1. Introduction
After more than one year into the COVID-19 pandemic, experts and policymakers across the world are expressing increasing concern about the psychological and emotional toll that the spread of the coronavirus disease and the consequent restrictions to mobility have imposed on populations in the most heavily affected countries. Several reports published since the pandemic outbreak in March 2020 have signaled an alarming increase in rates of diagnosed mental disorders such as depression, anxiety, and Post-Traumatic Stress Disorder in response to health-related concerns, prolonged lack of sociality and high uncertainty about the future.
The Pew Research Center reported fear, panic, anger, sadness, health anxiety, and economic/financial anxiety as the most highly prevailing emotions [
1]. Not only, in fact, the rapid surge of the pandemic has caused preoccupation about the disease’s short- and long-term health consequences, especially among the elderly, but the resulting economic downturn accompanied by the massive loss of jobs and the social isolation caused by lockdowns, mandatory mask wearing, and prolonged interruption of many leisure activities has triggered a whole spectrum of negative emotions associated with loneliness, economic uncertainty and financial insecurity The Pew Research Center reported evidence concerning the US population according to which at least one-third of Americans experienced high levels of psychological distress during the coronavirus outbreak. Overall, 36% of Americans said the coronavirus outbreak represented a major threat to day-to-day life in their community, according to a related survey [
1]. Almost one year later, another survey found that 49% of adults who are unemployed and looking for a new job report to be pessimistic about their possibility to find one in the near future [
2].
As far as the US situation is concerned, about half of non-retired adults (51%) say the economic impact of the coronavirus outbreak will make achieving their long-term financial goals harder [
3]. The situation in developing countries is even worse, given that, according to a recent WHO estimate [
4]. The Covid-19 pandemic is still causing the disruption of essential health services in 90% of the countries after one year. The crisis, in a nutshell, is having far-reaching and long-lasting implications for almost all aspects of the life of millions of people worldwide, and the increasing awareness among the general public of the magnitude and duration of its many effects is producing widespread emotional distress.
It is well known that our affective states, both positive and negative, have important behavioral consequences, strongly influencing our cognitive processes and deliberation abilities [
5]. Therefore, in times of turbulence such as the present one, in which our daily habits and routines have been disrupted by radical limitations to sociality and mobility, and the future is highly uncertain, it is plausible to assume that people’s peculiar emotional reactions, besides affecting their well-being, might also significantly affect their decision-making outcomes and processes in a significant way. In this study we focus specifically on how negative emotions may affect two variables that influence our decision-making activity by specifically impacting the extent to which we are willing to sacrifice present gratifications for future ones, i.e., our “time preferences” (or temporal discounting), and the extent to which we are willing to take risks, i.e., our “risk attitudes” (or probability discounting).
The importance of studying temporal and probability discounting comes from the observation that both variables have a remarkable influence on decisions involving important aspects of our lives (e.g., savings [
6], consumer decisions [
7,
8], employment [
9,
10], educational investment [
11], energy conservation [
12], and financial decisions [
13]), many of which bear long term consequences, for example, suboptimal behaviors (e.g., health behaviors [
14,
15], obesity [
16], smoke [
17,
18]) and clinical conditions (e.g., drug [
19,
20] and alcohol [
21,
22] abuse, and gambling disorder [
23,
24]).
We chose to test the impact of the entire range of the primary negative emotions (Primary or basic emotions are hypothesized to be a special class of emotions out of which all other emotions are compounded. According to most theorists, they are innate, universal, and distinct affective states which evolved to serve adaptive functions [
4,
25]) as they were all, to some extent, experienced by a relevant portion of the population during the pandemic. Further, we included an additional negative emotion to this spectrum, moral disgust: although moral disgust does not belong to the core set of primary emotions, we believe it is important to include it in our study, as many people may experience such emotion toward members of their community refusing to contribute to ending the pandemic (by, for example, refusing to wear a mask and/or refusing to get vaccinated).
In a nutshell, the results of our study showed that the elicitation of all emotions significantly made subjects more “impatient” and more risk-prone in comparison to a “neutral emotion” condition taken as baseline. In both tasks, all negative emotions were associated with the same direction of the effect, although differences were observed in their magnitude, especially for sadness. These results are at odds with the motivation appraisal theory [
26,
27] therefore we propose alternative explanations based on self-regulation theories.
These findings contribute to the literature on emotions and their effect on high-level cognitive functions, by directly comparing a wide range of negative emotions and partially explaining the rather inconsistent results observed in the previous literature.
Furthermore, they also contribute to a broader literature on economic decision making—which has traditionally been focused on the emotional arousal associated with intertemporal or risky decisions [
28,
29,
30], by showing the effect of prior emotions on decision-making abilities, which are often understudied.
Finally, they might help shed some light on the psychological drivers of many choices that individuals made during and immediately after the pandemic.
The rest of the paper proceeds as follows: the following section defines each target emotion precisely and illustrates previous research on emotions and decision making, with particular attention to intertemporal choices and choices under risk. We also briefly review the current neuroscientific evidence about the main drivers of risk propensity and time discounting. The next section defines the hypothesis of the current study. A method and Results section follows, describing in detail the experimental design and the main results, separately for experiment 1 and experiment 2. The last section offers some concluding remarks and directions for future research.
1.1. Negative Emotions and Decision Making
After several months into the COVID-19 pandemic, many of us are still suffering the consequences of the emotional toll that the spread of the virus brings about. We might have experienced physical disgust due to contamination threats arising from even the simplest daily activities (like going to the supermarket), moral disgust because we have been (and still are, depending on the peculiar situation of each country and region) constantly exposed to behaviors that put us at risk, such as violating social distancing, refusing to wear a mask, or refusing to vaccinate; anger because of restrictions that limit our sociality or, alternatively, lack of restrictions that expose us to the risk of infection, fear of contracting the disease and facing its immediate and long term consequences, and sadness because of isolation, loneliness, and, in the worst cases, loss of dear ones. Finally, in a non-negligible portion of the population, we must add the fear of vaccines and their potential side effects.
In the present research, we aim to compare whether and to what extent the above-mentioned negative emotions might affect our time and risk preferences in decision making compared to a control condition. Hence, we provide a short review for each emotion with particular attention to their possible effect on risk and time discounting.
Disgust is defined as a feeling of revulsion, sometimes accompanied by nausea, along with a strong desire to withdraw from the eliciting stimulus [
31]. From an evolutionary perspective, disgust may function primarily with the objective of rejecting offensive food [
32,
33,
34] or, more generically, to avoid contact with substances that transmit pathogens [
35]. The psychological and economic literature focused on the relationship between physical disgust and risk sensitivity: for example, it has been found a positive association between disgust and risk-aversion [
36], similarly, another study demonstrated that disgust sensitivity is correlated with heightened risk perception [
37]. More recently, a study found that disgust only dampens participants’ propensity to take the gain, but not the loss [
38]. With respect to intertemporal decision making, the effects of disgust are understudied and mostly unknown. To the best of our knowledge, the only study investigating the effect of disgust on temporal discounting indicated that disgust sensitivity predicted a preference for larger later rewards [
39].
Moreover, moral transgressions can trigger disgust responses [
40,
41,
42,
43], implying that situations involving a social threat may equally lead to efforts to protect oneself from more indirect potential sources of harm or to preserve social order [
35]. Hence, disgust elicited by abstract socio-moral transgressions can be defined as moral disgust [
44]. To the best of the authors’ knowledge, there is no previous research that systematically tested the relationship between moral disgust and risk-taking or intertemporal decision making. However, moral transgressions can elicit physical disgust responses, and therefore rejections or avoidance responses [
45]: hence, it is plausible to expect that morally disgusted individuals will be more risk avoidant and more inclined to choose larger-later rewards. Conversely, other studies juxtapose moral disgust to anger, identifying differences between elicitors of disgust and anger in moral contexts, with disgust responding more to bodily-moral violations such as incest, and anger responding more to socio-moral violations such as theft [
46]. However, the research so far does not allow to sharply distinguish between moral disgust, physical disgust, and anger. As a consequence, related predictions about the relationship between moral disgust and decision-making are difficult.
Anger is generally defined as a feeling of displeasure and hostility that arises after being exposed to any form of injustice or offense [
47]. Unlike other negative emotions, anger has been associated with motivational approach tendency [
48] which provides the impulse to “go toward” (i.e., approaching behavior; [
49]), as opposed to the avoidance system, which provides the motivational impulse to “go away”. The link between trait anger and approach motivation is particularly relevant in risk-taking. In fact, angry people tend to have optimistic risk assessments [
50] whereas other studies found that loss aversion is negatively influenced by anger [
51].
The effects of anger on intertemporal decisions are not well investigated, as only a few studies can offer some indications of the possible relationship between anger and temporal discounting. Specifically, one study has investigated the interaction between state and trait anger, showing that people low on trait anger preferred large delayed rewards over smaller-immediate ones when in an angry mood, while people with high trait anger, when in an angry mood, showed an inverse preference (i.e., preferring small-immediate reward over a large delayed one; [
52,
53]). Another study, conducted on adolescents with conduct disorder showed a significant relationship between irritability and an increased preference toward immediate rewards [
54].
Fear can be defined as the feeling of agitation and anxiety caused by perceptions of uncertainty in the environment, or by the presence of a physical danger [
27]. Fear responses are coherent with its theorized evolutionary function of avoiding/escaping danger [
55], as the main objective of fear is defending an organism from an imminent threat [
56,
57,
58]. Both precautionary and self-protective motivations [
59,
60] characterizing fear are well documented and established in the risk-taking domain: in fact, fear was consistently found to trigger pessimistic judgments and risk-averse behaviors [
61,
62,
63,
64,
65,
66,
67]. However, according to the evidence of increased risk-taking as a result of excitement [
62,
68], it was argued that depending on the characteristics of the task, fear can be reinterpreted as a state of excitement and consequently promote, rather than discourage, risk-taking [
69]. Hence, the role of fear in risky choices is still open to debate. No study, to the best of our knowledge, has investigated the effect of fear on intertemporal decision making.
Sadness is a typical response to loss [
26] that predicts disengagement from goals that can no longer be attended to [
70,
71], and it is characterized by avoidance motivation [
72]. Sadness may also sometimes be described as a psychological pain accompanied by additional feelings of loneliness, distress, depression, anxiety, grief [
72,
73,
74], and anguish [
75]. Previous research has compared anger and sadness showing that sadness attenuates the unrealistic illusion of control typical of anger [
50,
76], generating higher risk aversion [
77,
78]. However, it was found that when the typical rumination response of sadness is prevented, subjects, although still sad, are less risk averse [
79]. Hence, the evidence is still largely inconclusive.
A study conducted on a population of cigarette smokers has shown that, with respect to intertemporal decision making, sadness has the effect of increasing impatient choices for cigarette puffs, as a measure of delay discounting preference [
80]. Such an effect appeared to be mediated by an increased self-focus of attention which, in turn, is induced by the feeling of sadness [
81]. Along similar lines, it has been suggested that sadness increases impatience and creates a myopic focus on obtaining money immediately instead of later, a phenomenon known as myopic misery [
82].
1.2. Hypothesis
Based on the literature reviewed above, our hypothesis is that all target emotions affect temporal discounting and risk propensity significantly with respect to the neutral condition. As for the direction of the effect, in accordance with the literature on the motivation appraisal of emotion, we expect the emotions associated with avoidance motivation and withdrawing behavior—such as disgust, fear, and sadness—to be associated with increased risk aversion and preference for delayed gratification, while emotions associated with approaching motivations, such as anger, to increase risk taking and promote impulsivity in intertemporal choices. As for moral disgust, according to the previous literature, it is not entirely clear whether this emotional response falls within the approaching or withdrawing motivational system, thus, precise predictions are difficult to formulate.
However, the approaching/withdrawing motivation theory of emotions, albeit compelling, may not be as straightforward as it seems for some emotions. For example, fear—which is usually described as a withdrawing/avoidance emotion (i.e., flight response)—can actually trigger also responses of freezing or fight, depending on the cognitive appraisal of the situation (e.g., based on the course of action that is more likely to defend the individual from the threat that activated the emotional response of fear). Moreover, there are additional appraisal components that might affect the proposed negative emotions, as for example a feeling of certainty and a sense of personal control over the (threatening) situation [
83]. Based on these lines of reasoning and on the inconsistency of the previous results in the relevant literature, it is possible to observe different effects on decision making, as a result of different appraisals triggered by the target emotions. Nonetheless, we expect that all the emotions will significantly modulate both time and probability discounting.
3. Discussion
Across two experimental studies, we showed that when experiencing anger, sadness, fear, physical disgust, or moral disgust participants discount the future more steeply, i.e., they become significantly more impatient with respect to a neutral condition. This tendency is particularly pronounced in the case of sadness and moral disgust, and less pronounced for fear and anger, while physical disgust is the emotion yielding the less marked effect (Study 1). Secondly, anger, sadness, fear, physical disgust, and moral disgust generate a reduction of the probability discount rates, i.e., make individuals significantly more tolerant toward risk. The effect is more pronounced in the sadness condition and systematically decreases in the fear, physical disgust, anger, and moral disgust conditions (Study 2).
These results reverse the classical motivational appraisal approach according to which emotions associated with an approaching motivational appraisal, such as anger, should generate a higher risk propensity compared to the emotions associated with an avoidant motivational appraisal, as for example sadness, disgust, and fear [
50].
However, when people feel an emotion, especially in the case of unpleasant or unwanted emotions—they often use specific emotion-regulation strategies in order to try to downregulate them [
108]. The authors distinguish between two independent regulation strategies that can be differentially employed by individuals: antecedent-focused strategies versus response-focused strategies. The former are based on
cognitive reappraisal, i.e., the reframing of a situation in order to change its emotional impact [
109]. For example, one might reframe a reproach from one’s boss as an opportunity for personal growth, thus “transforming” a feeling of sadness into one of motivation. In contrast, response-focused strategies (i.e., expressive suppression) are based on the inhibition of the current emotion-expressive behaviors without acting upon the emotion itself [
109], such as when hiding behavioral hints of anxiety during a job interview in order to make a good impression: in the latter case, the emotion is left unaltered, but its behavioral consequences are suppressed. These two strategies were studied by Heilman et al., (2010) in the context of decision making under risk and uncertainty, and the authors found that, especially for negative emotions, response-focused strategies did not impact upon risk-taking behavior, probably because these strategies do not change the underlying emotional experience, but only its behavioral manifestations. Conversely, response-focused strategies had the effect of increasing risky choices; according to the authors, because negative emotions are associated with risk avoidance, when people are able to successfully down-regulate the emotional experience, the effects of the negative emotion on risk-taking behavior are not observed anymore, i.e., people became more risk prone [
110].
Nevertheless, an alternative explanation for these findings is that emotional reappraisal requires cognitive and attentional control, thus, “draining” cognitive resources from attending behavioral regulation in the risk choice paradigm. In other words, if attentional and cognitive resources are devoted to emotion reappraisal, participants may end up shorthanded in exerting control in other aspects of behavior [
111,
112]—such as deliberation and choice—as it is known that risk avoidance and intertemporal bargains require a certain degree of cognitive control [
113]. Thus, a possible explanation of our results in the light of cognitive reappraisal theories may be that when feeling a negative emotion, participants may have tried to downregulate such emotion in order to feel better, thus diminishing the availability of cognitive control resources to devote to behavioral regulation.
A further and even more compelling explanation of our results derives from the observation that people tend spontaneously to engage in self-rewarding activities as a strategy aimed at down-regulating unpleasant emotions [
114]. For example, it has been found that engaging in pleasant, rewarding activities is the most successful strategy for reducing negative affect [
115]. Along these lines, a literature review has shown that in the remediation of negative mood/affect, self-reward represents an obvious and anecdotally frequent response, which also appears to be effective in producing a change or a reduction in the negative state [
116,
117].
According to this view, bad mood and negative emotions represent a frustration of the affective balance of an individual, who may react by using self-reward as a self-regulation strategy. Immediate rewards in intertemporal choice tasks and risky alternatives in probability discounting tasks represent forms of rewarding behavior. Indeed, it has been reliably shown that the preference for risk-taking behavior [
118,
119], as well as the preference for immediate rewards at the expenses of larger-later ones [
96,
120], is related to the activation of the mesolimbic reward system, i.e., a dopaminergic circuit responsible for our feeling of reward and pleasure [
121]. According to this line of reasoning, the frustration of one’s emotional balance induced by a negative emotion should produce a tendency toward risky choices or immediate rewards, as a self-regulation strategy aimed at stimulating the mesolimbic reward system. Thus, the differences observed across the different emotions may be related to the extent to which a certain negative emotion threatens the individual’s emotional balance. Although the extent to which a specific negative emotion is considered as threatening for one’s emotional balance may be subject to individual differences [
122], as well as to cultural expectations [
123], we consistently found sadness to elicit the most pronounced effect across both experimental tasks. Sadness is often associated with a high degree of unacceptability, which is partially socially and culturally shaped. Cultural norms dictate that people are expected to strive for happiness and not to feel sad, as happiness has been enthusiastically promoted as important for personal well-being and a meaningful life [
124]. Hence, people often struggle with the feeling of sadness and may be particularly motivated to down-regulate such emotion. In the context of the current study, this would determine a higher tendency toward the selection of immediate and risky options.
The interpretation of increased preference toward risky and impatient choices as an emotional regulation strategy of self-reward finds additional support also in other lines of literature. For example, studies on eating disorders, and in particular on binge-eating, have shown that dysregulated eating behaviors are used as a mechanism that suppresses or ameliorates the experience of negative affect [
125,
126,
127], especially when the emotional state is experienced as unacceptable or wrong, thus determining the reliance on dysfunctional coping strategies [
128].
Further support comes from some trends that have been observed during the Covid-19 outbreak. According to a study conducted by the Pew Research Center, in the third quarter of 2020, about 28.6 million Baby Boomers (People born between 1946 and 1964.) reported that they were out of the labor force due to retirement and that since February 2020, the number of retired Boomers has increased by about 1.1 million [
129]. Although the mechanisms that led to this trend are yet to be tested, according to many financial advisors, the new attitude of “
life is too short” is spreading among their customers, pushing them to take the plunge. Thus, after months of restrictions, regulations, bad mood, and sacrifices, the prospect of returning to the old pre-COVID working day appears no longer attractive, pushing people toward the prospect of a self-rewarding pre-retirement. The phenomenon could be explained both in the light of the “cognitive appraisal” interpretation by which negative emotions led to a reframing of one’s beliefs (“
a life of continuous work and sacrifice is not worth living”) and in the light of the self-regulation strategy explanation (“
after months of distress I need a reward”).
Similar considerations might apply also to another observed pandemic trend, i.e., the sharp increase in the homeownership rate in the US since the onset of the virus outbreak in February 2020. More specifically, the rate rose 1.2 percentage points for households headed by someone age 65 or older, and the increase from 2019 to 2020 was more substantial for households with family incomes below the national median [
130]. Besides rational motivations (e.g., very low interest rates, foreclosure moratoriums, etc.) an “emotional” factor could have also played a role in pushing so many Americans to make such an important decision at this particular time.
Limitations and Further Research
Given the massive exposure to a whole range of negative feelings generated by the COVID-19 pandemic, in this study we have systematically tested the effect of all primary negative emotions on two related and important determinants of economic choices: time and probability discounting. We also added moral disgust to the emotions investigated, although it is not considered a primary emotion, because we deemed it especially relevant in the context of the pandemic. Our results reverse the classical motivational appraisal approach, consistently identifying in sadness the negative emotion more likely to determine a higher tendency toward the selection of immediate and risky options. However, a recent study has shown steeper time discounting functions and higher risk propensity in depressed participants as compared to healthy ones [
131], while an additional study showed no overarching tendency for the depressed to engage in either more or less risk-taking, and that differences in risk-taking behavior are largely explained by behavioral traits such as locus of control, optimism, and trust [
132]. The same goes for the emotion of fear which has been traditionally associated with avoiding/withdrawing behaviors. Nevertheless, studies found that anxious participants show a higher tendency toward the selection on immediate vs. delayed choices [
133], while others report no relationship between anxiety and risk preferences [
134]. Hence, in general, the inconsistency of the results calls for further studies that investigate the specific mechanisms underlying the relationship between sadness and risk and shortsightedness propensity. Moreover, we studied negative emotions most likely elicited by the pandemic: however, in doing so, we used pre-validated scripts with no direct reference to the pandemic itself. Further research can corroborate our results using more context-specific scripts.
Finally, our study is the first that analyzed the effect of moral disgust on risk aversion and intertemporal choice; additional research is required in order to better delineate the role of this emotion in decision making, for example, comparing it not only with other negative emotions but also with positives ones.
A possible limitation of the current study is related to the use of emotional scripts which—in some cases—elicited more than a single emotion; on the one hand, eliciting multiple emotions can make it more difficult to disentangle the specific contribution of each emotion; on the other hand, it must be noted that in real life, emotions are rarely felt independently. Additionally, the employment of different methods to elicit emotions—such as the utilization of images or facial expressions—would be difficult to apply in the case of complex emotions such as moral disgust.
An additional limitation of the current study can be found in the use of hypothetical choices. Although several studies have reported no differences between hypothetical and real choices (involving actual monetary gain) in discounting tasks [
135,
136,
137], further studies are necessary to corroborate our findings using a fully incentivized structure.