To dig holes in the ground," paid for out of savings, will increase, not only employment, but the real national dividend of useful goods and services,
wrote John Maynard Keynes in The General Theory of Employment, Interest, and Money. Of course, since Keynes was an economist, there was an on-the-other-hand:
It is not reasonable, however, that a sensible community should be content to remain dependent on such fortuitous and often wasteful mitigations when once we understand the influences upon which effective demand depends.
Unfortunately, much of the economic rationale for the two-thirds of the Stimulus Bill which did not consist of tax cuts seems to have been based on Keynes' first sentence minus the part about "paid for out of savings" – Stimulus (aka The American Recovery and Reinvestment Act or ARRA) was financed by increasing the national debt. Trouble is that if you borrow money to hire people to dig holes in the ground (and presumably to fill them up again), all that you're left with is disturbed dirt and debt, not a "real national dividend of goods and services" as Keynes claims. When you spend government money – just like private money – you need to concentrate on value created, not the number of people employed. Ultimately greater economic value leads to a more robust economy and more jobs. In the short term, debt-financed jobs which don't create value damage the economy and reduce future employment.
In a laudable attempt to make the government accountable for the results of its expenditures, Stimulus law and regulation called for counting jobs "created or preserved". Many people assumed that a "job" was something at least semi-permanent such as what you get when a new factory opens nearby. In fact, Stimulus had several definitions for jobs, but this wasn't one of them. All were measures of short-term employment. You can read more about the many ways to measure jobs and my attempts to explain them to Vermonters in How Many Jobs Were Stimulated?, which I originally published on my state blog while I was Vermont's Chief Recovery Officer responsible for reporting this data to the feds and explaining it to our citizens.
Measurement is an important part of management. It's true that you get the results you measure and reward. We weren't asked to count the jobs that were lost as businesses waited to see whether they or their competitors would get "free" government money, nor could we have done that measurement if we tried. We weren't asked to count permanent jobs, nor could we have. We weren't asked to measure the impact that the dollars we spent today would have on jobs available tomorrow. We weren't asked to measure the value of service received by the public for the dollars we spent. We didn't measure the effect of paying people not to work. We were measured against other states in how many "jobs were created or retained"; and there was some debate over whether a high number of jobs per million dollars meant that were doing well or badly – after all, if you create a lot of jobs per dollar, it may be because those jobs pay less well than jobs created or retained by other states.
Stimulus didn't literally hire people to dig holes and then fill them up again, but much of the money was aimed at preserving public-sector jobs and salaries. The intellectual rationalization for this way laid out in an Obama Transition Document which is cited and summarized in a May, 2009 report from the President's Council of Economic Advisers. This table of assumptions from the report tells much of the story:
Let's forget the absurdity of treating all spending in a category as equivalent and just subject these numbers to a thought experiment. If they are literally true, we can get down to zero unemployment by just having the government hire anyone who needs a job and raising taxes by whatever amount is required. Assuming no government overhead (yeah, sure), increasing taxes by the $92,136 required to fund government spending to create one job will only raise taxes enough to cost about two-thirds of a job somewhere else. We just keep doing that and everybody lives happily ever after. Socialists do believe that works; history doesn't support their view (just ask the Chinese).
By concentrating on counting "jobs" rather than what those jobs provided or created, we got short-term jobs we could count (those funded directly by government) at the expense of other jobs which may actually have been lost because of Stimulus and at the expense of building value for taxpayers. On the other hand, we probably got better value from Stimulus than from TARP, which concentrated on saving bankers' jobs and salaries and was accountable to no one.
There is a role for government spending in the economy – obviously. There is even a role for counter-cyclical spending. More on what could be in future posts in this series which begins with Confessions of a Stimulator.