Against the Social Discount Rate (Cowen & Parfit) - Weak refutations
I’ve been reading the paper Against the Social Discount Rate (1992) by Tyler Cowen and Derek Parfit for a related project. In the paper they present a series of arguments commonly used in support of a positive social discount rate (as defined in the beginning of their paper linked above) and their counterarguments refuting these arguments one by one. While I think the paper as a whole is sound and well argued, there are two refutations I find somewhat weak. In this post I present the two arguments in question, along with Cowen and Parfits refutations and my objections to the refutations. I ask readers to clarify if I misunderstand the refutations, improve on them to better respond to the arguments or—if Cowen and Parfit are mistaken - make the case that these arguments are indeed sound.
Argument from Special Relations
This non-economic argument is presented in the paper as:
(slightly rephrased)
We ought to give more weight to the interests of certain people, like people ought for their children or for a government for its own citizens, and to a lesser degree to their children, and even less to their children’s children and so on.
The full refutation may be read in the paper, but here is my summary:
(on page 149-150 in the pdf linked above)
This argument might support a new kind of discount rate, not based on time itself, but for degrees of kinship. But these two—time and kinship—might radically diverge. Degrees of kinship is not perfectly correlated with time. Further more, a discount rate based on kinship should at some point cease to apply or at least asymptotically approach zero. We ought to give some weight to the effects of our acts on mere strangers, but not more than to effects on our own descendants. These two notions are in conflict as only zero is strictly below an asymptote approaching zero, and zero weight is of course not “some weight”.
Nor should such a discount rate apply to all kinds of effects. Governments might priorities its own citizens for benefits, but this does not apply to social costs. Should a government discount harm affecting aliens over its own citizens? The special relations should make no moral difference.
My comments and objections
This refutation takes the same form as several others that Cowen and Parfit brings forward. Yes, there may be some correlation of time and this thing X (here: kinship), but using a discount rate of time to represent degrees of kinship is too crude and in some cases even counterproductive.
My objection to the refutation is that I don’t think Cowen and Parfit is right in their assertion that people thinks only benefits should be distributed favorably to members of their own special relations group, and not costs. In the paper they give an example of the US government moving nuclear tests from Nevada to the South Pacific so that fewer Americans are at risk of radioactive exposure, even if this increases the exposure to a greater number of non-US citizens. Cowen and Parfit argues that most US citizens believe benefits should be favored to other US citizens, but it is immoral to move costs to non-US citizens. This asymmetry of benefits and costs seem unsupported to me. If we take the utilitarian view that this is indeed immoral, we’ve already accepted the utilitarian view. Then we ought to distribute benefits in such a way that maximizes utility regardless of nationality and time of life, implying a zero percent discount rate from the start.
My point is that for those who believe members of their special relation group are entitled to more benefits than those outside of the group will not necessarily be convinced of the utilitarian argument that costs should be equally distributed.
Argument from Double Counting
This argument, which Cowen and Parfit calls an altruistic argument, is presented as the following:
(found on page 157, slightly rephrased)
The present generation is altruistic towards future generations, regardless of the discount rate. Using a zero discount rate on top of this constitutes a form of double counting and improperly weighing future lives too much.
Cowen and Parfits refutation
(follows on page 157)
What this argument calls double counting, however, appears to be a proper counting of interests. If a person benefits from the rescue of his best friend, for instance, it is not double counting to consider as benefits both the value of the life saved and the value of the friendship to the other party.
Furthermore, they provide an example to illustrate their point: A gift to future lives does not give the present generation preferential discount rates when sharing common goods. Giving an altruistic gift provides no privileges to the giver. In a pure utilitarian view where this gifting is in it self immoral, implementing a positive discount rate to compensate is a crude, inaccurate measure. Individuals of the present generation which are not providing immoral, altruistic gifts to future lives are thus, wrongfully, benefited with a positive discount rate—and conversely, future lives who are not receiving any benefits from this gifting are still disadvantaged by the positive discount rate.
Moreover, it is not obvious that the present generation is altruistic towards future lives. Unsustainable use of natural resources, polluting and increasing the risk of human extinction are all costs the present generation put upon future lives. If the present generation in sum steals more benefits from and incurs more costs upon future lives than they act altruistic towards them, this calls for a negative discount rate.
My comments and objections
This response also seems rather weak and isn’t responding to the strong version of this argument. Firstly, the example of the person rescuing his best friend is not comparable, because none of these two benefits—the value of the life saved and the value of the friendship --- are not transferable between them. Saving one persons life (in this example) does not shorten or deteriorate the other persons life, and the value of friendship is experienced by both parties.
A better steelmanning of the argument would be: Consider that you in your altruism give an apple to your brother, and then later you both found another apple, which you share equally. In sum, your brother got one and a half apple, while you got only a half. Your brother both benefited from your altruism and the equal distribution of the common benefit. A more equitable result would be if your brother gave his half of the second apple to you, such that you both end up with an apple each.
I agree that by some moral views, it is not right that a voluntary provider of gifts should be given any later privileges. However a pure utilitarian view calls for a redistribution of benefits in favor of the giver if, by giving the gift to future lives, the giver sets himself in a more disadvantaged position than the future lives. Note that in a pure utilitarian view, such a gift would be immoral in the first place. For this argument to be valid, three further claims must be substantiated:
Current lives are providing net positive, transferable benefits to future lives.
A positive social discount rate will offset these benefits from future lives and back to current lives.
Future lives must be expected to be at least as well off after applying the positive social discount rate.
I believe it is unlikely, but it could be the case that these three claims are true to an sufficient extent. If they are, then this is a valid argument for a positive discount rate, even if it is a crude tool. I don’t feel Cowen and Parfit do a good enough job of refuting this point.
End section
I’ve presented here two of Cowen and Parfits twelve set of arguments for a positive social discount rate and their refutations, of which I had objections to the refutations. I am largely convinced that a positive social discount rate, at least at the magnitude used in most countries today, is immoral, but I feel the weak refutation of the two arguments above could be improved.
I therefore ask for feedback on my comments on the arguments and refutations, and more importantly, suggestions for improved refutations on these two arguments.
In my humble opinion, you are totally correct about the first argument and Cowen and Parfit are totally correct about the second argument. (Note: I haven’t read the paper just your post).
1.
The first argument is philosophy. If a person genuinely believes that a state has a greater duty to its current citizens today than to future citizens then that person should probably apply a social discount. All that can be done to counter a philosophical intuition is to point out that there are intuitions that would suggest otherwise, clearly unpersuasive to someone who doesn’t share your intuitions you.
That said I think the Cowen and Parfit argument could be stronger by pointing out the mutual benefits or intergenerational trade, our place in history and how much we benefit form forward thinking ancestors, and the benefits to us of planning long term.
2.
Cowen and Parfit are correct this is not a case of double counting. The whole point of a social discount rate is to allow your economic models to map your goals. So the fact they align with your goals is not double counting, it is just counting. It would be like claiming that if I would intuitively buy tasty food, then decided to build a model to map out my preferences, I should discount the value of nice tastes because I already consider nice tastes. Which is rubbish as I would just end up using the model (rather than my intuition) to choose what to buy and then having less nice tasting food than I would ideally like.
Now you can use discounts to adjust for biases. For example if you know you always overestimate the value of tastiness of food compared to other factors, even after applying your model, then you could apply a factor to try to counter this intuition. (Real world example even after applying models people underestimate construction costs due to optimism bias etc so add a factor to increase estimated construction costs). But this is if you feel you have a bias that does not match your goals, even after using a model to make decisions. Making the case for this would require some empirical evidence that such a bias exists. But the evidence does not point in this way which leads to the second point that Cowen and Parfit raise that you do not discuss which is a key part of the argument.
All the evidence suggests that humans do not value the future as much as they would ideally like to. If anything an empirical examination comparing what we do for the future compared to what we want for the future should suggest (as Cowen and Parfit highlight) a negative discount rate, to push back against availability bias and political short-termism etc. (eg see: https://www.bbc.com/future/article/20190318-can-we-reinvent-democracy-for-the-long-term)
Hope that helps.
Thanks for writing this up! Minor point:
> I agree that by some moral views, it is not right that a voluntary provider of gifts should be given any privileges, but as Cowen and Parfit admits, this is not the case in a pure utilitarian view.
Maybe I’m getting confused by the double negatives, but isn’t this backwards? A pure utilitarian would argue that no one has any special privileges, right?
Apart from that minor point though, I would be interested in refutations to the objection.
Thanks for pointing out these unclear sentences. I’ve made some changes in this paragraph to make my point more clearly.
The first part of the sentence remains; in some views, it is not right that a giver of gifts get any privileges on other benefits. But in a pure utilitarian view, this might be the case in some sense. If one party provides a gift to another, otherwise equal party, this will create an inequality that decrease the total utility. A pure utilitarian view will demand that a redistribution of benefits should follow to restore the equal situation.
Of course, the utilitarian will not use the term “rights” or “privileges” to argue the case for a redistribution after the gift. Also, it is worth pointing out that in a utilitarian view the initial gift is immoral as it decreases total utility, but this is a bit beside the point as this gifting is an assumed fact with this argument.