One of the foundational questions libertarians need to have a response to is the ever-popular "who will build the roads?" The question has a sort of superficial plausibility to it; most of the roads we see in our daily lives, after all, were built by, and are allegedly maintained by, some government or other. There is, however, no substance to the idea; the best and most concise explanation of how absurd this question is was given by Tom Woods, who said:
"Who will build the roads?" is the question that belongs at the top of every libertarian drinking game. If we didn’t have forced labor, the argument runs, there would be no roads. There’d be a Sears store over there, and your house over here, and everyone involved would just be standing there scratching their heads.
Clearly roads are a socially desirable good, and, given that it’s rather a challenge for people to get to work or go shopping — which is to say: produce or consume — without them, there’s quite an obvious and powerful incentive to build them. It should go without saying that, even in the absence of coercion, roads would get built. As to the specific question of who would build them, here it’s important for libertarians to be cautious; there’s a natural tendency to push the argument too far and attempt to outline some scheme by which we imagine the roads could be built, but, in reality, the most a libertarian can say in response to "who would build the roads if not the government" is "somebody else."
There is an infinite multiplicity of ways society could be organized, and no single "planner" can possibly identify the best or most likely one. It is therefore an act of extreme hubris — a fatal conceit, to borrow a phrase — for any libertarian to jump from "roads could be built without government" to "and here’s how it would be done." None of us can know the optimal number of roads to build, the optimal way to build them, the optimal times or costs — nothing. All we can know is whether or not we’re willing to pay price X in return for road Y at this exact moment. In fact, if by "roads" we mean only actual, literal roads and not just "some method of facilitating travel," we can’t even be sure anybody would bother to build roads; the possibility exists that, if roads had to compete with other travel methods rather than being enforced by overseers, they would fall out of favor. Some other method of travel may outcompete them, just as the subsidized roads themselves decimated the rail freight industry in favor of long-haul trucks. So while we must admit the possibility that, in a libertarian society, nobody would build the roads, that would be the case if and only if something better came along.
We should differentiate here between two different versions of the road question. In one case, we’re wondering about who would continue to extend and maintain the existing road system if the government suddenly packed up and went home, while, in the other, we’re wondering how roads could come to be in the "state of nature" if not by coercion. In either case, the overriding concern appears to be that, without the government in charge, there would be no way to know how many roads to build and where to build them, but the situation is in fact almost precisely the reverse; the government, due to its nature as a coercive monopoly, is that institution least capable of judging where and when roads should be built. As Ludwig von Mises demonstrated quite powerfully in his book Socialism, it is impossible to act rationally in the absence of market prices. On the market, roads would be built, by and large, in places where the traffic they get justifies the cost; mistakes will be made, of course, but the developers who make those mistakes will take losses, the losses will force them out of the road-making business, and, in the long run, only those developers who do the best job of anticipating where and how many roads are needed will succeed — the same as in any other industry. In the coercive world of government monopolies, however, profit-and-loss incentives function exactly backwards, thus virtually guaranteeing that roads will be badly placed, badly maintained, and far too expensive.
Why is this? First of all, consider that, since the government doesn’t need to turn an honest profit to stay in operation (it just needs to step up the violence), the government doesn’t really care if its enterprises are profitable. Its decisions will be driven by political, rather than economic, considerations, and thus will tend to suit the whims of bureaucrats; as such, government road construction will result in absurdities like bridges to nowhere, constructed at tremendous taxpayer expense for the personal benefit of politicians. Making matters even worse is the fact that the government’s bureau of road construction, like all government agencies, will find that its budget is increased in proportion to its inability to perform its duties. If the roads, by some happy miracle, are all built well, on time, under budget, and in a sensible fashion, there will be no need seen to give more money to the bureau. Only if the roads are an absolute disaster — say they kill tens of thousands of people every year, for example — can the road bureau argue for more money. Given that we don’t assume that government road builders are some special, superior breed of man, and that they’re likely to be more or less the same actual individuals who would be hired to build roads by road entrepreneurs, it should be quite clear that the road consumers will be better served by an incentive structure that encourages them to make the best, rather than the worst, possible roads.
So we see that there are several approaches a libertarian can take when faced with the dread shibboleth of government roads. We could say that probably the same people would build the roads, but they’d do a better job. We could say that the roads would be built by people with a history of good judgement as regards where and when roads should be built. We could say that roads may even be obsolete already if they weren’t being propped up by massive government subsidies. The thread that ties the whole thing together, of course, is: we can’t really know who would be building the roads, but we can know with certainty that, regardless of who’s building them, they’d be better, safer, and cheaper. That’s probably enough, yes?