For the past 1 year, the world has faced a global outbreak of acute respiratory illness of unprecedented extent and impact. This illness, caused by the novel betacoronavirus SARS-CoV-2, has been designated COVID-19. Clinical manifestations of COVID-19 range from asymptomatic viral carriage to severe respiratory illness, with the latter outcome being more common in the elderly and those with significant medical comorbidities (1). At the time of writing this paper (January 12, 2021), over 86 million cases of COVID-19, and over 1.8 million deaths due to this disease, have been reported globally (2).

In an effort to contain the spread of this disease, local and federal authorities worldwide have resorted to large-scale containment measures such as lockdowns, “stay-at-home orders,” and restrictions on commercial, educational, religious, and other public activities. Though deemed necessary by governments and experts, these measures have often led to widespread socioeconomic disruption, difficulties in accessing healthcare, and shortages of food and other essential supplies (3).

Against this background, a significant proportion of the general population has experienced various forms of psychological distress, such as symptoms of anxiety, depression, and post-traumatic stress. The frequency, severity and correlates of these phenomena have already been documented extensively in several systematic reviews (4–6). Besides these well-recognized phenomena, reports of “panic buying” in response to the COVID-19 pandemic have been reported in the literature since March 2020, with the earliest published reports coming from Asian countries (7–9). Wang et al. (9). Panic buying, also referred to as “stockpiling,” is characterized by the purchase of “unusually large amounts of products,” in anticipation of or during a natural or man-made disaster, related to a fear of shortage of unavailability of the concerned products, which are usually food or hygiene-related items (10, 11). Panic buying has been documented as a local response to phenomena such disease outbreaks or typhoons (12, 13), but has occurred on an unprecedented scale in the context of COVID-19 and the attendant restrictions imposed in an attempt to contain the spread of the disease (14).

Various explanatory models have been advanced to account for this phenomenon (11, 14). Before examining these, a brief review of the existing literature on the scope of this phenomenon, and the factors associated with it, is in order.

Changes in purchasing behavior during the COVID-19 pandemic were commonly reported in the general population; in a study of Spanish consumers, over 60% reported such changes. However, only a small proportion of them exhibited actual panic buying (15). A review of media reports (14) found descriptions of panic buying from over 20 different geographical regions, encompassing both developing and developed countries. This paper found a predominance of reports (97.6%) from countries with a high level of urbanization and industrialization, though this may simply reflect a reporting bias. A study from Portugal reported stockpiling in 36% of respondents (16). In a report from the United States, the most common stockpiled materials were toilet paper (63%), canned foods (59%), staple foods such as rice and bread (53–57%), bottled water (57%), and medications (53%) (17).

A number of factors have been associated with an increased risk of panic buying. These may be conveniently classified as follows:

Alongside the descriptive research on panic buying, summarized above, several authors have speculated on the possible causes or mechanisms underlying this complex behavior. From the perspective of survival psychology, acquiring essential supplies during an actual or threatened disaster is an adaptive behavior; however, when this behavior is influenced by excessive anxiety or fear of the “unknown,” or of “losing control” over the situation, the result is irrational decision-making and panic buying (8, 11, 21). Alchin (25) has suggested that excessive exposure to displays of “panic buying” by others, either directly or through the media, can lead to the activation of evolutionarily primitive brain pathways which suppress higher-level “pro-social” behaviors, leading to a behavior is socially inappropriate. His model has also highlighted the potential role of social learning or imitation in reinforcing maladaptive cognitions regarding the risk of food shortage or scarcity of other essential supplies. From a broader socioeconomic perspective, Keane and Neal (13) have highlighted the role of local and international patterns of viral transmission, as well as government-imposed restrictions on internal (but not external) movement, in influencing panic buying.

While these explanations have a great deal of merit to them, they are deficient in three aspects. First, they are—to a certain extent—unduly reductive, and fail to consider the way in which biological, individual and social factors may interact. As a result, suggestions to improve panic buying that are derived from one model, such as excessive anxiety, may not be effective in areas where social factors are of greater significance. Second, these models fail to account for the significant variability in panic buying that has been reported across individuals, regions and countries. Third, these models do not explain the specific phenomenon of panic buying, which is a unique and discrete behavior pattern—at best, they provide general explanations for psychological distress and maladaptive coping during the COVID-19 pandemic.

In order to address these limitations, an outline for a more comprehensive model was proposed by Arafat et al. (21). In this model, the precipitating event (a disaster, such as a pandemic or earthquake) interacts with psychological processes (appraisal and processing of the event) and social factors (such as cultural and political variables) to produce a sense of threat (resource scarcity) leading to panic buying. However, they have noted that this model needs to be verified by further research.

This paper takes the position that the broad outline proposed by these authors is essentially correct. However, it may be possible to refine it further and define areas for research, prevention and intervention by taking into account recent advances in the biological and social sciences. These advances and their implications will be discussed in section Divergent Theoretical Perspectives on Panic Buying below, and an attempt to integrate them will be presented in section Summary: Integrated Biopsychosocial Model of Panic Buying.

From a biological perspective, Alchin (25) has attempted to explain panic buying in terms of the polyvagal theory proposed by Porges (26). According to this theory, perception of threat by the brain generates “primitive” fearful responses through activation of the autonomic nervous system and hypothalamic-pituitary-adrenal axis; however, if a situation is subsequently determined to be safe, this response can be overridden through a putative “social engagement system” (SES), mediated through the vagus nerve, which dampens stress responses and facilitates prosocial behaviors. However, exposure to continuous threatening stimuli—especially through the media, including social media—may overwhelm the SES, leading to maladaptive behaviors that are socially harmful. This theory, though plausible and backed by advances in neuroscience, functions better as a general theory of maladaptive behavior during the COVID-19 pandemic, and does not explain why a specific form of behavior (panic buying) should arise.

A more fruitful approach may be obtained by examining the similarities between panic buying and certain known psychological disorders—more specifically, obsessive-compulsive disorder (OCD) and related conditions such as hoarding disorder and compulsive buying. These latter conditions are considered to lie on the “OCD spectrum” and share genetic and neurobiological links with it (27–30). Hoarding disorder, whether considered as a discrete entity or as a subset of OCD, is characterized by the irrational accumulation of materials, usually of a non-essential nature, to an extent that causes functional impairment (31). Though the symptoms of these disorders are recognized as irrational and excessive by sufferers, they can also be understood from an evolutionary perspective as a dysfunctional variant of adaptive behaviors related to threat detection and harm avoidance (32). More specifically, hoarding behavior has been conceptualized as a form of altruistic behavior that helps in maintaining a supply of scare resources, particularly in the face of a large-scale disaster or other cause of scarcity (33, 34). The argument, in brief, runs as follows: individuals with an innate tendency to hoard essential materials during a crisis would ensure a survival advantage for both individuals and their community as a whole, particularly in more “primitive” or traditional societies. On the basis of analogies with animal behavior, a similar hypothesis has been proposed by Miguel and Ligabue-Brown (35); their proposal requires no putative link to OCD.

From this perspective, it is possible to understand the emergence of stockpiling behavior in the face of actual or anticipated disasters, and especially in the context of a global crisis such as COVID-19. However, this explanation is missing a crucial factor: what causes this behavior to cross the “threshold of rationality” and take on an excessive form, namely panic buying? A number of neurobiological factors may underlie this phenomenon, including a paradoxical increase in prefrontal gray matter (36) and over-activation of specific brain regions, such as the anterior cingulate and dorsolateral prefrontal cortices (37). A similar model implicating under-activity of the medial prefrontal cortex and overactivity of nucleus accumbens has been postulated for compulsive buying (38). These changes may be reflected in higher-order constructs such as difficulties in emotion regulation, and intolerance of uncertainty or distress. From this viewpoint, the numerous uncertainties and negative emotions associated with the COVID-19 pandemic and its attendant socioeconomic changes may trigger aberrant neural activity in susceptible individuals, leading to an aberrant form (panic buying) of an otherwise useful behavior (purchasing essential materials) (39).

In addition, regional or cross-national variations in panic buying may be explicable in terms of genetic variations influencing brain structure and function, which account for 36% of the variance in hoarding disorder (40); further, there is a correlation of over 40% between genetic vulnerability for hoarding disorder and OCD (41). In this context, it is worth noting a consistent association between the long (l) allele of the serotonin transporter functional polymorphism 5-HTTLPR, located on chromosome 17, and vulnerability to OCD in certain subgroups (42). This genetic variant is of particular significance to behavioral responses to the COVID-19 pandemic, as there is some evidence that it was subject to selection pressures caused by infectious disease, and that it may influence behavioral responses to the threat of infection (43). These aspects will be discussed further in section Summary: Integrated Biopsychosocial Model of Panic Buying.

It can be argued that though phenomenologically similar, panic buying may not necessarily exist on a continuum with OCD; to date, no literature has reported an association between the two, or an increased risk of panic buying in OCD patients. This could be explained by the alternate hypothesis that hoarding may have different evolutionary roots from OCD in general (35). In the model proposed by these authors, human genes homologous to those involved in hoarding behavior in birds may be activated via epigenetic modification during traumatic or threatening situations. This would trigger a tendency to accumulate possessions and derive a sense of safety from them. Another neurobiological hypothesis arises from the similarity of panic buying to compulsive buying. In situations where opportunities for natural rewards from the environment are lacking (due to containment measures or social distancing), repeated shopping or buying might activate the mesolimbic dopaminergic pathway, leading to a perception of reward and reinforcing this behavior.

If hoarding behavior has remote evolutionary origins, and can be triggered by a low tolerance to (or an inability to self-regulate) psychological distress, what are the specific psychological mechanisms or pathways linking distress and panic buying? In examining this association in the context of COVID-19, a useful perspective can be obtained from the field of attachment theory, as formulated by John Bowlby (44, 45). According to attachment theory, many psychological disorders have their origins in disruptions of attachment bonds in early childhood, and can be triggered or exacerbated by the disruption of attachment bonds later in life (46). Contemporary neuroscience has identified a certain degree of overlap between the brain regions involved in attachment behavior and those implicated in hoarding disorder, such as the anterior cingulate cortex (37, 47). Further, there is evidence that parental abuse, neglect or separation in early childhood, which are all associated with significant disruption of attachment bonds, are associated with an undue emotional attachment to possessions, which in turn is linked to the severity of subsequent hoarding behavior (48). Similarly, later experiences of social exclusion (49), trauma in interpersonal relationships (50), or poor social support (51) have been associated with the development and maintenance of hoarding behavior in adult life. This may be mediated through insecure adult attachment, leading to negative affective states which in turn trigger an increased attachment to, and desire to accumulate, certain possessions (51, 52). Similar phenomena have also been reported in compulsive buyers, though they are less well-characterized (53, 54).

The addition of these facts to the neurobiological framework discussed in section Neurobiology allows a more complete picture to emerge. In a given individual, genetic vulnerability interacts with disruption in early attachment bonds to influence the structure and function of discrete brain regions implicated in hoarding disorder, which is manifested in higher-order psychological constructs such as difficulties in emotion regulation, distress tolerance and tolerance of uncertainty, or a heightened attachment to possessions. However, the emergence and persistence of hoarding symptoms requires further disruptions or deficits in interpersonal attachment in later life (adolescence or adulthood). This is particularly relevant to the COVID-19 pandemic, where infection control measures have led to the disruption of existing interpersonal bonds on a global scale (3, 45). In such contexts, even individuals who had never previously exhibited hoarding behavior may develop an increased attachment to, and urge to accumulate, certain possessions—leading, not to hoarding disorder per se, but to panic buying. In this context, it is worth noting that many of the individual psychological variables identified as being linked to panic buying, such as high neuroticism and low openness and conscientiousness (17), have been specifically associated with disturbances in attachment both in children and in young adults (55).

Though this model is both evidence-based and plausible, it remains deficient in one significant aspect. Hoarding disorder is characterized by the accumulation of materials which are non-essential, while panic buying is characterized by the accumulation of objects which are actually or potentially essential; though excessive, it is not pathological because lacks the irrationality which is characteristic of hoarding disorder; neither is it unrelated to external threats, as in compulsive buying disorder. In order to explain this discrepancy, it is now necessary to turn to social-anthropological perspectives on these phenomena.

Hoarding disorder is characterized by a pathological form of attachment to material objects; however, object attachment varies significantly across individuals and cultures, and is influenced by a variety of factors. These include exposure to specific life events, individual and cultural beliefs, physical health, and cognitive functioning. From an anthropological perspective, an object that appears “non-essential” or even “worthless” may have a symbolic significance for a given individual (56, 57). In other words, it may be unduly reductive to create a dichotomy between “essential” and “non-essential” forms of object accumulation. Instead, these behaviors may exist on a spectrum or continuum, ranging from normal object attachment, to rational stockpiling of essential supplies in the face of threat, to excessive stockpiling of essential supplies in the same context, to overt panic buying, and finally to hoarding disorder or compulsive buying where the element of rationality is almost completely absent (16, 53, 57) (see Figure 1). In support of this notion, it is useful to note that the materials hoarded in panic buying may vary significantly in perceived “usefulness” across cultures. For example, the hoarding of guns and ammunition, which have symbolic significance in terms of safety and self-defense, may occur in certain Western countries (17) but not in Asian countries (21). Moreover, the value attached to possessions in general varies across cultures. Western countries generally assign a higher level of importance to material possessions, which are often portrayed as linked to success, happiness and popularity (58); compulsive shopping has been reported more frequently in such cultures (59), as has panic buying (14, 21).

These theories of symbolic meaning must be placed in the broader social framework of the COVID-19 pandemic, which has been characterized not only by large-scale disruptions of social bonds (as outlined in section Attachment Theory) but by economic disruptions leading to unemployment and poverty (60, 61). In turn, these disruptions have been associated with increased rates of specific social problems, such as domestic violence (62) and social exclusion or stigmatization (63) at the individual level. The large-scale social unrest and hardship occasioned by this pandemic is thus “translated” into a number of individual-level stressors, most of which have been associated with the initiation or maintenance of hoarding behavior (49–51). At a symbolic level, a lack of trust in local or federal authorities—which may be worsened by the disruption of social bonds (45)—can further increase the likelihood of panic buying, as can a social learning effect based on imitation (25). The interactions between these factors are likely to be complex and non-linear.

With these details in place, it is now possible to formulate a more comprehensive model.