Open Bilbilitano opened 4 years ago
We have census population data from 1900 to 2011, the population is calculated in the following years 1960, 1970, 1981, 1991, 2001, 2011 (file 01_data/03_raw/population/population1900to2011census.dta)
For padron so far we have population since 1996 to 2016 but 1997 is missing (we can interpolate that one) (file 01_data/03_raw/population/population_clean.dta)
Looking into Banesto data. Hopefully, we can get a better sense of population data before.
Panel data structure using matching based on month of birth, marriage and birth of last living child. Allows to follow up families over birth event times.
Include marriage data to also get observation before births.
Look at the marriage-1st birth gap and 2nd birth-1st birth gap. Do we observe an effect here? Is the gap larger or smaller than before? This could point out to budget constraints. If the new births come from families with larger gaps that suggest budget constraints.
How can we disentangle budget constraints from optimistic (maybe unrealistic) expectations? This is a key point for the paper to work. Budget constraints would be a very nice result.
Another possibility is that families have more kids because now they can specialize and one member of the couple can stay at home at taking care of the kids. Do we observe an increase in the number of home-stay mums? Even after controlling for employment prior to birth (using marriage certificates)?
The data should be structure in the following way: For each couple, we link marriage and birth certificates using a matching strategy based on months of births for each partner, marriage and births. I've done this in the past and it works remarkably well and it's quite simple to implement. However, one of the main lim