Austerity, political control and supplier selection in English local government: implications for autonomy in multi-level systems

ABSTRACT Analysis of 60,000 contracts awarded by English councils between 2015–19 reveals that austerity constraints are a key predictor of councils outsourcing services to for-profit suppliers, regardless of their political control. Conservative Party-controlled councils are also more likely to contract with for-profit suppliers, although we found no link between Labour-controlled councils and not-for-profit suppliers, nor evidence that political or budgetary factors influence whether councils contract with providers based in their own region. We argue that centrally imposed funding cuts, and a belief that for-profit suppliers represent a cheaper option, could be overriding Labour Party councils’ ideological preference for not-for-profit providers.


Introduction
In recent decades, this journal has highlighted the numerous ways in which public bodies have developed and adopted different service delivery models (Tavares and Camões 2010;Allen et al. 2016;Isaksson, Blomqvist, and Winblad 2018;Mazzei et al. 2020;Warner, Aldag, and Kim 2020). Studies have found that the default approach is now to outsource services in many countries, rather than rely on in-house provision (Nemec, Merickova, and Vitek 2005;Guo and Ho 2019). This has led scholars to investigate the determinants of contracting-out, especially in a local government context Fageda 2007, 2017). From these studies we know that political, financial and institutional factors are all relevant in explaining the decision to outsource services like social care, healthcare, education and transport. By contrast, we know very little about the factors that influence whom public bodies select as their outsourcing partner from the range of for-profit and not-for-profit suppliers available, or whether decisions over supplier choice are influenced by political or financial factors. We examine local government outsourcing in England between 2015 and 2019 to shed light on this question.
Using local government as the research context is appropriate for several reasons. Decisions over contractor selection are more likely to be affected by political factors at the local level because councillors are close to the operational frontline and able to directly communicate with procurement officers (Murray 2007;Gelderman, Semeijn, and Bouma 2015). Moreover, since the Byatt Review of local government in England, councillors have been actively encouraged to help formulate procurement strategies, oversee high-value projects and ensure democratic accountability in procurement decision making (Byatt 2001). There is also evidence showing that political party control of councils is deterministic of sourcing decisions. Specifically, leftist councils appear to prefer in-house service provision and rightist councils outsourcing (Dubin and Navarro 1988;Tavares and Camões 2010;Elinder and Jordahl 2013;Petersen, Houlberg, and Christensen 2015;Alonso and Andrews 2020). These observations are consistent with partisan politics theory (Hibbs 1977(Hibbs , 1992, which predicts that the decisions taken by political parties reflect their supporters' interests and ideology. Previous studies focus largely on the factors that determine a public body's initial decision to outsource. In this paper, we investigate whether political ideology might explain the type of supplier that English councils select, after they have decided to contract out a service. Type of supplier has two manifestations in this study. The first is for-profit versus not-for-profit suppliers. Despite the expansion of the supply marketplace to include third sector, public sector and public-private sector actors capable of delivering services on behalf of local government (Stein 1990;Boggio 2016;Mazzei et al. 2020), there is a lack of research into how partisan politics may affect supplier selection decisions. Our second manifestation of supplier type is whether providers are based in the same region of England as the council. Although scholars increasingly suggest that local government procurement can play an important role in supporting regional enterprise eco-systems (Preuss 2009;Walker and Brammer 2009;Thompson et al. 2020), there have been few large-N studies examining how and why public bodies may be engaged in this activity.
In terms of English councils' preferences for supplier types, we anticipate that the ideological instincts of Labour Party-controlled councils will lead them to select notfor-profit providers and the ideological instincts of Conservative Party-controlled councils will lead them to select for-profit suppliers. These positions reflect what partisan political theory says about parties taking policy decisions that advance the interests of their supporters (Hibbs 1977(Hibbs , 1992. We also anticipate that where councils are politically in-the-balance i.e., not controlled outright by either the Labour or Conservative Party, the probability of contracts going to same region suppliers increases. This is because politicians are more likely to use public contract awards to win the support of voters who own, manage or work for local suppliers. An analogy can be made with 'spatial selectivity' (Jones 1997) or 'pork barrel' politics, in which central government funding streams disproportionately benefit actors in marginal constituencies (John and Ward 2001;Denemark 2003;Kang 2015).
Resource constraints may restrict the extent to which councils can use contracting as a political tool. Studies have reported that fiscal stress, whether measured by funding cutbacks, budget deficits or financial dependency, is associated with councils' decision to privatize public services (González-Gómez, Picazo-Tadeo, and Guardiola 2011;Geys and Sørensen 2016;Zafra-Gómez et al. 2016). Using the same logic, councils experiencing fiscal stress are likely to favour suppliers offering the lowest prices, regardless of which political party is in control. The procurement literature shows that such circumstances will advantage large, for-profit suppliers (Loader 2007;Cabras 2011;Young, Nagpal, and Adams 2016). Where this fiscal stress is caused by reductions in central government grants, it could mean that councils are adopting procurement strategies that do not fit with their political objectives. This points to an underresearched way in which central government actors may be indirectly restricting local political autonomy in multi-level systems, alongside more visible constraints like performance monitoring frameworks or regulations.
To examine the determinants of supplier selection decisions, our paper analyses 60,000 service contracts awarded by 150 English councils, beginning in 2015 (when online publication of contracts awards became mandatory) and ending in 2019. This research design maximized the size of the dataset, provided information on contracting behaviour that overlapped with local government electoral cycles and brought our political control variable into play. Overall, the results offer qualified support for partisan politics theory while also indicating that budgetary cutbacks can override ideological influences on contract awards. A regional dimension to contract awards also emerges from the analysis, suggesting that the dynamics of the supply marketplace impact council procurement decisions independently of political and financial factors.
Our paper makes three key contributions to the public contracting literature. First, it extends the body of work on the determinants of outsourcing decisions (Elinder and Jordahl 2013;Petersen, Houlberg, and Christensen 2015;Schoute, Budding, and Gradus 2018) by examining how political and budgetary conditions lead councils to select for-profit versus not-for-profit contractors. This is significant: for many public bodies the question is less about whether they should outsource and more about which suppliers they should contract with (Alonso, Andrews, and Hodgkinson 2016). Second, it introduces regional sourcing as a new dependent variable in studying the politics of contracting -one for which the ideological and financial antecedents have yet to be empirically tested. Third, by focusing on the impact that fiscal constraints have on sourcing decisions our paper addresses a fundamental question: do central government austerity policies result in subnational bodies adopting procurement approaches that they might otherwise not endorse?
The paper proceeds as follows. The next section discusses partisan politics theory and associated studies on political control and austerity in local government contracting. It culminates in four hypotheses we set ourselves for investigation. We then set out our method, followed by the empirical results, a discussion of their relevance for the public management field, and our concluding thoughts.

Theoretical framework -partisan politics
Alternatives to in-house provision of services by the public sector like outsourcing emerged from the 1980s onwards (Savas 1987;Stein 1990). Since then, public administration scholars have interested themselves in the economic, service quality and accountability implications of these alternatives. The verdict on their effectiveness has been mixed. The cost argument for contracting out is appealing and citizens often view it as a fiscally responsible option (Warner and Hefetz 2002;Guo and Ho 2019). At the same time, evidence for outsourcing having a cost advantage is less than conclusive (Boyne 1998;Nemec, Merickova, and Vitek 2005). Similarly, while external providers may have a role to play in addressing social and health care service challenges, research studies caution against viewing them as a panacea for raising quality and accountability standards (Eckersley, Ferry, and Zakaria 2014;Allen et al. 2016;Isaksson, Blomqvist, and Winblad 2018;Mazzei et al. 2020). Apart from critically engaging with outsourcing, administration and management scholars are interested in the political, financial and institutional factors that lead public organizations to opt for particular service delivery models (Gradus and Budding 2020). It is these factors that we focus on in developing our hypotheses, starting with partisan political effects in local government.
In partisan politics theory, there is a direct relationship between the type of party in power -left versus right -and the policies that they implement (Hibbs 1977(Hibbs , 1992Häusermann, Picot, and Geering 2013). The relationship is explained by the constituencies the parties represent. As Hibbs (1992) states, 'parties behave to a significant degree "ideologically", meaning that they promote policies broadly consistent with the objective interests and revealed preferences of their core constituencies'. Traditionally, leftist parties like the Labour Party in the UK have represented workers, trade unions, minorities and less privileged groups. Their policies were designed to advance the interests of these constituent groups, whether through increased public spending, wealth re-distribution or welfare expansion. Rightist parties like the Conservative Party in the UK have traditionally represented capital, business owners and the professional classes. Supportive of these constituents, rightist parties have pursued low taxation, public spending reductions and limits on social insurance protections.
How partisan politics plays out in local government service provision has been subject to extensive inquiry. Much of this work has concentrated on the fiscal, economic, political and ideological antecedents of councils deciding to either keep operations in-house or outsource them to an external provider Fageda 2007, 2017). Partisan theory implies that leftist parties will want to keep operations in-house, due to their ideological commitment to state ownership and control of public services. Rightist parties will be open to outsourcing, as it aligns with their belief in the primacy of market-based solutions to public service challenges. Moreover, whereas state ownership benefits leftist constituencies like workers and trade unions, privatization benefits rightist constituencies like investors and business owners; and as partisan theory makes clear, political parties act in accordance with the 'economic interests and subjective preferences' of their constituencies (Hibbs 1977(Hibbs , 1468. Evidence for the influence of partisan politics on local government outsourcing decisions is mixed. Supportive of partisan theory, some studies have demonstrated that rightist-controlled councils are inclined to outsource services (Dubin and Navarro 1988;Tavares and Camões 2010;Elinder and Jordahl 2013;Schoute, Budding, and Gradus 2018) and leftist-controlled councils are inclined to retain in-house provision (Petersen, Houlberg, and Christensen 2015;Zafra-Gómez et al. 2016;Alonso and Andrews 2020). Against theoretical predictions, however, other studies have returned no empirical support for the argument that ideology acts as a determinant of outsourcing (Alonso, Andrews, and Hodgkinson 2016;González-Gómez et al. 2011). This led Bel and Fageda (2007, 528) to surmise that 'local governments are guided by pragmatic rather than ideological motivations'. Yet, a meta-analysis undertaken by the same authors ten years later detected an increase in support for partisan theory and its thesis that ideology explains decisions by councils to outsource services or keep them in-house (Bel and Fageda 2017).
Less attention has been paid to the effect of partisan politics on choosing for-profit versus not-for-profit contractors, after an organization has opted to outsource. Yet politics matters here too, because 'once the decision to contract out has been taken, there remain important political and economic influences at play that ultimately impact the outcome of the "buy" decision' (Alonso, Andrews, and Hodgkinson 2016, 258). In line with Picazo-Tadeo et al. (2012), we expect Conservative-controlled councils to incline towards for-profit suppliers and Labour-controlled councils to incline towards not-for-profit suppliers. The prospect of exposing public services to the competitive rigours of the private sector marketplace should be ideologically appealing to Conservatives and their supporters. The opposite holds for Labour politicians, where outsourcing to public or third sector contractors would represent the lesser of two evils. The fact that public and third sector suppliers are not profitdriven and espouse values rooted in citizen wellbeing and community cohesion (Mazzei et al. 2020) makes them a more natural fit for left-wing parties. These arguments lead to the following hypotheses: H1: Conservative-controlled councils are more likely to contract with for-profit suppliers.
H2: Labour-controlled councils are more likely to contract with not-for-profit suppliers.
Political control is also likely to affect decisions on contracting with same region suppliers. Studies into pork-barrel infrastructure projects suggest political actors direct public resources to specific constituencies (John and Ward 2001;Denemark 2003;Kang 2015) through a strategy of 'spatial selectivity' (Jones 1997). This research examines central government distributive policies -how funding is distributed to subnational institutions and for infrastructure projects -rather than patterns of local government spending with service providers. Nonetheless, we might assume that local politicians are similarly keen to ensure that suppliers based in their constituencies are awarded public contracts, notwithstanding the structures that may be in place to combat corruption (Neu, Everett, and Rahman 2015;Auriol, Straub, and Flochel 2016). Indeed, Walker and Brammer (2009) found that councils are more likely than other public bodies to buy from local suppliers. Furthermore, cities including Cleveland, Barcelona, Bologna and Preston have instituted policies that prioritize local sourcing (Thompson et al. 2020). This is perhaps unsurprising for reasons of convenience and to avoid the costs of transporting goods and services over long distances. Nonetheless, it has more sinister undertones. Kyriacou and Roca-Sagalés (2020) cautioned that decentralizing procurement spend can lead to increased corruption because special interests exert greater influence over decision-making in smaller units of government.
Empirical research in this area focuses largely on the US and, thus far, has not examined in any great depth links between political control and the awarding of public contracts to local suppliers, particularly using large datasets. Even procurement-based studies on local and regional sourcing are apolitical and ignore ideological influences on 'buy local' policies (see, for instance, Preuss 2009;Walker and Brammer 2009;Cabras 2011). Based on the pork-barrel principle, however, we might expect councils that are politically in-the-balance to contract with local and regional suppliers to a greater degree than councils that have been controlled by the same political party for a lengthy period. If we assume that citizens who own, manage or work for local suppliers are more likely to vote for politicians who approve contracts with their organization (Sundell and Lapuente 2012), we can see how the risks of not buying from local providers are higher in marginal constituencies. In line with this perspective, we can view public contracts as a form of patronage that elected representatives bestow on constituents for electoral advantage. These arguments lead to the following hypothesis: H3: Councils that are politically in-the-balance are more likely to contract with suppliers based within their region than councils that are politically stable i.e., controlled by either the Labour Party or Conservative Party.
Financial stress has consistently emerged as a factor associated with councils outsourcing services in European countries Fageda 2007, 2017). Its effect has been observed across various national and service delivery contexts, including infrastructure and support services in Norway (Geys and Sørensen 2016), water services in Spain (González-Gómez et al. 2011;Zafra-Gómez et al. 2016) and maintenance, social and professional services in the Netherlands (Schoute, Budding, and Gradus 2018). Interestingly, Warner, Aldag, and Kim (2020) report contrary results from US councils, showing that financial stress is more likely to be associated with inter-municipal co-operation than privatization. The answer to why financial stress leads to outsourcing lies in the efficiency gains that, at least in theory, result from having market actors compete for the opportunity to deliver services previously provided in-house by councils (Boyne 1998). This theory holds that competitive forces drive down the cost of delivering the service, thus making outsourcing a financially attractive option relative to in-house production (ibid). Therefore, it is not surprising that councils under financial strain view outsourcing as a remedy to their predicament.
As well as increasing the chances of outsourcing, it is possible that financial stress influences the type of suppliers selected by councils. If a council prefers to contract with third or public sector suppliers instead of private companies, or if a council prefers same-region providers over all other providers, it will restructure the market in a way that could lead to price increases. This is because not-forprofit providers are typically smaller than private operators and struggle to compete with them on price due to the latter's economies of scale advantage (Boyne 1998). It is also because favouring particular types of supplier limits the number of bidders and, ultimately, the competitiveness of bid prices. Councils under financial pressure, regardless of their ideological instincts, would want to avoid these outcomes. As such, we might expect them to contract more with for-profit suppliers and look beyond the regional marketplace in the interests of obtaining the lowest price possible. Findings from procurement studies aver to this happening, where public bodies under budgetary pressures end up favouring large firms over local SMEs because achieving short-run savings is paramount (Loader 2007;Preuss 2009;Young, Nagpal, and Adams 2016). These arguments lead to the following two hypotheses: H4a: Councils that have experienced significant fiscal constraint are less likely to have contracted with same-region suppliers than councils that have experienced less fiscal constraint.
H4b: Councils that have experienced significant fiscal constraint are more likely to have contracted with profit-making suppliers than councils that have experienced less fiscal constraint.

Research context
We decided to test these hypotheses in the context of English local government for several reasons. First, councils in England are responsible for the provision of services such as education, transport, social care, housing, libraries, environmental health and waste management. These are the types of services eligible for contracting out, alongside corporate functions such as human resources, finance and payroll. We examined the contracts of all 150 'top-tier' local authorities, excluding the 192 smaller district councils that sit 'below' the county level. These 150 bodies are made up of 55 unitary authorities (which cover a mixture of urban and rural areas but sit outside county boundaries), 32 London Boroughs (councils that sit within the Greater London area), 36 metropolitan districts (councils located in conurbations outside the capital) and 27 county councils ('upper tier' authorities in rural counties). This approach meant that we studied elected bodies across the entire geographical territory of England.
Second, between 2015 and 2019, all bar one of these 150 councils fell into one of three categories: controlled by the Conservative Party; controlled by the Labour Party; not solely controlled by either the Conservative Party or Labour Party. This makes it straightforward to categorize councils by political control. Third, between 2010 and 2018 there was a £16bn (49%) reduction in central government funding to local authorities and the UK government has said that it will phase out these grants entirely by 2025 (National Audit Office 2018a). In the financial year 2017-18, approximately 48% of local government's total income came through these grants, which underscores the significance of these cuts (Ministry of Housing, Communities and Local Government 2019). In addition, because more deprived areas are unable to raise as much revenue locally from business and domestic property taxes, they are much more dependent on central funding than wealthier municipalities. This meant that cuts in grants were implemented asymmetrically across England, with poorer, often Labourcontrolled, municipalities experiencing deeper reductions as a percentage of their overall budgets than other areas (Gray and Barford 2018; Atkins 2020). As such, we might expect the differentiated nature of austerity to have had a concomitant impact on contracting decisions.
Another reason for researching English councils is that successive governments have encouraged them to open-up public services to alternative providers for several decades, and particularly since 2010 (Gov.uk 2011;Eckersley and Ferry 2020). We anticipated that this policy would result in numerous examples of councils contracting out services to different types of supplier. In 2011, the UK government removed a swathe of regulatory reporting requirements and gave councils additional legal freedoms such as a 'power of general competence' as part of its localism agenda. In line with this policy, Ministers did not require local government to put services out to competitive tender. However, the increasing financial constraints within which councils operated may have resulted in them relying more on for-profit providers regardless of their political persuasion, in the belief that these suppliers would provide services at lower cost.
We limited our study to service contracts, which is consistent with the extant body of work on the determinants of outsourcing. Also, political ideology is arguably less relevant when it comes to procuring works and supplies. Unlike services, these have rarely been provided in-house and local authorities have traditionally depended on the private sector to meet these needs. Moreover, not-for-profit product suppliers and works contractors are rare. Our data bear this out. We found that English councils awarded 99% of works and products contracts to profit-making suppliers between 2015 and 2019. The corresponding figure for services was 86.5%. It is also significant that council expenditure on service contractors increased throughout England during this period, even as spending on supplies and product categories has fallen. In 2019, it stood at £30.79bn out of the £61.59bn total, which underlines the growing presence of external contractors in local government (Ministry of Housing, Communities and Local Government 2019).

Method
Information on the dependent, predictor and control variables used in this study is given below and summarized in Table 1. The data were sourced from Contracts Finder, Office for National Statistics (ONS), National Audit Office (NAO), council and media websites. Our study includes all types of services that English councils outsourced between 2015 and 2019. This is consistent with previous research that has examined the outsourcing of multiple service categories (Geys and Sørensen 2016; Schoute,

Same Region Suppliers
Total percentage of service contracts awarded by the local authority to suppliers based in their own region between May 2015 -March 2019.

Scale Tussell
Budding, and Gradus 2018; Guo and Ho 2019;Warner, Aldag, and Kim 2020), although other studies confine their analysis to a single service category (Alonso, Andrews, and Hodgkinson 2016;Elinder and Jordahl 2013;Zafra-Gómez et al. 2016). Including all services categories improves the generalizability of what we can say about the determinants of outsourcing. It is also important because the effect of fiscal stress on a single service category is likely to be marginal (Bel and Fageda 2007). Typical of studies in this area, our dataset stretches over several financial years and represents the entire population of councils within the country of interest (Tavares and Camões 2010;Geys and Sørensen 2016).

Dependent variables
Data on the dependent variables was obtained from Contracts Finder, which is the official web portal for public contracts in the UK (https://www.gov.uk/contractsfinder). Since May 2015, all UK public bodies have been required to publish details of contract awards on Contracts Finder if their value exceeds thresholds of £10,000 for central government bodies or £25,000 for local authorities and other public sector organizations. We commissioned Tussell, a company that harvests, repurposes and sells data for market intelligence purposes, to provide us with details of every contract awarded by UK local authorities between May 2015 and March 2019. The resultant Excel dataset comprised over 120,000 individual contracts. This reduced to approximately 60,000 after removing contracts that did not relate to services or English councils. Each line of data generated by Tussell contained information on council name, council address, contract description, awardee name, awardee address, awardee legal form and date when contract award was made. We used the dataset to create two dependent variables for the study [1]. The first variable relates to councils selecting for-profit or not-for-profit suppliers. For-profit suppliers included private limited companies, public limited companies, unlimited companies and limited liability partnerships. These were put in one group. Not-forprofit suppliers included charities, central or local government bodies, community interest companies, housing associations, education institutes and the National Health Service (NHS). These were put in a second group. For each council we summed the number of contracts that went to for-profit suppliers between 2015 and 2019, and calculated this figure as a percentage of the total number of contracts awarded. We did the same for not-for-profit suppliers. The second dependent variable relates to same region suppliers. We were able to identify if suppliers were in the same region as the council by examining their addresses, which were contained in the Tussell dataset. As before, we summed the number of contracts awarded to same region suppliers and then calculated this figure as a percentage of the total number of contracts awarded between 2015 and 2019. A similar approach to dealing with service contract awards is used in Warner, Aldag, and Kim (2020), except in their case the split is between privatization versus inter-municipal cooperation.

Predictor variables
The predictor variables in this study are political control and fiscal constraint. Political control is measured by reference to one of three categories: Conservative-controlled councils, Labour-controlled councils and No Single Party-controlled (NSP) councils.
Conservative and Labour are the two dominant political parties in the UK and approximate to the left-right dichotomy used when testing the effect of partisan politics on contracting decisions (Picazo-Tadeo et al. 2012;Petersen, Houlberg, and Christensen 2015;Schoute, Budding, and Gradus 2018). The breakdown of political control for English councils is as follows: 29.3% for Conservative (n = 44), 46% for Labour (n = 69) and 25% for NSP (n = 37). NSP councils were not under the sole control of either Conservative or Labour throughout 2015-19 (included in this category is the London Borough of Sutton, the only council controlled by a different party -the centrist Liberal Democrats -between 2015 and 2019).
Budget reductions, income-expenditure ratio, cash solvency and financial dependency are among the measures of fiscal constraint that can be found in the outsourcing literature (Tavares and Camões 2010;Boggio 2016;González-Gómez et al. 2016;Schoute, Budding, and Gradus 2018;Warner, Aldag, and Kim 2020). Our preferred measure is the percentage reduction in a council's 'spending power' during the period 2010/11 to 2017/18 -this is the amount of money at its disposal for annual revenue expenditure, generated from a combination of government grants and local taxes (Atkins 2020). Spending power is a more meaningful indicator of fiscal constraint than funding cuts because councils in England rely on central grants to different extents and their capacities to raise revenue locally through council tax, business rates, fees and charges vary significantly (National Audit Office 2018a; Gov.uk 2019; Atkins 2020). The National Audit Office's (NAO) financial visualization tool was used to extract information on cumulative spending power reductions between 2010-11 and 2017-18 for each council (National Audit Office 2018b). This tool draws on the UK government's calculations of each individual council's 'spending power': an indicator that takes account of income the authority receives from central government alongside its ability to raise revenue locally.

Control variables
Using data from the Office for National Statistics (ONS) Nomis database (www. nomisweb.co.uk), we control for earnings, job density, population and region in this study. Reflecting the prerogatives of their constituents, we expect councils where earnings and job density are low to favour not-for-profit service providers. Lower income households are known to be resistant to private sector involvement in public service provision and, as a corollary, more accepting of public or third sector providers (Warner and Hefetz 2002;Alonso, Andrews, and Hodgkinson 2016;Guo and Ho 2019;Warner, Aldag, and Kim 2020). This can be attributed to concern about future price hikes by for-profit providers and the unaffordability of the service for poorer citizens. Building on insights from the sustainable procurement literature (Preuss 2009;Cabras 2011), we expect the same low income, low job density councils to prioritize regional suppliers as part of their economic regeneration strategies. Procurement is among the most effective instruments poorer councils possess for directing assistance to local firms and stimulating multiplier effects in the local economy.
Studies have demonstrated that councils with big populations are more likely to outsource water and utility services (Picazo-Tadeo et al. 2012;González-Gómez et al. 2011;Zafra-Gómez et al. 2016). This is because they often require assistance from the private sector in providing services (ibid). Additionally, private sector contractors will be more willing to provide services to large councils because of economies of scale benefits (ibid). Arguably, council size also determines the type of contractor selected. Councils with big populations are more likely to perceive that for-profit corporations have the organizational capacity, technical expertise and client management capabilities to deliver services on their behalf compared to the third sector. If this is the case, the same councils may have to look beyond the local or regional marketplace to find corporations big enough to service their populations. We therefore control for any spatial dimension to supplier selection decisions with a Region of England variable. Its inclusion is necessary as Alonso, Andrews, and Hodgkinson (2016) and González-Gómez et al. (2016) showed that geographic proximity helped explain outsourcing decisions in England and Spain, respectively. The fifth and final control variable is the total number of contracts issued by each council.

Analytical approach
The dataset was analysed using linear regression in IBM SPSS Statistics. This meant testing the determinants of contract awards in the case of for-profit suppliers (model 1a-b), not-for-profit suppliers (model 2a-b) and same region suppliers (model 3a-b). The predictors and controls consist of categorical variables (political control, region) and scale variables (budget reduction, earnings, job density, population and total contracts). The control variables were entered first, followed by the predictor variables. Prior to running the regression, we carried out a series of diagnostic checks. These confirmed that multi-collinearity was not present. The highest Variance Inflation Factor (VIF) was 4.13 and the lowest Tolerance Value was .242. These are under accepted thresholds. Correlation coefficients for all variables are reproduced in Table 2.

Results
The three models tested are statistically significant and account for between 10% and 19% of the Adjusted R variance in contractor selection decisions. The results return some support for our hypotheses (see Table 3). H1 predicted that Conservativecontrolled councils are more likely to award service contracts to for-profit suppliers. This hypothesis is supported. The effect of Conservative controlled councils on selection of for-profit suppliers is statistically significant and positive (β = .292, p < .01). H2 is not supported, however. Labour-controlled councils are not more likely to engage not-for-profit service providers. In fact, the negative standardized co-efficient value indicates that Labour-controlled councils are marginally less likely to select not-for profit service providers than councils controlled by Conservatives or that are in-thebalance i.e., NSP councils.
These results provide partial validation for our contention that there is a partisan element to contract award decisions, as partisan political theory suggests (Hibbs 1977(Hibbs , 1992. Conservative-controlled councils are found to gravitate towards for-profit providers, which is consistent with their ideological belief that free markets represent the most efficient means of allocating scarce resources. Moreover, it is consistent with the interests of their core constituencies. The behaviour of Labour-controlled councils, on the other hand, is inconsistent with their ideological aversion to the privatization of public services and with the interests of their core constituencies. When outsourcing service contracts, we expected them to incline towards not-for-profit providers like charitable organizations, community enterprises and public corporations. This proved  68 .12 .18 .19 .19 .25 .08 .17 .14 .15 .22 .19 .17 .13 .16 .14 .17 1 Correlations at or above .16 are statistically significant at p < .05.
not to be the case. The reason may lie in the fiscal constraints under which they have operated, which we discuss further on.
H3 predicts that councils in-the-balance are more likely to award contracts to suppliers in their regional catchment area i.e., same region suppliers. No support is returned for H3. While NSP councils award more contracts to suppliers within their own region compared to Conservative or Labour-controlled councils, the effect is not statistically significant. Arguably, our use of same region suppliers as beneficiaries of 'pork barrel' politics is too geographically diffuse and would need to be limited to suppliers situated within local council boundaries. Unfortunately, however, UK postcode identifiers do not directly correspond to council boundaries and sometimes cover several councils or parts thereof. As a result, we were not able to determine if contract awardees were based inside or outside council boundaries based on automated text analysis. Instead, we could only identify whether the awardee was in the same region as the awarding council. Fiscal constraint is not statistically significant in explaining the award of contracts to same region suppliers, either. H4a is rejected on this basis.
However, alongside party political control, fiscal constraint emerges as a statistically significant predictor of the likelihood of councils awarding contracts to for-profit suppliers. The effect is positive, meaning councils that have suffered higher reductions in spending power are more likely to contract with for-profit suppliers (β = .297, p < .01). While there is no statistically significant relationship between fiscal constraint and awarding contracts to not-for-profit suppliers, the standardized coefficient value is, as expected, negative. Taken together, these findings support H4b and imply that financial as well as ideological factors are relevant in explaining why councils award contracts to for-profit suppliers. We contend that this is because for-profit suppliers are perceived to be a cheaper option, and contracting with them would thereby enable councils to contain or even reduce their expenditure (Picazo-Tadeo et al. 2012;Geys and Sørensen 2016;Gonzalez-Gomez et al. 2011;Zafra-Gómez et al. 2016;Schoute, Budding, and Gradus 2018;Warner, Aldag, and Kim 2020). Our analysis confirms Gray and Barford (2018) finding that English councils are affected by fiscal constraints to varying degrees. Labour-controlled councils experienced a 33% reduction in spending power over the study period, as against 27% for NSP councils and 23% for Conservative councils; and these mean differences are statistically significant (p < .001). The fact that Labour-controlled councils are operating under tighter financial constraints probably explains why we do not see them awarding contracts to third and public sector suppliers in the numbers expected. Relative to private sector suppliers, the up-front costs of contracting with not-forprofit suppliers tend to be higher. Even though Labour-controlled councils have a partisan preference for not-for-profit suppliers, the expedient option is to award contracts to cheaper suppliers from the private sector when funds are limited. In this way, economic necessity takes precedence over the Labour Party acting on their ideological convictions. To a large extent, this situation is a consequence of central government-imposed funding cuts. Whether by design or otherwise, these cuts are limiting procurement autonomy for Labour councils and creating a situation where private sector operators become suppliers by default rather than by choice. We explore the full implications of this finding in the next section.
The demographic variables have mixed effects on supplier selection decisions. Socio-economic characteristics are not statistically significant. Specifically, lower earnings and job density do not increase the likelihood that councils award contracts to not-for-profit suppliers or same region suppliers. Our a priori assumption was that poorer communities would prefer not-for-profit service providers because they perceive them to be less threatening to their economic and social interests. However, this assumption does not hold. If anything, communities in England that enjoy higher wages and plentiful job opportunities are more likely to have not-for-profit service providers working on behalf of their councils. It could be that affluent communities are open to the idea of third sector suppliers in the belief that they are sustainability-led and practice organizational citizenship. Others have found that affluent citizens appear more comfortable with the marketization of public services (Guo and Ho 2019; Warner, Aldag, and Kim 2020), and we extend these studies.
There is a strong spatial dimension to the destination of service contracts. Councils in six of the eight England regions are statistically less likely to select same region suppliers than councils in the London region. The two exceptions are North West and West Midlands regions. Explaining this occurrence, we can say that London councils form part of the biggest urban conurbation in England and are surrounded by thriving service and industrial sectors. This makes sourcing from local suppliers an eminently feasible strategy for them. Similarly, councils in the North West and West Midlands region of England have the benefit of choosing from a wide pool of service providers. West Midlands is home to Birmingham, England's second biggest urban conurbation, while the North West is home to Greater Manchester and Merseyside (including Liverpool), England's third and fifth biggest urban conurbations respectively. These findings lead us to infer that local sourcing is influenced by external marketplace factors, particularly supplier availability and local enterprise system dynamics, rather than by the political make-up of the council or its financial health.

Discussion
Our findings echo those of others who have found a link between political control of councils and outsourcing (Dubin and Navarro 1988;Tavares and Camões 2010;Elinder and Jordahl 2013;Petersen, Houlberg, and Christensen 2015;Alonso and Andrews 2020). This research has found that centre-right councils are more likely to outsource services than centre-left authorities. By analysing a comprehensive dataset of all service contracts agreed by top-tier English councils over a four-year period, we take these studies a stage further to demonstrate that, once a decision to outsource has been taken, centre-right, Conservative-controlled, councils are more likely to choose forprofit suppliers than non-profit providers. As such, we strengthen the case that political ideology and control of the council plays an influential role in local government procurement and contracting decision-making. However, we did not find that Labour-controlled councils were more likely to agree contracts with voluntary and public sector suppliers. A key reason for this may well be because Labour councils have been disproportionately affected by fiscal constraint since 2010, and we found strong correlations between reductions in spending power and a preference for profit-making providers.
Our study also takes previous research into the links between austerity and contracting out a step further. This literature has shown that fiscal constraint leads to more outsourcing of local public services (see, for example, González-Gómez et al. 2011;Geys and Sørensen 2016;Zafra-Gómez et al. 2016;Schoute, Budding, and Gradus 2018;Guo and Ho 2019;Warner, Aldag, and Kim 2020). We add to this by demonstrating that austerity also means councils are more likely to select for-profit providers once they have taken the initial decision to contract out. This throws up questions about the extent to which councils that are subject to external funding constraint feel that they can exercise autonomy in supplier selection. The UK Government made its preference for 'opening up' public service markets clear in 2011, but did not require councils to contract out services. Indeed, it gave English councils greater legal autonomy by removing performance monitoring frameworks and granting them a general power of competence as part of its 'localism' policy. However, given that the Conservative Party has formed every UK Government since 2010, and Conservative finance ministers introduced significant reductions to local authority funding during this period, we can see how these austerity cuts have led to councils' procurement decisions being more in line with centre-right ideological preferences on the size of the state and the role of private companies in providing public services.
In other words, ministers may be achieving their ideological objective by using austerity and budgetary policy -rather than regulation -to create incentives to outsource local services to the private sector. This is particularly the case for Labourcontrolled councils, which we would otherwise expect to prefer to contract with notfor-profit suppliers (if indeed they wish to outsource services at all). Labour councils in more deprived areas have been disproportionately affected by austerity cuts (Gray and Barford 2018;Atkins 2020), and we suggest that their financial situation led them to contract more with for-profit suppliers that then would have liked. This is because they did not have the budgetary 'slack' to take political or ideological considerations into account in procurement decisions and believed that private providers would be cheaper than non-profit suppliers. Despite the UK Government's rhetoric of localism, therefore, its budgetary policies may have undermined the ability of councils to take autonomous procurement decisions.
Critiques of central control over English councils have largely focused on ministerial attempts to incentivize and monitor them through target-based frameworks such as Comprehensive Performance Assessment (Pratchett 2004) and, subsequently, Comprehensive Area Assessment. This regulatory approach was abolished in 2011, but our study suggests that fiscal policies could also result in councils introducing strategies that fit with the central government's ideological preferences. Such an approach is probably less visible than regulation, but could have more significant implications in asymmetric multi-level systems in the UK and elsewhere -whether it forms part of a deliberate governing strategy or not. Indeed, it is somewhat ironic that UK government ministers initially encouraged councils to incorporate local political priorities into procurement decisions, but then introduced fiscal constraints that has led to them adopting a particular sourcing approach.
Previous research has highlighted how local authorities have responded to austerity cuts by becoming more reliant on external providers (Bel and Fageda 2017). Our study adds to this literature by demonstrating that councils experiencing significant funding reductions are also more likely to contract with for-profit suppliers than not-for-profit providers, even where this may conflict with local ideological priorities. By highlighting funding cuts as a mechanism through which central government could shape local procurement choices, these findings have wider relevance for scholars of multi-level governance, central control and budgetary decision-making, particularly in countries where subnational tiers are heavily reliant on central funding streams. Further research in this area is necessary, particularly to identify whether and how central actors seek to use fiscal policy constraints to direct subnational governments towards specific procurement and sourcing strategies.
To return to H3, councils that were politically in the balance during our period of analysis (not under the sole control of either the Conservative or Labour party between 2015-19) were not more likely to contract with suppliers based in their regions. Therefore, despite concerns that special interests can increase their influence over public spending when procurement is devolved (Kyriacou and Roca-Sagalés 2020), this does not appear to be particularly prevalent in English local government. Nonetheless, councils located in more densely-populated parts of England (the North West, the West Midlands and -particularly -Greater London) are more likely to source services from suppliers in their regions than those based elsewhere. We suggest that this is partly because they can: public service providers are more likely to be located in larger conurbations than rural areas. As such, external marketplace factors such as supplier availability and market dynamics may be more influential in sourcing strategies than internal factors such as political control or fiscal constraint. We would encourage further studies into this phenomenon, in the UK and other jurisdictions, where relevant data are available.
We would also encourage more research into supplier preferences in public bodies that are controlled by different political parties. Previous studies have often focused on the decision to outsource, rather than a preference for non-profit or for-profit providers, and it may well be that research into the latter in other jurisdictions produces more meaningful results. This is particularly the case at the subnational level, where single-country, large-N studies into the effects of political control are more feasible and could reveal more significant differences between political parties. Finally, studies that focus on the percentage of overall spending that is allocated to private and non-profit suppliers, rather than the share of contracts that are awarded to different types of provider, would also increase our knowledge of how procurement decisions may be influenced by political and strategic factors.
There are limitations to our study, which we acknowledge. Our examination of service contracts in aggregate, while comprehensive, means that possible differences in service categories are obscured (for example, Petersen, Houlberg, and Christensen 2015 show how this plays out in their comparison of social and technical services in Danish municipalities). This is important because party ideology is, arguably, more salient in areas like healthcare and social care than transport or recreation. Building on category specific studies by Alonso, Andrews, and Hodgkinson (2016) and Alonso and Andrews (2018), we encourage future studies to investigate if different service categories are affected by political, budgetary and socio-economic factors to varying degrees. Similarly, we only examined services where councils had already decided to employ a contractor to undertake the work. This has methodological benefits, because the data are more readily available: the local authorities concerned have already defined and packaged each individual service in order to put it out to tender and the identity of the contractor is publicly available. However, we did not examine cases where councils opted to keep a service inhouse. The ideological or financial factors that influenced such decisions may reflect our own findings, and those of other studies but further research is necessary in order to confirm this. Finally, it was also not possible to ascertain if the contract awardee was based inside or outside the council area. Studies with fewer contracts under examination than ours could manually check the addresses of awardees to make this delineation.
Our study is also missing qualitative insights into the relationship between politics and public contract awards. There is scope for researchers to undertake qualitative work, whether interviews or ethnographic observations, on the dynamics of this relationship, and factors like budget availability and constituent profile that mediate it. Elite interviews with senior officers and elected members could paint a more accurate picture of the reasons why certain councils might adopt specific sourcing strategies, and thereby address more general questions about the extent to which local politics shapes procurement decision-making. In particular, it would be interesting to examine how officer-member relationships play out in contexts where elected representatives ultimately endorse a sourcing strategy that conflicts with their ideological preferences. By highlighting the councils that are more reliant on private suppliers, our study has helped to identify potential case studies for such research. Murray (2007) and Gelderman, Semeijn, and Bouma (2015) did examine these issues in their qualitative studies, but similar research is necessary to develop the literature further and build theory around officer-member relations and the role of politics in public procurement.

Concluding thoughts
Investigations into the role that elected representatives play in procurement decision making have been noticeable by their absence (Murray 2007), despite the inherently political nature of public procurement. Instead, studies from Preuss (2009), Walker andBrammer (2009), Cabras (2011) and others besides have treated party politics and ideology as invariant when examining phenomena like sustainable procurement in a local government context. One of the underlying premises of our study was that the political party in control of a local council influences the types of suppliers it chooses to deliver public services. The results lend some support to this hypothesis, while at the same time indicating that budgetary constraints can nullify the effect of political preferences on procurement decision making (see also Bel and Fageda 2017). As such, its empirical contribution lies in extending existing lines of inquiry on the externalization of public services (Tavares and Camões 2010;Elston and Dixon 2019;Andrews et al. 2019;Warner, Aldag, and Kim 2020) and taking greater account of the political and financial factors that may affect public procurement. Theoretically, we have also highlighted how central actors might use fiscal policy to direct subnational bodies towards particular procurement strategies, thereby reducing local autonomy and the extent to which councils can use contracting as a political tool. There is scope for methodological refinement of our study and cross-national comparison, which we encourage researchers to undertake.

Notes
1. Initial analysis revealed that data on the value and duration of contracts were not available in a significant number of cases. We attributed this to the uncertain nature of 'payment by results' arrangements, as well as commercial confidentiality considerations. In addition, some councils agreed significantly more contracts than others, probably because they broke up larger deals into lots. As such, we used neither the share of overall spending on suppliers nor the total number of contracts agreed by a council as dependent variables. Instead, we felt that the percentage of contracts that councils awarded to for-profit versus not-for-profit suppliers and the percentage of contracts awarded to suppliers based in the same region as the council would provide us with the necessary information on spending patterns to conduct our analysis.

Disclosure statement
No potential conflict of interest was reported by the author(s).

Notes on contributors
Peter Eckersley is a Senior Research Fellow at Nottingham Trent University, UK, Research Associate at the Leibniz Institute for Research on Society and Space, Erkner, and an Editor of Local Government Studies. His research focuses on central-local government relations, public policy, climate change, public services, sustainability and public accountability.
Anthony Flynn is a Lecturer in Purchasing & Supply Management at Cardiff University. His research interest is procurement and supply chain management, with particular emphasis on SME participation in public sector supply chains.
Laurence Ferry is a Professor in Accounting at Durham University Business School, a Senior Distinguished Visiting Fellow in Accounting at Rutgers, the State University of New Jersey, and a qualified Chartered Accountant. His research, teaching and consulting covers accounting and accountability with a specific focus on the UK public sector and local government. In 2018/19 he was a prestigious Academic Fellow at the UK Houses of Parliament.
Katarzyna Lakoma is a Research Associate in the Centre for Economics, Policy and Public Management at Nottingham Trent University She has been undertaking a PhD in the area of public management since January 2018, examining the impact of governance changes on accountability in the context of English Fire and Rescue Services.