In my project I model the potential impact of regulation designed to manage systemic risk in the financial sector.
It is often the case that to allow the free market to operate unfettered is to submit to a suboptimal outcome. This may be true even in theory, consider for example Braess's paradox. There it should be clear that the construction of an additional road can only enlarge the set of possibilities available to a central planner; the optimal flow can only improve. Yet where agents are permitted to choose their own routes selfishly there may result in an increase in overall travel time!
If there were no interactions then there would be no issue. A system of banks, say, each acting to minimise its own risk, would indeed find its way to the system-wide optimum. But where interactions exist, just as in the traffic flow game, that is far less likely to be the case. Economists refer to the effects of such interactions as externalities. Thus Andy Haldane of the Bank of England has said, in criticism of currently ongoing regulatory efforts,
[A] conceptual problem with [Basel Accords] risk-weighting is that it takes no account of the collective consequences of banks' asset allocation decisions. For example, no account is taken of the externalities, positive or negative, that banks' portfolio choices may give rise to. [...] In an ideal world, these risk externalities would be taken into account in the setting of risk weights.
Continue to Mechanism Design.