Compare relative characteristics of June 2018 CPS to 2018 ASEC

[MaxGhenis]

This seeks to answer the question, does today's unemployed population resemble 2018's, just scaled up, or do they differ on important dimensions for distributional analysis? If it's the latter, we may want to do a more involved reweighting to simulate policy effects using the 2018 ASEC.

First step is comparing the distribution of DURUNEMP.

As a subsequent step, bucket DURUNEMP and compare it to other important factors for determining poverty status: marital status, existence of kids in the household, and some income concept, either directly (though this is endogenous to FPUC) or average wages by OCC occupational code.

[ngpsu22]

Struggling to figure out how count a person multiple times. I.e. if a person has a weight of 1200 DURUNEMP = 20, then I want 1200 people to appear in the histogram as unemployed for 20 weeks.

[MaxGhenis]

Try seaborn's distplot function, which accepts weights https://stackoverflow.com/a/32500044/1840471

On Thu, Jul 23, 2020 at 10:25 AM Nate Golden notifications@github.com wrote:

Struggling to figure out how count a person multiple times. I.e. if a person has a weight of 1200 DURUNEMP = 20, then I want 1200 people to appear in the histogram as unemployed for 20 weeks.

— You are receiving this because you authored the thread. Reply to this email directly, view it on GitHub https://github.com/UBICenter/covid_ui/issues/3#issuecomment-663132660, or unsubscribe https://github.com/notifications/unsubscribe-auth/ABOLNTYG2WZRSHQO2IZNHVLR5BW7DANCNFSM4PF5T4KQ .

[ngpsu22]

Unemployment seems to be getting shorter.

[ngpsu22]

These trends seem remarkably similar.