State measurements of benefit fraud: Why expert elicitations cannot be used to measure incorrect personal assistance payments

Research output: Contribution to journalArticle

Abstract

During the last decade, personal assistance programs have suffered from cutbacks, resulting in a public debate about the future of the service. This debate has largely been framed by public commission estimations of the level of incorrect personal assistance payments, where the idea that circa 10% of all assistance payments are incorrect repeatedly has been used to justify measures designed to decrease costs. These estimates have been generated with the Expert Elicitation Method (EEM), originally developed to assess uncertainty and risks based on hard scientific data. In this study, we aim to answer two questions: (1) whether the method is suited to estimate incorrect payments and (2) whether the application of the method follows scientifically agreed upon methodological recommendations. By a systematic literature review about the method and when it can be used, we conclude that the EEM is not applicable to estimate incorrect payments. In a second step, we examine whether the applications of the Swedish public commission follow common methodological recommendations in the research literature, concluding that assessments of incorrect personal assistance payments do not meet basic requirements of how the method is supposed to be used. Hence, our overall conclusion is that the application of the EEM to estimate incorrect payments in personal assistance should not be relied upon to guide policy decisions.

Details

Authors
Organisations
Research areas and keywords

Subject classification (UKÄ) – MANDATORY

  • Political Science

Keywords

  • Austerity, Expert elicitation, Expert elicitation method, Personal assistance
Original languageEnglish
Pages (from-to)158-167
Number of pages10
JournalScandinavian Journal of Disability Research
Volume22
Issue number1
Publication statusPublished - 2020
Publication categoryResearch
Peer-reviewedYes