A need for a program is identified when a market failure can be pinpointed.
Four major categories of government intervention are defined: public production, private production with taxes or subsidies, private production with regulation, and public-private partnerships.
One of these categories is best suited to address the program aim on either efficiency or distributional grounds.
Design features such as eligibility requirements must be set to maximize receipt among deserving individuals and minimize receipt among non-deserving individuals.
The public program may crowd out parallel private sector efforts to address the identified problem.
Both the short-term and long-term effects of government intervention on private sector activity must be considered.
Efficiency losses are associated only with substitution effects.
Any level of welfare achieved through a price subsidy could have been achieved less expensively through an income grant.
The distinction between intended and actual incidence, where the benefits can be shifted from the intended recipients to another group entirely, is drawn.
Programs’ distributional consequences on distinct groups are considered: poor versus rich, young versus old, cities versus suburbs, or urban versus rural.
The details of a program can offer different tradeoffs between efficiency and equity.
Any noneconomic goals of a program must be articulated so that the program can be evaluated in light of such goals as well as its economic consequences.
Political realities may affect the range of feasible programs.