This is a very weird post by Department of Transportation chief economist Jack Wells, entitled “Transportation Spending, an Inefficient Way to Create Short-Term Jobs.” It reads:

Whenever the economy hits a rough spot, politicians often say that we need to spend more on transportation infrastructure to create jobs.  They often cite numbers like “47,500 jobs are created for every billion dollars spent on infrastructure.”  The Federal Highway Administration has indeed done estimates of the number of jobs that are supported by spending on highway infrastructure, and the “47,500 jobs” number comes from one such study done in 1997.  But a billion dollars doesn’t buy as much as it used to, in highways as in most things, and, because that billion dollars buys less steel, concrete, and employment-hours, recent updates of those studies have cut the number of jobs supported by a billion dollars in federal highway spending to about 34,800 jobs.

Moreover, that number is based on a federal investment of $1 billion, assuming that it is matched by $250 million in state spending.  If we calculated the number of jobs supported from $1 billion in total federal and state spending, the jobs created would fall to about 27,800.  Also, it’s really more correct to say that the billion dollars “supports” 27,800 jobs, because the actual number of new jobs created depends on how much unemployment there is when the highway spending starts.  If most people already have jobs when the construction starts, people will just leave their old jobs to take a new job, and there might be very few new jobs created.  The highway construction jobs might be better jobs than people had before, but they won’t all be new jobs.  It’s also important to understand that not all of these jobs are construction jobs.  About half of the jobs are created in the construction industry and in supporting industries like steel and concrete production, but half of the jobs are in industries that produce consumer goods and services that construction workers and highway engineers buy with their increased incomes – everything from movie production to fast-food services.

Finally, it takes a long time for these jobs to be created.  Infrastructure construction requires a long series of steps to plan, design, get environmental clearance on, and construct infrastructure projects.  Only about 27 percent of the funds, on the average, are actually spent (“outlayed”) in the first year, while another 41 percent are spent in the second year.

A billion dollars spent on almost anything will create jobs.  John Maynard Keynes used to say that, if necessary, we should bury pound notes in bottles and bury them, so that people could dig them up.  It’s not very useful, but it does create jobs (digging up bottles).  The real question is, if we have a billion dollars to spend, what is the best thing to spend it on – better education?  Better health care?  Better infrastructure?  What will produce the greatest benefits, short-term and long-term, for our economy?  The real question to focus on for transportation infrastructure is what impact it will have on improving the long-run productivity of our economy, and how that compares with alternative uses of those tax dollars, rather than on the short-run impact on jobs.

There are several weird things about this. One is simply that at a time when gas tax revenues are depriving DOT of needed funds, they’d have an economist argue that a stimulus package designed to add to those funds would be unwelcome. Another is that he doesn’t show what the title says he’s going to show. He explains why transportation spending creates fewer jobs than it used to. He doesn’t explain why transportation spending is an inefficient means to create jobs relative to other ways of spending that money.

But he’s also being deliberately obtuse throughout the post. The number of new jobs created depends on the unemployment rate? Ok, why not note that unemployment is high and growing? Headline unemployment has risen to 6.1 percent, but broader measures that include discouraged and marginally attached workers are in the double digits. Surely that’s relevant information?

He recycles this popular long lead time critcism, but that assumes all the spending will be done on projects which have not yet begun. Far more likely, a stimulus would seek to get projects back on track that had been derailed by budget shortfalls due to the economic downturn. There are thousands of infrastructure investments in various stages of completion, and it’s entirely unhelpful to pretend in his post that that’s not the case.

Finally, he says we really ought to focus on the long-term impact of transportation spending. Fine. According to the Congressional Budget Office, we could justifiably spend tens of billions of dollars more each year on transportation infrastructure projects, based on the probable return to investment.

So, transportation spending is an excellent long-term investment, which might also serve as an efficient stimulus, given that unemployment is high and rising and that there are plenty of projects which could provide immediate employment were funding available. Isn’t that the story? Wouldn’t you expect a DOT economist to look at these facts and conclude that now would be an excellent time for a transportation spending stimulus?

Not in Ma Pete’s DOT.