Causality in Public Management Research: are we Humeans or are we dancers?

Causality in Public Management Research: are we Humeans or are we dancers?Public Management research seeks not to only develop new knowledge that has inherent value but also, as a ‘design science’, to make a difference to... [ view full abstract ]

Causality in Public Management Research: are we Humeans or are we dancers?

Public Management research seeks not to only develop new knowledge that has inherent value but also, as a ‘design science’, to make a difference to the governance, structures, processes and outcomes of public sector organizations. This approach assumes that causal arguments and evidence are identifiable, and can be applied to produce better results. Yet the notion of causality that underpins research in the field is opaque; causality is sometimes mentioned, but is rarely discussed explicitly.  At best we are ‘dancing around’ the topic of causality. The consequence is that we cannot be confident that we have identified organizational variables that are causally connected to the outcomes of interest, or that that the practical recommendations that we draw from our research will have the intended effects if they are implemented. 

The aim of this paper is to examine whether research on management and organizational performance is consistent with Humean criteria of causality.  Even though Hume’s work is almost three centuries old it continues to be the benchmark for new interpretations of causality.  Traditional readings of Hume suggest that his model of causality contains three simple criteria: temporal precedence (causes happen before their effects), spatial contiguity (causes are adjacent to their effects) and constant conjunction (causes are consistently associated with their effects).  However his work also contains a more sophisticated and complex model that relaxes the conditions of temporal precedence (and allows for cause and effect ‘cycles’), spatial contiguity (causal action can happen ‘at a distance’) and constant conjunction (such conjunctions can be ‘impeded’ by contingencies).

Ideas about causality in research on management and organizational performance are explored by reviewing empirical studies that have been published in journals such as JPART, PAR and PMR in the last twenty years, and drawing out the causal assumptions that underpin them.  These studies closely reflect ‘design science’ principles and seek to draw conclusions on how to improve performance.  The review shows that the conclusions for policy and practice are typically based on notions of causality that are implicit at best, and often characterised by the ad-hoc and inconsistent adoption of a mix of simple and complex Humean criteria.  Conclusions are drawn on the benefits of more explicit causal foundations for research on management and organizational performance and for recommendations for policy and practice.