In rich economies, stability matters a lot; small further increases in allocative efficiency matter less. Thus, policy should be heavily focused on e… - Adair Turner, Baron Turner of Ecchinswell

The proposition that markets drive economic efficiency is central to much of economics. Adam Smith illustrated that the “invisible hand” of the market drives efficient allocation of resources in a system of division of labor. Friedrich Hayek illustrated the central importance of the price system as an information processing mechanism more powerful than any centrally planned system could ever be. Kenneth Arrow and Gerard Debreu illustrated that complete and perfect markets deliver a Pareto-efficient equilibrium, in which no one person can be made better off without making someone else worse off. And the development of the efficient- market and rational- expectations hypotheses suggested that financial markets are in fact efficient, and that the conditions required for efficiency and for rational and stable equilibria apply even in contracts between the present and the future, which financial markets provide.

Imperfect markets are different. (...) Similarly, in Paul Krugman’s neat adaptation, “all perfect markets are perfect in the same way: all imperfect markets imperfect in their own different way.” But truly perfect markets exist only in economists’ models; in the real world, markets are imperfect. But within imperfect markets there is a range from those that work well enough for a laissez-faire approach to be broadly valid to those for which market failure or imperfection is extreme and inherent. The market for restaurants works pretty well. The best way to ensure a range of restaurants that provide us with variety, incentives for good service, and enjoyment of changing ambience and menu is to let entrepreneurship do its business—let thousands of flowers bloom, some to succeed and some to wilt.

In sum, therefore, many of the assumptions and analytical frameworks that underpin the instrumental argument for free markets and inequality are either invalid or much weaker than is commonly supposed.