Did Inequality Fall Between 1870 and 1910?

By Vincent Geloso. 

"The period from 1870 to 1910, which includes the Gilded Age and Progressive era, is depicted everywhere as one where there was rapid economic growth. This growth is commonly seen as rapidly and unevenly distributed with the poorest 90% enjoying far fewer improvements. 

This popular conception is probably wrong because of the way we are using the existing data on inequality. In fact, inequality between the top 10% and bottom 90% of the income distribution in America most likely declined. 

Let us look at that data. Right now, there are no actual estimates from 1871 to 1909. What there is an estimate in 1870 and another one for 1910. 

The estimate for 1870 comes from the work of Peter Lindert and Jeffrey Williamson. They used what is known as a “social table.” This approach estimates income distribution by assigning average incomes to different social or occupational groups based on historical records, such as census data. It simultaneously provides total national income and income inequality.  That estimate is not disputed, and if anything, it probably understates inequality because of the problems of census underenumeration of the poor. 

The estimate for 1910 is derived from the work of Thomas Piketty in his famed Capital in the 21st Century, and is not directly based on income data. Instead, he used tax records from 1917 to estimate the top 10% and data from 1913 for the top 1%, then backcasted these figures to 1910. 

The problem is that the estimates that Piketty (and his co-author Emmanuel Saez) created for 1913 and 1917 are now known to massively overestimate inequality. In an article in the Economic Journal with Phil Magness, John Moore and Phillip Schlosser and in a companion article with Phil Magness in Economic Inquiry, we corrected for these errors. In fact, we showed that their entire series from 1917 to 1962 was flawed because of the way missing filers were treated, how net income was converted into adjusted gross income, and how they forgot that state and local governments (5% of the workforce with incomes well above the national average income) were not required to file federal taxes until 1938.  

We also discovered that Piketty and his co-authors made an incorrect estimate of total income. They arbitrarily defined total income as 80% of personal income (as reported by national accounts) minus transfers. They justified this by claiming that “the ratio between total gross income reported on tax returns and personal income minus transfers in national accounts has been fairly stable since the late 1940s (around 75-80%).” However, according to their own datasheets, the actual average was 82.7%. While this difference may seem small, a higher proportion reduces the income shares of the rich. Using less arbitrary methods, we found a much larger denominator and, consequently, smaller income shares for the wealthiest groups.

Overall, we found that the income share for the top 10% was 5 percentage points lower than that of Piketty for 1917. If one uses the exact same backcasting method as Piketty did in his book, you also get a lower share of total income going to the richest 10% of Americans. When this is then combined with the estimates of Lindert and Williamson for 1870, we see a significant decline in inequality as can be seen in the figure below. 

This is rich in implications. Consider that economic growth, depending on the data series used, showed that Americans enjoyed income increases that averaged between 1.9% and 2.0% per annum between 1870 and 1910. By that point in history, never had such fast growth been observed. And when there had been growth that nearly matched that one, it was clearly marked by rising inequality. 

Finding that the bottom 90% got a bigger part of the pie implies that the bottom 90% probably saw improvements that were larger than the average gains. Given the numbers I obtained, this means that the income of the poorest 90% increased between 2.0 and 2.2% every year. 

This difference implies that a period often characterized as one of rising capital concentration and inequality was, in fact, not only the fastest and most sustained growth ever observed by that time but also the first in history where the poor saw their incomes grow faster than average, experiencing significant improvements in their standard of living. 

Some contemporary writers of the 19th century noted that there were exceptional gains at the bottom of the income ladder. It seems we overlooked them, in part, because we relied on data that misled us, obscuring the real progress made by those at the lower end of the income  distribution which contemporaries saw with their own eyes."