Exchange rates

Michael Berry looks at New Zealand's exchange rate in this weekend's Christchurch Press. I provided a few comments; he's quoted me accurately. I'm copying below what I'd sent Michael, not because he's gotten anything wrong, but rather because it seems a waste to lose the bits Michael didn't use. The paragraphs answer questions about which he'd asked me to comment.

“The exchange rate reflects a complicated mix of foreign demand for New Zealand assets and exports, domestic demand for foreign goods and assets, domestic savings rates, and trader expectations about what will be happening with real asset returns in New Zealand relative to other countries. Trying to push it around without thinking hard about the reasons that it’s currently high isn’t without risk. If there are structural problems in the economy that, if fixed, would reduce the exchange rate, that would be a good thing – but mostly because those problems are worth addressing regardless of the exchange rate. For example, we rely on capital imported from abroad because a reasonable proportion of domestic savings are invested in housing. Fixing land use policy to reduce the cost of housing would free up some capital for domestic investment, reduce demand for foreign capital, and help push down the dollar.”

“A lower dollar could help some exporters in the short to medium term, but we have to remember the mechanism by which this works. A low dollar helps exporters by reducing the real wage they pay to their employees. At the same time, it increases the cost of machines and equipment that manufacturers import – our manufacturers losing access to those goods hurts us. In the medium to longer term, wages, in New Zealand Dollar terms, are bid up. And then exporters suffer again the next time that the exchange rate rises and they’re stuck with a wage bill that’s high compared to their export earnings.”

“There is absolutely no good reason for New Zealand to be considering quantitative easing. Quantitative easing is a policy that you try when you’ve reduced nominal interest rates to zero, have indicated that long term rates will remain at zero until inflation expectations come back up, and still have inflation outcomes that are at or below the bottom of the Reserve Bank’s target band. We are not in that world. If we tried it now, with so many other options still available to RBNZ if needed, our central bank’s international credibility would be completely shot. It would be like cutting off your hand because you had a hangnail.”

“A pegged currency also is a last-resort kind of policy. If different countries are affected by different economic shocks, independent currencies give economies ways of easing those shocks. A pegged exchange rate effectively means that you’ve given up having an independent monetary policy. That’s not a bad idea if you have a terrible central bank and you’ve not otherwise been able to establish central bank credibility, but it’s a particularly bad idea for a small open economy subject to idiosyncratic resource-price shocks and with a decent central bank.”

“The Tobin Tax is another of those ideas that sound good on paper – stick it to the speculators! – but risk being pretty awful in practice. The transactions tax is meant to reduce currency volatility by requiring that the expected returns from any trade be higher before anyone make the trade. But, there is reasonable argument that this kind of tax instead can work to increase volatility – it opens up the bid-ask spread on a currency and thins out trading markets. Further, New Zealand Dollar exchange rate movements seem to be on longer cycles than we might expect would be smoothed by a transactions tax. Even if this kind of tax reduced volatility, it would seem likely to do more to reduce intraday volatility than the longer term increases and decreases in the dollar that New Zealand experiences. Those persistent swings seem more likely to reflect fundamentals.”

“Fundamentally, New Zealand has a high exchange rate because we’re an attractive place for foreign investors to put their money. Our relative lack of domestic savings in things other than housing means that the returns on other kinds of investment here are relatively high. Because we have decided to impose very tight limits on urban growth, preventing our cities from either increasing in density or expanding at the fringes, property prices have been something of a one-way bet. So it isn’t surprising that Kiwis choose to put a lot of their savings into housing. Changing land use policy so that households could choose to put a bigger portion of their savings into the real economy would reduce our need for foreign capital and would help reduce pressure on the dollar. It’s a policy worth doing for its own sake, and if you want a lower dollar, it would help to work to that end.”

“The most important thing that the government can do in the next year is start fixing land use policy. Current policy hurts young people trying to get into their first homes and consequently helps encourage them to move overseas; it also embeds a lot of fragility into our cities in case of disaster. Imagine what Christchurch would be like today if, after the earthquake, developers had quickly been able to get a pile of new subdivisions up on the edges of town. Instead, we’re only now seeing consents issued for developers to start building. Christchurch home owners are not even allowed to build a secondary flat with a kitchen into existing homes except under regulations that make it uneconomical to do so; letting them do that would have been one of the quickest ways of getting new housing supply into the market after the earthquakes. Instead, it was forbidden.”

Investing to please and anger others [updated]

I love Amihai Glazer's little model claiming that voters are motivated by the desire to anger or please others through their vote. The model explains a few anomalies that are otherwise hard to explain, like that voters pay more attention to national election contests than local ones even though the probability of decisiveness in the latter more than makes up for the smaller stakes.

Expressiveness considerations of this sort should not matter in investment markets; expressiveness there gets expensive. And yet we have ethical investment funds for investors who care about that sort of thing. And, yet we have this one.

Moa Beer has been carving out a little niche in getting free publicity by being outrageously blokeishly offensive. When Wellington proposed putting up a Wellywood sign (tacky in the extreme), Moa promised a beer bounty to anybody who knocked it over. They sponsored the Great Easter Bunny Hunt. Gay rights groups haven't been pleased with some of their marketing campaigns. And they've been a bit cute around ASA guidelines regarding swearing in ads. And the Pakistani cricket team isn't pleased either.

Moa's now heading for IPO. Here's their prospectus. If you're running an investments course, this is one to keep on file. They've advertising for Beretta shotguns inside the prospectus, among other things. The Herald has some highlights:

The 132 page blurb channels a mix of the Mad Men TV series and up-market men's magazine FHM, featuring black and white shots of Moa's all-male board and management striking macho poses in sharply cut suits, while attractive women in black skirts, white blouses and ties fondle cigars and bottles of Moa.

A section headlined "Investment Highlights" appears beside one such full page photograph, while in another, Moa general manager Gareth Hughes taps out a cigar into an ashtray a model is holding above her head.

Advertisements in the prospectus include plugs for Aston Martin sports cars and Beretta rifles, while a naked woman on a horse promotes Ecoya scented candles, which Ross floated on the NZX in late 2010.

The prospectus seems designed for those who get expressive benefits from their investment decisions. So, then: will expressive bias against them in the IPO from those who are angered by their strategy be outweighed by expressive bias in their favour?

I'm not considering investing. The University sends a hefty chunk of my pay cheque over to the NZ University Superfund for my retirement; I don't try picking stocks. But if you were a value investor, would you expect that expressive investors would unduly bid up the price, and so you'd be inclined to stay away, or would you expect that IPO-avoidance by ethical types would make this a nice play? I'd lean towards the former.*

I expect that Moa will wind up doing well. Their beer is pretty decent and they're pretty shameless. That's not a bad niche. A small brand can afford to have half the world hate it so long as it gets a few people who love it.

Update: The most obvious model of what's going on is that they're trying to select for investors who like the current marketing strategy so the shareholders won't get mad when they do outrageous things in future. It'll be interesting to see what happens at IPO.

Update 13 November: Their IPO was oversubscribed. Good for them!

* Not investment advice; I've only skimmed the prospectus; form your own judgments, etc.

The paper and the rhetoric

From today's New Zealand Medical Journal:

The University of Otago's John Langley and and Pauline Gulliver find that alcohol was "involved" in 23.6% of non-fatal assaults by males; 13.1% of non-fatal assaults by females. Table 1 shows assaults to be relatively concentrated among Maori (31.7% of male assaults, 47.8% of female assaults; Maori represent about 15% of the population by the 2001 Census) and among the poorest cohorts - the authors note decile 9 males have assault rates 26 times the average in decile 1. There is no econometric analysis allowing for the disentanglement of ethnicity and income effects; Maori are more heavily represented among the poorest cohorts. Note further that "involvement" here also refers to the victim, not necessarily the perpetrator, unless the victim and the perpetrator co-identify (two people in a fight, for example).

We also have no clue about base rate alcohol consumption. If assault is concentrated among the poorest, and among Maori, if we were to take a random snapshot of that cohort during the day and ask "Have you had alcohol recently?", we don't know what the answer would be. What is the rate of alcohol involvement in other non-violent activities among the at-risk cohort: alcohol-involved rugby-watching, alcohol-involved friendly gatherings, and so on. I'd expect it would be higher than the population average, but that's just a hunch. If a high proportion of activities undertaken in this population group could have the "alcohol involvement" tag added, it's hard to draw causal inferences about the effects of alcohol on assaults from simply observing that between a fifth and a quarter of incidents involve alcohol.

In noting the study's limitations, the authors write:

We have no way of knowing whether alcohol involvement played a casual role in the assaultive injuries described in this paper. It could well be, for example, that many of the males just happened to be drinking at the time. On the other hand there is evidence from elsewhere that being intoxicated increases one’s risk of becoming a victim of assault.⁹ Alcohol misuse is also a distal risk factor for assault, where a pattern of heavy episodic drinking has been associated with a threefold increase in victimisation involving drinking.¹⁰

And today's headline: "Booze is fuelling rise in serious assaults" . From the newspaper:

A rise in serious assaults among young Maori and Pasifika males to ''disturbingly high levels'' has researchers calling for the Government to take action over alcohol.

Emeritus Professor John Langley, of the University of Otago's Injury Prevention Research Unit, says assaults are rising steadily and blames the Government for their failure to curb alcohol consumption.

A study published today in the New Zealand Medical Journal, co-authored by Langley and Dr Pauline Gulliver, examined the 8006 serious non-fatal assaults between 2000-2009 and is the first comprehensive statistical analysis of this category of violence in over two decades.

 ...

Excessive alcohol consumption was a common factor in the assault incidents, which frequently featured young to middle-age males beating up similarly aged males, Langley said.

''One of the key messages from the paper is 80 per cent of the problem is males. Most of those males are being assaulted by other males, and guess what the fuel is: it's alcohol.''

He said the Government needed to make alcohol harder to buy, and said their lack of action to date had been ''inexcusable''.

The article continues in that vein, quoting the study's authors drawing an awfully long policy bow from their paper's findings.

Things that definitely are NOT in the NZMJ paper:

• Any assertion of causality regarding alcohol in this data;
• Any discussion of time trends in alcohol's involvement in assaults. They give time trends on the rate of assaults, but nothing on time trends for alcohol's involvement;
• Any reason to believe that alcohol consumption has increased among the at-risk cohort over the period.

Here's the last data I'd seen on hazardous drinking prevalence sorted by ethnic group:

Table H6.2 Age-standardised potentially hazardous drinking prevalence rate (%), for adult drinkers, by ethnic group and sex, 1996/1997, 2002/2003, 2006/2007

	European/Other	Māori	Pacific peoples	Asian	Total 15+
1996/1997
Male	31.0	46.1	48.2	11.6	30.9
Female	12.0	30.6	20.8	5.1	13.3
Total	21.6	38.3	38.1	9.4	22.3
2002/2003
Male	29.9	42.4	44.1	11.5	30.6
Female	13.3	24.1	24.3	4.8	14.2
Total	21.7	32.9	36.1	8.6	22.5
2006/2007
Male	32.1	46.8	46.6	12.9	31.2
Female	14.5	28.5	25.8	3.8	14.7
Total	23.1	37.5	37.7	8.9	22.9

Source: Ministry of Health
Notes: (1) Rates are age-standardised using the WHO world population. (2) People who reported more than one ethnic group are counted once in each group reported.

The prevalence of potentially hazardous drinking dropped from 1996 to 2002 among Maori before rising again to roughly its 1996 level by 2006. The NZMJ paper has serious assaults by bodily force (ie a fight without weapons) sharply increasing from 2003 onwards, so that's at least not inconsistent with the alcohol hypothesis. But look at the rates: prevalence of hazardous drinking went from 32.9% to 37.5% among Maori over the period (with much much smaller increases for other groups). A fourteen percent increase. Serious assaults using bodily force more than doubled (other categories were flat).

Further, where the rate of potentially hazardous drinking is pretty high for some cohorts, it would be more surprising to find an absence of alcohol use. What's the rate of alcohol involvement in watching rugby within that cohort? How can we say anything about causality or draw strong policy conclusions?

I guess the NZMJ piece is a convenient hook, but it sure doesn't look like it provides much scientific basis for the policy conclusions they're hanging on it. And I fully expect it'll be cited often as evidence favouring cracking down on alcohol.

Plain Packaging and Evidence

Otago's Public Health department and British American Tobacco seem to be talking past each other.

A few days ago, the usual bunch at Otago put out a press release about some surveys they conducted about tobacco labelling. The article is here; I've copied their main results table below.

This is the sum total of the paper, really. They ran the survey, these are the results. Again, I remain pretty amazed at how easy it is to get published in some outlets. Nothing on demographic covariates; nothing on education/smoking interactions on views... hard to see a there there.

What does the survey tell us?

1. Both smokers and non-smokers think there should be warning labels on tobacco products, though non-smokers are more likely to agree;
2. Non-smokers think attractive packaging encourages smoking initiation; smokers disagree. 
• Crampton's editorialising: Presumably smokers are more likely to know what got them started. 
3. Smokers do not think that tobacco packaging encourages smokers to try new brands; non-smokers disagree.
• Crampton's editorialising: Branded packs should encourage smokers to stay with their chosen brand - they're an effort to build customer loyalty in what's otherwise pretty easily a homogeneous commodity product.
4. Non-smokers are more likely to agree that plain packaging discourages smoking initiation; smokers just as strongly disagree.
• Crampton's editorialising: Again, why do we think that smokers would be wrong on this one? They remember what got them started on smoking.
5. Non-smokers lean slightly towards agreeing that plain packaging would encourage smokers to quit; smokers disagree pretty strongly.
6. Non-smokers think plain-packaging is fair; smokers are split.

The second survey, according to the press release, tells us that there are identities around brands. That isn't surprising; it's the point of branding.

Nothing in either survey really tells us what's likely to happen if plain packaging is imposed. Here, I put a lot more weight on basic theory than on surveys. And theory tells us that once you pull brands out, cigarettes become more homogeneous. Smokers then become more cost sensitive across brands and are more likely to shift downwards towards cheaper brands. When that happens, smoking rates among smokers may increase; Clarke, linked above, doesn't think the price shifts will be substantial enough to have large price effects - he reckons it can be offset by tax increases in any case. But neither he nor I really know the scale of likely effects. And Janet Hoek doesn't know either. 

The Otago Daily Times reports on BAT and Hoek's dispute. Recall that Hoek's evidence above is entirely survey stuff.

A University of Otago researcher has hit back at comments by a tobacco industry spokesman who said plain packaging would not reduce tobacco consumption.

On Wednesday British American Tobacco (BAT) general manager Steve Rush responded to the findings of Otago University's Aspire2025 research group, which found that two-thirds of respondents supported plain packaging. Mr Rush said the research did not prove plain packaging would reduce tobacco consumption.

"The research is attitudinal and is based on opinions. It's about how people think they might behave in the future. It's not about how people actually behave," he said.

Tobacco packaging did not influence why people started smoking or quitted, he said.

Prof Janet Hoek, from university's marketing department, hit back yesterday, saying Mr Rush's remarks about her team's research were "illogical, unsupported" and ignored a "well-established evidence base".

"He has trotted out the same tired arguments in a failed attempt to disguise the fact that there is very strong public support for plain packaging, and very little sympathy for the arguments his company is investing hundreds of thousands of dollars trying to promote.

She says people like plain packaging. He says there's no evidence that it'll work. She says he's ignoring the evidence that people like plain packaging. What evidence there is in the Hoek paper suggests, at least to me, that the cohort of people who are smokers don't think that plain packaging will do much to reduce uptake or to encourage quitting; I don't put much confidence in survey measures of this sort, but if I had to bet, I'd bet on the estimates from the folks who are smokers rather than the ones from the folks who aren't. And the rest of the international literature on which Hoek would draw suffers from the problem that, because nobody's ever tried plain packaging, we can't really tell what would happen.

I worry that most of the support for plain packaging comes less from potential effects on smoking rates but rather from this kind of sentiment:

@ericcrampton who cares if it works or not - any handicap put on tobacco companies is good in my books @metiria @dpfdpf @danylmc @keith_ng
— Stuart Maxwell (@stuartm) October 11, 2012

It's stuff like this that makes Baby Pareto cry.

Nobody really knows what the effects of plain packaging will be. If we and the Australians cared about finding out whether plain packaging had any effect, we'd be setting up a randomised control trial across the two countries such that some areas would get plain packages and others wouldn't. Instead, both countries seem to be bundling plain packaging with a host of other anti-smoking measures that will make it pretty hard to sort out later on whether any change in smoking rates was due to plain packaging, changes in tax rates, time trends, or something else.

I'm less than convinced by the argument that plain packaging encourages smuggling and black market cigarettes: you can always add anti-counterfeiting tax-stamp holograms to the packs to make it obvious which ones are legit.* But I can see a mechanism where black market cigarettes are encouraged: breaking brand loyalty and making smokers re-evaluate their consumption choice may lead some smokers to choose black market products as there will be less product differentiation. I have a hard time seeing other mechanisms driving it - it has to come from smokers being less reluctant to switch to black market products. If we start seeing divergences between excise counts on tobacco sales and self-reports of smoking behaviours in the households surveys, we just won't be able to tell whether increased use of illicit tobacco is due to plain packaging or the big set of price hikes that will hit the legal market.

Harry Clarke, who supports plain packaging, suggests that one of the main potential benefits of its adoption in Oz and NZ is figuring out whether the policy does anything:

This conclusion does not mean PP legislation will have strong effects in reducing demands. This depends on the dissuasive impact of PP on consumption. But it is very unlikely that such policies would be counterproductive in the sense of increasing demands. A weak argument endorsing Australia’s adoption of PP policies is that they will have some positive effect in reducing demand without causing any harmful effects. An important gain is determining whether such policies will be very effective or not. Such innovative policies provide a “global public good”. There is often a Prisoners’ Dilemma case against delivering such public goods but, because the policies are likely to be non-disadvantageous, there are no extra costs if Australia is the first country to try out such policies. Indeed Australia has advantages over other developed countries in performing such an experiment. It is remote and physically isolated compared with European, developed Asian or even North American nations. This reduces problems of illegal cigarettes and private importing. Possibly only New Zealand is better as a laboratory. Altruistic provision of global information on the effectiveness of PP should provide benefits in reducing smoking but, at worst, will cost Australia nothing and boost its image as a good global citizen.

But it sure will be hard to run that policy evaluation. We'll likely be able to say "the bundle of plain packaging and substantial excise hikes and other stuff led to X", but disentangling effects would require that there's reasonable ex ante price variation across Australian states so we'd have different proportional changes in total prices with the excise increases.

I'd also disagree pretty strongly with the "will cost Australia nothing" conclusion. Even if you just hate tobacco companies and count their utility negatively, every policy of this sort that's adopted increases the chances that it'll be adopted for other products. The UK Health Select Committee just a couple of months ago was talking about considering plain packaging for alcohol, while promising that they'll not consider it for food. I'm still nervous for Count Chocula. The policy buys some fights on trade policy that could well be pointless. And, it's another step down the road that says that even if we provide all the relevant information to consumers, consumers are incompetent to make their own choices and public health experts should get to choose for them. I don't like that road; every step on it is costly if you like consumer sovereignty.

* I generally concur [that's too strong as I've not gone through it in enough depth. Let's say "generally sympathetic to"] with the points Philip Morris makes in its submission. But the points on black market and plain packaging really seem to be an argument in favour of putting anti-counterfeiting measures on unbranded products. They really bring out the heavyweights on it: James Heckman is their consultant; he cites his Nobel on causal inference among his qualifications to be an expert. His report on youth smoking, the appendix to the submission, deserves its own post sometime soon.

Stadium subsidies - a new hypothesis?

Football is for Republicans, baseball for Democrats? 

StubHub is the secondary market for sports and concert tickets in the States. Their data analyst has been playing with the numbers and finds that states that have a lot more trade in football tickets than in baseball tickets break GOP while those with a higher proportion of baseball tickets like the Democrats.

Here's the infographic from StubHub:

They also put up some evidence showing that changes in the BFR correlate with changes in polling data in the swing states in the leadup to the 2008 election. 

So do baseball or football sales really cause states to move politically? No, not really. Just because the two are correlated doesn’t mean that one action caused the other. But it shows that sports and politics may have a deeper connection than previously thought. Political pundits sometimes say “As Ohio goes, so goes the nation”. I don’t think they’ll ever say “As BFR rankings go, so goes the nation”, but they’d be right if they did.

If the culture wars are a long game, it would be fun to check whether Republican governors/mayors are more likely to throw money at football stadiums while Democrats fund baseball stadiums. We'd then expect periods of unified state/local government to coincide with greater sports stadium subsidies and divided government to hinder it. You'd have to run the latter kinds of checks because if a Democrat state were also a baseball state, it would be no surprise if Democrat governors put money into baseball stadiums. Focus on swing states and on unified vs divided government.

Economists tend to say that stadiums are an incredibly wasteful public investment; do Republicans cite economists against baseball stadiums proposed by Democrats and vice versa?

It is the time of year when I have to start thinking of honours projects to propose for next year....

Starter Wages

The hyperbole around National's "Starter Wage" legislation has been pretty amazing, given how limited the Bill's provisions really are.

Recall that, when Labour abolished the Youth Minimum Wage, which allowed employers to pay 16-17 year olds 80% of the adult minimum wage, they kept a "New Entrant" Wage. A young worker could be paid a lower minimum wage in the first 200 hours or first three months of employment. Employers found it intractable: you needed to figure out an applicant's entire work history on the New Entrant Wage to see for how long the employee would be eligible for the New Entrant Wage if you hired him. So employer uptake was very limited.

The new bill lengthens the period of applicability from three months to six and extend it to 18 and 19 year old workers coming off of benefits and into work. As advertised a year ago. National is hardly enacting Sir Roger's proposed reinstatement of the full lower youth minimum wage.

The Mana Party apparently thinks that the lower youth minimum wage is slave labour. Maybe some employment arrangements with lower paid workers fail the conditions for euvoluntary exchange, but they're still at least voluntary.

I fielded a few calls yesterday from radio stations looking for comment on the legislation. I noted that the legislation seems pretty much in line with what was promised a year ago, that's it's far more a rejigging of the prior New Entrant Wage than a reinstatement of full youth minimum wages, that the potential harms are rather exaggerated, but that the gains from the policy aren't likely to be spectacularly large either. More employers will take it up as it is an improvement over the prior New Entrant provisions. In response to the worry that employers might use the provision to churn through cheap workers every six months, I noted this to be fairly unlikely where employees go through any kind of on-the-job learning: the wage savings would quickly be eroded by the new worker's lower productivity as compared to the worker who's already figured out how to do the job.

Whether you love minimum wages or hate them, it's pretty hard to get too excited about the new starter wage. I think it's a small step in the right direction.

Difference in Difference and the Alcohol Purchase Age: A Stillman Preview

Reducing the alcohol purchase age to 18 really doesn't seem to have increased youth drinking, or at least not on the aggregate statistics. I couldn't see any case for worsened outcomes on the aggregate trends. Steve Stillman kindly forwards the tables from a paper he's now finalising using proper difference-in-difference and regression discontinuity design techniques to look at the effect of the alcohol purchase age change. 

Difference-in-difference method compares changes in a control group's behaviour with those in the treatment group. The counterfactual then is that changes in drinking among youths would have mirrored changes in drinking among 22-23 year olds but for that the alcohol purchase age dropped to 18.

In Table 3, below, Stillman reports difference-in-difference results on drinking behaviours. And I'll apologise here for pasting in an image of Stillman's tables; when he has the full paper up, I'll link it.

In the models with statistical controls for income, ethnicity, deprivation index and the like, Stillman finds evidence of relative reductions in drinking among younger cohorts. Fifteen to seventeen year olds reduce their binge drinking rates, relative to the change among 22-23 year olds, and also reduce their likelihood of suffering alcohol-related injury.  By 2006/2007, that cohort also was consuming fewer drinks per drinking session and had reductions in the log of total consumption, again relative to consumption among 22-23 year olds. 

Looking at Part B of the table, which reports the results from specifications controlling for other covariates, we find some evidence of improved outcomes among 15-17 year olds and no evidence of worsened outcomes. 

The binge drinking rate among 15-17 year olds was 13.3% in 1996/1997. Steve kindly, and patiently, corrected my initial misreading of the results: the -0.15 coefficient suggests a 15 percentage point reduction from 13.3 rather than a 15% reduction. I had incorrectly assumed that the tables presented transformed results as the fifteen percentage point reduction seemed implausibly large - binge drinking rates obviously cannot be negative. The results then tell us that binge drinking must have increased among the 22-23 year old control group. 

Similarly, the rate of alcohol-related injury dropped substantially relative to that experienced among 22-23 year olds. 12.7% of kids in that age cohort reported ever having had an injury subsequent to drinking in 1996/1997. The 8.8 percentage point reduction in the injury report rate would seem to have dropped reported injuries to a third of their rate prior to the change in the purchase age but, again, the counterfactual here is that the injury rate among 15-17 year olds would have increased in line with the increase among 22-23 year olds. 

Relative to the counterfactual that 15-17 year olds' results would have mirrored those among 22-23 year olds, the difference-in-difference estimates tell us that 

more than 7,300 fewer kids presented with alcohol related injuries than we would have otherwise expected and that more than 12,000 fewer kids participated in binge drinking.

The model cannot distinguish between changes caused by the reduction in the purchase age and anything else that might have simultaneously affected binge drinking rates among kids but not among the comparison cohort, or that might have caused young adults in the 22-23 year old reference group to start binge drinking. But it’s hard to point to other policies that would have had the age-specific effects.

Results on reductions in drinks-per-session are also substantial: a one drink reduction where the mean otherwise was just over three. The rest of the table shows no significant effects of the reduction in the alcohol purchase age: no increase in the likelihood of drinking or in the frequency of drinking. I couldn't see anything in the aggregate stats suggesting increases in youth drinking after the reduction in the purchase age; Stillman can't find it using difference-in-difference either.

Doug Sellman had suggested that binge drinking rates could mask what's going on in the upper tail of that distribution: those who binge drink might be consuming a lot more harmfully consequent to the change. I suppose we can't rule it out, but per-session consumption and the log of total consumption are both down; the maths then require that group is fairly small or that everyone else's consumption is down by more than is suggested above.

Stillman finds evidence of increased hospital admissions for alcohol use disorder. But, the effects are not large. Figure 1, below, uses regression discontinuity design (RDD) with a break at the time of the law change. RDD isn't possible for binge drinking because we there only have three waves of annual data. The advantage of RDD over difference-in-difference is that we can look at the period right around the change in the purchase age, making it unlikely that other changes have affected outcomes. But, if it takes a while for behaviours to change subsequent to policy changes, RDD will have a harder time picking that up.

Yup, there's a break at the time of the law change. The biggest effect was among 18-19 year olds. Alcohol-related hospital admissions rose from about 6 per 100,000 to about 10 per 100,000 per month. There were about 55,000 kids in that age cohort at the time of the law change. So the RDD design suggests that the law change resulted in perhaps two additional 18-19 year olds per month showing up for treatment. It's a relatively large proportional change, so regressions will show large percentage effects. But again, the base rate is tiny. In difference-in-difference estimation, the largest estimated effect was a 49.6% increase in hospital admissions related to an alcohol use disorder among 18-19 year olds three years after the law change. That corresponds to about three people per year.

What about traffic accidents? Stillman finds a small initial decrease in alcohol-related car accidents among 18-19 year olds after the purchase age declined, followed by an increase. 

The regression discontinuity design finds no significant effect. When Stillman runs the difference-in-difference estimates, he finds no effect of the change in alcohol purchase age on alcohol-related traffic accidents in models with a full range of demographic controls that allowed for age-specific time trends. A model with a more limited set of control variables suggested no effects on traffic accidents until three years after the law change. At that point, accident rates increased by 13.7% among 15-17 year olds and by 25.3% among 18-19 year olds. While the percentage increase sounds large, again, the base rate is tiny: the estimates correspond to an increase from 0.5 to 0.57 car accidents per month for 15-17 year olds, or an increase of about one car accident per year. For 18 and 19 year olds, we went from 1.2 accidents per month before the change to 1.46 afterwards: 3.5 additional accidents per year. And, again, those effects disappear with the addition of age-specific time trends, which might just matter once we get to three or more years from the time of the policy change.

In short, it’s pretty hard to build a case for substantially worsened outcomes consequent to the drop in the alcohol purchase age; doing so requires relying on estimates without the full set of desired controls and putting a lot of weight on a pretty small absolute change in the number of accidents. Further, since negative effects on accidents really only seemed evident from three years subsequent to the alcohol purchase age, we have to worry whether anything else happened that differentially affected youth versus adult drink driving rates. And, the results from difference-in-difference estimation suggested pretty substantial reductions in youth binge drinking subsequent to the reduction in the alcohol purchase age, relative to the older comparison cohort.

We are very lucky that clever herestheticians prevented the drinking age from rising. It would have imposed substantial costs on 18-20 year olds for no real benefit.