When he worked in Africa, development economist Ariel BenYishay noticed small-scale entrepreneurs were often unable to grow their businesses beyond the initial stage, even when the opportunities were obvious.
"They told me they could do so much more if they had a piece of machinery – a car, a motorbike, or a sewing machine," says the 33-year-old lecturer in the Australian School of Business.
He assumed the urban street-sellers and small-service providers simply couldn't afford the items or they were unavailable. But it turned out robbery was common in many of the communities and people were reluctant to invest in new items.
"Why would they invest if it was likely to be stolen?" BenYishay says.
The realisation spurred the Israeli- American on to his next project, looking at the impacts of crime on micro-economic activity, especially the behaviours crucial to business growth.
Together with Dr Sarah Pearlman from New York's Vassar College, BenYishay began comparing crime statistics with data on the behaviours of poor people in urban Mexico.
Based on BenYishay's observations in Africa, the researchers first looked into robbery as a possible explanation for sluggish micro-enterprise growth in Mexico. As expected, higher rates of property crime in urban areas were associated with significantly lower rates of business expansion.
But what of other forms of crime, like homicide? Were they also having an effect? Again Mexico was an ideal case study.
Violent crime in certain Mexican regions has surged in recent years, largely as a result of a 2006 federal government crackdown on drug cartels. Five Mexican cities now rank among the top 10 most violent cities in the world, according to the Citizens' Council for Public Safety and Criminal Justice.
"The gruesome and public nature of many of the killings as well as the perception that cartel members can operate with impunity has generated a high level of fear within the population," BenYishay says.
This fear has a tangible economic impact, the researchers found, with an increase of 10 homicides per 100,000 in a year (or 0.01%) associated with a 1% drop in the number of hours worked per week. The effect was even greater on the self-employed, specifically those who worked from home.
The correlation is significant given that drug-related homicides in Mexico increased 480% between 2006 and 2011 (from 2,120 to 12,366) and homicides overall increased 45% from 25,780 to 37,375.
The results were even more pronounced in states most affected by drug cartel violence, where average hours worked by the self-employed had declined by up to two and a half hours a week.
The research is some of the first evidence that fear of violence leads to individual behavioural changes that lower economic activity.
"Even when the risk of violent crime is relatively small over a given year – less than 1% – private individuals may adjust their behaviour by as much as 30–50%. The marginal hours are the riskiest: going home in the dark, for example. Instead people leave work early."
BenYishay says we've long known violent internal conflicts like Mexico's war on drugs have serious impacts on short-term and longer-term economic development, largely due to effects on public services and institutions.
"But when it comes to the wider impact of violence and crime – in this case on productivity and labour market participation – it's not all about the state. It's also about the individual."
Explore further: No correlation between medical marijuana legalization, crime increase, study says