Emotions on the trading floor

40Several studies have investigated the physiological responses to market events and the role played by emotions for financial professionals. For instance, Lo et al. (2005) showed that professional traders exhibit emotional reactions to monetary gains and losses. Using professional traders in their own environment as the subjects of studies is a tricky task that limits the research materials to physiological measures such as heart rate and heart rate variability (HRV). For that purpose, Lo and Repin (2002) analysed the emotional decision-making process on ten professional traders, taking biofeedback measures such as heart rate, skin conductance and blood pressure as proxies for these emotions. They indicated that traders exhibit greater emotional arousal around important events such as volatility peaks. Furthermore, experienced traders feel these emotional arousals to a lesser extent than less experienced ones. The authors argued that making quick decisions based on their emotional arousal is a necessary condition in traders’ decision-making.

41This research highlights the emotional responses of traders with respect to market events and explains that experienced traders are less sensitive to emotional cues. However, this paper does not examine how professional traders learn to regulate their emotions in the span of their careers (i.e. emotion regulation is part of their expertise). Several papers have shown the usability of HRV as a good proxy for intentional emotion regulation (León et al., 2009; Denson et al., 2011).

42A physiological research with investment bank traders examined emotion regulation and traders’ experiences using HRV (Fenton-O’Creevy et al., 2012). The results showed that volatile market events are correlated with lower HRV while greater experience is correlated with higher HRV. Hence, the authors evidenced traders can hardly control their emotions during volatile market conditions and thus the importance of emotion regulation in traders’ expertise.

Stress and financial decisions

43The previous subsection examined how a stressful environment, such as the trading pit, and stressful decisions, such as market events, can generate emotions (Lo and Repin, 2002; Fenton-O’Creevy et al., 2012). The effects of stress on financial decision-making can now be tackled.

44Porcelli and Delgado (2009) studied the effects of acute stress on financial decision-making. In their experiment, stress is induced in subjects by immersion of one hand in cold water for several minutes. Non-stressed control subjects follow a similar procedure with hot water. The patients engaged in a gambling game where they faced two alternatives either presented as a gain or as a loss. During the task, skin conductance was measured. The results showed a significant increase in skin conductance levels in the stress condition. Stressed participants made riskier choices. Furthermore, participants under stress made riskier choices in the loss domain compared with non-stressed participants, and less risky choices than non-stressed participants in the win domain. The authors therefore argued that acute stress alters financial decision-making. They tend to use a somewhat automated risk bias. The authors concluded that stress may disrupt resources used by the executive functions of the brain, which may lead to an excess reliance on low level automated systems.

Regret

45In this review, we have underlined the fact that investors are far from behaving in a rational manner. On the contrary, they are driven by emotions. Indeed, investors try to limit or avoid possible future regrets by taking suboptimal decisions, even when they possess specific financial knowledge. Coricelli et al. (2007) observed that people with lesions in the orbitofrontal cortex (OFC) do not anticipate the negative consequences of their choices and decide according to actual situations only. Hence, they do not feel regret.

46In a stock market experiment using fMRI, Frydman and Camerer (2016) proposed a test of the implications of regrets. They showed that participants who observe a positive return on a particular stock they do not hold experience regrets. In addition, they highlighted the impossibility for participants to repurchase stocks that had increased in value even if it could be optimal for them, which can be referred to as the repurchasing mistake.

Positive anticipated emotions

47Knutson and Greer (2008) defined anticipatory affect as the emotional state people experience when anticipating significant outcomes. They reviewed the consequences of an anticipatory affect and its consequences over individual choices, and showed changes in brain circuits activation when subjects anticipate monetary incentives. On the one hand, NAcc activation increases when gains are anticipated. On the other hand, the anterior insula activation increases when both losses and gains are anticipated. Moreover, NAcc anticipatory activation is correlated with a self-reported positive arousal, a preference for risky gambles, and precedes the purchase of desirable goods. Conversely, the anterior insula is correlated with both positive and negative arousal, the desire for non-overpriced items and the choice for non-risky gambles.

48Knutson et al. (2008) investigated the effect of a positive anticipatory exciting environmental signal (anticipatory affect) on risk-taking behaviour in financial markets. They conducted an fMRI experiment on young student males. Before trials, they were shown suggestive female pictures. The authors found that risk-taking was increased after an activation of the subject’s NAcc via this prior picture. Therefore, the subjects were more likely to participate in lower expected value gambles and to make riskier choices.

49In addition, Wu et al. (2014) showed the affective traits of incentive anticipation of gains and losses using fMRI. The subjects played a monetary incentive delay task. During each trial, the participants began by seeing a cue indicating potential gains or losses of different amounts. They then watched a cross and waited for different intervals and responded with a button to a target appearing for varying time spans to see the outcome. The results established NAcc activation during positive arousal and anticipation of large gains, while the anterior insula was activated during negative arousal and anticipation of large losses.

50Andrade et al. (2016) tested the relationship between excitement and bubbles. Their experiment followed the experimental design of Smith et al. (1988). Before the experiment starts, the experimenter shows participants a video tape to introduce emotions. The emotional states that can be introduced are fear, calm and excitement. The authors noted that bubbles are much larger in the excitement state than the others. The bubble amplitude in the first round, meaning the difference between the average trading price of the asset and its fundamental value, is larger than in the calm and fear treatments. The results suggest that the excitement generated by increasing prices in real stock markets actually fuels bubbles.

51Most studies investigate the relation of emotions and “suboptimal” choices, leading to bad decisions. However, positive emotions can also induce altruism. Indeed, one of the persistent assumptions about human beings is that they are self-interested only. However, sometimes we may share things with people whom we do not know. Genevsky and Knutson (2015) showed that affective mechanisms can influence the success of microloans. Positive affective features of photographs increase the success of those requests in an internet microloans database request. With a small sample and fMRI imaging, the authors demonstrated that NAcc activation and self-reported positive arousal predicted the success of loan requests on the internet.

Positive beliefs

52The brain regions involved in processing risk and rewards are also in charge of processing emotions, thus we can wonder whether emotions affect financial choices and beliefs.

53Emotions play a role in the belief formation process of investors. Brunnermeier and Parker (2005) established that agents forming optimistic beliefs about their future outcomes are happier. This belief has an immediate impact on their subjective well-being via anticipatory emotions and encourages them to prefer skewed assets. In a physiological experiment, Baechler et al. (2018) use two equal lotteries except for their skewness. The first lottery has a positive skewness and the second one has a null skewness. The authors point out that subjects participating in the non-skewed lottery exhibit greater self-regulation than other subjects. Hence, they are less prone to optimism. Furthermore, they confirm that all participants feel positive anticipatory emotions during the waiting interval between the revelation of the lottery and the draw.

54Another paper (Kuhnen and Knutson, 2011) proved the emotional impact of both risk-taking behaviour and beliefs formation in individuals, using an experimental design where subjects need to update their beliefs about the distribution of a risky asset. The authors found that events associated with positive arousal led to riskier choices while those associated with negative arousal led to risk-averse choices. They also showed that affect transforms the participants’ belief formation process. Positive affects increase subjects’ confidence in their capacities to assess risky choices. Beliefs are updated in a way that preserve a positive state of mind, leading subjects to incorporate only news that does not interfere with their previous choices. Thus, they form incorrect beliefs.

Genetic differences

60Making investment choices is an individual task and may vary across individuals. Some agents are more willing to gamble (Grinblatt and Keloharju, 2009) or to invest in short-term assets. Therefore, under equal conditions, successful financial decisions strongly depend on investors themselves, with very few of them able to “beat the market”. Several studies have investigated the role played by financial knowledge (van Rooij et al., 2012), endowment (Agnew and Szykman, 2005), age (Li et al., 2013) and gender (Barber and Odean, 2001). However, these factors only partially explain risk-taking behaviours. Other factors such as genes may play a role.

61Several papers have studied the link between risk-taking behaviour and genetic components. Using a large sample of twins from Sweden, Cesarini et al. (2010) confirmed that around 25% of variation in portfolio risk is due to genetic components as well as the decision to invest in socially responsible assets.

62Barnea et al. (2010) studied a dataset on identical twins completing financial portfolio decisions. They found a genetic component explaining around 30% of the variation in stock market participation and asset allocation. The authors highlight the non-permanent effect of the family environment. It has a measurable effect on young individuals’ behaviour but does not last as the individuals gain experience. They argue that twins who grow up in similar environments as well as twins growing up in different environments exhibit the same investment behaviour beyond a common genetic component.

63These papers show the role played by genetic components in risk-taking behaviours. However, they fail to explain the specific genes at stake. Other research has focused on the specific gene configurations that affect risk-taking by modulating dopaminergic and serotonergic circuitries involved in decision-making processes. All these systems are known to modulate decision-making such as pathological gambling (Ibanez et al., 2003). Two genes in particular are involved in financial risk-taking: the dopamine receptor D4 gene (DRD4) and the serotonin transporter polymorphism (5-HTTLPR), which have previously been linked with emotional behaviour, anxiety and addiction (Kuhnen and Chiao, 2009).

64Kuhnen et al. (2013) also investigated neuroticism through the polymorphism in the promoter region of the serotonin transporter gene (5-HTTLPR). They found that carriers of short alleles invest less in equities, make less active financial decisions and have fewer credit lines. In addition, short allele carriers perceive stocks as riskier and tend to feel more negative emotions when faced with risky investment choices, the authors attribute this to neuroticism.

65Kuhnen and Chiao (2009) found that carriers of the DRD4 7-repeat allele take 25% more risks than non-carriers. On the other hand, Sapra et al. (2012) disclosed the role of dopamine in risk-taking by examining 60 professional traders. They showed that successful traders are carriers of COMT (catecholamine-O-methyltransferase) and DRD4P (dopamine receptor 4 promoter), gene alleles that moderate dopamine and hence they take more appropriate risks. Inversely, carriers of short 5-HTTLPR allele genes (which modulate serotonergic activity) take 28% less risks than others (Kuhnen and Chiao, 2009).

66Frydman et al. (2011) tested the effects of several genes on financial risk-taking behaviour with a simple gambling game. They spotted that people with the MAOA-L gene are more likely to take financial risks compared to MAOA-H carriers but only when they perceive it as advantageous. MAOA-L carriers exhibit higher connectivity between the prefrontal cortex and the amygdala and the gene is known to contribute to aggressive and impulsive behaviour. Hence, they are more willing to take risks.

67In a seminal paper, Cronqvist and Siegel (2014) studied the genetic foundation of several investment biases such as lack of diversification, home bias, turnover, the disposition effect, chasing for performance and the preference for skewness. They found higher correlation for the biases under their scope for identical twins compared with fraternal twins: the correlation for identical twins is about twice the correlation between fraternal ones. For the latter, correlations are higher for same sex fraternal twins. The authors also pointed out that some moderators exist to investment biases such as having work experience and knowledge in finance.

Hormone differences

68Several studies have established the role played by hormones in financial decision-making, especially testosterone [4] and cortisol [5]. Coates and Herbert (2008) analysed the “winning effect” of testosterone with seventeen professional traders in London, taking measures for testosterone and cortisol in the morning and the afternoon. On days when morning testosterone levels were high, traders experienced afternoon profits higher than on days when these levels were lower. Moreover, they showed that cortisol levels increase with portfolio variances and market volatility. This result is consistent with previous knowledge. Indeed, cortisol is known to influence brain regions linked with irrational financial decisions. Therefore, it fluctuates with risks and returns, and it may alter a trader’s ability to make optimal decisions.

69In experimental asset markets with induced testosterone, Nadler et al. (2017) showed that testosterone has an effect on both trading behaviour and price bubbles. It generates high bids and the slow incorporation of the asset fundamental value, causing longer lasting bubbles. These results are in line with Eckel and Füllbrunn (2015) among others, who analysed gender differences in price bubbles. They found that women generate less bubbles than men, and overall trade less than men (Barber and Odean, 2001).

70In another study, Cueva et al. (2015) explored the links between financial risk-taking, testosterone and cortisol using the experimental market design of Smith et al. (1988). They found that higher cortisol levels in aggregate and in individuals predict risk-taking and price volatility. In addition, testosterone increases optimism about changes in future prices. The authors then administered cortisol or testosterone to male subjects participating in an experimental game. In this experiment, subjects were shown plots about past prices of two stocks. Then, participants were asked to choose between two stocks either with high or low variance of returns. Subjects who were administered cortisol or testosterone were either way more likely to take the riskier choice, demonstrating that both hormones increase risk-taking and price destabilisation with market bubbles followed by a crash. Furthermore, testosterone administration resulted in increased optimism regarding future stock prices.

Anatomic differences

71The results about the second-to-fourth digit ratio are mixed. Apicella et al. (2015) pointed out that the field, which appeared ten years ago, is still emerging, despite the non-invasive character of this proxy to hormones measurement. However, several studies have provided positive results about risk-taking and 2D:4D. Coates et al. (2009) used 2D:4D as a predictor of future financial success among traders, and 2D:4D has been proved as a good predictor of future success in highly competitive sports already. The authors mentioned 2D:4D as a sign of long-term profitability and the number of years traders will remain in the business.

72Stenstrom et al. (2011) tested the impact of testosterone on risk-taking. They used the second-to-fourth-digit-ratio (2D:4D) and the length of the second finger relative to the sum of the lengths of all four fingers (rel2) as a proxy of prenatal exposure to testosterone across five topics: financial, recreational, social, ethical and health-related risk-taking behaviours. They found that lower rel2 is predictive of greater risk-taking in the financial, social and recreational domains and lower 2D:4D is predictive of greater risk-taking in the social and recreational domains.

73Cronqvist et al. (2016) examined the link between several prenatal environmental differences and the heterogeneity of financial decisions one may take later in life. In this purpose, they investigated birth weight and prenatal exposure to testosterone. For prenatal testosterone, the authors compared twins of opposite sex pairs with twins of the same sex pair. For women who grew up with a male twin, they noticed an increase in prenatal testosterone exposure which leads to a “masculinisation” of financial behaviour, meaning a high risk-taking and trading in adulthood. In addition, the study mentions that a higher birth weight is related to a higher participation in stock markets, while a lower birth weight is correlated with the tendency to prefer skewed assets and to choose portfolios with higher volatility.

Are all investors risk-averse?

74The neural basis of risk aversion was sparsely known until the rise of neuroimaging. Some studies using structural MRI show the relation between the grey matter (central tissue of the nervous system) volume of different brain areas managing risk-taking and investor differences in risk-taking decisions. In two voxel-based morphometry studies, subjects exhibiting greater risk aversion had altered grey matter volume (Nasiriavanaki et al., 2015), a lower grey matter volume in the posterior parietal cortex, a lower grey matter volume in the anterior insula (Gilaie-Dotan et al., 2014) compared with risk-seeking participants.

75Furthermore, risk-averse subjects have different brain activation during each stage of their financial decisions, anticipation of risks and rewards, choosing between different risky options and processing the outcome of a risky choice. During the anticipation phase, risk-averse subjects appear to overestimate risks. They exhibit a higher activation of the ventral striatum and the anterior insula (Rudorf et al., 2012), which are regions involved in risk processing (Preuschoff et al., 2006, 2008). While the NAcc is directly connected with the insula, this connection appears to be distorted in investors having a preference for skewed assets and gambles (Leong et al., 2016).

76According to Rudorf et al.(2012), risk-averse people still overestimate risks. They are unable to assess their expectation of risk whereas the results of the gamble are known to be less risky than they would have expected: a reduced risk prediction error (the tuning of the estimated risk of a hazardous choice when the outcome is known) is altered in their insular cortex.

Participation in financial bubbles

77Some studies suggest that some investors may trigger financial bubbles by trying to infer the investment intentions of other market participants. In an fMRI experiment De Martino et al. (2013) pointed out an increase in prefrontal cortex activity in subjects exhibiting a tendency to participate in bubbles (region involved in inferring intentions of others): they “ride bubbles”, taking into account the intentions of other players.

78Smith et al. (2014) found a positive correlation between aggregate NAcc activity and changes in prices during bubbles. In this experimental market, prices were entirely determined by twenty participants trading between each other while two subjects were scanned. The experiment was based on one risky asset (a stock) and a non-risky one (numeraire). One of the main findings of this study is that the evolution of the risky stock price follows the same patterns as the NAcc activation in all participants. Also, the peak of the bubble corresponds exactly to the peak in NAcc activity. Thus, NAcc activity represents risk streaks and the erroneous idea that prices will keep rising. “Smart traders” [6] however, exhibit greater anterior insula activation, a brain area known to be related with uncertainty (in this case the uncertainty about the continuous increase in prices). Traders with a superior anterior insula activation sell a few moments before the peak of the bubble, when traders with greater NAcc activation continue to buy. Hence, according to the authors, when they start selling stocks, smart traders use their intuition [7] rather than statistical signals.

79In an experiment, Efremidze et al. (2017) showed the role played by reinforcement learning in financial bubbles. For this purpose, the authors administered a drug known to inhibit learning to several participants. Compared with the control group, asset prices were 60% higher in the drugged participants.