Factors explaining inter-municipal cooperation in service delivery: a meta-regression analysis

Inter-municipal cooperation is an important public service delivery reform, whose drivers move beyond simple concerns with costs and economic efficiency, to policy issues related to governance structure and spatial context. We conduct a meta-regression analysis based on the existing multivariate empirical literature to explore what factors explain divergence in results in the existing empirical studies. We find strong evidence that fiscal constraints, spatial, and organizational factors are significant drivers of cooperation. Our meta-regressions do not yield results to explain divergence in results on community wealth, economies of scale, or racial homogeneity. More studies on these factors are needed to understand how these factors might affect cooperation. Future theoretical and empirical research should give more attention to spatial and organizational factors to develop a better understanding of factors driving cooperation, and how they differ across local government structures and regions.


Introduction
In the field of local government service delivery reform, most policy attention has been focused on contracting to the private sector (Boyne 1998;Hodge 2000). But another, equally important reform strategy, especially in the USA, inter-municipal cooperation, has begun to garner more research attention as it has grown in importance (Hefetz, Warner, and Vigoda-Gadot 2012). Inter-municipal cooperation involves contracts or joint production with other local governments as a means to gain economies of scale, improve service quality, and promote regional service coordination across fragmented local government regions (Agranoff and McGuire 2003). Cooperation tends to be voluntary in the USA (Feiock 2007;Holzer and Fry 2011) and more formal and structural (involving more regulation) in Europe (Hulst and van Montfort 2007;Lago-Peñas and Martinez-Vazquez 2013). It is typically found to be a more politically viable alternative than amalgamation (Feiock and Scholz 2010) and for this reason is an important alternative of interest to policy-makers of local government reform.
As the literature on cooperation has grown, one distinguishing feature is the broader set of factors that appear to drive or inhibit cooperation as compared to privatization. Theory and empirical evidence on local government contracting to the private sector has primarily focused on service delivery characteristics related to cost savings and transaction costs (Brown and Potoski 2003;Bel and Fageda 2008;Levin and Tadelis 2010;Wassenaar, Groot, and Gradus 2013;Gradus, Dijkgraaf, and Wassenaar 2014). But empirical and theoretical evidence on inter-municipal contracting gives more attention to factors related to the local government organization (structure, management and spatial factors) (Tavares and Camões 2007;Bel, Fageda, and Warner 2010;Hefetz, Warner, and Vigoda-Gadot 2012;Zafra-Gómez et al. 2013). Comparative studies of the USA and Europe have noted that while European studies are more focused on cost savings, US studies are more concerned with government organization factors, in part because the structure of responsibilities and finance in the USA creates more diversity in service competencies making cooperation both more needed and more difficult (Bel and Warner 2015).
Although shared services delivery is a widespread phenomenon, no meta-regression analysis of the literature has been conducted. By contrast, systematic evidence is available on other reforms, such as privatization, where meta-regression analyses have been published on the factors explaining privatization (Bel and Fageda 2009), and on the relationship between privatization and costs (Bel, Fageda, and Warner 2010). Our current analysis bears some similarity to Bel and Fageda (2009), because we also deal with factors explaining decisions related to government reform; here, cooperation instead of privatization. We do not attempt to deal with the issue of cooperation and costs as too few empirical studies exist in the literature (see Bel and Warner [2015] for a list).
A large enough number of empirical analyses on the motivations for inter-municipal cooperation are now available. Thus, our paper provides an extensive and in-depth analysis of the empirical evidence on the factors explaining inter-municipal cooperation. We conduct a meta-regression analysis based on the existing empirical literature to permit a systematic analysis of the similarities and differences in the results of empirical studies of motivations for inter-municipal cooperation and the implications for public management.

Theoretical background
To set up our meta-regression analysis we first outline the major theoretical components of interest. These can be grouped into two major categories: cost and fiscal factors, and organizational and governance characteristics. These two broad categories have been seen as key categories in the theoretical and empirical literature on service delivery reform. Because of this, we believe they may be especially important for policy on inter-municipal cooperation, and our meta-regression analysis of the available empirical literature can shed light on the relative importance of these factors.

Cost and fiscal factors
2.1.1. Cost structure: scale and density One of the challenges facing local governments is that size and fragmentation make it difficult to capture the benefits of economies of scale, or the need for coordination across the metropolitan region. While the economy has moved to primarily a city-regional scale, Lobao, Martin, and Rodriguez Pose (2009), local government organization reflects the historical development of the region and may result in sub-optimal jurisdiction size. This is especially a concern in Southern Europe and in the USA where there is a long tradition of many small municipalities (Holzer and Fry 2011;Lago-Peñas and Martinez-Vazquez 2013). The literature on local government has paid attention to the optimal size for service provision (Olson 1969;Mirrlees 1972;Oates 1972). Volume of service, size of population, and dispersion of population are the three dimensions on which the optimal geographic scale depends (Ladd 1992). Olson (1969) argued for fiscal equivalence that government boundaries should match the area enjoying the local public good. When municipal boundaries do not match the service area, some form of municipal cooperation is a natural alternative to achieve both economies of scale and fiscal equivalence. Amalgamation has been offered as one alternative to suboptimal size, but amalgamations are typically not voluntary and result from laws made at a higher jurisdictional level. While Reingewertz (2012) finds reduced expenditures resulted from amalgamation in Israel, most empirical papers on the topic find no evidence of positive impacts from consolidation either on costs/debt or on scale economies (i.e. Tyrefors Hinnerich 2009;Jordahl and Liang 2010;Allers and Geertsema 2014).
By contrast, Ostrom, Tiebout, and Warren (1961, 836) envisaged intermunicipal cooperation as an alternative to amalgamation when they suggested that small municipalities could make use of special arrangements to act jointly to provide services when the municipal boundary is suboptimal. Cooperation has proven more popular than amalgamation among local governments (Holzer and Fry 2011). Research across Europe and the USA has found cooperation can be especially important in helping small municipalities reach economies of scale (Bel and Costas 2006;Warner 2006;Mohr, Deller, and Halstead 2010;Hefetz, Warner, and Vigoda-Gadot 2012;Bel, Fageda, and Mur 2013;Wassenaar, Groot, and Gradus 2013;Zafra-Gómez et al. 2013).
The most common measure in the empirical literature for economies of scale is population, though studies also frequently include attention to community wealth and fiscal stress as drivers of cooperation. Sonenblum, Kirlin, and Reis (1977) focus on scale issues and extra-local spillovers in service delivery arrangements. Most studies find cooperation to be negatively related to population as larger cities enjoy internalscale economies (Levin and Tadelis 2010;Hefetz, Warner, and Vigoda-Gadot 2012;Bel, Fageda, and Mur 2014).

Fiscal factors
Fiscal factors have been an important driver of local government reform in the last decades, particularly as local government fiscal stress has grown in the 2000s. Whereas privatization was seen as the primary reaction to these restrictions, cooperation is also a tool to create cost savings. In the USA, cooperation has increased since the Great Recession and is now more common than privatization (Homsy and Warner 2014). Governments generally cooperate when it is in their self-interest, and may tend to exclude governments with higher costs or lower fiscal capabilities from cooperative agreements (Lowery 2000). Because of this, wealthy communities may engage less frequently in cooperation (Warner and Hefetz 2002). However, wealthy communities can also be desired partners for cooperative agreements, because of their ability to contribute to fund the service (Kwon and Feiock 2010).
Fiscal constraints, related to debt and expenditure burden or limits on local revenue raising, can be differentiated from fiscal capacity related to community wealth. In recent years, local governments have faced increased constraints due to austerity policy, and tax and expenditure limitations passed by higher levels of government (Clifton 2014). Understanding the limits of cooperation as a policy tool requires differentiating fiscal constraints from community capacity to understand their differential effects on cooperation as a tool of policy reform.

Institutional and governance structure
Fragmented local government systems face the challenge of providing services efficiently and addressing problems of service spillovers and tax exporting. In the USA, many scholars argue that when regional governance is voluntary and lacks sanctioning authority or taxation power, the ability to promote cooperation is underminedespecially in services where there is heterogeneity in need and resources across the region (Frug 1999;Lowery 2000;Warner and Hefetz 2002). Key differences between Europe and the USA make comparison across studies of cooperation especially important. US local governments are characterized by a higher level of fiscal autonomy and service responsibilities and lower levels of intergovernmental aid, while in continental Europe basic local services (solid waste, water, etc.) are compulsory for all municipalities, so service levels are more homogeneous (Lago-Peñas and Martinez-Vazquez 2013). These features facilitate cooperation in Europe and may explain why most European studies focus on cost savings. The US studies by contrast, focus more on the institutional challenges to cooperation, which result in part from differences in local government structure in the USA (Bel and Warner 2015).
Cooperation requires attention to political institutions and the structure of policy networks, which places an important role on the structure of management and markets among local governments in the region where cooperation takes place . Cooperation is easier when there is homogeneity in interests, needs and resources, and institutional homogeneity in budget rules and service requirements (Feiock 2007). Other important factors that may facilitate cooperation are longer tenure and more professional management (Brown and Potoski 2003;Hefetz, Warner, and Vigoda-Gadot 2015). Regional governance bodies also facilitate cooperation, as this gives time for reciprocal relationships to build both weak and strong ties across the network (Thurmaier and Wood 2004;Wood 2006;Bel, Fageda, and Mur 2013).
Unlike New Public Management reforms that emphasize competition, cooperation is based on collaborative relations that extend across time and space. The possibility of joint production with neighboring municipalities is driven not only by concerns with economic efficiency, but also by factors which include regional coordination and improved effectiveness of service delivery and reduced contracting risk (Joassart-Marcelli and Musso 2005;Vigoda-Gadot 2012, 2015). Theory also addresses geographic considerations related to metropolitan location. Suburbs, as similar sized localities in the metropolitan region, offer the most attractive market for cooperation (Warner and Hefetz 2002;Joassart-Marcelli and Musso 2005;Hefetz, Warner, and Vigoda-Gadot 2012). Cooperation is also high in rural areas, which lack private market alternatives (Warner and Hefetz 2003;Warner 2006;Mohr, Deller, and Halstead 2010) and may use cooperation to gain market power Mur 2013, 2014). Feiock (2007Feiock ( , 2013 has outlined a theory of institutional collective action that builds from Ostrom's (2010) studies of polycentric local government where cooperation is based on trust, reciprocity and collective benefits across the metropolitan region.

Governance costs of the cooperative agreements
Inter-municipal cooperation may be subject to lower governance costs than privatization because cooperating governments share similar objectives (Brown 2008). Cooperation is based on a public market of cooperating governments, not a competitive market of for-profit providers (Warner 2011). Recent theoretical and empirical work emphasizes the importance of the public partner in creating more stable service delivery arrangements (Hefetz, Warner, and Vigoda-Gadot 2014). However, Feiock (2007) has emphasized that transaction costs from cooperation involve substantial costs related to information, negotiation, and monitoring. Indeed Marvel and Marvel (2008) have found challenges with monitoring due to lower ability to sanction partners in voluntary inter-municipal cooperative agreements in the USA. Sørensen (2007) has found similar problems with dispersed ownership in cooperative agreements in Norway.
Several studies (i.e. Levin and Tadelis 2010; have shown that when service-related transaction costs are high, inter-municipal contracting is preferred to for profit contracting. Girth et al. (2012) point to the opportunity costs of creating a market in for profit contracting, costs that are avoided in inter-municipal contracting. Professional management can help overcome these transaction cost challenges and council manager forms of government have been shown to have higher levels of cooperation (Nelson and Svara 2011;Vigoda-Gadot 2012, 2015). However, differences in wealth, demographic makeup, and geographic location of participating communities may still produce problems creating a willing market of participating municipalities (Lowery 2000;Warner and Hefetz 2002;Warner 2006), and result in coordination problems after the cooperation is in place (Feiock 2007). Heterogeneity undermines local cooperation because it imposes higher transaction costs (Feiock 2007). In the USA, racial homogeneity/ heterogeneity have been found to be potential drivers of transaction costs in the governance of the cooperation (LeRoux and Carr 2007; Kwon and Feiock 2010).

Empirical studies on inter-municipal cooperation: what do they tell us about drivers and obstacles?
We have been able to find 49 articles (either published or forthcoming) and working papers including multivariate analysis of factors explaining cooperation, as Table 1 shows. We reviewed both published and unpublished papers from the fields of Economics, Public Policy, Public Administration, Political Science, Urban Studies, and Area Studies. The 10 unpublished papers included in our analysis were presented in international meetings specializing in public policy, and/or available in large working paper collections, such as Econlit, Social Science Research Network, Ageconsearch, Proquest, and Repec-Ideas. Furthermore, we searched for papers from data bases specialized in PhD Dissertations and Grey Reports such as OpenSIGLE¸European Science Research Council (ESRC) and E Thesis Online Services (ETHOS) in Europe, and US GAO and The National Technical Information Service (NTIS) in the US. Papers selected for our analysis had to have homogenous measures of cooperatione.g. a regression model with cooperation as the dependent variable.
To our knowledge, our database includes all published and unpublished papers that estimate factors explaining cooperation. In all, we have 38 journal articles, one book chapter, nine working papers, and one poster presented at a conference (where enough information was provided). The database was constructed by the authors. We used as Table 1. Main characteristics-multivariate studies on drivers of intermunicipal cooperation in service delivery. key words for the search "intermunicipal cooperation", "interlocal contracting", and "joint contracting". The search was completed on December 2014. Overall, we took into consideration the Meta-Analysis of Economics Research (MAER-net) reporting guidelines in Stanley et al. (2013). The 49 studies presented in Table 1 include a total of 171 estimations where the dependent variable is a measure of the frequency of inter-municipal cooperation. Typically, the measure is (1) a dummy variable with one for cooperative delivery and 0 otherwise in single service studies; and (2) a percentage of the services each jurisdiction provides via cooperative delivery for multi-service studies. Interestingly, most studies (38) focus on the USA, either the whole country or one/several states. Eleven papers study countries other than the USA. Next we provide a review of the most frequently used variables in these studies.

Authors
Fiscal constraints have been operationalized by means of different variables, among which the most frequent are debt per capita, own revenues per capita, laws limiting debt, etc. Table 2 shows that 70% of the estimations in our database included variables related to fiscal constraints. Among these, more than half the estimations in our database have shown fiscal constraints to have a significant effect on cooperation. 1 As expected, a large majority of them find a positive effect of fiscal constraints. Results showing the opposite (a negative influence of fiscal constraints on cooperation) are relatively rare, around one third the number showing positive influence. Non-significant results are obtained in little more than 40% the estimations. Notes: Community wealth is considered to be a driver of cooperation when per capita income has a negative and significant relationship with cooperation. Spatial effects measures geographic proximity as in city/suburb and is a driver of cooperation when it has a positive and significant relationship with cooperation. Economy of scale is considered to be a driver of cooperation when population has a negative and significant relationship with cooperation. Source: Authors' elaboration.
Community wealth: The diversity in theoretical expectations regarding community wealth is well reflected in the empirical results obtained. More than half of the available estimations considered community wealth as an explanatory variable for cooperation, usually specified as income per capita. No significant relationship was found in almost two-thirds of the estimations. Much less frequent are estimations with significant results for community wealth, and they are almost evenly split between positive effect (20%) and negative effect (15%). Therefore, theoretical expectations on community wealth are divergent, and empirical results so far do not help to solve the question.
Scale economies: Service and place characteristics are the most important drivers of economies of scale. Among these dimensions, size of population is the variable most frequently used in empirical studies analyzing factors explaining cooperation. This is explained by the fact that data on population are more readily available than data on volume of output when a single service is analyzed, and because population is the best indicator of volume of output when the analysis is of multi-service character. Around 70% of the estimations in our sample use number of inhabitants as an explanatory variable for cooperation. Half of these estimations do not find population to be significant. Among the estimations where this variable is significant, the most frequent result is that population has a negative and significant association with cooperation. Therefore, there exists some partial evidence that the frequency of cooperation decreases as population increases, as expected from theoretical views on economies of scale.
Spatial factors: Inter-local cooperation frequently involves geographic proximity of the cooperating municipalities. Cities and suburbs in metropolitan regions represent a market of similar sized municipalities in close proximity. One-third of the estimations in the database included variables related to cities and suburbs as drivers of cooperation, and more than half of these found the variable to be significant and positive, as shown in Table 2. Thus, results are largely consistent with theoretical expectations.
Organizational factors (manager): The council-manager form of city government is understood as more professional and less subject to political interference than the mayor-city model, and thus may serve as a driver for service delivery reform. More than one-third of the available estimations analyze the relationship between councilmanager form of government and frequency of collaborative agreements, of which half of the estimations find the manager variable to be significantly related to cooperation. Among these estimations where manager is significant, almost 80% find a positive association between manager and cooperation, consistent with theoretical expectations.
Racial homogeneity/heterogeneity: Little more than one-third of the available estimations (58) and little more than one quarter of the studies (13) have included specifications of racial homogeneity/heterogeneity. More than half of the estimations find racial homogeneity/heterogeneity significant, and among these a large majority find a positive effect of homogeneity. Note, however, that the majority of estimations showing a positive (negative) association between homogeneity (heterogeneity) and cooperation come from Blaeschke (2014). Also, two of these papers (Morgan, Hirlinger, and England (1988); Morgan and Hirlinger (1991), together include eight estimations) specify the variable as a proxy for private interests and political factors (rather than transaction costs-related factors). Therefore, there is divergence in what the variable measures.
Service level transaction costs: In spite of the important theoretical and conceptual work available on transaction costs, the empirical evidence is still quantitatively limited. Only 31 estimations have included variables reflecting transaction costs. However, among these estimations 42% find a positive and significant relationship (more service related-transaction costs are associated with more cooperation) and only six find a negative relationship. Overall, and consistent with theoretical expectations, cooperation is typically found to be higher when services are more asset specific and other factors related to transaction costs are present.
Finally, we note that "political factors", another type of variable usually considered in the empirical analysis of local government reform, is seldom used in empirical analysis on drivers of cooperation. Only 12 papersand 30 estimationsconsider variables reflecting political factors, and the estimations are divided between those that consider political orientation of elected politicians, and others that consider the type of election (i.e. election at large). The diversity in the way the political variables are specified is very wide. Therefore, we cannot specify expectations about the influence of political factors on cooperation.

Empirical strategy: data and methods
There are different reasons why analyses that focus on a single topic have a large variation of empirical results. Among these, Stanley and Jarrell (1989) emphasize three types of categories for these reasons: (1) uniqueness of the data sets used in each particular study; (2) biases induced by model misspecification, and (3) distinctive statistical methods. We use meta-regression analysis to analyze the pattern and diversity of findings in the empirical studies. In this way, we are able to appraise if significant relationships in papers that study drivers and obstacles of inter-municipal cooperation depend heavily on the individual characteristics of each study. Meta-regression analysis has been used frequently in economic and public policy research, after it was introduced in the late 1980s. Recently, it has been used to analyze variability of empirical results in the field of public services and local government (i.e. Bel and Fageda 2009;Bel, Fageda, and Warner 2010;Carvalho, Marques, and Berg 2012). We are not aware of any meta-regression analysis published to explain the differences between empirical results in studies on drivers and obstacles of inter-municipal cooperation, thus our research contributes to filling this gap.
The most frequently applied empirical strategy in meta-regression analysis is that suggested by Stanley and Jarrell (1989), where b j is the reported estimate of β of the jth study, β is the true value of the parameter of interest, Z jk are the meta-independent variables that measure relevant characteristics of an empirical study, α k are the coefficients associated with those independent variables, and e j stands for the residuals. The initial empirical strategy for implementing the meta-regression is to use the coefficients or t-statistic values estimated in each study as a dependent variable in the meta-regression. A first problem we confront is that we have neither t-statistics nor standard errors for many estimations. Moreover, we cannot compute t-statistics based on coefficients and standard errors in other cases, because these are given with all zeroes (i.e. 0.000). Therefore, the sample would be significantly reduced if we used t-statistics. Furthermore, the studies on factors explaining inter-municipal cooperation have used a wide variety of variables to test the relationship between cooperation and economic and organizational factors. Thus, the number of homogeneous t-statistics that we have been able to identify is very limited, which we are aware is a limitation of our research.
Our objective is to analyze whether the relationships we are interested in are significant. To do so, we follow the strategy applied by Bel and Fageda (2009). We construct a set of dependent variables as dummy variables that take a value of one if a study finds a significant relationship between inter-municipal cooperation and the corresponding set of explanatory variables: fiscal and economic, organizational, and spatial. Table 3 displays the set of dependent variables, and the set of independent variables (moderator variables) that concern particular characteristics of the studies. Note that we do not Note: Difference between number of observations of dependent and independent variables is due to the fact that the variables we take as dependent are not always present in all studies and estimations. include specific regressions for service level transaction costs 2 or politics, because the small number of available estimations does not allow us to conduct a sensible estimation. We specify the dependent variables as follows: Fiscal constraints: We find a significant influence of fiscal constraints on inter-municipal cooperation when variable(s) that capture this effect has(ve) a positive influence on cooperation. The variables most commonly used are debt burden, tax burden, own fiscal revenues, and supra-local regulation limiting local taxation and debt. 3 Community wealth: We find a significant influence of community wealth on intermunicipal cooperation when the variable that captures its effect has a negative influence on cooperation. The variable is usually specified as income per capita.
Economies of scale: We find a significant influence of economies of scale on intermunicipal contracting when the variable that captures this effect has a negative influence on cooperation. By far, the variable most commonly used is population size.
Spatial factors: We find a significant influence of spatial factors on inter-municipal cooperation when the variable that captures this effect has a positive influence on cooperation. The variable most commonly used is city in a metropolitan area or suburb.
Organizational factors: We find a significant influence of manager-type of government on inter-municipal cooperation when the variable that captures this effect has a positive influence on cooperation. The variable most commonly used is council-manager form of local government.
Racial homogeneity: We find a significant influence of racial homogeneity on intermunicipal cooperation when the variable that captures this effect has a positive influence on cooperation. The variable most commonly used is percent racial/ethnic majority in population.
The independent variables used are those common in meta-regression analysis, reflecting particular characteristics of the studies: year of data collection, number of observations, geographical area, and method of estimation. These variables reflect the type of categories used to explain variations, such as the uniqueness of the data-sets, or distinctive statistical methods. Furthermore, area and time differences may reflect institutional contexts or learning over time. As in Bel and Fageda (2009), we include two additional characteristics of the studies: whether the estimations are single-service or multi-service (a few studies consider different services but use single-service estimations), and whether small municipalities (less than 5000 population) are included in the sample or not. Finally, we include two additional variables. One indicates whether the study focuses on cooperation solely, or whether it is considered among other alternatives (e.g. private contracting) for service delivery. The other is a measure of research quality based on the latest available (2014) impact factor in the Journal Citation Reports of the journal in which the article was published. Journals not considered in JCR and other studies received a 0.

Results
We first regress the dependent variables against the independent (moderator) variables by means of probit estimation, robust to heteroskedasticity. We estimate the following set of relationships: Cooperation ¼ FðFC; SE; CW ; OF; SF; RHÞ; (2) Results from the six meta-regressions are provided in Table 4. In all cases, the estimations are robust and have clustered estimations from the same study. Recall that each observation is a study analyzing factors explaining inter-municipal contracting. The dependent variables are dummy variables that take a value of one when a study finds a significant relationship between inter-municipal contracting and that variable (except for community wealth and economies of scale which are coded 0 = negative relationship). Positive coefficients of the moderator variables indicate that studies with that characteristic tend to provide a positive significant relationship between the corresponding dependent variable and cooperation. A negative sign implies that the associated characteristic of the study is less likely to be associated with positive and significant results. Note the opposite interpretation holds for community wealth and economies of scale.
We obtain relatively high R 2 (recall we are conducting probit estimations) for most of our estimations: Fiscal Constraints, Economies of Scale and Racial Homogeneity. The explanatory capacity is smaller but still significant for Spatial Factors and for Organizational Factors, and is very low for Community Wealth. Overall, the explanatory results of our meta-regression estimations compare favorably with those of studies using similar methodology, such as Bel and Fageda's (2009) analysis of factors explaining local privatization.
The hypothesis that fiscal constraints would trigger more cooperation is well established in the literature. Results show that more recent studies (year) are less likely to find an effect of fiscal constraints. Also, studies done for the USA are less likely to find fiscal constraints as a driver of cooperation.
By contrast, studies that include smaller places (<5000 population) in the sample are more likely to find a positive effect of fiscal constraints, as small places are more likely to benefit from cooperation. The same is true of multi-service studies, as cooperation in these studies is measured on an organization-wide level, not just in a single service, so the potential for fiscal impact is greater. Furthermore, studies using logistic regression also tend to find a significant positive influence of fiscal constraints. Recall that logistic regression models are more robust than OLS when the dependent variable  is bounded between 0 and 1. Finally, studies that focus only on cooperation tend also to find a significant positive influence of fiscal constraints.
The surprising result in the fiscal constraint model is the lower likelihood of finding fiscal constraints as a driver in the US models. Prior meta-regression analysis on local government privatization found US studies were more likely to find fiscal constraints as a driver (Bel and Fageda 2009). What explains the divergent results between privatization and cooperation as regards fiscal constraints? Local government fragmentation in the USA is high and the motivations for cooperation extend beyond fiscal constraints to concerns regarding service coordination across the region and service quality improvements (Hefetz, Warner, and Vigoda-Gadot 2012). While privatization is driven heavily by cost and fiscal constraints, cooperation is used for a broader range of purposes in the USA, namely service quality and service coordination, but these variables are not captured well in the regression model literature. Our regressions cannot capture other motivators because the number of studies addressing them is too small. Furthermore, in Europe, small places often receive more service for the same price under cooperative agreements. A primary type of cooperation in Europe is joint governance of services, but US studies primarily measure inter-local contracting. The type of contracting in Europe is much more prone to cross subsidies among the partner municipalities.
Turning now to community wealth, the meta-regression studies with larger sample sizes (generally considered to be more robust) are less likely to find significant effects (recall the community wealth variable is reverse-coded). The meta-regression also shows that studies where cooperation is considered alone are more likely to find a negative relationship between cooperation and community wealth. Rich communities may feel less pressure to meet economic efficiency goals, or to cooperate with their neighbors. Also, studies published in higher quality journals tend to find this negative relationship between community wealth and cooperation.
Regarding economies of scale, the main operational expectation is that as population increases, cooperation would decrease. Recall that economies of scale exist when population has a negative and significant relationship with cooperation. We expected that studies containing more small places (<5000) would be more likely to find a scale effect. However, that moderator variable is positive, meaning that it is less likely that a relationship with scale is found. Studies conducted for the USA and studies focusing only on cooperation are less likely to find effects of economies of scale. This may reflect the wider range of factors driving cooperation in the US studiesbeyond economic considerations. Multi-service studies are more likely to find scale effect but this result is only significant at the 10% level.
With respect to spatial effects, we did not find any significant moderator. The equation itself did a poor job of explaining differences in results across studies. This result is interesting. A majority of estimations show that cities in metropolitan areas have higher rates of cooperation, and only a small minority of studies find cities with lower levels of cooperation (see Table 2). Our meta-regression shows these results are not differentiated by any moderator variable. Therefore, theoretical expectations are met by empirical results, 55% of the observations found this result, and differentiation across studies is not explained by the intrinsic characteristics of the estimations. City status in a metropolitan region, where proximity helps create a market for cooperation, is a relevant driver of inter-municipal cooperation.
The meta-regression on organizational factors (manager-type government) dropped the continent moderator variable as all the studies measuring this effect are in the USA.
Three variables show a positive relationship with manager influence: logistic method, sample including small municipalities, and studies focusing only on cooperation. Our theoretical expectation was that professional management would be positively associated with cooperation and indeed this is trueespecially in studies including smaller places (<5000 population). As in the case of spatial effects, our meta-regression shows that divergence of results in the literature is not heavily driven by the characteristics of the estimation. Forty two percent of the estimations have a positive association between managers and cooperation (this being a large majority among the results finding a significant association, see Table 2), thus, the theoretical expectation is met by the empirical evidence.
The meta-regression on racial homogeneity also dropped the continent moderator variable. In this case, almost all variables show a significant relationship with racial homogeneity. More recent studies, studies with small municipalities, and studies published in high-impact outlets tend to find a positive association between racial homogeneity and cooperation, and the opposite happens with almost all other moderators. This equation is the most powerful in explaining the divergence of results. The meta-regression shows that studies that include a wider range of governments are less likely to find racial homogeneity to be a factor. These include the larger sample sized studies, multi-service studies, and those which only look at cooperation.
It is common in meta-regressions for papers included to provide more than one observation (see Table 1). That can cause a problem of dependence across observations (Ringquist 2013). 4 To deal with this, we replicated the probit estimations by including a dummy variable to control for the observations obtained from the study by Blaeschke (2014), which isby farthe work that provides more estimations, up to 22. The results were unchanged. Furthermore, in order to take full account of within-study autocorrelation, we followed the suggestion in Ringquist (2013, 218) and used Generalized Estimating Equations (GEE) to estimate a random effects meta-regression model. Table 5 shows the results.
Results obtained from our GEE estimations are very similar for fiscal restrictions and for organizational factors; thus, we take the results for these factors as solidly established. The estimations on the other factors yield somewhat different results, which we discuss next.
The estimation for community wealth does not yield any significant variables and, furthermore, shows a very poor overall fit. There is no association between any moderator and community wealth. In this sense, it is worth recalling that two-thirds of studies found no effect and those that did were almost evenly split between positive and negative effects (see Table 2). Although community wealth has been described in the literature as an important factor, our meta-regression provides limited insights into why we find such divergent results across studies, and has a large majority of non significant results overall. Indeed, it might well be that community wealth by itself is not a relevant driver of cooperation.
Regarding economies of scale, only two variables are still significant and these only at the 10% level. Studies conducted for the USA and studies where only cooperation is considered are less likely to find economies of scale, as before. However, population size is not significant when we take into account dependence across observations. By contrast, the new estimation for spatial factors shows a better fit, and multi-service has a strong negative correlation with cooperation. Finally, in the estimation for racial homogeneity, several variables cease to be significant. However, we cannot conclude anything in this case, as this estimation did not converge.   ***1%; **5%; *10%; &Omitted because of collinearity.
In the case of racial homogeneity, the estimation did not converge.
Overall, when we take into account dependence across observations caused by studies providing many observations, several variables cease to be significant in several estimations. Fiscal constraints, spatial factors, and organizational factors appear to be the most robust models across estimations.

Meta-regression robustness tests
One concern in meta-regression analysis is studies might be more likely to be published if they find significant relationships between variables. Because of this, publication bias is an important limitation of meta-regression analysis (Stanley 2005). Publication bias in meta-regression may be examined by means of funnel asymmetry tests -FAT - (Stanley 2005(Stanley , 2008. FAT tests are based on the estimation of the reported effect of a study and its standard errors.
The studies included in our meta-regression analyses use different variables to test the relationship between cooperation and each set of factors. Because of this, we can implement our tests for publication bias only for the most common variables for each group of hypotheses, which reduces the samples available for the publication bias test. Furthermore, some studies report the coefficient and statistical significance, but not standard errors or t-statistics, so we cannot include these studies in the tests. All this forces us to be cautious when interpreting results of the FAT, because of the low number of observations that can be used.
With respect to publication bias, FAT (1) and FAT (2) show that we find strong evidence of publication bias for fiscal constraints, because in both estimations [using 1/SE and sqrt (sample size)], the intercept significantly differs from zero. We find weak evidence of publication bias for spatial factors, as the intercept is significant, but only at 10%. If there is publication bias, the direction is positive in both cases, which means that the studies are obtaining overly large effect of fiscal constraints and of city in metropolitan areas. We find even weaker evidence for racial homogeneity, as publication bias found with FAT (1) is contradicted by the results from FAT (2). Table 6 shows results of the funnel asymmetry tests for these three factors. 5 Therefore, we do not believe publication bias is an important problem for racial homogeneity. Publication bias may be filtered by means of estimating a multivariate FAT meta-regression model (Stanley 2005). We deal with this issue following methodological guidelines pointed out in Bel, Fageda, and Warner (2010). This will imply re-estimating equations for fiscal constraints (Equation (3)) and for spatial factors (Equation (6)), including square root of sample size, and excluding sample size. Our previous results with respect to moderator variables in both equations, for fiscal constraints (3) and for spatial factors (6) are confirmed, because the sign and explanatory significance of each independent variable do not change. 6 Thus, the results from our original equations are robust.
7. Discussion and agenda for future research Inter-municipal cooperation is a growing government reform and one gaining increasing theoretical and empirical attention. Some studies treat cooperation as another alternative service delivery mechanism in the arsenal of new public management style reforms, an alternative to privatization but driven by similar motivators, primarily cost reduction. This is most commonly found in the European studies. However, in the US studies, we find attention to a much broader array of motivating factors relating to spatial location and organizational factors. This raises the question of whether cooperation requires a broader theoretical framing than traditional studies of government contracting. We present theoretical expectations that extend beyond cost concerns (fiscal stress, economies of scale) to address governance structure and spatial differences.
Theoretically, economies of scale would be a primary motivator for cooperation; however, we find population is not significant in the majority of studies. Fiscal constraints are more likely to be found significant in studies that include small municipalities, but less likely in more recent studies and in studies from the USA. We believe this may be explained by differences in governance structure. Although small, fragmented local government is characteristic of the US and some European countries (France, Spain, Italy), more standard service delivery obligations and cross subsidies due to the wider use of joint production and municipal corporations in the European context (Tavares and Camões 2007;Gradus, Dijkgraaf, and Wassenaar 2014) may allow cooperation to better address fiscal constraints than in the USA.
While fiscal constraints are important drivers, we find some evidence of publication bias in study results. Policy makers typically assume that inter-municipal cooperation will yield cost savings, but the empirical literature on cost savings is limited and cost savings are often not found (Sørensen 2007;Holzer and Fry 2011;Bel and Warner 2015). The growing popularity of inter-municipal cooperation is not explained by fiscal drivers alone.
The majority of studies find higher rates of cooperation among suburbs (cities in metropolitan areas). Cooperation, unlike privatization, is rooted in space. The surrounding local governments create a public market of potential cooperation partners. Multi-service studies are less likely to show a positive effect of metropolitan location on cooperation. These studies may be capturing the potential for economies of scope within the municipalitywhich may explain lower rates of cooperation. Cooperation occurs both across space and across services and studies of factors driving cooperation need to be structured at the level of the local government, as a multi-product entity, in order to capture this effect. Single service studies may be appropriate for studies of cost, but they may not capture the broader array of factors driving local government cooperation.
Professional management is an important driver of cooperation and our meta-regression shows studies including small municipalities are more likely to find positive impacts of professional management. While much of the literature on transaction costs has focused on service characteristics (Brown and Potoski 2003;Levin and Tadelis 2010), our meta-regression shows the importance of organizational and spatial factors as part of transaction costs. These results confirm  call for a broader conception of transaction costs, which includes attention to spatial and managerial issues.
Our results suggest important implications for policy-makers. Local governments are facing greater fiscal stress since the Great Recession, and inter-municipal cooperation is being promoted as a solution. While fiscal constraints may drive cooperation, it is not clear that cooperation will result in efficiency gains. We do not find clear support for economies of scale in our analysis. We do find that professional management is important to manage the transaction costs of cooperation. It policy-makers wish to promote more cooperation, they should provide support for professional management and study the potential economic benefits of cooperation, as these are not always assured. Policy-makers also should recognize that cooperation is rooted in space. Spatial proximity facilitates cooperation in metropolitan areas, but what can policy-makers do to promote cooperation among more sparsely settled rural communities? Our spatial analysis also shows the potential tradeoff between cooperation and economies of scope. Cooperation across services within a community may be as important as cooperation across communities.
Our meta-regression analysis of factors explaining cooperation demonstrates that cooperation requires a broader theoretical framing that includes factors beyond the standard efficiency concerns. Studies of alternative service delivery reforms must move beyond individual service level analyses and focus on the policy challenges affecting local government as an organization in its spatial and structural contexts. Future scholarship needs to give more attention to these organizational and spatial concerns.

Disclosure statement
No potential conflict of interest was reported by the authors.