The evolution of higher incomes in the USA: 2019 and 2020 comparison

 Personal income (PI) is the key economic parameter. The aggregated personal income has to be equal to the GDP if to count all sources of income as personal. At the end of the day, all money belongs to physical persons in one of the numerous ways. In reality, there are so many official and unofficial ways to count personal income that it is hard to find a person with full knowledge of all details of all the measurements. Here, we compare two measures of personal income as reported by the US Census Bureau (CB) and by the Bureau of Economic Analysis (BEA).  Figure 1 presents the evolution of the (nominal) personal income reported by the CB and BEA since 1967. The deviation between these two curves in relative terms is shown in Figure 2 (the sources of personal income differ between the BEA and CB, e.g. the BEA includes the imputed rent). One can see that the CB/BEA ratio has been slowly growing from 0.8 in 1970 to 0.88 in 2019. The years after the Great Recession (2011-2019) are characterized by slightly higher ratios than at the beginning of the 21^st century. The years of the Great Recession are well seen in Figure 1 but they produce different effects on the CB and BEA curves. For the BEA curve, only the year 2009 is characterized by a fall and then the curve is back to growth. The CB curve has a shelf between 2007 and 2010: people reported no money income growth during these years, but the BEA includes some other income sources, e.g. special kinds of social security money transfer as shown in Figure 3.

 The Current Population Survey of the CB gives an expected result for 2020 – almost the same PI as in 2019 (Figure 1). This means that the personal income distribution, PID, should not be affected much. One of the most sensitive parameters of the personal income distribution is the number or share of people with high incomes. Figure 4 presents the age-dependent curves for the number of people with a nominal personal income above $100,000. In 2020, no overall growth is observed and the change in each group is likely within the uncertainty limits of the measurements in the CPS ASEC (March) Supplement. Figure 5 illustrates the fact that the growth in population with income >$100,000 in 2020 was small in all age groups. Only young people likely demonstrated reliable income growth. Elder people are rather characterized by income decline relative to 2019. 

Because of the population growth, inflation, and real economic growth, the number of people increases with time in all age groups between 2010 and 2019. Relative growth is a more reliable characteristic in this case. Figure 6 presents the share of people with income >$100,000 in the same age groups. The basis is the number of people with income (a total number less the number without income or loss) and it varies between the age groups as Figure 6 shows for 2020. Finally, the shapes of the relative shares of people with income >$100,000 are shown in Figure 7. All curves are close to each other, as expected for the power (Pareto) law income distribution for the higher incomes. 

The question is why the income from the government supposedly was not counted in the money income of the Census Bureau in 2020. The government transfer under “Social Security” and “Others” was larger than 1 trillion, and people obtained part of this money in cash. The income distribution was likely not affected. 

Figure 1. Personal income measured by the CB and BEA since 1967.

 

Figure 2. The ratio of the personal income estimates reported by the CB and BEA since 1967

 

Figure 3. Sources of personal income used by the BEA but likely not used by the Census Bureau.

Figure 4. The evolution of the number of people with income >$100,000 as a function of age between 2010 and 2020.

Figure 5. Share of people with income >$100,000 as a function of time in various age groups

 

Figure 6. The share of the number of people with income in various age groups.

 

Figure 7. The evolution of the share of people with income >$100,000 as a function of age between 2010 and 2020.  The curves in Figure 4 are normalized to their respective peak values.