Outsourcing as management strategy in local public services

As a contribution to the academic debate on the use of outsourcing as a formula for the provision of local public services, the article proposes to focus attention on political-institutional factors. The paper departs from the review of the main arguments highlighted by the literature and it contrasts their expected results in cross-sectional study based on data from all the municipalities with over 5,000 inhabitants in Catalonia, a Spanish Autonomous Community. A typology of municipalities is constructed through principal component analysis and a cluster analysis and the existence of differences in the percentage of outsourcing among types of municipalities is tested. The results show the relevance of political-institutional factors in the selection of the service provision formula and allow us to identify three strategies that propose outsourcing as a mechanism of flexibility of public management.


INTRODUCTION
Government and local administration have shown to be a productive field for the implementation of innovative experiences in public management (Ihrke, Proctor and Gabris, 2003;Osborne and Brown, 2005), especially in the processes oriented to the improvement in the services delivery to the citizenry (Torres and Pina, 2002;Baldersheim and Wollman, 2006), and even when the organizational structures and the strategies in the public sector have to adjust for encouraging the activities and the enterprising culture through the stimulated opportunities management (Kim, 2010).
The emergence of alternative formulas to the direct provision of public services has been marked for the increasing influence of international referents.These have been incorporated to the administrative reality of the South-European countries, like the Spanish case, with certain grade of adjustment.Among these referents there are the trends integrated in the New Public Management, denomination given by the academics working in the field of public management to the public sector reforms promoted during the 80s and 90s, mainly in the countries with an Anglo-Saxon administrative tradition (Hood, 1991;Osborne and Gaebler, 1992).Although there is a lack of consensus about the content of those reforms and the existence of remarkable differences regarding its application to different political-administrative realities (Pollit and Bouckaert, 2004;Torres, 2004;Ramió and Salvador, 2005), most of them are favorable to the incorporation of market mechanisms in the public management and especially to the intensive use of outsourcing as a mean for increasing the efficiency in the service delivery.
An important part of these approaches was conceptually based on the Public Choice referent, one of the most outstanding academic approaches to the analysis of public service delivery processes.According to this approach, overproduction and inefficiency are the consequences of the monopoly of services delivery that politicians and bureaucrats exercise (Niskanen, 1971).In addition, they tend to capture returns in the management of local public services (Savas, 1987).Based in this approach, it is assumed that through the externalization of services, governments could improve efficiency and could reduce expenses in public services delivery, and that the technical efficiency in the service delivery will advance (Boyne, 1998).
Both theoretical and empirical evidence (Niskanen, 1971;Williamson, 1981Williamson, , 1999;;Ferris and Graddy, 1986;Boyne, 1998;Levin and Tadelis, 2010) indicate that the characteristics of the service are a central factor in choosing a mechanism for provision.In other words, certain services are better catered for by a particular management formula.It should be noted that the transaction costs, which include administrative costs and the costs of incomplete contracts, are important when selecting between direct production or outsourcing (Williamson, 1979(Williamson, , 1999)).Moreover, monitoring and control are critical to the outcome (Sappington and Stiglitz, 1987).It follows that factors such as asset specificity and service measurability, which affect transaction costs, will be decisive in deciding when a public service can be outsourced (Brown and Potoski, 2003b;Lamothe, Lamothe and Felock, 2008;Fernández, 2009).However, other authors, with arguments from the field of public administration, point out that the context of new emerging public services are also important in shaping the mechanisms for providing them (Zullo, 2009).The present study recognises the explanatory weight of the nature of the service in the choice of mechanism, but assumes it to be comparable within distinct town councils in order to focus attention on the differences associated with certain institutional characteristics.
The emergence of applied research on this phenomenon ranges from the classic studies on the decision to produce public services directly or not (Ferris, 1986, Ferris andGraddy, 1986), to the approaches developed from various disciplines such as economic theory (Williamson, 1981), political science (Kakabadse andKakabadse 2001, Leiblein, Reuer andDalsace, 2002;Löffler, 2003), labor sociology (Hebson, Grimshaw and Marchington, 2003) or financial economics (Johnsson, 2002).However, as Fernández indicates, research on this field is "extensive and unsystematic [...] lacking generalizability and explanatory power" (2007:1120).This fact makes it difficult to obtain systematic results (Bel and Fageda, 2007) and provides contradictory evidence regarding the effectiveness and efficiency of outsourcing by public administrations (Boyne, 1998).
Some of these studies show significant variations in the weight attributed to the considered variables, being cautious in their conclusions.Such variations may be due, among other factors, to the sample used.For example, some authors disagree with the importance given to city managers in decisions concerning outsourcing by those studies that are based on survey data from the International Association of Municipal and County Managers (International City/County Management Association, ICMA), considering that such data tend to over represent the cities that have such managerial figure (Lamothe, Lamothe and Felock, 2008).Joassart-Marcelli and Musso (2005) insist on this criticism about the survey, arguing that the size of the sampled cities (those with more than 10,000 inhabitants) is not representative.Warner and Hefetz (2002) also point out to the bias introduced by the low response rate (31% in 1992 and 32% in 1997), especially in small local administrations whose governments face strong pressures in their financial capacity, a fact that affects their outsourcing strategies.
This article aims at contributing to this debate by providing evidence based on data from a survey of all Catalan municipalities with more than 5,000 inhabitants.The universe of selected municipalities allows the inclusion in the analysis of the management options adopted by cities, with important variations among them in terms of socioeconomic aspects, government policy options and institutional characteristics.
Starting from this complex analytical reality, the study tries to provide arguments and empirical evidence on the starting assumptions of the outsourcing of local public services from its reflection in a political-administrative reality different from the Anglo-Saxon one.The paper proposes the revision of some of the main arguments highlighted by the literature on the influence of institutional factors of local governments in the form of public services management through a comprehensive transversal study of outsourcing in the municipalities of a Region of southern Europe.
The article is structured in six additional sections to this introduction.The first one presents the lines of argument and the assumptions associated with the outsourcing of services based on the research developed at the international level, as well as the hypothesis that is intended to contrast.A second section describes the basic characteristics of local governments that constitute the empirical evidence used in the analysis (those of the municipalities of more than 5,000 inhabitants of the Autonomous Community of Catalonia, Spain).The third section describes the methodology and the data used to contrast the hypothesis, followed by the results of the empirical analysis, presented in the fourth section.The fifth is the discussion section, which dialogue with the results obtained in previous studies.Finally, the sixth section presents the conclusions and points out future lines of research.

THE THEORETICAL BASES OF OUTSOURCING AS MANAGEMENT STRATEGY OF LOCAL PUBLIC SERVICES
Private provision of government services is not an unknown reality in public administration (Metcalfe and Richards, 1989), but outsourcing is innovative insofar as it involves an intentional use of the market to achieve efficiency and effectiveness gains.From this perspective, it becomes a strategy of public management.
Although outsourcing emerged in the field of economic theory, it was the paradigm of the New Public Management that emphasized it as a superior strategy to the direct management of public services (Fortin and Van Hassel, 2000).What distinguishes the outsourcing of private provision is its intentional use to achieve certain objectives.
One of the main arguments for outsourcing as a public management strategy is the specialization or delimitation of the activities that constitute the core business of the organization (Grimshaw, Vincent and Hillmott, 2002, Leiblein, Reuer and Dalsace, 2002, Kakabadse and Kakabadse, 2000, 2001).In the present study, it is assumed that the local public administration can define its distinctive activities, critical to its performance, and in which it must concentrate its attention.In the rest of activities, which are not strategic to achieve success and for which the local government does not have special capacities, it is proposed to boost outsourcing.Following these arguments, through a strategic use of outsourcing the local administration can specialize in its critical activities and take advantage of suppliers' innovation capabilities, thereby obtaining quality goods and services and gaining efficiency and effectiveness.
Based on the initial argument, numerous researches have focused on the factors that drive and explain that a local administration makes the decision to outsource.We can synthesize them into four groups.
A first group of factors is associated with the financing of local governments and their restrictions as a driving force for outsourcing.The increase in outsourcing has been associated with difficulties in maintaining the growth of public services in contexts where there is a reduction in revenue from own taxes and transfers from other administrations.As pointed out by Bel and Fageda (2007) in their review of 28 investigations on this field, the variables related to financial stress, called "fiscal stress", were used as explanatory factors for privatization in US studies in the 80's (Bel and Fageda, 2009), but neither more recent studies nor in those conducted in European contexts are so clear about that.
A second group of factors is in the field of public management, highlighting the opportunity of outsourcing as a mechanism to introduce flexibility, especially in the management of human resources not linked to the central activities of public administration (Atkinson, 1984).At the same time, outsourcing could be used as a mechanism to strengthen the power of public managers in front of workers and trade union organizations (Pfeffer, 1992).
Also related to management, a third group of factors is associated with the change in the composition of public organizations and the internal agents operating in them.This area has been highlighted both from the logic of the interest of public managers to enhance their status (Dunleavy, 1991) and from New Public Management's own arguments about the role and internal functioning of public administrations (Osborne and Gaebler, 1992).Although a multiplicity of investigations have incorporated the existence of city managers as an explanatory variable for outsourcing, the conclusions about their impact on the outsourcing level are not unanimous (Brown and Potoski, 2003a;Hefetz and Warner, 2004).
Finally, a fourth group of factors associated with outsourcing are linked to political and ideological factors.This category includes, for example, motivations related to the distribution of power and the creation of alliances between bureaucratic and political elites (Greve, 2001(Greve, , 2003) ) or motivations linked to the ideology of the mayor or the party in charge of the local government, associating those with a predominantly right-wing orientation to higher levels of service outsourcing and the predominantly left-wing ones to greater use of internal production by the city council (Young, 2005).Some research includes in this section the reference to political attitudes on the part of the citizenship, as a greater demand of smaller governments (Fernández, Ryu and Brudney, 2008).In relation to the effective impact of this last group of motivations, Bel and Fageda point out that "it is sensitive to argue that motivations linked to political interests play a more or less important role depending on the particular institutional framework of government and the type of service under consideration" (Bel and Fageda, 2007:529).This conclusion is especially relevant for the analysis proposed in the article.
Based on these arguments, it is hypothesized that, considering that the characteristics of services (their asset specificity and their service measurability) are comparable in the set of municipalities, the differences in the use of outsourcing as a strategy of flexibility can be explained from political-institutional factors.These include variables of policy and public management of each municipal organization.Among the first is the strength of the government and the ideology of the party or parties in government; between the second, the fiscal situation, the existence of certain internal management positions and the composition of the human resources of the city council.The data and research strategy used to answer the proposed hypothesis are presented below.

PRESENTATION AND DESCRIPTION OF THE UNIVERSE UNDER STUDY
The data used in this article refer to all municipalities with more than 5,000 inhabitants of the Autonomous Community of Catalonia (Spain), with the exception of its capital, Barcelona.This section is dedicated to the contextualization of the study to the Spanish political reality, to the description of the universe of local Catalan governments and to the justification of its delimitation based on the characteristics and competencies of the municipalities they manage.
After an important process of decentralization initiated after the democratic period in 1978, Spain has defined three levels of government: the State, the Autonomous Communities (regions) and local governments.The three levels are autonomous in the management of their fields of action, although they usually produce overlapping competencies and problems of coordination in the provision of public services.
The Autonomous Community of Catalonia counted in 2007 with a population of 7,210,508 inhabitants 1 and it is administratively structured as an autonomic government (Generalitat de Catalunya), four provinces, 42 county councils and 947 municipalities.In relation to the municipalities, Law 7/1985, of April 2, regulating the Bases of the Local Regime constitutes an important normative reference that, among other subjects, establishes the services that, according to the population of the municipality, must be obligatorily provided by the local administration.In addition, for various reasons such as the proximity or effective inactivity of the other levels of administration, local governments have also assumed the provision of certain services that would not fit them.A direct result of this situation is the emergence of significant financial deficits in accounts in local entities, which assume responsibilities beyond their field of competence and without adequate financing systems.
As Montesinos and Brusca point out, "Spanish local entities face the challenge of adopting new management systems to allow them to fight financial imbalances while fulfilling the principles of economy and efficiency in their basic function of providing services to citizens" (Montesinos and Brusca, 2009:200).Although for this purpose the international references such as the New Public Management are particularly relevant, the evidence is intended to highlight the particularities of its adoption by the local administration of a country linked to the Napoleonic continental tradition of the rule of law.
Among the innovations related to the introduction of these management systems stands out the option of outsourcing public services.As it is an option, different municipalities have adopted it in different areas and with different intensities, giving rise to different realities that are reflected in the set of municipalities that are analyzed in this work.
The empirical evidence is based on the survey conducted in 2007 by the Carles Pi i Sunyer Foundation of Autonomous and Local Studies to municipalities in Catalonia of more than 5,000 inhabitants, a total of 189 municipalities.The city of Barcelona was excluded from the analysis because it constitutes an atypical case within the set of municipalities due to two factors: its comparatively high population (1,668,700 inhabitants, 23% of the total population of Catalonia) and, as a result, a special status that confers much more self-government than that of other municipalities.
The survey combined information on basic data of the municipality with data on its structure and functioning (at the political and management levels), its instrumental entities and related organizations, its public services and the management formula used for its production and on economic and budgetary information.The fact that the evidence is based on all of municipalities of the selected population segment (all the universe of this population) allows us to include in the analysis the different management options developed by cities with a significant internal variability, both in socioeconomic features and in political options in government as well as in institutional characteristics, among others.The data considered are not only opinion but evidences, which were collected through the survey but also checked by consulting complementary sources of information to confirm their validity.
For the analysis of the outsourcing of local public services, a list of 30 local public services has been considered, which are obligatorily established by law, 2 together with other services that, in practice, are usually provided from this level of administration in Spain.They are: public lighting, cemetery, waste collection, street cleaning, potable water supply, sewerage, access to population centers, paving and conservation of public roads, food and beverage control, conservation of public parks, Markets, selective collection, civil protection, social services, fire prevention and extinction, sports facilities for public use, abattoir, collective urban passenger transport, environmental protection, safety, housing, museums, popular and traditional culture, Conservatories of music/music schools, cradle schools, adult schools, public actions relating to health, residences for the elderly.
The fact that the survey covers the universe of municipalities of Catalonia from 5,000 inhabitants, allows us to consider the pattern of provision of municipal services as a constant.That is, it is considered that the set of municipalities share a similar pattern of public services, both mandatory and additional.
A first approximation to the mechanisms of provision of public services offers an aggregated image that emphasizes the importance of the direct provision of the public services considered in the research.In fact, 73% of the total services analyzed are managed directly by the municipality itself, either without a differentiated organization (67%) or with its own organizations (6%).

METHODOLOGY AND DATA
In order to contrast the effect of the political-institutional variables of local governments on the outsourcing of public services, a two-stage analytical strategy is applied.In a first phase, a typology of municipalities is constructed based on a set of variables that the literature considers affecting outsourcing.To do this, a standard methodology (Roldán, 1996) is used to perform a principal components analysis of the variables of interest followed by a cluster analysis of the principal components retained.In a second phase, once the typology is obtained and the local governments are classified in relation to it, an analysis of variance is used to contrast the existence of differences in the level of outsourcing among the municipalities of the different groups, which confirms the incidence of the variables selected in the outsourcing of local public services.

VARIABLES FOR THE CONSTRUCTION OF THE TYPOLOGY OF LOCAL GOVERNMENTS
The seven variables considered are: Population.Various researches on public service delivery options have incorporated indicators referring to the characteristics of cities (Brown and Potoski, 2003a;Fernandez, Ryu and Brudney, 2008).As an indicator, the variable "population of the municipality" is used, with a continuous range of 5,008 to 252,884 inhabitants, an average of 24,446 and a standard deviation of 36,539.
Political orientation of the local government.Traditionally, the use of New Public Management's formulas has been associated with conservative ideological positions, although the transfer of these influences to the Spanish political reality also allows progressive governments to use these instruments as mechanisms for transforming public administration.But more importantly, the issue of public management in Spain has traditionally remained outside the agenda of political debate, reserving itself to more "technical" positions, especially at the sub-state levels of government and administration (Salvador, 2005).In order to define the variable "political orientation" of the team at the head of the municipal government, the positioning of the political parties in the left-right axis has been taken according to the results of the Barometer of Political Opinion of the Center for Opinion Studies (http://ceo.gencat.cat/ceop/) of July 2008.The measure of the "political orientation of the government" is the average of the positioning of the parties weighted by the proportion of members that each of them contributes to the government. 3The result is a continuous variable4 ranging from 0 (maximum left orientation) to 10 (maximum right orientation).The extreme values 0 and 10 are theoretical and would only be reached in the highly improbable assumption that all individuals responding to the survey would have placed all government parties at position 0 or position 10 on the left-right axis.But this assumption does not occur.The ideological position of the municipal governments of our study ranges from a minimum of 2.93 to a maximum of 8.27, with an average value of 4.71 and a standard deviation of 1.03.
Majority Government.According to the result of the local elections, the resulting government can count on more or less support in the full body of the corporation (institution where all political parties are represented according to their local electoral result).Although in the local political logic the figure of the mayor plays a very important role, for the purposes of this analysis his/her effective power is assimilated to the support that counts his/her government.Majority governments or majority governing coalitions are distinguished, which are associated with a greater capacity of political leadership of the government team and of robustness in order to manage the public services, minority governments or minority coalitions, in opposite situation.Although there are different options in terms of their impact on the service delivery formula, strong governments are seen as tending to assume the management and provision of public services (either directly or through instrumental entities) while weaker governments tend to opt by outsourcing as a way to simplify its political agenda.To obtain a variable that reflects the "strength of government", a dichotomous variable has been created, differentiating majority or majority coalitions (greater strength, with 161 cases), assigning them value 1, from minority or minority coalitions (lower strength, with 28 cases), assigning them value 0.
Fiscal stress.Economic crises and their impact on the public sector have forced the search for referents to maintain the expenditure containment and the provision of public services.The reforms associated with New Public Management have tended to develop in contexts of reduction of the public expense.This type of market-oriented reforms have been highlighted as strategies to 3 The formula used is: C i improve efficiency in service production and cost reduction in response to situations of Anglo-Saxon terminology known as local government fiscal stress (Joassart-Marcelli and Musso, 2005).The same arguments are proposed from conceptual frameworks such as public choice in some of its applications to the field of outsourcing, especially when it links the supposed efficiency of private enterprise and the assumption of rigidity of public employment (Moore, 1987).In these contexts, although it is difficult to create certain instrumental entities, to the extent that some of them have greater flexibility to borrow independently from the parent city council, they represent an option also appropriate to situations of financial stress.To incorporate this dimension into the analysis model, an indicator is proposed referring to the difficulties that a municipality has to borrow.The indicator used is the one proposed by Bel, Fageda and Mur (2010) that consists of the quotient between the sum of interests plus amortizations and the total income. 5The result is a continuous variable with an average of 0.09, standard deviation of 0.05, ranging from a minimum value of 0.01 to a maximum of 0.33.
Private-type local public employment.Spanish local governments, like all public administrations in the country, follow the continental tradition of public administration or "Rechtsstaat" tradition (Pollitt and Bouckaert, 2004), associated with closed civil service systems, with specific and differentiated regulation of the private sector for its management, which sets from admission through opposition to its lifetime, including standardized remuneration and career systems.However, for certain positions, municipalities may opt for hiring personnel in the private sector labor regime (private-type local public employment).In fact, this option has been particularly used prominently in the local administration, as opposed to the state and regional administrations.Although there are significant risks and doubts about their effectiveness, recourse to the hiring of private-type employment personnel (under the same regime as private sector personnel) has been conceived as one of the clearest expressions of the escape from administrative law for greater flexibility in human resources management.When there are changes in the provision of public services, the option of employing staff through the hiring of private-type employment increases the flexibility of municipal government management as it allows the temporary hiring and adjustment of human resources.Several investigations into outsourcing of services in Spain, both at the local and regional levels, show that the use of alternative formulas of service provision, such as outsourcing, are basically conceived as a way of avoiding the rigidities of the public function (Ramió and García, 2006;Ramió, Salvador and García, 2007).In this context, it is reasonable to assume that municipalities with a higher proportion of workers with a contract of employment in the private sector have greater flexibility in personnel management and, consequently, have less need to opt for outsourcing in the provision of public services than the others.That is, the municipalities with the highest percentage of private-type employment are expected to have a lower degree of outsourcing.To contrast this relationship, the variable "Private-type local public employment" has been proposed, which reports the percentage of employees with a private employment relationship among the total number of public employees of the municipality (as opposed to employees with a civil servant relationship).The average is 56.17, with a standard deviation of 19.41 in a range from 3.92 to 93.55.
State-enabled officials.This variable informs about the existence or not in each city council of officials of the state administration that occupy the positions of secretaries and interveners, who assume the functions of control of the legality of administrative acts and that watch over the validity of the budgetary management, respectively.Both posts, which must be covered in each of the local administrations, are also known as positions of state enabled officials because the persons who occupy them must be career officials who have passed an examination designed by the General State Administration (in collaboration with the Autonomous Communities) and therefore tend to receive a different training (and socialization) different from the personnel recruited and selected by the local administration.In order to support the local administration, the group of secretaries and auditors registers considerable independence from both the political and technical positions of the local corporation (Soto, 1995;Ramió, 1999).However, due to the lack of qualified personnel who can occupy such positions as well as an undisclosed will to circumvent supervision and legal and economic control, some municipalities resort to the appointment of professionals to cover these positions on an interim basis.The positions covered by this resource, supposedly of an exceptional and transient nature, actually end up being extended in time.Both by the type of designation and by the heterogeneous profile and trajectory of the professionals who occupy this position in this way, they have been considered as a professional group different from the previous one.With a dichotomous character, the state-enabled officials variable distinguishes with value 1 those municipalities that have the Secretary, Auditor or both state-enabled (147 cases) and with value 0 those that do not have any of these figures (41 cases).
City managers.From the end of the 90's numerous cities introduced the figure of the municipal manager.The incorporation of this figure implies a commitment to a model of local direction and management that is closer to the postulates of New Public Management (Longo, 2003;Brugué and Gallego, 2003).The concretion of this figure shows important differences in scope (some city councils opt for the general manager of the whole organization and others choose managers only in some areas) and their ability to act (with varying competencies, determined in each case) (Ramió, 1999).Although these figures could not be equated directly with the City Manager of the local administration of the United States, there are notable parallels to establish comparisons.In order to include this variable, 1 has been assigned to municipalities that have either a unitary manager for the whole organization or managers in any of the sectoral areas (62 cases) and with value 0 those cases in which no figure appears with this profile (121 cases).

STATISTICAL TECHNIQUES APPLIED
The principal components analysis is a statistical technique that linearly transforms an initial set of variables into a smaller number of uncorrelated variables that represent most of the information contained in the original variables (Dunteman, 1989).Its usefulness in this case is in the elimination of the redundant information existing in the initial data (since the original variables were correlated) and in that being the principal components standardized variables all have the same weight in the cluster analysis that is performed later.The technique uses the correlation matrix, which is why the variables used should be continuous or dichotomous with values 0 and 1.In this analysis we have used four continuous variables and three dichotomous variables that have been summarized in three principal components.These three components synthesize the existing variability among municipalities in terms of the political-institutional externalizing factors highlighted by the literature and are the variables used in the cluster analysis.
The cluster analysis is a classification technique that aims at forming groups of objects that are as similar as possible within each group and, in turn, that the resulting clusters are so different from each other as possible, in relation to a series of variables of interest (Aldenderfer and Blashfield, 1984).Since the technique is based on the calculation of distances, it is desirable that all variables that enter the analysis have a same range of variation to avoid that the ones of greater variation dominate over the others.Applied to our study, this technique allows the formation of relatively homogeneous groups of municipalities with regard to the factors that affect outsourcing.
The analysis of variance is a multiple comparison test of means.It allows contrasting the null hypothesis that the means of several populations are equal, against the alternative hypothesis that at least one of the populations differs from the others in terms of their expected value.It requires compliance with the assumptions of normality and equality of variance for populations for the dependent variable.Although the F-test of the variance analysis is robust to the lack of normality and not sensitive to the violation of the homoscedasticity assumption (Moore, 2005), we have chosen to use the robust statistics of Welch and Brown-Forsythe asymptotically F distributed) instead of the F statistic in those cases in which the equality of variances between groups cannot be assumed.
All statistical analyzes were performed with the SPSS program.

RESULTS OF THE EMPIRICAL ANALYSIS
A principal component analysis has been performed with the seven variables that characterize the local governments described in the previous section.Three components accounting for 59% of the total variance of the seven variables included in the study have been retained, without any of them being poorly represented. 6An orthogonal rotation was applied to the retained components in order to obtain a simpler structure of correlations that facilitates interpretation.The purpose of the rotation is that each of the new components has high correlations with some variables and low correlations with others, so that the component can be interpreted as a factor associated to the set of variables with which it strongly correlates.
Table 1 shows the matrix of rotated components (also called weights matrix) whose cells contain the existing correlations between the three rotated components and the original variables.To facilitate the reading of the table, the most important correlations between each component and the original variables are in bold.
6 The percentage of explained variation of each one of the variables oscillates between 0.41 and 0.72.The first component is a factor positively associated with the variables population size and the existence of city managers and, in relation to the composition of public employment, negatively associated with the percentage of workforce hired with private-type local employment.Municipalities with high values in this first component are, therefore, municipalities of large dimensions, generally with presence of city managers and with few personnel hired by the work path following the benchmark of the private sector.
The second component can be interpreted as a factor associated with the characteristics of the municipal government and its economic situation, since it correlates negatively with the strength of the government and positively with its ideological position and the fiscal stress of the city council.Municipalities with high values in this second component have, therefore, governments in a situation of weakness (in minority or minority coalitions), of right wing and with high level of indebtedness.
The third component only correlates in a significant way with the existence of state secretaries and /or auditors authorized in the city council.Therefore, high values in the third component correspond to municipalities with presence of state-enabled officials in charge of the legal and economic supervision of their actions.
A cluster analysis applied to these three components allowed us to establish a typology of municipalities based on their characteristics of a) population size, presence of managers and composition of public employment, b) ideology and strength of government and financial stress, and c) Presence of enabled personnel at the state level that controls and supervises their actions. 7s a result of this analysis, three groups of municipalities have been identified.The first type of municipalities, which we propose to call "Private-type employment city councils", contains just over two-thirds of the total number of cases (69%) and is composed of small towns with strong governments with a certain left-wing ideology.At the administrative level, they have a significant presence of workforce with contracts in the private sector labor regime; very few have city managers and are those with lower fiscal stress.These are local governments that have maintained the basic characteristics of the continental administrative tradition, looking for flexibility options in the management of human resources through the recruitment of personnel through private sector formulas.
The second group of municipalities of this typology is constituted by the "managerial city councils", counting on 18% of the cases.They are characterized by large dimensions and by strong and leftwing governments.At the administrative level they tend to have city managers and little workforce hired below the private sector regime.And, comparatively, have a low level of debt.The distinctive feature of the local governments included in this group is the option for the managerial model, which approximates the approaches of the New Public Management as a mechanism to transform the municipal administration.
The third group of this typology would be formed by "indebted city councils", with 13% of the cases.As main features, they generally have minority or minority coalition governments, with a certain right-wing ideological orientation, with less presence than in others of state-enabled officials for the supervision of legal and economic issues, and heavily indebted.A distinctive feature of the municipalities included in this grouping is their relatively small strength, attending both to the characteristics of the municipal government and to the lack of state officials that can supervise its operation.
An analysis of the variance applied to each of the variables that have been used for the construction of the typology allows to establish that, in fact, there are significant differences in the mean values of the variables considered among the three groups of municipalities identified, confirming the validity of the typology and the characterization of the three groups described in the previous paragraphs.Table 2 presents the mean values of the numerical variables.Table 3 presents the mean values of the dichotomous variables expressed in percentages.This allows the interpretation of their values as a percentage of municipalities in each group that present each of the characteristics indicated.Since these variables do not meet the assumption of equality of variance between groups, the robust statistics of contrast of Welch and Brown-Forsythe have been used.The presence of state-enabled officials is the only variable among the seven variables considered that had no significant differences observed in the mean value of each group.
Once the typology was constructed, the level of outsourcing of each of the three types of municipalities has been calculated.For this purpose, the average value for municipalities in each group of the percentage of outsourced services in the list of 30 local public services mentioned above has been taken as an indicator.8For this variable, the assumption of variance homogeneity is satisfied and the F statistic of the analysis of the variance can be used.The result indicates that there are significant differences between the three types of city councils as regards the percentage of outsourced services (table 4).A more detailed analysis of multiple comparisons indicates that in the "Private-type employment" city councils the percentage of outsourcing is significantly lower than in the "managerial" and "indebted" groups, although there are no significant differences between the last two groups of municipalities.9

DISCUSSION
The analysis of the influence of political-institutional factors on the outsourcing of local public services has allowed us to establish typologies and point out divergent strategies regarding the outsourcing of local public services that can contribute to the academic debate in this field.
The analysis of the Catalan case provides evidence on the arguments highlighted by the international literature regarding the outsourcing of public services (outlined in the section on "The theoretical bases of outsourcing as management strategy of local public services" of this article): 1.The arguments associated with the financing of local governments only appear as a prominent reference to explain outsourcing in the third group of municipalities (the "indebted", which make up 13% of the total), which seems to be more clearly committed to the outsourcing of services.Such result is in line with the conclusions of Bel and Fageda (2007) since that, beyond its explanatory potential in Anglo-Saxon realities, in the European institutional context the relationship between fiscal stress and outsourcing is not so clearly manifested.2. The evidence provided by the analysis also favors the arguments that link externalization to the composition of public organizations and the emergence of figures such as city managers.Although the Anglo-Saxon literature on the figure of the city manager (which is complicated to equate the manager figure) does not attribute decisive conclusions in the promotion of outsourcing (Brown and Potosky, 2003a;Hefetz and Warner, 2004), other references in the Spanish context do point to a clearer linkage (Ramió, 1999).On the other hand, the existence of state-enabled officials does not seem to have a significant relationship with the outsourcing, specifically the secretaries and interveners who are figures more settled in the Spanish local administration model.
3. Ideological factors do not seem to have a clear impact on the adoption of externalizing strategies.In fact, according to the data provided, the two groups of municipalities with the highest outsourcing of services have government teams with different ideological composition.These findings are in opposition to those recorded in Anglo-Saxon realities (as in the Australian case analyzed by Young, 2005) but they correspond to the conclusions pointed out in analyzes conducted in the Spanish context such as that of Bel and Miralles who suggest that "the decisions to contract out seem to have been motivated by pragmatic rather than ideological reasons" (Bel andMiralles, 2003:1323).4. The strength of local government appears as an ambivalent factor to explain outsourcing, since it characterizes both "managerial city councils" and "Private-type employment city councils", municipalities with countervailing strategies regarding the use of outsourcing services.This result that can be interpreted in terms of clarifying a strategy (as a reflection of the weight of institutional factors pointed out by Bel and Fageda, 2007), but not in a single direction in terms of outsourcing.5.One factor that has emerged as particularly significant in explaining outsourcing is the percentage of workforce employed under private sector regulation.An increase in this group is associated with the search for flexibility against the rigidity of the human resources management system in the Spanish public administration.In this sense, the hiring of personnel in the private sector labor regime (private-type employment) is considered as an alternative to outsourcing.These results can be interpreted in terms of incorporating instruments of private management to increase flexibility in the management of public sector human resources (Ramió and García, 2006;Ramió, Salvador and García, 2007).
Considering the results obtained, the article contributes to the academic debate about the provision of public services at the local level focusing attention on the role played by political and institutional factors.The support of the study hypothesis suggest that, beyond the nature of the service provided, there are certain political-institutional factors that contribute to explain the option to outsource local public services.In connection with the academic debate in this area, the contributions made correspond to the existing literature regarding factors linked to the fiscal situation and the role of local government ideology, yet highlighting the differences between Anglo-Saxon realities and those that are separated from this administrative tradition (as the case analyzed).In the same vein, the article also provides evidence regarding specific factors from the Spanish institutional context, such as the existence of a city manager or the rigidities in the management of human resources, which play a decisive role in the choice of outsourcing services.

CONCLUSIONS
The approach to the outsourcing processes of the local public services in a political-administrative context linked to the tradition of public administration Continental or "Rechtssaat" allows glimpsing complementary visions in its configuration.The evidence provided by the municipalities of more than 5,000 inhabitants of the Autonomous Community of Catalonia allows covering a diversity of institutional configurations and management options.This universe of local governments has made it possible to use a series of variables that are difficult to use in comparative studies, such as the pattern of service delivery.
These results allow illustrating the existence of three differentiated lines of action in relation to the external contracting of public services as a formula of flexibility in the local public management.While two of them clearly opt for outsourcing, the third makes a limited use of that option.
The first strategy is the use of outsourcing of services as a mechanism to escape the complexity of managing public services in situations where there is no strong government or there are budget deficits that hamper other management options.This route is used by the "indebted city councils." The second strategy is based on the incorporation of professional management figures such as city managers (both general for the entire city council and sectoral for certain areas).It is a commitment to an in-depth transformation of the management model that includes, among others, the outsourcing of local public services.This is the path used by "managerial city councils." The third strategy, alternative to outsourcing, is the hiring of personnel under the labor private sector regime (private-type employment) as the main flexibility formula in public management, maintaining the direct provision of most public services.This route is the one used by "Private-type employment city councils." The identification of these three differentiated options in relation to the use of outsourcing as a management strategy in local public services represents a contribution of the article to the academic debate related to the role of political and institutional factors in the determination of the modalities of provision of public services.This contribution is especially relevant to interpret the phenomenon in administrative realities typical of a non-Anglo-Saxon context such as the region of southern Europe is.
The characterization and development of these strategies, however, require more in-depth studies to investigate their scope and effectiveness, as well as their implications in relation to the transformation of local public management.
of parties that form the government Ci = number of government members of party i Wi = positioning of party i = mean of the position on the left-right axis of party i E = total number of government members = k Σ i = 1

TABLE 4 LEVEL
OF OUTSOURCING BY TYPE OF MUNICIPALITY Source: Elaborated by the authors.