Statistics matter. Take the newly popularized statistic, the “under­employment rate,” that the media began to report during the Great Recession. This rate adds part-time workers who can’t find the full-time work they want and people who have given up looking for work altogether to the officially unemployed. In April the underemployment rate reached 17%, indicating that workers are feel­ing the economic downturn more severely than suggested by the 9.9% official unemployment rate. The official poverty line similarly misses the mark in measuring economic suffering in the United States. Poverty experts have long recognized that the poverty income threshold published by the Census Bureau marks an excessively severe level of economic deprivation. But the problem isn’t just that the pov­erty line is too low. Anyone can multiply the poverty line by two and use that to define poverty; that’s how eligibility for subsidy programs like reduced-price school lunches is determined. The offi­cial poverty line, however, actually rep­resents a living standard that deterio­rates year after year.

The problem is that the Census Bureau adjusts the poverty line every year for overall inflation with the “Consumer Price Index for All Urban Consumers” (CPI-U). To calculate the CPI-U, a team of Census Bureau sur­veyors asks consumers about their purchases and determine the cost of the typical “basket” of goods and ser­vices purchased. The change in the price tag of this basket from one year to the next is the overall inflation rate. The flaw in using the CPI-U to adjust the poverty line is that the budget of a low-income household is different from that of a middle-in­come household. We can see this by comparing the Basic Family Budget produced by the Economic Policy Institute (EPI) and average consumer basket used by the DOL to track in­flation (see pie charts).

Two expenses—health care and child care—take up greater propor­tions of EPI’s Basic Family Budget than in the DOL’s average consumer bas­ket. This is because the DOL’s basket reflects spending by the average household, not specifically house­holds with children. The Basic Family Budget also assumes, reasonably, that all adults in the household need to hold jobs (thus the high child care expenses). Moreover, non-essential items such as recreation activities and extra clothing that an average con­sumer may buy do not fit into a bud­get restricted to basic needs.

These differences between the bud­gets are crucial. Child-care and health-care costs have been rising dramati-cally faster than overall inflation. According to the DOL, between 1999 and 2009 the cost of child care rose by 57% and the cost of health care rose by 50%. This compares to a general rise in prices of 29%.

Here’s a simple exercise to demon­strate what this means. The 1999 official poverty line was $16,895 for a family of four. Double this number is a more rea­sonable poverty line of $33,790. The Census Bureau raised the poverty line by 29% between 1999 and 2009--the amount of inflation measured by the CPI-U. Twice the 2009 poverty line there­fore is $43,590. The income required to support the same standard of living in 2009 as in 1999, however, requires a big­ger, 39% adjustment up to $46,970 to reflect the rapidly rising costs of child-care and health care. In other words, in 2009, a low-income four-person house­hold would need to bring in another $3,400 over and above the inflation-ad­justed $43,590 to maintain the same standard of living in 2009 as in 1999. This $3,400 deficit in the CPI-adjusted figure is equal to over 2/3 of this family’s medi­cal care budget.

This past March, the Census Bureau announced that it is undertaking a major project to reconsider how it measures poverty. This is a welcome development. Without revising the official poverty line to reflect the actual costs of families’ basic needs, the key statistic we use to understand economic deprivation in the United States will not only undercount the poor, but it will do so by a larger mar­gin every passing year. As the saying goes, “Statistics can be made to prove anything—even the truth.”