Families That Work: Policies for Reconciling Parenthood and Employment

Appendix C: Construction of Policy Indexes

Appendix C Construction of Policy Indexes THE INDEXES PRESENTED in chapter 8 were constructed as follows: First, we converted the policy data presented in chapters 5, 6, and 7 to twenty-two indicators. We entered quantified data (for example, ECEC enrollment rates) numerically, and we coded qualitative data (for example, ECEC quality) into categories (high, medium, low). We included all of our major policy measures, with the exception of those that regulate part-time work, as nearly all current policies governing part-time work in Europe were enacted after our most recent data on labor market outcomes. For all twenty-two policy indicators, coded data correspond to the data in chapters 5, 6, and 7, with one exception: Because the current regulated workweek in France (thirty-five hours) was implemented after our outcome data on working time, we used the value that was current in France in the mid-1990s. The twenty-two indicators (with units) are as follows: EARLY CHILDHOOD EDUCATION AND CARE • v1 guaranteed slot for some children 0-1-2 (yes, no) • v2 enrollment in public care ⬍ age 1 (percent of age group) • v3 enrollment in public care age 1–2 (percent of age group) • v4 cost to parents if children in public care age 1–2 (percent of total cost) • v5 enrollment in public care age 3-4-5 (percent of age group) • v6 cost to parents if children in public care age 3-4-5 (percent of total cost) 316 Families That Work • v7 typical hours age 3-4-5 (full-day, mixed, part-day) • v8 enrollment age 6 (percent of age group) • v9 quality (low, medium, high) • v10 tax relief for ECEC (yes, no) SCHOOL SCHEDULING • v11 starting age (age) • v12 hours per day (hours) • v13 days per year (days) • v14 continuity of school day (yes, no, sometimes) FAMILY LEAVE • v15 weeks of full-pay available to mothers (weeks) • v16 paid paternity leave (yes, no) • v17 gender equality scale / incentives for fathers (see chapter 6) • v18 some paid leave after third birthday (yes, no, some) • v19 paid sick-child leave (yes, no) • v20 expenditures on leave (2000 $U.S. per employed woman) WORKING TIME • v21 normal weekly hours (hours) • v22 normal vacation time (days) The raw data on these twenty-two indicators are presented in tables C.1 and C.2. Second, we converted all qualitative values to quantitative values (for example, high, medium, and low were coded as 1.0, .66, and .33). We then rescaled all indicators such that a higher value signified more policy support. For example, cost to parents if children in public care was converted to cost to government (for children in public care), with a higher value signifying more government support. School starting age and normal weekly hours were converted such that higher values signify an earlier starting age and shorter weekly hours. Third, we rescaled all indicators again so that values fall be- Policy Indexes 317 tween 0 to 1. We did that using one of three methods: for percentages the original value was used (for example, enrollment rates); other values were divided by the observed maximum (for example , weekly school hours) or by the theoretical maximum (for example , the gender-equality scale). Fourth, we created seven subindexes by combining the rescaled indicators. We weighted some items, based on our expectation of the share of the families affected by individual components or based on their relative importance. The subindexes were constructed as follows: • sA. ECEC, youngest child under the age of three: v1  [.33  v2]  [.66  v3]  v4  v9  [.25  v10] • sB. Family leave, youngest child under the age of three: v15  v16  v17  v19  v20 • sC. ECEC, youngest child aged three to five: v5  v6  [1.5  v7]  v9  [.25  v10] • sD. Family leave, youngest child aged three to five: v17  v18  v19 • sE. Family leave, youngest child six or older: v19 • sF. School schedules: [.10  v8]  [.10  v11]  v12  v13  v14 • sG. Working time: v21  v22 We then rescaled these subindexes to values of 0 to 1 by dividing each by the observed maximum. Next, we converted the subindexes into Indexes A, B, and C as follows: • Index A  all policies that affect families with children under the age of six  sA  sB  sC  sD  sG • Index B  all policies that affect families with children ages six and older  sE  sF  sG (weighted to give school schedules [sF] 50 percent) • Index C  all...