Carl Schmertmann

In NJ suburbs I refer to it as “hedge fund fours” — rich enough to have the luxury of 4 children (and a lot of help)

We (@thehauer and I) have an alternative estimate of TFR > 3 for those million-a-year-plus folks. We're still checking it carefully, though.

Try it for zip code 07078

I'm sure you're right, but there's no way to get zip-level data for super-high income people in an ACS sample.

I could just drive by and take photos of some families for you In front of their mailbox?

I thought we'd agreed you'd stop doing that, Laura.

reasons???????

The cost of children varies with income. Women in "middle" income households are most likely to be in the labor market, so they face higher $ costs for each hour spent on child-raising.

... and really rich women can hire help

Is that household income in the pre-birth year? Otherwise income would naturally fall with births due to leave taking. (The very high incomes obvs have something different going on))

Excellent point. ACS Qs refer to births and total HH income over the 12 months before the survey date. On avg births would have happened 6 months into the period for which income is reported. In short: you're right that recent births might cause lower recent incomes.

... but we see an even stronger U-shaped pattern in other estimates that are based on last-year income and births over the last five years. So it's unlikely that the interaction between recent births and recent incomes is the main story for the left side of the graph.

Wow I've been making this case to friends and family. What is the source - is there a link? Thx.

This is from the American Community Survey, a US Census product. This is accumulated data from monthly samples over 2012-2016, and uses the Qs on HH income and whether/not women had a birth in the past year. Raw data is at https://usa.ipums.org/usa/ 

Was this American Community Survey data recently released? Is there a link to this graph at their website or did you create the graph from their data?

I created it from ACS microdata, available from @ipums -- at https://usa.ipums.org/usa/  . (Microdata = anonymized sample of individual ACS responses).