Steffen Eckhard

International organizations (IOs) rely on locally hired staff to translate global mandates into local action, yet the performance implications of localization remain contested. Addressing this gap, this study combines natural language processing (NLP) with a novel dataset of 5,230 project evaluations and staffing information from five United Nations organizations (2007–2022). Building on current research on local staff employment in IO projects together with insights from expert interviews, we test two dominant competing arguments: local capture versus local knowledge. Multilevel and fixed-effects analyses show that projects with higher shares of local staff perform significantly better, but only once local employees constitute at least 60% of the team. The results advance theory on IO effectiveness by shifting attention from local discretion to staff composition, demonstrating that locally embedded knowledge, when institutionally represented in staffing, constitutes a key organizational resource for adaptive implementation.

Verbal communication between bureaucrats and citizens crucially determines the dynamics and outcomes of public encounters. However, so far, scholars have not sought to quantitively measure it, which limits our knowledge of the role language plays in shaping interactions between bureaucrats and clients. Addressing this gap, we develop and systematically validate computational linguistic measures of complexity and emotionality as two foundational dimensions along which bureaucratic communication varies. We thereby rely on a novel speech‐data set of 154 conversations between frontline bureaucrats and citizens, recorded at four local agencies in Germany. Content validation shows that the two measures serve to accurately capture these two concepts, whilst convergent validation yields that they conform with two alternative measures. Finally, construct validation is established by showing that they generate theoretically plausible results. The paper contributes to the literature by innovating the methodical approaches for analysing the language used during bureaucrat‐client interactions. It thereby paves the way for advancing our understanding of the manifold ways in which these interactions affect the administrative treatment and outcomes citizens experience.

Street-level workers are crucial to promoting equity and equality in public service delivery. Their discretionary decisions can either perpetuate biases and stereotypes or enhance equity by acknowledging citizen-clients’ diverse social backgrounds. While current research often focuses on the micro-level behaviors of street-level bureaucrats (SLBs), there is limited exploration of how societal structures, historical patterns, and cultural beliefs shape these decisions. This Roundtable article argues for incorporating sociological imagination into the study of equity in public service delivery. Three key areas are proposed: (1) societal structures and dynamics—examining how social stratification and political conditions influence equity and equality in service delivery; (2) historical patterns of inclusion and exclusion, and the long-term effects of inequity and inequality on citizen-clients’ attitudes and behaviors; and (3) cultural beliefs about (in)equality and (in)equity—exploring perceptions of just inequalities and inequities among workers and the public. Together, these dimensions provide a richer understanding of equity and equality in public service delivery that can advance the public administration and public management knowledge base.

During crisis management, public administrators typically adjust bureaucratic rules to become more flexible. In this paper, we go one step further, asking about the factors that explain rule-breaking behavior among public servants during crisis management. We report novel qualitative evidence and findings from a preregistered randomized quasi-experiment surveyed among 448 public servants in Germany. Our findings reveal that bureaucrats’ prosocial attitudes, whether the affected population is part of their in-group and higher levels of red tape affect significantly their willingness to engage in rule-breaking. Effects from decisive leadership are weak and conditional on other explanatory factors. These findings provide valuable insights for theory and practice in crisis and human resource management, demonstrating that organizations have limited leverage in influencing whether their employees engage in rule-breaking behavior, which is, in the main, conditional on individual factors and the characteristics of the crisis.

Governments spend hundreds of millions on evaluations to assess the performance of public organizations. In this article, we scrutinize whether variation in the institutional design of evaluation systems leads to biases in evaluation findings. Biases may emerge because influence over evaluation processes could enable the bureaucracy to present its work in a more positive way. We study evaluation reports published by nine international organizations (IOs) of the United Nations system. We use deep learning to measure the share of positive assessments at the sentence level per evaluation report as a proxy for the positivity of evaluation results. Analyzing 1082 evaluation reports, we find that reports commissioned by operative units, as compared to central evaluation units, systematically contain more positive assessments. Theoretically, this link between institutional design choices and evaluation outcomes may explain why policymakers perceive similar tools for evidence‐based policymaking as functional in some organizations, and politicized in others.

International organizations (IOs) of the United Nations (UN) system publish around 750 evaluation reports per year, offering insights on their performance across project, program, institutional, and thematic activities. So far, it was not feasible to extract quantitative performance measures from these text-based reports. Using deep learning, this article presents a novel text-based performance metric: We classify individual sentences as containing a negative, positive, or neutral assessment of the evaluated IO activity and then compute the share of positive sentences per report. Content validation yields that the measure adequately reflects the underlying concept of performance; convergent validation finds high correlation with human-provided performance scores by the World Bank; and construct validation shows that our measure has theoretically expected results. Based on this, we present a novel dataset with performance measures for 1,082 evaluated activities implemented by nine UN system IOs and discuss avenues for further research.