Elevator pitch
Expenditures on job placement and related services make up a substantial share of many countries’ GDP. Contracting out to private providers is often proposed as a more efficient alternative to the state provision of placement services. However, the responsible state agency has to design and monitor sufficiently complete contracts to ensure that the private contractors deliver the desired quality of services. None of the recent empirical evidence indicates that contracting out is necessarily more effective or more efficient than public employment services.
Key findings
Pros
From a theoretical point of view, contracting out job placement services opens this market up to competition, which might decrease costs compared to the public delivery of such services.
With well-designed contractual arrangements and performance measurement, contracting out might be able to improve job placement, at least for certain population groups.
Contracting out allows the state to expand or reduce service capacity and to hire specialists for particular target groups while avoiding the long-term commitments that are often found in the public sector.
Cons
For the responsible state agency, ensuring the quality of private employment services puts great demands on contract design and monitoring systems, and a successful contracting out process is by no means guaranteed.
Many public employment services have replaced bureaucratic input-based structures with output-oriented performance management in an effort to increase their efficiency.
Rigorous empirical studies for several countries indicate that the public provision of placement services performed equally well or even better than the private provision of such services.
Author's main message
Contracting out job placement services might save costs and provide a service capacity buffer in times of increasing unemployment. However, the responsible state agency must ensure a suitable balance between services and quality, and must carefully monitor and evaluate private provider outcomes. Evidence finds that public employment agencies were at least equally as successful in placing the unemployed as private providers. Experience shows that high performance-based payments for successful job placements increases private provider performance in both the short and long term, while high upfront payments decrease the likelihood of re-employment for certain worker subgroups.
Motivation
In many welfare states, the provision of placement services for job seekers has been an important task for public employment services (PES). Since the end of the last century, the state provision of such services has come under increasing criticism for a presumed lack of efficiency. Thus, many states have enacted reforms involving the contracting out of at least part of the placement services to private providers [1]. In the US, for example, the privatization of welfare services has increased significantly since the 1990s, when individual states were given more autonomy to formulate policies. The most resolute approach has been adopted by Australia, where placement services for all unemployed persons have been tendered out to private providers since 1998. Other prominent examples are the UK and the Netherlands, which started to contract out portions of their placement services at the beginning of this century.
Meanwhile, many countries supplement public placement services with private services. However, few rigorous economic evaluations have been conducted to investigate whether private providers of placement services are in fact more efficient and effective than in-house services provided by the PES.
Discussion of pros and cons
The delivery of social services
One might ask why the supply of social services—such as placement services for job seekers—is not entirely left to the market. The reason is potential market failure in the supply of such services. There are four sources of market failure [2]:
External effects: Social services may generate benefits and costs that are greater than those occurring for individual service recipients. For example, ineffective placement services might increase unemployment, with all kinds of negative side effects on society (e.g. increasing crime rates). Thus, there is a public interest in the provision of effective placement services.
Distributional concerns: If placement services were a purely private endeavor, then the individuals most in need of such services would probably not be able to afford them.
Agency problems: Recipients of the social service might have limited capacities to make choices (e.g. to identify a capable private provider of placement services), and private providers might not necessarily be trusted to act in the recipients’ best interests (in particular, if contractual relations leave room for interpretation). As such, a state agency might represent job seekers’ interests better than they could do themselves.
Informational asymmetries: In many social service areas, recipients may have difficulties judging the quality of services. And, even for a state agency, it might be difficult to measure the performance of a private provider regarding placement services.
In the case of placement services in particular, there is general consensus that externalities and distributional concerns are important. With regard to agency and information problems, the picture is less clear. Several modes of social service provision can be distinguished depending on which of the above market failures are in effect; the three listed below are relevant for the problem at hand [2]:
Privately operated services with regulation and public funding: If job seekers are fully capable of identifying a suitable provider of placement services, if providers only have their clients’ best interests in mind (i.e. no agency problems exist), and if differences in provider quality are easily observable, then placement vouchers can be used.
Public ownership and contracts with private providers for service delivery: If agency problems arise but quality is readily observable, then privatization by means of contracting out is an appropriate way to provide services.
Publicly operated services and delivery: Service provision by a state agency might be preferred, in principle, if all four kinds of market failure co-exist. Even if it might be more costly in the long term, the state may be able to ensure the desired service quality in this way.
Some countries make use of placement vouchers, which job seekers can redeem at the placement agency of their choice; the service firm is then paid by the state agency for successful placements. The Netherlands used this type of scheme from 2004 to 2012; it was referred to as reintegration arrangements and was very popular [1]. In practice, such vouchers are generally used as a complement to public or contracted-out placement services rather than as a substitute. This is likely due to the fact that, even with observable output quality, it is not possible to overcome all agency problems, as mentioned above.
The most frequently used mode of interaction between the public sector and private providers of placement services is, however, the contracting-out of services. Thus the following discussion primarily focuses on these types of setups, rather than on voucher schemes. From a theoretical point of view, contracting out requires that output quality is observable and that contracts can be concluded that are contingent upon output quality. Payment for such services is often made up of a combination of an upfront fixed payment per job seeker and subsequent premiums given for successful placements, though the weight of these components varies significantly between countries.
The contents of contracting out
Contracting out takes place on so-called quasi-markets [3], which (to some extent) replace monopolistic state providers with competitive independent providers. In the context of job placement services, a state agency specifies the tasks to be performed by private firms. Competition for entry into the market (but not within the market) is achieved by means of a bidding process among potential service providers. One or more providers win the right to supply a specific set of services, while a state agency maintains some control over the respective activities. Competition is then suspended until the state initiates a new bidding process.
Contracting practices vary between countries in many respects, including [1]:
Scope: Contracting out might encompass all unemployed persons (as in Australia), selected groups of unemployed persons (as in the Netherlands or the UK), or shares of all or selected groups of unemployed persons (as in Germany). Contracts can also be confined to particular regions within a country. Services might be restricted to placement and counseling, or they may also include elements of training or other services.
Bidding process and design of contracts: Contracts may be short-term or cover very long periods, and the actual remuneration design could encompass several components.
Private contractors: These could be non-profit organizations or for-profit firms. One or several private providers might cover a specific region, while main contractors might be selected to manage and monitor a network of smaller subcontractors (as in the UK).
Potential advantages of contracting out
Efficiency entails that activities are organized in such a way that the costs of providing any given quantity or quality are minimized. Proponents of contracting out argue that it increases efficiency by creating competitive incentives for job placement providers [3]. Furthermore, they presume that private providers react more flexibly to the requirements of the clients and the market, and that they are more innovative than traditional welfare bureaucracies. Whether contracting out increases choice depends on the particular mode of service delivery. If the unemployed are assigned to one particular contractor, without the possibility to switch, then they do not have more choice than they would with a public provider. Voucher schemes, in contrast, provide clients with the option to choose from a number of providers. However, not all job seekers are able to cope with a wide spectrum of opportunities (for instance, due to their lack of knowledge on the merits of different potential providers).
One major argument in favor of contracting out is clearly that it is supposed to save money compared to the public provision of similar services. This does not necessarily require private providers to be able to organize their work more efficiently, set better incentives, or be more innovative than their public counterparts: they might instead use less manpower to provide less intensive services, pay lower wages, or offer fewer benefits to employees than state agencies. A study for the Netherlands analyzed contracting-out decisions made by Dutch municipalities and found that they seem to be driven predominantly by cost considerations—in particular, municipalities with tight budgets have a larger share of private contracting. Hard budget constraints should increase the likelihood of contracting out services, as a poorer government will probably place less emphasis on social goals or the quality of provided services. A further reason for contracting out is that this practice makes it possible for states to rather quickly expand or reduce service capacity and to purchase specialists for particular target groups without engaging in the long-term commitments frequently found in the public sector [1]. Thus, private services can provide a helpful capacity buffer in times of increasing unemployment. Furthermore, additional options for job seekers to take advantage of private instead of public placement services can help to maintain some competitive pressure on case workers in public services.
Conditions for the successful implementation of a quasi-market
Contracting out job search can only achieve the goal of enhancing efficiency—compared to service provision by a monopolistic state agency—if a quasi-market can be successfully established [3]. However, this is no trivial matter.
First, it requires that barriers to market entry be low. If expected profits are small or vary considerably with the state of the labor market, and if market entry involves non-negligible costs, then the level of ex ante competitiveness will be low.
Second, the state agency has to draw up sufficiently complete contracts to ensure that the desired level of access for all groups of unemployed persons is achieved and to make certain that the desired quality of services is delivered. Poor service quality might result if the tendering process places too much emphasis on low-priced bids. Furthermore, incomplete contracts give contractors scope for cutting costs and quality. In particular, service providers might deploy the following strategies:
“Creaming” or “cherry-picking”: A strong emphasis on performance-based payments may encourage private providers to focus on individuals with rather good labor market prospects and—if possible—to refuse services to unemployed job seekers with poor employment prospects. However, this can be avoided if providers are obliged to accept all individuals assigned to them.
“Parking”: A high proportion of upfront payments, which are not contingent upon successful job placement, may encourage service providers to neglect unemployed individuals who are difficult to place and to reduce service quality. This could be mitigated by regular contract monitoring and participant surveys, or by assigning more homogeneous groups of unemployed persons to private providers.
“Gaming”: Service providers might try to exploit further weaknesses in program designs. If, for instance, performance-related bonuses were paid for placing job seekers in employment for a fixed period of time, then there would be an incentive to mainly find jobs for clients for exactly that period of time (instead of permanent positions) and to repeatedly collect the bonus available for each individual assigned to the provider.
Third, monitoring the quality of the provided services is essential for the successful implementation of a quasi-market. Monitoring systems must enable the responsible state agency to track participants and their employment outcomes. However, there is often a lack of information about the effectiveness and efficiency of individual service providers. Gross integration rates (i.e. the rate of job placements) are therefore often used as a proxy for net impact (i.e. the additional share of job placements achieved by the provider, compared to a situation without an assignment to private services). An example of a rather sophisticated approach is the Job Services Australia Star Rating, which uses regression analysis to calculate what providers can reasonably be expected to achieve. Ratings measure relative performance of job placement services by means of two performance indicators taken over the most recent three-year period: (a) the average time taken by placement services to find employment for relevant participants in comparison with other providers, and (b) the proportions of relevant participants for whom placements and outcomes are achieved. These indicators are used by job seekers to assess the comparative performance of job placement services in their local area and by the responsible state agency to allocate business shares to individual providers.
Fourth, the implementation of an efficient contracting system requires a longer-term development process [1]. Gains might emerge only over a longer period of time, after public officials have been able to exclude poor performers and to improve their own performance management. It could be useful to favor well-performing providers in subsequent rounds of tendering, for instance, by using reputation indicators. As this creates entry barriers for new entrants, thus reducing competition between providers, an option to terminate contracts if performance drops below a pre-defined threshold is necessary with this setup.
The other side of the coin: Modernization of public employment services
In many countries, the concepts of new public management have been applied to reform public employment services with the intent of replacing bureaucratic input-based administrative structures with output-oriented performance management in order to increase PES efficiency. For instance, a performance management system based on management by objectives has been implemented in the German PES [4]. Local branches of the public employment agency are afforded more flexibility to allocate resources, but are also held accountable for the success or failure of their policies. Objectives are established and monitored by means of performance indicators (i.e. the share of integrations of unemployed persons into employment), and wage premiums are awarded to managers depending on the degree to which these objectives are fulfilled. In this situation, the comparative advantage of quasi-markets might not be as strong as under the traditional bureaucratic structures of PES.
Randomized controlled trials on contracting out
Theoretically, placement services delivered by private providers may be more or less effective and/or cost-efficient than services delivered by the PES. Thus, the relative benefits must be investigated empirically to determine the real-world consequences of job placement privatization. Randomized controlled trials that assign individuals to either public or private services avoid any problems of unobserved heterogeneity in both groups. The number of such trials conducted so far, however, is surprisingly small [5], and most of them have focused on particular groups of workers. In the experiments described below, public providers were not allowed to refuse unemployed persons assigned to them, so cherry-picking could not occur.
The largest study—with around 44,000 participants—was conducted in France in 2007 [6]. Job seekers were assigned to either a private or a public program offering intensive job search assistance, or to a control group receiving standard services from the PES. Compliance rates (the share of those who were assigned to a particular treatment and did in fact take advantage of it) amounted to approximately 50% in both intensive treatment groups. The results show that intensive job search assistance has a positive impact on exit rates to employment, and the effect is considerably higher for the public than for the private program. Furthermore, a basic cost–benefit analysis strongly favors the intensive public program.
Some recent field studies have analyzed the effects of contracting out on the labor market prospects of hard-to-place unemployed persons. A randomized field experiment in two German labor market agencies from the years 2009–2010 assigned hard-to-place unemployed individuals to either intensive in-house placement services or to contracted-out job placement services [4]. Compliance was around 80% for assignments to privately provided services and 100% otherwise. Over a period of 18 months from assignment, and compared to contracted-out intensive services, assignment to the intensive public employment services reduced the accumulated number of days in unemployment by one to two months. However, two-thirds of this effect can be attributed to labor force withdrawals. One explanation would be that in-house (i.e. PES) teams have been more successful in encouraging individuals to deregister from unemployment. From a fiscal point of view, the public provision of services performed slightly better.
Two further studies refer to Swedish unemployed persons with particular labor market impediments. The first analyzes the effects of contracting out for groups of hard-to-place unemployed persons during the period 2007–2008 in three regional labor markets; more specifically, the focus was on unemployed persons under the age of 25, migrants, and people with disabilities [7]. Compliance rates were below 30%. Overall, the study did not find any effects on the probability of employment due to enrolment in private versus public placement services, and the same holds for the subgroup of disabled persons. However, evidence did show positive short-term effects for migrants and negative effects for young unemployed individuals who participated in private services. No cost–benefit analysis was provided. The second Swedish study investigates the effects of vocational rehabilitation conducted by public or private providers on the labor market prospects of individuals who had been out of work long-term due to sickness during the period 2008–2009 [8]. Compliance rates were around 80%. No differences could be determined in the employment rates and the average costs of rehabilitation.
A study for Denmark [5] analyzes the effect of contracting out placement services for highly educated job seekers in two metropolitan areas who hold a university degree and have been unemployed for less than three months. Private providers delivered more intense, employment-oriented services, which were offered earlier during unemployment. However, the effectiveness of public and private services did not differ regarding subsequent outcomes, such as shares in regular and subsidized employment, in non-benefit receipt, and in unemployment. Furthermore, the two modes of service delivery were found to be equally as costly from a public spending perspective. Note that this study also provides a systematic overview of the features of other available experimental studies.
Finally, an early study for the State of Michigan analyzed an experiment where hard-to-place participants in the Aid to Families with Dependent Children program were randomly assigned to private providers, where they received more intensive services. It reported no substantial effects on job-finding rates; the compliance rates of the individuals assigned were below 50%.
All in all, none of the experimental studies cited above indicate that contracting out is necessarily more effective or more efficient—results are either indifferent or in favor of public services. The design of the experiments, however, usually does not permit disentanglement of potential channels that may have had an impact on program effectiveness. Candidates for such channels are, for instance, the contents of services, differing caseloads (number of unemployed individuals per caseworker), and the performance standards in the private and public settings [4].
Non-experimental impact studies on contracting out
A number of non-experimental impact studies utilize regression models or statistical matching techniques to compare individuals assigned to private placement services with similar “statistical twins” that received in-house services from the PES. The drawback of these studies is that they can only control for observed heterogeneity, while important unobserved differences between the groups might still exist.
For Germany, results dealing with the early era of contracting out indicate that placement services provided by private agencies might, at best, be as effective as those provided in-house. One such study finds significant small and negative short-term impacts on the employment prospects of individuals assigned to a private placement agency during 2004. Another that deals with recipients of means-tested benefits who were assigned to a private provider during 2005 shows that the assignment was generally ineffective and even counterproductive in some subgroups [9]. Only for selected groups of rather hard-to-place job seekers was the assignment found to be effective. The rather discouraging results might, however, be due to deficits in contract management during the time period investigated.
A study for Belgium estimates the effectiveness of contracted-out services for the long-term unemployed that were delivered by for-profit and non-profit firms [10]. The intervention reduced unemployment duration, but also increased employment instability and labor market withdrawals. Overall, for-profit firms performed best. Another study for Sweden compares unemployment durations in areas that cooperate with private contractors versus areas that do not. Cooperating regions had lower transition rates into employment, which the author accredits to poor contract design.
The design of contracting out schemes and public performance standards
One empirical study uses detailed information on the contract design and payment of German private employment service providers in 2009 and 2010; it provides estimation results that are in line with expectations [11]. It indicates that, on average, high performance-based payments have positive effects on private provider performance in the short and longer term. In contrast, high upfront payments decrease the likelihood of re-employment for certain subgroups of clients and private providers.
A number of studies investigate design issues for the Netherlands. One study compares the impact of “no cure, less pay” and “no cure, no pay” contracts (i.e. payment methods that primarily or only reward agencies for successful job placement) and finds that the latter were associated with some cherry-picking. However, it also finds that the share of shorter-term job placements for workers with good placement prospects increased somewhat [12]. These effects could not be found for hard-to-place unemployed persons. Another study investigates the effects of changes in scoring weights in tenders for contracted-out services (i.e. how important different criteria are in the tender selection process). Increasing the quality components resulted in higher priced bids, while increasing the weight of indicators for reputation and for intended integration plans improved placement rates [13].
Limitations and gaps
There is still little rigorous empirical evidence on the effectiveness and efficiency of competing modes of delivery of job placement services. Moreover, findings from one region will not necessarily hold for others. For instance, the characteristics of the unemployed population and the labor market situation might vary between regions, programs might consist of heterogeneous components, or the contract design and monitoring system might differ. In addition, observed differences in the outcomes of private and public providers might only be partially attributable to the inherent aspects of each respective method. Examples include frictions in the transfer of unemployed individuals from public employment agencies to private providers, and the possibility that case workers in public services use sanctions against the unemployed as an incentive device [4]. Contract structures for private placement services might create relatively weak incentives for effective placement, while new internal performance monitoring systems in the PES might create strong in-house incentives. More carefully tailored research is needed to fill these knowledge gaps.
Summary and policy advice
Many countries have contracted out at least a portion of their job placement services to private providers, largely based on the idea that this will be more cost-efficient for the state. However, the few available randomized controlled trials that compare private and public placement service delivery methods indicate that job placement by the PES is neither less effective nor less cost-efficient than services provided by private placement agencies. In fact, some of these trials even indicate that public provision has been both more effective and more efficient than the private alternatives existing in the investigated regions.
Having said that, these findings cannot be generalized; each country has unique circumstances that must be taken into account when designing subcontracting protocols and for determining an optimal mix of public and private placement services. If well-designed contractual arrangements and reliable performance measurements are ensured, then contracting out might be able to improve outcomes, at least for specific worker groups. However, before introducing a large-scale contracting out program, countries should assess the effectiveness and efficiency of the desired program by conducting a pilot study. Ideally, this would be done using a randomized controlled trial accompanied by a qualitative implementation study, which would allow policymakers to adjust the program as soon as problems with contract design or placement management become obvious.
Acknowledgments
The author thanks an anonymous referee and the IZA World of Labor editors for many helpful suggestions on earlier drafts. The author also thanks Pia Homrighausen and Sarah Bernhard. Previous work by the author (together with Gerhard Krug) contains a larger number of background references for the material presented here and has been used intensively in all major parts of this article [4].
Competing interests
The author is employed at the Institute for Employment Research (IAB), which conducts independent research. It is funded by the German Federal Employment Agency and the German Federal Ministry of Labor and Social Affairs. The IZA World of Labor project is committed to the IZA Guiding Principles of Research Integrity. The author declares that she has observed these principles.
© Gesine Stephan