Infrastructure Projects Create Jobs

There’s quite a hefty debate about whether the stimulus is creating jobs or not. A headline from a recent article went so far as to make the claim that stimulus funds have not created any significant number of jobs.

Despite the headline, some economists were quoted in the article saying that it was too soon to see any real effects from the stimulus, or that the stimulus is most likely counter-acting job losses. In other words, while there’s no job growth, the stimulus stopped job losses from happening.

At the same time, some economists were quoted as being skeptical about the stimulus.

Either way, I’m shocked that any university economist would agree with findings from such a weak analysis – or at the very least that the headline is so misleading.

I decided I would run some of my own tests, using stronger economic data and methods.

The important question is this: how does public spending on roads affect unemployment rates and jobs?

For my own analysis, I used county-level data of all Wisconsin counties between 1990-2006.  For the road spending variable, I looked specifically at all municipal capital spending on roads within a county and compared it with that county’s unemployment rate. Email me at jpkovari@uwm.edu if you have questions about the data and output.

Overall, I found a statistically significant relationship between spending on roads and unemployment rates (Pearson coefficient = -0.273, p <0.000). Take a look at the relationship in Milwaukee County.

It’s not clear as day, but the correlation is negative. That is, the more Milwaukee County municipalities spend on local roads, unemployment trends downward. For example, the graph shows how unemployment rates in 2002, 2003, and 2004 range between 6.5-7.0%, when road spending was lowest. Compare those years with 1990, when spending was at its highest and unemployment under 5%.

I also used some more advanced statistical methods. I ran a few panel regression models that better isolate the impact of road spending on unemployment and construction jobs, controlling for such variables as population, per capita income, and property values.

The regression models tell us that higher rates of road spending results in more construction jobs – which makes common sense. When government spends money to build a road, someone needs to build it. A construction firm hires employees. The more road projects, the more employment that occurs.

More importantly, when I ran the same regressions using unemployment rates as the dependent variable, a similar effect was found.

My models estimate that a $10 million increase in spending on local roads translates into a 1% drop in the unemployment rate.

Taken together, the evidence suggests that the stimulus is working. We can also expect positive effects from Milwaukee County’s expansion of capital projects, as Scott Walker champions here.  If only he thought similarly about government spending in general.

The AP analysis couldn’t be more off. And the critics of stimulus need to take a stronger look at the economic data.

Bottom line: infrastructure projects create jobs and reduce unemployment.

Guest post by: John Kovari

John Kovari is a Ph.D. student in political science at the University of Wisconsin-Milwaukee and the 2008-2009 Norman N. Gill Fellow at the Public Policy Forum.  Additionally, he has served as a legislative assistant to city of Milwaukee Alderman Michael Murphy.