As the president prepares to announce another stimulus (possibly through an infrastructure bank), this time in the name of job creation, my colleagues Garett Jones and Daniel Rothschild have published two new — and very important — papers, which look at why the previous stimulus failed to create as many jobs as the administration promised. Jones and Rothschild have both conducted extensive field research for both studies. For the first one, they “collected over 1,300 anonymous, voluntary responses from managers and employees that allow us to better understand what happened at the organizations that received contracts funded by ARRA spending.” For the second one, their “[r]esearchers interviewed representatives of 85 different organizations, drawn from a random sample in five different metropolitan areas, and learned about their experiences applying for and receiving contract and grant funding under ARRA.” 

The papers primarily examine how people elected to use their stimulus dollars, and why. A key finding is that hiring unemployed workers with stimulus money wasn’t as common as hiring workers from other firms.

Hiring isn’t the same as net job creation. In our survey, just 42.1 percent of the workers hired at ARRA-receiving organizations after January 31, 2009, were unemployed at the time they were hired (Appendix C). More were hired directly from other organizations(47.3 percent of post-ARRA workers), while a handful came from school (6.5%) or from outside the labor force (4.1%) (Figure 2). Thus, there was an almost even split between“job creating” and “job switching.” This suggests just how hard it is for Keynesian job creation to work in a modern, expertise-based economy: even in a weak economy, organizations hired the employed about as often as the unemployed

They also provide a real life illustration of what economists call the ‘crowding-out effect.’ In this case, however, it was labor, and not capital, being crowded out. And the paper confirms that there is no such thing as a shovel-ready project:

As President Obama conceded in an October 2010 New York Times interview, “There’s no such thing as shovel-ready projects. Our interviews provide a thorough, reality-based vindication of the president’s statement.

Based on the information acquired in the interviews, the most vociferous boosters of fiscal stimulus should discount their benefits calculations by a significant factor.

Both papers are a must read. They provide a good explanation of why the ARRA stimulus didn’t deliver on its promises, and why fiscal stimulus generally doesn’t. This work is timely, as we can expect talk in the coming weeks of increasing on government spending to “create jobs.”

Based on his comments this morning, I can only presume that The New Republic’s Jon Chait hasn’t paid close attention to either paper:

I read the paper in the vain hope of finding the authors’ explanation of what they think happens when a new job is filled by moving a worker from another job. I did not see one. Nor did I see any attempt to demonstrate, or even suggest, that the newly-opened jobs of workers moving into stimulus-created jobs were going unfilled. They genuinely seem to assume that “job shifting” is simply the opposite of job creation. (They also seem to think that hiring half of new workers from the ranks of the unemployed, when some 90% of the potential workforce has jobs, is a wildly low figure.)

Here are Jones and Rothschild on page 8 of No Such Thing as Shovel Ready Projects.

Perhaps, one might say, those other companies whose workers were poached ultimately hired people from the unemployment line, and perhaps future research can find out the extent to which that actually happened. In any case, the process of hiring good workers takes time, and that eats up the short time wherein free-lunch Keynesianism can work. And if, as is the case of the construction engineering firm discussed above, only a third of a company’s new hires come from unemployment, then it is quite a lot to hope that some other firm will actually hire the unemployed.  

Moreover, job switching is not costless. When companies lose workers to stimulusfunded firms, they lose valuable skills and experience, what economists call―organizational capital So when a mid-level manager who understands the company’s database program switches jobs, or when an engineer with valuable contacts moves to the ARRA-funded engineering firm, the old firm is left weaker. The unseen effects of ARRA need to be counted, and the employees lost by other firms are among those costs. A reality-based stimulus accounting would certainly calculate at least two numbers: ―jobs create and ―jobs shifted For the engineering firm from the previous section that hired 20 new people, there would have been 6 ―jobs created and 14 ―jobs shifted. If the  raided firms ultimately hired unemployed workers—a speculation at this point—then second-round effects of stimulus might raise the ―jobs create figure.  

Our interviews did not provide enough information to create a full, firm-by-firm account of job creation versus job shifting. The organizations we interviewed often didn’t reveal or didn’t know if their new hires were unemployed beforehand; but in some cases, they pointed out that they either hired workers from the private sector or brought retirees back into the labor force. More often, firms just told us they hadn’t created that many jobs—they just used their own workers more and just hired some temps for a few days or weeks. 

And here they explain this in the other paper (page 8):

This is similar to the amount of job shifting that goes on in relatively normal economic times. Eva Nagypal (2008, p.1) notes that “employer-to-employer transitions…ma[de] up 49 percent of all—a separations from employers” in the decade prior to her study. Robert Hall (2005) finds a similar number, roughly 40 percent. Since on average separations equal hires (the minor factor of 9 net job growth aside), there is little difference between the recent U.S. average and our sample average. In other words, we find little evidence that stimulus spending was particularly effective at moving the unemployed into work. During the worst recession in generations, the ARRA-receiving organizations in our sample hired away employed workers at roughly the same rate as in normal economic times.

As Jones explains, “most employers said hiring good people was as hard or harder than pre-crisis. So worker experience and employer opinion tell the same story: stimulus was targeted at sectors where good help was hard to find.” 

Update: Here is a good post by Marginal Revolution’s Tyler Cowen about the papers. Cowen notes that:

One major problem with ARRA was not the crowding out of financial capital but rather the crowding out of labor.  In the first paper there is also a discussion of how the stimulus job numbers were generated, how unreliable they are, and how stimulus recipients sometimes had an incentive to claim job creation where none was present.  Many of the created jobs involved hiring people back from retirement.  You can tell a story about how hiring the already employed opened up other jobs for the unemployed, but it’s just that — a story.  I don’t think it is what happened in most cases, rather firms ended up getting by with fewer workers.

There’s also evidence of government funds chasing after the same set of skilled and already busy firms.  For at least a third of the surveyed firms receiving stimulus funds, their experience failed to fit important aspects of the Keynesian model.