Oh you’re not talking about the FRED data, which examines real median personal income over time (which has been constantly increasing).
You’re talking about why the BEA data doesn’t line up with the US Census Bureau data for 2019. That’s probably because the BEA data is average growth while the US Census data shows the median income growth. For year 2019, we noticed that wage growth was largely driven by the lower quintiles: https://www.hiringlab.org/2019/03/05/february-jobs-report-pr...
This can largely explain the discrepancy. In other words, the assertion that “median would have increased even less” actually turned out to not be true, empirically.
Sure, we have seen "anomalous" years for median earners from time to time, but it's important to remember that over time real median income has still been increasing as a function of the business cycle — I.e. the first derivative has always been positive during bull economies (see: FRED historic data). Sometimes the second derivative is also positive, which isn't uncommon. In 2019, the second derivative was not only positive, it was very positive.
You’re talking about why the BEA data doesn’t line up with the US Census Bureau data for 2019. That’s probably because the BEA data is average growth while the US Census data shows the median income growth. For year 2019, we noticed that wage growth was largely driven by the lower quintiles: https://www.hiringlab.org/2019/03/05/february-jobs-report-pr...
This can largely explain the discrepancy. In other words, the assertion that “median would have increased even less” actually turned out to not be true, empirically.