In cognitive psychology and decision theory, loss aversion refers to people's tendency to prefer avoiding losses to acquiring equivalent gains: it is better to not lose $5 than to find $5. The principle is very prominent in the domain of economics. What distinguishes loss aversion from risk aversion is that the utility of a monetary payoff depends on what was previously experienced or was expected to happen. Some studies have suggested that losses are twice as powerful, psychologically, as gains. Loss aversion was first identified by Amos Tversky and Daniel Kahneman.
Loss aversion implies that one who loses $100 will lose more satisfaction than another person will gain satisfaction from a $100 windfall. In marketing, the use of trial periods and rebates tries to take advantage of the buyer's tendency to value the good more after the buyer incorporates it in the status quo. In past behavioral economics studies, users participate up until the threat of loss equals any incurred gains. Recent methods established by Botond K?szegi and Matthew Rabin in experimental economics illustrates the role of expectation, wherein an individual's belief about an outcome can create an instance of loss aversion, whether or not a tangible change of state has occurred.
Note that whether a transaction is framed as a loss or as a gain is very important to this calculation: would you rather get a $5 discount, or avoid a $5 surcharge? The same change in price framed differently has a significant effect on consumer behavior. Though traditional economists consider this "endowment effect" and all other effects of loss aversion to be completely irrational, that is why it is so important to the fields of marketing and behavioral finance. The effect of loss aversion in a marketing setting was demonstrated in a study of consumer reaction to price changes to insurance policies. The study found price increases had twice the effect on customer switching, compared to price decreases. Similarly, users in behavioral and experimental economics studies decided to cease participation in iterative money-making games when the threat of loss was close to the expenditure of effort, even when the user stood to further their gains.
Humans may be hardwired to be loss averse due to asymmetric evolutionary pressure on losses and gains: for an organism operating close to the edge of survival, the loss of a day's food could cause death, whereas the gain of an extra day's food would not cause an extra day of life (unless the food could be easily and effectively stored).
Loss aversion was first proposed as an explanation for the endowment effect--the fact that people place a higher value on a good that they own than on an identical good that they do not own--by Kahneman, Knetsch, and Thaler (1990). Loss aversion and the endowment effect lead to a violation of the Coase theorem--that "the allocation of resources will be independent of the assignment of property rights when costless trades are possible" (p. 1326).
In several studies, the authors demonstrated that the endowment effect could be explained by loss aversion but not five alternatives: (1) transaction costs, (2) misunderstandings, (3) habitual bargaining behaviors, (4) income effects, or (5) trophy effects. In each experiment half of the subjects were randomly assigned a good and asked for the minimum amount they would be willing to sell it for while the other half of the subjects were given nothing and asked for the maximum amount they would be willing to spend to buy the good. Since the value of the good is fixed and individual valuation of the good varies from this fixed value only due to sampling variation, the supply and demand curves should be perfect mirrors of each other and thus half the goods should be traded. The authors also ruled out the explanation that lack of experience with trading would lead to the endowment effect by conducting repeated markets.
The first two alternative explanations--that under-trading was due to transaction costs or misunderstanding--were tested by comparing goods markets to induced-value markets under the same rules. If it was possible to trade to the optimal level in induced value markets, under the same rules, there should be no difference in goods markets.
The results showed drastic differences between induced-value markets and goods markets. The median prices of buyers and sellers in induced-value markets matched almost every time leading to near perfect market efficiency, but goods markets sellers had much higher selling prices than buyers' buying prices. This effect was consistent over trials, indicating that this was not due to inexperience with the procedure or the market. Since the transaction cost that could have been due to the procedure was equal in the induced-value and goods markets, transaction costs were eliminated as an explanation for the endowment effect.
The third alternative explanation was that people have habitual bargaining behaviors, such as overstating their minimum selling price or understating their maximum bargaining price, that may spill over from strategic interactions where these behaviors are useful to the laboratory setting where they are sub-optimal. An experiment was conducted to address this by having the clearing prices selected at random. Buyers who indicated a willingness-to-pay higher than the randomly drawn price got the good, and vice versa for those who indicated a lower WTP. Likewise, sellers who indicated a lower willingness-to-accept than the randomly drawn price sold the good and vice versa. This incentive compatible value elicitation method did not eliminate the endowment effect but did rule out habitual bargaining behavior as an alternative explanation.
Income effects were ruled out by giving one third of the participants mugs, one third chocolates, and one third neither mug nor chocolate. They were then given the option of trading the mug for the chocolate or vice versa and those with neither were asked to merely choose between mug and chocolate. Thus, wealth effects were controlled for those groups who received mugs and chocolate. The results showed that 86% of those starting with mugs chose mugs, 10% of those starting with chocolates chose mugs, and 56% of those with nothing chose mugs. This ruled out income effects as an explanation for the endowment effect. Also, since all participants in the group had the same good, it could not be considered a "trophy", eliminating the final alternative explanation.
Thus, the five alternative explanations were eliminated in the following ways:
Recently, studies have questioned the existence of loss aversion. In several studies examining the effect of losses in decision making under risk and uncertainty no loss aversion was found. There are several explanations for these findings: one, is that loss aversion does not exist in small payoff magnitudes; the other, is that the generality of the loss aversion pattern is lower than that thought previously. Finally, losses may have an effect on attention but not on the weighting of outcomes; as suggested, for instance, by the fact that losses lead to more autonomic arousal than gains even in the absence of loss aversion. This latter effect is sometimes known as Loss Attention. Even in a non-risky domain, prospective affective judgments about gains and losses show that loss aversion is magnitude dependent such that for low magnitudes there is no loss aversion 
Loss aversion may be more salient when people compete. Gill and Prowse (2012) provide experimental evidence that people are loss averse around reference points given by their expectations in a competitive environment with real effort.
Loss aversion and the endowment effect are often confused. Gal (2006) argued that the endowment effect, previously attributed to loss aversion, is more parsimoniously explained by inertia than by a loss/gain asymmetry.
In 2005, experiments were conducted on the ability of capuchin monkeys to use money. After several months of training, the monkeys began showing behavior considered to reflect understanding of the concept of a medium of exchange. They exhibited the same propensity to avoid perceived losses demonstrated by human subjects and investors. While a subsequent study suggested that the 2005 results were not indicative of loss aversion because of timing differences in the presentation of gains and losses to the monkeys, a follow-up 2008 study by Laksminaryanan, Chen and Santos ruled out this alternative explanation.
Expectation-based loss aversion is a phenomenon in behavioral economics. When the expectations of an individual fail to match reality, they lose an amount of utility from the lack of experiencing fulfillment of these expectations. Analytical framework by Botond K?szegi and Matthew Rabin provides a methodology through which such behavior can be classified and even predicted. An individual's most recent expectations influences loss aversion in outcomes outside the status quo; a shopper intending to buy a pair of shoes on sale experiences loss aversion when the pair she had intended to buy is no longer available.
Subsequent research performed by Johannes Abeler, Armin Falk, Lorenz Goette, and David Huffman in conjunction with the Institute of Labor Economics used the framework of K?szegi and Rabin to prove that people experience expectation-based loss aversion at multiple thresholds. The study evinced that reference points of people causes a tendency to avoid expectations going unmet. Participants were asked to participate in an iterative money-making task given the possibilities that they would receive either an accumulated sum for each round of "work", or a predetermined amount of money. With a 50% chance of receiving the "fair" compensation, participants were more likely to quit the experiment as this amount approached the fixed payment. They chose to stop when the values were equal as no matter which random result they received, their expectations would be matched. Participants were reluctant to work for more than the fixed payment as there was an equal chance their expected compensation would not be met.
Loss aversion experimentation has most recently been applied within an educational setting in an effort to improve achievement within the U.S. Recent results from Program for International Student Assessment (PISA) 2009 ranked the US ranks #31 in math and #17 in Reading. In this latest experiment, Fryer et al. posits framing merit pay in terms of a loss in order to be most effective. This study was performed in the city of Chicago Heights within nine K-8 urban schools, which included 3,200 students. 150 out of 160 eligible teachers participated and were assigned to one of four treatment groups or a control group. Teachers in the incentive groups received rewards based on their students' end of the year performance on the ThinkLink Predictive Assessment and K-2 students took the Iowa Test of Basic Skills (ITBS) in March. The control group followed the traditional merit pay process of receiving "bonus pay" at the end of the year based on student performance on standardized exams. However, the experimental groups received a lump sum given at beginning of the year, that would have to be paid back. The bonus was equivalent to approximately 8% of the average teacher salary in Chicago Heights, approximately $8,000.
Methodology--"Gain" and "Loss" teachers received identical net payments for a given level of performance. The only difference is the timing and framing of the rewards. An advance on the payment and the re framing of the incentive as avoidance of a loss, the researchers observed treatment effects in excess of 0.20 and some as high as 0.398 standard deviations. According to the authors, 'this suggests that there may be significant potential for exploiting loss aversion in the pursuit of both optimal public policy and the pursuit of profits'.
Utilizing loss aversion, specifically within the realm of education, has gotten much notoriety in blogs and mainstream media.
The Washington Post discussed merit pay in a 2012 article and specifically the study conducted by Fryer et al. The article discusses the positive results of the experiment and estimates the testing gains of those of the "loss" group are associated with an increase in lifetime earnings of between $37,180 and $77,740. They also comment on the fact that it didn't matter much whether the pay was tied to the performance of a given teacher or to the team to which that teacher was assigned. They state that "a merit pay regime need not pit teachers in a given school against each other to get results".
Science Daily specifically covers the Fryer study stating that the study showed that "students gained as much as a 10 percentile increase in their scores compared to students with similar backgrounds -- if their teacher received a bonus at the beginning of the year, with conditions attached." It also explains how there was no gain for students when teachers were offered the bonus at the end of the school year.
Thomas Amadio, superintendent of Chicago Heights Elementary School District 170, where the experiment was conducted, is quoted in this article stating "the study shows the value of merit pay as an encouragement for better teacher performance".
Education weekly also weighs in and discusses utilizing loss aversion within education, specifically merit pay. The article states there are "few noteworthy limitations to the study, particularly relative to scope and sample size; further, the outcome measure was a 'low-stakes' diagnostic assessment, not the state test--it's unclear if findings would look the same if the test was used for accountability purposes. Still Fryer et al. have added an interesting tumbling element to the merit-pay routine".
The Sun Times interviewed John List, Chairman of the University of Chicagos' department of economics. He stated "It's a deeply ingrained behavioral trait. .. that all human beings have--this underlying phenomenon that 'I really, really dislike losses, and I will do all I can to avoid losing something'." The article also speaks to only one other study to enhance performance in a work environment. The only prior field study of a "loss aversion" payment plan, they said, "occurred in Nanjing, China, where it improved productivity among factory workers who made and inspected DVD players and other consumer electronics". The article also covers a reaction by Barnett Berry, president of the Center for Teaching Quality, who stated "the study seems to suggest that districts pay 'teachers working with children and adolescents' in the same way 'Chinese factory workers' were paid for 'producing widgets'. I think this suggests a dire lack of understanding of the complexities of teaching."
There has also been other criticism of the notion of loss aversion as an explanation of greater effects. Indeed, all of the noted findings in education can be explained simply by the additional attention to a task when it includes losses (i.e., loss attention), independently of the weighting to gains and losses.
Larry Ferlazzo in his blog questioned what kind of positive classroom culture a "loss aversion" strategy would create with students, and what kind of effect a similar plan with teachers would have on school culture. He states that "the usual kind of teacher merit pay is bad enough, but a threatened 'take-away' strategy might even be more offensive".
In earlier studies, both bidirectional mesolimbic responses of activation for gains and deactivation for losses(or vica versa) and gain or loss-specific responses have been seen. While reward anticipation is associated with ventral striatum activation, negative outcome anticipation engages the amygdala. However, only some studies have shown involvement of amygdala during negative outcome anticipation but not others which has led to some inconsistencies. It has later been proven that inconsistencies may only have been due to methodological issues including the utilisation of different tasks and stimuli, coupled with ranges of potential gains or losses sampled from either payoff matrices rather than parametric designs, and most of the data are reported in groups, therefore ignoring the variability amongst individuals. Thus later studies rather than focusing on subjects in groups, focus more on individual differences in the neural bases by jointly looking at behavioural analyses and neuroimaging.
Neuroimaging studies on loss aversion involves measuring brain activity with functional magnetic resonance imaging (fMRI) to investigate whether individual variability in loss aversion were reflected in differences in brain activity through bidirectional or gain or loss specific responses, as well as multivariate source-based morphometry (SBM) to investigate a structural network of loss aversion and univariate voxel-basedmorphometry (VBM) to identify specific functional regions within this network.
Brain activity in a right ventral striatum cluster increases particularly when anticipating gains. This involves the ventral caudate nucleus, pallidum, putamen, bilateral orbitofrontal cortex, superior frontal and middle gyri, posterior cingulate cortex, dorsal anterior cingulate cortex, and parts of the dorsomedial thalamus connecting to temporal and prefrontal cortex. There is a significant correlation between degree of loss aversion and strength of activity in both the frontomedial cortex and the ventral striatum. This is shown by the slope of brain activity deactivation for increasing losses being significantly greater than the slope of activation for increasing gains in the appetitive system involving the ventral striatum in the network of reward-based behavioural learning. On the other hand, when anticipating loss, the central and basal nuclei of amygdala, right posterior insula extending into the supramarginal gyrus mediate the output to other structures involved in the expression of fear and anxiety, such as the right parietal operculum and supramarginal gyrus. Consistent with gain anticipation, the slope of the activation for increasing losses was significantly greater than the slope of the deactivation for increasing gains.
Multiple neural mechanisms are recruited while making choices, showing functional and structural individual variability. Biased anticipation of negative outcomes leading to loss aversion involves specific somatosensory and limbic structures. fMRI test measuring neural responses in striatal, limbic and somatosensory brain regions help track individual differences in loss aversion. Its limbic component involved the amygdala(associated with negative emotion and plays a role in the expression of fear) and putamen in the right hemisphere. The somatosensory component included the middle cingulate cortex, as well as the posterior insula and rolandic operculum bilaterally. The latter cluster partially overlaps with the right hemispheric one displaying the loss-oriented bidirectional response previously described, but, unlike that region, it mostly involved the posterior insula bilaterally. All these structures play a critical role in detecting threats and prepare the organism for appropriate action, with the connections between amygdala nuclei and the striatum controlling the avoidance of aversive events. There are functional differences between the right and left amygdala. Overall, the role of amygdala in loss anticipation suggested that loss aversion may reflect a Pavlovian conditioned approach-avoidance response. Hence, there is a direct link between individual differences in the structural properties of this network and the actual consequences of its associated behavioral defense responses.
The neural activity involved in the processing of aversive experience and stimuli is not just a result of a temporary fearful overreaction prompted by choice-related information, but rather a stable component of one's own preference function, reflecting a specific pattern of neural activity encoded in the functional and structural construction of a limbic-somatosensory neural system anticipating heightened aversive state of the brain. Even when no choice is required, individual differences in the intrinsic responsiveness of this interoceptive system reflect the impact of anticipated negative effects on evaluative processes, leading preference for avoiding losses rather than acquiring greater but riskier gains.
Individual differences in loss aversion are related to variables such as age, gender, and genetic factors affecting thalamic norepinephrine transmission, as well as neural structure and activities. Outcome anticipation and ensuing loss aversion involve multiple neural systems, showing functional and structural individual variability directly related to the actual outcomes of choices.
In a study, adolescents and adults are found to be similarly loss-averse on behavioural level but they demonstrated different underlying neural responses to the process of rejecting gambles. Though adolescents rejected the same proportion of trials as adults, adolescents displayed greater caudate and frontal pole activation than adults to achieve this. These findings suggest a difference in neural development during the avoidance of risk. It is possible that adding affectively arousing factors (e.g. peer influences) may overwhelm the reward-sensitive regions of the adolescent decision making system leading to risk-seeking behaviour. On the other hand, although men and women did not differ on their behavioural task performance, men showed greater neural activation than women in various areas during the task. Loss of striatal dopamine neurons is associated with reduced risk-taking behaviour. Acute administration of D2 dopamine agonists may cause an increase in risky choices in humans. This suggests dopamine acting on stratum and possibly other mesolimbic structures can modulate loss aversion by reducing loss prediction signalling.