Policy under a zero social discount rate

Professor John Quiggin has argued for years that a high private discount rate, or an individual’s preference for present over future benefits, does not imply that society as a whole should apply such discounting to evaluate long term projects.

The good professor has been kind enough to formalise this principle to demonstrate its internal consistency.  He concludes

Unless society is willing to discriminate in favour of earlier born individuals against their later born contemporaries, the pure rate of social time preference must be zero, regardless of whether or not individuals display pure time preference with respect to their own lifetime consumption.

Before I consider how one logically reaches this conclusion, I will start by quickly highlighting the importance of this principle.  Professor Quiggin highlights the issue in the context of action to mitigate climate change.  Because costs of mitigation occur in the ‘present’ (roughly speaking), and benefits occur in the future, a positive discount suggests forgoing present costs and enduring whatever future costs arise from climate change.

For me, a zero rate of social time preference (meaning a total discount rate in the order of the real rate on government bonds, or around 2 per cent), has implications for role of public spending and investment more generally – infrastructure, social programs, or any government spending directed towards long term benefits.  It suggest that ‘gold plating’ of infrastructure may not be such a terrible thing (provided it reduces future maintenance costs or improves the lifespan of infrastructure).  Also, it suggests that preventative measures are usually socially justifiable – whether they be ‘soft’ social investments such as support programs for struggling youths, or physical investment in flood protection infrastructure.

To follow the logic, it is best to start by considering why time preferences exist?  The short answer is opportunity cost.  An individual has a finite life, and needs to allocate their time in a way that provides them appropriate benefits.  They may not maximise in the strict economic sense, but they do trade-off various options.  Forgoing some kind of pleasurable activity now requires a chance at something even better in the future, since you never really know if the expected future benefits will arrive, or if you will be around to enjoy them.

But society does not have a finite life in any tangible sense.

So here’s the professor’s basic idea.  If society values current benefits to each individual equally, at whatever age they happen to be, then it must also value these benefits equally when the current younger generation are older.

For simplicity imagine a two person society – one child age 10 and one ‘verge-of-retirement’ boomer at age 60.  The government is evaluation a project, say a rail line, that is likely to provide community benefits in the form of reduced travel time, freight costs and so forth, in the order of $10million per year.  There are valid reasons why a private firm won’t invest in this line, and the rail line is expected to last 50 years.

Traditionally, one would conduct a cost-benefit analysis with the cost in the current period, and the benefits spread over future periods, and discounted back to the present at some rate, say 5% for simplicity.

If we discount an expected 50 year future stream of $10million benefits each year we find that the project is worth about $183million.  If it costs more than that, the costs outweigh the benefits, and it is apparently not socially optimal to invest.

But now let’s think about intergenerational fairness.  We valued $1 of benefit this year for both the 10 year old and 60 year old at $1.  But we value the benefit to the 10 year old in fifty years time, when they are 60, at a mere 9 cents.  Sure, we discount future costs as well, such as maintenance, but we can’t ignore the conundrum that benefits to individuals alive today, when they are the same age, are valued differently.

If we think at a macro level and consider costs in resources terms – labour, materials, capital, land and so on – then we say that diverting these resources from other productive uses in the future, is less costly than in the present.  Better to divert society’s resources to maintenance in ten years time, than to building a more durable piece of infrastructure in the first place.  But what this simply means is that the people who won’t be alive in ten years time incur none of those future costs, while younger people who are still around, incur them all.

From the perspective of society as a whole, our opportunity cost is all other production being undertaken at any point in time.  Divert 300 workers to fix a rail line, reduce output in other areas by whatever those 300 worker would have contributed (yes, assuming stable unemployment).

So what does this mean? Does it mean we should spend $500million building that rail line instead, which is the value of the un-discounted expected benefits for 50 years?

What it really means is that cost-benefit analysis should use a zero rate of time preference, and the resulting very low total discount rate.  But more importantly, cost benefit analysis with a near zero discount rate should be used as a tool for ranking potential public infrastructure or social investment to see which candidate projects are likely to deliver the best returns.   This ranking will be different to the ranking when using a larger positive discount rate that would be suitable for evaluating private projects.

This principle also suggests that the ‘gold-plating’ of public infrastructure can often be socially beneficial if it reduces future costs significantly. Meaning that in the long term, maintenance costs (including closures etc) are significantly reduced.

It means that public spending should look uneconomic when evaluated by measures suitable to private enterprise.  But ideally should facilitate private markets where private markets can produce better social outcomes.

A zero social discount rate raises plenty of new questions about selling State owned assets to pay off debts, even beyond the fundamental economic irrationality.  Exactly how much were these assets worth from a social perspective? New Queensland Premier Campbell Newman intends to sell a number of government office buildings in Brisbane, meaning future governments will have to lease space from private building owners – a cost that shouldn’t be substantially discounted.

Economists are typically terrible at aggregating from micro principles to the macro economy.  But the principle of a zero rate of social time preference, in a society that values each year of each persons life equally, is one of those cases where aggregation from the individual to the society has not merely been assumed away.


  1. But how does 0% factor in innovation and redundancy? That rail line of 50 years might be superseded by flying cars in 10 years time, and yet at 0% the cost will still have to be borne by the following generations.Surely, the 5% rate accounts for at least some sort of technical progression?

    • drsmithyMEMBER

      Conversely, how to account for it becoming more valuable over time ? For example, peak oil making long-distance trucking and near-distance flying vastly more expensive ?

      • Yes, a non-zero rate doesn’t take the innovation and redundancy into account any more than it does side-effects and unforseen benefits. An electrified track running on solar power would counteract rising oil prices, for instance.

    • But innovation cannot occur if there is a steep discount rate – funds and manpower are tied up in maintaining an aging infrastructure, and the ‘savings’ from not investing in innovation or upgrades in the short-term outweigh the costs.

  2. To clarify, I’m advocating a zero rate of pure time preference. That still allows for a positive discount rate, determined by opportunity cost, which will in turn be influenced primarily by technical progress. The evidence suggests that the appropriate real discount rate is around 2 per cent, which happens to be close to the real rate of interest on government bonds

    • Rumplestatskin

      Thanks for clarifying John. Glad to see you reading MB. The post has been updated to more clearly reflect your principle.

      A question. How do you measure opportunity cost for government spending except in terms of other spending? Why compare with saving or bond yields? Investing is how the economy ‘saves’ for the future.

      As a side note, what are your thoughts on hyperbolic discounting in cost-beefit analysis of very long term projects?

    • The rail example is interesting, but rather basic. Is it possible for you to present an argument on the NBN using this?
      Is it possible to tie the PMG (Telecom) rollout to support the argument as a proof?

  3. Just Dismal 2

    Honestly, this is an aparently logical, but realy wacky idea. In fifty years time, the same one dollar benefit will be worth 9c in relative terms, because the economic growth would have given us many more choices. How do you know that in fifty years my preferred method of comunication is physical transportation? Economists can’t get their head around the value of growth and innovation. Al these grand theories can’t handle it.

    • Are you willing to bet on continuing real economic growth over the next 50 years of >2% p.a? I know it’s only a small hurdle, but I have my doubts.

      BTW, this is an attempt to rectify the value of growth and innovation – the steeper your discount rate, the less likely you are to spend money now to benefit in the future.

      • Just Dismal 2

        That’s wrong. The discount rate doesn’t determine the amount of future benefits. It just favours projects with faster returns, therefore increase your choice of reinvestments. Look at it the other way around, the greater the discount, the greater the future value for a given dollar of today. So the future benefits will be greater!

        We need to get our head around correctly, and not swayed by slogans that appear logical, but actually are not.

        • Rumplestatskin

          You are right. The choice of discount rate doesn’t change the actual returns on investment or the quantity of future benefits. It changes the ranking, or choice, of competing alternatives when there are differences in timing between costs and benefits.

          A low (zero or near zero) discount rate still favours projects with higher total returns in the long run, not just fast returns.

          Looking at it both ways implies the same thing – high discount rate means you should have spent less in the past for a dollar of benefit today. Not really sure what’s you’re thinking there.

          Maybe I will try an explain this better in another post.

          • Just Dismal 2

            There is a fundamental error in the good professor’s assumption that the later borns want the same type of benefits as the earlier borns. Hence his incorrect conclusion.

        • Lighter Fluid

          Let me clarify:

          the steeper your discount rate, the less likely you are to spend money now to benefit in the DISTANT future compared to the NEAR future.

          Which is what statski is getting at, and what I was trying to refer to, albeit in a rushed manner.

  4. Likewise, the annual benefit to society of a piece of land in 100 years time, ie a year’s rent, can be argued to be the same as the annual benefit today, ie a years rent.
    Applying zero discount to land implies that it has an almost infinite value (the non discounted sum of future rents) and is certainly vastly undervalued today.

    • That depends – land for building a dwelling on or a factory, for instance, would (should) be valued differently to fertile land with the ability to produce other valuable goods. The degradation of arable land is a discount factor; for static (building) land, infrastructure and surrounding economic opportunities changes the value of the land over time – static land in a mining town might be valuable today, but in 20 years time when the mine is emptied out it won’t be worth nearly as much.

    • You only get an infinite value if you look at it under an infinite timeframe, but you do have a point that zero discount rates would show land to be undervalued today.

      If you used a 2% discount rate today (which is seen as zero in this article) and added up the rental benefits of a piece of land over the next hundred years, along with the expected capital value of that land after that time period, and then compared that against the initial capital outleigh and ongoing expenses in collecting that rent – then I think it is fair to say that *any* land will bring positive net benefit.

      But I don’t see those results as controversial. I think anyone would agree that if a piece of land is held for a hundred years it is going to produce benefits that outweigh the costs. Doesn’t this just show that the zero discount rate can reflect a reality that a 7% DR can obscure.

      I mean, under 7% DR we have Governments selling off assets for a quick buck. We all know this benefits us with cash today, but hurts our kids by reducing income tomorrow.

      I think the zero discount rate might be too extreme, but it does make sense to me that governments should use extremely low discount rates to place value on the preferences of future generations.

  5. A really interesting post Cameron. I suppose this is how you can get such differing views of the costs v benefits of the NBN with both being correct!

    • Tiges in 2012

      the line item budget for the NBN is:

      $1 billion for capital works

      $39 billion for government advertising telling us how great it is


      Cost $40 billion. Benefit reliable cashflow for TV stations.

      Seriously, how many newspapers could be saved if the government (Federal and States) spent as much money on corporate welfare/propaganda on them as they do on electronic media?

    • Just Dismal 2

      Interesting, that we mentioned the internet. It just happened not to exist 50 years ago. If we had spent all our resources under zero discount on the then known infrastructure, we might not have got internet after all. Which then of course would have proven the correctness of the zero discount policy. Some smart guys long figured out that by eliminating choices to just one, it will always be right!

      • Rumplestatskin

        Maybe I wasn’t clear enough. Capital won’t last forever simply because you evaluated competing spending options with a zero discount rate.

        It might appear that one can justify spending an almost infinite amount on, say, a bridge, if it guarantees it will last ‘forever’, instead of a spending less and getting bridge with an expected 80 year life. But that’s not true, as there are competing potential projects (although it may justify some cost-effective ways of prolonging the expected life, or reducing ongoing maintenance costs).

        It also doesn’t imply spending all of your resources.

        It simply changes the ranking of projects away from those that provide immediate benefits and long term costs, towards those that provide longer term benefits.

        • How are we to know that in 50 years we still want the same infrastructure? Your logic works only by assuming the choices we face in 50 years will be the same. The internet has exactly proven that to be wrong. The point is that discounting takes this into account, allows us to recoup our investment more quickly, and then we will have new choices when the innvators provide us with new technologies.

          • Rumplestatskin

            I would argue the internet is a great example of the potential upsides of durable infrastructure. After all, most connections run through phone lines, that in your world, may never have been built because high discounting led to future high take up of phones, faxes etc being undervalued.

            I would argue that it is more common to find new higher value uses for infrastructure in the long run, than for society to move technologically beyond it. There are also plenty of lock-in/ path-dependency effects that lead to long term demand, especially for infrastructure networks.

            Which of these can you imagine was not around 50 years ago, and won’t be around in 50 years time?

            Parks, roads, storm water, sewer, drinking water (e.g. recycling/desalination), airports, rail lines, electricity, phone lines (yes, we still need them for phone calls too), ports, dams, government offices…

            Got any other examples?

            Also, in a social setting, surely the pay off to society from programs that have known benefits of helping youths stay out of prison, for example, will still be desired in 50 years time.

            Of course, in many markets where capital and products change rapidly, with much shorter lifespans, and zero discount rate is inappropriate, but government involvement in these markets is probably in appropriate in any case.

          • Just Dismal 2

            how do you know in 50 years we will still prefer roads to rail or telecommuting? Will current desalination technology still the best? Will sea transport remain the same? Will battery technology change electricity distribution?

            Ah, the phones won’t be there.

          • They are certainly decisions with a degree of uncertainty. 50 years is a long time, and a lot can change.

            The way I see it, if you look at a 50 year project and place a high discount rate you are making the bet that the infrastructure will not be needed before the end of the project. If you use a low discount rate you are making the bet that the infrastructure will still be needed.

            Still, if I was a betting man I’d say that parks, roads, storm water, sewers, drinking water, airports, rail lines, electricity, phone lines, ports, dams and government offices would *all* be used in 50 years. Perhaps I might be wrong on one or two of them, but I reckon I’d still have a better pay off than a person who bet against them all by using a high discount rate.

            I guess my point is, a modicum of sense and deliberation is required on a case by case basis, instead of blanket rules for all circumstances.

          • Just Dismal 2

            Easy to say we will need roads at an abstract level. But the reality is about which road. It is also about alternatives. Do we build road A, road B, or NBN? Or hospital? Or spend on education? Discounting helps us select the highest returns and provides us with greater flexibility with investing the returns.

      • The internet was built as a research project on top of existing infrastructure – phone lines between universities. There wasn’t a “should we build The Internet As It Will Exist in 2012 vs should we just build more phone lines” question. Providing infrastructure to support innovation has its own unquantifiable value.

  6. Perhaps the rate of discount should depend on the location we are on the bell curve of our rate of exergy production (energy + productivity gains).

    This is difficult, as we don’t know what future productivity gains we will get (even though some limits e.g. for electricity production are already well understood)

    However I think it would be fair to say we will not get the continued exponential growth from the extraction of finite, high quality energy sources for the next 50 years, let alone the next 20 years. In fact looking at the bell curve for the extraction of finite resources, it would be more prudent to expect energy production from coal, oil and gas will decline exponentionally within the next 10-20years.

    This would mean without huge productivity gains elsewhere in the economy, infrastructure such as electric rail and renewable energy that is built now will have a negative discount rate i.e. will be far more valuable in the future than it is now, as energy flows becomes more scarce and harder to control.

  7. Diogenes the CynicMEMBER

    In my view the effects of peak cheap oil will be with us sooner than the effects of climate change – how would your discount rate affect decisions on priority spending between these two current+future problems?

  8. Rumplestatskin, congratulations, you are obviously smart student and you can challenge very well your professor on a public economics topic. One simple argument in favor of using discount rates is the fact we are living in a capitalist society, where the time value of money is there capital value and everything in our reality is measured by this capital value, which actually is presented by the marginal rate of capital return, which is the interest rate of the central bank of federal reserve bank. The government projects are not in any way exclusion of this rule, especially in the most developed stage of capitalism – its globalization. The only adjustment for social preferences of the capital value, or time value, of any government project costs and future benefits can be the use of social weights. Only the social weights reflect the actual social preferences regarding the distribution of costs and benefits among different socioeconomic groups or different generations.

    We can see how those weights are working in different societies. Take as an example USA government spending during the last 5 years. It is a brilliant example of how government uses social weight to express its preferences towards different socioeconomic groups and generations regarding public spending and the resulting public debt.

    The only agent who can express and use social preferences as a toll for cost/ benefits adjustments of the capital value of public projects is the government itself. there is no such thing as a social preferences based on individual preferences. If it was possible to measure and know what are the agregate social preferences per se, we would have been living in an absolutely different society and economic system. As long as there does not exist such thing as a real measure for social preferences outside of the process of political process and choices, the whole discussion about it is fruitless and pure sophistic. Ask your professor how he measure social preferences and who decides they can be at 0 rate discounted? Maybe it is because the political choice of the Federal reserve about the capital value of money is almost 0 rate???

  9. As I understand it, the O’Farrell government in NSW is also looking to privatise as many of it’s assets as it can too.

  10. “Society is a contract… the state is a partnership not only between those who are living, but between those who are living, those who are dead, and those who are yet to be born.” [Edmund Burke]

    Talking of intergenerational considerations, the 2012 Reith Lectures (this year Niall Ferguson) questions merit of cost burdens imposed on future generations by current generations and continually expanding public debt.


    Links to podcast:

  11. bhandleyMEMBER

    Great article and linked article. Discount rates appear so seductively accurate but due the assumptions stated and unstated surrounding them and the sensitive models they produce seem to me to be much more like being a “fudge factor” in the absence of better theory.

    The main point raised out how to evaluate long term project for policy is incredibly important, but with the number of disasters in economic theory and policy coming to light of late and the inability of the theorists, leaders, policy makers and the mainstream to query long held beliefs around these things, I’m not optismistic that as a whole we’ll do better than flipping a coin when making such choices.

  12. Correct me if I’m wrong, but as I understand this we are talking about two things:

    – the financial discount rate applied to multi-year projects, and
    – the possibility of also applying a “social discount”, which, were it to be applied, would modify the financial discount rate.

    I think it is worth asking why a social discount rate would be anything other than zero. If this rate were not zero, it would be – as the Professor points out – a way of saying that a unit of value derived by a person today must necessarily be different to the same value derived by another person at another time.

    It might be reasonable to suppose that such differences in values exist, but considered from the point of view of an independent observer, we have no way of determining what such derived-values are for any given person or set or persons at any time – whether now, in the past or the future, or for any project, no matter when it was or will be commenced. These values are unknowable and therefore cannot be ranked either.

    Since we cannot determine what these differences might be and cannot rank them, they have to be assigned a neutral value. As the Professor points out, the only neutral discount value to assign them is zero.

    It follows the best we can do is assign an arbitrary discount to the financial (measurable) costs and benefits that will flow in the future as a result of a decision today.

    The question of what this rate should be, surely, relates to risk: to the question of how likely is it that the perceived financial benefits and costs will flow as expected.

    To use 2% – the supposed long-run real bond rate – is to choose to minimize the funding risk attaching to a project. But there are always other risks, even if projects are funded at the lowest possible cost.

    So while I think it is completely logical to use a zero “social discount” rate, I don’t think it necessarily follows that the real bond rate is the right one to select for all public investments. Apart from anything, the use of this rate tends to suggest that the supply of public funding is close to being limitless – something that is obviously a false assumption in itself over the long run.