"Making good policy requires a clear understanding of the constraints under which such policy operates. While we are familiar with maximization of an objective subject to a budget constraint, there are other constraints with which real-life policy must contend. Most importantly, governments differ in their capability to implement policies and these differentials should be factored into choices between them in the first place. The same market failure that motivates an intervention could well have different optimal responses depending on what kinds of policies a government find more or less challenging.
In fact, whether or not to intervene at all should be part of the calculation. When considering a policy intervention, the welfare loss of a market failure, the degree to which an intervention can help poorer people (if that is considered as something separate from market failure) and the ability of government to carry out the intervention (including the loss of welfare of collecting the taxes to finance it) should all be quantified as well as possible and weighed accordingly. This takes some degree of honesty on the part of government to admit the possibility that things will not necessarily go as planned. Well written laws are not worth much if they can’t be enforced.
Social sector interventions related to health, education and social protection are good examples of where this balancing or weighing of alternatives are likely to change perspectives on appropriate policies. On the one hand, there are a wide range of market failures they address, from true public goods to subtle, hard-to-identify market failures and from interventions such as subsidies that are more or less beneficial to the poor. On the other hand, there are a wide range of policies that differ substantially in the degree of complexity required for their implementation. Unfortunately, many social sector policies are particularly difficult due to two common characteristics: transaction intensiveness and discretion.
“Transaction intensiveness” refers to the fact that many policies require many person-to-person interactions. School teachers in the tens of thousands have to deal with millions of children on a day to day basis. Further, they must employ “discretion” if they are to do a good job: they must tailor their activities to the needs and abilities of their students. Similarly in health care (to be distinguished from public health): each patient (many transactions) obviously needs a different treatment (discretion). Many antipoverty programs need to screen beneficiaries to determine eligibility – again, many transactions with a lot of discretion and with requisite updating as circumstances change. Both of these characteristics lead to opportunities for bribery since the more places a decision needs to be made, such as providing a service at all, the more that the services can be disrupted without the bribe. Less extreme, simply the monitoring and supervision of providers to do a good job requires more resources when there are millions of transactions requiring independent judgment.
One thing a market is good for is to have the individual be able to hold a provider accountable for adequate service. While pedagogical theories may be beyond the ability of parents to judge or while clinical knowledge is beyond patients, certainly absenteeism is less of an issue (if a doctor does not show up for work s/he does not get paid; if a teacher doesn’t, it is easy to observe). For this type of service, then, an administrative system may not work as well as a market.
In the area of health, these principles can be applied with some force. When market failures, the degree to which alternative policies affect poor people and the relative difficulty of implementing different policies related to health are all taken together, much of the conventional wisdom concerning health policy is shown to be questionable. In particular, the commonly advocated policy of providing free primary curative care does not fare as well as ensuring basic preventive public health interventions such as sanitation and water supply on the one hand, and finding a way to insure against expensive, catastrophic, hospital-based care on the other.