The standard case against monopoly is strong. A firm with no competitors can restrict output and charge a price above what a competitive market would produce, so consumers pay more and receive less — the quantity traded is below the socially best level, which is the definition of a misallocation. Without competitive pressure the firm also has weaker incentives to control costs, innovate or improve quality, since customers have nowhere else to go. And the profits earned represent a transfer from consumers to the owners of the firm, which may worsen the distribution of income.
But monopoly is not always a failure, and there are several reasons.
Some industries are natural monopolies, where the fixed costs are so large that a single firm supplying the whole market is genuinely cheaper than several duplicating the network. Water pipes, electricity grids and railway track are the standard cases: forcing competition would mean building parallel networks, which would raise costs rather than lower them. Here monopoly is the efficient structure, and the failure would be in creating competition, not in permitting the monopoly.
Monopoly profits can also fund innovation. Research is expensive and risky, and a firm facing intense competition may be unable to finance it. Patents deliberately create temporary monopolies for exactly this reason — society accepts higher prices for a period in exchange for inventions that would not otherwise have been made. A pharmaceutical monopoly on a new drug is a designed outcome, not an accident.
A large firm may also achieve economies of scale that a small one cannot, producing at a lower average cost. If some of that saving reaches consumers, they may pay less than they would under fragmented competition despite the absence of rivalry.
And a monopoly may be contestable. A firm that is the only supplier today but could face entry tomorrow behaves quite differently from one protected by legal barriers — the threat of competition disciplines it even though no competitor exists.
My judgement is that monopoly is a market failure when it results in output restricted below the socially best level with no offsetting gain, and not otherwise. That test distinguishes the cases properly: a natural monopoly supplying a whole market efficiently is not restricting output in the relevant sense, and a patent holder is restricting output now in exchange for a product that would not have existed at all.
What matters is therefore the behaviour and the circumstances, not the market structure by itself. A monopoly that raises prices, neglects quality and blocks entry is a failure. One that exists because duplication would be wasteful, that is regulated to prevent abuse, or that holds a temporary patent as the price of innovation, is not. The policy conclusion follows: the response to monopoly should usually be regulation of conduct rather than automatic attempts to break it up, because breaking up a natural monopoly would raise costs for everyone.