it's not that i'm debating the multiplier effect. that idea worked nicely when we were all in closed economies with these nice, hefty tariffs protecting us from the outside. but, that world is not with us anymore.
the construction worker might or might not buy a car, but let's look at more pressing targets of income. first, the construction worker is probably in a lot of debt. credit card debt. student loan debt. mortgage debt. before any new cars are purchased, all that debt needs to be addressed. and, then where does the construction worker go to spend what's left over?
the assumption justifying deficit-spending was that the money would go into local companies, which would invest in local manufacturing, which would create local jobs. the chain of logic is as valid as it's ever been. but, if the construction worker buys goods at walmart (likely, lowest cost) then the investment goes overseas and creates jobs in asia. it pulls the rug out from under the policy.
if we're smart reasonable people, we realize that jobs are created through increases in aggregate demand. but, when you apply that to the service sector it translates merely to urban sprawl. increased sales at walmart don't create jobs, they just make the workforce work harder.
it's in some senses an irrelevant concern in terms of half of the policy. somebody has to fix the infrastructure, and that somebody is invariably going to be the federal government. but, the reality is that free trade abolished keynesian deficit-spending as a coherent economic idea. the effects will multiply out into the global economy, sure. but, the only way that the government can create jobs is directly. deficit-spending just can't create jobs in a system of global free trade and offshored manufacturing.
this debate over whether the government ought to influence the economy is anachronistic. today, the government is unable to influence the economy. and that requires some serious adjustment on the pseudo-left.