HumBio 4B Spring 2013

Environmental and Health Policy Analysis: Some Key Evaluation Criteria

This note offers a look at several key criteria for evaluating public policies. The first part illustrates and contrasts the criteria of cost-effectiveness and efficiency. The second part discusses the avoidance-of-risk criterion. The note concludes with a few comments about the relevance of these criteria to your group projects.

A. Cost-Effectiveness and Efficiency To start out, some terminology. We’ll be referring to policy instruments and targets. Instruments are the specific measures employed to achieve a policy goal. Examples of environmental policy instruments are taxes on pollution, mandated technologies (e.g., catalytic converters in automobiles), limits on pollution concentrations (e.g., automobile “tailpipe” emissions limits), subsidies to clean technologies, and tradable emissions permit (or “cap and trade”) systems. Health policy instruments could include mandated medical visits (e.g., mammograms for women of certain ages under certain conditions), mandated water treatments (e.g., fluoridation), or subsidies for certain health services. Targets are the policy goal. In the environmental context, a target might be a given reduction in the concentration of a certain pollutant. In the health context, it might be saving a given number of lives, a reducing by a given amount the number of diabetes cases.

1. Cost-Effectiveness We say that instrument A achieves a given target more cost-effectively than instrument B if it achieves that target at lower cost. In the diagram below, instrument A is more cost-effective. It reaches the target pollution reduction R1 at lower cost than instrument B (compare cost at point x with cost at point w). Instrument A might be, for example, a tax on pollution emissions, while instrument B might be the requirement that certain firms install pollution-reduction equipment of a particular kind. We can compare cost-effectiveness at other levels of pollution reduction, such as R2. In our example, Policy A involves the instrument of a pollution tax. To achieve the larger reduction R2, the pollution tax rate would presumably be higher than the tax that achieves the reduction R1. Policy B involved the instrument of a mandate of a particular form of pollution-reduction equipment. To achieve the reduction R2, this might involve a broader mandate – that is, more firms would be required to install this equipment. Note that in the diagram Policy A is more

which instrument or combination of instruments can achieve the goal at the lowest cost? You might note that in addition to costeffectiveness.g. when they compare the costs of achieving a given level of health care. one policy could be more cost-effective for one target. B Cost instr. One of its key considerations is cost-effectiveness: among the various policy options.. But it didn’t indicate the specifics about how to achieve this goal. administrative ease. (This doesn’t always have to be the produces larger net 2 . Policy analysts also apply cost-effectiveness criteria in the health policy context. A policy involves both an instrument and a level of pollution reduction (or number of lives saved). The Act committed the state to reducing greenhouse gas emissions to 1990 levels by the year 2020. and simplicity in deciding how to achieve the goal.cost-effective both when the target is R2 as well as when the target is R1. in 2006 California enacted the Global Warming Solutions Act. the policy of achieving R1 with instrument A is more efficient than the policy of R1 with instrument B -. 2. If the cost-curves crossed. and less cost-effective for another. cost-effectiveness is a key consideration. It went into effect in January 2012. Efficiency Efficiency is a measure of the aggregate net benefits (total benefits minus total costs) from a policy. the Air Resources Board also considers fairness. A w z x R1 R2 Reduction (e. in lost lives or pollution) In many contexts. The California Air Resources Board is charged with determining how best to reach the goal.) Cost y Cost instr. For example. In the diagram on the next page.

in lost lives or pollution) 3 . or through a system of tradable emissions permits. A w z x R1 R2 Reduction (e. As another example: many researchers in the Los Angeles region continue to seek the least-cost way of meeting the air quality standards for criteria pollutants (that is. policy analysts settle for cost-effectiveness comparisons. or through taxes on emissions. when benefits are exceptionally hard to measure. For example. In the diagram. In contrast. The “efficiency-maximizing” policy is the one that maximizes net benefits. Finding the least-cost way of meeting the given target is a cost-effectiveness effort.g. Benefit y Benefit Cost instr. No other combination of instrument and reduction-level yields a higher distance between the benefit and cost curves. many public health analysts focus on the relative cost of saving given numbers of lives (or quality-adjusted life-years) under different policies. B Cost instr. They might contemplate meeting the standards through emissions quotas to given stationary sources. keeping concentrations below the maximum level)..benefits. since this avoids the need to measure the benefits. The policy of achieving R1 with instrument A is more efficient than the policy of achieving R2 with the same instrument – the net benefits (total benefits minus total costs) are larger in the former case. Note that efficiency requires considering both costs and benefits. costeffectiveness only requires comparing costs (of achieving a given target). it appears that the policy that uses instrument A and arrives at reduction R1 is the efficiency-maximizing policy. This is a cost-effectiveness comparison. Sometimes. Cost. We can also compare the efficiency of two policies involving the same instrument.

g. and net benefits (benefits minus costs) would be negative. the net benefits) of two policies. Policy A might be the distribution of an experimental new drug.” Let’s define risk as the amount of uncertainty surrounding the impact of a given policy. In the environmental context. won’t offer such large environmental benefits. Policy A involves more risk than Policy B because its probability distribution is wider. In contrast. if successful. Efficiency Another evaluation criterion is “reduction in risk. Probability Policy B Policy A 0 µB µA 4 . Policy B could be a safer. but it could turn out to be ineffective or even lethal. equity and political feasibility are other significant evaluation criteria. in which case the costs of the investment would not be accompanied by any environmental benefits. but also a smaller likelihood of serious adverse health effects. once adopted. As we discuss in class. a fuel-cell car). The risk is greater. Policy B could represent the distribution of a different drug with less “upside” potential.. would have large environmental benefits. The diagram on the next page shows the probability distributions of the outcomes (in this case. Risk vs.Efficiency or cost-effectiveness are not the only bases for ranking different policy options. the larger is the variance in the outcomes of the policy. B. In the health policy context. In contrast. even though the new technology. this drug has potentially very large life-saving benefits. more conservative investment in a technology that almost certainly will work (a slight improvement to the hybrid car). This investment. Policy A could be an investment in a possible new technology for a low-polluting vehicle (e. but this investment might not pan out at all (there might be no technological breakthrough).

Very highly risk-averse people will prefer Policy B to Policy A. The fact that this policy is more risky is less important to them. You need not apply all of the criteria – which ones seem most pertinent will depend on the project involved. In the example given here.Net Benefits From the graph. A given policy’s net benefit could lie anywhere from the left-hand edge to the right-hand edge of its probability distribution. and prefer Policy B despite the fact that there’s a greater chance that the net benefits will be lower under Policy B than under Policy A (the expected net benefit is lower). If we took a weighted average of all these possible net benefits. one might prefer Policy A or Policy B. there was a trade-off between expected net benefit and avoidance of risk. Relevance to Your Group Projects In your group projects. In cases like this. A person with no risk aversion is called a “risk-neutral” individual. 3. They do not like Policy A’s higher risk. that even though Policy X is slightly less efficient than Policy Y. this weighted average is the policy’s mean net benefit (or expected value of the net benefit). It often happens that some policies will emerge as better according to one criterion and worse according to another. the expected value of net benefits – the mean (or expected value) of the probability distribution – is higher under Policy A. The mean net benefit is the same thing as the expected value of the net benefit. Depending on one’s risk aversion. Risk-neutral individuals consider only the expected value of net benefits in comparing policies – the variance or level of risk doesn’t matter. or is much more politically 5 . Policy B entails less risk. People with very little risk aversion will prefer Policy A. On the other hand. it’s easier to support this policy over other policies. it’s a subjective matter to decide what policy is best overall. In some cases. for example. it is preferable overall to Policy Y because it involves much less risk (or is much more equitable. you will likely want to apply one or more of the criteria we’ve discussed. (Subjective judgment also is involved in deciding how much weight to give to the equity criterion as compared with the efficiency criterion. The mean net benefit under each policy is represented by µΑ and µΒ in the diagram. using the probabilities (or the height of the distribution corresponding to each possible net benefit) as the weights. Which policy is better? That depends on how much weight one puts on the avoidance-ofrisk criterion as compared with the efficiency criterion (here applied in expected value terms). In this case.” it offers higher net benefits has more importance. You could decide. Thus a risk-neutral person will prefer Policy A. there is no trade-off: one policy might involve lower risk and offer a higher expected net benefit.) Risk-aversion is a psychological characteristic that refers to the extent to which people will desire to avoid risk. The fact that. This is a subjective judgment. “on average. Judgments of this kind are quite acceptable.

the particular policy selected is less important than the analysis that shows how well that policy does compared to alternatives according to relevant evaluation criteria. Although we’d like you to end up endorsing a particular policy option (and one policy option is “no policy change.realistic).” that is. sticking with the status quo). What’s most important is that you make clear in your report how the policy options stack up along the relevant dimensions. 6 .