Improving Stylised Working Time Estimates with Time Diary Data: A Multi Study Assessment for the UK

Abstract

Accurate working time estimates represent an important component of the statistical toolbox used for economics forecasting and policy-making. The relatively good availability of such estimates may sometimes induce researchers to take them for granted and see their reliability as largely unproblematic. There is however a growing body of evidence showing that measurement errors may affect their robustness and quality, especially as far as specific but policy relevant subgroups of the population such as part-time or atypical workers are concerned. Against this background, the goal of this paper is to investigate the reliability of paid weekly working-time measurement instruments commonly available in a key UK social survey, the Labour Force Survey. It focuses on the discrepancies between estimates obtained by self-assessed/aggregated instruments—also known as stylised—and those recorded using time diaries which have been found more truthful to the time spent working in ones paid job(s). It is also to explore ways to improve the reliability of stylised estimates in datasets for which no time diary instruments are available, contrasting those where ’usual’ and ’actual’ hours of work are recorded. It does so by creating calibration weights based on the Work Schedule recorded in the 2000 and 2015 UK Time Use Surveys and using them to up/down scale stylised estimates in the 2000 and 2015 UK Labour Force Survey using statistical matching. Such techniques could enable significant improvements of measurement errors in large scale social surveys at a minimal cost.