earned income by linear age.R

library(survey)
library(lodown)

cpsasec_cat <- get_catalog( "cpsasec" , output_dir = tempdir() )

lodown( "cpsasec" , subset( cpsasec_cat , year == 2017 ) )


plot( 
	c( 15 , 75 ) , 
	c( 0 , .03 ) , 
	type = 'n' , 
	axes = F , 
	xlab = '' , 
	ylab = '' , 
	main = 'share of total nationwide earned income\nreceived by individuals of different ages' ,
	sub = '2016 earnings from the 2017 cps-asec'
)

axis( 1 )
axis( 2 , at = c( 0 , 0.01 , 0.02 , 0.03 ) , labels = c( '0%' , '1%' , '2%' , '3%' ) )

cpsasec_df <- 
	readRDS( paste0( tempdir() , "/2017 cps asec.rds" ) )

cpsasec_design <-
	svrepdesign(
		weights = ~ marsupwt ,
		repweights = "pwwgt[1-9]" ,
		type = "Fay" ,
		rho = ( 1 - 1 / sqrt( 4 ) ) ,
		data = cpsasec_df ,
		combined.weights = TRUE ,
		mse = TRUE
	)

overall <- data.frame( svytotal( ~ pearnval , cpsasec_design ) )
	
by_age <- 
	data.frame(
		svyby( ~ pearnval , ~ a_age , cpsasec_design , svytotal )
	)

share_by_age <- by_age[ by_age$a_age %in% 15:75 , 2 ] / overall$total

lines( 15:75 , share_by_age )