A dynamic linear modelling approach to public policy change. (16th October 2017)

• Record Type: Journal Article
• Authors: Loftis, Matt W.; Mortensen, Peter B.
• Is Part Of: Journal of public policy. Volume 38:Number 4(2018)
• Page Start: 553
• Page End: 579
• Publication Date: 2017-10-16
• Subjects: policy dynamics, -- public policy, -- public spending, -- Punctuated Equilibrium Theory; Policy sciences -- Periodicals; 351.007205
• DOI: 10.1017/S0143814X17000186
• Languages: English
• ISSNs: 0143-814X

Abstract

Abstract: Theories of public policy change, despite their differences, converge on one point of strong agreement: the relationship between policy and its causes can and does change over time. This consensus yields numerous empirical implications, but our standard analytical tools are inadequate for testing them. As a result, the dynamic and transformative relationships predicted by policy theories have been left largely unexplored in time series analysis of public policy. This article introduces dynamic linear modelling (DLM) as a useful statistical tool for exploring time-varying relationships in public policy. The article offers a detailed exposition of the DLM approach and illustrates its usefulness with a time series analysis of United States defense policy from 1957 to 2010. The results point the way for a new attention to dynamics in the policy process, and the article concludes with a discussion of how this research programme can profit from applying DLMs.