Liquor ban: Is there a rational
Posted June 18, 2008
on:After reading this excellent post on the liquor store regulation idea on Kiwiblog (which aggregates the thinking of a number of other blog authors posts on the issue are found *,*,*,*), I’ve decided to do a little thinking out loud about the issue.
Now, to analyse what it going on we have to ask why we want to have tighter controls on liquor outlets in the first place. From what I can tell, liquor outlets aren’t the direct cause of harm – the consumption of alcohol is. As a result, these measures are based on the causal link: More liquor outlets -> more liquor consumption -> more crime.
For fun, lets take as given that more liquor consumption leads to more crime. We still need the prevalence of more liquor outlets to cause more liquor consumption for this story to float. How does this work?
Well the fact that there are more liquor outlets DOES NOT increase demand for alcohol. It does however reduce the transaction cost associated with buying alcohol. As a result, if we have more liquor outlets popping up around the place, we can expect more alcohol consumption in the areas where the outlets turn up – fair enough.
However, how do we calculate the transaction cost? Well, there is the cost of getting to the outlet (say the cost of driving there) but more importantly there is an opportunity cost associated with getting there and buying it – namely the cost of time.
As a result, if the government’s policy was to reduce the density of liquor outfits in a uniform fashion across the country, everyone would suffer a certain increase in the cost of grabbing any liquor solely from their opportunity cost of time. Now it would seem to me that people who are going to commit the sort of violent crime that this policy is reacting to (robbing liquor stores) would probably have a lower opportunity cost of time than the rest of society – implying that it will have less impact on their liquor consumption decisions in the first place.
If the policy instead focused on lowering the density of liquor stores in areas that are susceptible to this type of crime then we have to ask – what sort of reduction in liquor consumption would we see from this increase in costs. I suspect that the answer will be, a very small decrease.
This reminds me of back in the day when King Country was a “dry” area. The liquor outlets in Kihikihi had a great time of it, as people would just travel north to get their alcohol, or would brew their own. If full scale liquor bans struggled to prevent consumption back then, having to drive an extra couple of km’s to get booze now isn’t going to stop many people.
As a result, if the link between the number of liquor stores and the quantity of liquor consumption among the target group is very small, then this policy seems really quite pointless.
Although Mrs Clark stated that:
New Zealand research, soon to be published, would show outlet density was associated with binge drinking and alcohol-related harm to young people, Miss Clark said.
the correlation between these two variables in recent times does not imply causation without a believable story behind it.
The existence of liquor outlets DOES NOT create demand for liquor – it can only influence the cost of getting liquor. If we really want to reduce alcohol consumption as the result of an externality we should just tax it (although I realise this is far from a perfect solution, given the way the externality is created), rather than messing around regulations – or potentially we could do nothing at all, and realise that whatever we try to do will just make matters worse.
14 Responses to "Liquor ban: Is there a rational"

My favourite example of the “thinking outside the square…when trying to justify govt policy”: I coedited a book with Cowen on market failure. I soon after got an email from somebody in UK Trade and Investment (variant on our MED) asking me what market failure argument could be used to support some initiative they’d come up with. I still use that example when I teach market failure and how we have to watch out for market failure arguments justifying policies ’cause often folks start from the policy and work backwards rather than the other way round.


Perhaps the thinking goes that reducing the number of stores, reduces the supply of liquor, the supply curve shifts to the left, reducing quantity supplied and increasing prices. However, if we assume that its the smaller stores that shut down as a result of a law change, we can probably assume that alot of the demand goes to those larger stores, who then, also with an increase in demand at that individual store, can take advantage of economies of scale moving the supply curve back to the right; sure security guards at larger stores may stop liquor stores being robbed but liquor is the cause of a lot more crime than just robbing liquor stores. This is a pretty poor argument because it doesn’t reduce the quantity at all, yet I expect this is the underlying assumption by the government rather than an assumption about a reduction in demand.
I only have one other argument which I can’t say is particularly good, but I’m having trouble arguing against it.
Perhaps this is about a long term strategy to influence our culture away from alcohol. I agree that the ones who pay the most for the proposed policy are the ones who enjoy only the occasional alcoholic beverage. Alcoholics and heavy drinkers probably don’t care about the slight increase in transaction costs so it doesn’t stop them from drinking. However, by reducing the number of stores, this may reduce the influence the alcohol industry has on low drinkers and young drinkers thereby reducing the number of people who move from the low drinker category into this heavy drinker category. i.e. in the long run there is a reduction in the demand curve; whereas in the short term there is only a material increase in transaction costs for those who don’t place a high value on personal drinking (perhaps those who percieve a high negative externality of heavy drinking – this is another discussion altogether but in short I kind of like the idea that those who see the need for change also pay the costs for implementing that change; and those who cause the negative externality to society pay the costs of duties etc).
I guess this whole argument comes with an assumption that our heavy drinking culture in NZ comes from the high avaliability of alcohol – (This is prob more anthropology than economics but would be interesting to find out). I would assume that avaliability is probably only a small factor – prohibition, the 6 oclock swill etc, were all initiatives that reduced availability, yet I think haven’t reduced this drinking culture.


Hey that was great, I was having trouble agreeing with the argument in my head myself but I hadn’t found a specific reason.
My thoughts were that prohibition, 6 oclock swill has had little effect (if not a negative effect) previously so this should be similar for a policy that reduces the number of stores but why? You’ve answered this with the taboo argument, and that consumers will buy more per transaction, increasing binge drinking (same as the 6 oclock swill; where drinkers consume more in the shorter time avaliable). Completely fleshed it out, Cheers.


That’s the 6 o’clock swill. Years ago in NZ, pubs had to close at 6 pm so that fathers would go home and be with their families. Of course it meant drinking as much as possible between 5 and 6 and driving home drunk.


I did a quick lit search on alcohol availability and crime. There’s not much out there, but what there is does suggest a link. I don’t like how they handled endogeneity though. Gyimah-Brempong, 2001, “Alcohol availability and crime”, Southern Ec J. Uses census tract data to show that the elasticity of crime with respect to liquor outlets ranges from 0.25 to 1. But he argues that the OLS estimates are underestimates because of endogeneity and that the higher IV estimates are the right ones. I just don’t get that. If an underlying variable is causing both alcohol outlets and crime (and in the same direction), OLS ought to overestimate the effect ’cause it’s bundling in the effect of the underlying variable. So, for example, if poverty increased both # liquor outlets and crime, and we had numbers on liquor outlets but not poverty, then our liquor outlet coefficient would be biased upwards ’cause it’s also pulling in the effect of poverty. But he’s saying that OLS is downwards biased.
He does make that rather reasonable point that the full cost of purchasing alcohol includes the transactions costs of getting it, and increased availability reduces those transactions costs.
Know of anything better on trying to work out causality?


A person necessarily help to make severely articles I might state. That is the first time I frequented your website page and thus far? I amazed with the research you made to create this actual post incredible. Wonderful task!

June 18, 2008 at 3:45 pm
Here’s an argument for it that I do not support.
Specify that there are certain fixed costs involved in setting up security systems. The larger the scale of the shoppe in question, the greater will be its cash on hand and so too the optimal security system. So long as the returns to scale in security systems outweigh the increased return to robbing larger outlets, it’s better to have fewer shoppes that have armed guards rather than lots of small shoppes with security cameras.
Again, I do not think that the gains to security outweigh the costs of regulation inducing the shift to fewer larger shoppes, so I do not support this argument. But at least it’s a plausible one, unlike the ones I’ve heard from the govt thus far.
Too bad Rodney Hide and Heather Roy couldn’t coordinate amongst themselves to make sure they weren’t both out of the House at the same time.