Modeling Fiscal Stress and Contracting Out in Local Government: The Influence of Time,Financial Condition,and the Great Recession

Abstract

Various studies have analyzed the relationship between fiscal stress and contracting out, but have failed to achieve conclusive results. In this article, we take a broad view of fiscal stress, addressed in terms of financial condition and studied over a lengthy period (2000-2010). The relationship between fiscal stress and contracting out is studied using a dynamic model, based on survival analysis, a methodology that enables us to take into account the effect of time on this relationship. As this study period includes the years of the Great Recession (2008-2010), we also highlight the impact of this event on the fiscal stress–contracting out relation. The results obtained suggest that taking into account the passage of time and conducting a long-term assessment of financial condition enable a more precise understanding of this relation. We also find that the Great Recession reduced the probability of local governments’ contracting out public services.

Keywords

local government fiscal stress contracting out Great Recession survival analysis

Introduction

When local authorities are in poor financial health, local managers must seek appropriate solutions. Fiscal stress may be addressed in various ways, for example, by raising taxes (Bel & Fageda, 2007), by reducing or eliminating public services, or even by making changes to the organizational structure of the local authority (Christoffersen & Bo Larsen, 2007; Zafra-Gómez & Muniz, 2010). However, when such situations must be faced and the municipal authorities make the political decision not to implement any of the above measures, the fiscal stress experienced by a public entity can also be overcome through the implementation of market mechanisms, such as the contracting out of public services (Osborne & Gaebler, 1992), which would reduce the cost of providing public services and thus alleviate fiscal stress.

Adopting the latter measure as a means of overcoming an unfavorable financial situation is supported by the public choice theory, which favors the use of market mechanisms to improve public sector efficiency (Niskanen, 1971; Zafra-Gómez, Plata-Díaz, Pérez-López, & López-Hernández, 2016b). Moreover, the application of private forms of management would raise the quality of service provision, according to the theory of property rights (Alchian, 1967), while achieving greater flexibility in resource management, and hence increased efficiency in the provision of services, as postulated by organizational theory (Earle, 2006).

The decision to contract out municipal services may also be influenced by political or socioeconomic factors, the level of information available to local public managers, the type of service (Brown & Potoski, 2003; Rodrigues, Tavares, & Araujo, 2012), and the characteristics of the service provider (Bel & Fageda, 2011). In addition, before contracting out is undertaken, the transaction costs of the operation should be determined; if these exceeded the cost savings achieved, fiscal stress would increase rather than decrease (Bel & Fageda, 2010). Similarly, conditions like asset specificity or the difficulty of performance monitoring are of crucial importance in determining whether a local service can be successfully contracted out, because if they are unfavorable the transaction costs may be prohibitive (Bel & Fageda, 2010; Brown & Potoski, 2003).

Most of the above factors, therefore, may affect the economic and financial situation of a local authority, and should be taken into account in considering how to respond to fiscal stress. However, the empirical evidence is unclear as to whether situations of fiscal stress are responsible for provoking changes in the management of public services¹ (Bel & Fageda, 2007; Levin & Tadelis, 2010). Moreover, previous studies of this question included variables that measured only a partial aspect of fiscal stress and did not truly reflect the financial health of local entities (Boyne, 1998). This was also observed by DiNapoli (2013), who concluded that fiscal stress cannot be measured by one single indicator, as some factors will have greater relevance for certain classes of local government and many indicators measure different aspects of the same problem.

Another important consideration is that most previous studies have focused on cross-sectional estimation, an approach that provokes what is termed economic myopia, which fails to analyze a phenomenon beyond a short-term horizon. This condition undoubtedly affects the conclusions drawn regarding the fiscal stress–contracting out relation, as a contracting out decision is very unlikely to be adopted taking into account just a single year of fiscal stress.

In this article, and unlike most previous research in this field (Bel & Fageda, 2007), we aim to measure fiscal stress by using a set of indicators—taxable value by debt, the importance of transfers from central to local government, budgetary sustainability, and short-term cash solvency (Zafra-Gómez, López-Hernández, & Hernández-Bastida, 2009a)—to measure the intensity with which each element affects the contracting out process. Furthermore, and for the first time, we incorporate the passage-of-time effect in analyzing decisions regarding the contracting out of local public services—decisions that are normally taken in response to the continuing impact, over several years, of fiscal stress (González-Gómez & Guardiola, 2009). This approach enables us to measure each of the components of financial condition for a broad time horizon—in other words, we consider the local authority’s long-run solvency—and so there is no need to evaluate the contracting out decision on the basis of the situation at a given moment.

In this study, therefore, we examined the contracting out processes used by 1,563 Spanish local authorities² during the period 2000-2010, using survival analysis. As we wished to study the long-term relationship between fiscal stress and contracting out, we examined the effect of the Great Recession (GR) on the latter decision and also that of the above-mentioned financial condition variables, before and during GR (2000-2007 and 2008-2010, respectively).

The results obtained show that some indicators of financial condition may be sufficient in themselves to justify contracting out, if we take into account the effect of time (short-run solvency and financial independence). However, not all indicators are individually capable of influencing the contracting out decision; in some cases, there must be a worsening of several indicators at once. In addition, the inclusion of GR in the analysis decreases the likelihood of public services being contracted out, while the effect of financial condition during GR remained unchanged from the pre-GR period (2000-2007), except for the cash surplus index. Furthermore, the more complex the service and the more specialized the assets required, the greater the probability of its being contracted out.

Does a Local Authority Resort to Contracting Out in Response to Fiscal Stress? Theoretical Framework

The main idea underlying public choice theory is that public entities should operate and be structured in the same way as private organizations (Savas, 1987), because this is how services will be provided most efficiently (Niskanen, 1971). To achieve this goal, public choice theory seeks, among other aspects, to improve public sector management by applying formulas derived from the private sector, changing the organizational structure and how services are provided (Hood, 1991, 1995; Laughlin & Pallot, 1998). The aim of these changes is to achieve a market-oriented outlook, and thus increase efficiency, effectiveness, and productivity. To do so, one of the most commonly used formulas is for a private operator to provide the services formerly supplied by the public operator, that is, through contracting out (Carboni, 2015).

Contracting out can be defined as a form of privatization in which a private company obtains residual gains from the service delivery (Vickers & Yarrow, 1991; Warner & Bel, 2008) and is the most important of the alternatives available in which services are provided by operators other than the public entity itself (Pallesen, 2004; Plata-Díaz, Zafra-Gómez, Pérez-López, & López-Hernández, 2014). Several theoretical frameworks have been defined to model the contracting out decision. When public services are contracted out, this can produce various benefits typically associated with private organizations, such as increased service quality, following new investment to improve service provision and, at the same time, cost savings, in accordance with the theory of property rights (Alchian, 1967). Furthermore, the welfare of the public administration itself may benefit from the introduction of private mechanisms, as a private entity may be more aware of and more familiar with appropriate systems to supervise and motivate its employees (according to organization theory), because the success of the company depends on this factor, among others, in a competitive market (Hatch & Cunliffe, 2013). Furthermore, contracting out also provides greater flexibility in the management of resources, thanks to the greater efficiency achieved in service provision (Earle, 2006; Meyer & Scott, 1983; Zucker, 1987). To achieve this enhanced efficiency, organization theory holds that the municipal government, as an organization (Levitt & March, 1988; North, 1990; Williamson, 1996), should seek to improve decision making by applying the above-mentioned systems of supervision and motivation. Organization theory can also be approached from other standpoints; thus, Rubin (2005) believes that organizations do not depend on themselves alone, but can be influenced by external social or institutional forces. They can also be viewed as the intersection of various contracts, or as complex systems or organisms; alternatively, they can be regarded as complex, decision-making hierarchies. Nevertheless, as Rubin (2005) states, the environment in which a global economic and financial crisis takes place makes it impossible to predict the behavior of local public managers, due to the complexity of this situation in the municipalities. A good example of this is the decision-making process concerning which municipal public services should be provided in times of crisis. Nevertheless, when local managers decide to contract out municipal services, they may also be influenced by political factors such as ideology (Pérez-López, Prior, & Zafra-Gómez, 2015; Pérez-López, Prior, Zafra-Gómez, & Plata-Díaz, 2016; Pérez-López, Prior, Zafra-Gómez, 2017). However, according to Dubin and Navarro (1988), Dijkgraaf, Gradus, and Melenberg (2003) and Zafra-Gómez, Plata-Díaz, Pérez-López, and López-Hernández (2016b), this factor does not significantly influence the provision of public services by local authorities.

The final outcome of the contracting out process may also be affected by the level of information available to public managers and the private companies involved regarding the costs of the operation (Brown & Potoski, 2003). This is because the municipality sometimes has only incomplete and asymmetric information about the contract (Bel & Fageda, 2006), and this can generate increased transaction costs and hence less cost saving than expected, together with reduced efficiency in the external provision of services (Máñez, Pérez-López, Prior, & Zafra-Gómez, 2016; Nogueira & Jorge, 2016).

Another factor that should be taken into consideration is the availability of service providers. In this respect, Bel and Fageda (2011) analyzed suppliers of contracted out services, according to the size of the municipality, and found that large municipalities attract more suppliers, due to economies of scale (Plata-Díaz, Zafra-Gómez, Pérez-López, & López-Hernández, 2014). Thus, the contracting out decision is influenced by a wide variety of factors, many of which may increase service provision costs and compromise local finances.

For all these reasons, we agree with the postulates of public choice theory, according to which the main reason underlying the contracting out of public services is the presence of financial stress in local government. Similar conclusions were reached by Bel and Fageda (2007), who observed that one of the hypotheses most commonly analyzed and empirically tested is that of the relationship between fiscal stress (tax burden, legal limitations on local tax levels, and the size of transfers from central to local government) and contracting out by local government (Ferris, 1986; Kodrzycki, 1998; Miranda, 1994). Thus, various studies have concluded that local officials respond to fiscal problems by this means (International City/County Management Association [ICMA], 1989; Morgan & Hirlinger, 1991; Mouritzen & Nielsen, 1988; Touche-Ross Company, 1987). In short, as remarked by Zullo (2009, p. 461), “Empirically, measures of fiscal stress, or any institutional constraint on local government to tax and spend, should positively correlate with private contracting.” However, empirical evidence is unclear as to whether situations of fiscal stress are responsible for provoking changes in management attitudes (Bel & Fageda, 2007; Zullo, 2009). This doubt has been expressed, especially, in studies conducted in the United States. As observed by Bel and Fageda (2007, p. 525), “ . . . in more recent studies in the US, those where data collection occurred from 1992 to 2004, fiscal stress appears as a significant explanatory factor only in the works of Kodrzycki (1998) and Hebdon and Jalette (2008),”³ whereas other studies, such as those by Levin and Tadelis (2010) and Zullo (2009), have failed to find any relation between fiscal stress and contracting out. This apparent nonrelation may be due to one or more of several factors.

First, most studies are cross-sectional, despite the need to evaluate municipal performance over an extended period of time (Bel, Fageda, & Warner, 2010). Second, previous measures used to determine the presence of fiscal stress were very imprecise and may not accurately reflect the true financial situation of the municipality (Boyne, 1998). In this respect, Boyne (1998, p. 152) observed that overall, “the evidence provides little support for the view that fiscal stress is a significant constraint on decisions to contract out,” attributing the lack of support for this hypothesis to various factors, including poor measures of the revenue shortfall aspect of fiscal stress, a disregard for measures of fiscal need, or the growing demand for services. In summary, the results obtained from prior research encourage us to present a new proposal to measure the fiscal stress–contracting out relation.

Proposed Model of the Fiscal Stress–Contracting Out Relation

As discussed above, we seek to establish a methodology to characterize the relationship between fiscal stress and contracting out. To do so, we must first obtain an adequate measure of fiscal stress (Boyne, 1998). In this respect, according to Hendrick (2011, p. 22), “Fiscal stress can be defined as a worsening of financial condition.” In previous empirical work, researchers have measured fiscal stress exclusively in terms of discretionary tax increases and of reduction in expenditure. However, these are only two areas of financial condition, and other aspects that may characterize fiscal stress may be omitted. If this were so, their absence from consideration might lead to the question of fiscal stress, and hence the fiscal stress–contracting out relation, not being properly evaluated. This is why many studies have concluded that this relationship is not significant.

Financial condition is the state of equilibrium or balance that exists between different dimensions or components of the financial health of an entity (Zafra-Gómez, López-Hernández, & Hernández-Bastida, 2009a; Hendrick, 2011). It can be grouped into two main elements: the environmental economic conditions and the internal conditions of local finance (Groves & Godsey, 2003). The environmental factor affects both the creation of demand and the provision of public services. According to Berne and Schramm (1986), to correctly evaluate financial condition, a series of contextual magnitudes related to society’s needs and preferences must be analyzed; these local conditions affect the provision of public services, the cost of productive factors, the resources available to society and the public policies that have an effect on the local authority. But as well as these environmental issues, there is the most important factor of all those comprising financial condition with respect to the internal finances of the local entity, namely, the need to balance revenues collected and the funds required for municipal spending (Jordan, Yan, & Hooshmand, 2015).

We believe the following aspects of local internal finance should be considered: cash surplus solvency, defined as the organization’s ability to generate sufficient liquidity to pay its short-term debts (Groves & Godsey, 2003; Abad-González & Gutierrez-López, 2016); flexibility, or its capability to respond to changes in the economy or in its financial circumstances, through modifications to public debt (Canadian Institute of Chartered Accountants [CICA], 1997; Greenberg & Hillier, 1995); budgetary sustainability (or service-level solvency), that is, the organization’s ability to maintain, promote, and protect the social welfare of the population, using the resources at its disposal (CICA, 1997; Greenberg & Hillier, 1995; Groves & Godsey, 2003); and financial independence, or the level of dependence on external funding received through transfers and grants (Honadle, 2003; ICMA, 2003; Zafra-Gómez, López-Hernández, & Hernández-Bastida, 2009a).

Finally, we consider long-run solvency, using a broad time horizon to measure each of the above elements (cash solvency, flexibility, budgetary sustainability, and financial independence). Thus, we are able to record the cumulative effect over several years of negative positions of each of the elements of financial condition. This is important because a public manager is more likely to decide to contract out public services when a position of fiscal stress persists over time, and not when it is limited to a particular year. Furthermore, we can specifically take into account the influence of the passage of time on such a decision, through the variable “duration dependence” (Guardiola, González-Gómez, & García-Rubio, 2010). This approach to the analysis of the fiscal stress–contracting out relation enhances the precision of the judgment reached.

In view of these considerations, we propose the following hypotheses:

Hypothesis 1A (H1A): The passage of time (long-run solvency) influences the probability of public services being contracted out.

Hypothesis 1B (H1B): The contracting out–fiscal stress relationship can be measured by means of any element of financial condition (short-run solvency, flexibility, sustainability, and independence), but taking into account the passage of time (long-run solvency).

Hypothesis 1C (H1C): The contracting out–fiscal stress relationship can be defined by combining two or more elements of financial condition (short-run solvency, flexibility, sustainability, and independence), taking into account the passage of time (long-run solvency).

However, a broader debate is currently being held on the suitability of applying market mechanisms to municipal services. In this respect, the social choice theory, for example, postulates that to improve the public sector, decisions must be taken via a dialogue between the population and the market (Hefetz & Warner, 2007). In other words, there should be a balanced position, in which both the benefits provided by the market and the commitment of the populace are recognized (Sager, 2002). For this reason, the hypotheses we propose are intended to verify whether this new approach is being put into practice by local administrations.

To measure the environmental factor, we included several socioeconomic indicators that may influence the contracting out of public services. One such indicator is the rate of unemployment in the municipality. According to Zullo (2009), a high rate of local unemployment reflects a weak regional economy, as it reduces governments’ ability to generate tax revenue and finance public services (Balaguer-Coll, Brun-Martos, Forte, & Tortosa-Ausina, 2015). Moreover, it is an indicator of stress, and therefore should be positively associated with contracting out. The index of economic activity is another factor that can influence contracting out, because a reduction in economic activity can provoke a corresponding reduction in tax revenues and thus increase the likelihood of public services being contracted out (Bastida, Benito, & Guillamón, 2009; Easterly & Rebelo, 1993). Finally, we included the variable “population” because it can be considered a proxy for the demand for local services. Thus, a rising population, requiring increased municipal services, may increase the likelihood of contracting out (Bel et al., 2010).

The Effect of the GR on Fiscal Stress and Contracting Out

If any event of recent years has had a truly international impact, that event is the global financial crisis, or GR. Therefore, in analyzing any recent phenomenon of an economic nature, we must separate or measure the effect of GR on this phenomenon (Moss, 2010). Both public and private entities need to be aware of the organizational disruption caused by GR and realize that the appropriate response may be different from that corresponding to a “traditional” problem or an internal one (James & Wooten, 2010). Therefore, it is essential to distinguish fiscal stress, in the sense of financial problems that are private or restricted to the municipal context, from the concept of “transboundary crisis” (Ansell, Boin, & Keller, 2010), which has a global impact and is capable of “infecting” many countries and affecting all levels of government, including local entities. In this respect, Nelson and Baku (2014) observed that since the start of GR, fiscal shortfalls have resulted in reduced public services in 46 U.S. states, with a severe impact on the services provided at the local government level (Johnson, Oliff, & Williams, 2011).

In this context, we consider it necessary to determine how GR has affected the contracting out of local authority services, as the fiscal stress–contracting out relation may well have changed in recent years due to the impact of this crisis. Most previous studies have hypothesized that public managers use contracting out to cut costs and to relieve fiscal stress. If we accept this, then in view of GR and the ensuing pressure to reduce budget deficits, at all levels of government, local authorities would be expected to vigorously promote contracting out. This was the view of Funkhouser (2012), who noted that the impact of the recession had clearly diminished municipal revenues and spending, and thus “the pressure to outsource services is greater than it has ever been” (p. 1). Accordingly, we propose the following hypothesis:

Hypothesis 2 (H2): GR increased the likelihood of public services being contracted out.

If GR may affect the contracting out decision, it may also affect the influence of fiscal stress on contracting out, through the elements that comprise financial condition. Accordingly, during GR, when economic activity declined, the indicators of financial condition would have worsened, and so their influence on contracting out would have varied. In consequence, we propose the following hypothesis:

Hypothesis 3 (H3): The influence of indicators of fiscal stress on contracting out decisions was greater during GR (2008-2010) than during the pre-GR period (2000-2007).

Service Measurability and Asset Specificity

In addition to the above, the type or nature of service involved is also relevant to the contracting out process (Ferris & Graddy, 1986, 1991), as the degree of asset specificity, that is, the quantity of specialized investments—which would be very difficult to adapt to the provision of other services (Brown & Potoski, 2003), for example, in areas such as water distribution and sewage collection and treatment—and the capacity to measure the outcomes achieved, thus enabling managers to evaluate the quality of the services provided, or the capability to monitor the performance of the provider (i.e., service measurability) are of decisive importance in determining the form in which local public services will be provided (Brown & Potoski, 2003, 2005). The greater the service complexity (as regards measuring outcomes or monitoring activities) or asset specificity (i.e., the need for specialized investment; see Table 4), the higher the transaction costs (Bel & Fageda, 2008; Brown & Potoski, 2003; Brown, Potoski, & Van Slyke, 2008; Levin & Tadelis, 2010; Rodrigues, Tavares, & Araujo, 2012; Walls, Macauley, & Anderson, 2005), and therefore the lower the likelihood of the service being contracted out. In other words, local governments will more frequently contract out services associated with low transaction costs, that is, with low asset specificity and whose performance is easy to measure (Brown, Potoski, & Van Slyke, 2006; Levin & Tadelis, 2005; Ménard & Saussier, 2000; Walls, Macauley, & Anderson, 2005). In accordance with these considerations, the following hypotheses are proposed:

Hypothesis 4 (H4): Public services which are more difficult to measure are less likely to be contracted out.

Hypothesis 5 (H5): Public services with a high degree of asset specificity are less likely to be contracted out.

The Fiscal Stress–Contracting Out Relation in Spanish Local Governments

The Case of Spain

The 1970s and 1980s were marked by growing frustration with the public production of services, and by a corresponding expansion of private production of public services, in Europe and the United States (Warner & Bel, 2008). For example, in Sweden, Norway, Finland, Denmark, and Spain 60% to 80% of municipalities privatized their refuse collection service during this period. Water distribution services also presented a high rate of privatization in the United Kingdom, France, and Spain (Bel, 2006; Organisation for Economic Co-Operation and Development, 2000a, 2000b). Subsequently, in the 2000s, the number of municipalities contacting out public services increased significantly; thus, in Portugal, around 100 such operations were conducted in 2005 alone (Tavares & Camöes, 2007).

The search for new forms of managing public services was accentuated by the global economic and financial crisis that broke out in 2008. In Spain, the response to GR included measures such as budget restrictions on budgets and on borrowing (Zafra-Gómez, López-Hernández, & Hernández-Bastida, 2009a; 2009b; 2009c). These limitations, together with falling municipal revenues and increased costs, due to increased demand for public services (Cabeza, 2009; Paulais, 2009; United Cities and Local Governments, 2009), in many cases severely affected municipal finances.

In Europe, various empirical analyses have been made of local finances in the context of GR (Lungová, 2010; Paulais, 2009; Peters, Pierre, & Randma-Liiv 2011; Steytler & Powell, 2010). In Spain, in particular, López-Hernández, Zafra-Gómez, and Ortíz-Rodríguez, (2012) confirmed that GR had a negative impact on the financial condition of Spanish municipalities, reducing revenues and prompting a search for cost savings. As a result, many municipal managers opted for contracting out. Figure 1 shows the evolution of various indicators of financial condition and of levels of contracting out.

Figure 1.

Annual relation between indicators of financial condition (Cash Surplus Index, Financial Independence, Taxable Value divided by Financial Charge Index “TVFCI” and Non Financial Budgetary Result Index “NFBRI”) and contracting out.

Data

Our empirical analysis is based on annual data obtained from various databases, for all Spanish small- and medium-sized municipalities with more than 1,000 inhabitants, a total of 3,245 municipalities.⁴ The compilation of these data allowed us to include observations of the dependent variable over an extended period,⁵ and thus time-varying covariates are incorporated in our estimations. After a filtering process required by the information heterogeneity of the databases examined,⁶ the sample was finally reduced to 1,563 municipalities for the period 2000-2010. Of the sample analyzed, 488 (31.2%) municipalities contracted out local services during the study period and 1,075 municipalities did not. These municipalities presented a total of 1,302 operations to contract out services for the period 2000-2010, leading to 21,608 observations. These variables are summarized in Table 1, the descriptive statistics are shown in Table 2, and the correlation matrix is shown in Table 3. Spanish municipalities constitute a study area of particular importance due to the very large volume of data, corresponding to an extended time period, made available by the Ministry of Finance.

Table 1.

Description of the Independent Variables.

Variable			Description	Source
Fiscal Stress^a	Short-run solvency	Cash Surplus Index	Difference between net short-term receivables, liquidity, and net short-term liabilities	Directorate General for Financial Coordination with Regional and Local Authorities (DGCFCAEL, Ministry of Finance and Public Administration)
	Flexibility	Taxable value divided by Financial Charge Index	Fiscal receivables divided by annual amortization payment—interest and principal
	Sustainability	Nonfinancial Budgetary Result Index	Nonfinancial current budgetary receivables and nonfinancial capital budgetary receivables divided by current budgetary payables and nonfinancial capital budgetary payables
	Independence	Financial Independence Index	Budgetary payables divided by budgetary receivables except grants
Unemployment			Ratio of registered unemployed adults/total population aged 15 to 64 years	Spanish Economic and Social Yearbook (Anuario Económico y Social de España): La Caixa
Population –log-			Logarithm of the number of inhabitants	National Institute of Statistics
Index of Economic Activity			Index of all economic activity. Obtained from the tax imposed on all economic and professional activities (industrial, commercial, and services)	Spanish Economic and Social Yearbook (Anuario Económico y Social de España): La Caixa
GR			Dummy variable that takes the value 1 for the years of the GR (2008-2010) and 0 for the other years of the period considered (2000-2007)	The authors
Pre-GR			Dummy variable that takes the value 1 for the years of the period (2000-2007) and 0 for the years of the GR (2008-2010)	The authors
Service measurability			Dummy variable that takes the value 1 for services that are more difficult to monitor and measure and 0 otherwise	The authors
Asset specificity			Dummy variable that takes the value 1 for services whose delivery requires specialized asset investment and 0 otherwise	The authors

Note. DGCFCAEL = Directorate General for Financial Coordination with Regional and Local Authorities; GR = Great Recession.

Model 6 incorporates an interaction effect between the variables fiscal stress and GR, to distinguish the value of each element of financial condition in the two study periods (pre-GR and GR).

Table 2.

Descriptive Statistics.

Variable	M	SD	Minimum	Maximum
Contracting out (dependent variable)	0.0603	0.238	0	1
Cash Surplus Index	3.315	48.296	−192.811	6,564.841
Taxable value divided by Financial Charge Index	29.235	522.165	0.206	51,853.79
Nonfinancial Budgetary Result index	1.029	0.171	0.310	3.991
Financial Independence Index	0.603	0.209	0.070	2.765
Unemployment	4.835	2.882	0.400	28.8
Population-log-	9.407	1.723	6.908	15.001
Index of Economic Activity	2.613	11.997	0	120.2
Great Recession	0.199	0.400	0	1
Service measurability	0.388	0.487	0	1
Asset specificity	0.592	0.491	0	1

Note. n = 21,608 observations.

Table 3.

Correlation Matrix.

	Cash Surplus Index	Taxable value divided by Financial Charge Index	Nonfinancial Budgetary Result Index	Financial Independence Index	Unemployment	Population-log-	Index of economic activity	Great recession	Service measurability	Asset specificity
Cash Surplus Index	1
Taxable value divided by Financial Charge Index	−.0003	1
Nonfinancial Budgetary Result Index	.0079	−.0003	1
Financial Independence Index	.0129*	−.0083	.4403***	1
Unemployment	−.0152**	−.0071	−.0282***	−.2517***	1
Population-log-	−.0089	−.0257***	.0039	.3031***	.0567***	1
Index of Economic Activity	−.006	−.0085	−.0239***	.0684***	−.0365***	.5338***	1
Great Recession	−.0117*	−.0060	−.0965***	−.2079***	.5079***	−.1143***	−.0379***	1
Service measurability	.0114*	−.0131*	.0198***	.1617***	−.0583***	.3590***	.1250***	−.1334***	1
Asset specificity	−.0053	−.0139**	.0240***	.1930***	−.0610***	.2463***	.0338***	−.1381***	.0557***	1

Note. ***, **, and * are significant at 1%, 5%, and 10% probability levels, respectively.

In addition, Table 4 lists the local services contracted out during the study period, based on the classification that local authorities are obliged to publish, under Local Government Act 7/1985 of April 2, which also establishes a list of minimum basic services that all local governments must provide (the exact content of this list varies depending on the population of the municipality). During the study period, the services that were most commonly contracted out were social services, waste collection services, access to population centers, and domestic water supply.

Table 4.

Services Contracted Out Among Those Which All Local Authorities Were Required to Provide During the Study Period (2000-2010).

Services	Asset specificity	Difficulty of measure	2001	2002	2003	2004	2005	2006	2007	2008	2009	2010	Total
Domestic water supply	High	Low	9	13	7	8	12	11	6	5	13	12	96
Access to population centers^a	High	Low	9	10	7	13	13	12	11	13	13	14	115
Drains and sewers	High	Low	5	5	7	7		3	6	3	5	2	43
Street lighting	High	High	1	2	6	6	3	4	7	1		2	32
Public library	High	High		2	1	1			1				5
Cemetery and funeral services	Low	Low	1	6		2		3	7	5	2	5	31
Public sports facilities	High	High	9	12	6	3	7	10	5	12	15	12	91
Street cleaning	Low	Low	12	9	3	14	5	9	9		1	2	64
Maintenance of street infrastructure	High	Low	4	2	2	4	5	9	11	2			39
Market	Low	High		1	1	1	1	2	2			3	11
Public park	Low	Low	8	8	7	8	1	6	9			1	48
Social services	Low	High	42	30	46	42	43	48	57	19	19	28	374
Fire prevention and firefighting	High	High					1	3	1		1		6
Civil protection	High	High				1		1					2
Environmental protection	Low	Low	1	7	7	5	3	10	7	1		3	44
Waste collection	High	Low	28	18	18	20	20	19	10	25	15	22	195
Selective waste collection	High	Low	3	2	4	5	1	3	3	7	4	3	35
Public transport	High	Low	1	1	2	1		4	1	2	3	3	18
Waste treatment	High	Low	6	1	8	4	5	4	3	8	4	10	53
Total			139	129	132	145	120	161	156	103	95	122	1,302

Source. Author analysis with a private database (see Note 3).

Management, regulation, and control of vehicle access to cities.

Taking into account the above and the fact a municipality may contract out various services in one or more of the years during the study period, a dichotomous dependent variable (0 if no contracting out takes place, and 1 if it does) used in the models under evaluation was constructed for each of the services studied. This means that if a municipality contracted out two or more services in a given year, the same number of separate cases were analyzed, taking into account the same socioeconomic and financial variables in the year in question. For example, in the municipality of Burgo de Osma, three contracting out events took place during the study period (in 2003, the domestic water supply, and in 2004 the waste collection service and the provision of social services). Thus, the database incorporated three contracting out events and three cases were addressed by the study method: one in 2003 and two in 2004 (see Table 5).

Table 5.

Example of Application of the Survival Model to the Contracting Out of Public Services.

Year	Municipality	Id	Sequence	Contracting out (dependent variable)	Short-run solvency	Flexibility	Sustainability	Independence	Population-log-	Unemployment	Index of economic activity
2000	BURGO DE OSMA	1	1	0	9.0525	3.9840	1.2899	0.8015	8.5104	3.7	0.1
2001	BURGO DE OSMA	1	2	0	2.9423	5.4561	0.8849	0.4079	8.5226	3.4	0.1
2002	BURGO DE OSMA	1	3	0	1.3437	5.2641	0.9755	0.4106	8.5342	2	0.09
2003	BURGO DE OSMA	1	4	1	2.3445	5.6694	0.9854	0.4765	8.5204	1.9	0.11
2000	BURGO DE OSMA	2	1	0	9.0525	3.9840	1.2899	0.8015	8.5104	3.7	0.1
2001	BURGO DE OSMA	2	2	0	2.9423	5.4561	0.8849	0.4079	8.5226	3.4	0.1
2002	BURGO DE OSMA	2	3	0	1.3437	5.2641	0.9755	0.4106	8.5342	2	0.09
2003	BURGO DE OSMA	2	4	0	2.3445	5.6694	0.9854	0.4765	8.5204	1.9	0.11
2004	BURGO DE OSMA	2	5	1	4.0604	8.2408	1.1406	0.6031	8.5236	2.4	0.08
2000	BURGO DE OSMA	3	1	0	9.0525	3.9840	1.2899	0.8015	8.5104	3.7	0.1
2001	BURGO DE OSMA	3	2	0	2.9423	5.4561	0.8849	0.4079	8.5226	3.4	0.1
2002	BURGO DE OSMA	3	3	0	1.3437	5.2641	0.9755	0.4106	8.5342	2	0.09
2003	BURGO DE OSMA	3	4	0	2.3445	5.6694	0.9854	0.4765	8.5204	1.9	0.11
2004	BURGO DE OSMA	3	5	1	4.0604	8.2408	1.1406	0.6031	8.5236	2.4	0.08

Note. Id identifies the municipality performing each contracting out operation. Consequently each municipality is repeatedly measured over time until it moves from public to contracted out delivery, as many times as there are contracting out events in the municipality concerned.

Method

To test the above hypotheses, the methods commonly used in studies of the fiscal stress–contracting out relation must be evaluated. Accordingly, we propose an initial model containing just one variable measuring fiscal stress, together with the socioeconomic variables within the environment in question (Table 6; Models 1.1-1.4). We then introduce a model that includes all the variables addressed in previous models (Table 6; Model 1.5). All of these variants of Model 1 are evaluated with respect to a single year, using a logit model in which the dependent variable takes the value 0 if no services were contracted out and 1 if contracting out took place, during 2001.⁷ The second model proposed is similar to Model 1, but instead of a single year, this model evaluates the variation of all the independent variables over several widely separated years. Specifically, we implemented the model to reflect the variations in the financial condition with respect to the socioeconomic environmental variables for the period from 2000 to 2010. Models 2.1 to 2.4 (see Table 6) incorporate a single variable measuring fiscal stress, whereas Model 2.5 (see Table 6) includes all the variables, considered jointly. As in the first case, Model 2 uses a logit design in which the dependent variable takes the value 0 if no services were contracted out in 2010 and 1 if contracting out took place in that year, following the approach described by Zullo (2009). These models were then assessed against Model 3 (Submodels 3.1-3.7) in Table 7, which takes the form of a survival (or duration) analysis, to take into account that the probability of the municipality deciding to contract out its services may vary over time. According to Gómez, Polo, and Fuentelsaz (2004) and Jenkins (2005), the above models are more appropriate than discrete models (cross-sectional ones, such as logit or probit) for analyzing situations in which the focus is on the time elapsed or duration until the occurrence of a certain event, because they take into account the information on the moment of time when the contracting out decision is made and thus provide information on the evolution over time of the explanatory variables. Accordingly, by means of these models, we were able to determine the factors that influence the timing of the contracting out decision.

Table 6.

Results for Models 1 and 2.

	Model 1: Individual measures of fiscal stress for discrete choice model (only 1 year; e.g., 2001)					Model 2: Individual measures of fiscal stress for discrete choice model (variation from 2000 to 2010)
Submodels	1.1	1.2	1.3	1.4	1.5	2.1	2.2	2.3	2.4	2.5
Duration dependence^a
Short-run solvency
Cash Surplus Index	0.972 (0.040)				0.9467 (0.049)	0.997 (0.009)				0.997 (0.009)
Flexibility
Taxable value divided by Financial Charge Index		0.999 (0.0002)			0.999 (0.0002)		0.998 (0.014)			0.999 (0.015)
Sustainability
Nonfinancial Budgetary Result Index			1.400 (0.721)		0.437 (0.292)			0.910 (0.379)		0.948 (0.395)
Independence
Financial Independence Index				1.652 (2.031)	2.34 (5.867)				0.925 (0.249)	0.941 (0.252)
Unemployment	0.919* (0.046)	0.920* (0.046)	0.922 (0.046)	0.971 (0.049)	0.985 (0.050)	1.084 (0.074)	1.084 (0.074)	1.085 (0.074)	1.084 (0.075)	1.084 (0.075)
Population-log-	2.925*** (0.251)	2.917*** (0.251)	2.922*** (0.251)	2.773*** (0.244)	2.713*** (0.249)	1.10** (0.205)	1.10** (0.204)	1.10** (0.205)	1.10** (0.206)	1,097** (0.206)
Index of Economic Activity	0.969** (0.014)	0.970** (0.015)	0.969** (0.014)	0.975 (0.016)	0.978 (0.017)	0.838** (0.075)	0.839** (0.074)	0.838** (0.074)	0.838** (0.074)	0.838** (0.074)
Pseudo-R²	.2544	.2541	.2544	.2238	.2368	.0533	.0534	.0533	.0534	.0534

Note. Model 1 reports the estimated coefficients transformed to odds ratios. The robust standard errors are shown in parentheses. Model 2 also reports the estimated coefficients transformed to odds ratios, with the robust standard errors in parentheses. However, in this case, all the variables are calculated in terms of the variation, and the log transform is not applied to the variable “population.”

This variable is intrinsic to the methodology used. It measures the time sequence (number of years, or survival time) until the contracting out decision is taken. This variable was created as the ln (survival time) and is included in the estimations to reflect whether the hazard rate is monotonically increasing, decreasing, or constant. Survival time is the variable that uniquely identifies each time period during which local services might be contracted out in each municipality.

p < .10. **p < .05. ***p < .01.

Table 7.

Results for Models 3 to 5.

	Model 3: Survival analysis, financial condition, passage-of-time effect (complementary log–log, 2000-2010) hazard ratios							Model 4: Survival analysis, financial condition, passage-of-time effect (complementary log–log, 2000-2010) with Great Recession	Model 5: Survival analysis, financial condition, passage-of-time effect (complementary log–log, 2000-2010) with Great Recession, Service measurability and asset specificity
Submodels	3.1	3.2	3.3	3.4	3.5	3.6	3.7	Hazard ratio	Hazard ratio
Duration dependence^a	2.349*** (0.111)	2.351*** (0.111)	2.355*** (0.111)	2.357*** (0.111)	2.341*** (0.110)	2.353*** (0.111)	2.343*** (0.110)	2.544*** (0.140)	2.979*** (0.177)
Short-run solvency
Cash Surplus Index	0.998* (0.002)				0.998* (0.002)	0.998* (0.002)		0.998* (0.002)	0.998* (0.002)
Flexibility
Taxable value divided by Financial Charge Index		0.999 (0.0001)			0.999 (0.0001)			0.999 (0.0001)	0.999 (0.0001)
Sustainability
Nonfinancial Budgetary Result Index			1.130 (0.203)		0.456*** (0.093)	0.634** (0.098)	0.455*** (0.092)	0.425*** (0.088)	0.501*** (0.111)
Independence
Financial Independence Index				2.419*** (0.296)	3.983*** (0.732)		3.965*** (0.727)	4.012*** (0.740)	2.801*** (0.553)
Unemployment	0.974*** (0.010)	0.974*** (0.010)	0.974** (0.010)	0.990 (0.010)	0.996 (0.010)	0.974** (0.010)	0.996 (0.010)	1.011 (0.011)	1.044*** (0.011)
Population-log-	1.684*** (0.029)	1.684*** (0.030)	1.684*** (0.030)	1.632*** (0.031)	1.607*** (0.031)	1.684*** (0.030)	1.608*** (0.031)	1.594*** (0.031)	1.342*** (0.029)
Index of Economic Activity	0.977*** (0.003)	0.977*** (0.003)	0.977*** (0.003)	0.978*** (0.002)	0.979*** (0.002)	0.977 (0.003)	0.979*** (0.002)	0.979*** (0.002)	0.988*** (0.002)
Great Recession								0.780*** (0.069)	0.932** (0.084)
Service measurability									3.148*** (0.202)
Asset specificity									4.358*** (0.283)
Prob > χ²	Prob > χ² = .000	Prob > χ² = .000	Prob > χ² = .000	Prob > χ² = .000	Prob > χ² = .000	Prob > χ² = .000	Prob > χ² = .000	Prob > χ² = .000	Prob > χ² = .000
Log pseudolikelihood	−4,400.1945	−4,400.1735	−4,400.0734	−4,380.9211	−4,372.8613	−4,399.9296	−4,373.2859	−4,369.3262	−3,928.0092

Note. Models 3, 4, and 5 present the hazard rate for each covariate and the robust standard errors are shown in parentheses.

This variable was created as the ln (survival time) and is included in the estimations to reflect whether the hazard rate is monotonically increasing, decreasing, or constant. Survival time is the variable that uniquely identifies each time period during which local services might be contracted out in each municipality.

p < .10. **p < .05. ***p < .01.

Municipalities are observed repeatedly, and so the observations are not independent. The vce(cluster clustvar) option offered by the Stata software provides a way to fit this model and obtains correct standard errors.

The survival models implemented in this study are of the discrete-time proportional hazard type, in which duration (time) is treated as a discrete variable, not because it is discrete (here, the time is continuous) but because the data are available only annually and therefore, the observations for each municipality are observed in a discrete way. In discrete-time models, the discrete-time hazard function is the probability of transition at discrete time t_j, j = 1, 2, . . ., given survival up to time t_j (Máñez, Pérez-López, Prior, & Zafra-Gómez, 2008). The risk associated with the discrete outcomes of the event can be set out in the concept of discrete-time hazard probability. It is essentially the conditional probability of the event being experienced by a randomly selected municipality during a particular time period, conditional to the fact that he had not already experienced it within the period considered (Graham, Willett, & Singer, 2013). Let T_i be the value of the survival time T for the i municipality and j represents the time period in which the local government i experiences the target event. Then, the hazard probability can be expressed as follows:

h (t_{i j}) = \Pr ⌈ T_{i} = j | T_{i} \geq j) ⌉,

where h(t_ij) represents the conditional hazard probability of an municipality i experiencing a first failure during the time period j.

As the rate of discrete-time hazard is a probability in Equation 1, the same is generally modeled as logistic or complementary log−log (cloglog) functional specifications. Given the binary responses y_ij and the hazard probability h_ij, Singer and Willett (1993) derived the log-likelihood function for the discrete-time hazard process as follows:

L = \prod_{i = 1}^{n} \prod_{j = 1}^{j_{i}} h_{i_{j}}^{y_{i j}} {(1 - h_{i j})}^{(1 - y_{i j})} .

The log-likelihood in Equation 2 involving discrete-time hazard probability can alternatively be modeled as a complementary log−log (cloglog) specification $h_{i j} = 1 - \exp (- \exp (α_{j} + β' X_{i}))$ , which can be rewritten as follows:

c l o g l o g (h_{i j}) = l o g (- \log (1 - h_{i j})) = α_{j} + β' X_{i},

where $α_{j}$ is the baseline hazard. This baseline hazard corresponds to the hazard attributed to the passage of survival time, which is common to all municipalities, and allows us to determine the influence of the passage of time on the probability of contracting out taking place (González-Gómez & Guardiola, 2009; Pradhan & Das, 2016).

Taking into account these considerations, together with the data specification obtained, we estimated three complementary log-log models (Models 3, 4, 5, and 6), that is, the clog-log model,⁸ to associate the municipality’s hazard rate with the time-varying covariates (see Prentice & Gloeckler, 1978). These models allow us to study factors that influenced contracting out during the period 2000-2010, taking into account the influence of the passage of time on the probability of contracting out. We present hazard rate ratios, which are easier to interpret. The hazard rate for each covariate is interpreted assuming that a value greater than 1 corresponds to a positive effect on the hazard rate, and that a value of less than 1 corresponds to a negative impact. Our study also included the effects of GR. Finally, we incorporated the characteristics of the municipal services considered, to determine whether they influenced the contracting out decision. Specifically, we followed the approach described by Brown and Potoski (2003), Brown, Potoski, & Van Slyke, (2008), and Rodrigues, Tavares, and Araujo, (2012) on asset specificity and service measurability, but adapted to the Spanish case following the criteria used by González-Ramírez, Gascó, and Taverner (2011) with respect to the importance or priority granted by each municipality to its different services, and also taking into account the degree of contracting out performed. Figure 2 summarizes these models.

Figure 2.

Different temporal models to explain financial stress–contracting out relationship (with other factors).

Results

To test the first hypothesis, it was necessary to study the results of the first three models (see Tables 6 and 7).

This analysis shows that neither the models with a single variable of financial condition nor the joint models—whether measuring a single year (2001; Model 1) or the variations over an extended period (2000-2010; Model 2)—are able to explain the contracting out of local public services. This finding confirms the conclusions of recent studies, conducted in the United States (Levin & Tadelis, 2010; Zullo, 2009), in which the fiscal stress–contracting out relation was tested by applying a model similar to our Model 1, as in the study carried out by Levin and Tadelis (2010), or in which a data structure was applied in terms of variations (Model 2), as proposed by Zullo (2009). On comparing these models with Model 3, which replicates Models 1 and 2 but takes into account the passage of time (i.e., with survival analysis), it can be seen that with the inclusion of this consideration in evaluating each submodel with a single indicator of fiscal stress (individual measurement, from Submodels 3.1 to 3.4), the individual measures of financial condition are statistically significant in half of the submodels considered, whereas the variable “duration dependence” is significant in all models (Guardiola, González-Gómez, & García-Rubio, 2010). Specifically, in Models 3.1 and 3.4, both short-run solvency and independence influence the contracting out decision. This explains why the cash surplus index was inversely associated with contracting out (hazard rate = 0.998; p = .086, in Submodel 3.1), that is, a decrease in the local authority’s cash position increased the likelihood of its contracting out public services. These results confirm that an organization’s cash position can be used to define its fiscal stress–contracting out relation, an aspect that previous results had not addressed. The financial independence index is the most significant factor influencing the contracting out decision (hazard rate = 2.419; p = .000, in Model 3.4). Thus, when a local authority receives less funding from central government, in relation to its revenue structure, to cover spending needs, it is more likely to contract out public services. This finding corroborates those of other studies carried out in Europe, where the level of financial dependence of local authorities tends to be very high (30%-40% of the revenue budget; Plata-Díaz, Zafra-Gómez, Pérez-López, & López-Hernández, 2014). However, studies conducted in the United States have reported no such evidence. These results confirm the need to measure the passage of time to properly understand the contracting out process, and hence H1A is confirmed. However, the passage of time does not affect all the elements of financial condition equally, and so we were unable to confirm H1B, because the variables budgetary sustainability and flexibility, in Models 3.2 and 3.3, were not significant, and therefore do not affect the contracting out decision as an individual measure of financial condition.

Nevertheless, the fact that, individually, they do not affect contracting out does not mean that the simultaneous presence of several variables of financial condition is required for contracting out to take place (DiNapoli, 2013). Submodels 3.6 and 3.7 address this consideration, as the variable sustainability, in itself, is never significant (Model 3.3), but when it is taken together with short-run solvency (Model 3.6) or with financial independence (Model 3.7), then it is. Accordingly, we conclude that not all elements of financial condition have the same weight/influence on the contracting out decision, and that in some cases, such as the existence of budget deficit, another element of financial condition must be present.⁹

The maximum possible development of combinations of indicators is to incorporate all the elements of financial condition into a single model, as in our Models 3, 4, and 5 (see Table 7). Model 4 shows that when this is done, three of the four elements are significant. In this case, budgetary sustainability also has a negative and significant effect on the probability of contracting out (hazard rate = 0.425; p = .000). This implies that the larger the public sector deficit, the greater the probability of services being contracted out, provided this factor is combined with short-run solvency and financial independence. Similar indicators have been used in studies carried out in the United States, but with different results. Thus, Kodrzycki (1998) found similar results for cities and towns in terms of variation from 1987 to 1992. However, Zullo (2009) did not find this indicator to be significant in terms of variation for the period 1992-1997.

It should be noted that the only element of financial condition that bears no significant relationship with contracting out is flexibility; in other words, if the organization becomes less able to repay its debt from its income, this does not influence the contracting out of services. This finding is consistent with the results obtained by Kodrzycki (1998), Levin and Tadelis (2010), Zullo (2009), and Pérez-López, Plata-Díaz, Zafra-Gómez, & López-Hernández, (2013), none of whom measured any significant association between the debt interest/revenue relation and the likelihood of public services being contracted out.

The results of Models 3.5, 3.6, 3.7, and 4 confirm H1C, that some elements of financial condition need to be combined with others for them to have a significant influence on contracting out. The socioeconomic environment variables that constitute financial condition all present similar results. Thus, the “population” variable is significantly related to contracting out in all the models analyzed, and an increase in the population is associated with a greater likelihood of public services being contracted out. However, the index of economic activity bears a significant inverse relationship with contracting out, which means that when economic activity decreases, there is a greater likelihood of services being contracted out.

Another hypothesis considered was that of the influence of GR on contracting out (H2). Models 4 and 5 both reflect the existence of this relationship; thus, GR produced a significant negative effect on contracting out (hazard rate = 0.780; p = .005, in Model 4; hazard rate = 0.932; p = .022, in Model 5), that is, contracting out declined during this period. These results lead us to reject H2 according to which crises constrain local decision makers to rely more on in-house options and thus make the provision of local public services more readily controllable. These results are also consistent with those of Hefetz and Warner (2007), who reported that reverse contracting (reinternalization) now exceeds the level of new contracting out, because public managers have expanded their concerns beyond transaction costs and efficiency to pay more attention to citizens’ voices in the service delivery process.

Given this situation, it would be useful to analyze whether the influence of financial condition on contracting out decisions changed during GR, which would verify H3. Accordingly, Model 6 (see Table 8) presents the results obtained concerning the effect of fiscal stress variables, differentiating the pre-GR period (2000-2007) from that of GR (2008-2010).

Table 8.

Results for Model 6.

	Model 6: Survival analysis, financial condition, passage-of-time effect (complementary log-log, 2000-2010) with financial condition/Great Recession interaction
	Hazard ratio
Duration dependence	2.962*** (0.174)
Short-run solvency
Cash Surplus Index × Pre-GR (2000-2007)	0.997 (0.003)
Cash Surplus Index × GR (2008-2010)	1.014* (0.010)
Flexibility
Taxable value divided by Financial Charge Index × Pre-GR (2000-2007)	0.999 (0.0001)
Taxable value divided by Financial Charge Index × GR (2008-2010)	0.999 (0.0002)
Sustainability
Nonfinancial Budgetary Result Index × Pre-GR (2000-2007)	0.598** (0.142)
Nonfinancial Budgetary Result Index × GR (2008-2010)	0.329*** (0.089)
Independence
Financial Independence Index × Pre-GR (2000-2007)	2.280*** (0.524)
Financial Independence Index × GR (2008-2010)	5.114*** (1.651)
Unemployment	1.045*** (0.011)
Population-log-	1.340*** (0.029)
Index of Economic Activity	0.988*** (0.002)
Service measurability	3.159*** (0.203)
Asset specificity	4.354*** (0.282)
Prob > χ²	Prob > χ² = .000
Log pseudolikelihood	−3,925.7023

Note. Model 6 presents the hazard rate for each covariate and the robust standard errors are shown in parentheses. GR = Great Recession.

p < .10. **p < .05. ***p < .01.

As can be seen, the financial condition values used to measure fiscal stress present similar behavior patterns except in the case of cash solvency. In other words, the variables that measure sustainability and independence presented a similar pattern of behavior in both periods. Nevertheless, as mentioned above, cash solvency did indeed vary from one period to the other. During the pre-GR period, this variable presented a negative, nonsignificant association with contracting out (hazard rate = 0.997), but during GR the effect became positive and significant (hazard rate = 1.014; p = .077). This may be because a major proportion of municipal revenues is paid to contractors for the services provided, and during GR these contractors preferred local entities to have liquid funds, to reduce the risk of nonpayment (Zullo 2009; Zafra-Gómez, Pedagua, Plata-Díaz, & López-Hernández, 2014).

In view of these results, we partially reject H3, because the indicators that worsened during GR were financial independence and budgetary sustainability (in compliance with H3), whereas the relation of cash solvency with contracting out was reversed, to become positive (contradicting H3). The exogenous components of financial condition behaved in a similar way.

Finally, as discussed above, the complexity in measuring the service and the presence of specific assets (as reflected in Models 5 and 6) are significantly and positively associated with contracting out, and so H4 and H5 are rejected, because in the Spanish municipalities analyzed, the presence of specific assets and of measurement complexity makes local authorities more likely to contract out public services (Brown & Potoski, 2003). We explain this association as arising from the fact that highly specific assets are difficult to adapt for use in other types of service (Carr, LeRoux, & Shrestha, 2009).

Conclusion

This article presents new evidence on the fiscal stress–contracting out relation, within the framework of a new methodology. In this analysis, the passage of time is considered a decisive factor in the contracting out process. Furthermore, fiscal stress is defined over a broad time span, incorporating the concept of financial condition. By examining the issue over an extended period, this study considers the impact made by GR on the contracting out process, taking into account that a transboundary crisis may require a particular diagnosis and solution, differing from the policies that may be suitable for resolving financial problems during periods of normal economic activity.

We applied a survival model for the period 2000-2010 and compared its results with those obtained by discrete choice and cross-sectional models (Model 1, for a single year; Model 2, the variation from 2000 to 2010) that do not take into account the effect of the passage of time, and thus lack a broad outlook on the question of fiscal stress. This approach allowed us to test the hypotheses proposed and thus characterize more precisely the relationships between financial stress and contracting out, by focusing on the impact of financial stress both before GR and during this period. We conclude that to measure the influence of fiscal stress on contracting out, the passage of time must be taken into account; moreover, an extended period of time (long-run solvency) should be considered. However, not all the elements of financial condition have the same influence on the contracting out decision; some (short-run solvency and independence), by themselves, have sufficient weight to lean public managers toward contracting out, whereas others (sustainability) are only influential in combination with other elements.

However, when our analysis is focused on an extended period of time, the impact of transboundary crises can affect the results obtained, and it has been observed that during GR, public managers made less use of the contracting out option. Moreover, there were changes in the influence of the indicators of financial condition. These findings suggest that during GR public managers granted more importance to the question of intergovernment subsidies than the deficit, and that a positive cash balance is necessary to undertake contracting out. However, the results obtained show that when services are more complex and/or require more specific assets, there is a greater probability of services being contracted out. With respect to the socioeconomic parameters analyzed, as part of the financial condition, we conclude that local services are more likely to be contracted out when the population is increasing and when the economic activity is worsening.

Discussion

Several theoretical frameworks have been proposed in which market mechanisms are adopted by local authorities to respond to a situation of fiscal stress. These approaches include the theory of property rights (Alchian, 1967), organization theory (Earle, 2006), and that of public choice. The latter is perhaps the most widely analyzed and therefore has had most influence on studies of the relationship between contracting out and fiscal stress. The fiscal stress–contracting out relation, thus, has been the object of intensive research. Nevertheless, the results obtained in this field have been inconclusive, and a more exhaustive analysis is needed. By studying this relationship with respect to the situation in Spain, we hope to clarify its impact in the European and U.S. contexts, too, because, irrespective of specific regulatory conditions, a poor financial situation may arise in any local economic and political context. The contracting out decision is influenced by diverse factors, including the characteristics of the service provider, the information available to local public managers and, possibly, political aspects. The latter question is not analyzed in this article, as Dubin and Navarro (1988) reported its influence to be unclear. Asset specificity and service complexity should also be taken into account, as services that need more specific assets and more complex measurement are more likely to be contracted out. Other limitations that have been observed in previous studies are the lack of a suitable instrument to measure fiscal stress, the failure to consider a sufficiently broad time horizon, and the absence of the passage of time as an explanatory factor of the decision to contract out municipal services. In our opinion, all of these aspects must be taken into consideration if we are to properly evaluate the fiscal stress–contracting out relation (Bel & Fageda 2007; González-Gómez & Guardiola, 2009).

Analysis of our methodological proposal shows that to determine the relationship between fiscal stress and contracting out, the influence of the passage of time must be taken into account. Moreover, if the model is to be specified properly, a comprehensive set of indicators should be evaluated, over a lengthy period of time, even if, in some cases, individual factors may sway the contracting out decision, but the impact/weight of each such factor on the probability of contracting out will vary. Therefore, studies should be conducted with larger numbers of indicators and models, with diverse combinations of the latter, to determine the fiscal stress–contracting out relationship with greater reliability. However, when the time period examined includes subperiods during which a transboundary crisis occurs, influencing the contracting out decision, this relationship changes. In this respect, the present study shows that GR led many public managers to restrict the use of contracting out in 2008-2010. In other words, during this period public managers did not view contracting out as the best means of providing public services.

GR modified the attitudes of public managers, making them less inclined to place municipal policies in the hands of the market, due to the financial difficulties experienced, such as declining revenues and budget constraints (López-Hernández, Zafra-Gómez, & Ortíz-Rodríguez, 2012). In consequence, current theoretical models incorporate the arguments put forward by proponents of the social choice theory (Ruiz-Villaverde, Picazo-Tadeo, & González-Gómez, 2015), which are critical of public choice, claiming that it is incapable of safeguarding collective interests and only serves the special interests of the market.

With respect to these questions, the model we present is incomplete, as it makes no analysis of other factors that may influence contracting out, according to other theoretical frameworks. In particular, a more detailed analysis is needed of service provider characteristics and of the type of information available to local managers in terms of costs. Furthermore, political factors such as ideology, political fragmentation and political cycles need to be addressed, together with the possible influence of personal interests of local public managers, which may give rise to improper behavior in the decision-making process regarding contracting out. Finally, it would be very interesting to examine whether fiscal stress occurred as a result of rising costs or falling income during the study period.

Footnotes

Declaration of Conflicting Interests

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

Funding

The author(s) disclosed receipt of the following financial support for the research, authorship, and/or publication of this article: We are grateful for the financial support received from Ministerio de Educación y Ciencia (Spain; ECO2013-48413-R and ECO2016-76578-R), Proyectos de Excelencia de la Consejería de Innovación, Ciencia y Empresa, Junta de Andalucía (Spain; P11-SEJ-7700) and Beca del Ministerio de Educación, Cultura y Deporte Formación Profesorado Universitario (FPU14/01403).

Notes

Author Biographies

Antonio M. López-Hernández is a University Professor of Granada and President of the Chamber of Accounts of the Andalusian Government, in addition to having several publications in journals indexed in first and second quarters

José L. Zafra-Gómez is a Professor of the University of Granada with several publications in journals indexed in JCR of first and second quartile.

Ana M. Plata-Díaz is an Assistant Professor Doctor of the University of Granada with diverse publications in journals indexed in JCR with high index of impact.

Emilio J. de la Higuera-Molina graduated in Finance and Accounting, Master in Management and Public Management and PhD student at the University of Granada funded by the grants of the Ministry of Education, Culture and Science FPU.

References

Abad-González

Gutiérrez-López

(2016). Modelización de la solvencia bancaria en escenarios adversos: Aplicación a los «PIIGS». Revista de Contabilidad, 19, 227-238. doi:10.1016/j.rcsar.2015.11.002

Alchian

A. A.

(1967). Pricing and society. London, England: Institute of Economic Affairs.

Ansell

Boin

Keller

(2010). Managing transboundary crises: Identifying the building blocks of an effective response system. Journal of Contingencies and Crisis Management, 18, 195-207. doi:10.1111/j.1468-5973.2010.00620.x

Balaguer-Coll

M. T.

Brun-Martos

M. I.

Forte

Tortosa-Ausina

(2015). Local governments’ re-election and its determinants: New evidence based on a Bayesian approach. European Journal of Political Economy, 39, 94-108. doi:10.1016/j.ejpoleco.2015.04.004

Bastida

Benito

Guillamón

M. D.

(2009). An empirical assessment of the municipal financial situation in Spain. International Public Management Journal, 12, 484-499. doi:10.1080/10967490903328139

Bel

(2006). Economía y política de la privatización local [Economics and politics of local privatization].Madrid, España: Marcial Pons.

Bel

Fageda

(2006). Between privatization and intermunicipal cooperation: Small municipalities, scale economies and transaction costs. Urban Public Economics Review, 6, 13-31.

Bel

Fageda

(2007). Why do local governments privatize public services? A survey of empirical studies. Local Government Studies, 33, 517-534. doi:10.1080/03003930701417528

Bel

Fageda

(2008). Reforming the local public sector: Economics and politics in privatization of water and solid waste. Journal of Economic Policy Reform, 11, 45-65. doi:10.1080/17487870802134884

10.

Bel

Fageda

(2010). Partial privatisation in local services delivery: An empirical analysis of the choice of mixed firms. Local Government Studies, 36, 129-149. doi:10.1080/03003930903435856

11.

Bel

Fageda

(2011). Big guys eat big cakes: Firm size and contracting in urban and rural areas. International Public Management Journal, 14, 4-26. doi:10.1080/10967494.2011.547746

12.

Bel

Fageda

Warner

M. E.

(2010). Is private production of public services cheaper than public production? A meta-regression analysis of solid waste and water services. Journal of Policy Analysis and Management, 29, 553-577. doi:10.1002/pam.20509

13.

Berne

Schramm

(1986). The financial analysis of government. Englewood Cliffs, NJ: Prentice Hall.

14.

Boyne

G. A.

(1998). Bureaucratic theory meets reality: Public choice and service contracting in U.S. local government. Public Administration Review, 58, 474-484. doi:10.2307/977575

15.

Brown

T. L.

Potoski

(2003). Transaction costs and institutional explanations for government service production decisions. Journal of Public Administration Research and Theory, 13, 441-468. doi:10.1093/jpart/mug030

16.

Brown

T. L.

Potoski

(2005). Transaction costs and contracting: The practitioner perspective. Public Performance & Management Review, 28, 326-351.

17.

Brown

T. L.

Potoski

Van Slyke

D. M.

(2006). Managing public service contracts: Aligning values, institutions, and markets. Public Administration Review, 66, 323-331. doi:10.1111/j.1540-6210.2006.00590.x

18.

Brown

T. L.

Potoski

Van Slyke

D. M.

(2008). Changing modes of service delivery: How past choices structure future choices. Environment and Planning C: Government and Policy, 26, 127-143. doi:10.1068/c0633

19.

Cabeza

(2009). Reflexiones sobre la crisis económica y el papel de la auditoria pública. Auditoría Pública, 47, 27-45.

20.

Canadian Institute of Chartered Accountants. (1997). Indicators of government financial condition. Toronto, Ontario, Canada: Canadian Institute of Chartered Accountants.

21.

Carboni

(2015). Ex post contract market structure: Implications for performance over time. American Review of Public Administration. Advance online publication. doi:10.1177/0275074015608753

22.

Carr

J. B.

LeRoux

Shrestha

(2009). Institutional ties, transaction costs, and external service production. Urban Affairs Review, 44, 403-427. doi:10.1177/1078087408323939

23.

Christoffersen

Bo Larsen

(2007). Economies of scale in Danish municipalities: Expenditure effects versus quality effects. Local Government Studies, 33, 77-95.

24.

Diefenbach

(2009). New public management in public sector organizations: The dark side of managerialistic “enlightenment.” Public Administration, 87, 892-909. doi:10.1111/j.1467-9299.2009.01766.x

25.

Dijkgraaf

Gradus

R. H. J. M.

Melenberg

(2003). Contracting out refuse collection. Empirical Economics, 28, 553-570. doi:10.1007/s001810200145

26.

DiNapoli

T. P.

(2013). Fiscal stress drivers and coping strategies. Albany, New York: Office of the New York State Comptroller, Division of Local Government and School Accountability.

27.

Dubin

Navarro

(1988). How markets for impure public goods organize: The case of household refuse collection. Journal of Law, Economics & Organization, 4, 217-241.

28.

Earle

J. S.

(2006). Employment and wage effects of privatization: Evidence from transition economies. Employment Research, 13, 3-6. doi:10.17848/1075-8445.13(1)-2

29.

Easterly

Rebelo

(1993). Fiscal policy and economic growth. Journal of Monetary Economics, 32, 417-458. doi:10.1016/0304-3932(93)90025-B

30.

Evans

Heath

Lalljee

(1996). Measuring left-right and libartarian-authoritarian values in the British electorate. British Journal of Sociology, 93-112.

31.

Ferris

J. M.

(1986). The decision to contract out: An empirical analysis. Urban Affairs Quarterly, 22, 289-311. doi:10.1177/004208168602200206

32.

Ferris

J. M.

Graddy

(1986). Contracting out: For what? With whom? Public Administration Review, 46, 332-344. doi:10.2307/976307

33.

Ferris

J. M.

Graddy

(1991). Production costs, transaction costs, and local government contractor choice. Economic Inquiry, 29, 541-554. doi:10.1111/j.1465-7295.1991.tb00845.x

34.

Funkhouser

(2012). Outsourcing and the critical need for access to capital. Public Administration Review, 72, 900-901. doi:10.1111/j.1540-6210.2012.02655.x

35.

Gómez

Polo

Fuentelsaz

(2004). Aplicaciones del análisis de supervivencia a la investigación en economía de la empresa. Cuadernos de Economía y Dirección de la Empresa, 19, 81-114

36.

González-Gómez

Guardiola

(2009). A duration model for the estimation of the contracting-out of urban water management in southern Spain. Urban Affairs Review, 44, 886-906. doi:10.1177/1078087408329274

37.

González-Ramírez

Gascó

J. L.

Taverner

(2011). Relaciones entre outsourcing y estrategia en los ayuntamientos españoles. Auditoría pública: revista de los Organos Autónomos de Control Externo, 53, 33-46.

38.

Graham

S. E.

Willett

J. B.

Singer

J. D.

(2012). Using discrete-time survival analysis to study event occurrence. In Newsom

Jones

R. N.

Hofer

S. M.

(Ed.), Longitudinal data analysis: A practical guide for researchers in aging, health, and social sciences (pp. 329-371). New York, NY: Routledge

39.

Greenberg

Hillier

(1995). Indicators of financial condition for governments. Paper presented at the 5th Conference of Comparative International Governmental Accounting Research, Paris-Amy.

40.

Groves

S. M.

Godsey

W. M.

(2003). Evaluating financial condition: A handbook of local government. Washington, DC: International City/County Management Association.

41.

Guardiola

González-Gómez

García-Rubio

M. A.

(2010). Is time really important for research into contracting out public services in cities? Evidence for urban water management in southern Spain. Cities, 27, 369-376. doi:10.1016/j.cities.2010.05.004

42.

Hatch

M. J.

Cunliffe

A. L.

(2013). Organization theory: Modern, symbolic and postmodern perspectives. Oxford, UK: Oxford University Press.

43.

Hebdon

Jalette

(2008). The restructuring of municipal services: A Canada–United States comparison. Environment and Planning C: Government and Policy, 26, 144-158.

44.

Hefetz

Warner

M. E.

(2007). Beyond the market versus planning dichotomy: Understanding privatisation and its reverse in US cities. Local Government Studies, 33, 555-572. doi:10.1080/03003930701417585

45.

Hendrick

R. M.

(2011). Managing the fiscal metropolis: The financial policies, practices, and health of suburban municipalities. Washington, DC: Georgetown University Press.

46.

Honadle

B. W.

(2003). The States’ role in U.S. local government fiscal crises: A theoretical model and results of a national survey. International Journal of Public Administration, 26, 1431-1472. doi:10.1081/PAD-120024405

47.

Hood

(1991). A public management for all seasons? Public Administration, 69, 3-19. doi:10.1111/j.1467-9299.1991.tb00779.x

48.

Hood

(1995). The new public management in the 1980s: Variations on a theme. Accounting, Organizations and Society, 20, 93-110. doi:10.1016/0361-3682(93)E0001-W

49.

International City/County Management Association. (1989). Service delivery in the 90s: Alternative approaches for local governments. Washington, DC: Author.

50.

International City/County Management Association. (2003). Evaluating financial condition: A handbook for local government (4th ed.). Washington, DC: Author.

51.

James

E. H.

Wooten

L. P.

(2010). Leading under pressure: From surviving to thriving before, during, and after a crisis. New York, NY: Routledge.

52.

Jenkins

S. P.

(2005). Survival analysis. Colchester, UK: Institute for Social and Economic Research, University of Essex. Retrieved from https://www.iser.essex.ac.uk/files/teaching/stephenj/ec968/pdfs/ec968lnotesv6.pdf

53.

Johnson

Oliff

Williams

(2011). An update on State budget cuts: At least 46 States have imposed cuts that hurt vulnerable residents and the economy (Report). Center on Budget and Policy Priorities. Retrieved from http://www.cbpp.org/cms/index.cfm?fa=view&id=1214

54.

Jordan

M.M.

Yan

Hooshmand

(2015). The role of State revenue structure in the occurrence and magnitude of negative revenue variance. American Review of Public Administration. Advance online publication. doi:10.1177/0275074015608265

55.

Kodrzycki

Y. K.

(1998). Fiscal pressures and the privatization of local services. New England Economic Review, 39-50.

56.

Laughlin

R. C.

Pallot

(1998). Trends, patterns and influencing factors: Some reflections. In Olson

Guthrie

Humphrey

(Eds.), Global warning: Debating international developments in new public financial management (pp. 376-399). Oslo, Norway: Cappelen Akademisk Forlag.

57.

Levin

Tadelis

(2005). Profit sharing and the role of professional partnerships. The Quarterly Journal of Economics, 120, 131-171. doi:10.1162/0033553053327506

58.

Levin

Tadelis

(2010). Contracting for government services: Theory and evidence from U.S. cities. The Journal of Industrial Economics, 58, 507-541. doi:10.1111/j.1467-6451.2010.00430.x

59.

Levitt

March

J. G

. (1988). Organizational learning. Annual Review of Sociology, 14, 319-340. doi:10.1146/annurev.so.14.080188.001535

60.

López-Hernández

A. M.

Zafra-Gómez

J. L.

Ortíz-Rodríguez

(2012). Effects of the crisis in Spanish municipalities financial condition: an empirical evidence (2005-2008). International Journal of Critical Accounting, 4, 631-645. doi: https://dx-doi-org.web.bisu.edu.cn/10.1504/IJCA.2012.051457

61.

Lungová

(2010). Estimated impacts of economic crisis on local municipalities: Comparison within Europe.

62.

Máñez

J. A.

Pérez-López

Prior

Zafra-Gómez

J. L.

(2016). Understanding the dynamic effect of contracting out on the delivery of local public. Services, Regional Studies, 50, 2069-2080. doi:10.1080/00343404.2015.1090561

63.

Máñez

J. A.

Rochina

M. E.

Sanchis

J. A.

(2008). Using survival models with individual data. In Congregado

(Ed.), Measuring entrepreneurship: Building a statistical system (pp. 227-246). New York, NY: Springer.

64.

Ménard

Saussier

(2000). Contractual choice and performance the case of water supply in France. Revue d’économie industrielle, 92, 385-404. doi:10.3406/tiers.2001.1503

65.

Meyer

J. W.

Scott

W. R.

(1983). Centralization and the legitimacy problems of local government. In Meyer

J. W.

Scott

W. R.

(Ed.), Organizational environments: Ritual and rationality (pp. 199-215). Los Ángeles, CA: Beverly Hills.

66.

Miranda

(1994). Explaining the privatization decision among local governments in the United States. Research in Urban Policy, 5, 231-274.

67.

Morgan

D. R.

Hirlinger

M. W.

(1991). Intergovernmental service contracts: A multivariate explanation. Urban Affairs Quarterly, 27, 128-144. doi:10.1177/004208169102700107

68.

Moss

(2010, July). The 13 largest oil spills in history. Mother Nature Network. Retrieved from http://www.mnn.com/earth-matters/wilderness-resources/stories/the-13-largest-oil-spills-in-history

69.

Mouritzen

P. E.

Nielsen

K. H.

(1988). Handbook of comparative urban fiscal data. Odense, Denmark: Dansk Data Archives.

70.

Nelson

A. A.

Baku

(2014). Local government responses to fiscal stress: Evidence from the public education sector. Public Administration Review, 74, 601-614. doi:10.1111/puar.12211

71.

Niskanen

W. A.

(1971). Bureaucracy and representative government. Chicago, IL: Aldine Atherton

72.

Nogueira

S. P. S.

Jorge

S. M. F.

(2016). Explanatory factors for the use of the financial report in decision-making: Evidence from Local Government in Portugal. Revista de Contabilidad, 19, 216-226. doi:10.1016/j.rcsar.2015.09.002

73.

North

D. C.

(1990). Institutions, institutional change and economic performance. Cambridge, UK: Cambridge University Press.

74.

Organisation for Economic Co-Operation and Development. (2000a). Competition in local services: Solid waste management. Paris, France: Author.

75.

Organisation for Economic Co-Operation and Development. (2000b). Global trends in urban water supply and waste water financing and management: Changing roles for the public and private sectors. Paris, France: Author.

76.

Osborne

Gaebler

(1992). Reinventing government: How the entrepreneurial spirit is transforming the public sector. New York, NY: Penguin.

77.

Pallesen

(2004). A political perspective on contracting out: The politics of good times experiences from Danish local governments. Governance, 17, 573-587. doi:10.1111/j.0952-1895.2004.00258.x

78.

Paulais

(2009). Local governments and the financial crisis: An analysis. Washington, D.C: EEUU, The Cities Alliance.

79.

Pérez-López

Plata-Díaz

A. M.

Zafra-Gómez

J. L.

López-Hernández

A. M.

(2013). Análisis de los factores determinantes y de las formas de gestión [Deuda viva municipal en un contexto de crisis económica]. Revista de contabilidad, 16, 83-93. doi:10.1016/j.rcsar.2012.12.001

80.

Pérez-López

Prior

Zafra-Gómez

J. L.

(2015). Rethinking new public management delivery forms and efficiency: Long-term effects in Spanish local government. Journal of Public Administration Research and Theory, 25, 1157-1183. doi:10.1093/jopart/muu088.

81.

Pérez-López

Prior

Zafra-Gómez

J. L.

Plata-Díaz

A. M.

(2016). Cost efficiency in municipal solid waste service delivery. Alternative management forms in relation to local population size. European Journal of Operational Research, 255, 583-592. doi:10.1016/j.ejor.2016.05.034

82.

Pérez-López

Prior

Zafra-Gómez

J. L.

(2017). Temporal scale Eficiency in DEA panel data estimations. An application to the solid waste disposal service in Spain, Omega. The International Journal of Management Sience, doi:10.1016/j.omega.2017.03.005

83.

Peters

B. G.

Pierre

Randma-Liiv

(2011). Global financial crisis, public administration and governance: Do new problems require new solutions? Public Organization Review, 11, 13-27. doi:10.1016/j.rcsar.2012.12.001

84.

Plata-Díaz

A. M.

Zafra-Gómez

J. L.

Pérez-López

López-Hernández

A. M.

(2014). Alternative management structures for municipal waste collection services: the influence of economic and political factors. Waste Management, 34, 1967-1976. doi: https://dx-doi-org.web.bisu.edu.cn/10.1016/j.wasman.2014.07.003

85.

Pradhan

J. P.

Das

(2016). Regional profile of export duration of firms. In Manufacturing exports from Indian states: Determinants and policy imperatives, 107-129. Springer India.

86.

Prentice

R. L.

Gloeckler

L. A.

(1978). Regression analysis of grouped survival data with application to the breast cancer data. Biometrics, 34, 57-67. doi:10.2307/2529588

87.

Rodrigues

Tavares

A. F.

Araujo

J. F.

(2012). Municipal service delivery: The role of transaction cost in the choice between alternative governance mechanisms. Local Government Studies, 38, 615-638. doi:10.1080/03003930.2012.666211

88.

Rubin

E. L.

(2005). Images of organizations and consequences of regulation. Theoretical Inquiries in Law, 6, 347-390. doi:10.2202/1565-3404.1110

89.

Ruiz-Villaverde

Picazo-Tadeo

González-Gómez

(2015). The “social choice” of privatising urban water services: A case study of Madrid in Spain. Journal of Policy Modeling, 37, 616-629. doi:10.1016/j.jpolmod.2015.04.005

90.

Sager

(2002). Democratic planning and social choice dilemmas: Prelude to institutional planning theory. Aldershot, UK: Ashgate.

91.

Savas

(1987). Privatization: The key to better government. Chatham, NJ: Chatham House.

92.

Singer

J. D.

Willett

J. B.

(1993). It’s about time: Using discrete-time survival analysis to study duration and the timing of events. Journal of Educational and Behavioral Statistics, 18, 155-195.

93.

Steytler

Powell

(2010, September 16-18). The impact of the global financial crisis on decentralized government in South Africa (Local Government Working Paper Series 1, November). Presented at the annual conference of the International Association of Centres for Federal Studies, Philadelphia, PA.

94.

Tavares

A. F.

Camöes

P. J.

(2007). Local service delivery choices in Portugal: A political transaction costs framework. Local Government Studies, 33, 535-553. doi:10.1080/03003930701417544

95.

Touche-Ross Company. (1987). Privatization in America. New York, NY: United States Department of Commerce, Bureau of the Census, County and City Data Book.

96.

United Cities and Local Governments (UCLG). (2009). The impact of the global crisis on local governments. Cavagraf, China: Author.

97.

Vickers

Yarrow

G. K.

(1991). Economic perspectives on privatization. The Journal of Economic Perspectives, 5, 111-132.

98.

Walls

Macauley

Anderson

(2005). Private markets, contracts, and government provision what explains the organization of local waste and recycling markets? Urban Affairs Review, 40, 590-613. doi:10.1177/1078087404273342

99.

Warner

M. E.

Bel

(2008). Competition or monopoly? Comparing privatization of local public services in the US and Spain. Public Administration, 86, 723-735. doi:10.1111/j.1467-9299.2008.00700.x

100.

Williamson

O. E.

(1996). The mechanisms of governance. Oxford, NY: Oxford University Press.

101.

Zafra-Gómez

J. L.

López-Hernández

A.M.

Hernández-Bastida

(2009a). Evaluating financial performance in local government. Maximising the Benchmarking value. International Review of Admiinistrative Science, 75, 151-67. doi: 10.1177/0020852308099510.

102.

Zafra-Gómez

J.L.

López-Hernández

A.M.

Hernández-Bastida

(2009b). Developing a model to measure financial condition in local government. Evaluating service quality and minimizing the effects of the socioeconomic environment: an applicationto Spanish municipalities. American Review of Public Administration, 39, 425-449. doi: 10.1177/0275074008320710.

103.

Zafra-Gómez

J. L.

López-Hernández

A.M.

Hernández-Bastida

(2009c). Developing an alert system for local governments in financial crisis. Public Money & Management, 29, 175-181. doi: 10.1080/09540960902891731.

104.

Zafra-Gómez

J. l.

Pérez-Muñiz

M.A.

(2010). Overcoming cost-inefficiencies within small municipalities: improve financial condiciton or reduce the quality of public services? Environment and Planning C: Government and Policy, 28, 609-629. doi: https://doi.org/10.1068/c09118

105.

Zafra-Gómez

J. L.

Pedagua

L. E.

Plata-Díaz

A. M.

López-Hernández

A. M.

(2014). Do local authorities use NPM delivery forms to overcome problems of fiscal stress? Spanish Journal of Finance and Accounting/Revista Española de Financiación y Contabilidad, 43, 21-46.

106.

Zafra-Gómez

J. L.

López-Hernández

A. M.

Plata-Díaz

A. M.

Garrido-Rodríguez

J. C.

(2016a). Financial and political factors motivating the privatisation of municipal water services. Local Government Studies, 42, 287-308. doi:10.1080/03003930.2015.1096268

107.

Zafra-Gómez

J. L.

Plata-Díaz

A. M.

Pérez-López

López-Hernández

A. M.

(2016b). Privatisation of waste collection services in response to fiscal stress in times of crisis. Urban Studies, 53, 2134-2153. doi:10.1177/0042098015586697

108.

Zucker

L. G.

(1987). Institutional theories of organization. Annual Review of Sociology, 13, 443-464. doi:10.1146/annurev.so.13.080187.002303

109.

Zullo

(2009). Does financial stress induce privatization? Correlates of private and intermunicipal contracting, 1992-2002. Governance: An International Journal of Policy, 22, 459-481. doi:10.1111/j.1468-0491.2009.01447.x