Public Employment Services (PES) increasingly use automated statistical profiling algorithms (ASPAs) to ration expensive active labour market policy (ALMP) interventions to those they predict at risk of becoming long-term unemployed (LTU). Strikingly, despite the critical role played by ASPAs in the operation of public policy, we know very little about how the technology works, particularly how accurate predictions from ASPAs are. As a vital first step in assessing the operational effectiveness and social impact of ASPAs, we review the method of reporting accuracy. We demonstrate that the current method of reporting a single measure for accuracy (usually a percentage) inflates the capabilities of the technology in a peculiar way. ASPAs tend towards high false positive rates, and so falsely identify those who prove to be frictionally unemployed as likely to be LTU. This has important implications for the effectiveness of spending on ALMPs.

Concerns about the unintended effects of marketised public employment services are increasingly expressed because there is mounting evidence that such services are frequently characterised by various gaming practices on the part of their providers. To prevent these unintended consequences, payment-by-result approaches have been progressively strengthened.

The aim of the research reported in this article was to investigate the extent to which such approaches are able to make service providers accountable for client outcomes. The study used two Italian regional cases in order to compare different and alternative contracting arrangements: Lombardy (outcome-based payments), and Emilia-Romagna (fixed payments). Drawing on rich administrative databases, the analysis relied on a quantitative methodology based on propensity score matching and logistic regressions. Even if outcome-based contracting can make service providers financially accountable for the service outcome, the results of the analysis show that it does not sufficiently prevent gaming practices, resulting in inequity among the services provided.

The exercise of administrative discretion by street-level workers plays a key role in shaping citizens’ access to welfare and employment services. Governance reforms of social services delivery, such as performance-based contracting, have often been driven by attempts to discipline this discretion. In several countries, these forms of market governance are now being eclipsed by new modes of digital governance that seek to reshape the delivery of services using algorithms and machine learning. Australia, a pioneer of marketisation, is one example, proposing to deploy digitalisation to fully automate most of its employment services rather than as a supplement to face-to-face case management. We examine the potential and limits of this project to replace human-to-human with ‘machine bureaucracies’. To what extent are welfare and employment services amenable to digitalisation? What trade-offs are involved? In addressing these questions, we consider the purported benefits of machine bureaucracies in achieving higher levels of efficiency, accountability, and consistency in policy delivery. While recognising the potential benefits of machine bureaucracies for both governments and jobseekers, we argue that trade-offs will be faced between enhancing the efficiency and consistency of services and ensuring that services remain accessible and responsive to highly personalised circumstances.

The purpose of this study is to analyse determinants of resources devoted to public employment services (PES). It has proven difficult to disentangle ‘carrots’ (placement and services) from ‘sticks’ (sanctions and monitoring) when tracking the development of PES spending. This has contributed to ambivalence concerning the role of partisan politics, especially since welfare states have been argued to increasingly emphasize the ‘sticks’ aspect of the PES, irrespective of ideological orientation. This suggests that the role of partisan politics should be analysed together with ‘demanding’ activation, which is made possible with novel data on unemployment benefit conditionality. The analysis includes 16 welfare states and results indicate that left and secular centre-right parties are associated with increased resources devoted to PES, but that effects of partisan politics are contingent on the form and extent of benefit conditionality. Increased conditionality is associated with higher PES spending, thus nuancing the cost-containment argument for activation.

This article examines how jobcentre advisors judge the individual deservingness of their clients and investigates what role these judgements play in translating activation policies into practice. The analysis draws on qualitative data gathered in a research project on back-to-work agreements that involved several German jobcentres. It is shown that advisors vary the intensity of ‘demand’ depending on whether benefit recipients fulfil reciprocity expectations and are deemed to have control over the unemployment situation. Whether or not they choose an enabling, supporting approach, and the extent to which they do so, depends on judgements about the client’s amenability to change and the presumed outcome of labour market policy instruments. Moreover, the analysis indicates that frontline workers use experiential cues such as appearance and self-presentation in the face-to-face encounters in order to interpret the client. The findings highlight the relevance of the normative dimension of frontline work.

Artificial intelligence (AI) is increasingly popular in the public sector to improve the cost-efficiency of service delivery. One example is AI-based profiling models in public employment services (PES), which predict a jobseeker’s probability of finding work and are used to segment jobseekers in groups. Profiling models hold the potential to improve identification of jobseekers at-risk of becoming long-term unemployed, but also induce discrimination. Using a recently developed AI-based profiling model of the Flemish PES, we assess to what extent AI-based profiling ‘discriminates’ against jobseekers of foreign origin compared to traditional rule-based profiling approaches. At a maximum level of accuracy, jobseekers of foreign origin who ultimately find a job are 2.6 times more likely to be misclassified as ‘high-risk’ jobseekers. We argue that it is critical that policymakers and caseworkers understand the inherent trade-offs of profiling models, and consider the limitations when integrating these models in daily operations. We develop a graphical tool to visualize the accuracy-equity trade-off in order to facilitate policy discussions.