Jay Stewart

Data Watch: The American Time Use Survey

We discuss the new American Time Use Survey (ATUS), an on-going household survey of roughly 1,200 Americans per month (1,800 per month in the first year, 2003) that collects time diaries as well as demographic interview information from respondents who had recently been in the Current Population Survey. The characteristics of the data are presented, as are caveats and concerns that one might have about them. A number of novel uses of the ATUS in economic research, including in the areas of macroeconomics, national income accounting, labor economics, and others, are proposed to illustrate the magnitude of this new surveys possible applications.

How to Think About Time-Use Data: What Inferences Can We Make About Long- and Short-Run Time Use from Time Diaries?

Time-use researchers are typically interested in the time use of individuals, but time use data are samples of person-days. Given day-to-day variation in how people spend their time, this distinction is analytically important. We examine the conditions necessary to make inferences about the time use of individuals from a sample of person-days. We also discuss whether and how surveys with multiple household members or multiple days are an improvement over single-diary surveys.

Adjusting body mass for measurement error with invalid validation data.

We propose a new method for using validation data to correct self-reported weight and height in surveys that do not measure respondents. The standard correction in prior research regresses actual measures on reported values using an external validation dataset, and then uses the estimated coefficients to predict actual measures in the primary dataset. This approach requires the strong assumption that the expectations of measured weight and height conditional on the reported values are the same in both datasets. In contrast, we use percentile ranks rather than levels of reported weight and height. Our approach requires the weaker assumption that the conditional expectations of actual measures are increasing in reported values in both samples. This makes our correction more robust to differences in measurement error across surveys as long as both surveys represent the same population. We examine three nationally representative datasets and find that misreporting appears to be sensitive to differences in survey context. When we compare predicted BMI distributions using the two validation approaches, we find that the standard correction is affected by differences in misreporting while our correction is not. Finally, we present several examples that demonstrate the potential importance of our correction for future econometric analyses and estimates of obesity rates.

The Timing of Maternal Work and Time with Children

The author investigates how maternal employment affects when during the day that employed mothers engage in enriching childcare and whether they adjust their work schedules to spend time with their children at more-desirable times of day. Using data from the American Time Use Survey and focusing on mothers of pre-school-aged children, he finds that both full- and part-time employed mothers shift enriching childcare time from workdays to non-workdays. On workdays, full-time employed mothers shift enriching care time to evenings, whereas part-time employed mothers shift care time very little. The author finds no evidence that mothers working full time adjust their work schedules to spend enriching time with their children at more preferred times of day. In contrast, part-time employed mothers shift their work hours to later in the day in order to spend time with their children at more-desirable times of day.

Tracking the Returns to Education in the 1990s. Bridging the Gap between the New and Old Current Population Survey Education Items.

The Current Population Survey (CPS) is used for many studies examining trends in the returns to education. The CPS changed its education item in 1992. This paper develops adjustment factors for earnings at different education levels to make pre- and post- 1992 earnings comparable. To accomplish this, contradictory results from alternative data sources are analyzed and, to the extent possible, reconciled. The adjustments reduce the estimated growth in the College/High School earnings ratio between 1989 and 1993 by between 29 and 48 percent for men and between 44 and 73 percent for women.

Adjusting Body Mass for Measurement Error with Invalid Validation Data

We propose a new method for using validation data to correct self-reported weight and height in surveys that do not weigh and measure respondents. The standard correction from prior research regresses actual measures on reported values using an external validation dataset, and then uses the estimated coefficients to predict actual measures in the primary dataset. This approach requires the strong assumption that the expectations of actual weight and height conditional on the reported values are the same in both datasets. In contrast, we use percentile ranks rather than levels of reported weight and height. Our approach requires the much weaker assumption that the conditional expectations of actual measures are increasing in reported values in both samples, making our correction more robust to differences in measurement error across surveys. We then examine three nationally representative datasets and confirm that misreporting is sensitive to differences in survey context such as data collection mode. When we compare predicted BMI distributions using the two approaches, we find that the standard correction is biased by differences in misreporting while our correction is not. Finally, we present several examples that demonstrate the potential importance of our correction for future econometric analyses and estimates of obesity rates.

How Does Household Production Affect Earnings Inequality? Evidence from the American Time Use Survey

Although income inequality has been studied extensively, relatively little attention has been paid to the role of household production. Economic theory predicts that households with less money income will produce more goods at home. Thus extended income, which includes the value of household production, should be more equally distributed than money income. We find this to be true, but not for the reason predicted by theory. Virtually all of the decline in measured inequality, when moving from money income to extended income, is due to the addition of a large constant--the average value of household production--to money income. This result is robust to alternative assumptions that one might make when estimating the value of household production.

Male Nonworkers: Who Are They and Who Supports Them?

Although male nonworkers have become a larger fraction of the population since the late 1960s, very little is known about who they are or who supports them. Using data from the March Current Population Survey and the 1979 National Longitudinal Survey of Youth, this article fills that void. The picture that emerges is that there is a small cadre of marginal workers who often do not work for periods of a year or more. The vast majority of nonworking men (men who do not work at all during the year) receive unearned income from at least one source, and the amount of unearned income they receive varies significantly by their reason for not working. Family members provide an important alternative source of support for nonworking men who have little or no unearned income of their own.

Adverse Selection and Pay Compression

A fundamental result of the principal-agent literature is that pay will be linked to performance when it is difficult for the principal to monitor the agent’s actions. However, performance pay can lead to adverse incentives. In these models, high-powered incentives encourage workers to neglect some aspects of their job or to sabotage their coworkers’ efforts. This paper offers another explanation for the weak link between pay and performance: worker heterogeneity. When workers are heterogeneous and labor contracts are contests, the Nash equilibrium often pools workers. I show that this implies that the link between pay and performance is weaker than would be the case if firms could observe workers’ types before contracting and offer each type their respective optimal contests.