How to Assess the Census Income and Poverty Numbers
August 28, 2006
Today, the Census Bureau will release findings regarding household income and poverty for 2005. It is possible these figures will show that median income increased in 2005 and poverty declined; that is the typical pattern for years well into an economic recovery. And if this is the case, Administration officials likely will hail the figures as good news (and seek to portray them as evidence of the success of Administration policies).
But such an assessment would be much too simplistic. To assess the new poverty and income figures entails examining not only the changes that occurred between 2004 and 2005, but also the degree of progress (or lack thereof) in income and poverty during the current economic recovery, and comparing any such progress to the progress made during comparable periods of past recoveries.
2005 marked the fourth year of an economic recovery. History shows that by the fourth year of a recovery, median household income always is well above -- and poverty always is below -- the levels attained when the economy was in recession and the recession hit bottom. This occurred in each one of the previous four recoveries (the recoveries for which both poverty and median household income data are available). In these recoveries, median household income in the fourth year of the recovery was an average of $2,127 in today's dollars -- or 5.5 percent -- above its level in the year when the last recession hit bottom. Poverty was an average of 1.0 percent point lower than its level at the last recession's trough.
How will 2005 measure up? It is distinctly possible that the new Census data will show that poverty declined and median income rose in 2005, but that poverty still was higher -- and median income lower -- than in 2001, when the last recession hit bottom. If so, this will be unwelcome news and hardly something to celebrate. It will mark the worst performance in recent decades for poverty and median income during an economic recovery.
Unfortunately, this could occur. The poverty rate increased each year from 2002 through 2004, even though those were recovery years. In addition, real median household income either fell or failed to rise in each of these years. 2004 marked the first time on record that median income was lower, and poverty higher, in the third year of a recovery than when the previous recession hit bottom.
Simply to return to recession levels (i.e., 2001 levels), the poverty rate for 2005 will need to fall from 12.7 to 11.7 percent, and real median household income will need to rise by 1.5 percent, to $46,569. If this does not occur, and poverty remains higher -- and income lower -- than in 2001, that will hardly be evidence that current policies "are working" (or that providing much larger tax cuts than in previous recoveries has produced superior results).
Moreover, if poverty remains higher and median income lower than in 2001 when the economy hit bottom, it will be the latest evidence that the current economic recovery has been remarkably uneven, with an unusually small share of the gains from the recovery reaching middle- and low-income families. Data recently issued by the Commerce and Labor Departments tell such a story; they show that a smaller share of the gains from the current recovery is going to workers' wages and salaries, and a larger share is going to corporate profits, than in any other recovery since World War II.
As Table 1 shows, in previous economic recoveries since the 1970s (the period for which data on median household income are available), real median income rose by between 2.0 percent and 8.7 percent by the fourth year of the recovery, relative to where it stood at the bottom of the previous recession. The average increase in real median income was 5.5 percent by the fourth year of the recovery. Similarly, poverty rates declined by between 0.4 percentage points and 1.4 percentage points by the fourth year of past recoveries, with an average decline of about 1.0 percentage points.
If current policies represent an improvement over the policies of the past and are producing better results, then the gain in median income and the decline in poverty should exceed the average improvement in previous recoveries. This is very unlikely to be the case when the new Census data come out tomorrow.
|Table 1: Changes in Poverty and Median Income During First Four Years of Economic Recovery|
|Recovery Period||Median Income (Percentage Change by the Fourth Year)||Poverty (Percentage Point Change
by Fourth Year)
|1960s (1961 - 1965)||NA||- 4.6|
|Early 1970s (1970 - 1974)||+ 2.0%||- 1.4|
|Late 1970s (1975 - 1979)||+ 8.7%||-0.6|
|1980s (1982 - 1986)||+ 8.3%||- 1.4|
|1990s (1991 - 1995)||+ 2.9%||-0.4|
|Current recovery (2001-2005)||To be released||To be released|
|First 3 years of current recovery (2001-2004)||-1.5%||+1.0|
 This analysis treats 1970-1974 as a single recovery period, although the recovery officially peaked in 1973. If the end point of the recovery, 1973, were used instead of 1974, the average improvement for recoveries shown here would be still higher. The analysis omits the brief one-year recovery of July 1980-July 1981.
 In the recovery of the 1960s, the first for which poverty data are available, the improvement was even greater.