The late Nobel prize-winner Daniel Kahneman and his late research partner Amos Tversky provide the answer. This is set out in part 4 of Kahneman’s excellent book Thinking, Fast and Slow, which I attempt to summarise in this article. (I hope to do it justice and apologise for the over-simplification!)
Losses loom larger than gains of the same amount
On average, losses feel twice as bad as the positive feelings of a gain of the same magnitude (they range between 1.5 and 2.5 times greater). Imagine your positive feeling if you find a $100 note on the footpath. However, if you lose the $100 note, your negative reaction to the loss will be about twice as strong as the positive feeling of the gain. In other words, you would need to find around $200 (for most people, between $150 and $250) to compensate for losing $100.
This relates to the concept of ‘loss aversion’ – we are driven more strongly to avoid losses than to achieve gains (probably for evolutionary reasons). The reference point can be the status quo (such as current wealth) or a future goal or target (say, a desired settlement amount). Exceeding the target is experienced as a gain, while not achieving the target is experienced as a loss.
This suggests that the more realistic the settlement target is and the less any concession is viewed as a loss, the greater the chances and the higher the utility of settlement.
Decision weights, possibility effect and certainty effect
We over-weight unlikely events and under-weight likely events. Kahneman refers to decision weights. For example, people give the probability of a 5% gain a decision weight of 13.2, overweighting it by more than 2.5 times. This is known as the ‘possibility effect’ – the possibility of a rare event is over-emphasised in our minds.
At the other end of the scale, a 95% probability of a gain is given a decision weight of 79.3: a 5% reduction from a certain outcome feels like a 20% reduction. This is called the ‘certainty effect’, where highly probable outcomes are under-weighted.
Diminishing sensitivity to gains and losses
Kahneman invites us to consider two problems. Which option would we prefer in each?
- Get $900 for sure OR a 90% chance to get $1,000?
- Lose $900 for sure OR a 90% chance to lose $1,000?
In problem 1, most people are risk averse and prefer the sure thing: subjectively, we value the gain of $900 at more than 90% of the value of a $1,000 gain (even though the expected value is the same for both options, at $900).
In problem 2, most people are risk-seeking and take the gamble: subjectively, the negative value of a $900 loss is more than 90% of the negative value of a $1,000 loss (even though the expected value is the same for both options, at -$900).
Fourfold pattern
The above factors combine to produce Kahneman and Tversky’s ‘four-fold pattern’
| GAINS | LOSSES |
HIGH PROBABILITY (Certainty Effect) | 95% chance to win $100,000 Fear of disappointment RISK AVERSE Accept unfavourable settlement | 95% chance to lose $100,000 Hope to avoid loss RISK SEEKING Reject favourable settlement |
LOW PROBABILITY (Possibility Effect) |
5% chance to win $100,000 Hope of large gain RISK SEEKING Reject unfavourable settlement |
5% chance to lose $100,000 Fear of large loss RISK AVERSE Accept unfavourable settlement |
In each quadrant:
- The top row is a possible gamble or outcome.
- The second row is the operative emotion.
- The third row shows how most people act when faced with a choice between the gamble and a sure thing of the same expected value, for example, between a 95% chance of winning $100,000 and $95,000 with certainty.
- The fourth row shows the expected attitudes of parties to a dispute.
The possibility effect influences the bottom row. In the bottom left-hand quadrant, the overweighting of the chance of a large gain explains why we buy lottery tickets and overpay for a very small chance of winning. The bottom right-hand quadrant is where insurance is bought. People overweight the possibility of a large loss and pay more than its expected value to avoid it.
The certainty effect influences the top row. In the top left-hand quadrant, people are risk-averse and will pay a premium to lock in a certain gain.
On the other hand, in the top right-hand quadrant, people are risk-seeking and will gamble to avoid a certain loss now, even at a high risk of a larger loss later.
This is due to (a) diminished sensitivity: a $100,000 loss does not feel much worse than a $95,000 loss; and (b) decision weight: for a 95% chance, the decision weight of 79.3 feels much less than the probability.
This is where bad options and desperate gambles are taken. The small hope of a favourable outcome is magnified, and the high chance of disaster is minimised. As we feel losses more than gains, we are hard-wired to avoid the pain of a certain loss now, and the remote possibility of relief is powerfully attractive.
Consequences for litigation
Kahneman cites work done by legal scholar Chris Guthrie and concludes that a party with a strong case is likely to be risk averse and be attracted to the certainty of a settlement and the avoidance of the (low but magnified) risk of disappointment. Conversely, a party with a weak case will be loss-averse and risk-seeking, prepared to gamble (overweighting the chance of success) rather than accept a certain loss.
As Kahneman says: In the face-off between a risk-averse plaintiff and a risk-seeking defendant, the defendant holds the upper hand. This is reflected in negotiated agreements where the plaintiff settles for less than the trial’s expected value (expected outcome x probability). It explains why vexatious litigants may win a more favourable settlement than the merit of their case justifies.
It also explains why the other party, with a weak case, won’t accept your perfectly reasonable settlement offer!
Unfortunately, you may need to pay ‘over the odds’ to settle. As a one-off proposition, this may be acceptable, but Kahneman sounds a note of warning. In each quadrant of the four-fold pattern, systematic deviation from the expected return (for example, paying a premium to avoid a small risk of losing) for many similar decisions is expensive in the long run. It leads to sub-optimal outcomes for that party.
