Has the Treasury produced a lemon?🍋
New discount rates for cost-benefit analysis
The New Zealand Treasury has recently published new guidance on what discount rate to use in cost-benefit analysis (CBA) of government proposals. This is a stark change from previous practice — a change that seems problematic.
Specifically, the Treasury now recommends distinguishing between commercial proposals and those with environmental or social outcomes. For commercial proposals, it proposes an updated discount rate of 8%, consistent with the approach taken by the Treasury in previous years, but for social and environmental proposals it now proposes a discount rate of 2% for short-term proposals of less than 30 years, declining to 1.5% and 1% for proposals of long duration (30+ years).
Before this change in policy, all proposals were discounted at the same rate, except for adjustments to reflect different risk profiles. So, the main change is a very large reduction in the discount rate for both short- and long-term social and environmental proposals.
The Treasury has not disclosed its justification and its thinking in writing. However, at a recent seminar,1 its representatives said that they had placed heavy reliance on a paper by Professor Arthur Grimes, which is on Treasury’s website.
There are three main reasons why I think this change is problematic. I will explain these in turn.
1. The distinction between commercial, and social or environmental proposals, is muddy
Professor Grimes’ paper says that “the use of dual (or, more generally, multiple) discount rates embodying different rates of pure time preference for different goods can reflect differing societal weights placed on future streams of services from differing goods.” He recommends that dual discounting should be investigated further. Further investigation seems sensible, since neither he nor the Treasury identified any empirical research that would provide a basis for a 2% discount rate for social or environmental proposals.
In any case the distinction between “commercial” and “social” proposals is far from clear. Most government initiatives have elements of both. The guidance in the Treasury Circular is difficult to apply. This creates scope for inconsistent application, and for agencies to classify proposals strategically to benefit from the lower rate.
To illustrate, consider the construction of a new road. The main benefit is usually travel time savings. There is no market for these. Is it therefore a “social” investment? Applying the criteria in the Treasury circular, I find “industry and trade” impacts, and “business and economic” impacts. But there is no expectation of a commercial financial return, the road will not compete with commercial private sector firms, risk is not commercialised, and investments are not “partially privately financed on a commercial basis”. There are distributional, equity, public safety, health, environmental and climate change impacts. On balance, it would therefore seem to be a social investment.
But if consideration is being given to whether the government should build the road itself or whether to let a contract under a public-private partnership (PPP), then using a 2% discount rate would massively bias the choice towards government provision.
2. Opportunity cost ignored
Neither the Treasury guidance nor Professor Grimes’ paper, nor the technical paper on Treasury’s website that sets out the calculation of the social rate of time preference (SRTP), seem to take into account the opportunity cost of resources employed.2
Opportunity cost is a fundamental concept in the theory of value. It refers to the value of the next best alternative foregone. In other words, it’s what you give up to get something else.
The discount rate is in effect a hurdle rate of return. It means that proposals with a rate of return exceeding the discount rate are acceptable for investment. If another proposal has a higher rate of return, then that means that it too would have a positive net present value (NPV), even under a higher discount rate.
If the discount rate is set so that it represents the rate of return of the most valuable proposals available to government, then in effect it becomes equal to the proposal’s opportunity cost. 2% does not represent any investment’s opportunity cost, as clearly it is possible to invest at a higher rate of return.
It is important to remember that the thrust of the Treasury’s long-standing CBA guide is to provide methodologies to adjust all intangible and other non-market values to a common metric so that they are comparable to market values. So, why should one accept a proposal with a 2% rate of return, when there clearly are other ways of investing resources that yield more value?
3. Lack of transparency
At the recent seminar referred to above, the Treasury representatives indicated that the purpose of the change was “to promote long-term social and environmental outcomes relative to purely commercial proposals”. However, it is clear from the Treasury circular that the low rate also applies to social and environmental proposals that are short-term.
The Treasury, and Professor Grimes’ paper, justify promoting certain kinds of proposals over others with an appeal to ethical considerations. But the ethical views espoused by the Treasury will not necessarily be shared by other officials, economists, the government of the day, or by the population at large. It seems inappropriate for such views to be buried in a technical device that most non-economists don’t understand. Decision-making would be more transparent if decision-makers are presented with a neutral analysis. Ethical or political considerations should be introduced into decision-making separately.
A low discount rate can also represent a hidden subsidy. Consider the example of a museum or a park, which might have a low rate of return even when visitors’ valuations (willingness to pay) and other intangible values are taken into account. The Treasury would presumably classify these as social investments. It is legitimate for government to make such investments. But if the CBA is given a low discount rate and the cost of capital is not recognised, then not only will the project seem cheaper than it is in opportunity cost terms, it will add to Crown debt, and create an interest rate and repayment liability that is not recognised by the CBA. It will involve a subsidy that is partially hidden at the time of making a decision, reducing fiscal transparency.
Other considerations
Much discussion about the choice of discount rate seems to be motivated by a concern about the impact of climate change on future generations, or more generally by a concern about environmental degradation and species extinction.
The consequences of climate change are a unique and unprecedented problem for the world. A CBA, which is a partial equilibrium analysis, is not the right tool for this. This is well argued in a 2022 paper by Stern, Stiglitz and Taylor.3
A common rationale for using lower discount rates is that high rates "de-value" the future, leading to little value being placed on environmental degradation that impacts on future generations. But this ignores that scarce or irreplaceable resources tend to increase in value. A well-conducted CBA should reflect this in its forecasts. Often this has not been done very well, and maybe this is where the Treasury should look to develop more guidance. Lowering the discount rate is a blunt and imprecise way to distinguish between replaceable and irreplaceable resources.
A more general observation is that people often seem wedded to a particular solution or outcome, and don’t like the fact that discounting “de-values” that solution. As Grimes illustrates in his paper, any value discounted over 100 years becomes almost negligible. But this is looking at the issue through the wrong end of the telescope. A more positive approach is to remember that a proposal usually involves a consumption sacrifice (an investment of some kind), which it is hoped will provide positive future outcomes. The aim is surely to find the proposal that maximises the value of those future outcomes relative to the sacrifice made. To illustrate, if we invest $1 today in a project that has a 5% rate of return, and the benefits accumulate over 100 years, then that $1 will grow to more than $130 in 100 years’ time. But if the project only has a 2% rate of return, then over the 100 years the $1 invested will only grow to $7.24!
Conclusion
It is not clear what the Treasury is basing the 2% discount rate for social and environmental proposals on. On the one hand, they derive it from a theoretical calculation of the SRTP, which however ignores the opportunity cost of capital; on the other they claim it to be motivated by ethical considerations.
Moreover, the distinction they are drawing between commercial and other kinds of proposals is far from clear. And it ignores the long literature on economic valuation methodologies to enable comparison of social and commercial values side by side.
Decision-making would be more transparent if decision-makers are presented with an analysis that is as neutral as possible, as has been the practice to date, using a discount rate that has some empirical basis and reflects the opportunity cost of capital. Any ethical or political considerations should be presented to decision-makers separately.
By Dieter Katz, ex-Treasury official.
Treasury / Motu Joint Seminar: The Treasury's Social Discount Rate, 16 March 2025.
See, for example, Moore, Mark A., Anthony E. Boardman, & Aidan R. Vining. (2013). More Appropriate Discounting: The Rate of Social Time Preference and the Value of the Social Discount Rate, Journal of Benefit-Cost Analysis, 4(1):1-16. Also Newell, R., W. Pizer, & B. Prest. (2023). The Shadow Price of Capital: Accounting for Capital Displacement in Cost Benefit Analysis, NBER Working Paper, w31526.
Stern, N., J. Stiglitz, & C. Taylor. (2022). The Economics of Immense Risk, Urgent Action and Radical Change: Towards New Approaches to the Economics of Climate Change, Journal of Economic Methodology, 29(3).
I have spent my entire life in genetic research, which involves large upfront costs and slow yet permanently accumulating benefits over time. We typically aimed for an IRR in the 20-30% range after inflation adjustment. Of course, part of this was to account for the risk involved. Hence, my surprise when I became involved in genetics to mitigate the impact of global warming led me to Stiglitz and others' 1% discount rates and the implicit "hack" embedded in GWP100. The change Dieter is referring to here will also result in unintended consequences and confusion. In the case of global warming, the usual additional cheat is to also penalise past emissions retrospectively, much like adding a benefit to the analysis of tar sealing the road 20 years ago.
My biggest criticism about the change is that Treasury tried to have it both ways - which will lead to the exact perverse 'usage decisions' that Dieter refers to in the article.
I would have much rather preferred a consistent update to the discount rate - either stick with the conventional wisdom on a 6%, SOC-based rate, or have insisted that everything utilised a 2% SRTP rate. As is, it's now going to be gamed substantially.
Treasury also requires the sensitivity test of the other discount rate, which is just going to confuse Ministers, rather than help them make decisions ("What do you mean the BCR is negative, now?")
That said, it is remarkably difficult to get large scale, intergenerational infrastructure investments to stack up with a 6% discount rate - this was not the only reason, but undoubtedly a contributing factor to our gaping infrastructure deficit.