The political hourglass: opportunistic behavior in local government policy decisions

Abstract

The aim of this article is to determine whether there is opportunistic behavior in local government decisions related to contracting out, public-private partnership (PPP) and/or public externalization. Analysis of the results obtained from applying a fixed effects panel data model to a sample of 2,274 Spanish municipalities supports the existence of asymmetric opportunistic behavior. Specifically, there is strong evidence that in preelection years, there is less likelihood of contracting out and PPP decisions being approved, and a greater probability of public externalization taking place. In post-election years, an increase in capital spending is associated with a lower probability of contracting out and of PPP. An increase in current spending in preelection years reduces the probability of decisions being taken on contracting out and PPP. These empirical results highlight the extent of opportunistic behavior in selecting management forms for the delivery of local government services.

Introduction

A major strategic decision that public sector organizations must address when designing a public policy is that of which organizational form will be used to implement it (Esteve et al. 2013; Prior et al. 2019). In this respect, the literature has mainly focused on the determinants of collaboration, that is, on why certain organizations implement their policies by working with other organizations rather than doing so alone (Krueathep, Riccucci, and Suwanmala 2010; Lundin 2006; McGuire and Silvia 2010; Smith 2007). Other studies have analyzed the influence of environmental, organizational and individual variables in this field (Bryson, Crosby, and Stone 2006; Koppenjan and Enserink 2009; Krueathep et al. 2010; Lundin 2006; Mandell and Steelman 2003; McGuire and Silvia 2010; O’Toole 1997; Smith 2007). However, previous research has not examined the question of how these determinants may lead public managers to decide on particular organizational forms to implement their policies, such as contracting out, public-private partnerships (PPP) or public externalization. In the present study, we analyze an important factor that is likely to affect strategic decisions concerning the management form adopted for the delivery of public policies. This factor is the consideration given to the political cycle – that is, the proximity of elections.

Traditionally, the literature about delivery forms has centered on the public vs. private dichotomy. However, in recent years, governments have increasingly used different organizational forms to implement public policies, and many of these forms are collaborative ventures involving the private sector to a greater or lesser degree (Pérez-López, Prior, and Zafra-Gómez 2015; Gradus, Dijkgraaf, and Wassenaar 2014). As Schoute, Budding, and Gradus (2018) stated, “whereas the main distinction made was traditionally between in-house and out-house (contracted to a private provider) service delivery modes, other forms, such as inter-municipal cooperation and municipally owned firms, can now be commonly observed” (2018: 503). Accordingly, studies of public-service delivery forms should take into account alternative arrangements, such as intermunicipal cooperation, mixed firms and private intermunicipal cooperation. At the same time, and paradoxically, private sector participation in the delivery of public services, in the form of contracting out, PPP or privatization, tends to be unpopular among voters (Battaglio and Legge 2009; Boyer and Van Slyke 2019; Boyer, Rogers, and Van Slyke 2018; Thompson and Elling 2000). Little is known about how considerations of voters’ reactions may affect policymakers’ strategic decisions in this respect. The notion of opportunistic behavior (Nordhaus 1975; Blais and Nadeau 1992; Pentecote et al. 2004) provides a theoretical framework for examining why public managers prefer to adopt certain organizational forms rather than others to implement their policies. According to the theory of opportunistic behavior, as elections draw near, politicians will increase public spending, cut taxes and/or adopt certain management forms for the delivery of public services in order to boost their own popularity and enhance their chances of reelection (Foremny and Riedel 2014). In this sense, the local population may oppose certain decisions, such as the contracting out of municipal services. This is based on the citizens’ belief that private sector involvement would raise the delivery cost of these public services (González-Gómez, Picazo-Tadeo, and Guardiola 2011; Garrone and Marzano 2015; Schoute et al. 2018). Accordingly, policymakers’ decisions may be affected by electoral considerations, i.e. the stage of the political cycle. Studies have analyzed the effects of opportunistic behavior on contracting out decisions (Fernández, Ryu, and Brudney 2008; Garrone and Marzano 2015; De la Higuera-Molina et al. 2019), but to our knowledge none has expanded this focus to include other organizational forms and thus better reflect the reality addressed by public managers. The aim of this study is to advance our understanding of how public sector decisions are made, by examining whether politicians engage in opportunistic behavior in choosing management forms for the delivery of public services.

To identify and measure the presence of political opportunism in the delivery of public services, we developed a model for each of these organizational forms, thus enabling us to establish whether policymakers employ opportunistic behavior in their strategic decisions about how to implement a public service. In this analysis, we took into account the level of risk acceptable to the local government in the provision of public services and the possible variations in the price charged to the public for these services. Using a fixed effects panel data method, we analyzed 2,274 Spanish municipalities, each with a population of over 1,000, for the period 2002–2014, during which three municipal elections took place (in 2003, 2007 and 2011).

The results obtained reflect the presence of opportunistic behavior associated with the political decisions taken regarding delivery forms for local public services. In the preelection period, fewer decisions were taken to contract out services or to establish a PPP, and public managers were more likely to resort to public externalization during this period. Furthermore, the increase in current spending that often took place prior to an election reduced the probability of contracting out and PPPs. Finally, our results also indicate that in post-election years, the capital spending of the municipality becomes important, as it is associated with a lower probability of the government entering into contracting out or PPP arrangements.

Opportunistic behavior and the use of public and private delivery forms

Building on Down’s (1957) notion that politicians are driven by their own interest, the theory of opportunistic political cycles argues that political decisions may be linked to opportunistic behavior (Nordhaus 1975; Alesina 1989; Binet and Pentecôte 2004). The latter term refers to the motivations underlying certain decisions taken by public managers, according to the stage of the electoral cycle (Foucault, Madies, and Paty 2008). In this respect, Nordhaus (1975) and Bastida, Beyaert, and Benito (2013), among others, claim that political decisions may change according to the current moment of the political cycle. Specifically, policymakers will prioritize their own interests when elections are imminent, in order to secure reelection (Veiga and Veiga 2007; Benito, Bastida, and Vicente 2013), and so their decisions will not necessarily be taken with a view to the general interest of the population (De la Higuera-Molina et al. 2019). In this respect, public managers can make use of various instruments and policies to further their own interests (Foremny and Riedel 2014), such as decisions related to public debt (Geys 2007; Benito et al. 2013) or to budgets and taxes (Persson and Tabellini 2003; Klomp and De Haan 2013; Bastida et al. 2013). However, there is very little empirical evidence analyzing how this behavior influences the choice of which organizational form should be used to implement a public service. Research has shown that citizens tend to have a negative perception regarding the participation of the private sector in the management of local public services. There seems to be the perception that the cost and quality of the service will be worse when this is implemented by the private sector (Donahue and Miller 2006; Garrone and Marzano 2015; Baekgaard and Serritzlew 2016; De la Higuera-Molina et al. 2019). Hence, the choice of the organizational forms that governments use to implement public services may be influenced by the opportunistic behavior of the public manager or the politician in charge of the decision.

Fernández et al. (2008) concluded that contracting out processes are opposed both by voters and by elected officials. Citizens usually oppose the participation of the private sector in the delivery of public services in the belief that the services provided through this type of management have a higher price than those provided via a public-sector organization (Rubin 1998). In this vein, another recent study suggested that public officials may engage in opportunistic behavior when contracting out the delivery of public services (De la Higuera-Molina et al. 2019). Public officials are aware of the negative perceptions aroused among the electorate by the participation of the private sector and so they prefer to avoid such measures when elections are imminent (Garrone and Marzano 2015). Another argument that has been proposed to explain this negative perception of the private sector is that service quality falls when the private sector becomes involved in the service delivery, because the private provider will seek to reduce costs in order to increase revenues. Another significant factor is that many contracting out operations are accompanied by the introduction of a new fee for the service, or an increase in the existing one (López-Hernández et al. 2018; De la Higuera-Molina et al. 2019). In addition, it has been argued that changes in management methods to incorporate private sector operators in the provision of local public services may also enable opportunistic behavior by the operator, such as deficient service and/or abusive pricing (De la Higuera-Molina et al. 2019). On the other hand, Fernández et al. (2008: 452) observed that strict supervision by the public authority can reduce the incidence of such opportunistic behavior. However, to date, there is a lack of empirical evidence as to whether opportunistic behavior arises when the private sector is invited to participate in the delivery of public services.

To understand better the influence of opportunistic behavior on the political choice of service delivery form, we must first discuss the main differences gamong the organizational arrangements public managers can choose One form is that of contracting out¹, i.e. the complete transfer to the private sector of the delivery of public services. In this case, the public authority transfers all the risk to the private operator and this may result in increased prices being charged to the end user. An alternative arrangement is that of indirect management, with a mixed approach to the delivery of public services (the public-private partnership, PPP²), in which ownership of the service is shared between the public administration and private enterprise. In most cases, the political authority retains control of the agreement and the private entity operates under the commercial regulatory system (Warner and Bel 2008; Andrews, Esteve, and Ysa 2015). In this study, we also consider whether indirect management, via a mixed economy, is affected by opportunistic behavior, and whether the latter is associated with the transfer of risk. By establishing a PPP, the government may enter into a longer-term relationship with private entities, thus creating a more durable and reliable association for the provision of public services.

However, there is an alternative to contracting out or to PPPs, which does not imply the participation of a private organization. This option is that of public externalization³, i.e. direct management by a local commercial company, an organization that is owned by the municipality; therefore, no part of the business risk is transferred to a third party (Warner and Bel 2008). According to the postulates of New Public Management, public companies have a similar status to that of agencies, since they have autonomy and a high degree of budgetary independence (Pérez-López et al. 2015). Moreover, in comparison to the public sector, they can more readily expand or contract the workforce, they are less subject to hierarchisation and their decision-making process is faster and more proactive (Andrews, Boyne, and Walker 2011). In consequence, these public firms achieve greater flexibility in the provision of local public services.

Taking into account the above arguments, we hypothesize that in post-electoral years public managers will prefer, first, contracting out and, otherwise, PPPs as management forms for the delivery of local public services. In the first of these options, all risk is transferred to the private operator, and in the second, the transfer is partial. However, since these measures are poorly perceived by voters (as they fear consequent price increases), policymakers make less use of these forms before and during an election year. Therefore, in the pre-electoral period, any change in this area is more likely to take the form of public externalization. In view of these considerations, we propose the following hypotheses:

H₁: In post-election years, there is a greater probability of local public services being delivered via contracting out or private-public partnership.

H₂: In pre-election years, there is a greater probability of these services being delivered via public externalisation.

Some of the factors that could most strongly influence contracting-out decisions associated with opportunistic behavior are the political ideology of the governing party and the degree of fragmentation in the composition of the local government (Dubin and Navarro 1988; Ni and Bretschneider 2007; Chortareas, Logothetis, and Papandreou 2016). In relation to the first of these, if we assume that political-public managers reach decisions on local issues in accordance with the positions held by their parties, we would expect their decisions to be taken in accordance with the theory of partisan political cycles (Hibbs 1977; Veiga and Veiga 2007; Benito et al. 2013; Guillamón, Bastida, and Benito 2013). On the other hand, this approach conflicts with the theory of opportunistic political cycles that we propose. According to the partisan approach, regardless of the political cycle, conservative governments will prefer to contract out local public services, and evidence in this respect has been reported (Elinder and Jordahl 2013; Zafra-Gómez et al. 2016). However, not all the studies made in this regard have supported the existence of a relationship between ideology and contracting out (López-de-Silanes et al. 1997; Lane 2000; Dijkgraaf, Gradus, and Melenberg 2003; Ohlsson 2005; González-Gómez et al. 2011).

In contrast, the theory of opportunistic political cycles suggests that there are no differences between conservative and progressive political-public managers in relation to the introduction of market mechanisms (Dubin and Navarro 1988; Dijkgraaf et al. 2003; Walls, Macauley, and Anderson 2005; Ni and Bretschneider 2007; Zullo 2009). In relation to the partisan model, progressive governments are more likely to implement the direct provision of public services (Bel and Fageda 2007), in preference to market mechanisms such as contracting out or PPP, while conservative political parties are strongly inclined to contract out these services (Zafra-Gómez et al. 2016). Hence, we expect that the association between the proximity or otherwise of elections and policymakers’ decisions regarding contracting out or the establishment of PPPs for the delivery of local public services will be moderated by political ideology. In short, local governments controlled by a party with a conservative ideology will be less affected by the proximity of elections. Progressive governments, on the contrary, are more likely than conservative ones to adopt contracting out agreements and/or establish PPPs during the preelection period, and therefore we propose the following study hypothesis:

H₃: The relation between the stage of the electoral cycle and the decisions to contract out local services or to participate in a PPP is moderated by political ideology in the following way:

H_3a: Conservative political ideology positively moderates the effect of pre-election periods on decisions to contract out local services or participate in a PPP.

H_3b: Progressive political ideology negatively moderates the effect of pre-election periods on decisions to contract out local services or participate in a PPP.

Factors underlying the use of different delivery forms for local government services

Public spending and its influence on delivery form decisions

In many cases, the main reason for incorporating market mechanisms into the provision of local public services is to achieve cost savings (Bracci and Llewellyn 2012). Interestingly, while citizens perceive that the participation of the private sector in the delivery of a public service will increase service costs, public managers seem to think the opposite. Accordingly, it may be useful to analyze and characterize the relationship between public spending and management forms such as contracting out, PPP, and public externalization, and how their use varies during the electoral cycle. In this respect, we hypothesize there may be a direct relationship between variations in public spending during the political cycle and these forms of management, distinguishing between capital and current spending. Another plausible argument to support this theoretical model can be found in the Fiscal Illusion theory, according to which citizens lack information not only about the fiscal income to which they contribute but also about the public spending decisions taken by the government (Dollery and Worthington 1996; Sausgruber and Tyran 2005; Baeekgard, Serritzlew, and Blom-Hansen 2016). Therefore, we seek to determine the influence of each type of spending on the management form adopted.

This examination of public spending is based on the theory of budgetary political cycles (Nordhaus 1975; Galli and Rossi 2002; Bhatti, Olsen, and Pedersen 2009), according to which as an election period approaches, local public managers tend to provide better quality public services and increase public spending in order to improve their likelihood of reelection (Schuknecht 2000; Drazen and Eslava 2010).

In this vein, empirical evidence has revealed the existence of such opportunistic behavior in the management of capital spending⁴ by local governments (Veiga and Veiga 2007; Foremny and Riedel 2014). Public managers employ this type of spending at the beginning of the term of office, expecting its effects to be apparent to the voting population before the electoral cycle ends. The introduction of the private sector into the delivery of local public services is normally expected to decrease public spending in this regard, i.e. there is an inverse relationship between the presence of contracting out/PPP management forms and the level of municipal capital spending. Accordingly, public managers will seek to incorporate these management forms in the delivery of the public services that are most expensive to provide (Shrestha and Feiock 2004; López-Hernández et al. 2018). Therefore, if changes in the provision of public services present a dynamic behavior, we hypothesize that capital spending may influence this relationship.

Opportunistic political behavior may also be reflected in current spending⁵. Empirical studies have shown that in the preelection period this type of spending tends to increase, presumably so that citizens will perceive a higher level of service provision (Veiga and Veiga 2007; Vergne 2009). At this stage of the electoral cycle, public managers prefer not to take measures that will have a negative impact on their chances of reelection, such as developing agreements with the private sector in the provision of local public services (Fernández et al. 2008; Garrone and Marzano 2015). Instead, they will opt for increasing current spending and provide services diretly, instead of engaging in collaborative agreements with the private sector. In view of these considerations, we believe there exists an opportunistic relationship, in political terms, between current spending and its influence on the choice of private forms of service provision (i.e. contracting out and PPP).

The influence of taxation and socioeconomic factors on the choice of public service delivery forms

One of the most important factors taken into account by local governments is the effect of taxation and of changes in this respect according to the stage of the political cycle. This consideration is of major importance with regard to the provision of local public services. According to the theory of opportunist political cycles (Dahlberg and Mörk 2011), in order to secure their own reelection, public managers will often take finance-related measures such as reducing taxes. As well as decisions related to spending and taxation, the stage of the political cycle may also influence other political issues, such as the preferred management form for the provision of local public services. As the theory of opportunistic political cycles predicts, taxes would be reduced during the preelection period in order to lighten the fiscal burden on the population (Revelli 2002; Binet and Pentecôte 2004; Foucault et al. 2008; Dubois and Paty 2010; Foremny and Riedel 2014).

Furthermore, electoral considerations might also influence the way in which local public services are provided. Specifically, management forms such as contracting out, PPP and public externalization may be adopted in order to reduce costs and to alleviate fiscal stress, but the effects produced by each one will vary, and so political considerations, in terms of whether such changes benefit or prejudice the possibility of reelection, will be taken into account. Thus, tax rebates associated with private management formulas will be produced to a greater extent following elections, while tax rebates and public externalization are more likely to take place in the preelection period. Hence, it is useful to analyze these questions from the standpoint of political opportunism (Binet and Pentecôte 2004; Foremny and Riedel 2014; De la Higuera-Molina et al. 2019).

Another important factor underlying the decision to contract out public services or to participate in a PPP is that of the population size. Previous studies have argued that when population size increases, so does the need for municipal services (Christoffersen and Bo Larsen 2007). Policymakers may try to cope with this increase in their populations by developing more contracting out operations, PPPs, or public externalizations, as politicians seek to minimize the costs involved and avoid tax increments (González-Gómez et al. 2011; Zafra-Gómez et al. 2016).

Other factors affecting delivery form decisions during the political cycle

The level of political fragmentation is also relevant to the strategic decisions taken by local governments (Geys 2007; Benito et al. 2013). In general, greater fragmentation is associated with higher levels of public spending, and therefore with a greater likelihood of public services being contracted out or provided via a PPP. Ashworth et al. (2005) concluded that political fragmentation increases the possibility of higher long-term public spending, and therefore municipalities in this situation would seek to eliminate the more expensive services from their budgets. In consequence, the existence of political fragmentation may influence the management form adopted at different stages of the political cycle.

The election of a new city manager/mayor is another variable that may influence delivery form decisions (GAO 1997; Chi and Jasper 1998; Ni and Bretschneider 2007). This is because elected city managers are more likely to implement unpopular measures at the beginning of the term of office, as they will not face new elections usually for the next four years. Moreover, the decisions that elected city managers make at the beginning of their mandates are linked to the political ideology of their political party, as described by the partisan political cycle’s theory (Hibbs 1977).

The final variable considered is that of the mayor’s gender, which may also be an explanatory factor in decision making regarding the choice of management form (i.e., contracting out, PPP or public externalization). There is a substantial body of empirical evidence relating gender and risk aversion in strategic decisions (see, for example, Hudgens and Fatkins 1985; Johnson and Powell 1994). These authors describe that famel managers tend to be more risk averse than their male counterparts. When considering how this could affect the specific decision of involving the private sector into the delivery of public services, one could argue that gender will play an important role. Contracting out or engaging in a PPP entails a higher degree of risk than implementing a policy within the public sector, via direct proivision of the service. Hence, we expect that female public managers will be less likely to contract out or to engage in PPPs when compared with male public managers.

Methods

The Spanish public sector is composed of three major areas of government: the General Administration of the State (national government), the Autonomous Community Administration (regional government) and local administration. The legal and organizational configuration of the latter is regulated by the Local Government Act (LRBRL) 7/1985, of 2 April, amended in 2013 by the Local Administration Rationalization Act, which stipulates the local public services that must be provided, according to the size of the municipality. However, the form of service delivery of local public services is at the discretion of local public managers (elections are usually held every four years). In the present study, we use the words politician and public managers indistinctively, because in the Spanish municipal government, most public managers are, in fact, elected politicians, as the figure of a professional public manager is only present in very few municipalities (usually only in big cities). And even where the figure of a public manager does exist, the post is highly politicized (see Losada and Esteve 2008 for further details on the relationship between public managers and politicians in Spain).

Our analysis is based on the use of a Fixed Effects Panel Logit model, after confirming the importance of the panel-level variance component, by the likelihood-ratio test of rho. This test formally compares the pooled estimator (logit) with the panel estimator for each of the three models considered. To justify the use of the Conditional Fixed Effects model, the Hausman test was applied to each of the three models implemented (Contracting out, PPP and Public externalization). In view of the test results obtained, the null hypothesis of independence of the study variables was rejected. This finding led us to employ the fixed-effects estimator rather than the random-effects estimator (Frondel and Vance 2010).

Study data

Our empirical analysis is based on annual data obtained from various databases, for all Spanish municipalities with a population exceeding 1,000 inhabitants (3,129 municipalities in all). From these data, we obtained observations of the dependent variables (contracting out, PPP and public externalization) over an extended period, from 2002 to 2014, as well as the temporal variation of the covariables.

The study variables were obtained from budgetary and balance sheet information. However, although all local authorities are obliged to provide budgetary information to the Directorate General for Financial Coordination with Regional and Local Authorities (DGCFCAEL, part of the Ministry of Finance and Public Administration), some do not comply with this requirement, which limits the sample size available. Furthermore, in line with Bastida, Benito and Guillamón (2009), our analysis of the municipal situation in Spain is limited to municipalities with more than 1,000 inhabitants because, as the latter authors point out, for very small municipalities the reliability of financial data is doubtful; moreover, the information for certain variables was only available for municipalities with more than 1,000 inhabitants.

Accordingly, municipalities with fewer than 1,000 inhabitants were excluded from this analysis, as were those for which financial information was not available for one or more of the years of the study period (2002–2014). The final sample used in the study was composed of 2,274 municipalities. This proportion is highly representative of the total of 3,129 municipalities in Spain with more than 1,000 inhabitants.

Among the public services analyzed were public lighting, cemetery services, refuse collection, street cleaning, drinking water treatment and supply, drainage and sewers, access to population centers, and street paving. These constitute a suitable range of services for analysis. The time period addressed (2002–2014) reflects the effects on these services of three successive local elections (2003, 2007 and 2011).

The dependent variable in this study is the organizational form used for each of the municipal services analyzed (see Tables 1 and 2). To construct the final database, each case was coded to reflect when the service in question was provided by contracting out, by PPP or by public externalization. We obtained this information by examining the Official Bulletin of the Province (BOP) and then constructing a database of all the announcements published in the BOP by the municipalities that sought to contract out public services. Hence, the dependent variable is a binary variable in which we have assigned the value 1 when one of the above management forms is used, and the value 0 otherwise. The remaining independent variables were obtained from various data sources, as shown in Table 1. The data compiled include the year in which the service was derived to one of the above three management forms, the existence (if any) of fees charged for contracting out, PPP or public externalization, and when applicable the duration of the contract, and a description of the contract conditions and the contracting entity. Taking these considerations into account, as well as the fact that a municipality may have had one or more services subject to contracting out, PPP or public externalization during one or more of the years of the study period, a dichotomous dependent variable was created for each year studied. This variable was awarded the value 0 if there was no contracting out, PPP or public externalization, and the value 1 if any such operation took place).

Table 1. Description of the variables.

Variable	Description	Source
Contracting out (dependent variable)	Dichotomous variable that takes the value 1 when the service is contracted out and the value 0 otherwise	Official Bulletin of Province (BOP)
Public-private partnership (dependent variable)	Dichotomous variable that takes the value 1 when the service is PPP and the value 0 otherwise	Official Bulletin of the Province (BOP)
Public externalization (dependent variable)	Dichotomous variable that takes the value 1 when the service is public externalization and the value 0 otherwise	Official Bulletin of the Province (BOP)
Election	Dichotomous variable that takes the value 1 in the pre-electoral years and the value 0 in the post-electoral ones	Derived by the authors
Capital spending-Postelectoral years	Interaction between capital spending and postelectoral years	Derived by the authors from data published by the General Secretariat for Local and Regional Coordination
Current spending-Preelectoral years	Interaction between current spending and preelectoral years	Derived by the authors from data published by the General Secretariat for Local and Regional Coordination
Political ideology	Dichotomous variable that takes the value 1 when the party in government has a conservative ideology and the value 0 when it has a progressive one	Ministry of the Interior
Conservative absolute majority	Dichotomous variable that takes the value 1 when a conservative party governs with an absolute majority and the value 0 when it governs in coalition or when another party governs.	Derived by the authors
Progressive absolute majority	Dichotomous variable that takes the value 1 when a progressive party governs with an absolute majority and the value 0 when it governs in coalition or when another party governs.	Derived by the authors
Change of mayor	Dichotomous variable that takes the value 1 when a different mayor takes office and the value 0 when the same mayor continues in office	Derived by the authors using data supplied by the Ministry of the Interior
Change of mayor-Postelectoral years	Interaction between political change and postelectoral years	Derived by the authors
Gender	Dichotomous variable that takes the value 1 when a woman serves as mayor and the value 0 when the mayor is male	Derived by the authors using data supplied by the Ministry of the Interior
Tax pressure	Tax revenues (Chapter 3 of the spending budget) divided by the number of inhabitants	Derived by the authors
Tax pressure-Preelectoral years	Interaction of tax pressure and preelectoral years	Derived by the authors
Mandatory tax 1 (Property)	Local tax on property (real estate)	General Secretariat for Local and Regional Coordination
Mandatory tax 2 (Economic activity)	Local tax on economic activity	General Secretariat for Local and Regional Coordination
Mandatory tax 3 (Vehicles)	Local tax on vehicles	General Secretariat for Local and Regional Coordination
Population	Ln of municipal inhabitants	National Institute of Statistics (INE)

Source: Compiled by the authors.

Table 2. Features of delivery forms of public services.

Delivery forms / Features	Private Sector participation	Risk transfer	Responsibility	Citizens’ perception
Contracting out	Total	Total	Low	Negative
PPP	Partial	Partial	Shared	Negative
Public externalization	None	None	Total	Positive

Source: Compiled by the authors.

Finally, following the arguments exposed early based on the Budgetary Political Cycles Theory (Veiga and Veiga 2007), it is necessary to capture any temporal lag between the moment in which the decision is made and its effects on the budged. To do so, we have incorporated a variable that interacts the capital expendidure with the poselectoral period. As the execution of this type of expense is dilated over time due to the fact that its concretion requires more time for the creation of investment and budgeting, as well as to reflect that the mayor will be interested in seeing its effects in periods closest to their reelection. In addition, we have also added to the model the interaction between current spending and the preelection moments. This is because it is where the public manager is presumed to increase this part of local public spending more, in order to guarantee their reelection.

Analysis

Discrete-choice models are appropriate when the aim of the study is to determine the probability of the explanatory factors of an individual economic agent choosing a particular course of action from a (usually finite) set of options. Such models have been used in many cases to analyze the factors underlying the choice of management forms for the delivery of public services (Dijkgraaf et al. 2003; Tavares and Camöes 2007; González-Gómez et al. 2011). Among the various options, the logit discrete-choice model is the most commonly used.

In the present study, the dependent variable is binary for each organizational form considered (contracting out, PPP and public externalization), and we have a data panel for a broad time horizon (2002–2014) (see Table A1 in the Appendix for the descriptive statistics). Therefore, in our analysis a fixed-effects conditional logistic regression is used for each of the three management alternatives within the panel data. The specification of fixed effects makes it possible to control all the individual characteristics that are constant over time, even when they are unobserved, which means that these characteristics can be correlated with the observed ones that vary in time, see Table 3 –Correlation matrix-.

Table 3. Correlations matrix.

	Elect.	Inter. 1	Inter. 2	Progr. Ideology	Inter. 3	Inter. 4	Political change	Inter. 5	Gender	Tax pressure	Inter. 6	Mandat. tax 1	Mandat. Tax 2	Mandat. tax 3	Populat.
Election	1
Inter. 1	−.502*	1
Inter. 2	−.363*	−.407*	1
Progr. Ideology	.003	.008	−.004	1
Inter. 3	.21*	−.007	−.02	−.012	1
Inter. 4	.006	−.01	−.002	.003	−.274*	1
Political change	.422	−.190*	−.175*	.020*	−.055*	−.077*	1
Inter. 5	−.058*	.114*	−.070*	.056*	−.063*	−.044*	.356*	1
Gender	.006	.013	−.017	−.002	−.051*	−.052*	.027*	.003	1
Tax pressure	−.031*	.005	.030*	.004	.009	−.031*	−.006	.005	−.009	1
Inter. 6	−.172*	−.289*	.477*	−.002	−.010	−.015	−.080*	−.033*	−.010*	.633*	1
Mandatory tax 1	−.088*	.132*	−.081*	.009	−.065*	.073	−.03*	.041*	−.006	.355*	.088*	1
Mandatory tax 2	−.024*	.022*	−.002	.009	−.041*	.023*	−.011	.003	−.008	.403*	.153*	.378*	1
Mandatory tax 3	−.025*	.003	.01	.006	−.053*	.042*	−.013	−.002	−.011	.426*	.171*	.315*	.547	1
Population	−.001	.035*	.024*	.000	.033*	−.055*	−.008	−.003	.024*	.016	.006	−.019*	.062*	.031*	1

*= p<.01.

Results and discussion

Table 4 presents the estimated coefficients, transformed to odds ratios, of the conditional fixed-effects logistic regression for each delivery form of public services used (Model 1- Contracting out; Model 2- Public-Private Partnership; Model 3- Public externalization).

Table 4. Results.

Variables	Model 1	Model 2	Model 3
	Contracting Out	Public-Private Partnership	Public Externalization
	Odds ratio^1	Odds ratio^1	Odds ratio^1
Election	.019***(.010)	.219**(.606)	8.801*(11.585)
Capital Spending-Postelect.	.776***(.029)	.922**(.161)	1.220**(.107)
Current Spending-Preelect.	.704***(.024)	.917***(.168)	1.181**(.098)
Progressive ideology	.978(.091)	.572(.246)	.895(0.188)
Absolute conservative majority	1.101*(.135)	1.122**(.109)	.897(.238)
Absolute progressive majority	.687***(.077)	1.950(1.154)	1.724**(.369)
Political change	.845(.134)	.374(.425)	.750(.332)
Political change-Postelect. years	.889(.296)	1.330(.002)	.666(.590)
Gender	1.361**(.170)	.990(.654)	.454**(.140)
Tax pressure	.994***(.001)	.994*(.003)	.999(.001)
Tax pressure-Preelectoral years	1.010(.001)	.999(.003)	.999(.001)
Mandatory tax 1 (property)	.430***(.130)	.639(.570)	1.622*(.460)
Mandatory tax 2 (economic activity)	1.728***(.295)	2.022(1.428)	1.236(.299)
Mandatory tax 3 (vehicles)	1.031***(.004)	1.018(.019)	1.006(.007)
Population	1.000***(.000)	.999(.000)	.999***(.000)
Test of Rho	.8437***	.5811***	.7131***
Hausman test	7.67 (.000)	1.40 (.000)	1.47 (.000)

***= p < 0.01; **= 0.01<p < 0.05; *= 0.05<p < 0.1 Standard errors are in parentheses.

The results obtained confirm the existence of asymmetric opportunistic behavior in relation to the management methods used, which is in line with previous research in this respect (Nordhaus 1975; Foucault et al. 2008). We tested our first hypothesis by examining the two stages of the electoral cycle, i.e. the value of the variable ‘Election’. In relation to the provision of public services with private sector participation, via the formulas considered (Model 1- Contracting out; Model 2- PPP), our results show that in preelection years these management forms are less commonly employed (Model 1- odds ratio 0.019, p = 0.000; Model 2- odds ratio 0.219, p = 0.011). In contrast, they are more commonly adopted in post-election years . These results support our first hypothesis and confirm the postulates of opportunistic behavior proposed by Garrone and Marzano (2015). Furthermore, the results indicate that, as suggested in our second hypothesis, in preelection years there is an increased use of public externalization (Model 3- odds ratio 8.801, p = 0.098). This initial analysis shows that local government decisions on management forms for the delivery of public services are subject to opportunistic behavior by policymakers.

Various political factors – the presence of an absolute majority, the ideology of the party in power and the occurrence of political changes – were included in the study model to determine their relation with opportunistic behavior by policymakers. Confirming our third hypothesis, analysis shows that political decisions on management forms are taken in line with partisan considerations. According to the study results, when the local authority is governed by a conservative party with an absolute majority, policymakers are more likely to adopt a management form for public services related to the private sphere, such as contracting out or PPP. However, when a progressive government has an absolute majority, there is a lower probability of the services being contracted out (odds ratio 0.687; p = 0.001), and a greater likelihood of public externalization taking place (odds ratio 1.724; p = 0.011). These findings confirm the presence of asymmetric behavior when changes are made to the management form for the provision of local public services, according to the ideology of the governing party when it has an absolute majority.

Another group of variables that may influence the management form decision and its relation or otherwise with opportunistic behavior is that of public spending. It has been hypothesized that public managers are likely to adopt expansive economic policies, based on increased public investment, in order to gain voters’ favor (Foucault et al. 2008). Again, our study results reveal asymmetric behavior in relation to the type of spending, the political cycle and the forms of service provision. Thus, capital spending is inversely related to the probability of services being contracted out (Model 1) or provided via PPP (Model 2) in the post-election period, with a stronger impact in the first of these cases (odds ratio 0.776; p = 0.000) than in the second (odds ratio 0.922; p = 0.02). On the other hand, capital spending is directly related to public externalization (Model 3) in post-electoral years.

These results suggest that politicians behave opportunistically in post-election years, by reducing capital spending and by adopting contracting-out and/or PPP options to incorporate private operators into the management of public services. On the other hand, when capital spending increases, services are more likely to be externalized to public agencies. Overall, we conclude that opportunistic behavior exists in politicians’ decisions as to which organizational form should be used to deliver a public service.

The relationship between current spending and service delivery forms is also subject to opportunistic behavior by policymakers. Increased current spending in the preelection period is associated with a decrease in the number of contracting out operations and PPP agreements performed, reflecting the fact that public managers prefer to take electorally popular measures during this period (Veiga and Veiga 2007), whilst avoiding less popular ones such as contracting out and PPP. This inverse relation is particularly strong in the first of these cases. These results are in line with those reported by Fernández et al. (2008), Garrone and Marzano (2015) and de la Higuera-Molina et al. (2019), who all observed that contracting out is unpopular with the population in general and (during this period) with politicians in particular. The results for public externalization indicate that an increase in current spending in the preelection period is associated with greater use of this management form (Model 3- odds ratio 1.181; p = 0.045).

Other political variables, such as the election of a new mayor, have no influence on any of the management forms analyzed, with respect to interaction with the variable “Political cycle (Post-electoral period)”. Nor does the ideology of the governing party present any significant influence, although this influence may be partly determined by its inclusion within the variable “Absolute majority”.

According to our results, the mayor’s gender does influence policymakers’ decisions, which is in line with previous research in this area (Fox and Schuhmann 1999). Oposite to what we expected, our results show that when a female mayor is elected, there is a greater likelihood of public services being contracted out (Model 1) and a lower probability of public externalization (Model 3). While we initially had linked this decision with the literature considering the relation between gander and risk aversion, it should also be mentioned that there is some empirical evidence suggesting that female managers tend to manage in a more collaborative way. When comparing male and female managers’ decision making, it seems that females are more willing to involve stakeholders in the process (Fox and Schuhmann 1999). Meier, O’Toole, and Goerdel (2006) argue that females manage organizations in a more flexible and participatory way, whereas male managerial styles tend to be more hierarchical and rigid. Therefore, this could explain why female managers seem to have a stronger tendency to engage with collaborative ventures to implement public services when compared to their male counterparts.

We apply two strategies in our analysis of taxation: first, we examine the relationship between fiscal pressure and the stage of the electoral cycle, considering the argument proposed in various theoretical frameworks that prior to an election the fiscal burden tends to be reduced. We then analyze the impact of the main taxes imposed, according to their nature. The results obtained from these analyses show that the stage of the electoral cycle has no significant influence on changes in management forms, and hence there is no opportunistic behavior in this respect. However, we also show that the most important tax revenue obtained by local corporations in Spain is from the property tax (Mandatory tax 1), and that this revenue is inversely associated with contracting out decisions (Model 1- odds ratio 0.430; p = 0.005). Thus, when the property tax is reduced, more contracting out decisions are taken. In contrast, when the property tax increases, public externalization is more likely to be adopted (Model 3- odds ratio 1.622; p = 0.088); when this tax is raised, more public agencies are created. This relationship highlights the existence of opportunistic behavior by public managers with respect to the main tax item available to local governments. The second and third taxes considered (economic activity and vehicles) present a direct relationship with contracting out, indicating that with increasing economic activity and therefore higher tax revenues in this respect, there is a greater likelihood of local public services being contracted out. However, these two taxes represent a relatively low proportion of the local authority’s total income.

Finally, the size of the local population is significantly associated with contracting out and public externalization (in both cases the odds ratio is practically equal to one). However, for the creation of public-private partnerships, the odds ratio is not statistically significant.

Conclusions

In the field of public management, many studies have addressed the question of why public managers implement public services via certain organizational arrangements, such as contracting out, PPPs or public externalization. Various theoretical approaches have been proposed to model the behavior of public managers. One such is that of opportunistic behavior, according to which policymakers’ decisions may be influenced by electoral considerations. In the present study, we examine whether public managers make their decisions motivated (at least in part) by the wish to gain political credit, and not simply by what is best for the community. In our analysis, we test whether opportunistic behavior, arising from electoral considerations, is a significant factor in policymakers’ decisions regarding public service delivery. Specifically, our research aim is to determine whether changes in the management form of public services, in terms of transferring risk and partnership from the public to the private sector, via contracting out, PPP or public externalization, reflect opportunistic behavior by policymakers. Our main argument is then when deciding which organisaitonal form should be used to implement a public policy, policymakers seek to obtain political benefits by orienting their decisions according to the stage of the electoral cycle.

The theoretical model we describe provides a new understanding of asymmetric opportunistic behavior in political decision making, according to the degree of private sector participation and risk exposure in the provision of public services. In accordance with the theory of opportunistic political cycles, we believe it essential to take into account the current stage of the electoral cycle in order to understand the motivations underlying public managers’ decisions on the form of management to be employed in service provision. Thus, during the preelection period, public managers will seek to improve their electoral prospects by making strategic decisions that, in their opinion, will be aligned with the priorities of their voters. However, once the elections have concluded, the behavior of the public manager changes and the strategic decision as to which organizational form should be used to implement a public service is no longer influenced by the perception of voters’ wishes.

In this study, we obtain empirical evidence in support of the theoretical model presented, showing that the choice of management form is influenced by the current phase of the electoral cycle. Thus, in the preelectoral period, the organizational forms by which the private sector participates in service delivery (contracting out and PPP) are less commonly used, probably because policymakers believe there exist unfavorable perceptions among voters regarding the involvement of the private sector in the delivery of public services (Fernández et al. 2008; Garrone and Marzano 2015).

According to the results obtained in our study, the service management form based on a high level of public participation and no transfer of management risks (i.e., Model 3- Public externalization) is especially likely to be adopted in preelectoral years. This pattern, too, would reflect the existence of opportunistic behavior in public sector decision making. In contrast, contracting out and PPP are much more prominent in the post-electoral period, when the municipal government’s activities are less strongly oriented toward swaying voters’ opinions.

Although these results corroborate the theory of opportunistic political cycles regarding the choice of organizational form, another two theoretical arguments should also be taken into consideration, to better understand the relationships addressed, namely the Partisan model and the Budgetary Cycles theory. According to the Partisan Model, the choice of organizational form is also influenced by political ideology, which can moderate the effect of the political cycle discussed above. Our results show that the political ideology of the governing party is an essential element in the relationship between the political cycle and the decision to allow the private sector to participate in the delivery of a public service. Thus, progressive parties strongly favor service delivery exclusively by the public sector, whereas conservative ones prefer to allow the private sector to participate, and hence are much more likely to contract out municipal services or to create a PPP. Thus, the Partisan Political theory is supported by our results, and complements the theory of Opportunistic Political Cycles.

The opportunistic behavior of public managers is also reflected in public spending policies, as argued by the theory of budgetary political cycles. Our empirical analysis reveals that in preelectoral years, when public managers often use current spending to raise the visibility of their actions and thus enhance voters’ perceptions of their political effectiveness, increased public spending is related to decreased private sector participation in the delivery of public services. These results are in line with those obtained in previous research into the relationship between public spending and contracting out (de la Higuera-Molina et al. 2019), thus underpinning the robustness of theoretical models proposing the existence of opportunistic behavior by public managers.

Arguably, opportunistic behavior arises from policymakers’ belief that voters, on the whole, do not agree with the participation of the private sector in the delivery of public services. This study offers empirical support for the latter argument, and shows that politicians’ interests affect certain decisions on the provision of public services. It should be noted that the present study was carried in Spain, where the role of public manager is highly politicized, particularly at the municipal level. Further study is necessary to determine whether our results are equally applicable to other countries, for example those with an Anglo-Saxon administrative culture, where public managers’ decisions are not as strongly influenced by the political orientation of the mayor.

To conclude, the results presented in this article offer practical information to public and private sector managers involved in the delivery of public services. There is a widespread view among policymakers that the price of engagement with the private sector in the delivery of public services is a loss of support in subsequent elections. Further research is needed to establish whether this is actually the case or whether the perception is misplaced. In addition, we show that private sector service providers need to make more effort to convince citizens of the benefits of their involvement in the delivery of public services. Finally, our results provide a very useful basis for further analyses of service provision in the public sector, raising the question of what other effects might be produced by opportunistic behavior in this field. For example, studies are needed to examine in detail the behavioral mechanisms that lead policymakers to engage in opportunistic behavior, and to analyze the factors that might reduce its presence in the public sector.

Additional information

Funding

AGAUR (Agency for Management of University and Research Grants), Grant/Award Number: 2017-SGR-1556; Ministerio de Economía, Industria y Competitividad, Gobierno de España, Grant/Award Number: CSO2016-80823-P.

Notes on contributors

Emilio J. de la Higuera-Molina

Emilio J. de la Higuera-Molina ([email protected]) is an Assistant Professor Doctor of the University of Granada with diverse publications in journals indexed in JCR. He did his PhD at the University of Granada.

Marc Esteve

Marc Esteve ([email protected]) is an Associate Professor in International Public Management in the School of Public Policy at University College London, and a Visiting Professor at ESADE-Ramon Llull University. His primary research interests focus on understanding how individual characteristics influence decision making, specifically in collaborations. He has published widely in journals such as Journal of Public Administration Research and Theory, Public Administration Review, International Public Management Journal, and Public Administration.

Ana M. Plata-Díaz

Ana M. Plata-Díaz ([email protected]) is an Assistant Professor Doctor of the University of Granada with diverse publications in journals indexed in JCR with high index of impact in the area of public administration and finance.

José L. Zafra-Gómez

José L. Zafra-Gómez ([email protected]) is a Professor of the University of Granada. He has published widely in journals such as Omega, Journal of Public Administration Research and Theory, Local Government Studies, International Public Management Journal, and Public Administration.

Notes

1 For a more extensive discussion of the concept of contracting out, see López-Hernández et al., (2018), and Table 2.

2 For a more extensive discussion of the concept of PPP, see Warner and Bel (2008), and Table 2.

3 For a more extensive discussion of the concept of public externalisation, see Warner and Bel (2008), and Table 2.

4 In the Spanish context, these type of spending includes real investments (the creation of new equipment and infrastructures, and the acquisition of assets (property, plant and equipment) and capital transfers (amounts received, usually by other administrations, to finance capital operations and real investments.

5 In the Spanish context, these are the expenses used for current operations, which include the personnel expenses, the current expenses in goods and services, and the current transfers received primarily from other administrations, among other concepts.