Thomann, van Engen and Tummers (2018) used fuzzy-set qualitative comparative analysis of two large Dutch surveys to show that frontline implementers' perceived discretion is a quasi-necessary condition for high self-reported willingness to implement policy, supporting bottom-up over top-down implementation theory. The authors themselves left the decisive question open: their evidence concerns implementation intention, not observed implementation behavior, and their cross-sectional, single-source designs cannot establish that perceived discretion causes implementation. This problem asks for a causal-behavioral account of the discretion-implementation link: whether experiencing discretion causally produces more or better implementation behavior, measured on observed implementation behavior and identified causally rather than by correlation, and under which contextual conditions (sector, workload, resources, national context) the relationship holds. This core remains open: later work extends the willingness framework to new countries and adjacent attitudes, but no study known to date manipulates or longitudinally tracks discretion while measuring observed implementation behavior; a recent systematic review of the field concludes that discretion's consequences remain unsettled (Choi, Park and Lee 2025).

Background

Public policies are largely delivered by street-level bureaucrats — frontline public professionals such as teachers, nurses, social workers, and tax inspectors who interact directly with citizens and must translate abstract policy into concrete decisions. A defining feature of their position is discretion: the freedom to choose among courses of action when implementing a policy. Implementation theory has long been split on whether this discretion is a liability or an asset. Top-down theories treat frontline discretion as a control problem that lets policies drift from what lawmakers intended, while bottom-up theories treat it as an inevitable and potentially beneficial resource that lets implementers tailor policies to local circumstances. Lipsky's classic street-level bureaucracy work, and later syntheses such as Evans's review, note that this tension has never been fully resolved empirically. Thomann, van Engen and Tummers (2018) gave this debate a behavioral twist: rather than asking whether discretion should exist, they asked whether experiencing discretion motivates frontline workers to implement a policy at all. They conceptualized perceived discretion as 'policy powerfulness' at three levels — strategic (perceived influence on national policy decisions), tactical (influence on how the policy is translated into local programs), and operational (influence in day-to-day execution) — and measured implementation willingness by survey. Their method was fuzzy-set qualitative comparative analysis (fsQCA), a set-theoretic technique that calibrates each case's membership in conditions and in the outcome as scores between 0 and 1 and then searches for configurations of conditions that are necessary or sufficient for the outcome; a condition is (quasi-)necessary when cases with the outcome almost always also have the condition (high consistency), covering a substantial share of those cases (coverage). Analyzing two Dutch datasets — 1,004 mental-healthcare workers confronted with a diagnosis-related-group (DRG) hospital-financing policy, and 1,087 secondary-school teachers facing education policies — they found robustly that perceived discretion is a quasi-necessary condition, hence a prerequisite, for high implementation willingness. The implied 'flipside' — that the simultaneous absence of strategic, tactical, and operational powerfulness coincides with unwillingness — was supported in the healthcare data but not in education. The authors flagged decisive limitations: the outcome is self-reported willingness (intention), not observed implementation behavior; the cross-sectional, single-source survey design precludes causal claims and invites common-method bias; aggregation and calibration choices were forced by missing data ('optimistic' aggregation, two calibration strategies); and the two Dutch sectors leave generalizability open. An, Yu and Tan's later question-order survey experiment shows how strongly rule-focus versus discretion attitudes matter for frontline service delivery, and Choi, Park and Lee's 2025 systematic review of 162 street-level-discretion studies confirms the field remains fragmented, with attitudinal, cross-sectional designs still dominating.

Problem Statement

Thomann, van Engen and Tummers (2018) established that frontline implementers' perceived discretion is a quasi-necessary condition for high self-reported willingness to implement policy. The question they left open, which is the problem here, is: does experiencing discretion causally produce more or better implementation behavior? In this question the causal exposure is discretion as experienced by the frontline implementer — the construct the original study operationalized as 'policy powerfulness,' the degree to which frontline workers perceive to possess discretion — and an answer must state which behavioral aspect of implementation it causally establishes (the quantity or the quality of implementation behavior) and with respect to which policy or policies. Because only the objective grant of discretion can be manipulated while perceived discretion can only be measured, an answer that relies on an experimental manipulation must make explicit how the manipulated grant relates to the implementers' experienced discretion, so that the causal claim addresses the same discretion-implementation link as the original finding. An answer must convert the correlational, intention-based finding into causal-behavioral knowledge by meeting four linked requirements that jointly define one objective. (1) Causal identification: the effect of perceived discretion on implementation must be identified with a design supporting causal inference, rather than with a cross-sectional correlation. (2) Behavioral outcome: implementation must be measured directly as observed behavior (administrative records, documented implementing actions verified independently of the respondent's own account, observation of practice, or independently assessed implementation quality or quantity), not as self-reported willingness or intention, and the behavioral outcome must not be drawn solely from the same respondents whose discretion is measured, so as to avoid single-source common-method bias. (3) Contextual moderators: the answer must establish under which conditions the relationship holds, weakens, or reverses — including policy sector, workload, available resources, and national context. (4) External validity: findings must be shown to generalize, or to differ, beyond the original Dutch mental-healthcare and education datasets. The single scientific objective uniting these requirements is a causal-behavioral account of the discretion-implementation link; an answer that restores a causal claim only about implementation willingness, or only within one cross-sectional Dutch survey, does not meet it. What can settle the question is a causal-inference research design that identifies the effect, specifies the behavioral outcome, and lays out the contextual moderators and generalization checks, rather than further analysis of the already-collected cross-sectional survey data.

References

  1. Eva Thomann, Nadine van Engen, Lars Tummers (2018), 'The Necessity of Discretion: A Behavioral Evaluation of Bottom-Up Implementation Theory', Journal of Public Administration Research and Theory 28(4):583-601, doi:10.1093/jopart/muy024, https://doi.org/10.1093/jopart/muy024
  2. Nadine van Engen, Bram Steijn, Lars Tummers (2019), 'Do consistent government policies lead to greater meaningfulness and legitimacy on the front line?', Public Administration 97(1):97-115, doi:10.1111/padm.12570, https://doi.org/10.1111/padm.12570
  3. Thorbjørn Sejr Guul, Mogens Jin Pedersen, Niels Bjørn Grund Petersen (2021), 'Creaming among Caseworkers: Effects of Client Competence and Client Motivation on Caseworkers' Willingness to Help', Public Administration Review 81(1):12-22, doi:10.1111/puar.13297, https://doi.org/10.1111/puar.13297
  4. Shuai Yuan, Zhixia Chen, Mei Sun (2022), 'Discretion: Whether and How Does It Promote Street-Level Bureaucrats' Taking Charge Behavior?', Frontiers in Psychology 13:805872, doi:10.3389/fpsyg.2022.805872, https://doi.org/10.3389/fpsyg.2022.805872
  5. Carolin Schütze, Håkan Johansson (2020), 'The importance of discretion for welfare services to minorities: Examining workload and anti-immigration attitudes', Australian Journal of Public Administration 79(4):426-443, doi:10.1111/1467-8500.12410, https://doi.org/10.1111/1467-8500.12410
  6. Chao An, Jinhai Yu, Xi Tan (2026), 'Public service motivation and policy implementation attitudes: survey experimental evidence from Street-level Bureaucrats in China', Public Management Review 28(6):1486-1510, doi:10.1080/14719037.2025.2501734, https://doi.org/10.1080/14719037.2025.2501734
  7. Heasun Choi, Soohyun Park, Sungyoon Lee (2025), 'Unpacking street-level discretion: a systematic review of its conceptualizations, antecedents and consequences', Public Management Review, doi:10.1080/14719037.2025.2559852, https://doi.org/10.1080/14719037.2025.2559852
  8. Kosal Chuong, Seunghoo Lim (2026), 'Street-level tax officials' willingness to implement Cambodia's value-added tax policy for e-commerce', Journal of Asian Public Policy, doi:10.1080/17516234.2026.2655877, https://doi.org/10.1080/17516234.2026.2655877
  9. Tony Evans (2020), 'Street-Level Bureaucrats: Discretion and Compliance in Policy Implementation', Oxford Research Encyclopedia of Politics, doi:10.1093/acrefore/9780190228637.013.1422, https://doi.org/10.1093/acrefore/9780190228637.013.1422
  10. Ioana-Elena Oana, Carsten Q. Schneider, Eva Thomann (2021), 'Qualitative Comparative Analysis Using R', Cambridge University Press, doi:10.1017/9781009006781, https://doi.org/10.1017/9781009006781