Loss aversion

From the inside it looks like this: you are offered a coin flip that pays $150 on heads and costs $100 on tails, and you turn it down even though the expected value is positive. Or you hold a falling stock because selling would make the loss real. Kahneman and Tversky built this into prospect theory in 1979. People value outcomes as gains and losses from a reference point, not as final wealth, and the value curve is steeper on the loss side, so a loss of a given size weighs more than a gain of the same size. The curve is also convex for losses, which is why someone who is already down will accept a gamble they would refuse when ahead.

The reference point does the work. Whether a change registers as a loss depends on what you compare it to, and that comparison can be set by a price you paid, a salary you had, or a number you were just shown. Framing the same outcome as a loss rather than a forgone gain changes behaviour. Fryer, Levitt, List and Sadoff used this in Chicago Heights schools: teachers who were paid a bonus up front and told they would have to give it back if their students fell short produced larger gains than schemes that promised the same bonus at year end. Free trials that lapse, and 'you will lose your streak' notifications, run on the same mechanism.

The size and even the existence of a general loss aversion are contested. Gal and Rucker reviewed the evidence in 2018 and concluded that it does not show losses weighing more than gains on balance: the effect appears for large stakes and in some tasks, and fades for small amounts or when the reference point is unclear. Their paper drew replies from prospect theory's defenders, and the argument is open. The safe reading is that loss framing is a real and usable effect, while 'losses loom larger than gains' as a universal law overstates what has been measured. Check which version your decision relies on.