What is crucial, however, is reciprocity. It is only this requirement that guarantees the harmony of interests. Correspondingly, Smith does indeed see cases in which this mechanism doesn't work. In the Theory of Moral Sentiments, he warns us not to give in to vanity and self-conceit, for example, and in the Wealth of Nations he describes the impairing effect of privilege. Undeniably, business people will always be tempted by monopolies or cartels. This is held in check by their inherent instability. Monopolies or cartels can be stable only when they are supported by a privilege granted by the authorities. A similar point concerns selfishness and greed. Selfishness is self-interest out of bounds. Usually, in market exchange, selfishness and greed are checked by the simple necessity of finding a transaction partner. Often, however, political measures are taken that boil down to making people more greedy than is natural-for example, through "quantitative easing" in monetary policy, through a generous social policy without any regard to what is affordable or not, through a lack of tight controls. When that is the case, the natural check on greed breaks down. We then observe excessive behaviour and its devastating social results.

Regrettably, Smith was unable to conclude the third large project that he had been working on throughout his life: a theory of government. In his "obvious and simple system of natural liberty", this is the only gap. It is not a small one. The answer to this question is fundamental for the current debate - in a situation where our Western civilisation seems to have come to a crossroads as more interventionist power is reclaimed for government, in which nationalisation and re-regulation are going unquestioned, and in which politics is generally gaining a new self-confidence. The question has to do with the natural harmonisation of interests. Is there any way in which the interests of government are spontaneously aligned with the interests of the governed? Is political action possible without destroying the checks and balances of the spontaneous order? Smith doesn't give us an answer. Maybe he had found that there was no such way.