Question of the Day: Should We Just Let Murderers Do Their Thing?

A reader named Mark Kozel writes to say:

I heard that Chicago will be pouring up to $14 million into police overtime to prevent murder and violent crime.

It got me thinking: is it cheaper to prevent this kind of crime, or to just let it happen and clean up the mess afterwards?

It would be hard to find many people, even economists, who would arguing that “just letting it happen” isn’t an outcome that society should even think about accepting.

On the other hand, this is the kind of question that economists do tend to pose. It also reminded me of a recent Marginal Revolution blog post that mentioned John Broome‘s new book, Climate Matters: Ethics in a Warming World, and highlighted this excerpt:

What is the role of experts in democracy?…Their views, supported by arguments and evidence, help individuals and their representatives to form judgments.

This is not how economists typically see their democratic role.  They do not see themselves as participants in public deliberation, helping people to make their judgments.  Instead, they think their role is to help ensure that the preferences of the people prevail.  They do this by basing their valuations on market prices, which reflect people’s preferences.

It strikes me that Broome is obviously correct in writing that economists pay a lot of attention, maybe even too much, to people’s preferences; but it strikes me as obviously incorrect that “they do not see themselves as participants in public deliberation.” In fact, the world seems absolutely brimming with activist economists of one stripe or another.

Anyway: what do you make of Mark Kozel‘s question?

Addendum: This post was written and scheduled before we became aware of the recent Colorado shooting; we regret the timing.


The "mess afterwards" is more than just bodies, it's the loss of expectation and reliance on others, especially the state, to watch each others' backs. That kind of social capital is difficult to value accurately and to regain fully, e.g. failed states.


We are better off with fewer homicides and other violent crimes, unless you place a very low value on human life and health.

It's not plausible that cleaning up the mess would change behavior in a way that would reduce violent crimes.


If the police aren't paid to stop crime, individuals will start paying to stop it themselves. This would probably cost more than funding the police department, not work as well, result in more innocent people getting shot, etc. Law enforcement definitely seems to be one of those areas where government is more efficient than the free market.

As to the question of should we just let crime happen, it already does. There's not much police can do to prevent a person intent on murdering someone from doing so, especially with the ease with which someone can get a gun. If we were serious about actually preventing crime, rather than just jailing criminals after the fact, we'd push for having better inner city schools.

Simon K

Curious. A related point, in some way, is prison policy: should we be rough-handed, intending to punish criminals, or should we be easy on convicts, attempting to rehabilitate them? As Scandinavian countries tend to find, it is in fact on an economic level cheaper, in the long run, to splurge on rehabilitation programmes and comfortable prison existences - e.g. Norway's Halden - so as to not further alienate and embitter people who have committed crime. Criminals who have been treated such tend not to fall back into crime and become productive citizens instead.

Of course, it does very much bother you on a philosophical level: the victim of a crime may very well leave the court room back to a more difficult life, whereas the criminal will go to a state-afforded hotel. You want to feel like people get what's coming to them, in some way. But pragmatically, it makes no sense to do as the people would prefer.


Herman Brodie

I guess there would be a lot more crime once we decide to just 'let it happen', so it is not possible to estimate those costs

David Bley

I don't believe that experts serve any role in the functioning of the US government for several reasons. Typically, for every expert who espouses a certain opinion, there is another with a contrary opinion. Our government operates reactively rather than proactively. When the government only reacts to things after it is clear that something went wrong, predictions of experts don't serve any purpose. Lawmakers will not get re-elected following experts recommendations. Some justify the existance of lobbyists as part of our legislative process by maintaining that the lobbyists bring expertise to that process. I say that their expertise is tainted to the point that it is not objective science.


If it comes back to cost per life then things like making driving safer, or preventive medicine are the better option. There is also the issue that while maybe not contributing to a large reduction in murders the number of police officer deaths could rise. The money could also be used for education as to reduce poverty and the lower the associated risk of drug abuse. 61.2 percent of murders (FBI stat) get solved so it would be interesting to see if cleaning up the rest would help society better off then spending more to limit the initial murder rate.


Well, the lower limit on "letting it happen" is a simple back of the napkin calculation. Insurance companies tend to value a human life at ~10x their annual income. Median income in the US is ~$30k. So for a person killed, society loses ~$300k of value. So if the $14mil poured into police overtime saves >40 lives, it's worth it. That of course only counts loss of productivity by the person who got killed. If you consider other factors, like time spent by police investigating, loss of productivity by the victim's family members, etc, far fewer lives than that need to be saved for the investment to be worth it. Let's call it 20. In 2012 so far there have been 286 homicides, so to be effective, over the rest of the year, the $14mil needs to lower homicide rate by 7%.

Ben Wheeler

Prevention takes many forms. There is not only prevention in terms of policing, but earlier prevention in terms of social investment. A study I believe in the 80s by the rand corporation found that dollar for dollar, you stop a ton more crimes with investment in education and health care and job training then you do in policing or prisons.

Tim White

As witnessed by today's shooting here in Colorado, the economic impact of people being afraid to go outside, to the theater, to malls, or schools is significantly more impactful than $14M.


People aren't going to express their "preferences" if they live in fear of a lawless society. Just look around the world and see how capitalism thrives when there's no law and order.

People need to feel safe in order to exercise their choices. Otherwise the market is going to take a back seat to security.

Interesting, uh timing on the question, guys.


It's hard to put an amount on it, but yes, there is some value of cost where it's less efficient to spend more to reduce murder rates. The key is in where the money would go otherwise. If spending the money on crime fighting saves X lives, but spending the same amount of money on some other activity saves more than X+20 lives, then choosing to prevent the crime-related murders is actually just like choosing to kill 20 people.

And of course, it's very difficult to know what would have happened otherwise. I can imagine an argument that the money should be spent on schooling, to reduce the future murder rates, rather than the present rates. Or should simply be not taxed, and the corresponding economic growth saving the lives of people on the margin for health/lifestyle related decisions. And when you start bring in future people, since not taxing the money or spending it on education will change people's wealth, which impacts how many kids they'll have and the quality of those kid's lives, it becomes even more complex and hard to quantify.

And then there's the weird stuff, like how spending it on preventing murders might actually save fewer lives than an alternative use now, but because people might not understand that and murders are noisy data while a small drop in average education is quiet data, people will react differently to how it gets spent, and their reactions will have long term consequences also. It's like how we spend money on homeland security preventing almost no deaths, instead of on lightning rods which would prevent many deaths, but because we do, people and businesses feel safer to live, grow, and move to here, possibly outweighing the inefficiency of the expense (though I think that one case is unlikely, tbh).



I understand there is a fairly solid but very unpopular argument that to save the maximum number of lives in the case of terrorist attacks, the best place to invest money is in first responders and not in prevention. I expect this idea wins or fails on the length of time you judge it over, classified data and so on, which make it almost impossible to judge. That said, if we abandon the rather false either/or question and ask the more productive "to what extent" question about the balance of spending between either one, we could prompt an interesting examination of whether funding goes where it's most effective or too much towards the areas where its ineffective but popularly supported. Might be some prevention funds ought to be nudged over to the clean-up column of the budget. It's the security equivalent of your analysis into why soccer goalkeepers don't just stay in the centre of the goal during penalty shootouts. Better to be seen to be decisive and active than mistaken for passive and weak, right?



Morality aside, tolerating murder would prove costly to society. Murder itself is generally not a productive activity, at least economically speaking, but what is more costly is the threat of murder. The threat of murder raises the cost of engaging in any economically productive activity because an individual would have to account for the possibility of being murdered in every transaction whether its the fear of provoking murder or merely stepping outside your door and exposing yourself to greater risk of murder. I think the real question is what is the optimal level of murder? In other words, where does the the marginal costs of prevention outweigh the benefits. $14 M worth of overtime probably isn't worth it if it will only prevent one murder (from an economic standpoint); however, if it were to stop 50 murders, it might be.


"Cheaper"? It almost implies that you're just looking at pure dollar terms. No offense guys, but why do you continuously assume that monetary value is the only value worth considering? It's not. Most of the murders in Chicago probably occur in the city's poorest neighborhoods and thus I imagine you wouldn't see a major drop0ff in economic output from each death. We should not, however, engage in such morbid calculi. If you view welfare more broadly than it's much more difficult to weigh $14 million against the value people receive from security and the cost of grief. To be sure, there's always a tipping point and we can never exist in a risk-less world. Where we find the balance between cost and broader social welfare is the fascinating debate you refer to at the end of your post.

When we talk about exerting preferences through the democratic process we begin to delve in other social sciences. Personally, I don't think the people in crime ridden neighborhoods have as much of a say in the democratic process as those with more money and influence. In that case, maybe we are under-funding crime prevention.

Finally, I just want to point out that we see more reductions in crime from added police presence than from harsh sentences. Criminals don't rationally weigh likelihood with severity to reach an expected value. Instead, they overweight likelihood and underweight severity (I am extrapolating a bit from prior studies). Maybe we can reap some savings in our system by shortening prison sentences and increasing police presence.



@A that calculation breaks down for people with negative income, such as benefit recipients.


What we need is data! How effective are the preventative measures? What is the cost of cleaning up afterwards (investigations, prosecutions, prison, costs associated with the victim and family)? Then we can put a dollar sign on it and then you have your answer.

But then it gets tricky, as Mr. Dubner suggests, in measuring that dollar sign against our values as a society. It seems to me that this is a question is one that our society is very uncomfortable answering and therefore rarely gets asked. And maybe it should, just so that we're being honest with ourselves. For instance, when we as a nation go to war, there are serious dollar signs attached. If we were to ask this question, maybe it would illuminate some of the real costs, such as in lives lost, and create a picture of where the money flows. Could this enlighten us as to who pays the price and who stands to benefit from the decision to go to war? Also, might it also turn up some interesting relationships between the decision makers and those that stand to benefit the most?

Measuring dollar signs against human life is perhaps the most difficult thing a society has to do. Everybody likes to think that each life is priceless, but in practice these types of decisions are being made. Since we can't discuss it, then the decision makers can determine whether or not these things are worth it and their decisions will thus be one not supported by data.

At the same time, the day we come up with an exact monetary value for a life, is the day we lose a very important part of our humanity.



Here's another (repugnant) idea. Pour 14 million into a massive marketing campaign to ask pregnant women in neighborhoods where these violent crimes tend to happen, and just use some economist tested ways (guilting, agreeing to do what the masses does, etc) and just ask those women to abort their babies.

This is also a repugnant idea. But if Levitt's study is to be believed (and I do), then aborting these future criminals before they are born is way cheaper than letting people be shot or paying for overtime.


It already happens to an extent, although not to the extreme to let it all happen and prevent nothing. The police prevent the crime they can given the budget that they have.

If you were given the options of:

1) Let it happen and pay 0% tax.
2) Prevent some murders, have a reasonable police force and pay 20% tax
3) Have a near zero murder rate, have a brutal, intruvise police force, pay 60% tax.

Most people, in my opinion would choose option 2, because the cost of option 3 (both in terms of money and liberty) is too high vs the risk of being murdered.


While there may be some short-term savings, in the long-term, higher murder and general crime rates in the city will simply drive people and businesses out. The eventual and significant loss of revenue won't even come close to what I presume is a small, at best, savings at the start of such a scenario.