Delay Discounting

What it is

The idea does its main work in explaining why a person with an addiction is not making a stupid choice so much as making a choice on a differently shaped curve. The reward that is available now is not merely preferred. It is worth more, on the curve that person is actually using.

Popular accounts of this material tend to introduce the concept as though defining it fresh, and the formal study of discounting curves is usually absent from them. This publication records its own identification, and records it as unverified: hyperbolic discounting and the preference-reversal result are conventionally credited to George Ainslie, with the hyperbolic functional form usually given as Mazur's and the wider behavioural framework as Rachlin's. No year, journal or title for any of the three has been traced here, so none of those names should be printed in an article without going to the papers first.

Two behavioural findings carry the applied claim. In a lottery-ticket task reported by Bretteville-Jensen, people were offered less money now against more money later. About one in five active drug users took the smaller immediate sum, against roughly one in twenty-five former users and one in fifty matched controls. And in a study of temporal horizon by Petry, Bickel and Arnett, participants asked to imagine the future gave horizons averaging nine days when opioid-addicted, against nearly five years for controls. That second finding is the most quotable in the set and it is properly cited. Note the author order, because secondary coverage routinely gives the paper to Bickel and the first author is Petry.

In effect

Almost every practical device for self-control is an answer to the shape of this curve. Precommitment, self-binding and the Ulysses pact all work the same way: they move the decision to a moment at which both options are distant, because at that distance the larger delayed reward still wins. The curve has not been flattened. The choice has been relocated to a part of it where the ordering is correct.

The same logic explains why the opening fortnight of any abstinence attempt is the hardest part, and why plans made in the evening for the following morning are systematically more ambitious than the morning will honour.

On causation, both readings stay on the page. Steep discounting may be a trait that precedes and predisposes to addiction, it may be a consequence of it, or both. The ladder in the lottery-ticket task, with former users sitting between active users and controls rather than at either end, is suggestive of change with recovery, and it is a cross-sectional comparison, so it cannot establish that. This publication leans to the both-directions reading for that reason and does not claim the design can decide it.

What it does not say

It does not say that steep discounting causes addiction. The comparison studies are cross-sectional and cannot separate cause from consequence.

It does not say that discounting is a defect. Everyone discounts, and a future reward really is less certain than a present one, so some discounting is rational. The question is always about the steepness and never about whether the curve exists.

It does not support the claim that we are all now vulnerable to prefrontal cortical atrophy from modern abundance, which is a sentence built on McClure and colleagues' imaging study and which that study does not support. A contrast of brain activation between choices involving immediate rewards and choices involving delayed ones tells you which regions were more engaged during which decisions. It says nothing about tissue loss, cortical thinning or structural change of any kind, because atrophy is a structural finding requiring a structural measure and none was reported. The generalisation from a laboratory choice task to a whole population under modern conditions is a further step again, with nothing behind it. Use the McClure result for what it is, which is evidence that immediate and delayed rewards recruit partly different systems, and that is interesting enough.

---