Negativity Bias and Loss Aversion: The Fault Is Not in the Reflex but in the Signal

You are sitting on the train, phone in hand, and on the screen a city is burning that you did not know existed until three seconds ago. A woman fills the frame, in slow motion, the audio is mixed onto her voice, and your body does something strange in that moment, because it releases what it would release if the fire were in the house next door. Only it is six thousand kilometers away, it changes nothing about your afternoon, and you could do nothing anyway, and still the tension sits in the back of your neck as though you were about to get up and run.

The first reflex is to treat this as a weakness, a kind of over-sensitivity you ought to be able to switch off with a little more discipline. The reflex is not a defect, though, but a sensor doing exactly what it was built to do across hundreds of thousands of years, and the interesting part begins only with the question of what this sensor actually responds to and who has their hand on the dial today.

Two sensors that were built exactly right

A human being carries two old sensors around, both with the same bad reputation, and both undeservedly so. The first is the negativity bias, the tendency to weight negative signals more heavily than positive ones, and Roy Baumeister and his colleagues spelled it out across psychology in 2001 in a much-cited paper with the dry title "Bad Is Stronger Than Good." The variable this sensor was calibrated to was nearness, because a bad event within one's own band, within a few meters and a few people, was once genuinely life-threatening, while the same event three valleys away was nobody's concern. So the sensor does not measure importance, it measures nearness, and it does that remarkably well as long as nearness is real.

This is exactly where the modern information environment goes to work, without meaning any harm. A news format supplies the face, the slow motion, the on-location audio, the emotional loading, meaning precisely the cues the sensor once used to recognize real nearness, and the sensor cannot separate simulated nearness from real nearness because it never had to learn how. In the world it was calibrated in there were no images of catastrophes on the far side of the earth, there was only what one saw with one's own eyes, and what one saw was by definition near.

The second sensor is loss aversion, the tendency to weight a loss more heavily than an equally large gain, ever since Daniel Kahneman and Amos Tversky described it in their prospect theory in 1979. How large this overweighting is exactly is still disputed today, because the often-cited rule of thumb, that a loss weighs roughly twice as much, comes from a single measurement in 1992, and Gal and Rucker showed in 2018 that the effect appears less often and more weakly than the textbooks suggest. That losses weigh more at all is barely disputed, only the neat number behind it is shakier than it sounds.

This sensor too was calibrated to something real, namely to actual scarcity, because resources really were scarce and not freely replaceable, and whoever lost a handful of grain may have lost the meal for which there was no resupply. Today scarcity is manufactured where there is none, through the countdown on the booking page, through the "only three left," through the "whoever does not move now misses the boat," and the sensor responds to the manufactured signal just as reliably as to real scarcity, because it cannot tell one from the other.

It would be convenient to point at media or marketing here, but that would miss the point, because both are only doing their job, and the job is to adapt to whatever works on the audience. A format that simulates nearness gets attention, a countdown that claims scarcity sells, and whoever refrains loses to whoever does it. So the problem is not moral but structural, because the environment can manufacture the signal the old sensors respond to, artificially and as often as it likes, while the reality behind it fails to keep up.

Why knowing about it does no good

At this point almost every text wants to reach the same conclusion, namely the advice that you should just look more consciously, choose your intake more carefully, become more aware of your own distortion. That very advice repeats the mistake it claims to fix, because Kahneman himself stressed again and again that knowing about a bias barely changes behavior, and he observed it in himself after half a lifetime of writing about these distortions. The fast, automatic system that responds to the nearness and the scarcity signal does not read what the slow, reflective system has learned, and so awareness helps against the reflex about as much as knowing about an optical illusion helps against the illusion itself, because you still see it, you merely know that you are seeing it.

If awareness does not work, though, then the solution must not land on the individual again, because asking him to think against his own sensor means giving him a task that fails exactly where a human being is simply built the way he is built. So the responsibility has to lie somewhere else, and the interesting question is where.

Taking away does not work, adding does

There are two fundamentally different ways to intervene in a mechanism like this, and the difference between them decides everything. The first way is subtractive, it asks the recipient to actively filter out the manufactured signal, to separate the real from the simulated, to ignore the countdown, to recognize the slow motion as production, and it therefore demands attention, knowledge and willpower, meaning precisely the three things that, by premise, are not reliably available. It fails for the same reason awareness fails.

The second way is additive, and it turns the task around. Whoever designs the environment, meaning the leader, the change manager, the marketing team, the format developer, feeds a real, competing signal into the same automatic channel on which otherwise only the manufactured one arrives, and the recipient has to do nothing for it. That is the core of what Richard Thaler and Cass Sunstein called choice architecture, namely not to change a person's will, which does not succeed anyway, but to change what reaches them automatically before the will is even consulted.

For practice this becomes concrete the moment you look at how organizations talk about change. Crisis communication in change works almost reflexively with exactly the two signals from the first part, because "this affects all of us very directly" is a manufactured nearness signal, and "whoever does not move now will soon no longer belong" is a manufactured scarcity signal, and both work reliably because the old sensors fire. Whoever proceeds additively instead places a second, real signal beside them, one that does not rely on threat but on something a person can identify with, and does not leave it to the individual to reason the fear away by himself.

The crisis signal has an inflation built in

And here the matter turns economic, because the crisis signal really does work, it even works fast, only it carries an inflation the purpose signal does not have. A crisis or loss appeal activates what the self-determination theory of Edward Deci and Richard Ryan describes as controlled motivation, meaning action out of avoidance and compliance, and Gagné and Deci gathered in 2005 that this form takes hold quickly but persists more poorly and leaves more exhaustion behind. A purpose or potential frame, by contrast, activates autonomous motivation, meaning action out of identification, which starts up more slowly and then carries itself.

The real difference, though, lies in the wear curve. A fear signal dulls with repetition, so it has to grow louder to achieve the same effect, and Kim Witte showed in her Extended Parallel Process Model that a fear appeal without a credible message that you yourself can do something does not tip into action but into defense, into looking away, into defiance. An identification-based signal does not know this wear curve, it can even reinforce itself through a person's self-perception, because someone who has once acted out of conviction experiences himself the next time as someone who acts that way.

This holds for employee retention just as much as for customer loyalty, because a workforce that goes along out of fear only goes along as long as the fear is fresh, and a customer who buys because of a discount comes back as long as the next discount comes. The real value, the customer lifetime value on one side and genuine commitment on the other, arises where someone stays without having to be frightened or baited afresh every time, and that is precisely the value the crisis signal systematically erodes while it works in the short term.

Why you cannot switch over from one day to the next

It would be tempting here simply to advise the switch, away from the crisis signal, toward purpose, and the JCPenney case shows why that would be the most dangerous recommendation of all. When Ron Johnson switched the chain in 2012 from the endless discount battle to consistently fair, honest prices, he did in principle exactly the right thing, because he removed the manufactured scarcity signal the customers had been trained on for years. Only the customers had indeed been trained for years, and an audience that has learned to measure the value of a purchase by the discount experiences the loss of the discount not as honesty but as a loss, so that comparable store sales fell by a good twenty-five percent in fiscal 2012, the company lost close to a billion dollars, and Johnson was out of his post by April 2013.

The lesson is not that the switch was wrong, but that it was made cold. An audience conditioned over years to a particular signal cannot be switched over by decree, and the exit from crisis communication is therefore not a cut but a migration, in which you introduce the new signal step by step and pull the old one back slowly, long enough for the expectation to catch up. With that the circle closes in an uncomfortable way, because even the designer who wants to rebuild the environment deliberately cannot do so by decision, and that too is again a problem of system design, only one level up.

Who calibrates the signal

At the start there was someone on the train with a sensor in the back of the neck they cannot switch off, and the whole time it stayed open whether this someone is only the recipient of the signals or also the one who sends them. For the recipient there is no good news, because he cannot think his way out of his own sensor, and every piece of advice that he should try anyway mistakes how a human being is built. The good news exists only for the other one, for whoever builds the communication and incentive environment, for leadership, for HR and change, for marketing and format development, because with him lies the only dial that actually moves anything, and that is the real work on an organization's action infrastructure.

The bias is fine, it works exactly the way it was calibrated. The only question is who calibrates the signal it responds to.