Working Paper: NBER ID: w31526
Authors: Richard G. Newell; William A. Pizer; Brian C. Prest
Abstract: Government analysts have long used discount rates based on investment rates of return to approximate the effect of capital displacement. However, we show how this approach is not well grounded in economic theory and produces highly biased results, particularly in the context of decisions involving long-lived impacts such as climate change. We demonstrate how analysts can use the conceptually correct shadow price of capital (SPC) approach in a straightforward manner to account for concerns about capital displacement in federal regulatory analysis. We derive a formula for the SPC as a function of four key parameters and propose a central SPC value of 1.1, with a reasonable range of 1.1 to 1.2. We then illustrate how the SPC approach could be easily implemented in practice using the example of the 2015 Clean Power Plan Regulatory Impact Analysis, showing that estimated net benefits are far less sensitive to capital displacement concerns under the analytically correct SPC approach as compared to the inappropriate approach of using a 7 percent investment rate or return. Our work is particularly important given the ongoing efforts to revise federal guidance for benefit-cost analysis and discounting.
Keywords: shadow price of capital; capital displacement; cost-benefit analysis
JEL Codes: D61; E2; H43
Edges that are evidenced by causal inference methods are in orange, and the rest are in light blue.
| Cause | Effect |
|---|---|
| 7% investment rate of return (G11) | biased results in benefit-cost analysis (BCA) (H43) |
| 7% investment rate of return (G11) | misrepresentation of value of benefits to households (H31) |
| SPC approach (C23) | more accurate results (C52) |
| SPC approach (C23) | consumption-equivalent values (D11) |
| SPC approach (C23) | net benefits less sensitive to capital displacement concerns (F21) |
| SPC approach (C23) | positive net benefits (D61) |
| 7% investment rate of return (G11) | negative net benefits (J32) |
| current method (B50) | misapplied in federal regulatory analyses (L98) |