A 20-year-old recently became internet famous for punching a
who joked about raping a drunk girl on the street one

When is it ethical--and more importantly, efficient--  for individuals to take justice into their own hands? This question is particularly salient in Ghana, where vigilante justice is a common response to crime, and that justics goes up to and includes death for perpetrators of rape or severe property crime.
Actually, I did see what you did there. And I have a machete.

Justice generally serves two economic purposes: to make an injured party whole, and to provide a disincentive against bad behavior. Vigilante justice tends to be focused on the latter. If you doubt this, just wait for the next Batman movie-- I guarantee my dear Bruce Wayne will wreak more economic havoc as the caped crusader than he will provide in redress. Vigilante justice comes at a high price, to the executioner, society, and to innocent parties who fall mistakenly fall victim to it, and is limited in the benefits it can provide. 

In order to be justified then, vigilante justice must have a very high deterrent benefit. This will generally only be the case when legal justice systems completely fail to act as deterrents to morally reprehensible acts.

It seems to me that the woman's decision to punch the rape joker fits this criteria. No court in the United States would ever provide any kind of consequence for comments like this man made. To me, the welfare gain from the chilling effect (amplified by publicity) this will have on such jokes, which quickly turn a fun night out into a stomah-clenching affair that can haunt a woman for months (if not years), clearly outweigh one guy having a sore nose for a couple weeks. 

Similarly, in Ghana, legal remedies are a poor deterrent to crime. (Especially legal remedies that are actually legal, NOT including police beating suspsects.) Even in clear cases where suspects are caught red-handed with eyewitnesses present, criminals often walk free. Two friends of mine were beaten with bricks a couple years ago, and one of the assailants was captured at the scene of the crime, but was released after several months of courtroom circus.  Death by beating, however horrific, not only provides a clear disincentive to crime but permanently removes a criminal from society.

So can vigilante justice be morally and efficiently right? Yes. Does that mean it should go unpunished? No. The fact that vigilante justice can have a high cost for falsely accused victims means that executioners must have a disincentive to engage in vigilante justice unless they are sure of its benefitcs. Punishing, or potentially punishing, perpetrators of vigilante justice provides that disincentive. So the woman who punched the joker should be potentially liable for damages, and those who engage in mob justice beyond that necessary for defense should be liable under assault laws, to prevent vigilantes from risking targeting those not truly guilty. 

This may seem like a rather extreme position; it is certainly influenced by living in situations where law enforcement seems ineffectual. However, the truth is, society engages in vigilante justice all the time. In the U.S., however, it is usually psychological, not physical: when someone cuts in line, they are often shamed, but not punched. Vigilante justice is an appropriate response to those inevitable situations where the law has no reach, but how it is exercised must be limited by disincentives against its most extreme forms.
I often fantasize about plots to confound muggers and highway robbers. (I know, this probably suggests I need to go on more dates.) One of my favorite—and less disturbing—plots involves buying a knock off designer purse, filling it with rocks, and prancing around with it at Gumani intersection after dark.  I take particular pleasure in picturing how the inevitable purse snatcher will stagger under the bag’s surprising weight, and the look of dismay on his face when he opens his hard-won prize.

This paper by Cormac Hurley of Microsoft suggests this might not be such a bad idea.  The paper models the economic decision made by an attacker, taking into account the cost of initiating an attack, the payout if successful, the total number of viable targets, and the density of viable targets among the general population.  The model shows that to make attacks economical, ability to identify a subpopulation with a high density of viable targets is key.  

The paper’s findings have important implications for how we might approach crime reduction strategies in Tamale.  Generally, one can try to deter crime by increasing the cost of engaging in it, or by reducing the benefits. 

Increasing cost: You can do this in a few ways. First, you can increase the penalty if the attacker is caught.  You can also increase the likelihood that the attacker is caught.  Either way, you increase the “expected”, or average, penalty for engaging in crime.  (The rarity of catching criminals in Tamale is the main reason I am sympathetic the high penalties imposed in vigilante justice here.) You could also do this by increasing the cost of attempting an attack regardless of penalty, say by forcing attackers to sustain injury in order to get a payout.

Reducing benefits: You can reduce the benefits of attacks by reducing the frequency with which criminals get a payout (reducing the number of viable victims), or by reducing the amount of the payout if the attack is successful.

While this may all seem obvious, the Hurley paper’s model sheds some light on the efficiency of different approaches. The model suggests that reducing the frequency with which attackers get a payout can have a much larger effect than increasing the cost of crime or reducing the amount of the payout.  For instance, in one example in the paper, if an attacker can successfully identify viable victims 99% of the time, and the ratio of the payout to cost is 100, the attacker will end up attacking 32% of viable victims. If you reduce the payout to cost ratio to 20 (you could do this by reducing the payout by 80% or by increasing costs by 400%), then the attacker will attack about 15% of viable victims. However, if you reduce the attacker’s ability to accurately identify victims from 99% to 95%, the attacker will only attack 4% of viable victims.

The challenge with crime in Tamale is that the cost of committing it is pretty low, due to lackadaisical policing, and the ease with which attackers can identify viable victims: white people out late are almost always a viable target.  Increasing police vigilance will be difficult to effect, especially without risk of repercussions such as curfews and travel warnings that might lead organizations to avoid Tamale. The good news is that a strategy we can more easily implement—decreasing the percent of Tamale expats who are out late and are actually viable victims—is likely to be a much for effective strategy anyway.

So what can you do? Make sure you are not a viable victim. Don’t carry anything that could be a payout for an attacker.  Don’t do things that make you an easy target, like get drunk or carry bags that are easy to take away.  Travel in groups, and in areas where help will come quickly.

You have probably realized that a corollary to this is that increasing the total number of people who are out late, and are not viable victims, could be a very effective strategy to deterring crime.  However, in the case of Tamale muggings, being a non-viable victim still carries risks, and I am not advocating that anyone go out at night with the goal of being a “false positive” in an attempt to reduce general crime.  But if you see me at Gumani junction with a fancy bag that looks awfully full, you’ll know what’s up.  

After my purse was stolen in Burkina Faso, I called Wells Fargo to get my cards cancelled and order new ones. The hotel staff told me I could use the hotel front desk phone, since my phone was stolen and my money supply was limited.

I had to call two different lines to cancel and re-order my debit and credit cards. The debit card was fast to cancel, but they didn't offer the option of mailing the new card to me in Africa. The credit card was also quick to cancel, with one snag. 

When you first call the line for lost credit cards, the automated system asks for your credit card number. I didn't know it. The system gives no immediate option for this. I stayed on the line dumbly for several minutes before the automated voice informed me I could say "I don't know." Given that this is the line for lost cards, it seems like they should mention this sooner. 

I was able to cancel the card quickly, and the agent then mentioned i could have the card sent to an emergency address, for a cost of $50 for international mail. Considering the alternative of having no access to money, this seemed like a great deal. Then the trouble began.

I was put on hold for extended periods, despite telling the agent I was calling from a country with no toll-free line. At some point, while on hold the hotel staff started yelling at me in French that I was taking too long on the phone. I struggled to explain in frazzled, broken french, that I was on hold. When I did get taken off hold, I was told that I would have to talk to a manager to get the card sent to a different address.  On hold again, with the hotel concierge poking me with his pen. 

When I did talk to the manager, she told me that she was going to ask me some security questions, and if I didn't know the answers, that was okay, I could just say "I don't know" and she would ask another. She started with a couple I didn't know, and some I did. I didn't know my last card transaction amount-- it was a cash advance in cedis, and I didn't know the ATM fee or the exchange rate that would have applied. After I failed to give her the first eight base pairs of the DNA in my grandmother's 11th chromosome, she informed me that she couldn't change the mailing address because I failed to get enough of the questions right. 

I was livid. Why didn't she tell me this before? Why did she say it was "okay" to say "I don't know"? If I had known that I needed to get a certain percent correct, I could have looked up my last card transaction, or ordered a gene sequencing test. By now the hotel concierge was seriously skewering me with his pen, so I told the manager that I could not stay on the line but that I was very frustrated with how this call went. I hoped that the call was indeed being recorded to monitor quality. 

A week later, I called again. This time, when I mentioned that I was calling from abroad on a non-toll free line, I was directed to the collect line. I was able to get the card sent to the emergency address by answering one security question. I was put on hold, but I didn't mind, since Wells Fargo was paying for it. That is how a customer who is alone in a strange country and has just lost her access to funds should be treated. I don't know what went so terribly wrong the first time. 

I tried to write to Wells Fargo to tell them this story, and suggest that they change their automated call response; instruct agents to either not put international callers on hold, or give them the collect number; and inform people when failing a security question will result in not getting a needed service. It turns out, I also need to suggest that they allow more characters in their email contact form. So Wells Fargo, if you are reading this, that is why this is online for everyone to see instead of in your inbox. 

Also, dear concierge: I stole your pen and traded it to a small boy for a sticker of Qaddafi. 

You should not ever experience the Kintampo Police Escort Caravan. It only happens in deep night, and long-distance travel at night is generally unadvisable. I shouldn’t have had to experience it either, but because of a series of inconveniences, beginning with a driver and ticket seller lying to me about the destination of the car I bought a ticket for, my travel took about 10 hours longer than expected.

The experience of the Kintampo Police Escort Caravan begins with waking up to find that your bus (or in my case, trotro) has stopped at the side of the road in the middle of the bush, in Kintampo—the district in Ghana most notorious for armed robberies on the road.  The first-timers wonder why we are making sitting ducks of ourselves, before noticing the other vehicles also pulled over. The veterans hop out to buy jollof rice and jugs of palm oil from the vendors capitalizing on the situation.

The Kintampo Road, 2am
After waiting 30 minutes to an hour, the police arrive with the caravan coming from the other direction. Everyone scrabbles to load back into their cars.  Everyone starts off, and the ensuing scene looks less like an orderly security convoy than it does Mario Kart: with traffic going one direction, the cars drive all over the road, the faster ones overtaking the slower.  The impression is only strengthened by the peppy hiplife music emanating from the trotros.  As an overloaded trotro hit bumps in the road, it would not be that farfetched to imagine plantains flying off its roof and landing in the path of an NGO truck driven by Yoshi, causing it to go into a cartoonish tailspin.

The police leave the cars to travel through Kintampo township on their own. Once on the other side of Kintampo, the bus, trucks, trotros and cars gather to wait for a second escort.  I have never heard of a car being robbed while in the convoy.  The convoy also likely increases safety be reducing the risk of head-on collisions at night, since traffic only moves one way at a time.  It adds about 2 hours to the 5 hour drive, though.  By the end of the night, I was very familiar with the Dagbani phrase “Ti chema!”—“Let’s go!”