📈 economics

Natural Disasters and Criminal Activity: Evidence from Ecuador

This paper analyzes the 2016 Ecuador earthquake to demonstrate that severe ground shaking caused a sharp, temporary surge in property crime arrests—driven by offenders shifting criminal activity to the present due to lowered enforcement costs—while violent crime remained stable and alternative socioeconomic factors were ruled out as primary causes.

Original authors: Kaveh Sanjabi Malayeri

Published 2026-08-12

📖 5 min read🧠 Deep dive

CC BY 4.0

Original authors: Kaveh Sanjabi Malayeri

Original paper licensed under CC BY 4.0 (https://creativecommons.org/licenses/by/4.0/). ✨ This is an AI-generated explanation of the paper below. It is not written or endorsed by the authors. For technical accuracy, refer to the original paper. Read full disclaimer

The Crime Wave That Wasn't: Why Disasters Don't Always Mean Mayhem

Imagine the world of crime as a giant, invisible game of chess played by people, police, and the environment. For decades, scientists studying this game—known as criminologists and economists—have tried to figure out what makes the pieces move. They often look at two big ideas: the "cost" of breaking the rules and the "opportunity" to get away with it. Think of the cost as the fear of getting caught, and the opportunity as an open door left unlocked. Usually, when a huge disaster hits a town, people expect chaos. The old saying goes, "When the cat's away, the mice will play," suggesting that if the police are busy fighting fires or rescuing people, the "mice" (criminals) will rush in to steal everything. But is that always true? Does a disaster always mean a spike in crime, or does the human heart react differently depending on what kind of bad thing happens? This is the big question researchers ask when they look at how natural disasters change the behavior of a community.

The Earthquake That Made Thieves Rush, But Neighbors Hug

This paper dives into a massive earthquake that hit Ecuador on April 16, 2016. It was a 7.8-magnitude quake, a terrifying event that shook the ground, destroyed homes, and left thousands displaced. The author, Kaveh Sanjabi Malayeri, wanted to see what happened to crime in the weeks and months after the shaking stopped. He didn't just guess; he looked at real police records from 219 different towns (called cantons) across the country, comparing the towns that got hammered by the quake to the ones that barely felt it.

Here is the surprising twist the study found: Property crime went up, but violent crime went down.

Think of it like this: When the earthquake hit, the "police shield" was temporarily broken. The officers were busy saving lives and fixing roads, so the chance of getting caught for stealing dropped. The author argues that thieves, acting like rational players in a game, saw this open door and rushed through it. This is called temporal displacement. It's like a thief deciding, "I was going to steal that bike next month, but since the police are busy right now, I'll do it today." The result? In the towns that shook the hardest, property crime arrests jumped by about 2.2 per 100,000 people every month. That might sound small, but it was actually two and a half times higher than the average for towns that weren't hit hard.

However, this crime wave was a flash in the pan. The data shows a sharp spike in May 2016, where arrests were roughly ten times higher than before the quake. But by August, just four months later, the numbers had dropped back down to normal. The "window of opportunity" closed as soon as order was restored.

Why Thieves Ran, But Neighbors Hugged

The paper also explains why these two types of crime moved in opposite directions. Property crimes, like stealing cars or breaking into houses, are often about opportunity. When the system is chaotic, the risk of getting caught is low, so the "rational" choice is to steal.

But violent crimes, like assaults or fights, are different. They are often "crimes of passion." The study suggests that after a disaster, something magical happens: people come together. Neighbors help neighbors, and the community bonds over the shared trauma. This surge of "social glue" actually made people less likely to hurt each other. In the hardest-hit towns, violent crime didn't spike; it actually declined slightly. It's as if the disaster turned the town into a temporary "utopia" where people focused on survival and helping each other, putting aside their personal grudges.

What It Wasn't (And What We Can Be Sure Of)

You might wonder, "Did the crime go up because people lost their jobs and got desperate?" The author checked this carefully. The data shows that unemployment actually went down in the hit areas (probably because of all the rebuilding work), and incomes didn't drop. So, the "desperate poor stealing to eat" story doesn't fit the facts here.

Could it be that the police just stopped showing up? The study found that police hiring actually stayed steady or even went up slightly. The rise in crime wasn't because the police left; it was because the police were busy doing rescue work, which temporarily lowered the "risk of getting caught" for thieves.

The author is very confident in these findings. They used a method called "difference-in-differences," which is like a scientific control group. They compared the hit towns to the safe towns before and after the quake, and they even ran a "fake" test using data from 2015 (when there was no earthquake) to make sure the crime spike wasn't just a seasonal fluke. The spike only happened in 2016, right after the quake.

The Takeaway for the Future

So, what does this mean for the real world? The paper suggests that if a disaster happens again, police shouldn't panic and assume crime will stay high forever. The spike is short-lived. Instead, the best strategy is hotspot policing. This means sending extra officers to specific, high-risk areas (like markets or shopping centers) for a few weeks right after the disaster. This closes the "open door" for thieves quickly. Once the police are back in their normal rhythm, the crime rates naturally fall back down.

In short, the study tells us that while a disaster might give thieves a brief, rational reason to strike, it also gives communities a reason to stick together and stop fighting each other. The chaos creates a temporary window for crime, but the human spirit creates a temporary shield against violence.

Technical Summary: Natural Disasters and Criminal Activity: Evidence from Ecuador

Problem and Motivation
Natural disasters create immediate chaos and disorganization, yet their secondary impacts on social stability and criminal behavior remain under-explored, particularly in low-to-middle-income contexts. While existing literature offers mixed evidence—some studies suggesting crime rises due to reduced opportunity costs (Belasen and Polachek, 2008), while others indicate a decline due to strengthened community ties (Dynes and Quarantelli, 1985; Hombrados, 2020)—there is a need for causal evidence isolating the specific mechanisms driving these changes. This paper addresses this gap by examining the impact of the April 16, 2016, 7.8-magnitude earthquake in Ecuador on property and violent crime. The study seeks to determine whether the post-disaster environment fosters a surge in criminal activity and, if so, to identify the behavioral mechanisms (e.g., economic distress vs. temporal displacement) responsible for such changes.

Data and Empirical Strategy
The study utilizes a high-frequency, monthly panel dataset covering 219 cantons in Ecuador for the years 2015 and 2016. The primary data source consists of daily police arrest records from the National Police of Ecuador, detailing the type of crime, arrest date, location, and demographic characteristics of the arrested. These are merged with geological data from the U.S. Geological Survey (USGS) regarding Modified Mercalli Intensity (MMI) shaking levels in each canton.

The identification strategy employs a Difference-in-Differences (DiD) design combined with an event-study framework.

Treatment Definition: A canton is classified as "treated" if it experienced shaking intensity greater than MMI 4, a threshold where damage becomes noticeable. Control cantons experienced lower intensities.
Model Specification: The baseline model estimates the interaction between a post-earthquake time indicator and the treatment status:
$C_{kit} = \beta_0 + \beta_1 Post_t + \beta_2 Treated_i + \beta_3 (Post_t \times Treated_i) + \varepsilon_{kit}$
where $C_{kit}$ is the crime rate (arrests per 100,000) for crime type $k$ in canton $i$ at time $t$ .
Dynamic Analysis: An event-study model with leads and lags is used to trace the dynamic response over time, testing the parallel trends assumption and visualizing the temporal evolution of the shock's impact.
Robustness: The analysis includes placebo tests using 2015 data, long-run pre-trend analysis (extending back to January 2015), and checks for alternative channels (unemployment, inequality, income, police hiring) using quarterly administrative data from 2010–2018.

Key Results
The analysis reveals a distinct divergence in how property and violent crimes respond to the earthquake:

Property Crime Surge: There is a statistically significant increase in property crime arrests in severely shaken cantons. The DiD estimate indicates an increase of approximately 2.2 arrests per 100,000 inhabitants per month, roughly 2.5 times the mean of the control cantons.
Temporal Displacement: The event-study estimates demonstrate that this increase is sharp but temporary. Property crime arrests peaked in May 2016 (one month post-quake) at roughly ten times the pre-earthquake baseline level. The effect decays rapidly, returning to baseline levels by August 2016 (four months post-quake).
Violent Crime Decline: In contrast, violent crime shows no significant increase; if anything, it declines slightly in the months following the disaster, with coefficients becoming negative and significant in the event-study estimates for September and December.
Mechanism Verification: The paper rules out economic distress as the primary driver. Data on unemployment, inequality (Gini coefficient), and per-capita income show no deterioration in affected cantons; unemployment actually fell slightly, and income rose, likely due to reconstruction spending. Furthermore, police recruitment did not decline, suggesting the crime spike was not due to a permanent reduction in enforcement capacity but rather a temporary disruption.

Mechanisms and Interpretation
The author argues that the rise in property crime is driven by temporal displacement rooted in rational-choice theory. In the immediate aftermath of a destructive earthquake, the "cost" of criminal behavior is lowered due to:

The diversion of police resources to rescue operations.
The breakdown of social order and informal monitoring.
The perception of a "window of opportunity" where the probability of apprehension ( $p_t$ ) is temporarily reduced.

Rational offenders are hypothesized to shift criminal activity from the future to the present to exploit this transient reduction in enforcement. This explains the sharp, short-lived spike in opportunistic property crimes (theft, burglary, robbery). Conversely, the decline in violent crime is attributed to increased social cohesion and community solidarity in the face of shared trauma, which suppresses interpersonal conflicts and "crimes of passion."

Significance and Policy Implications
The paper contributes to the literature by providing causal evidence from a South American context, moving beyond the US-centric focus of much prior research. It distinguishes between the mechanisms driving property versus violent crime, challenging the assumption that all crime types respond uniformly to disasters.

The findings suggest that post-disaster crime surges are not necessarily indicative of long-term social collapse or economic desperation but are often transient responses to temporary enforcement gaps. Consequently, the author posits that hotspot policing—the strategic, temporary deployment of police resources to high-risk areas immediately following a disaster—is an effective mitigation strategy. Because the window of opportunity for offenders is short-lived, targeted interventions can close this window before patterns return to baseline, without the long-term displacement concerns often associated with such policing tactics.

Limitations
The author notes that the study relies on arrest data rather than reported offenses, meaning the estimates represent a lower bound on the true change in criminal incidence. Additionally, the analysis focuses on a single event in one country; while the mechanism of weakened enforcement is generalizable, the specific magnitudes may vary across different disaster scales and institutional settings.

The Crime Wave That Wasn't: Why Disasters Don't Always Mean Mayhem

The Earthquake That Made Thieves Rush, But Neighbors Hug

Why Thieves Ran, But Neighbors Hugged

What It Wasn't (And What We Can Be Sure Of)

The Takeaway for the Future

More like this