2.1. What Is a Disaster?

The question “what is a disaster” has been addressed for decades ¹⁵. One of the first scholars to put forth a definition was Fritz ², who in 1961 said:

\(\ldots\) an event, concentrated in time and space, in which a society or a relatively self-sufficient subdivision of a society undergoes severe danger and incurs such losses to its members and physical appurtenances that the social structure is disrupted and the fulfillment of all or some of essential functions of the society is prevented.

Barton ¹⁶, too, offers an early conceptualization of the term, emphasizing a collective stress framework. Here, the focus is on the source of the threat and the system level impacted. A wide range of social scientists have offered definitions for example, Oliver-Smith and Hoffman ¹⁷ point to how disasters reveal what is already there. Tierney ¹⁸, too, emphasizes the social aspect, claiming that disasters are socially constructed and socially constituted. Despite these multiple definitions of disaster, there is still not complete consensus on the concept. The lack of consensus is partly due to widely varying operational and legal uses of the term, which are applied by a variety of international groups, national states, government agencies, and private groups.

Arguably, there are some common features of events, and there is not one all-purpose term. If we situate disasters as part of social process, we cannot omit crime from the equation because crime is part of the pre-event context. However, much of the preexisting empirical work on disaster has not taken crime into account because it was focused on debunking commonly held misconceptions about human agency. When the field of emergency management was new, emergency managers were most often perceived as “air raid wardens” who would sound an alert when an attack came from outside the United States ^19,20. At the time, the main concern was over nuclear powers, which prompted actual air raid drills, the creation of bomb shelters for protection, and the distinct possibility of threat from the outside.

Because assumptions had been made about human behavior, the Disaster Research Center (DRC) conducted a series of studies about human behavior after disaster. DRC findings debunked many myths about human behavior after disaster like the assumption that people will panic and engage in any behavior deemed necessary at the moment to facilitate escape ^1,21,5,22. Studies have shown that people do not panic, and in fact that the greater issue is fighting normalcy bias, where people underestimate the potential dangers ^14,23,24,25. DRC studies debunked the role abandonment myth, where it was once assumed that first responders would be too focused on their homes and families to do their jobs. The myth of widespread looting has too been debunked ¹¹. Yet, some places where disaster occurs experience looting.

This body of research focuses on prosocial behavior following a disaster event. These data are meant to illustrate the broad patterns in response during “typical” disaster events, and arguably overemphasize consensus after disaster, fail to consider that disaster events are changing, and that contemporary events are anything but typical. Arguably, severe social vulnerability and chronic social strain pre-event create social problems and may set the stage for criminal behavior.

2.2. The Case Against Crime

In addition to debunking dangerous myths about human behavior during disaster, one of the most significant early findings is what Fritz ² calls the therapeutic community, where post-disaster behavior is positive, aimed at helping others, and restoring normal community life. Barton ¹⁶, too, puts forth that the efforts to help following a disaster can be described as a therapeutic response. There are several reasons why post-disaster altruism triumphs over anti-social behavior. First, a disaster stops or disbands social discord. The trauma of the event is so great that people put aside their differences in the effort of survival and recovery, because the collective trauma of the disaster promotes social solidarity. Second, visible suffering also promotes empathy, which triggers social cooperation. Finally, a disaster event may present the opportunity for social change ². The therapeutic community, then, overrides any significant amount of criminal behavior.

Quarantelli and Dynes’ ¹⁴ research supports the therapeutic community too, and they put forth several reasons why crime does not happen. First, disasters come from the outside, which promotes solidarity. Second, the disaster event is an identifiable threat, which makes it easier for communities to mobilize for action. Third, people agree on priorities in disaster situations, and there is more emphasis on activities that benefit the whole. Fourth, the triage situation means people address those needs most paramount first. Fifth, disasters lead to focusing attention on the present, and there is a sense of “liberation” from everyday concerns. Sixth, disasters are leveling events, and there is a temporary breakdown in status distinctions. Finally, the disaster event strengthens community identification. These findings highlight the focus on disaster studies at the time, which was on how communities restore normalcy, how people do not lose agency, and how people are more likely to help others.

2.3. The Case that Crime Does Happen

Arguably (and respectfully), Quarantelli ¹¹ and Quarantelli and Dynes ¹⁴ focus too heavily on consensus, which misses how pre-existing structural inequality and strain may contribute to post-disaster antisocial behavior, and especially crime. The idea that disasters are great levelers is noble, but likely not the case when there is severe inequality and strain prior to the disaster. For example, a large percentage of the population in New Orleans could not evacuate before Hurricane Katrina struck because they did not have cars. These are the people—New Orleans’ poorest and most vulnerable—who ended up in the deplorable conditions of the Superdome. Indeed, this is directly related to structural inequality. A more recent (2003) middle-range theoretical framework for understanding disaster called social vulnerability acknowledges an unequal distribution of resources in the pre-disaster context. Generally defined, vulnerability is the potential for loss of life or property due to hazards. The hazards-of-place model ²⁶ combines the physical and social vulnerability to determine an overall place vulnerability. Social vulnerability is represented as the social, economic, demographic, and housing characteristics that influence a community’s ability to respond to, cope with, recover from, and adapt to environmental hazards and supports the claim that disasters only exacerbate pre-existing inequality.

Disasters, then, are directly influenced by pre-event patterns of inequality and strain. To focus only after-the-fact omits critical pre-event information that influences the post-disaster context and behavior. The social vulnerability perspective positions the disaster as both an independent variable responsible for triggering different responses, and as a dependent variable directly influenced by pre-event social patterns. Thus, the causes and the effects of disaster are not neutral with respect to social structure and processes. Disasters, then, are products of and contributors to ongoing social arrangements.

Criminology has much to contribute to our understanding of antisocial behavior in disaster, and cascading disasters and seminal events have opened the door for the possibility of a criminology of disaster ¹³. Strain theory in concert with social vulnerability considers pre-existing social factors and is therefore useful when working to understand patterns of crime during disaster. While environmental criminology (routine activities theory, rational choice theory, and broken windows theory) offers important explanations for how crime opportunities shift due to disaster, this paper focuses on strain and social vulnerability to illuminate the possible influence of pre-disaster conditions on crime patterns during disaster.

Strain theory has a robust history. Merton ²⁷ argued that the discrepancies between culturally defined goals and the institutionalized means that are available to achieve those goals determine if individuals engage in criminal behavior. Agnew ^28,29 attempted to clarify which stressors are more likely to lead to crime. Among them were failure to achieve goals when those goals are easily achieved through crime, poverty, experiences of prejudice or discrimination, and experience of criminal victimization. An abundance of strain over time may lead to a negative view of others, followed by negative emotions such as anger, which may in turn lead to the tendency to respond to affronts with aggression. We can consider a disaster an affront. Strain may be more likely to lead to delinquency when the strain is perceived to be unjust and/or perceived to be high in magnitude. The strain, then, may either create a predisposition for delinquency and crime, or it may function as a situational event. That is, existing, chronic strain may contribute to people committing crime, or, an event such as a disaster may be an acute situational circumstance that leads people to engage in criminal behavior.

Strain may be the mechanism through which macrolevel characteristics manifest in criminal activity during disaster. There are several studies that have shown that social vulnerability affects crime occurrence after disaster ^8,9,10. Genevie et al. ⁹ argue that when social problems (social vulnerability indicators) like crime, underground economic activity, unemployment, poverty, delinquency, drug abuse, and social inequality are high, community members do not feel like they belong to the community, and therefore may exhibit antisocial behavior like crime during disaster. Byrd and Matthewman ³⁰ found that power outages in particular provide an opportunity for certain types of crime, such as fraud, theft, and exploitation. Businesses found in poorer areas were more likely to be looted than businesses located in areas with greater economic prosperity ⁸. Frailing and Harper ¹³ argue that there is predictive potential to reduce burglary (proxy variable for looting) with an increase in formal guardianship (law enforcement presence).

We do not yet know how much each of these conditions needs to be present for burglary or other types of crime. If prior to the hazard, a large percentage of the population is under strain by way of discrepancy between goals and means, there is likely a pre-existing crime problem. If a community is under strain prior to the hazard—heavy pockets of concentrated poverty, lack of access to healthcare, political corruption, strained relations between law enforcement and the community, and high crime rates, it makes sense that that community would experience crime during a disaster. In a community already under strain, there is more likely to be criminal behavior as an extension of the pre-event strain. A key emphasis here is to avoid criminalizing poverty, as that is incomplete and uncritical.

Dynes and Quarantelli ⁴ themselves assert that when responding to an emergency, a community’s “primary resource is its pre-disaster social organization.” This claim further supports that if a community is experiencing strain prior to the disaster event, that community is likely to be less resilient. In this thinking, this same community is arguably more likely to experience criminal behavior, as part of the milieu of strain. There is strong empirical evidence that crime did occur during and after Hurricane Katrina, including looting, fraud, theft, burglary, domestic violence, sexual assault, and drug trafficking ⁷. There is evidence of crime occurring during St. Croix after Hurricane Hugo, burglary and fraud after Superstorm Sandy, robbery and gang-related criminal activity following the Haitian earthquake, among others ⁷. These communities have in common a history of vast social inequality and other indicators of social vulnerability.

Studies also demonstrate that violent crimes like domestic violence and sexual assault increase in the wake of disaster, despite the multiple barriers to reporting ^31,32. Reported accounts of domestic violence increased after a major flood event in Australia, and after the Mt. Pinatubo eruption in the Philippines ^31,33,34. Requests for temporary restraining orders increased by 50% after the Loma Prieta Earthquake in Santa Cruz ³⁴. Calls for service increased 50% after Hurricane Andrew ^35,31. Reported domestic violence and sexual assault increased following the Indian Ocean tsunami ^36,37. Intimate partner violence is a clear indicator of family violence, and the number of traumatic brain injuries in children—the strongest indicator of child abuse—increased following Hurricane Floyd ³⁸.

Important to acknowledge too is the potential influence of disaster response conditions. Indeed, some crimes are opportunistic and not strain-driven. Fraud, looting by non-locals, scams targeting survivors of the disaster, and burglary resulting from reduced guardianship (routine activities theory) cannot be fully explained through strain theory. Factors such as initial police capacity, communications breakdowns, emergency shelter management, or evacuation procedures can influence crime during disaster. In 1989, Hurricane Hugo destroyed much of the island’s infrastructure. Looting was widespread. Quarantelli ³⁹ contends that this looting was in part the result of the response to the disaster; government agencies that were tasked with responding were ineffective. Authorities’ loss of control immediately following Katrina’s impacts may have played a role in the disaster crime picture too ⁴⁰.

While there were many pre-existing conditions in St. Croix and New Orleans that can be considered regarding their influence on crime patterns, it is certainly important to consider the potential influence of disaster response conditions.

There may be disaster-specific strains, too, that may be useful in further understanding offender motivation as well. Loss of loved ones, loss of property, loss of income, inadequate provision of basic needs, poor official response arguably fall into Agnew’s ^41,29 categories of removal of potentially valued stimuli (losses) and presentation of negatively valued stimuli (inadequate provision of basic items), and, could lead to feelings of anger, and consequently, crimogenic behavior. Negative emotions may equate to anger, and then to criminal coping mechanisms or new offenders ⁴².

3.1. Inclusion Criteria

This systematic review includes document materials about each event such as government reports, community organization reports, police reports, and minutes from community meetings where available. For the 1977 NYC Blackout documents had to have been published after the event in order to be included. However, the time frame extended to after the 2003 NYC Blackout occurred in order to include comparisons of the 1965, 1977, and 2003 NYC blackouts. For Hurricane Katrina, documents published after landfall in 2005 are included. Because Katrina is one of the most well-documented disasters in our history, there are quantitatively more resources available on this event versus the 1977 NYC Blackout. Similar to the blackout sources, the timeframe for including sources about Katrina extends to present day to include any comparative pieces. A written document analysis worksheet (Excel) was used for internal validity. Interview data was used when available, mostly in the context of Hurricane Katrina in New Orleans. Analysis of these data yields a much richer picture of how disaster scholars can understand human behavior during crisis. The goal is to find the balance between not overlooking the differences between two places, and at the same time, acknowledging any patterns so to better anticipate where crime may emerge during disaster.

3.2. Exclusion Criteria

In this study, scholarly articles and government sources that did not include relevant topics, which are the pre-existing social context of NYC and New Orleans, were excluded. Sources focused on geographic areas outside of NYC and New Orleans were excluded. Although the search strategy did not impose any language limitations, sources in languages other than English were excluded due to accessibility issues.

Mass media sources were largely excluded due to additional bias that compromises validity. One source from Time Magazine was included as a major, long-standing global media brand that adheres to a journalistic code of ethics. It is noted that Time Magazine has a political left lean.

3.3. Methodological Limitations

There are several methodological limitations when conducting secondary data analysis. Secondary data analysis means that there may be data quality issues such as poorly documented data, incomplete records, and sources with unreliable methods. There also may be sources with irrelevant themes; this study is specifically examining crime after two disaster events. There are many sources available about these events that do not examine crime and thus are excluded from study. There are also access barriers—data that cannot be accessed ethically or legally for secondary use (sensitive data without proper de-identification).

Some of these limitations are not new to disaster research. First, the scope is limited to two disasters: the 1977 NYC Blackout and Hurricane Katrina in 2005. Second, comparing events that are roughly three decades apart, is methodologically challenging and brings validity issues into question. The social and political context of NYC in 1977 was very different than that of New Orleans in 2005. Emergency management, law enforcement, and academic research differ, making comparison challenging.

4.1. Social Vulnerability and Structural Strain

Hewitt ⁴⁴ has suggested that the pre-existing interface between humans and the environment is paramount to what happens during and after the disaster, rather than the post-disaster restorative period. The social vulnerability approach incorporates pre-hazard event social context, a variable that is critical in understanding the relationship between pre-hazard event social context and post-event human behavior. In this framework, those who are socially vulnerable are more likely to be adversely affected by disaster ⁴⁵. To operationalize social vulnerability, researchers have developed the social vulnerability index (SoVI) as a way to measure inequality ^46,47.

The SoVI has encouraged emergency management and key community stakeholders to know their communities better and has helped identify neighborhoods that may need additional support during disaster. New Orleans is one of the poorest and most racially segregated cities in the United States ⁴⁸. Pre-Katrina, 68% of the population were poor and African-American ⁴⁹. A majority (62%) of the city’s children lived in single-parent households. The child poverty rate was the highest in the nation, and the general poverty rate was 8th in the nation. The high percentage of single-parent households, a high unemployment rate, and very low family incomes created an extremely vulnerable population, with very high and concentrated poverty ⁴⁸. Census data about New Orleans reveals a poverty rate of 29%, which is nearly double the national rate ⁴⁹. The median household income in 2004 Metropolitan New Orleans was $37,146, one of the lowest in the country ⁴⁹. The city has one of the highest percentages of people without health insurance coverage, one of the lowest literacy rates, and only one-third of the population has a college degree. These data point to structural strain and social vulnerability prior to Hurricane Katrina ^26,50,51.

NYC, too, was a community under great stress leading up to the 1977 blackout. Despite gains in employment, income, and educational attainment, NYC experienced a rise in poverty from 1969 to 1979, and a continued high poverty rate from 1979 to 1999 ⁵². Growth in the Black and Hispanic populations and an increase in the share of those living in female-headed households are associated with the rise in poverty from 1969 to 1979 ⁵³. In 1969, 14.5% of the population lived below the poverty line, but by 1979, 20.2% lived in poverty ⁵³. While there was a national recession, the city itself had its own financial crisis as well. From 1969 to 1979, the populations fell from almost eight million to nearly 7 million people. Payroll fell from 3.8 million to 3.3 million jobs, and median family income plummeted from $36,543 to $29,878. Despite the continued growth in median and per capita incomes, there was little change in the poverty rate because of an increase in female-headed households and a rise in income inequality ⁵⁴.

Both New Orleans and NYC were experiencing great strain leading up to Katrina and the blackout. Disaster does not happen overnight; it exacerbates pre-existing social problems and may even provide opportunities for antisocial behavior.

4.2. Pre-Event Crime Picture

Expert disaster scholars argue that a disaster event reveals what is already there ¹⁷. Though it is not an absolute, if there is a crime problem pre-event, some level of post-event criminal behavior can be anticipated. More research is needed to understand the nuances of crime during disaster. Arguably though, it is incorrect to assume that a disaster will suspend criminal activity and/or that a therapeutic community “wins,” especially in a community that is already struggling. The disaster can change crime patterns (for example, drug trafficking routes) and invite new types of crime to emerge (for example, disaster relief fraud).

Moreover, crime patterns in an urban context can differ from crime patterns in non-urban contexts ^55,56,57. Park and Burgess ⁵⁸ argued that “a great city tends to lay bare to the public view in a massive manner all the human characteristics and traits which are ordinarily obscured and suppressed in smaller communities” (pg. 1). In other words, social phenomena—both good and bad—are more visible in an urban context. The Uniform Crime Report (UCR) is a database of crime statistics that paint what the crime picture looks like in an area. UCR data confirms that crime is less frequent in rural areas, and one explanation is that there are greater informal social controls in rural areas ⁵⁹. The lack of reported crime during the 2016 Louisiana Flood supports this ⁶⁰. While there may be some unreported crime, these rural communities affected by the flood certainly fall more closely with DRC findings that a therapeutic community supersedes crime. The Buffalo Creek flood is another example of prosocial behavior overriding criminal behavior in a rural place ⁶¹.

Disaster exacerbates validity limitations with crime data, creating additional barriers to relying on reported crime during disaster. For example, under-reported crimes like domestic violence are even less likely to be reported in the aftermath of disaster ³¹. Criminal codes may change following disaster, too. After Hurricane Katrina, police officers used the code “21K” to report nearly half of the city’s looting complaints. The “K” stood for Katrina, and the “21” stood for lost or stolen property, which is a standard pre-storm designation used when criminal activity is not clear-cut. The post-storm disruption made it hard to separate legitimate looting from actual storm losses ⁷. Despite the limitations of crime statistics, these numbers can help inform emergency management and policy decisions.

In 2004, New Orleans held the title for murder capital of the country with fifty-nine murders per 100,000 people, or 264 total murders ^62,63. While this number was down from the 424 found in 1994, it was still high compared to the rest of the country. By comparison, NYC had seven murders per 100,000 people in 2004. Violent crime overall was on the decline but still very high compared to the national average. Property crime, too, was down compared to previous years, but high compared to the rest of the country. Despite the mass evacuation of the city prior to and during Hurricane Katrina, the murder rate actually increased in 2005 ⁶⁴. While there was a drop in some crimes like murder, all other violent crimes had increased.

By 1977, while the population had decreased by nearly one million, the crime rate in NYC had increased ⁶⁵. Just prior to the blackout in 1977, the Son of Sam murders had happened, and the collective fear around this only exacerbated an already precarious city. Crime trends reveal a city on the decline. Both New Orleans and New York City reveal there was a crime problem leading up to the events. Without considering additional variables, it is no surprise then that crime happened during these events. The exact relationship is difficult to pinpoint, but the pre-event crime picture needs to be considered.

4.3. Status of Law Enforcement

Today, policing is especially complicated. The law enforcement community is subject to more intense scrutiny with smartphones at the ready to record their every move. There are changing standards of police accountability ⁶⁶. Many reforms have happened ⁶⁷. However, some law enforcement agencies are more functional than others ⁶⁶. Both function and effectiveness are extremely complicated variables to measure. Quarantelli ¹¹ mentioned that there was some level of dysfunction in law enforcement departments in places where we have seen criminal behavior following a disaster but did not fully explore or explain how. While the author of this paper does not have access to data directly measuring police corruption, integrity test data, or lifestyle monitoring data, crime rates can be an indicator of law enforcement effectiveness. Public opinion polls can also function as an indicator of law enforcement “health;” if the public largely does not have good confidence in law enforcement, there may be nefarious reasons why.

Law enforcement in New Orleans and NYC reveals a history of dysfunction, and, in both cases, corruption. New Orleans earned the title of murder capital of the country not only in 2024 but also in 1992 and 1994, and New Orleans Police Department (NOPD) solved only 37% of murder cases at the time—a clearance rate only half of the national average ^64,68. The root of police department corruption may be because of the historically low salaries of NOPD officers. Low salaries cultivated a culture of moonlighting and private security details in bars and strip clubs. The culture was one of “pseudo-organized crime,” and Moore ⁶⁹ argues that the loyalty became more to those “seedy” after-hours establishments rather than to the police department. There are many examples of corrupt behavior of officers; in 1993, NOPD’s vice squad was disbanded after allegations of theft. Three quarters of the members were assigned to desk jobs. Prior to Katrina, NOPD was suspected of beating an unarmed Tremé man, but the flood quickly washed those allegations away. Then, Katrina seemingly triggered a long list of corrupt behavior on the part of police officers, including Henry Glover’s still unsolved death, the Danziger Bridge case, and many smaller scale allegations of theft, fraud, and assault.

There were signs of corruption in the New York Police Department prior to the blackout in 1977 ⁶⁵. Because police were constantly angling for appointments and promotions from elected officials, “partisan wrangling” shaped the force from the very beginning ⁷⁰. Plus, the entrepreneurial parts of police work may have expanded with the growth of the city ⁷⁰. Police collected fees for serving warrants, detaining suspects, and appearing in court. Detectives often received rewards for reclaiming stolen property, and police regularly provided special services for citizens and businesses ⁷⁰. Thus, the boundaries between actual police work and rewards for those with political pull became fuzzy. Appointments for officers to obtain command positions required money and political influence. The Knapp Commission of the 1970s and the Mollen Commission in 1994 were tasked with investigating the department and revealed corruption in both years ^71,72.

If police corruption is problematic during routine times, corrupt behavior may be likely to extend post-disaster. Said differently, if corruption is normalized during routine times, it is likely to happen during collective crisis. And, there may be additional opportunities for corrupt behavior to emerge and even less oversight.

4.4. Socio-Historical Context

When a disaster happens, the social context informs—perhaps even determines—post-disaster behavior and outcomes. Neither New Orleans nor NYC had a stable, organized social structure prior to the disasters, and both places were under social stress leading up to each event. NYC in the 1970s was a place of disillusionment and distress. During the 1970s, prostitution steadily increased, the economy collapsed, and crime in the subway system dramatically increased ⁷³. In the late 1960s, district attorneys and borough presidents declared that police corruption contributed to the city’s rising crime rate. By 1969, murders were up to 1,043, compared to 746 just two years earlier ⁷⁴. The Knapp Commission reported severe and deep seated corruption in New York Police Department ⁷¹. In 1975, NYC was on the brink of declaring bankruptcy ^75,76. In 1977, David “Son of Sam” Berkowitz murdered six and injured seven more people over a thirteen-month span, worsening an already bleak social current of distress and hopelessness ⁷⁷. Then, in July of 1977, the blackout happened. There were looting and fires, more than 3,700 arrests, and a 300-million-dollar cost to the city of New York ⁷³.

When Katrina was threatening the Gulf Coast, the majority of residents in New Orleans did not have the means to evacuate the city. Compounding this was the fact that the homes of those who could not evacuate were the most vulnerable to weather events, especially flooding. While the rate of homeownership in New Orleans was relatively low, in some of the poorest neighborhoods like the Lower Ninth Ward, homeownership was higher than the national average. These homeowners had lived in their homes for many generations, which reflected the uniquely New Orleans emphasis on strong community ties and generational continuity. However, by government standards, this ownership was unofficial, as the homeowners did not have written documentation of the ownership (a deed to the home). This was particularly problematic during recovery because if the home was still standing, residents had to prove ownership to be eligible for relief. In other words, no deed meant no disaster relief monies.

The 1977 Blackout left New York City powerless in terms of electricity, exposing both engineering flaws in the power network to deeper rooted and long-lasting issues of racial inequality and oppression. The blackout presented context for the people who were powerless during routine times to suddenly seize some amount of power—or at least provided that perception. The National Urban League argued that disadvantaged people did not feel as if they needed to follow routine social rules because those routine social rules do not apply to them ⁷⁸. While it cannot be said that darkness causes crime, the convergence of pre-existing conditions of strain combined with the darkness may have enabled it.

Crime happened following Hurricane Katrina, and the patterns are similar to what happened in NYC. The city of New Orleans has a rich history of culture—food, music, architecture, and celebration—yet, it is one of the poorest and most racially segregated cities in the United States. Because of the pre-existing inequality, Hurricane Katrina caused extreme and unprecedented consequences. A high percentage of single-parent families, high unemployment, very low family incomes, and white flight to the suburbs made very high and concentrated rates of poverty. Immediately after landfall, over one million people were forced—many at gunpoint—to leave their homes and were forbidden from returning to their homes. The storm only exacerbated the effects of long-term social injustices.

4.5. The Hazard Events

Of course, the hazard itself plays a role in whether or not crime may happen during disaster. However, the role the event plays is related to the other variables pre-hazard. The 1977 NYC Blackout and Hurricane Katrina were both “atypical” disasters because of their impact on an already vulnerable place. Human negligence allowed a Category 3 hurricane to overwhelm the levee system in New Orleans, and over 80% of the city’s infrastructure—stores, school, homes, police and fire stations, museums, government buildings, restaurants, highways, and bridges—were totally destroyed. Katrina was one of the costliest and one of the deadliest hurricanes in U.S. history ⁷⁹. Severe destruction spanned from central Florida to Texas, and damage mostly came from flooding. Hurricane Rita happened just a few weeks later.

The 1977 NYC Blackout was localized in NYC, unlike the 1965 and 2003 blackouts. A lightning strike tripped two circuit breakers and subsequently, the system could not be restored. Additional lightning strikes caused the entire Con Edison power system to shut down nearly an hour after the first lightning strike. Power was not fully restored until the following day ^81,72,80. While these are hazard events, they cannot be divorced from the social context.

5.1. Future Research

Extensive future research is needed to fully understand crime and criminal behavior during disaster. While this manuscript contributes to the still burgeoning criminology of disaster subfield, more systematic research is needed to better understand the disaster and crime context. Specifically, further examining environmental criminology frameworks could contribute to this understanding. For example, routine activities theory ⁸⁶ is an extant criminological theory that can help explain disaster crime. This theory of crime examines three factors that must be present together in time and space for crime to occur: motivated offenders, suitable targets, and a lack of capable guardianship. In fact, several studies already use routine activities theory to explain disaster crime (for examples, see ^{87,42,88,89,10,90,91,92}). A disaster may create suitable targets, increase the number of motivated offenders, and may reduce guardianship. Analyzing opportunistic crime patterns, so we can clearly and with empirical evidence, distinguish between pro- and anti-social behavior can contribute to our collective understanding of disaster crime.

Future research can also compare disaster phases. Each phase—mitigation, preparation, response, and recovery—presents various challenges that disrupt the normal routine. Examining each phase more systematically and comparing phase and place can contribute to a better explanatory and perhaps predictive understanding of disaster crime. Finally, incorporating real-time response data would strengthen our collective understanding of disaster crime.