Blow This Graph, Matt Yglesias!

Yglesias Blows Graphs, Part II: Matthew Yglesias says he never said private jobs losses “since the beginning of the recession” were smaller than public job losses, though I also never said he did. Yglesias does argue that the public sector has taken a bigger hit lately, specifically since the start of the Obama administration. But why pick that period? It’s often possible to choose a period where what you’d like people to think has been tragically declining (e.g. public sector employment) has in fact been declining even though in a slightly wider view exactly the opposite is true. It’s still deceptive. That was Yglesias commenter Peter Schaeffer’s point. To see how deceptive, it helps to look at the whole post-2000 picture. Here it is (from Schaeffer**):Obviously the recent public sector cutback is a modest correction to the dramatic growth in government employment since 2000. As Schaeffer notes, there’s often a time lag between when a recession hits and when government employment finally contracts. You can always take a snapshot of only the period of lagged decline (the far right portion of the chart), ignore what went before, and say “see, the public sector is dangerously shrinking.” And I guess Yglesias-style Democrats always will.

Meanwhile, in the bigger picture, the larger private sector has been down, up, way down, and is now clawing its way back only slowly. Why would Democrats choose to pump money into propping up government jobs in this circumstance? Well, argues Yglesias, it would have a stimulative effect. As things stand “private sector gains are being partially offset [by] steady job losses in the public sector” (though his latest graph–as with his previous effort hyping a mild 4% decline in non-education employment–is dramatic only if you ignore the actual numbers on the left-hand axis).

Maybe this jobs “offset” is currently happening. But that still doesn’t explain why a Keynesian stimulus should focus on the public sector. After all, as Yglesias notes, the private sector remains much larger than the public sector. There are lots of areas in the private sector where employment has been declining even faster than the recent mild public sector retrenchment. Construction jobs are still down almost 6% over the period Yglesias chose (compared with 3% for all government jobs, and a bit less than 4% for state & local government jobs).  Manufacturing jobs are down 5%. Furniture manufacturing jobs are down 16%. Why not stimulate furniture manufacturing? Its decline is also a “partial offset” to gains elsewhere in the private sector.***

Nor is there a good argument for not stimulating the private sector generally, even if  it’s finally growing (all-too-slowly). There might be such an argument, as Ygelsias suggests, if there were no idle resources in the private sector for Keynesians to put to use. But there are. Plenty. Especially in the furniture industry, apparently. If you want to go Keynesian, Yglesias’ graphs do very little to indicate where you should do it–and to the extent they do, they don’t point to government jobs.

Plus there are very good arguments for not subsidizing the continued growth of government–the growth that Schaeffer’s bigger picture documents. Government employment (and attendant spending on wages, pensions, benefits, discrimination lawsuits, etc.) has proven very hard to cut. In good times, when tax money is flowing freely, it’s even harder. Recessions are the last, best hope of those who would put the bureaucracy on a diet. If not then, when?

Liberals, especially, should be almost desperate to shrink government to a size that’s sustainable in the long run. Here’s a graph–one of many documenting the public’s declining faith in government efficacy, in this case that of the federal government. It shows that in 1986 Americans estimated that the government wasted 38 cents of every dollar it raises. Now they estimate it wastes 51 cents. It’s not hard to read the inevitable death of even Clinton-style liberalism in the rising slope of that chart, and others like it. Surely at least part of the voter’s disdain for government comes from their (accurate) perception that it’s bloated and keeps bloating, with new agencies and inefficiencies simply layered in on top of the old ones. Meanwhile, the private sector has used computer technology to streamline itself, eliminating layers rather than adding.

Democrats, as the party of government, need to turn that perception around by. Obama, by dedicating his stimulus to preserving bureaucracy as we know it, has by and large missed this opportunity. (Clinton would not have missed it.)

Public sector employment is also pretty much the worst sector to stimulate if you care about government deficits, since the unnecessary jobs saved will linger after the recession ends, protected by powerful public worker unions, where they will add to government’s long term costs. If we stimulate the construction industry, the tire industry or any other private sector industry,  there’s not that same budgetary downside. If all those jobs linger after the recession ends, so much the better.

Even direct government employment in new WPA-style jobs would be better in this respect, I think, than simply paying to preserve existing government jobs. It’s easier to wind a WPA down, when recession ends, than it is to shrink the entrenched education bureaucracy.

And every dollar spent on a neo-WPA is a dollar spent on a job. This may make a big difference, stimulus-wise–as Paul Krugman has at least hinted. Why, after all, didn’t Obama’s original stimulus–which also sent money to state and local governments to subsidize their existing labor forces–work as well as expected? One reason might be that cities and states (much like consumers) effectively diverted some of that money to pay off their debts instead of spending it on teachers and sex-harrassment counselors, etc.. (Even if money is earmarked for state jobs, states can divert other money, that would have been spent on those jobs, to debt payment.)  If cities and states had no deficits, this wouldn’t happen.

 If Yglesias wants his favorite union-friendly Keynesian stimulus to work efficiently in the future, he should want those deficits eliminated as quickly as possible–which means subsidizing employment that isn’t going to cause them to stay high for decades. Which means subsidizing some type of employment other than the state and local government jobs he’s (deceptively!) campaigning to subsidize.


**:  Data in Schaeffer graph (like data in Yglesias’ graphs) is from Fred.

***: These stats are from this handy array of tables.