Hacker Newsnew | past | comments | ask | show | jobs | submitlogin
Exponential Discounting (johncarlosbaez.wordpress.com)
20 points by headalgorithm on Oct 25, 2020 | hide | past | favorite | 5 comments


> Ainslie’s research showed that a substantial number of subjects reported that they would prefer $50 immediately rather than $100 in six months

I'd probably choose the same thing, but is this because they really prefer the $50 today, or is it because of uncertainty about the $100 in 6 months? I feel like this could be either due to uncertainty about it even materializing, or uncertainty about its value at that point.

Another experiment to consider, to rule out the latter possibility: what happens if you replace dollars with gold? Offer me 1g of gold today vs. 2g of gold in 6 months and I would almost certainly choose 6 months, despite choosing the payment to be made today if in dollars.


If I ask you today, would you choose A or B?

A) 5 chocolate bars on January 1st 2021

B) 6 chocolate bars on January 2nd 2021

Almost anyone would choose B, because:

- the risk of not getting paid is almost insignificantly different between those two days, and

- the value of a chocolate bar will not drop much in a day

But, if offered the choice on January 1st, many people would choose to have 5 bars immediately, rather than wait a day for 6 bars.

Now, you're right, this might be because you're worried about getting paid, and 0 days' risk vs 1 days' risk is quite different. The famous marshmallow test is apparently an example of this: for the kids who had experienced less stable/reliable circumstances, it seemed crazy to wait and risk getting nothing.

But, even when the counterparty risk is ~zero, people exhibit similar time-inconsistent preferences.


Oh definitely. I'd even go farther and say once January 1 comes, if I'd chosen B earlier, I'd be happy I hadn't chosen A—I'd already waited a while, so an extra day would seem like nothing. But if the choice was first given on that day, I would still choose A... even though the situation I'd subsequently end up in would be exactly the same in both cases moving forward.


Transaction costs say it's not worth double the transaction costs for 1 more chocolate bar


Useful in theory, not in practice.

Choosing a model for analytic tractability is a strange quirk of theoretical economists.

There's plenty of excellent software to do the maths for you, so why not choose a model that matches reality more closely?




Guidelines | FAQ | Lists | API | Security | Legal | Apply to YC | Contact

Search: