Daniel J. Acland

Estimates of the parameters of the quasi-hyperbolic discounting model have been used for quantitative purposes in a range of contexts. I show these estimates are sensitive to time horizon (largest value of t in the data) and timescale (months, years, etc.), and argue that when the quantitative value of the parameters matters, estimates should be used only with great caution, if at all. Using online survey data, I show that β is decreasing in both time horizon and timescale, and vice versa for ẟ. The time-horizon effect is clearly at least partly the result of estimating a discontinuous discounting model using data that are generated by a continuous discounting process. The time-scale effect appears to be caused by a genuine difference across scales in how individuals discount. I discuss implications for positive and normative analysis as well as possible approaches to proceeding with quantitative analysis in the face of these challenges.

A commonly expressed concern about distributional weighting in benefit‐cost analysis is that the informational burden is too high and the practical challenges insurmountable. In this paper, we address this concern by conducting distributional weighting on a number of real‐world examples, covering a range of different types of policy impacts. We uncover and explore a number of methodological issues that arise in the process of distributional weighting and provide a simplified set of steps that we believe can be implemented by practitioners with a wide range of expertise. We conduct sensitivity analysis and Monte Carlo simulation to test the robustness of our estimates of weighted net benefits to the various assumptions we make, and find that, in general, distributional weighting is no more vulnerable to modeling assumptions and parameter selection than unweighted benefit‐cost analysis itself. We conclude that the concern about the practicability of distributional weighting is, at least in a range of important cases, unfounded.

Distributional weighting to address concerns about diminishing marginal utility of income in benefit-cost analysis has been the topic of increased interest in recent years. Concern has been expressed about the practicability of distributional weighting, given limitations on data and on the analytical capacities of agencies. This paper contributes to a small but growing literature that attempts to provide guidance and real-world examples of distributional weighting. We develop a methodology for calculating what we call “population weights,” which, once computed for a given population by an analyst, can be used by other analysts to implement distributional weighting on similar populations, without those analysts needing information on income distribution or the cost or benefit experienced by households at different income levels within those populations. These population weights can be calculated without knowing the costs or benefits received by households at different income levels, using proxies for cost or benefit that may be observable or about which, in the absence of data, assumptions can be made in some cases. We implement the methodology on an example regulation and present results that we believe provide useful information to decision makers, even in the absence of estimates of unweighted costs and benefits.

There are increasing calls for concrete suggestions on how to account for distributional impacts in policy analysis. Within the context of benefit-cost analysis, per se, one possibility is to apply “distributional weights,” to inflate costs and benefits experienced by poor or disadvantaged groups. We distinguish between “utility-weights,” intended to correct for the bias in willingness to pay caused by diminishing marginal utility of income, and “equity-weights,” intended to account for the possibility that decision makers might have disproportional concern about the welfare of the poor or other disadvantaged groups. We argue that utility-weights are appropriate and necessary to maintain the legitimacy of BCA as a measure of aggregate welfare, but that equity-weights are inappropriate because they involve moral judgments that should remain in the domain of democratically accountable decision makers, and because they conflate information about both the welfare and equity impacts of policies, making it impossible for decisionmakerstoapplytheir ownmoralvaluestotheassessmentoftradeoffs betweenwelfare andequity. Weoffer concrete suggestions regarding the application of utility-weights and the calculation of a set of metrics to provide intuitively comprehensible and useful information about, and allow decision makers to quantitatively assess the tradeoffs between, welfare and equity caused by specific policies.