Paul Krugman gets a lot of flak from the right about cherry-picking numbers and leaving out inconvenient facts when he makes his case for Keynesian stimulus. I tend to disagree with these critiques, to put it mildly, but I do think nice-looking charts and graphs can disguise a poorly-sourced argument. Even when I agree with Krugman’s overall point — that Republican-driven austerity, particularly at the state level, is damaging the economy — I sometimes question the evidence he marshals to support it.
This graph has been popping up on several left-wing websites, from The New Republic to Greg Sargent’s blog at the Washington Post.
Krugman posted it on his blog in advance of his Monday column, in which he argues that the narrative of President Obama as a big-spending socialist is an invention of the conservative message machine. In fact, the Republican prescription for the economy — lower spending and lower taxes — is “precisely the policy we’ve been following the past couple of years.” He explains:
What do I mean by saying that this is already a Republican economy? Look first at total government spending — federal, state and local. Adjusted for population growth and inflation, such spending has recently been falling at a rate not seen since the demobilization that followed the Korean War.
This isn’t an opinion; it’s a fact. Stimulus money ran out by 2010, and nearly 610,000 public jobs — teachers, police officers, etc. — have been lost since 2009, contradicting Mitt Romney’s claim that the stimulus package “didn’t help private sector jobs, it helped preserve government jobs.” (The 145,000 added “government jobs” were almost entirely at the federal level in departments like defense and homeland security, which Romney actually wants to expand.)
Yet the graph Krugman offers doesn’t necessarily support this point. Government spending per capita seems to be an accurate enough measure of how much the spending is occurring under Obama, but the chart shows not absolute amounts but “percent change from year earlier.” When the line dips into negative territory, the casual viewer (at least this casual viewer) assumes that spending per capita has decreased below whatever baseline the x-axis represents. (In this case, it looks like spending in the year 2000 is the baseline.) But all we’re really seeing is how much spending changed from one year to the next. Zero means no change from the previous year; when the graph dips to negative 3%, it doesn’t necessarily indicate that spending is lower than the 2000 baseline. It’s simply lower than whatever it was the year before. Even at right-hand side of the graph, in 2012, when the line descends precipitously, per capita spending is still higher than it was in 2000 — the ostensible baseline, when the line hits at zero instead of negative 3.
Though relying on my high-school math skills is always a dangerous proposition, I think this scenario is an accurate interpretation of what Krugman’s graph shows:
- Assume the government spends a certain amount — say, $100 — on each person.
- At the graph’s highest point, toward the end of 2009, per capita spending increases by 6% from the previous year. It’s now $106.
- By 2012, per capita spending dips by 3%. Three percent of $106 is $3.18, so spending now stands at $102.82.
- $102.82 > $100. Obviously.
- That means per capita spending is still higher, by about 20%, than at the beginning. Not exactly the massive austerity that Krugman purports to show.
Even though the graph is technically correct — after all, the y-axis is clearly labeled “percent change” — it seems misleading. When I see a chart going into negative territory, I assume (perhaps naively) that I’m seeing an absolute decrease. However, Krugman’s graph disguises the fact that spending, while lower than at the height of the stimulus, has actually increased.
This is not an argument for lower government spending. On the contrary, it’s only reasonable that spending increase during a recession, when demand for safety net programs like unemployment insurance and food stamps soars. And when the private sector is pulling back, government should step in and spend even more to make up the difference. Add in the effect of a growing elderly population, which naturally requires increased outlays for Medicare and Social Security, and it’s no surprise that government spending as a percent of GDP is going up. The Republican proposal to cap it at 20% is not only unreasonable but cruel. So don’t take this as a case for lower spending. Take it as a case for better, more responsible charts.
. . . So Well Read. . . 