Street-Level Bureaucrats as Individual Policymakers: The Relationship between Attitudes and Coping Behavior toward Vulnerable Children and Youth

ABSTRACT

Lipsky (1980) pointed out that street-level bureaucrats (SLBs) are important policymakers due to the discretion they exercise and argued from a structural perspective that these workers manifest relatively similar coping behaviors owing to their shared working conditions, characterized by chronically limited resources and non-voluntary clients. Using data from a national survey of municipal child welfare caseworkers in Denmark, we further develop Lipsky's theory from an agency perspective by focusing on variation in coping among SLBs and examining the extent to which such variation is explained by SLBs’ attitudes towards the target group, the objectives and content of their jobs, and their perceptions of the capacity of their institutions. We find that SLBs’ aversion to and tolerance of the client group, their perceptions of institutional capacity in terms of municipal resources and local political inefficacy, and their conceptual modification of job contents are all related to their use of coping.

ACKNOWLEDGMENTS

We would like to thank two anonymous reviewers and the editor for their helpful comments, and also the municipal caseworkers who responded to our survey.

Notes

“Coping” means “to deal with and attempt to overcome problems and difficulties” or “to contend with difficulties and act to overcome them.” Although the term was suitable for the behavior in the context of the structural factors presumed to cause it as described by Lipsky, it can be argued that it is not the most appropriate term to exclusively describe behavior purged of any causal linkage. In other words, the problem with using the term “coping” outside of the specific context described by Lipsky is that it mixes presumed cause (a chronic shortage of resources) and effect (street-level bureaucrats’ behavior aimed at rationing the supply of and reducing the demand for their services). See also Handler (1986).

Note that the structural imbalance that is argued to motivate coping at the individual level may itself be partly a consequence of resource cutbacks of various types at the macro (agency or higher) level. As Lipsky (1984) documents in his article on “bureaucratic disentitlement,” policymakers may resort to policy retrenchment in a variety of ways that are reflected in inadequate welfare staffing levels and growing staff responsibilities, and the rationing of services and benefits.

Coping can also be influenced by “policy alienation,” a concept related to, but distinct from, work alienation (see Loyens 2015; Tummers 2012).

In fact, Nielsen and Lindemann (2006) found that New Public Management reforms had, to varying degrees, been adopted in Danish municipalities’ services for vulnerable children but found no impact of these on SLBs’ use of coping. One reason might be that NPM reforms were never adopted and implemented in Denmark to the same extent as in the US and many other countries.

After deletion of duplicate cases (i.e., involving the same caseworker) and cases with values missing on any one of the variables in the analysis.

At the same time, one can choose to distinguish further between sub-dimensions of this supply-side coping. A principal component factor analysis indicated that our five coping items load on two sub-dimensions, of which the main—and, to our knowledge, only—one identified in the literature is “creaming” or “cream-skimming” (the first three items). However, for theoretical reasons, which are also supported by Lipky's treatment of rationing services as one theoretical dimension, we prefer to use a single formative index of coping tapping rationing services. This is similar to democracy indices that often combine different dimensions in one index for theoretical reasons. Our sensitivity analyses have shown that the relationships between our explanatory variables and the reflexive creaming sub-index remain largely the same as when using one overall formative coping index.

In a Danish context, there is reason to expect concordance between the two items tapping callousness and indifference toward both young persons and their parents. Until quite recently, the focus of child protection services was not primarily on the child but on the family. In addition, it is important to note that the primary client in the case of youth cases are the young persons themselves, while the client in the case of child cases are the parents (the mother, in the vast majority of cases). Thus, we are referring to two separate sets of clients: young persons and children (who are represented by their parents). Caseworkers are more likely to hold parents responsible for their children's condition, but are more likely to blame young persons for their own condition. (The bivariate correlation between the two items is high (0.72) in our sample, and provides empirical support to the hypothesis that the insensitivity and indifference that sets in presumably as a result of prolonged stress on the job is a generalized phenomenon robust to client type (as described earlier)).

An examination of the five items in the scale shows, as expected, considerable variation in the degree to which caseworkers resort to these methods, with the combined percentage of “agree” and “strongly agree” responses ranging from just 3% (more time on cases more likely to be a success) to 27% (priority to easy cases). The variation is also reflected in the summary statistics: the means (std. deviations) of the five items ranged from 1.52 (.85) to 3.17 (1.26). This underscores the overall finding that SLBs’ self-reported coping as measured by these items is low.

We also conducted an ordinal logistic regression (after confirming that the proportional odds assumption held) as a robustness check. The results of the logistic regression were similar to those obtained with OLS regression in terms of the sign and statistical significance of the coefficients. We therefore retain the results from the OLS regression since interpretation of coefficients is easier. In addition, using the log of the coping variable as the dependent variable (to correct for the skew in the distribution) did not substantively alter the OLS results.

Note, however, that our test relates only to objective capacity and not subjective or perceived capacity, which previous research suggests is strongly related to coping (Winter 2002; Beer et al. 2008).

An index of the types of supervision/professional support available to the caseworker was also tested in the models and found to be non-significant.

This result is robust to controls for case characteristics including, most notably, the seriousness of the case.

Additional information

Notes on contributors

Siddhartha Baviskar

Siddhartha Baviskar ([email protected]) is senior research fellow at SFI–Danish National Centre for Social Research. His research examines policy implementation and street-level bureaucracy in the areas of vulnerable children and youth and education in Denmark.

Søren C. Winter

Søren C. Winter ([email protected]) is professor at SFI–Danish National Centre for Social Research. His research focuses on policy implementation, street-level bureaucracy, and public management.