July 14th, 2023
Because of a one-click Google search short cut, many are claiming the wages in the Great Depression were higher than today after adjusting for inflation.
As you can see Google links to an IRS sourced number claiming a little north of $4k for annual income in the early 1930’s. After adjusting for inflation, this becomes $91k in 2023 dollars, which is far higher than the $75k or so average wages & salaries in 2021 (the most recent year of data available).
The issue, of course, is the IRS data is NOT average salaries for all workers, but rather the average taxable income for those who filed. In the 1930’s a relative small portion of the population filed taxes, and that proportion was a far richer slice. (See details in this video).
Beyond the fact that the Great Depression was far worse than anything we’ve seen in the economy since (See this video for a plethora of details), the message has resonated with millions. Why? My guess is because there still exists hardship for many people, despite all of the progress we have made since. Second, the level of inequality we currently experience is, indeed, comparable to the Great Depression. The bottom 50% of the population receives around 14% of the income. This compares to 20% of the income in 1970.
But, as this viewpoint continues to gain in viral popularity, it requires further careful data work to discern the truth.
Recently @NickPowers, a “data-guy and educator” claims “we are currently making less money than they did during the Great Depression.”
Let’s go through the claims. Thankfully, Nick Powers has done some work to cite his sources on his website.
First: “Median Household Income.”
This is the crux of the whole argument. He compares 2022 median household income with the 1940s in the context of specific purchasing power. But, what are the data sources, exactly?
From 1950 through 1980, he cites the National Center of Educational Statistics’ compilation of Census Data for Median Family Income. For 1940, he cites Prologue Magazine’s description of the Census. $1,368 is the “average income”. This is not the median, but the mean. Given society’s wage inequality, the mean will be significantly higher than the median, thus making 1940 appear more relatively affluent than later years. This is not a comparable definition to the rest of the table.
After 1980, it appears Nick Powers is using “Real Median Household Income,” rather than “Nominal Median Household Income.” In 1990, Powers has $57,677. This figure only comes after adjusting the nominal values of $29,943 by the 2020 CPI-U-RS. If you go to the current publications of Real Median Household Income, you’ll find they are now adjusted using 2021 dollars, so the FRED link will not feature the same numbers. However, lucky for our verification project, you can find a historical table, here, from the Census that has the 2020 dollar adjustment. The numbers from that table line up with those in Nick Powers’ video. Finally, the 2022 figure, $54,132, is not the median household income, as the Census has yet to publish that number. Rather, it comes from median individual wages & salaries. In the 2Q 2022, median weekly earnings were $1,041. Multiply this by 52 weeks, and you get $54,132.
There are two issues. First, these are individual median wages, not household wages, the earlier years will be higher by virtue of dual income households. The second is far more serious for the purposes of comparison with the 1940 figure. The median earnings are much lower than the mean earnings. If we use the same data source as the 1940 average annual wages & salaries, we find in 2021, that figure to be $75,447. The difference is substantial! And because the levels of inequality are similar today as they were then, the difference between median and mean would likewise be as dramatic in 1940.
Next Item: Rent.
Again, Nick Powers compares apples and oranges in the same data series. From 1940 through 2000, the gross rent data comes from the Historical Census of Housing Tables. For 2005, the gross rental figure comes from the American Housing Survey. For example, in 2005, you can find the number $694 by going to table 4-13 and look at monthly median housing cost for renters. The number for 2015 also matches: $923. However, the American Housing Survey only runs every odd year. Therefore, it isn’t clear where the 2010 and 2020 figures. The cited source is IPropertyManagement.com. But the currently published numbers differ from his. Using the WayBackMachine, we can see that his numbers line up with this back in a 2022 capture. Unfortunately, while IPropertyManagement does give a group of sources at the bottom of the article, I was unable to find the exact numbers.
The issue, for our purposes, is the appropriate 2022 comparison. His cited figure of $2,007 comes from the November 2022 Rent.com data. Unforunately, they do not use a comparable definition. The Rent.com figure is defined as newly contracted rental units. It does not include any existing rental contracts, which form the vast majority of the economy’s rental payments. We can see that the 2020 figure from Rent.com (above $1,600) do not line up with the IPropertyManagement data ($1,104). This is a stark contrast and not a comparable measurement.
The American Housing Survey has a median rent of 2021 figure of $1,184 and the American Community Survey has a median gross rent of $1,163. These are in the ballpark for estimated median rent reported by ApartmentList.com. Therefore, a reasonable comparable median rental number for 2022 is $1,349, which is still an extraordinary annual increase.
Now let’s compare the rental affordability in 1940 to 2022.
For income, we will use the National Income Accounting’s Wage and Salary per Full Time Employee Equivalent. This is better than GDP per Capital or Disposable Personal Income, as it only includes labor income, rather than capital income. It is unfortunate that we do not have median income data going back to 1940, but we need to compare apples with apples, so we will use mean wages from 2021. The 2022 figure has yet to be published. Because wages also experienced fast nominal growth during the inflationary period of 2022, our current analysis, here, will be an upper bound for affordability. It’s important to use the 2022 rent figures, as that is when the large imajor increase occurred.
| Average Wage & Salary | Monthly Median Gross Rent | Rent Income Share | |
| 1940 | $1,317 | $27 | 24.6% |
| 2021/22 | $75,447 | $1,349 | 21.5% |
The claim that folks had it easier in 1940 compared to today does not hold with careful comparing similar definitions. We also see a similar 22% housing share of income for the average worker when we look at the Consumer Expenditure Survey data.
A further issue is not controlling for the quality of housing. Since 1940, houses are now larger, better insulated, we nearly universally have refrigerators, indoor plumbing, flushing toilets, electricity, air conditioning and central gas/electric heating, as well as modern ovens and ranges. In 1940, 43.8% of housing did not even have their own private bathtub or shower. Not to mention the internet connectivity. The BLS takes these improvements into account when tracking inflation through time. If the quality improves, but the nominal price does not, then consumers are better off (the same can work in reverse, if quality deteriorates, but the nominal price remains, consumers are worse, ie: shrinkflation). For further details, see the the BLS explanation of hedonic pricing techniques and other technical methods to control for quality improvements through time.
If we compare the CPI for Rent since 1929, we see that mean nominal wages and salary grew far faster than rents. Keep in mind that this measure of rent is attempting to keep quality constant. It has only been in the past 10-20 years that housing rents have grown faster than average wages
Now, none of this is to say everything is Utopiaic today. Our economy is performing far below potential, and we have far too much economic suffering, given our immense capacity and national wealth. Real median wages were basically flat for 30 years, and while they have grown some since 2014, it is not enough to catch up with our potential. There is growing economic research showing that labor markets are not competitive and that firms have more bargaining power compared to employees.
But, to address these issues, we need to accurately describe them. Polemic and inaccurate comparisons about the Great Depressions and today do not help.
Econ Chris Clarke 