On Balance: Efficiency without Apology: Consideration of the Marginal Excess Tax Burden and Distributional Impacts in Benefit–Cost Analysis
An important and difficult issue in benefit-cost analysis is how to deal with the distributional impacts of policies. An approach to this issue is described in a recent article published in the fall 2020 issue of the Journal of Benefit-Cost Analysis by Anthony Boardman, Aidan Vining, David Weimer, and me. This blog summarizes our analysis.
There are several arguments for considering the distributional impacts of policies in benefit-cost analysis. The declining marginal utility of income or wealth is probably the most frequently invoked. Most economic models addressing valuation assume that higher-income persons receive less welfare from a given increase in consumption (or income) than do lower-income persons. A second argument is that many people have a preference for a more equal income distribution, perhaps because that would be fairer, a form of inequality aversion. It might also be based on a willingness to pay for minimizing the indignity of those at the bottom of the income distribution. Another possible rationale is that some far-sighted wealthier individuals have preferences for redistribution to reduce civil unrest.
Many suggestions have been made for directly incorporating at least some of these concerns into benefit-cost analysis. An approach that addresses diminishing marginal utility is to use distributional weights based on proxies for individuals’ utilities. Another approach, one that addresses individuals’ preferences for a more equal income distribution, is to estimate the willingness-to-pay for a more equal income or wealth distribution, probably through contingent valuation, and then add that value to the other benefits and costs.
Neither approach addresses all the arguments for incorporating distributional considerations into benefit-cost analysis and that they are challenging to implement—there are no generally accepted set of distributional weights and the distributional impacts of different policies vary, thereby requiring policy-specific estimates of willingness-to-pay. A further critique of these approaches can be found in the previously mentioned article.
In our view, benefit-cost analysis best contributes to good public policy by systematically assessing the allocative efficiency of policies and programs—that is, whether there is a potential Pareto improvement. However, the approaches outlined above attempt to take account of distributional and efficiency impacts within a single metric. As a result, determining whether there is a potential Pareto improvement will be lost and tradeoffs between efficiency and distribution will be blurred. In our view, allocative efficiency should be determined independently of other social values. After that, tradeoffs between efficiency and other social values can be transparently assessed. Thus, any distributional or equity weighting scheme should be in addition to, rather than directly incorporated into a core benefit-cost analysis.
An approach that facilitates the examination of trade-offs between policy goals is multi-goal analysis, which can take factors beyond efficiency into account—for instance, distribution, budgetary impact, and political feasibility. Such impacts, although very important to decision-makers in the real world, are often ignored by benefit-cost analysts. The multi-goal framework has several virtues:
- It does not require using controversial weights
- It is “doable” by analysts in government, consulting companies, or academia with basic training in benefit-cost analysis
- It is not prohibitively expensive
- Decision-makers can understand the information and make their own judgments about how to trade-off different goals.
Multi-goal analysis can help make the trade-offs between goals apparent. Sometimes there is no trade-off. For example, an examination of 26 BCAs of welfare-to-work programs, programs in which distributional impacts are of obvious importance, found that higher income taxpayers who effectively funded the programs and lower income participants of the programs were both better off in nine and worse off in seven. Tradeoffs between efficiency and desirable redistribution occurred in only 10 of the 26 programs (Boardman et al., 2018, Chapter 19).
When trade-offs do exist with distributional issues, they can be treated in the following manner within the multi-goal framework. First, unweighted net benefits should always be computed. That is, analysts should conduct core benefit-cost analyses that computes the net social benefits to society. This indicates whether, given the existing distribution of wealth, the policy being analyzed would potentially permit a Pareto improvement. Then, if distributional considerations appear relevant and important, and the necessary data exist, analysts should also provide benefit and cost estimates for different income groups. Finally, in order to provide further guidance to policymakers, distributional weighting might be considered.
Because distributional weights are difficult to derive in practice and there is little consensus concerning what they should be, we suggest two simple alternative approaches. The first is based on work by Edward Gramlich (1990) who proposed a weighting scheme based on his estimate that it costs around $1.50 to $2.00 to redistribute a dollar through a representative income transfer program. These estimates are over three decades old, and Gramlich considered them preliminary at the time. Thus, they need updating.
Nonetheless, if the $2 value is provisionally accepted as a benchmark, then this would imply that if the net benefits to low income persons from a program under consideration—say, a training program for low-wage workers—are a third of the net costs to the higher income persons who fund the program, then a simple welfare program can then transfer income to the poor less expensively. Moreover, assuming that participants and those paying for the program are the only persons affected by the program, its net social benefits are negative. If instead the net benefits to low income persons were three-quarters of the net costs to higher income persons, the training program is superior to a welfare program as an instrument of redistribution, even though the program’s net social benefits are still negative. Thus, existing welfare programs provide a standard to which programs under consideration can be compared.
The second approach, one that works best when there are just two groups, advantaged and disadvantaged, is to compute internal or breakeven distributional weights when a trade-off between allocative efficiency and distributional impacts occurs—that is, when the social net present value (NPV) and the NPV for the disadvantaged group are of opposite sign. Under this approach, the weight for the advantaged group is first is set to one and then the weight for the disadvantaged group is computed by dividing the estimated NPV for the former group by the NPV for the latter group. Thus, if the net benefits to disadvantage persons from a program under consideration are a third of the net costs to advantaged persons, the breakeven weight for the former would be three. Given this weight, the program being analyzed would just break even. If net benefits to disadvantaged persons were instead three-quarters of the net costs to advantaged persons, the weight for the former would be 1.33, again the weight at which the program would just break even.
Policymakers can use these breakeven weights to formulate a judgment as to whether the benefits received by disadvantaged persons should be given a higher or lower weight than that implied by the internal weight. To illustrate, consider the 10 welfare-to-work programs mentioned above for which a trade-off between efficiency and distribution occurred. Three of these programs had a breakeven weight greater than two and a positive unweighted social NPV, but a negative NPV for participants. Gramlich’s benchmark weight of two implies that these programs should be adopted even though the income distribution would be less equal. In contrast, three programs had a breakeven weight in excess of two and a negative unweighted social NPV, but a positive NPV for participants. In these cases, the Gramlich benchmark implies that these programs cannot be defended even though they improve the income distribution. A policymaker may, of course, choose to use weights other than Gramlich’s.
In summary, this blog argues that core benefit-cost analysis should focus on determining whether a policy or program increases allocative efficiency. Tradeoffs between efficiency and distributional considerations should be considered explicitly by first examining each separately within a multi-goal framework. Only then should distributional weighting schemes be implemented.
References
Anthony Boardman, David Greenberg, Aidan Vining, and David Weimer, 2020. Journal of Benefit-Cost Analysis. 11(3), 457-478.
Boardman, Anthony E., David H. Greenberg, Aidan R. Vining, and David L. Weimer. 2018. Cost-Benefit Analysis: Concepts and Practice. 5th ed. New York, NY: Cambridge University Press.
Gramlich, Edward, 1990. A Guide to Benefit-Cost Analysis, 2nd ed. Englewood Cliffs, NJ: Prentice Hall.