Prospect Theory & Loss Aversion

Hi, You have obviously done a fair amount of background on this. I will give you my spin on it all. Firstly I disagree that the figure on pg5 does not imply loss aversion. If you focus only on the right side, an investor starting from zero would get a certain amount of utility from a (say) £1,000 gain, but would not get twice that utility from a £2,000 gain. So if someone has a £1,000 gain (and associated utility / pleasure / delight) and they are offered a "double or quits" they will likely turn it down. That is where I see loss aversion coming into this theory.
As you say, prospects theory is bigger than that. It looks at someone faced with a loss. If you have a £1,000 loss, you have a certain negative utility, and associated sadness. If someone offers you a double or quits gamble, you might go for it because the utility from a £2,000 loss is not twice that of a £1,000 loss. So a weighted average of £2,000 loss and £0 loss is better than the utility of a certain £1,000 loss.
I hope this helps. I dont think in reality that prospects theory suggests that people will definitely be risk seeking for losses. But some will, whereas very few are risk seeking with their profits or gains.

 

I certainly agree that this level of detail is unlikely to be required for SP5. But I think we are actually saying the same thing in different ways. I said a gain of 1000 would be better than a 50/50 outcome of 2000 or 0. So I am saying what you are saying above, but from the start point of +1000 on the x axis. I agree that what you have said above, starting with someone who has nothing on the x axis (and therefore no gain to lose), the chart doesnt show loss aversion as the negative and positive utility appear to be the same. But it seems to be just your starting point. Start the individual with a gain to lose, and the individual is risk averse. Start the individual with a loss to gamble and the individual is risk seeking.

 

I havent seen a version of prospects theory where the graph is extended/amplified on the negative side as you have done above. It has always been symmetrical.

 

Yes, as you mention, it should indeed by steeper on the negative side than the positive. I will see if our chart drawing skills are up to the challenge and change the graph for the next update.
My mention of "gain to lose" was just a way of describing the original Kahneman and Tversky experiment detailed in the notes.
1. Alternative 1: an 80% chance of winning $4,000 and a 20% chance of winning nothing
2. Alternative 2: a 100% chance of winning $3,000.
Option 2 was the "gain" that someone could "lose" if they take the gamble. It is a way of describing why a concave utility function leads to risk-averse decision making and a convex utility function leads to risk seeking decision making.
Having a steeper negative utility function will affect the decision when a person is faced with either zero, or a 50/50 gamble to win £1000 or lose £1000.