The Stimulus Dissected:

For example, as a straight-up jobs bill, $787 billion could have created 10 million jobs per year for two years, well above the now 6.7 million jobs lost. So why is it that the current bill hasn't achieved those kind of numbers? My intuition is that this stems from a fundamental weakness in orthodox "fiscal" Keynesianism versus "social" Keynesianism which is that "fiscal" Keynesianism re-routes demand through the existing structure of the economy, even if the existing structure is maladjusted.

Let's take each part by itself: we know that tax cuts are an inefficient way to boost demand, although to give the Obama administration credit, they did try very hard to make the tax cuts as demand-enhancing as possible (by reducing withholding for example). The problem I think was that the tax cuts on their own simply weren't powerful enough to counter-act the liquidity preference caused by huge and continuing job losses - people are still afraid of losing their jobs, so they squirrel money away against that possibility, thereby creating a paradox of thrift.  (Here's an area in which I think the cross-effects of a jobs program with traditional Keynesianism might get quite interesting; if people are seeing massive hirings instead of firings, if the newspapers are trumpeting "FOUR MILLION WORKERS NEEDED" instead of "500,000 more laid off," would that affect the liquidity preference of the employed?)

I think the aid to states has probably been the most effective part of the bill, which has significantly counter-acted (although not completely successfully; witness the way in which California used a "trigger" mechanism to sap the stimulus of its impact on the California state budget) the pro-cyclical effects of state budgets. Indeed, the expansion of SCHIP and other social programs in the face of increasing need, backed by Federal stimulus funding, I think is the exact image of what 50-State Keynesianism is all about.

What hasn't yet kicked in fully is the public investments - and here I think the Obama administration made a huge mistake in focusing on "shovel-ready" projects as opposed to "people-ready projects." And here I think we can learn from history.

Background:

In the First Hundred Days of the New Deal, one of the chief priorities of the progressive faction within the Roosevelt Administration was to fund massive public works projects to put people back to work and boost demand - exactly the strategy that Congressional progressives like Senator Robert Wagner had pushed through the Congress, only to be repeatedly vetoed by President Hoover. Thus, when the Roosevelt Administration went to work crafting the National Industrial Recovery Act of 1933, the centerpiece of the so-called "First New Deal," one of the elements of the bill designed to mollify progressives who were cool to the idea of business self-regulation was Title II, which established the Public Works Administration (PWA) and gave it $3.3 billion to put towards building public works.

And yet, just like the Obama stimulus, it proved more difficult to make public works an economic stimulus than had initially been thought. Now, the problem is the decision to focus on "shovel-ready" projects, but in 1933, the problem was "self-liquidating" projects. the PWA's Administrator, Harold Ickes, was a dyed-in-the wool progressive, a veteran of Teddy Roosevelt's 1912 campaign, a staunch member of the progressive faction of the New Deal who were committed to proto-Keynesian ideas, and probably the most racially egalitarian New Dealer. He was also probably the wrong man for the job, because Ickes was also very much a traditional good government progressive - as a result, Ickes made it his first priority, not to spend the money quickly, but to make sure that the taxpayer's money was spent properly, and to focus on the public works not the public workers. As a result, Ickes made two fateful decisions - the first was that the PWA would only support "self-liquidating" projects, i.e, public works that could generate enough income to pay back the cost of construction, which Ickes felt would prevent the public's money from being wasted on pork-barrel projects; the second was to spend the PWA's money through the traditional contracting process, which would allow him to closely vet all applications to make sure that the public's money would be spent honestly and that the project was of sufficient public value.

This kneecapped public works as an economic stimulus. By mandating self-liquidation, the PWA restricted itself to projects that would generate revenue streams - dams that could sell electricity to consumers, bridges and tunnels that could charge tolls, etc. - which meant that projects that couldn't generate revenue (schools, libraries, post offices, etc.) couldn't be done, and it also meant that projects went to where there was enough people to generate revenue, not to where unemployment was highest. Because of the different engineering requirements (the Hoover Dam versus your local library), self-liquidating projects used most of their money on land, materials, and machinery, rather than manpower. By mandating contracting, the PWA slowed down the process of project approval to a crawl, ensured that money would be spent on the lowest bid (a good thing from the point of view of frugality, but not from the view of trying to generate economic stimulus), and ensured that the amount of new employment generated would be relatively low, because most contractors already had an employed workforce, would generally expand only at the margins, and preferred to use a few skilled workers and more labor-saving machinery than thousands of unskilled workers (to keep productivity and profits up).

As a result, by the end of1933, the PWA had only awarded $900 million in contracts (27% of the allotment), and had spent only $110 million (3% of the allotment). As late as the end of 1934 and the beginning of 1935, the PWA's workforce had stabilized at only 650,000 jobs a a time when 8 million men were still out of work.

What ultimately saved the New Deal's public works program was Harry Hopkins' suggestion to go around the PWA. In late October of 1933, Harry Hopkins, then head of the Federal Emergency Relief Administration, went to FDR with a proposal to create 4 million jobs through "force account," or direct hiring by the government. FDR, not wanting to go into the first Depression winter of his presidency without having done something, authorized Hopkins to take $400 million of Harold Ickes' PWA money and establish a Civil Works Administration (CWA). Hopkins took the president's authority and ran with it, such that by January of 1934, only three months later, the CWA employed 4,270,000 workers.  Hopkins' CWA, by directly hiring the unemployed and initiating projects on its own instead of waiting for projects to come to the government, was able to lap the PWA - it spent through the $400 million almost instantly, and would eventually spend about $1 billion in six months. Of the 4.27 million workers, paid $48 a month ($780 in 2008 dollars), one-half of them came from people who had previously accepted poor relief, and the other half came from any unemployed worker who applied for a job, eliminating the hated and demeaning process of having to declare yourself as a pauper. In the six months that they were employed on more than 4,000 projects, these CWA workers built a half-million miles of road, built or improved 40,000 schools, employed 50,000 laid-off teachers, built 500 new airports and improved another 500. It was a dress rehearsal for the Works Progress Administration, which would become the workhorse of the New Deal.

History Repeating:

And so we find ourselves in a very similar position today - we've allocated a lot of money, but it's not getting spent fast enough or to employ enough people. Public works have changed dramatically since the New Deal's time, and one of the changes has been that the process has gotten a lot longer between the initial appropriation and breaking ground. Now you have to go through planning, environmental impact reports, public comments, permits, possibly lawsuits, then contracts, then finally somewhere down the road, you get jobs - but the focus, and thus the marginal dollar, is always on the thing being built.

The beauty of shifting from a traditional public works model to a WPA-style "public employment" model is that it flips the process on its head. With "public employment," you start with creating a certain number of jobs, and then figure out how many projects you can complete with that amount of labor power; then all that remains is to work out the non-labor financing (the WPA required states and localities to pay part of the costs of the projects, usually the materials, tools, and land), and put the workers to work.  But there are two hurdles here -first is the eternal, undying, and unremitting hostility of contracting firms and their political allies to any kind of "force account." More on that in a later installment. The other is how to make this kind of project fit through the modern permitting process.

Squaring the Circle:

In an ideal world of course, the combination of Job Insurance's permanency and Swedish-style Labor Market Policy allows one to construct a permanent "shelf" of public projects that obviate the need to go through the normal planning process of designing new projects. However, the success of the CWA and WPA does show that you can bypass the planning process with an ad-hoc approach that focuses on simple, easy-to-duplicate projects (hence, the "light construction" of roads, sidewalks, schools, libraries, post offices, municipal buildings that can be easily designed and done with mass hand labor) and still wind up with quality public works, especially if you use outside groups like the National Mayors Association or the National Association of Engineers to construct "shelves" of necessary projects.

That gets one around the planning process, but what about the longer permitting process? The solution here is to move from individual project Environment Impact Statements, permits, and hearings, to categorical permits. The idea here is to develop categories of model projects that can be approved ahead of time, allowing proposals that cleave to the category's guidelines to bypass this step and move directly into operation.

• In-Fill vs. Sprawl - projects that involve building in developed areas could be automatically accepted under EI standards, whereas projects that would involve building in previously undeveloped areas would have to go through the standard process. By this means, development could be pushed towards "in-fill" versus "sprawl," making construction and real estate more generally more sustainable.
• Rehabilitation/Renovation/Maintenance vs. New Construction - similarly, projects that involve the re-development and upkeep of existing infrastructure could also be automatically accepted, whereas projects that involve new construction would have to go through the standard process. By this process, we would also make our use of resources more efficient and sustainable, and also prevent sprawl (since new construction on outlying land is generally cheaper than new construction in developed areas).
• Green Construction vs. Non-Green Construction - finally, establishing a general criteria for "green construction" (energy efficiency, use of green materials, and so forth) projects that could go through automatically as opposed to going through the normal measures would further speed the process of approving projects, while making it economically advantageous to build sustainably.

In 1933, when the most sophisticated administrative technology was carbon paper and the rotary phone, we were able not merely put 4.27 million men to work in three months, but build an enormous amount of public works which remain of value more than seventy years later. Surely we can find a way to cut down the time from appropriation to construction to three months or less?

Conclusion:

The urgency of getting this right is heightened by another concern. As we have seen in the last twenty years, the economy is becoming less job-productive over time, as can readily be seen in the increasingly sharp recessionary layoffs and the increasingly common phenomena of "jobless recoveries." During the last recovery, between 2001-2007, we saw a large increase in productivity without a large increase in employment; employers learned how to extract further value from their existing workers, and without a tight labor market, workers lacked the leverage either to preserve their time or to demand increased wages.

This change means that Okun's law is increasingly going to break down, and its application to classical "fiscal" Keynesian policy harder and harder to achieve through remote manipulation of demand. A government may well attempt and succeed in boosting consumer demand - the cash for clunkers program was a very good micro-example of this - but if " billion in Federal grants only adds 1,500 workers at GM (after laying off 75,000), the amount of public funds spent per job created is going to get way too high.

Luckily, John Maynard Keynes was no fool, and his General Theory did not rely entirely on fiscal policy; indeed, Keynes was quite critical of efforts to use interest rates as the sole lever of policy. Instead, Keynes emphasized complementing demand policy with government intervention into investment, at points calling for the complete socialization of the investment function - and it was this element of Keynes' theory that the so-called "social" Keynesians emphasized. Thus, if there was an economic maladjustment that meant that stimulated demand wasn't going to produce enough jobs, the government could produce the necessary investment to re-adjust the makeup of the economy.

This is my final point - just as FDR was able to get around the halting progress of the PWA by turning to Harry Hopkins' CWA, Obama can make the choice to leapfrog the current "jobless recovery" by pursuing a straight-up jobs bill.