In the forty years between 1970 and 2010, real GDP per person doubled, the U.S. spent trillions of dollars on anti-poverty measures, but the poverty rate increased by two percentage points. Just yesterday the Census reported 46.2 million people living in poverty in 2011.

That's the official story. But a new paper from Bruce D. Meyer and James X. Sullivan says it's missing everything. "We may not have won the war on poverty, but we are certainly winning," they write.

When they looked at poorer families' consumption rather than income, accounted for changes in the tax code that benefit the poor, and included "noncash benefits" such as food stamps and government-provided medical care, they found poverty fell 12.5 percentage points between 1972 and 2010.

The graph below tells the story. The official poverty rate (shown in DARK BLUE) is higher today than it was in the early 1970s. But when you measure after-tax income (RED) or consumption (GREEN), the story changes: Poverty rates have declined steadily, and dramatically, since the 1960s.

What's the story? Tax cuts for the low-income combined with tax credits for parents (such as the child tax credit) and working poor (like the Earned Income Tax Credit) accounted for much of the collapse in poverty, the authors find. Increases in Social Security also helped poverty rates fall under 10 percent by the new measure. For all the attention we give to taxes falling on the top 1% since the Reagan administration, it is equally true that effective tax rates on the bottom quintile have been halved in that time.

 
It sounds disastrously wrong to claim that we're winning the war on poverty 24 hours after the Census reported more people living in poverty than the combined populations of California and Missouri. But this report isn't suggesting that the war against poverty is over, merely that it is advancing. Even after a generation of stagnating market wages for men and lower-class workers, it turns out that taxing the poor less and spending more money on them is having its desired effect.