Socio-economic interactions pose many coordination problems to individuals. Generally social agents face problems of sharing and distributing limited resources in an optimal way. Examples range from the use of public roads and the Internet, to exchanging what we produce with what we consume. Solution which invoke the intervention of a public authority who finds the social optimum and imposes it on agents are easy to find but hard to enforce. Rather self-enforcing solutions -- where agents achieve optimal allocation of resources while pursuing their self-interests, without explicit
communication or agreement with others -- are of great practical importance. An analytic approach to these kind of problems has been recently made possible using tools of statistical mechanics of disordered systems. We review the main results and implications.