“There are two economic realities in America today. There’s been a record six straight years of job growth, and new census numbers show incomes have increased at a record rate after years of stagnation. However, income inequality remains significant, and nearly half of Americans are living paycheck to paycheck.”
– Lester Holt at the first presidential debate on Sept. 26, 2016.
For Lester Holt, the economy has been performing well, but there is a problem in the distribution of income. This is the liberal position. The conservative position is that the economy has been underperforming, growing at a slow rate, with many in the middle class losing purchasing power, and many losing faith in their ability to improve their lives.
It is fair that a liberal brings his perspective to a debate when he serves as moderator. One might hope that the mix of moderators reflects a range of perspectives. But, what is not fair is for the moderator to confuse their perspective with objective truth. With regard to the statement that the new census numbers show an increase in income, expressed in the opening question, and then repeated as a “correction” of something Donald Trump said during the debate, it is not true. The truth is the new census numbers indicate that median income has probably increased, based on a “bridge estimate” for 2014, but remains lower than 2007 even with the new methodology.
For 2015, the report for which was released immediately prior to the debate, the census changed the way income is “calculated.” It now imputes income (makes it up) for those who don’t fully answer their questionnaire. Mostly, we’re talking of imputations on “unearned income” (interest, retirement income, and so forth). Indeed, the new census numbers show no real growth for “earned income” (wages and salaries).
The imputations are based on correlations of the respondent with a blank to other respondents who have provided answers. The assumption is that some people are simply not remembering the amount, so that the blank doesn’t represent a zero, but represents missing data. With the large number of people having blanks, these imputations can have a significant impact on the resultant numbers. According to an analysis of a prior year’s survey of income, the imputations increased median household income by 3 percent (with a standard error of 2 percent), so that most if not all of the 5 percent increase in median household income in 2015 was due to the change in methodology.
The change in methodology is not clearly stated in the 2015 report, but is discussed in a supplementary publication, which says that the Bureau of he Census will soon (i.e., after the election) publish an analysis of how the change in methodology affected their numbers. Indeed, to the contrary, the report itself prominently states (on p. 5):
“Median household income … increase[d] in real terms … 5.2 percent from the 2014 median … This is the first annual increase in median household income since 2007 … “
Given the fact that most if not all of the “increase” was due to a change in methodology, this report reinforces the conservative position; namely, that in spite of the recovery, the economy is underperforming, and many people are losing purchasing power.
The change in methodology is reflected in the report in the charts which distinguish calculations for 2013 to 2015 (with the figures for 2013 and 2014 based on “bridge estimates”), from the series up to 2013. In the long run, the exact timing of the 3 percent or so of new, imputed income during 2013 to 2015 might not matter much, although the odd increase during this period may make difficult analysis of the income dynamics of the period. But, in an election year, the reported change can be significant.
Lester Holt wasn’t the only liberal in the news media to jump to the conclusion that the economy actually is functioning well. The New York Times headlined, “U.S. Household Income Grew 5.2 Percent in 2015, Breaking Pattern of Stagnation.” As for the implications of this income growth for Trump: “Mr. Trump, the Republican nominee, has repeatedly cited the stagnation of household income as evidence of a broader economic malaise.”
When somebody pointed out that, in spite of the overall turn-around in income, farm income had fallen (and maybe this was why Trump is resonating with rural voters), the people at the census were quick to explain the dip was only due to a changes in the boundaries of metropolitan areas. When a comparison is made based on consistent definitions of rural and urban areas, the change in income was found to be nearly uniform across the country. The bureau therefore was removing comparisons of income by type of community between 2015 and 2014.
If nothing else, we can see that Donald Trump is already responsible for a recovery of manufacturing in this country, of statistics. To create income by imputation when that aligns with the liberal perspective, and to uncreate income by type of community when that doesn’t so align.
# # #
 Bernadette D. Proctor, Jessica L. Semega, and Melissa A. Kollar, “Income and Poverty in the United States: 2015.” September 2015. U.S. Bureau of the Census. http://www.census.gov/content/dam/Census/library/publications/2016/demo/p60-256.pdf
 See Appendix D of Carmen DeNavas-Walt and Bernadette D. Proctor, “Income and Poverty in the United States: 2014.” September 2015. U.S. Bureau of the Census.
 “Source and Accuracy of Estimates for Income and Poverty in the United States: 2015 and Health Insurance Coverage in the United States: 2015.” http://www2.census.gov/library/publications/2016/demo/p60-256sa.pdf