Market solutions are superior to coercive ones because voluntary exchange offers the assurance that social interactions are mutually advantageous. However, transaction costs prevent some potentially profitable voluntary exchanges from taking place. Through the use of appropriate rules, government can provide feasible alternatives. In the standard examples of roads and national defense, the transaction costs of individual exchange are high and the free-rider problem is substantial. Thus, there is at least some potential for using tax-financed provision of these public goods as a corrective mechanism. Of course, government provision of public goods is fraught with numerous problems, and one ought not to be overly optimistic that government will get it right. However, we should not automatically rule out all government intervention.

The observer may introduce an efficiency criterion only through his own estimate of his subjects’ value scales. Hence the maximization criterion which the economist may employ is wholly in terms of his own estimate of the value scales of individuals other than himself. Presumptive efficiency is, therefore, the appropriate conception for political economy.

In political economy the observer isolates an "illness" or rather what he believes to be an "illness" through his knowledge of the system. He presents a possible change. But this change is a "cure" only if consensus is attained in its support. The measure of “wellness” for the political economist is not improvement in an independently observable characteristic but rather agreement. If no agreement can be attained, the presumed "illness" persists, and the political economist must search for still other possible solutions.

• The potential gains should greatly outweigh the costs and foreseeable free market solutions should be almost non-existent.
  
• The policy should be implemented and enforced at the lowest level of bureaucracy possible.
  
• The bureaucracy enforcing a given policy should have only the absolutely minimum amount of coercive power necessary to carry out its function and there should be "chinese walls" between separate bureaucracies. A given bureaucracy should not have access to any information or any power that does not directly pertain to its function.
  
• The burden of proof should always be on those proposing government action. Something like a 75% majority required in favor of collective action. Such a high burden would discourage repeated fishing with ballot initiatives, etc.
  
• Every bureaucracy should have a defined life span. A given policy should only be re-initiated through a full review of policy goals and all policies associated with those goals. A given policy should never be (re-)initiated without a reevaluation of all similar policies. Essentially a one-goal, one-policy, one bureaucracy rule. This should apply to all levels of gov't. Perhaps another rule should be that a higher level of gov't cannot enact a policy unless all lower levels of gov't abdicate any authority over how to achieve that goal.