Students eat lunch inside the school's gym at Little Fort Elementary School in Waukegan, Ill. For schools in the poorest 25 percent of Illinois districts, per-pupil funding stalled at around $13,500 in 2014. (Kamil Krzaczynski/Associated Press)

America is on the mend. Witness the good news in the latest version of the nation’s “economic report card”: the Census Bureau’s annual estimates of the median household income and the poverty rate.

Here are the crucial numbers. In 2015, median household income — the midpoint, with half of households above and half below — rose 5.2 percent to $56,500 from $53,700 in 2014. That’s the largest increase since the 2007-2009 recession, though household income remains below the pre-recession peak of $57,423 in 2007. The story is similar for the poverty rate. It fell to 13.5 percent in 2015 from 14.8 percent in 2014, though (again) that compares unfavorably with the 2007 level of 12.5 percent.

Still, not all the news is good. There’s one number in the “report card” that always shocks. It’s the child poverty rate: the share of those younger than 18 living in households with incomes beneath the official poverty line ($24,257 for a family of four). The child poverty rate for 2015 was 19.7 percent.

This is a horrifying figure, if believable. A fifth of the nation’s children are growing up in households that, at best, are living hand-to-mouth. The poverty rate for children is the highest of all age groups. By contrast, the poverty rate for people 65 and older is only 8.8 percent. If children had the same poverty rate, only 1 in 11 would be poor, not 1 in 5.

Could this be? Can we trust the official figures? Maybe not. Anti-poverty experts on both left and right agree that the official poverty rates — for children and the entire population — are misleading, because they exclude large sources of actual income. Economist Scott Winship of the right-leaning Manhattan Institute recently estimated that the child poverty rate might be 5 percent or lower if all income sources are counted.

Here are Winship’s six steps that reduce the child poverty rate. He starts with the official poverty rate of 21.1 percent, which is based on households’ receipts of cash income: wages, salaries, Social Security, unemployment insurance and other programs that pay benefits in cash. (All the figures are for 2014 and would be somewhat lower for 2015.)

1. He first adds non-cash benefits — not counted by Census — such as food stamps, subsidized school breakfasts and lunches, housing subsidies and energy subsidies. These cut the child poverty rate to 16.2 percent.

2. He then adds tax refunds, which further reduce the poverty rate to 13.2 percent. Both the Earned Income Tax Credit and the Child Tax Credit are refundable: People get payments from the Treasury even if their incomes aren’t high enough to be taxed.

3. The next adjustment involves cohabitation — two adults living together, though they aren’t married. This affects about 15 percent of households with children. The extra income of the second adult reduces the child poverty rate to 11.4 percent.

4. The choice of an inflation indicator to adjust for price changes cuts the poverty rate to 9.7 percent. Census now uses the consumer price index (CPI), which many economists think overstates inflation. Substituting another index based on personal consumption expenditures (PCE) corrects this alleged bias.

5. Winship next adds in the value of government health insurance for poor children, mainly through Medicaid and S-CHIP (State Children’s Health Insurance Program). This cuts the child poverty rate to 7.8 percent.

6. Finally, studies find that poor households underestimate the value of some means-tested government benefits. Including the extra value reduces the child poverty rate to 4.5 percent.

Naturally, some of Winship’s changes are controversial, but a fair reading is that the true child poverty rate is probably between 5 percent and 10 percent — not the 20 percent that Census finds.

What we learn from this is that the official poverty figures give a useful picture of year-to-year changes — but not necessarily of long-term trends, which are often not captured by the official statistics.

There’s good news and bad news here. The good news is that the safety net succeeds in protecting against extreme hardship. This should be a source of some satisfaction and self-congratulation. The bad news is that this doesn’t automatically empower people to escape poverty and join the middle class. This should be a source of much worry and regret.

Read more from Robert Samuelson’s archive.