The Family Policy Database, Version 2 (2003) contains five files:

- the detailed policy data are presented in three Excel workbooks, 
- a set of composite indexes is presented in a fourth Excel workbook, and
- the reference list is in a Word file (Reference List.doc)

The Policy Data workbooks:

The index workbook:

The indexes were constructed as follows (see Gornick and Meyers 2003, for more detail):

1) We converted the policy data presented in the detailed policy data tables to 22 indicators. We entered quantified data (e.g., ECEC enrollment rates) numerically and we coded qualitative data (e.g., ECEC quality) into categories (e.g., high, medium, low.)

We included all of our major policy measures, with the exception of those that regulate part-time work, as implementation of most of those measures remains ongoing.

For all 22 policy indicators, coded data correspond to the data in the policy tables, with one exception. Because the current regulated work week in France (35 hours) was implemented so recently, we used the value that was current in France in the late 1990s.

 The 22 indicators include (with units):  

The raw data on these 22 indicators is presented in IN. Table 1.

2) We converted all qualitative values to quantitative values (e.g., high, medium, low were coded as 1.0, .66, and .33.)   We then re-scaled all indicators such that a higher value signified more policy support.  For example, cost to parents if children are 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 work hours were converted such that higher values signify an earlier starting age and shorter weekly work hours.  

 3) We re-scaled all indicators so that their values ranged from 0-1.  We did that using one of the following methods: using the original value if it was a percentage (e.g., enrollment rates), dividing by the observed maximum (e.g., weekly school hours), or dividing by the theoretical maximum (e.g., the gender equality scale).

4)   We created seven sub-indexes by combining the individual indicators.  We weighted some items based on our expectation of the share of the families affected by individual components.  The subindexes were constructed as follows.

We then re-scaled these subindexes to 0-1 by dividing by the observed maximum.   

5) We converted the subindexes into Indexes A, B, and C as follows:

We then re-scaled Indexes A, B, and C to 0-1 by dividing by the theoretical maximum (5, 3, and 7, respectively).  These index values are presented in IN. Table 2.

Finally, note that variable v17 presents a “gender equality in paid family leave” scale.  It was constructed as follows:

 We assigned countries one point on the “gender equality scale” if they have any paid paternity leave, two points if fathers have non-transferable parental leave rights (either “use or lose” portions of share-able leave or individual entitlements) and up to three additional points depending on wage replacement (three points if benefits are wage-related and at 80 percent or higher, two points if benefits are wage-related but at less than 80 percent, and one point if benefits are paid but only at a flat rate).


Reference List