In partnership with the Church, IPUMS (Integrated Public Use Microdata Series) has recently made the entire 1850-1890 set of census data available in tabular (spreadsheet) form for analysis. While individual records have been available for some time, as has a 1% sample of the quantitative data, this new development allows us to download all of the census responses for the 19th century at once. As you can imagine, this is a fairly large file, but if you subset Utah it is much more manageable. The wonderful IPUMS folks have harmonized the different questions asked across time so that you can make comparisons across decennial censuses.
In a previous post I discussed race in early Utah. In this post I’ll discuss what the Census Bureau data has to say about inequality (or equality, as we’ll see) in early Utah. As far as I can tell, the IPUMS data doesn’t have much in way of economic variables that extend all the way from the 19th into the 20th centuries. The exceptions are occupation scores, these are numeric socioeconomic scores that are assigned to particular occupations based on 1950s data. Here we use the Duncan Socioeconomic Index in particular. There is some controversy about these measures that I’m not terribly well read up on, but I see them enough that I assume they have some validity. I’m sure cross-century comparisons also complicate some things, so here I’m including various additional states as a sort of point of comparison. For inequality in particular I will look at the variation in the Duncan SEI. High variation obviously means greater inequality, with lower variation meaning more equality. Since we are looking at the entire census and not just a sample I will be using a simple standard deviation to measure variation/inequality.
As seen in the above graph, Utah had relatively high levels of equality as measured by occupational status for the first couple of decades of Latter-day settlement. Ultimately this isn’t super surprising since (I assume) just about everybody was a poor farmer. Once the railroad came in and the Utah economy started diversifying there was probably additional room for socioeconomic differentiation and by 1900 Utah had become more unequal than the comparison states. It is worth noting that looking at total assets would probably tell a loss rosy story since there is room for highly leveraged points (i.e. very wealthy religious/social elites increasing the overall variation) than there is a on a scale from 1-100. There is some census data for property values, but only for a few decades in the 19th century. Another post for another day.
Code
library(dplyr)
library(tidyverse)
if (!require(“ipumsr”)) stop(“Reading IPUMS data into R requires the ipumsr package. It can be installed using the following command: install.packages(‘ipumsr’)”)
setwd()
ddi <- read_ipums_ddi(“usa_00016.xml”)
data <- read_ipums_micro(ddi)
#Per codebook, 0s are NAs
df<-subset(data, SEI!=0)
summary_df <- df %>%
group_by(STATEFIP, YEAR) %>%
summarise(
mean_SEI = mean(SEI, na.rm = TRUE),
sd_SEI = sd(SEI, na.rm = TRUE)
)
Very interesting…I think what we’re mostly seeing here is diversification of occupation in all states. As the US economy modernized, the proportion of the population who were famers fell and people entered a huge array of other occupations instead. It makes sense that Utah started low, as they basically had to build an economy from scratch starting in 1847. I’d say that by 1940 the states had pretty much converged and their rankings at that point probably don’t mean much: if we could generate confidence intervals for those numbers I’ll bet they’d overlap.
That story is obscured by what I suspect are some data issues (i.e. issues in the source data, not anything wrong with what Stephen C did). I don’t trust the point for Wyoming in 1870 at all, and I wonder if something changed in the data collection in 1880. But I think you can see the effect of the end of slavery in the dip for Virginia and Texas in 1870: a lot of enslaved people who presumably didn’t have occupations in the 1860 census became farmers in the 1870 census. That probably doesn’t explain all of 1870 though, as Vermont dips too (but a lot less).
I’m not sure we can use these data to say much about how inequality changed over time. For example, there was (and is) a lot of inequality among farmers–just ask Joseph Smith and Martin Harris. Also, Duncan’s estimate of farmers’ socioeconomic status in 1950 probably doesn’t hold up very well in 1850. But it’s not like there are a lot of great alternatives in 1850.
(Data geek note: R 4.1 added a pipe operator to base R that acts pretty much the same as the tidyverse pipe operator, but because it’s built into base R it’s faster. Just replace “%>%” with “|>”.)
I agree completely; this is mostly a story of economic specialization and industrialization. I do think it means something that in the mid-19th century Utah didn’t have much room for differentiation of labor, but your critiques are well taken re variation among farmers and the iffiness in applying the 1950 SEI scores to the 19th century. Again, I’m not an early Utah historian, but until the miners came I suspect in a sentence Utah could be described as a bunch of subsistence farmers with some rich farmers thrown in.
Thanks for the note about the new operator, I had no idea!