Some of us like to say that the state can spend a large sum of money - say $ 810 million on a train - and "create" jobs. As statements go, this one is fairly meaningless. The only thing we can say for certain is that it has employed - directly or indirectly - however many people that amount of money has hired. Want 8100 jobs paying $ 100,000 in 2011? Give me that money. I'll find something for those folks to do - or not. To say that I created jobs is, in one sense of the word, a truism but a meaningless one.
In another sense of the word, it is almost certainly a false statement. It is false if you mean by "create" a net addition of 8100 jobs. That $ 810 million must come from somewhere. Had it remained where it was, it would save for the rare possibility of something like a Keynesian "liquidity trap") have been put to some use. That use also would have resulted in people being hired. I have not, therefore, added jobs to the economy in the numbers that I have claimed for I have also caused other jobs not to be created.
This is why statements about how many people would be hired (jobs "created") by some public works project are, if not entirely meaningless, not particularly helpful. The relevant question is whether the project creates additional value, i.e., gives people something they desire or increases the productivity of the economy, in an amount greater than some alternative use of the funds expended upon it.
Sharp readers will note that this statement implies that there are circumstances in which government might "create" jobs in the sense of causing there to be more jobs than there otherwise might be. This is true. Public education can create jobs although this does not mean that every dime spent on public education does so or that any additional spending on public education is justified. Highways may create jobs but, again, this doesn't mean that all highways are justified. Even consumer rail might create jobs but, given its cost and the demographics in most parts of the United States, that is unlikely.
The real test of any expenditute of public funds is whether it constitutes a better use of dollars than alternative uses - including private ones. In addition, that question must take into account the fact that state expenditures involve coercion. One does not have to be a Randian or regard private property rights as absolute to believe that regard for the intrinsic value of human beings and their autonomy creates a certain presumption against taking the fruits of their labor.
Of course, a presumption can be overcome but it is also true that, in the great run of cases, market mechanisms are ging be more efficient in directing funds to the most productive or desired use than centralized bureaucracies.
Not always, but mostly.