers and own-source revenues. This study differs from that of Dick-Sagoe and Tingum (2021) and most of the previous empirical works in that it employs a more robust methodology—the system GMM that caters for endogeneity and updates the data to cover 2009 to 2020.

Methodology

There exist various empirical studies that have applied different econometric techniques to examine the impact of intergovernmental transfers on local government expenditures, such as the fixed effects (FE) and random effects (RE) models, as well as the system generalised method of moments (GMM). Several studies have employed these methodologies to address heterogeneity and endogeneity concerns in panel data analysis, particularly in investigating the flypaper effect across different countries or local governments.

A key advantage of the fixed-effects model is that it controls for unobserved heterogeneity by allowing for correlation between individual-specific effects and independent variables (Hsiao, 2014). According to them, this ensures that time-invariant characteristics do not bias the estimates, making the model particularly useful for examining the fiscal behaviour of local governments. Interestingly, since local governments have unique institutional and economic structures that remain relatively constant over time, the FE model could be well-suited for capturing these specific influences. However, one limitation of the FE model is that it does not estimate the effects of time-invariant regressors, which makes it less suitable when such variables are of interest. Notable studies that employed the fixed-effects model include Amusa et al. (2008) for South African municipalities and Dick-Sagoe and Tingum (2021) for Ghanaian local governments. These studies found varying effects of intergovernmental transfers on local government expenditure, with Amusa et al. (2008) observing no flypaper effect, whereas Dick-Sagoe and Tingum (2021) found evidence supporting the flypaper effect in Ghana.

Secondly, the random-effects model assumes that individual-specific effects are uncorrelated with explanatory variables, making it more efficient than the FE model when this assumption holds (Wooldridge, 2010). The RE model is advantageous when studies have a large number of cross-sectional units and relatively few time periods because it allows for generalisable conclusions beyond the sampled entities. Kang and Setyawan (2012) on Indonesian municipalities utilised the random-effects model to explore the fiscal response of local governments to intergovernmental transfers. Their findings suggested a stronger effect of own-source revenue on local expenditures, indicating the absence of the flypaper effect. A key methodological consideration in determining when to use the FE or RE for a given dataset is the Hausman test. If the test rejects the null hypothesis, the FE model is preferred; otherwise, the RE model is used. However, both the FE and RE do not cater for endogeneity concerns arising from reverse causality or omitted variable bias.

It is on this premise that this study employs only the FE and RE as the control methodologies and uses the system GMM (S-GMM) estimator to address these concerns. The S-GMM approach is widely used in fiscal policy studies because of its ability to handle endogenous regressors through instrumental variables. For example, Arellano and Bover (1995) and Blundell and Bond (1998) introduced the S-GMM to improve efficiency by incorporating both first-differenced and level equations in a system framework. The S-GMM estimator integrates a system that involves both first-difference regression and level regression. This entails instrumenting differences with lags of levels, and level variables are instrumented with lags of differences.

The S-GMM has gained global recognition in panel data estimation, partly because of its capability to address endogeneity issues and concerns related to omitted variables, thereby enhancing the accuracy of parameter estimates (Soto, 2009; Roodman, 2009). Employing Monte Carlo simulations, Soto (2009) demonstrated that in smaller sample sizes (less than 100), the S-GMM estimator exhibits lower bias and greater efficiency compared with all the other panel estimators examined, including the conventional first-differences GMM estimator. The S-GMM technique is particularly relevant for studies analysing fiscal decentralisation and intergovernmental transfers because it accounts for the dynamic nature of public finance data (Roodman, 2009). Notably, Ferreira (2021) applied the S-GMM method to examine the flypaper effect in Brazilian municipalities and found strong evidence of its presence.

Model specification

Following the models for estimating the flypaper effect on Ghana’s Central Region local governments follow a panel data approach of fixed/random effects, generalised least squares (GLS), and a system generalised method of moments (S-GMM). These models account for the unobserved heterogeneity among the 17 local governments in their estimation and therefore produce efficient, reliable, and stable estimators (Greene, 2012).

The key dependent variable is total expenditure, which represents the composite expenditure of local governments. With log-linearisation (which linearises the data and gives straightforward interpretation to the coefficients (elasticities)) and the addition of other control variables, the empirical panel model is specified as follows:

1

$\begin{array}{c}\log t{\exp }_{it}={\beta }_0+{\beta }_1\log transfer{s}_{it}+{\beta }_2\log ownre{v}_{it}\hfill \\ +{\beta }_{3}logpop+{\beta }_{4}logdistance+{\alpha }_i+{\delta }_t+{\mu }_{it}\hfill \end{array}$Where $\mathit{texp}$ is the total expenditure by the local governments; transfers is the grants/transfers from central governments to local governments, ownrev is the revenue raised by the local governments, $\mathit{pop}$ is the population of the local government areas within the central region of Ghana and $\mathit{distance}$ is the physical distance between the central government and the local government receiving the funds. According to Seabright (1996), remote municipalities may have fewer regular interactions with central authorities, potentially leading to fewer checks on spending, allowing local governments more discretion over how grants are used. The subscript $\mathit{t}$ stands for ‘time’ and i stand for the local governments, which implies that ${\alpha }_i$ and ${\delta }_t$ are local government-fixed and time-fixed effects, while μ_it are the random errors, representing the errors linked with the ith group in period (t) ${\beta }_0$ stands for the intercept and ${\beta }_1$ to ${\beta }_4$ stands for the coefficients of the variables. The per capita terms were used at 2015 prices and deflated using the consumer price index.

Panel estimates of the flypaper effects are probably biased because the flow of intergovernmental transfer from central to local governments is expected to autocorrelate over time. This may lead to serial autocorrelation and heteroscedasticity among the estimates, which is confirmed using the Hausman Specification test (See Table 4). To address the concerns of serial autocorrelation and heteroscedasticity, we employed the Generalised Least Squares (GLS) estimator. The GLS technique estimates the unknown parameters in a linear regression model (Eq. (1) above) when there is a certain degree of heteroscedasticity and correlation between the residuals in the model (Green, 2012).

To explore the dynamic relationship between total revenue, transfers, and own revenue, this study introduces the lagged levels of total expenditure in Eq. (1). A dynamic model is more appropriate because total expenditure is probably highly persistent over time. In this case, it is reasonable to assume that total current expenditure depends on past expenditures. The dynamic equation that captures this relationship is specified as follows:

2

$\begin{array}{c}\log t{\exp }_{it}={\beta }_0+{\beta }_{1}logtex{p}_{it-1}+{\beta }_2\log transfer{s}_{it}\hfill \\ +{\beta }_3\log ownre{v}_{it}++{\beta }_{4}logpop\hfill \\ +{\beta }_{5}logdistance+{\alpha }_i+{\delta }_t++{\mu }_{it}\hfill \end{array}$

Equation (2) is estimated using the system generalised method of moments (S-GMM) initiated by Hotz et al.

Data and variables

To empirically estimate the fly-paper effect, the study employed a panel of 17 local governments with data from 2012 to 2019 sourced from the official website of the Ministry of Finance of the government of Ghana, under the fiscal data section. The data are constructed at the ministerial level based on financial reports from municipalities that allow for revenue disintegration between revenue transferred from the central government (grants) and revenue raised through taxation and other sources. Data on population and distance from the central government (Accra) to local government areas were obtained from the official website of the Ghana Statistical Service (GSS). The study was limited to 2019 due to the disintegration of some local governments and the unavailability of data on some variables beyond that period. Nevertheless, the ample number of time series and cross-sectional observations provided sufficient degrees of freedom, ensuring that the methodology produced robust and reliable results.

The 17 local governments covered are Cape Coast, Ag. Swedru, Dunkwa, Saltpond, Winneba, Elmina, Assin Fosu, Twifo Praso, A. Beraku, Diaso, Apam, A. Dunkwa, Nsaba, Afransi, Ajumako, Nsuaem, and B. Asikuma, representing 74% of the Central Region of Ghana’s local governments. Thus, the sample size is balanced, meaning that the same number of data points is used for each local government. The variables used are central government transfers to local governments (also known as ‘central transfers’), local governments’ revenues (locally generated revenues), and districts’ population. Table 1 presents the descriptive statistics of the variables. The correlation matrix in Table 2 shows weak average relationship between the variables.

Table 1. Descriptive statistics of the nominal variables.

Statistics	Texp	Transfers	Ownrev	Population	Distance
Mean	2725893	2698824	555733.9	136214.8	146.3118
Median	2488749	1996862	298490.8	129641	129.3
Std. Dev.	1691275	4525201	880752.6	49229.8	78.31936
Minimum	1.17E + 05	95889.22	8228.89	68713	65.5
Maximum	1.02E + 07	6.67E + 07	7748625	342103	357.8
Skewness	1.896873	10.51546	4.885265	1.764636	1.481758
Kurtosis	8.404974	118.5173	35.47882	7.045682	4.393571
Observation	136	136	136	136	136

Authors’ calculations, 2023.

Table 2. Pairwise correlation matrix of the variables.

	Texp	Transfers	Ownrev	Population	Distance
Texp	1.0000
Transfers	0.4250	1.000
	(0.0000)
Ownrev	0.3362	−0.0574	1.0000
	(0.0001)	(0.5072)
Population	0.1547	0.1478	−0.0810	1.0000
	(0.0722)	(0.0860)	(0.3485)
Distance	−0.0600	−0.0592	0.0301	−0.3750	1.0000
	(0.4875)	(0.4937)	(0.7276)	(0.0000)

Authors’ calculations, 2023.

A rigorous test of Hausman’s (1978) specification is estimated to choose between the FE and RE models appropriately. In the Hausman test, the null hypothesis is as follows: H0: random effects are more consistent; against the alternative hypothesis: H1: only fixed-effect estimates are consistent. Thus, we reject the null hypothesis if a significant figure is recorded for the Hausman test in favour of the fixed-effects model being consistent Table 3.

Table 3. Panel results.

	(1)	(2)	(3)	(4)	(5)	(6)
Variables	Pooled OLS		Fixed-effects model		Random-effects model
Cotransfers	0.471***	0.434***	0.393***	0.354***	0.467***	0.414***
	(0.048)	(0.061)	(0.058)	(0.073)	(0.048)	(0.063)
Logownrev	0.306***	0.210***	0.294***	0.220***	0.322***	0.216***
	(0.042)	(0.050)	(0.056)	(0.071)	(0.043)	(0.054)
Logpop	−0.054	−0.079	1.804**	1.001	−0.021	−0.066
	(0.152)	(0.146)	(0.755)	(0.866)	(0.182)	(0.179)
Distance	−0.001	−0.001			−0.000	−0.001
	(0.001)	(0.001)			(0.001)	(0.001)
2013. year		−0.210		−0.212		−0.206
		(0.166)		(0.164)		(0.161)
2014. year		0.111		0.095		0.118
		(0.168)		(0.170)		(0.164)
2015. year		0.302*		0.263		0.308*
		(0.173)		(0.182)		(0.169)
2016. year		0.428**		0.246		0.399**
		(0.202)		(0.235)		(0.202)
2017. year		0.137		0.050		0.141
		(0.174)		(0.196)		(0.171)
2018. year		0.340*		0.287		0.356*
		(0.190)		(0.222)		(0.189)
2019. year		0.385**		0.319		0.402**
		(0.195)		(0.234)		(0.195)
Constant	4.705**	6.579***	−15.976*	−5.145	4.163*	6.625***
	(1.908)	(1.979)	(8.127)	(9.898)	(2.211)	(2.332)
Observations	136	136	136	136	136	136
R-squared	0.507	0.573	0.519	0.566
Number of groups with ID			17	17	17	17

Standard errors in parentheses.

***p < 0.01, **p < 0.05, *p < 0.1.

Table 4 presents the results of the Hausman test. The results show that the two regressions reject the null hypothesis in favour of the fixed-effects model, which is more consistent. Time and year dummies are not included in the estimation of the first regression, whereas regression 2 adds year and time dummies.

Table 4. Hausman test results.

Regression	Test	Chi2	Prob	Decision
1	Random vs. Fixed effect	19.35	0.0002	Reject H0
2	Random vs. Fixed effect	35.28	0.0000	Reject H0

Authors’ calculations, 2023.

Because of the time-series and cross-sectional dimensions of panel data, estimations usually encounter the problem of serial and heteroscedasticity correlation (Beck, 2001; Hsiao, 2007). As widely discussed by Beck (2001), problems may arise within the same region as a result of shocks in one local government area spilling over to another. When this occurs, one local government’s shocks are transferred to another in the same year. In addition, shocks in the current year may be correlated with shocks in the previous year within the same local government. Therefore, the study used the Modified Wald Test (Baum, 2001) to test for heteroscedasticity and the Wooldridge Test (Drukker, 2003) to test for serial autocorrelation. Table 5 presents the results of these analyses.

Table 5. Panel heteroscedasticity and autocorrelation.

Regression					Decision
Modified Wald test for heteroscedasticity in the FE model
1	F (16, 111)	2.10	chi2 (17)	7164.41	Heteroscedastic
	Prob > F, where	0.0125	Prob > chi2	0.0000
2	F (16, 109)	1.67	chi2 (17)	824.43	Heteroscedastic
	Prob > F, where	0.0623	Prob > chi2	0.0000
Wooldridge test for autocorrelation in panel data
1	F (3, 16)	415.86	F (1, 16)	2.523	Serial correlation
	Prob > F, where	0.000	Prob > F, where	0.0318
2	F (2, 16)	8.37	F (1, 17)	11.726	Serial correlation
	Prob > F, where	0.0082	Prob > F, where	0.0006

Authors’ calculations, 2023.

The value recorded for Prob > F statistics was used to interpret the test results. The null hypothesis (which states that there is no heteroscedasticity) is rejected in the heteroscedasticity test, indicating the presence of heteroscedasticity in the models.

The null hypothesis was further rejected by the Wooldridge test for serial correlation. The conclusion is that serial correlation exists in the error terms of the models. As a result of serial autocorrelation and heteroscedasticity, the generalised least squares (GLS) estimator is applied alongside the S-GMM.

Empirical analysis

The results in Table 5 produce biased estimates in the presence of heteroscedasticity, serial or residual autocorrelation (Keele and Kelly, 2006), and/or endogeneity. To account for potential sources of bias and robustness, this study employed both the GLS and Blundell–Bond System GMM (SGMM) models (Blundell et al., 2000). The GLS produced optimal unbiased results in the presence of serial and heteroscedasticity correlations in panel data (Muthama, 201; Hansen, 2007). The SGMM estimation exploits both the lagged differences of the endogenous variables and the lagged levels of the equation as instruments to circumvent endogeneity.

The results of the GLS and dynamic panel (S-GMM) models are presented in Table 6. The results in the first two columns are for the GLS model, whereas the last two columns are for the S-GMM model with total expenditure as the dependent variable. Columns 2 and 4 present results for time or year dummies. The diagnostic tests of these models suggest that the Wald chi2 value for the four models is significant at the 1% level of significance, indicating that the independent variables altogether explain the dependent variable. Regarding the S-GMM, the two models validate the AR (2) tests, as designated by their p values, which are all above 0.05 and therefore not significant at a 5.0% level of significance. This implies that the serial correlation of the error term is not second-order. The number of instruments for the two models is 12 and 16, respectively, which is less than the number of local governments (17). covered in the study. To further demonstrate the validity of the instruments, the p-value from the Hansen over-identification test is presented. The p values were >0.05 (insignificant p values), indicating that the S-GMM instruments were robust and validated. Overall, we conclude that the estimators are sufficiently robust and reliable for policy inferences on flypaper effects.

Table 6. Parameter estimates of the GLS and S-GMM models.

	(1)	(2)	(3)	(4)
Variable	GLS model		S-GMM model
Logtransfers	0.471***	0.434***	0.432***	0.344**
	(0.047)	(0.059)	(0.057)	(0.154)
Logownrev	0.306***	0.210***	0.375***	0.227
	(0.041)	(0.048)	(0.122)	(0.341)
Logpop	0.054	0.079	0.435	1.744*
	(0.149)	(0.140)	(0.631)	(0.990)
Distance	−0.001	−0.001	0.008	-0.005
	(0.001)	(0.001)	(0.014)	(0.005)
2013. year		−0.210
		(0.158)
2014. year		0.111		0.099
		(0.160)		(0.226)
2015. year		0.302*		0.146
		(0.165)		(0.303)
2016. year		0.428**		−0.149
		(0.193)		(0.482)
2017. year		0.137		−0.101
		(0.166)		(0.369)
2018. year		0.340*		0.035
		(0.182)		(0.480)
2019. year		0.385**		−0.031
		(0.186)		(0.535)
Logtexp (−1)			0.196**	0.155
			(0.093)	(0.131)
Constant	4.705** (1.872)	6.579*** (1.890)	−5.413 (6.690)	−17.815 (14.377)
Test for AR(1) errors (z)			−1.90 [0.058]	−1.66 [0.096]
Test for AR(2) error -z			0.09 [0.929]	0.45 [0.652]
Sargan overidentification test: chi2(8)			15.19 [0.056]	15.00 [0.059]
Hansen overidentification test: chi2(8)			6.22 [0.622]	5.68 [0.683]
Number of instruments			12	16
Wald square test	139.85 (0.000)	182.19 (0.000)	30858.78 (0.000)	1480.89 (0.000)
Observations	136	136	119	119
Number of groups with ID	17	17	17	17

Standard errors in (…), p values in […]: ***p < 0.01, **p < 0.05, *p < 0.1.

The results show that the coefficients of central government transfers are all significant and consistently greater than the coefficient of one’s own revenue, except for model 4, in which the coefficient of one’s own revenue is insignificant. At the 1% level of significance, the coefficients of locally generated revenue and central transfers are both positive and significant. Under the GLS model, a 1% change in central transfers significantly causes a 0.471% and 0.434% change in local governments’ total expenditure (with and without year dummies, respectively). This assumption implies that higher central transfers to local governments for a particular year match higher local governments’ expenditure for the same year. On the other hand, a 1% increase in revenue from local taxes changes local governments’ total expenditure by 0.306% and 0.210%, respectively. Therefore, higher local own-source revenues for a particular year match higher local governments’ expenditure for the same year. In Sijabat (2016), Ğbafi and Saruçş (2004), and Kang & Setyawan (2012), the authors observed similar results.

The same pattern emerges in Table 6 when the S-GMM model is estimated. First, the estimated (positive and significant) effects of transfers on total expenditures are larger in magnitude than the effect of revenue with or without time dummies. This argument is supported by an array of existing literature that established a positive and significant relationship between central government transfers and total expenditure in developed countries with a relatively higher magnitude than the revenue generated by local governments (Canare, 2019; Dick-Sagoe and Tingum, 2021; and Dick-Sagoe et al., 2022). However, the estimated coefficient for own revenue contributes positively to total expenditure, but does not significantly. The one-period lag of total expenditure significantly and positively affects total expenditure in the S-GMM model without year dummies. The results agree with a priori theoretical and empirical expectations. Increases in the previous year’s total expenditure stimulate higher current expenditure through the expenditure multiplier effect. This finding implies that an initial increase in spending cycles repeatedly throughout the economy has a larger impact than the initial amount spent. These results are empirically supported by Masaki.

The coefficients for the main control variable were compared when all variables (own source revenue and central transfers) were regressed simultaneously, while time, expressed in terms of effects of years, was controlled. For the period considered in this study, the coefficient of central transfers is higher than the coefficient of the country’s own source revenue. This finding demonstrates that central transfers make significant contributions to total local expenditures rather than their own revenues. This instance clearly proves the presence of the flypaper effect because central government transfers are higher in determining total expenditure than local government-generated revenue.

According to fiscal interest models, the flypaper effect reverses decentralisation, thereby causing re-centralised decentralised local governments (Dick-Sagoe and Tingum, 2021). This results in local governments being accountable and loyal to central governments rather than to their local residents. Local governments respond better to the needs and instructions of the central government to attract more funding. Thus, the magnitude of central transfer becomes higher than its source revenue in local government finances (referred to as the ‘flypaper effect’). According to Gorodnichenko (2001), the flypaper effect manifests in two ways: first, it results in a rise in local taxes and exorbitant local government budget spending. Second, it results in a higher elasticity in expenditure to central transfers than in the elasticity of expenditure to local tax revenue of local governments. Therefore, this study supports the flypaper effect.

The implications of having a significantly higher impact of central transfers relative to own transfers lie in the accountability and effectiveness of these grants, as well as meeting the priority needs of local governments. Politicians at the local government level prefer central transfers because there is less oversight and commitment to monitoring than there is with locally generated revenue, where the contributors reside in the environment and can monitor all activities. OECD (2015) submitted that when citizens pay taxes, they have a direct stake in how the government spends their money, which creates a greater sense of accountability and transparency in the use of tax revenues. This level of sensibility is higher for individual revenues than for central transfers, although some central transfers are also derived from taxes at the national level. Second, when central transfers have higher impacts, one of the objectives of decentralisation, which is to allow local-level governments to govern themselves based on their priorities, is defeated. This is because central government transfers in Ghana are usually conditional transfers that project the priorities of the central government over the local government.

For the other control variables (distance from central government (Accra) to local government areas and population of the local government areas), the distance variable is insignificant for all the models. The population variable is significant for the S-GMM model while controlling for year dummies, which implies that a 1% increase in population size increases the total expenditure of the local government by 1.74%. Thus, an increase in population generates an increase in local public expenditures because the degree of rivalry between local public goods is high and the range of local public goods is broad. This is based on the observation by Oates (1999) that highly populated areas or cities require and provide a broader range of services than smaller ones, causing an increase in total expenditure.

These findings highlight the urgent need for a more robust and equitable system of fiscal federalism in Ghana. Currently, local governments remain largely subservient to the central government, resulting in an imbalance in intergovernmental financial relations that disproportionately favours the central government. To address this, the central government must reduce its dominance and financial control while empowering local government bodies. Such reforms would not only bring governance closer to the people but also enhance accountability, ensuring that local officials are answerable to their constituents. Additionally, it would mitigate the central government’s political manipulation of local governance, fostering a more transparent and effective decentralised system.

The interpretation of the result from this study shows that the performance of local governments increases when there is autonomy in revenue generation and decision-making. In this case, local government officials act in the best interest of their constituents in accordance with the public choice theory. The current situation shows that local governments in Ghana are more likely to tailor public services to meet central governments’ priorities, which leads to misallocation of resources and inefficiencies. Further, the finding speaks to the fiscal interest model from the reduced incentive to mobilise local revenues, implying that local governments invest little effort in mobilising local revenues in the presence of central transfers. This suggests that local governments in Ghana may be less responsive to the needs of their communities, leading to inefficiencies in service delivery and fiscal complacency. Furthermore, the presence of the flypaper effect indicates a lack of institutional capacity for effectively mobilising local revenue. As a result, local governments become heavily reliant on central transfers, making them vulnerable to political manipulation, where financial allocations are used to sway political loyalty and influence local decision-making.

Conclusion and policy implications

Based on the theoretical positions of public choice theory, the fiscal interest model, and the flypaper effect, this paper argues that more fiscal autonomy (share of local revenue in total local expenditure) remains critical for local governments to be efficient in local service delivery. This statement was also made in 2021 by Otoo and Danquah, who stated that fiscal autonomy improves the efficiency of local service provision in Ghana. The more local governments can finance their local expenditures from their local revenue (meaning less dependence on central transfers), the more fiscally autonomous they become, and this improves the efficiency of local service delivery.

The ability of local governments to determine their own development priorities without any restrictions or controls from the central or state government was deduced from the public choice theory. Such autonomy promotes the active participation of local people in local development decisions that reflect local priorities, as argued by the fiscal interest model. Thus, the active participation of local people in the process of local decision-making and the high demands for accountability put local government officials on their toes to work better because they are under frequent monitoring by the local people. This creates the ideal environment for decentralisation to work, as epitomised by the delivery of efficient and effective pro-poor services, as supported by studies by Pöschl and Weingast (2013), Timmons (2005), Brollo et al. (2013), Gadenne (2017), Dick-Sagoe, Tingum, and Asare-Nuamah (2022), and Otoo and Danquah (2021: 415). To these authors, higher dependence on own-source revenues leads to increased fiscal autonomy (measured as the percentage of own-source revenues to local expenditure). Higher fiscal autonomy for local governments has a positive impact on the efficiency of local service provision because local governments with higher fiscal autonomy tend to behave well in terms of local expenditure to ensure compliance with local taxes and fees, thus ensuring efficient use of public resources. On the other hand, local governments that depend on unearned sources of revenues (intergovernmental sources of revenues) tend to use such revenues in a way that does not ensure compliance with local taxes and fines from the local people (Brollo et al., 2013; Gadenne, 2017). The explanation for this behaviour is that the payment of local taxes and fines motivates citizens of local governments through their civil organisations to demand accountability from local government authorities. This will put the authorities of local governments on their toes to enhance the provision of local services and goods that meet the needs of local citizens and encourage the payment and compliance of local taxes and fines. The presence of flypaper effects further imply that the assignment of revenue sources and tax structures across local governments in Ghana significantly affects local decision-making processes on fiscal matters. This is because local expenditure is linked to central transfers (the flypaper effects); thus, any fiscal reform should focus on the revenue sources assigned to local governments.

Musgrave, Oates and Tiebout argued in favour of decentralisation with the mindset that making the central government less bulky, by breaking down some of its functions and giving them out to the lesser forms of government through decentralisation, will make service delivery efficient and effective (Musgrave, 1959; Tiebout, 1956; Oates, 1972). However, the works of the new generation of decentralisation thinkers, such as those from the fiscal incentive model and the flypaper effects, argue that decentralisation can only promise efficient and effective service delivery and attain its ideals, as held by Musgrave (1959), Tiebout (1956) and Oates (1972), when its design is critically considered, as political actors can use the same design of decentralisation to work in their favour. One of the many tools political actors use is the design of central transfers (also referred to as intergovernmental transfers). This study investigated this phenomenon by using local governments in Ghana’s central region to demonstrate the likelihood of political influence, as observed in the analysis results, which showed the presence of the flypaper effect.

According to the findings of this study, central transfers, which account for the majority of local governments’ revenue and explain a larger portion of local expenditures, send incentive signals to local government officials. Being in a position to understand local finances better than the local people, these local government officials tend to depend on central transfers rather than increasing the mobilisation effort of locally generated revenue to finance local expenditures. This situation puts the officials of local governments in a trap where they are forced to follow the central government’s instructions to receive continued and sustained funding. This position erodes local autonomy and the essence of decentralisation, which we describe as ‘recentralised decentralisation’.

Based on the results, the study proposes the following actionable recommendations to enhance local government fiscal autonomy and reduce the flypaper effect in Ghana:

1. Incentivize own-source revenue generation by introducing programmes such as capacity-building programmes to improve the revenue mobilisation skills of local government officials, and establishing performance-based grants that reward local governments based on their success in mobilising internal revenue.

2. Reform the intergovernmental transfer system by shifting from unconditional transfers to conditional transfers, requiring local governments to meet specific performance metrics related to revenue generation and service delivery.

3. Enhance local government financial autonomy to give local governments greater control over determining their own development priorities and expenditure decisions.

4. Promote local economic development initiatives by supporting local businesses and microenterprises and encouraging investment in local economies to increase internally generated funds.

5. The government of Ghana could institute a robust monitoring and evaluation framework for tracking local government revenue collection and expenditure efficiency, and conduct regular impact assessments to evaluate whether fiscal decentralisation reforms are achieving their intended outcomes.

By implementing these recommendations, Ghana’s local governments can reduce their dependence on central transfers, improve fiscal autonomy and enhance the efficiency of local service delivery.

Acknowledgements

Open-access funding provided by Queen Mary University of London.

Data availability

Data requests must be submitted in writing to the corresponding author.

Competing interests

The authors declare no competing interests.

Ethical approval

The article does not contain any studies with human participants.

Informed consent

The article does not contain any studies with human participants.

Publisher’s note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.

References

Amusa, H; Mabunda, RR; Mabugu, RE. Fiscal illusion in a local sphere: empirical test of flypaper effects from South African municipal data. South Afr J Econ; 2008; 76, pp. 443-465.

Aragon FM (2009) The flypaper effect revisited. LSE Suntory and Toyota International Centres for Economics and related disciplines (STICERD) research paper no. EOPP004. London: London School of Economics (LSE)

Arellano, M; Bover, O. Another look at the instrumental variable estimation of error-components models. J Econ; 1995; 68, 1 pp. 29-51.0831.62099

Bailey, SJ; Connolly, S. The Flypaper Effect: Identifying Areas for Further Research. Public Choice; 1998; 95, pp. 335-361.

Baum, CF. Residual diagnostics for cross-section time series regression models. The Stata Journal; 2001; 1, pp. 101-104. [DOI: https://dx.doi.org/10.1177/1536867X0100100108]

Baskara T (2016) Intergovernmental transfers, local fiscal policy, and the flypaper effect: evidence from a German state. FinanzArchiv/Public Finance Anal 72(1). https://doi.org/10.1628/001522116X14462054864882

Beck, S. Autoregressive conditional heteroscedasticity in commodity spot prices. J Appl Econ; 2001; 16, pp. 115-132. [DOI: https://dx.doi.org/10.1002/jae.591]

Bird, R. Subnational taxation in developing countries: a review of the literature. J Int Commer Econ Policy; 2010; 2, No. 1 pp. 139-161. (2011)

Blundell R, Bond S, Windmeijer F (2000) Estimation in dynamic panel data models: improving the performance of the standard GMM estimator. Nonstationary, panel cointegration, and dynamic panels. Adv Econ 53–91. https://doi.org/10.1016/S07319053(00)15003-0

Blundell, R; Bond, S. Initial conditions and moment restrictions in dynamic panel data models. J Econ; 1998; 87, 1 pp. 115-143.0943.62112

Bradford DF, Oates WE (1971) Analysis of revenue sharing in a new approach to collective fiscal decisions. Quarterly J Econ 85(3):24

Brollo, F; Nannicini, T; Perotti, R; Tabellini, G. The political resource curse. Am Econ Rev; 2013; 103, pp. 1759-1796.

Canare T (2019) The effect of revenue shares on local government spending: evidence from Philippine provinces. Singapore Econ Rev 68(4). https://doi.org/10.1142/S0217590819500206

Careaga M, Weingast BR (2003) Fiscal federalism, good governance, and economic growth in Mexico. In: Rodrik D (ed) In search of prosperity: analytic narratives on economic growth. Princeton: Princeton University Press

Chakraborty, L; Jha, S. Central transfer: a curse or blessing? Evidence from the relative revenue effort of Indian states. Econ Rec; 2019; 95, 310 pp. 117-134.

Dahlberg M, Mörk E, Rattso J, Ågren H (2006) A discontinuous grant rule to identify the effect of grants on local taxes and spending. Cesifo Working, 92(12):2320–2335

Dash, BB; Raja, A. Fiscal autonomy and public expenditure performance: Some panel-data evidence from Indian states. Bull Econ Res; 2022; 74, 1 pp. 45-68.

Delali CR (2019) The effects of intergovernmental transfers on the local governments to generate own revenue in Ghana. https://ugspace.ug.edu.gh/items/b0884601-ffb5-46a3-ac56-c54debc49214?utm_source=chatgpt.com

Dick-Sagoe C, Tingum EN, Asare-Nuamah P (2022) Flypaper effects of Central Transfers on the spending behaviour of Ghana’s central region local governments: Does status matter? (Masyarakat, Kebudayaan Dan Politik, 35(3):297–309. https://doi.org/10.20473/mkp.V35I32022.297-309. Available online: https://e-journal.unair.ac.id/MKP/article/view/33280/22235

Dick-Sagoe, C; Tingum, E. Flypaper effect of intergovernmental transfers and incentives to improve own-source revenue mobilisation of local governments in the central region of Ghana. Open J Soc Sci; 2021; 9, pp. 434-447. [DOI: https://dx.doi.org/10.4236/jss.2021.98030]

Dick-Sagoe C (2017) Towards quality service delivery outcomes: What can local governments’ own-source revenue do? Paper presented at the 2nd International Conference on Public Administration and Development Alternatives in Gaborone, Botswana (26–28 July 2017). In: MP Sebola JP Tsheola (eds) The independence of African states in the age of globalisation. University of Limpopo publication, South Africa, p 223–228

Dick-Sagoe C (2013) Performance and challenges of internally generated revenue mobilisation of the Abura-Asebu-Kwamankese District Assembly for local level development. University of Cape Coast

Drukker, DM. Testing for serial correlation in linear panel-data models. Stata J Promot Commun Stat Stata; 2003; 3, pp. 168-177. [DOI: https://dx.doi.org/10.1177/1536867X0300300206]

Dzansi J, Jensen A, Lagakos D, Telli H (2019) Improving local government tax collection capacity: evidence from Ghana. International Growth Centre Policy Brief, p 33417

Ferreira, LOG; Serrano, ALM; Revelli, F. The effects of federal constitutional grants in Brazilian municipalities. Rev Urban Reg Devel; 2019; 31, 3 pp. 203-233.

Ferreira LOG (2021) Constitutional Tax Transfers in Brazilian Municipalities: the fiscal illusion and the flypaper effect. A Ph. D. thesis submitted to the Post Graduate Program in Accounting, Department of Accountancy and Actuarial Sciences, Faculty of Economics, Administration and Accounting, University of Brasília. https://repositorio.unb.br/handle/10482/42292

Gadenne L (2017) Tax me but spend wisely: the political economy of taxes. Theory and Evidence from Brazilian Local Governments. ULC Working Paper

Gamkhar S, Shah A (2007) The Impact of Intergovernmental Fiscal Transfers: A Synthesis of Conceptual and Empirical Literature. In: Boadway R, Shar A (eds) Intergovernmental fiscal transfers: principles and practice. World Bank, p 225–528

Ğbafi, S; Saruçş, NT. Intergovernmental Transfers and the Flypaper Effect in Turkey. Turkish Studies; 2004; 5, pp. 79-92.

Gorodnichenko (2001) Effects of intergovernmental aid on fiscal behaviour of local governments: the case of Ukraine. Master's Thesis, National University of Kyiv-Mohyla Academy

Greene WWH (2012) Econometric analysis (7th ed), Prentice Hall

Gruening, G. Origin and Theoretical Basis of New Public Management. Int Public Manag J; 2001; 4, pp. 1-25. [DOI: https://dx.doi.org/10.1016/S1096-7494(01)00041-1]

Hansen, CB. Generalised least squares inference in panel and multilevel models with serial correlation and fixed effects. J Econ; 2007; 140, pp. 670-694. [DOI: https://dx.doi.org/10.1016/j.jeconom.2006.07.011]

Hausman, JA. Specification Tests in Econometrics. Econometrica; 1978; 46, pp. 1251-1271.5136920397.62043

Hines, JR; Thaler, RH. Anomalies: the flypaper effect. J Econ Perspect; 1995; 9, 4 pp. 219-226.

Hsiao, C. Panel data analysis—advantages and challenges. TEST; 2007; 16, pp. 1-22.2368444 [DOI: https://dx.doi.org/10.1007/s11749-007-0046-x]

Hsiao C (2014) Analysis of Panel Data. 3rd Edition, Cambridge University Press, New York. https://doi.org/10.1017/CBO9781139839327

Inman RP (2008) The flypaper effect. NBER Working Paper No. w14579. Cambridge, MA: National Bureau of Economic Research

Kang Y Setyawan D (2012) Intergovernmental transfer and the flypaper effect: evidence from municipalities/regencies in Indonesia. KDI School of Pub Policy & Management, p 12–06, Available at https://doi.org/10.2139/ssrn.2065638

Kaur A, Mohanty RK, Chakraborty L, Rangan D (2021) Ecological Fiscal Transfers and State-level Budgetary Spending in India: Analysing the Flypaper Effects. Working Paper No. 990, Levy Economics Institute of Bard College, London, UK

Keele and Kelly (2006) ‘Dynamic models for dynamic theories: the ins and outs of lagged dependent variables’

Lu T, Wu W (2022) Empirical on the Influencing Factors of Local Government’s Online Response. https://doi.org/10.1155/2022/8741336

Musgrave R (1959) Public Finance. New York: McGraw-Hill

Muthama, V. Modelling panel data: comparison of GLS estimation and robust covariance matrix estimation. Am J Theor Appl Stat; 2015; 4, pp. 185-191. [DOI: https://dx.doi.org/10.11648/j.ajtas.20150403.25]

Oates WE (2003) An essay on fiscal federalism. In: Baimbridge M, Whyman P (eds) Fiscal federalism and European economic integration, Routledge, p 13–47

Oates, WE. An essay on fiscal federalism. J Econ Lit; 1999; 37, 3 pp. 1120-1149.

Oates WE (1972) Fiscal Federalism. New York: Harcourt Brace Jovanovich

OECD (2015) Issue paper: linking revenue to expenditure. OECDIssue_Paper_on_Linking_Revenues_and_Expenditures.pdf (theicpa.org)

Osei-Owusu J (2022) Report of the committee of the whole on the proposed formular for the district assemblies’ common fund for 2022. Second session of the Eight Parliaments of the Fourth Republic of Ghana. Available online: https://ir.parliament.gh/handle/123456789/2162

Otoo, I; Danquah, M. Fiscal decentralisation and efficiency of public service delivery by local governments in Ghana. Afr Dev Rev; 2021; 33, pp. 411-425. [DOI: https://dx.doi.org/10.1111/1467-8268.12590]

Rao GV, Singh N (1998) Intergovernmental transfers: rationale, design and indian experience. Working Paper, Centre for International Development, Stanford University

Rodden J (2003) Fiscal decentralisation and the challenge of hard budget constraints. Cambridge, M.A: MIT press

Roodman, D. How to do xtabond2: an introduction to difference and system GMM in Stata. Stata J; 2009; 9, 1 pp. 86-136.

Rubinfeld D (1987) Economics of the local public sector. Handbook of Public Economics, Vol. II, ed. AJ Auerbach and Feldstein, M New York: Elsevier. p 571–646

Sa, S; Saruç, N. Intergovernmental transfers and the flypaper effect in Turkey. Turkish Stud; 2004; 5, pp. 79-92. [DOI: https://dx.doi.org/10.1080/1468384042000228602]

Seabright, P. Accountability and decentralisation in government: an incomplete contract model. Eur Econ Rev; 1996; 40, 1 pp. 61-89. [DOI: https://dx.doi.org/10.1016/0014-2921(95)00055-0]

Shah A (2007) A practitioner’s guide to intergovernmental fiscal transfers. In: Boadway R, Shah A (eds), Intergovernmental fiscal transfers: principles and practice. Washington, D.C. The World Bank

Sijabat, R. Panel analysis on the association of fiscal decentralisation and the size of government in Indonesia. J Econ Bus Manag; 2016; 4, pp. 206-214. [DOI: https://dx.doi.org/10.7763/JOEBM.2016.V4.392]

Singh N, Srinivasan TN (2006) Federalism and economic development in India: an assessment. SCID conference on the ‘challenges of economic policy reform in Asia’, Stanford University, p 1–3

Singhal, M. Special interest groups and the allocation of public funds. J Public Econ; 2008; 92, pp. 548-564.

Slavinskaite, N. Fiscal decentralization and economic growth in selected European countries. J Bus Econ Manage; 2017; 18, pp. 1-13.

Slavinskaitė, N; Liučvaitienė, A; Gedvilaitė, D. Theoretical analysis of the fiscal federalism. Int J Contemp Econ Adm Sci; 2019; 9, 2 pp. 250-267. [DOI: https://dx.doi.org/10.5281/zenodo.3595926]

Soto M (2009) System GMM estimation with a small sample. Barcelona Economics Working Paper Series 395

Tiebout, C. A pure theory of local expenditures. J Political Econ; 1956; 64, pp. 416-424.

Timmons, JF. Fiscal contracts: states, taxes, and public services. World Politics; 2005; 57, pp. 530-567.

Wallis J, Sylla RE, Legler JB (1994) The interaction of taxation and regulation in nineteenth century U.S. Banking’,. In: Goldin C, Libecap GD (eds) The regulated economy: a historical approach to political economy. Chicago: University of Chicago Press

Weingast BR (2013) The fiscal interest approach: the design of tax and transfer systems. https://web.stanford.edu/group/mcnollgast/cgi-bin/wordpress/wpcontent/uploads/2013/10/Weingast-Poschl-BRW.13.1007.pdf

Witterblad M (2007) Income heterogeneity and the flypaper effect. Umeå: Department of Economics, Umeå University

Wooldridge JM (2010) Econometric analysis of cross section and panel data (2nd ed). MIT Press