Wealth gaps in the Survey of Consumer Finances

In this exercise, we will look at the distribution of wealth across U.S. households in 2019, using data from the Survey of Consumer Finances. The file scf.csv contains data which were prepared by this code. We define wealth as household net worth. For future reference, you can access the full dataset through Berkeley Survey Documentation and Analysis.

Data and helper functions

You can load data with this line.

scf <- read_csv("https://info3370.github.io/sp23/assets/data/scf.csv")

You will want a weighted quantile function that accepts

• a vector x to be summarized
• a quantile q to report (e.g., q = .5 for the median)
• a vector w of survey weights (e.g., w = weight)

No need to code that yourself. By running the line below, you’ll get a weighted.quantile() function in your R environment.

source("https://info3370.github.io/sp23/assets/code/weighted.quantile.R")

Ratio of Black median to white median

First, we will calculate

1. the median wealth of Black households
2. the median wealth of white households
3. the ratio (2) / (1)

What do you think that ratio is? Write down a guess.

Point of reference: If we were looking at income, the ratio would be 1.7.

Now calculate the ratio.

Proportion with low and high wealth

Within each category of race, what proportion of have wealth

• less than $10,000?
• greater than $100,000?

Write down a guess, then calculate an estimate.

Challenge

For each wealth value $10k, $20k, …, $200k, what proportion of households have income above that cutoff

• among Black households?
• among White households?

Try making a ggplot() where the

• x-axis is the wealth cutoff
• y-axis is the proportion of households exceeding that cutoff
• color indicates race

Summary video: What we covered today