Three Short Essays on Eric Garner


If you asked me to make the best possible argument in favor of the police action that led to the death of Eric Garner, it would go like this:

  1. Cigarettes are taxed.
  2. You can’t have taxes without enforcement. In this case, the enforcers are the police.
  3. Where there are enforcers, there will be confrontations.
  4. When sellers refuse to cooperate, the enforcers have only two options: Walk away, or resort to violence.
  5. Enforcers who walk away soon lose their credibility and their effectiveness. This is more than doubly important for a police officer, who needs that credibility when he confronts far more dangerous criminals.
  6. Therefore, we cannot fault the police for resorting to violence.
  7. Violence is sometimes catastrophic. That’s sad, but it’s not news.

If you asked me to make the best possible counterargument, it would go like this:

  1. You could say exactly the same thing about a protection racket.

That is, every protection racket needs an enforcer. When shopowners don’t pay up, the enforcer has only two options: Walk away or resort to violence. To walk away would sacrifice credibility. Therefore we cannot fault the enforcer for resorting to violence. Sometimes violence gets pretty messy. So it goes.

The force of that reductio ad absurdum depends on the analogy between taxation of cigarettes and the demand for protection money. I think that reasonable people can disagree about the depth of that analogy.

But the lesson remains that every law must occasionally be enforced through potentially catastrophic violence, or, to put this more succinctly, all legislation is deadly. Violence is part of the cost of making laws, and it’s a cost the makers of new laws would be well advised to contemplate.


Why, exactly, did the police try to take Eric Garner into custody in the first place? Did he refuse to comply with some order? If so, what order? To hand over his stock of untaxed cigarettes? To move to another corner? Or was it their intention to take him into custody right from the moment they arrived?

I’ve watched the same video everyone else has, and I can’t find the answer to this question — because there’s a clear break in the video somewhere around 1:18, and the answer appears to be hidden in that missing footage.

Prior to that break, we see Garner vigorously protesting police harassment — but we don’t know the nature of that harassment, because we don’t actually see the police making any demands. After the break, they’re suddenly trying to put cuffs on him. What happened in between?

I can imagine many possible answers. Maybe they intended to cuff him all along and were just waiting for backup, which arrived during the break in filming. Or maybe they intended to issue some sort of warning, but turned nasty when they thought Garner disrespected them. Or maybe Garner became far more belligerent during that break in filming.

Presumably the grand jury heard ample testimony about all this. Possibly there is additional video that hasn’t surfaced on the web — or that I’ve failed to notice. Perhaps this has all been asked and answered elsewhere, and I’ve just missed it. But from the video I’ve seen, I feel like it’s quite impossible to know the full story here.


Suppose you are a typical street vendor of an illegal product, such as, oh, say, untaxed cigarettes.

Suppose the police make a habit of harassing such vendors, by confiscating their products, smacking them around, hauling them off to jail, and perhaps occasionally killing a few.

I have good news: The police can’t hurt you.

Oh, I know it seems like they’re hurting you — it’s no fun to get smacked around, go to jail, and even risk death in an arrest-gone-horribly-wrong. But it’s no fun for your competitors either, which keeps those competitors from multiplying and helps to maintain your profits. So harassment has both costs and benefits — and they have to just about exactly cancel out.

Here’s why: Street vending can never be substantially more rewarding than, say, carwashing. If it were, car washers would become street vendors, driving down profits until the rewards are equalized. If car washers were happier than street vendors, we’d see the same process in reverse. (The key observation here is that it’s very easy to move back and forth between street vending and other occupations that require little in the way of special training or special skills.)

Because police harassment of street vendors has no effect on the happiness of car washers, and because street vendors are always just as happy as car washers, it follows that police harassment has no effect on the happiness of street vendors.

In a sense, you can think of the police as your employees, working to keep your competition down. Like any employees, they extract a cost, but you’re getting what you pay for.

Of course, if becoming a street vendor means signing on for a one-in-a-million chance of getting killed and you turn out to be that one-in-a-million, you’re a huge loser. But a million others who signed on for the same chance and didn’t get killed are, collectively, a huge winner. As in any risky occupation, some get lucky and some get unlucky — but also as in any risky occupation, you’re fully compensated in advance for your risks.

So if you’re a street vendor, the police can’t hurt you. On the other hand, when the police go around putting people in deadly chokeholds, they’re clearly hurting someone. So the question is: Who?

Answer: Not the vendors, but their customers. Fewer vendors means higher prices. That hurts consumers, and the sum total of that harm adds up to the harm that we see in the viral videos.

There might also be a second class of losers, namely crime victims who get inadequate protection because the police are out busting cigarette vendors instead of, say, purse thieves or rapists. [I used the word "might" because somebody might want to counterargue that the best way to prevent major crimes is to crack down on minor crimes, perhaps even including those that shouldn't have been crimes in the first place.]

Just as the War on Drugs is a war not on sellers but on users, and just as the War on Prostitution is a war not on prostitutes but on their customers, so the War on Cigarette Vending is a war not on vendors but on smokers. That makes it no less a war, and no less destructive.


55 Responses to “Three Short Essays on Eric Garner”

  1. 1 1 David Sloan

    “Because police harassment of street vendors has no effect on the happiness of car washers [...]”

    Sure it does; by your own reasoning it temporarily makes carwashing “substantially more rewarding than” street vending, driving street vendors to become carwashers, thus “driving down profits [for carwashers] until the rewards are equalized”.

    The corollary to “street vending can never [in the long run] be substantially more rewarding than, say, carwashing” is not “police harassment doesn’t hurt street vendors”. It’s “police harassment hurts, say, carwashers too”. This generalizes to “police harassment hurts everyone”.

  2. 2 2 Keshav Srinivasan

    Steve, I don’t follow your logic when you say “On the other hand, when the police go around putting people in deadly chokeholds, they’re clearly hurting someone. So the question is: Who?” Why is it that they’re clearly hurting someone? Before carrying out the rest of your analysis, in terms of costs to consumers and the like, how can you tell that the policy’s actions are imposing ex ante costs on someone? (Clearly they impose ex post costs on some people, but it’s less clear to me that they’d impose ex ante costs on anyone.)

  3. 3 3 Bennett Haselton

    I think the flaw in Essay I is the assumption that the number of police confrontations is proportional to the number of laws. This sounds superficially plausible, but it rests on the assumption that without cigarette taxes (and without the opportunity make money by skirting them), people like Eric Garner would have instead made their living at a regular job.

    I think it’s far more likely that for low-level non-violent criminals who don’t care about getting arrested (Garner had already been arrested 30 times), they would just find some other hustle or resort to shoplifting. All of which could just as easily lead to a confrontation with police.

    Besides, the premise fails an empirical test. Every developed country (I assume) has cigarette taxes, and lots of other laws which libertarians oppose, but the United States has far higher than average rates of incarceration and of police killing of civilians.

  4. 4 4 Jim W K

    Steve, have you considered that the jury (and maybe generally wider in American society) still harbours (perhaps subliminally) a series of prejudices – pro-white, pro-authority – that played out in the trial, culminating in the acquittal of an over-zealous white male police officer strangling a black man to death with an unnecessary and fatally dangerous choke hold?

  5. 5 5 Harold

    “But the lesson remains that every law must occasionally be enforced through potentially catastrophic violence.” This may be true, but there is a lot we can do to prevent that potential from becoming reality. There are many contributing factors that determine whether the potentially deadly confrontation ends up being deadly. “Therefore, we cannot fault the police for resorting to violence.” We can fault the police for resorting to certain sorts of violence in certain situations. Otherwise we would throw our hands up if the police routinely gun down unarmed child shoplifters. It makes little sense to say we cannot prevent all violence, so we must accept all violence.

    So for me, the best possible case does not end with a necessity for some violence. The actual violence used must be justified also. The best case for the police thus also concludes that the force used was reasonable, necessary and proportionate. The Grand Jury apparently believed it was. The video you mention in essay 2 has led many to believe that it wasn’t.

    If laws against violence are not enforced, it may be that we end up with more violence, not less. So a failure to enforce limits on the violence used in enforcement will predictably lead to more of that action.

  6. 6 6 Steve Landsburg

    David Sloan: Great point, but not, I think, convincing. When you drive street vendors out of street vending, some become car washers, some become ditch diggers, some become shoe shiners. So that even when the number of street vendors goes down substantially, the number of people in any one of those alternative occupations goes up just a little, so there’s a negligible impact on rewards to each alternative occupation.

  7. 7 7 Steve Landsburg


    The actual violence used must be justified also. The best case for the police thus also concludes that the force used was reasonable, necessary and proportionate. The Grand Jury apparently believed it was. The video you mention in essay 2 has led many to believe that it wasn’t.

    What alternative (lower) level of violence do you believe would have allowed the police to effect an arrest?

  8. 8 8 Steve Landsburg

    Keshav Srinivisan:

    Why is it that they’re clearly hurting someone?

    Are you seriously suggesting that there’s no social cost to killing people?

  9. 9 9 CMOT

    Garner had just been freed from a 6 month jail term for selling loose cigarettes. He then immediately went back to selling loose cigarettes in a way (in front of a store that was sure to complain about him cutting into their business) that was guaranteed to cause another arrest.

    I think it’s safe to assume that his reasons were not economic in the raditional sense, in that the satisfaction he got from selling 75 cent smokes must have meant far more to him than 75 cents obtained in any other way.

    So a lot of your analysis simply does not apply in this case.

    BONUS points: Garner was never in a chokehold and was not choked during the takedown. He likely died as a result of positional asphyxia, resulting from being left in a prone position will handcuffed.

  10. 10 10 James Kahn

    #9 CMOT I have read in more than one place (for example, that the he died of cardiac arrest on the way to the hospital. Which is not to say that the stress of the confrontation wasn’t a contributing factor.

    Steve, great points in all three essays. I think the protection racket analogy is weak because I believe one can show that they are more costly than governments due to the lack of coordination (turf wars, people being subject to potentially large numbers of rackets, etc.). I think this was part of Buchanan and Tullock’s theory of government. But the “final lesson” of essay I is still correct.

  11. 11 11 khodge

    #10 Kahn: I think that the protection racket analogy is spot on because (1) for a protection racket to work they need to establish an effective monopoly and (2) for government to work it needs to protect its (legally recognized) monopoly.

    Having not followed the case I cannot really speak to the specific issue but police departments follow rules e.g. in my part of the woods the conditions for an allowable police chase are supposed to be limited. The fault lies with the government if the established rules require arresting people for either selling cigarettes or being obnoxious to police officers.

  12. 12 12 Ken B

    Great post, and great comments 1 and 6.
    This is just another incarnation of Steve’s tax incidence arguments.

  13. 13 13 Ken B

    I want, more to see Steve’s response than as a rebuttal, to question his comment 6.
    Your counter in 6 seems to rely on a different analogy, water level perhaps in a narrow stream and a wide lake. Is that the right water analogy here? Think instead of two cylinders filled with fluid, one narrow, one wide, that are joined. Apply pressure to the end of the narrow one. What do you feel at the far end of the wide one? People don’t wash cars for fun, they feel a need to do even a dull job for money. “Pressure” seems relevant here.

  14. 14 14 Major.Freedom


    You didn’t take into account the happiness/gain/utility that comes not from just the payoffs(ends) of street vending and car washing, but the utility from the constitutive ends of street vending and car washing itself.

    Given the fact that with a choice between street vending and car washing, we see some people choosing street vending and other people choosing car washing, it is not correct that both employments have equal, even risk adjusted, payoffs.

    Utility is based on the individual.

    Even if the dollar denominated (risk adjusted) return is the same for both street vending and car washing, perhaps I personally would still choose street vending over car washing, or car washing over street vending, because I personally enjoy doing one or the other. Perhaps I derive a particular enjoyment from selling cigarettes without taxes because I am anti-state politically. Can’t get that from car washing! Does this mean that I welcome and consent to being beaten up by the cops? Of course not. I would rather, if I had the choice, live in a world without any state, and take my chances with hoodlums who would have a much more difficult time acquiring resources in the street, than from the tax revenues of an entire economy.

    It is not correct to make the blanket statement that that cops harassing street vendors “has no effect” on the happiness of car washers. I know I *am* made unhappier compared to the counter-factual, when there are cops running around beating street vendors up. I am not just a “liberty for myself” libertarian.

    It would not be correct to say that because I could just walk away from street vending and wash cars instead, that my happiness is improved, or the same as before. If I had to leave street vending because of risk of being beaten up and imprisoned, then I would not pretend my liberty has not been infringed upon as I sulk to the car wash to work, and pretend that my happiness is unchanged. I will feel bad. I already feel bad because I can’t sell what I want to sell without such risks. I would be happier if that risk of being beaten up isn’t there, and I would not necessarily be made worse off just because my profit margins were lower because of legalization. Money payoffs is not the only source of utility gain. Humans don’t engage in the division of labor only because of profit margins. They do so because not only are we diverse in terms of happiness in particular payoffs, but also happiness in terms of particular work.

    I do feel unhappier relative to the counter-factual, with the fact that I know those engaging in perfectly peaceful activity on the streets are being harassed and beaten and thrown in prisons by the state. How can you say my happiness is not affected? We may work in a division of labor, but that doesn’t mean we can’t feel compassion and empathy for others, and feel sadness when our fellow humans suffer.

    It is also not correct to say that street vendors are “just as happy” as car washers. That kind of a statement burns an Austrian’s eyes. One cannot compare inter-personal utilities. Each person brings forth a unique, incomparable subjective value scale. A ranked order of one person’s values cannot be compared to a ranked order of another person’s values such that one is “higher” or “lower”, for that presumes an external objective standard for these things, which does not exist. I am not “happier” than you, nor are you “happier” than me. My happiness is qualitatively different than yours. We may have the same dopamine running through our veins, but the whole process that makes you and I who we are, is not the same. You can only be happier or unhappier relative to what you yourself desire and feel. Same with myself.

    Finally, I cannot but help write in all caps my response to this:

    “On the other hand, when the police go around putting people in deadly chokeholds, they’re clearly hurting someone. So the question is: Who?”


    I think the fundamental flaw in your essays is that you are setting enforcement as the standard from which we move around and dodge and pretend we’re no worse off by shifting from one job to another, rather than holding enforcement up to a standard that actually determines people’s happiness.

    If I want to street vend, and there are cops beating people up, myself included, then I am being hurt, because what I want, which is peaceful selling, is being violently oppressed.

    It isn’t all about the final payoff. MEANS are supremely important, and constitute sources of utility as well.

  15. 15 15 Keshav Srinivasan

    Steve, I can understand there being ex post social costs to killing people. But why must there be ex ante social costs?

  16. 16 16 Dmitry

    I agree with Keshav’s comment #2. Mr. Landsburg, your response in 8 is misleading: you also don’t discuss the costs of one individual death. If ex ante information about the risk of being killed is known to everybody, nothing will change when a death occurs. One vendor goes, another comes into his place (or doesn’t), prices don’t change. You could make a point that if some car washer enters the market instead of the dead vendor, then cars that won’t be washed constitute opportunity cost of death, but you didn’t go in that direction.

  17. 17 17 James Kahn

    #11 Khodge But that’s the point: It’s very costly for protection rackets to establish monopolies. They can’t enforce contracts, either with their “clients” or with their competitors. I suppose you could argue the difference is just of degree, and that the Garner arrest was an example of a turf war, but there’s something about constitutions, elections, “consent of the governed,” that makes governments qualitatively different.

  18. 18 18 David Wallin

    r/14 “Given the fact that with a choice between street vending and car washing, we see some people choosing street vending and other people choosing car washing, it is not correct that both employments have equal, even risk adjusted, payoffs.”
    I am quite sure that when Steve says “Street vending can never be substantially more rewarding than, say, carwashing,” he is note talking about risk-adjusted net pay, but how rewarding it is. That takes into account payoffs, risks, and the personal preferences you seem to think he left out. Thus we can suggest that members of the armed forces find their jobs at least as rewarding as a civilian counterpart because they have revealed that preference. Nobody is arguing that the risk-adjusted monetary compensation in the military is anywhere near that of civilian equivalents for the majority of the men and women who serve. They seem to find some non-monetary benefits to serving. And thanks for that.

  19. 19 19 khodge

    #17 Kahn: I actually would argue that it was very costly for government to establish its monopoly. Perhaps that is one reason for the fact that laws – good, bad, effective, useless, counterproductive – never go away; it’s just part of the “legal” protection racket that can never be tampered with.

  20. 20 20 Brian

    “Are you seriously suggesting that there’s no social cost to killing people?”


    As Keshav points out, it’s not clear that there are additional ex ante social costs. Isn’t it possible that those social costs are already factored into the decision to be a street vendor of illegal items? After all, most of the social costs are borne by those who interact most frequently with Mr. Garner, in particular his family, associates, and customers. Given the clear risks in engaging in street vending (as indicated by his many arrests), wouldn’t those in his social circle have put pressure on him to stop so as to avoid the social costs? Wouldn’t his customers have stopped buying? Since these things did not happen, they had clearly already factored in the social costs and still preferred the street vending activity.

    OR, if they did not prefer the activity itself, it might still be the case that costs of making him stop outweighed the benefit of having him stop. From that perspective, the police did them all a favor by doing for Garner’s social circle what was too costly to do themselves. In any case, it’s not clear that there are social costs not already accounted for.

  21. 21 21 GabbyD

    “every law must occasionally be enforced through potentially catastrophic violence”

    i dont understand this. if the law is enforced through catastrophic violence, even if only on occassion, then either:
    1)law is wrong and should b changed, or
    2) there are better enforcement methods, or… etc…

    another way to put it, arent there MORE steps to take before inevitable catastrophic violence?

  22. 22 22 Ken B

    Keshav & Brian,
    A virus mutates creating a deadly disease. No ex ante social cost?
    Oddly enough this disease, by some miracle, kills only those breaking some cigarette sales laws who would resist a cop if he were there enforcing it, when he isn’t.

  23. 23 23 David Sloan

    Steve Landsburg (#6): “When you drive street vendors out of street vending, some become car washers, some become ditch diggers, some become shoe shiners. So that even when the number of street vendors goes down substantially, the number of people in any one of those alternative occupations goes up just a little, so there’s a negligible impact on rewards to each alternative occupation.”

    “Negligible”, and yet enough to (in the long run) make each of those other professions no better than street vending under the threat of police harassment.

    As you’ve said, your entire argument about one profession not being more significantly attractive than another only holds when workers can easily move between professions. However, if you introduce police harassment to the workplace conditions of street vending, street vendors cannot easily move between today’s profession (street vending under the threat of police harassment) and yesterday’s profession (street vending without the threat of police harassment). In the absence of a way to move from one profession to the other, your argument that they must end up equally attractive (which they must, in order to claim that police harassment is not harmful to street vendors) falls apart.

  24. 24 24 Ben

    @Prof L, David Sloan is right. When the police hurt a street vendor the hurt is spread to street vendors and other occupations which they could easily migrate to, as well as to the customers. Is the aggregate cost to such occupations really negligible compared to the aggregate cost to the customers?

  25. 25 25 Scott H.

    “I have good news: The police can’t hurt you.”

    This is wrong. You are assuming a uniform police response to illegal cigarette sales. But what if the police decide to dramatically increase their enforcement against the sellers? This is what seems to have been taking place with Garner. I’ve seen it written that NYC’s top officer had recently made it a priority to better enforce laws against illegal cigarettes.

    This “changing of the rules” could explain Garner’s protests (and, frankly, miscalculations), as well as explain how illegal cigarette sellers in general felt hurt by a new and not-necessarily-optimal-for-them, risk/reward ratio for selling illegal cigs.

  26. 26 26 James Kahn

    #23,24 Seems to me you both are arguing that a tax increase (in the form of increased enforcement) is costly. But increasing any tax (with uninteresting exceptions) is costly. The issue is whether this form of taxation is more costly than others, holding fixed revenues. It might be, but simply pointing out costs fails to make that case.

  27. 27 27 David Sloan

    #26 No, I’m trying to refute Steve Landsburg’s claim that there is *no* cost to harassed street vendors.

  28. 28 28 Eric

    Steve (#6),

    I agree with David Sloan. It depends on elasticities (just as it does for tax incidence). Police violence makes Street Vendors temporarily worse off, so they move out into other professions, driving the comparable utility downward. It could be that elasticities of supply and demand pass this through to consumers. It could also be, as you say, that the external impact is “negligible” if the number of people in external occupations is very large. In that case, though, the harm is the “negligible” per person harm multiplied by the “very large” number of people harmed, giving a decent-sized real harm to “producers”.

    Another point is that, by your argument, producers are not harmed in expectation (risk-adjustments and so on…). On the other hand, the *specific* individual in the choke-hold is harmed. It’s the same as saying that since stockholders are not on average harmed relative to money market investors, I wasn’t harmed when the stock market went down today. Clearly, though, I *was* harmed ex-post.

  29. 29 29 Steve Landsburg

    David Sloan (#1 and #23) and Ben (#24) and Eric (#28):

    My claim is that with constant costs (i.e. no special skills or training) and free entry/exit, a tax on one industry (or equivalently the harassment of one industry) has an arbitrarily small effect on sellers as the number of industries gets large. I stand by that claim.

    I could write out a very general theory, but let’s stick for now with a perfectly representative example. If you think the example is somehow contrived, I’ll be glad to see your equally detailed counter-example.

    Suppose there are N suppliers distributed across K industries. In any given industry, one supplier supplies one unit of output. The demand curve in any given industry is P=A-Q.

    Initially it’s easy to see that in equilibrium the suppliers distribute themselves equally, with N/K per industry. The price of each good is therefore A-N/K, which is what each supplier earns.

    Now let’s add a tax of 1 in the first industry only. We now get X suppliers in the first industry and (N-X)/(K-1) in each of the others. The after-tax price of the first good is A-X-1 and of any other good is A-(N-X)/(K-1). In equilibrium these prices must be equal, so we can solve for X and get X=(1+N-K)/K. Sellers in the first industry — and hence in any industry — then earn A-X-1= A-(N/K) – (1/K).

    In other words, the tax costs each seller 1/K, which becomes arbitrarily small as K grows large.

  30. 30 30 Eric

    Steve (#29).

    I agree that the tax costs each seller 1/K and the cost to each seller gets negligibly small. However, since there are K industries it costs “producers” as a whole 1/K * K = 1.

    I thought you were claiming the tax harms consumers not producers. It now appears you are claiming it negligibly harms each individual producer, which may or may not say anything about how it harms consumers relative to producers. After all, if there are an infinite number of consumers who buy only infinitesimally small amounts of your product, the harm to each individual consumer is also negligible. The elasticities determine how the sum of these “negligibles” are distributed between producers and consumers.

  31. 31 31 David Sloan

    Steve Landsburg (#29): “In other words, the tax costs each seller 1/K, which becomes arbitrarily small as K grows large.”

    I don’t dispute your math up to this point, and I will assume it’s correct for the rest of my post, but I think you stopped too soon.

    If this cost of 1/K is applied to each seller, we still need to multiply by the number of sellers to get the total cost. Since there are N suppliers, the total cost is N/K.

    If K grows arbitrarily large and N remains constant, I agree that N/K becomes arbitrarily small. However, N > K (and perhaps N >> K), so K can’t grow arbitrarily large without N growing too.

    The total cost N/K is as small as possible when every seller holds a monopoly in their industry, thus N=K. In this case, the total cost is 1, the original size of the tax. As soon as any industry contains more than a single supplier, it only gets worse from there.

    (Caveat: I’m not sure how much of either your or my math makes sense when N=K; what does it mean for fractional sellers to move industries?)

  32. 32 32 Steve Landsburg


    I agree that the tax costs each seller 1/K and the cost to each seller gets negligibly small. However, since there are K industries it costs “producers” as a whole 1/K * K = 1.

    Nope. There are N sellers, not K. The total cost is N/K. Holding N fixed and letting K get large (which is exactly the right thought experiment for this issue), the total cost goes to zero.

  33. 33 33 Keshav Srinivasan

    Steve, how can K be larger than N? By the pigeonhole principle doesn’t that mean that some industries employ no one? And if you allow K to grow arbitrarily larger than N, then doesn’t that mean you’re assuming a large number of industries that employ no one? Maybe I’m missing something obvious.

    And in any case, in real life isn’t the number of sellers usually far larger than the number of industries?

  34. 34 34 Steve Landsburg

    David Sloan: As long as N/K is large (i.e. as long as there are a substantial number of suppliers in each industry — which is manifestly the case for things like car washing), we don’t need to worry about fractions. In the examples we’re considering, both N/K and K are very large, which is possible because N (the number of unskilled workers) is very large.

  35. 35 35 David Sloan

    Steve Landsburg (#32, #34, regarding my #31): you seem to be in agreement with my claim that “the total cost is N/K”, and you’ve addressed my “Caveat: I’m not sure [...]” to my satisfaction.

    It looks like you’ve skipped over the paragraphs in between, which is where I think the meat of my objection lies. Am I missing something?

  36. 36 36 Keshav Srinivasan

    Steve, in comment 34 you assume N/K is large, and in comment 32 you talk about N/K becoming arbitrarily small. Isn’t there at least a prima facie tension between those two things?

  37. 37 37 Steve Landsburg

    David Sloan (#35): We agree that the total cost is N/K. We agree that the cost to any given producer is 1/K. My claim is that if you’re a producer, a tax can’t significantly hurt you, provided K is large, which in practice it is.

    If there are, say, 5 million unskilled workers spread over 20,000 industries, then a tax of $1 per worker in one industry actually costs those workers $1/20,000 each — in other words, 1/20,000 as much as you might naively expect.

    So: I said that a tax of $1 can’t hurt you. You’re saying no; a tax of $1 actually costs you .005 cents. I think it’s reasonable to sum that up by saying “the tax doesn’t hurt you”. You think it’s important to mention that it still costs you .005 cents. That, as far as I can tell, is our only disagreement.

    The fact remains that for every $1 worth of damage the police appear to be doing to that seller, the seller is in fact suffering only .005 cents worth of damage. That, I suspect, is not obvious to a non-economist, and is well worth pointing out.

  38. 38 38 Steve Landsburg

    Keshav Srinivasan (#36):

    Steve, how can K be larger than N?

    Nobody said that it has to be. With millions of unskilled workers, K can easily be in the thousands, which means that $1 worth of harassment per seller actually causes less than a penny’s worth of damage to each seller. In other words, if you’re a seller, the police can’t significantly hurt you.

  39. 39 39 Keshav Srinivasan

    Steve, if a one dollar tax in the first industry gives rise to a total cost of N/K across all sellers industries combined, and N/K is larger than 1, then what happened to your point that the class of consumers is the one that has to pay the cost of the tax, not the class of sellers? It seems like the class of sellers as a whole is already incurring a higher cost than the actual monetary amount of the tax, so why must any of the burden of the tax be borne by the consumer?

  40. 40 40 David Sloan

    Steve Landsburg (#37, #29, original post):

    First, let me say that I turned out to have been confused earlier (and am perhaps still confused) by the units of the “tax of 1″ mentioned in #29.

    In your example, we have 5 million unskilled workers over 20,000 industries, and then we tax one industry by $1 per worker. Let me assume the industry we’re taxing has 250 workers (1/20000th of the 5 million). So, naively, a cost of $250 is imposed.

    Following your logic, instead each of the 250 workers in this industry only incurs a cost of $1/20000, but then so do the other (5 million – 250) workers in all the other industries. This puts us right back at the $250 total cost from earlier.

    When you said “I have good news: The police can’t hurt you.”, I think it would have been more fair to say “I have good news: The police aren’t hurting you as much as you might think, because in the long run the pain is shared by every other unskilled worker.”

    I see our only disagreement as whether or not splitting the pain 5 million ways makes the total pain negligible; you say yes, I say no.

    Furthermore, as someone working in an industry with a relatively high skill requirement to enter, and as someone not being harassed by the police, I’m more than a little uncomfortable with telling unskilled workers that one of their own being strangled by the cops causes them negligible pain because in the long run in hurts *all* unskilled workers.

  41. 41 41 Steve Landsburg

    David Sloan: Actually, let me amend my earlier comments — I now see and acknowledge your point. A tax of $1 per worker in industry A, which is roughly a tax of $(N/K) altogether, imposes a total cost of $(N/K) — in other words, the total damage is equal to the total visible damage. You’re right about that.

    What remains true is that the total damage *per worker* is negligibly small, so I continue to stand by the statement that if you’re one of those workers, police harassment can’t (significantly) hurt you. But again, you’re right that producers as a class feel a burden equal to the tax revenue.

    Consumers in industry A *do* feel non-negligible costs, far greater (under any plausible assumptions) than any individual producers.

    (Edited to add: I wrote this comment before I read your comment #40. Now I’m sure we’re both saying the same thing. I apologize for not getting your point quicker.)

  42. 42 42 Steve Landsburg

    Keshav: Did you read the example before responding? The tax is one dollar per worker in the first industry, i.e. approximately N/K dollars altogether. As to why any of the burden must fall on consumers, try drawing a demand curve and looking at the area under the curve down to two different prices. I think you’ll find that those areas are different. In fact, the difference in the areas is greater than the tax revenue.

  43. 43 43 Steve Landsburg

    David Sloan:

    I think we should also note the following: The total cost to producers increases linearly with the tax per worker, whereas the total cost to consumers increases as the square of the tax per worker.

    So even if you want to add up the damage to all the workers, there remains the point that as the tax gets bigger, the total damage to consumers grows much faster than the total damage to producers.

  44. 44 44 Capt. J Parker

    Re: Essay II
    My guess, and it is only a guess with one tiny bit of supporting data is that the police were called because someone reported a fight. The viral video of Mr Garner has audio of someone repeatedly saying “all he did was break up a fight” where “he” obviously referred to Mr Garner. So, from this it follows that the police intended to arrest Mr Garner because they suspected he was involved in a fight not because he was selling “loosies”. The police are motivated to tell a suspect he is being arrested for the least serious infraction they suspect has occurred in order to maximize the chance the suspect will be cooperative. So, it’s not a big stretch to imagine that the police told Mr Garner he was being arrested for the misdemeanor loosie selling when in fact their intent was to get him off the street while investigating if he may have been involved in possibly felony assault. If my guesses are correct then what the video shows is the police using force to subdue a felony assault suspect who is acting at the very least cooperatively and at worst belligerently. Now, Mr Garner did not deserve to die for breaking up a fight, being in a fight, starting a fight or belligerently telling the police to buzz off. The people responsible for his death should be held accountable. But, in addition, our system of taxes and state regulation of some forms commerce while far from optimal shouldn’t suddenly be at risk of nullification because of the claim that Mr Garner was choked to death by police for not complying with the tax laws for cigarettes. There is adequate reason to believe that was not what brought the police down on Mr Garner.

  45. 45 45 Keshav Srinivasan

    Steve, I don’t doubt that if you actually calculate the cost to consumers, you’ll find that consumers will incur a cost. What I’m asking is, before you even do such a calculation, how can you tell right off the bat that someone other than sellers must bear some of the cost? In your post you said sellers don’t incur a cost, and yet there is a social cost, so someone else must bear a cost. Now that we’ve established that the class of sellers bears a cost of N/K, doesn’t that mean that the argument that someone else must bear a cost no longer holds?

  46. 46 46 Steve Landsburg

    Keshav: Okay, fair enough — unless, of course, one knows enough general theory to know that the social cost of a tax must exceed the revenue it raises. But point taken.

  47. 47 47 Keshav Srinivasan

    Steve, so then your statement “That hurts consumers, and the sum total of that harm adds up to the harm that we see in the viral videos” is not completely accurate, right? There is also the sum total of the harm done to the class of sellers. As a practical matter, how do those sum totals compare with each other?

  48. 48 48 Harold

    “What alternative (lower) level of violence do you believe would have allowed the police to effect an arrest?”
    Letting off the pressure after he said he couldn’t breathe? Seems reasonable to me. According to wiki “After the incident, city medical examiners concluded that Garner was killed by neck compression from the apparent chokehold, along with “the compression of his chest and prone positioning during physical restraint by police.”"

    A less confrontational approach? There are all sorts of ways of affecting an arrest without killing people. The best case for the police to my mind is not that arrests must sometimes involve violence, so people are bound to get killed.

    The equivalence of car washing and street vending depends on the expectations of the traders. If the level of violence in this particular case was higher than the street vendor could reasonably expect, then he personally does bear some of the cost, since he could not have included this in his calculation of whether to wash cars instead. Possibly it matters not whether he could reasonably expect it, but whether he actually expected it.

  49. 49 49 nobody.really

    [A]ll legislation is deadly.

    Let us all remove our hats and take a moment in silent reflection, remembering the fate of our comrades and fellow countrymen who have fallen in the course of the legislatively triggered annual bloodletting known as House Ref. No. 102-68, declaring National Pickle Month.

    It’s all so senseless. With the problems of sodium sensitivity so widely known, why would anyone commit such a heinous act? Similarly, given that everyone knew that the citizens of Gnome were spoiling for any excuse for a parade, why pick February? Isn’t the frostbit deathtoll of the Black History Month parade bad enough?

    But just as Congress learned nothing from Viet Nam, they’re now preparing to rename a post office, permit electronic service of documents in federal court, any myriad other actions. Oh, the humanity….

  50. 50 50 Steve Landsburg

    Keshav: Touche.

  51. 51 51 John Becker

    Death is to marginal utility analysis what black holes are to the mathematics of relativity and quantum mechanics. The equations break down as values become infinite.

  52. 52 52 Bob Murphy

    Steve, Keshav does that to me on my blog too. Sometimes I consider washing cars.

    (Joking aside, you guys arguing this out really clarified the issue. I think Steve is making a good point but he oversold it initially in a way that would make the casual reader think that there was no point.)

  53. 53 53 Steve Landsburg

    John Becker:

    Death is to marginal utility analysis what black holes are to the mathematics of relativity and quantum mechanics. The equations break down as values become infinite.

    Fortunately, policy decisions almost never involve particular deaths; instead they involve subjecting people to various probabilities of death, which are perfectly amenable to cost-benefit analysis.

  54. 54 54 Floccina

    I am not saying the police were right but one more point needs to made:
    8. Overwhelming force minimizes collateral damage.

  55. 55 55 vikingvista

    The argument actually falls apart at 5, not 1. The credibility of a police officer or anyone else rises or falls on whether he follows discovered law, not legislation. Everyone, police or not, must ask himself before he acts, and independent of any legislation, if this is an appropriate way to treat another person.

    The credibility of legislation and legislators is judged the same way.

    E.g., a credible cop would attempt to disarm an armed man threatening to hurt innocents-irrespective of any legislation. Likewise, a credible cop would refuse to physically threaten or assault a nonthreatening person engaging in peaceful voluntary trade of his own property–again irrespective of any legislation.

    You consider legislation only to be safe, not to be good.

  1. 1 Potpourri
  2. 2 The Liberty Herald – Potpourri
  3. 3 The Liberty Herald – Corrections
Comments are currently closed.