Research & Projects

Project 1

Estimating CPI bias using Engel Curve for the Elderly

Abstract

This paper estimates bias in the Consumer Price Index (CPI) for elderly individuals in the United States using Engel curve methodology and household-level data from the Panel Study of Income Dynamics (1999–2023). The analysis focuses on how well the CPI reflects changes in the cost of living for older Americans, with particular attention to differences by gender and marital status. Results show that from 1999–2011, the CPI overstated the increase in the cost of living for both elderly and non-elderly households. In contrast, from 1999–2023, the CPI understated cost increases, with the bias generally larger for non-elderly households. Within the elderly group from 1999-2021, the CPI overstated the annual rate of increase in the cost of living for unmarried and female heads of household, but understated the annual rate of increase in the cost of living for married and male heads of household. By identifying these discrepancies, the study contributes to a better understanding of how inflation affects different segments of the population and informs policy efforts aimed at improving the accuracy of cost-of-living adjustments used to index Social Security benefits.

JMP.AudertaAmoako-Nuamah (1).pdf

Project 2

Maternal Employment and Overweight Children

Abstract

Using the 2014, 2019, and 2021 waves of the PSID Child Development Supplement, this study examines how maternal work hours impact children’s overweight status and how this relationship varies with household income. The analytic sample includes 3,404 children aged 3–17 linked to their primary caregiver. A benchmark linear probability model shows that a ten-hour increase in maternal weekly hours is associated with a 7.7 percentage-point higher probability of child overweight, while the interaction between hours and log income is negative, indicating weaker effects at higher incomes. A semiparametric partially linear varying-coefficient model allows the hours effect to vary with income and reveals clear heterogeneity: the association is positive and statistically significant at low incomes, attenuates near the middle, and turns negative at the top of the income distribution. Maternal obesity is a strong positive predictor of child overweight; higher birthweight is similarly associated with elevated overweight risk, while the presence of a spouse or partner in the household significantly reduces it. The findings point to systematic heterogeneity in the maternal-employment–child-overweight relationship across the income distribution, with direct implications for the targeting of nutritional assistance and childcare subsidy programs.

Project 3

Estimating CPI bias using Semiparametric Engel Curves

Objectives

We estimate semiparametric Egnel curves to infer CPI bias as a cost of living index. Utilizing PSID household level data over 1999-2023, we compare household with a senior head with the same level of CPI deflated total expenditure, controlling for demographic differences in marriage status, gender and race in a novel partially linear single index model. The estimated time drift parameter captures the CPI bias, and we find that CPI overstates the cost of living by less than 1 percent annually, with a much larger bias of 2.72 percent during 1999-2011.

Page updated

Report abuse