Median Household Income in October 2019

Posted: Dec 04, 2019 10:40 AM

Median household income in the United States reached a new record high in October 2019, with the typical American household in $66,465 according to demographic analytics firm Sentier Research. That figure is up slightly from the 66,214 the firm reported for September 2019. The new record high applies both in nominal terms and after adjusting for the effects of inflation.

The following chart shows the nominal (red) and inflation-adjusted (blue) trends for median household income in the United States from January 2000 through July 2019, with July 2019's estimate marking the top of both scales. The inflation-adjusted figures are presented in terms of constant October 2019 U.S. dollars.

Year over year, median household income rose by 5.1%, or 2.4% after adjusting for inflation, from the figures recorded in October 2018.

Analyst's Notes

Our alternate methodology for estimating median household income from data reported by the U.S. Bureau of Economic Analysis would put the figure at $65,752 for October 2019, which is about 1.0% lower than Sentier Research's estimate for the month.

Sentier Research has published a comparison of its inflation-adjusted monthly median household income estimates with the annual estimates published by the U.S. Census Bureau. This chart is taken from that report as it appeared on 2 December 2019.

Gordon Green discusses what the comparison of Sentier Research's monthly estimates and the U.S. Census Bureau's annual estimates reveals:

Because the Census Bureau estimates are the official estimates and are considered to be the gold-standard, the operative question is, “How closely do the Sentier Research estimates approximate the Census Bureau estimates?” The answer is “pretty darn close.” The following chart shows a trend line of the Sentier Research monthly estimates of median annual household income in red and a trend line of the Census Bureau official estimates of median annual household income in blue. The Sentier Research trend line hows the actual month-to-month movement of the series whereas the Census Bureau trend line appears as a step ladder function in which each month is essentially the same value for the entire calendar year. It is readily apparent from a visual inspection of the graph that the Sentier Research series closely parallels the Census Bureau series, moving down, flattening out, and moving up in tandem during the same time sequence. The Sentier Research series naturally shows more volatility than the Census Bureau series because it is measuring the actual month-to-month movements, which is to be expected, whereas the Census Bureau series is flat for the entire calendar year because there is no information on month-to-month movements. The conclusion is incontrovertible – the two series are closely measuring the same phenomenon.

If the Sentier Research series has closely paralleled the Census Bureau series for the past 18 years, it is reasonable to conclude that it is giving a good approximation of income movements in more recent time periods for which the Census Bureau income data are not available. Therein lies the true value of the Sentier Research series because it can closely measure income changes in real time, with a lag of only about one month, whereas one has to wait almost a year or substantially more (as time passes) to obtain the Census Bureau’s annual estimate of median annual household income for the previous calendar year. Moreover, the Sentier Research monthly estimates can detect important turning points in the time series, such as the beginning or ending of a recession, or the precise month when income bottoms out or tops off, something which the annual income series is incapable of doing. The Sentier Research series has value because decisions by policy makers and the business community are made in real time, not by looking back to what happened last year.

In the chart, the greatest deviation between Sentier Research's monthly estimates and the U.S. Census Bureau's annual estimates occurs during 2015, which coincides with when the U.S. Census Bureau was implementing changes to the questions asked in its Current Population Survey questionnaire related to household income. Here, although the redesigned survey was fully implemented in the Census Bureau's March 2015 questionnaire, which was used to generate its annual estimate for this year, the Census Bureau didn't provide an updated interviewing manual for training its field employees about the redesigned survey until April 2015, which we suspect may account for the relatively large discrepancy that developed between Sentier Research's monthly estimates and the Census Bureau's annual estimate during this time.

We'll also observe that Sentier Research's estimates track very well with the data we use in developing the median household income estimates we generate using our alternate methodology, where the aggregate income data we use in generating those estimates is produced by the U.S. Bureau of Economic Analysis.