By Albelda, Randy; Clayton-Matthews, Alan
Research in Brief Series
Table of Contents Introduction Data Behavior Estimations Simulating Unknown Behavior The Simulator Flow Reliability of Estimates Figures
In developing a simulation model to estimate the cost of paid family and medical leave programs in a given state, we rely on data documenting known leave-taking behavior. Where this is not possible, we provide a set of reasonable assumptions about unknown aspects of behavior in the presence of a paid leave program. To obtain the estimates about known leave-taking behavior, we use the Department of Labor's Family and Medical Leave 2000 Survey of Employees (hereinafter referred to as the DOL survey) to estimate behavioral models of leave-taking conditional on the demographic characteristics of individuals, combined with the Census Bureau's March Annual Demographic sample of the Current Population Survey (hereinafter referred to as the CPS) to capture the demographic characteristics of individuals in individual states.
The DOL survey is the best available source of information on leave-taking behavior. It is a representative national sample of leave takers, leave needers (those persons who said they needed but did not take a leave), and other workers who did not take a leave. (1) The survey, which covers the 19-month period January 1999 through July 2000, includes extensive information on the number and types of leaves taken, how long they were, whether and to what extent the employer provided pay during the leave, and whether or not some or additional pay during the leave would result in a decision to take a leave or to take a longer leave. The DOL survey includes several demographic characteristics related to leave-taking behavior, including sex, race and ethnicity, age, martial status, the presence of children, education, family income, and whether or not the respondent was paid on an hourly basis. We use the DOL survey to estimate several aspects of leave-taking behavior, conditional on demographic characteristics and leave type. These include the probability of needing a leave, taking a leave, getting paid for a leave, and extending a leave if some or more pay were received.
The CPS is a nationally representative sample of households, families, and persons. It is of sufficient size at the state level to obtain reliable estimates of total paid leave program costs and of the distribution of program benefits using multiple years of survey data. The CPS also provides a rich array of demographic characteristics that closely match those in the DOL survey, which means that the behavioral models estimated on the DOL survey can be used to predict the leave-taking behavior of any state as represented by the CPS. We draw from data on employed persons who are not self-employed--the entire universe of potential paid program users.
The CPS surveys people in households. Therefore, for our Massachusetts cost estimate, the data include workers who live in Massachusetts, regardless of the state in which they are employed. There may be workers eligible for a paid leave program who work in Massachusetts but do not live in the state. Our simulation data does not capture them. Conversely, there may be people who live in Massachusetts but work in another state who would not be eligible for a Massachusetts paid leave program. Our simulation data do include them. We assume that the net difference is negligible.
Concatenating and "Cloning" CPS data
Aside from errors in estimates of the behavioral equation parameters and errors in assumptions (discussed below), there are two sources of statistical error that are important to consider. One is sampling error in the CPS. The magnitude of this sampling error is approximately inversely proportional to the square root of sampling size and can be reduced by concatenating successive years of the CPS together. We have done that with our Massachusetts estimates by using data from the March 1999, 2000, 2001, and 2002 surveys. …