bd CPS: benchmark of family count fails

[bdecon]

Fixing issues related to the family variables has to be a priority. As one example, the benchmark example that tries to calculate how many families have kids in 2017 is way off.

[bdecon]

There were lots of issues with the previous attempt to benchmark the BLS report on families. The BLS report is based on annual averages of the monthly data and counts only primary families, meaning secondary families in the same household should be excluded from the benchmark. The latest attempt yields results that are reasonably close to the published estimates.