Commissioned report

Declining Discount Rates in Cost Benefit Analysis

Formats and related files

Executive Summary#

This report discusses the role of declining discount rates (DDR) in cost-benefit studies. In this report we use the term DDR rather than the expression “hyperbolic discounting” for consistency with the literature and to emphasise that we are addressing the question whether the social discount rate (SDR) should broadly decline over time, rather than the merits of any specific equation or discounting methodology.

In this Executive Summary we provide an outline of the Report and our main conclusions.

Report Outline#

Section 1 outlines the two main forms of DDR. In one alternative, the discount rate falls continuously over time according to an equation (this is known as hyperbolic discounting). In the other alternative, the period-on-period discount rate profile falls in discrete steps periodically over time.

Section 2 describes the two main approaches to discounting: the social time preference rate (STPR) and the social opportunity cost of capital (SOC), and how these concepts are formalised to provide a single social discount rate. It is noted that there are two forms of the SOC approach: the weighted cost of funds and the return-on-investment model. Reference is also made to the social cost of capital approach to discounting. When, as is usual practice in CBA, a single discount rate is adopted over the whole period of the project or policy, this is described as exponential discounting.

It is a premise of the report that DDR should be consistent with the underlying principles of the discounting method being used. Section 3 discusses the various conditions or views of the future under which the concept of a DDR may be integrated, and consistent, with the two main conceptual approaches (STPR and SOC) to discounting.

Section 4 discusses three other arguments commonly advanced in favour of a DDR that are less connected with the underlying discounting approach. This includes dual discounting where some outcomes, notably for life and major health outcomes and the environment, are discounted at a different (lower) rate.

Section 5 discusses the various advantages and disadvantages of employing a DDR in these various contexts.

Section 6 describes international approaches to the choice of SDR, noting the NZ Treasury (SOC) approach, and some international adoptions of DDR. A comprehensive review of international practice is outside the scope of this report.

Section 7 provides a further possible approach to valuing long-term non-market goods. This can allow for significant values to be attributed to long term outcomes, especially for environmental impacts, separately from the choice of discount rate and discounting methodology.

The final section summarises the main conclusions of this report.

An Appendix sets out the general background to the issues: the aims and process of cost-benefit analysis (CBA) and the key role of valuing (discounting) future costs and benefits in the CBA process.

Main Conclusions#

If applied, a DDR should be integrated with a core discounting approach, i.e.: with the STPR approach or with one of the two SOC approaches to discounting. It is not a separate discounting method.

Of the 18 countries and international agencies reviewed in this report, 12 (including New Zealand) adopt a SOC approach to discounting; 6 use the STRP approach. The discount rate with the SOC is usually 5% to 6% and sometimes higher. The discount rate with the STRP is usually around 3% to 4%.

It appears that all six countries or agencies using the STRP approach adopt some form of DDR, but often only after 50 or 75 years. Only two countries using the SOC method appear to adopt a DDR, and they do so weakly.

Under the STPR approach to discounting, the main technical arguments for adopting a DDR are (i) uncertainty about the rate of future growth of incomes (consumption), (ii) a view that the long-term rate of growth may decline, so that more value should be placed on the well-being of future generations and (iii) putting greater precautionary emphasis on long-term external risks.

With the SOC approach, the main technical arguments for a DDR are based on (i) uncertainty about the long-term rate of return on investment and (ii) a view that long-run rate of return on investment may decline. But applications of these points do not lead to DDRs significantly lower than the usual core discount rate of around 5% to 6%, and therefore would have little impact.

The main practical arguments against introducing a DDR are:

  • Difficulty in calibrating or estimating the “correct” schedule.
  • The complexity introduced to the CBA modelling process given that the DDR will make little difference in most project evaluations. This applies especially to the SOC approach to discounting.

A counter argument is that it is better to be approximately right, than precisely wrong on the issue of DDRs.

And, once an approach is determined, application in routine CBA need not be overly complex. An approximate stepped profile is easily implemented in Microsoft Excel. And a continuous profile for most specifications can be expressed in a simple formula, such as Weitzman’s (2001) gamma discounting.

On the other hand, other approaches may be able to achieve most of the objectives behind the concept of DDRs. These include:

  • Sensitivity testing at high and low discount rates to reflect the impact of uncertainty. For example, discounting at a high discount rate and a low discount rate, then averaging the NPVs, provides very similar outcomes to a DDR based on the expected net present value approach discussed in the Report.
  • Switching from a SOC/ROI approach to discounting to a shadow price of capital approach. In this approach, long-term benefits are valued more highly due to the lower overall level of the discount rate, irrespective of the term structure.
  • Dual discounting, or relative price adjustments, to ensure that the CBA allows for appropriate valuation of major health and environmental outcomes.