Levy Institute Measure of Time and Income Poverty: United States, 2007–2022
Sources, Methods, and Assessment
In this paper, we present the empirical methodology used to estimate the Levy Institute Measure of Time and Income Poverty (LIMTIP) for the United States over the period 2007–2022. We provide a step-by-step account of the statistical matching procedure employed to construct a synthetic dataset by combining the American Time Use Survey (ATUS) for year t with the Annual Social and Economic Supplement (ASEC) for year t + 1. We describe in detail how records were matched using a combination of principal component analysis, propensity score, and clustering methods. We then assess the quality of the match, focusing on the 2022 data. Specifically, we examine the alignment of the ATUS weekday and weekend samples with the synthetic dataset across key demographic characteristics and summarize the performance of the matching algorithm. Finally, we compare the marginal distributions of time use between the original ATUS data and the synthetic dataset. Our findings indicate that the statistical matching procedure produced a high-quality match, rendering the synthetic dataset suitable for time poverty analysis. Although not discussed in detail here, we also evaluated match quality for each year from 2007 to 2021.