The standard model in economic theory assumes that arbitrarily complex contracts can be written at zero cost, and contractual arrangements can be enforced by a third party (the judiciary) at zero cost to the contracting parties. When these conditions hold, it is impossible for agents to cheat successfully, and no moral issues are involved in executing their side of the contract. When there are aspects of performance for one of the partners to a contract that cannot be third-party enforce, principal-agent models suggest that the principal can costlessly set complete incentives for the agent that induces optimal performance, so the principal is indifferent to the moral character of the agent.
The assumptions that imply the irrelevance of moral character in market exchange are, however, widely violated, and often gravely violated. For instance, an employer may prefer an honest and trustworthy worker to a self-interested worker, because the cost of monitoring will be lower and the level of effort and care delivered by the worker will be higher if the worker behaves morally. Similarly, when two firms contract for services, the honesty and trustworthiness of each may be a prime concern of the other because it is infeasible to write complete contracts.
The proof of the importance of morality is that economies in which the moral element is absent will generally operate at a low level of efficiency. Corruption and malfeasance do not mere transfer benefits from the moral to the immoral agents, they are contagious: integrity unravels when moral individuals see that they are being scammed, and respond by in turn abandoning their moral high ground. This is known as the unraveling of cooperation.
David Rose's fine analysis of the morality of economic behavior is inspired by the economist Friedrich Hayek, who stressed that information in the economy is radically decentralized and distributed across economic agents (see his American Economic Review paper of 1945), as well as the philosopher Immanuel Kant, who stressed that moral behavior is based on adhering to moral rules, not acting benevolently on behalf of others. For Kant and Rose, and I think this is the correct view, is that one is not honest because one cares about one's exchange partner, but because being honest is the right thing to do.
Rose's support for the Kantian position is not philosophical, but psychological. In very small societies, one can care about all of one's trading partners, he argues, but in modern society, in which one has many trading partners and one does not really know more than a few personally, contract enforcement cannot be expected to depend on mutual positive feelings, he argues.
One of the attractive features of this closely argued book is that Rose argues that trust is not a moral virtue, but that trustworthiness is. He correctly notes that in a society in which individuals generally trustworthy, then even amoral agents will be trusting, and there is nothing moral about trusting others in a generally untrustworthy environment.
My only quibble with Rose's argument is that it is just too Kantian. Rose argues that honesty and trustworthiness must be unconditional, whereas I believe that no moral value is truly unconditional. This is because moral behavior is in fact governed by the rational actor model: individuals behave morally because this maximizes their utility. Individuals in general will trade off some moral values against others, and will even trade off moral behavior against selfish interests. When the cost of behaving morally is very high, individuals will abandon their morality. In fact, we do not expect unconditional compliance from exchange partners, but rather only generally dependable integrity.