Monday, August 22, 2016

What does it mean to say that 'Walmart has a crime problem'?

Bloomberg Businessweek has a new cover story out, titled Walmart's Out-of-Control Crime Problem is Driving Police Crazy. The article is fascinating--in part because of how wrong it is.

Although the reporting is interesting, the article's conceptual premise is not well-established, leading to some confusion about the causal relationship between Walmart and crime. Let's dive in.

The article makes the following claims, which I'll take in turn:
  1. There's a huge amount of crime at Walmart stores
  2. Local police departments and stakeholders notice that fact
  3. Walmart could easily remedy the situation with different funding and policy choices
  4. Police (and maybe the article's authors) are annoyed that they don't see faster progress

There's a huge amount of crime at Walmart

This is the most straightforward claim, but also the most complicated, upon which the entire premise of the article relies.

Right off the bat it's worth noting that descriptions and anecdotes about individual crimes and situations do not constitute a representative data set. Journalism is all about details and texture, which is great, but in terms of justifying the claim that Walmart has a crime problem, they are not enough.

The crime statistics that are included are mostly presented without context: a local police call log is 126 pages long (compared to what?); there were hundreds of thousands of petty crimes at Walmarts this year (compared to what?); there is almost one violent crime every day at Walmarts around the country (compared to what?).

The one comparative statistic, between police calls from Tulsa's Walmarts and Tulsa's Target stores, is instructive. What is the takeaway here, exactly? The article pushes a narrative that Walmart is especially ineffective and unconcerned by crime, but perhaps Target is simply very effective or focused on crime-control. Certainly Target--as mentioned later in the piece--serves a higher-end consumer base. Would it be a surprise if charging higher prices resulted in fewer customers from crime-prone groups, resulting in less crime? Maybe Target's Tulsa locations are at lower-risk for crime, as the article implies. These questions are asked but not answered.

The failure to clearly define Walmart's 'crime problem' is reminiscent of the mini-scandal a few years ago over worker suicides at Apple's Foxconn factories in Shenzhen. In this controversy, anecdotal data created a media splash that was later undermined by comparative analysis showing that suicide rates were actually lower among Foxconn workers than in similar population groups.

The Bloomberg article's total absence of any mention of crime rates--as opposed to absolute statistics--is bewildering given their centrality in academic and applied crime analysis. It leads to the question: does Walmart really even have a crime problem, compared to other retail businesses? Who knows!

Walmart is a huge corporation, with thousands of stores, tens of thousands of employees, hundreds of thousands of customers, millions of square feet of retail space, and billions of dollars in revenue. Is their per-customer rate of shoplifting especially high? Per-hour-spent-on-premises? What is the rate of murder or violent crime compared to other businesses, residences or public spaces? Does Walmart report and prosecute illegal activity at higher or lower rates compared to other businesses? Who knows!


Local police departments and stakeholders notice crime at Walmart

Although local police officials might not know the actual crime facts in a rigorous statistical sense, some have apparently noticed something. That makes sense. Walmart is, after all, huge. In countless cities and jurisdictions across the country, the forces of zoning, land use, financing rules, regulation and political incentives have effectively ceded all public space and physical retail to suburban big box stores.

What we have being reported in this article is basically the mother of all selection bias effects. Setting aside the fact that journalists at Bloomberg--whose sourcing is overwhelmingly corporate rather than blue-collar--probably came to this story because Walmart itself is making a big push to reduce crime, the impression that Walmart has a crime problem is understandable.

If a huge proportion of retail activity and social interaction occurs at Walmart (especially if among disproportionately at-risk groups not subscribing to Amazon Prime), then naturally we would expect crime to occur there also.

In a recent article for Strong Towns' Big Box series, I made the case that big box stores are often blamed for social problems that actually have deeper structural causes, such as public policy, economics, technology and population demographics. Attributing crime to the institutional qualities of big box stores--or to Walmart specifically--is a much tricker causal inference challenge, and unfortunately Bloomberg fails in its identification strategy. Bloomberg sees crime at Walmart, and assumes Walmart is to blame. Wrong!


Walmart could easily remedy the situation with different funding and policy choices

The article quotes an expert saying that reducing crime at Walmart is easy: they just have to spend more money on better cameras, data systems and higher wages for better employees. Perhaps that's true. But this claim irks me for two reasons.

First, the sorts of core trade-offs that are suggested--higher wages, more capital investment, higher agenda priority for crime--are almost exactly the things that every competitive business specializes in. Strip away everything else, and the almost literal definition of a market economy is private businesses innovating via a Hayekian discovery process disciplined by profits.

Outside observers might study Walmart's corporate decisions viz. crime and have the benefit of impartiality. But they also suffer from a lack of deep knowledge and awareness about local incentives. It's safe to assume that Walmart employees know more about how to reduce crime at their stores than do police officers and business journalists. When it comes to backseat driving a huge transnational distributed retail corporation, it's best to take recommendations with a heavy grain of salt.

What assurance is there that increasing various corporate resources would in fact significantly reduce crime? The article contains no mention of a return-on-investment style of thinking, and instead assumes that resource infusions simply materialize out of thin air from Arkansas corporate.

The repeated counterfactual with Target is also strange: Walmart is open 24 hours a day, typically allows squatters in their parking lots, and employs relatively few people. The implication is that if Walmart scrapped these policies, it could reduce crime. Perhaps. But might this reduce its margins as well? Might this reduce its product differentiation and bring it into tighter competition with Target and other higher-end retailers? What if a store goes bust and closes? Would this reduce crime? Increase it?

This feeds into my second problem with assuming crime reduction is merely a question of Walmart's choices: it disregards the conceptual problems described in the previous two sections. Most of the crime at Walmart has nothing to do with Walmart as a decisionmaking institution per se; crime is really complicated, and it tends to happen at places of commerce and social mixing. To place the onus of crime reduction on Walmart merely because it's a key regulator of these deep structural factors--land-use, poverty, transportation, environmental lead exposure, psychological health--is a mistake.


Police (and maybe the article's authors) are annoyed to not see faster crime reduction.

Bloomberg does a pretty good job at skirting that weird professional line between reporting/analysis and outright advocacy. But the implicit message that I draw from the authors is that Walmart in fact should get its act together and do more to reduce crime. Setting aside the previously-raised empirical and conceptual problems (plus the fact that there's nary a mention of time series data that could be used to actually evaluate potential changes in store crime), does Walmart really have a public obligation to get better on crime? Well, perhaps.

First, it's important to understand the historical context here: liberals hate Walmart. Much of the initial furor was actually misdirected rage over downtown economic malaise, but after Walmart really took off the critique morphed into a vague elitist cause célèbre, covertly aligning with NIMBYs and cultural snobs fearful of downmarket retail.

In recent years anti-Walmart folks have emphasized low wages and the extent to which employees rely on public assistance programs like food stamps (therefore 'subsidizing' Walmart). Although that last critique is incoherent, it does highlight how economically massive Walmart has become in some communities, and how much it dominates the public realm in a practical sense.

Indeed, this distinction between public and private is central to understanding Walmart's social obligation. Every business is accountable to the public, in some sense. I'm not familiar with the legal and regulatory situation surrounding retail and crime, but it doesn't appear that Walmart is in blatant violation of existing rules. If Walmart does have an unmet moral obligation regarding crime and public safety, it is currently not reflected in law and regulatory norms.


Cops versus Walmart. Really Bloomberg? Really?

My uninspiring conclusion is that Walmart isn't obviously obliged to improve its crime situation beyond what is competitively rational. Our legal and government structures clearly must do more to confront the reality that private corporations now function as town squares. But the ideal role that Walmart should play in controlling crime outside of any police or regulatory context is unclear to me.

Indeed, my suspicion is that many residents who are personally exposed to Walmart's crime risks are more apt to demand action from local police and government officials than from Walmart itself. If Walmart is where crime and commerce happens, presumably there is a decent case for increased police integration and local government oversight. Instead, the article bypasses this perspective entirely, framing the issue as basically one of cops griping and corporate analysts analyzing.

It's not a good look.

Bloomberg is of course a business publication aimed at business readers, so maybe it's unfair to be this harsh. From a certain perspective, even articles as muddled and detached as this one can become self-fulfilling analyses and have positive secondary effects on social justice. By merely raising the profile of this issue, it exerts competitive pressure on Walmart to improve its safety performance. Let's hope so.

But for readers who are concerned with crime and social justice, I fear the easy takeaway will be that Walmart is still evil and now it even doesn't care about crime and murder! If only Walmart's executives would just get woke and start caring about the communities they serve, things would improve, right? Guhh.

Walmart is obviously a high-profile target for activism, and tactically it makes some sense to aggressively leverage its unique qualities to pursue change. But shoddy analysis and misdiagnosing social justice problems does a disservice to those seeking effective solutions. Crime is an unbelievably complex issue that is overwhelmingly driven by bottom-up factors unrelated to Walmart's security and loss-prevention policies. Scapegoating our own political failings when it comes to crime and the management of public space isn't what we need right now.

No comments :

Post a Comment