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. 

Paying for earthquakes

Matt Nolan walks us through why quantitative easing to pay for the Christchurch Rebuild isn't particularly good policy. If we need a monetary push, the place to start is interest rates - and then only if RBNZ thinks we're going to be below the 1% lower inflation bound over the medium term. If we need a fiscal push, which is highly debatable, doing it through earthquake spending might not be as helpful as the Greens might like - especially as regulatory bottlenecks and regulatory capacity constraints in Christchurch seem to be holding things up at least as much as money.

Let's recall how standard public finance says to pay for things like earthquakes. If the status quo ex ante had about the right mix of spending and taxation, then you cut spending in non-earthquake areas, increase spending on earthquake rebuilding, perhaps slightly increase taxes in the short term but definitely take on debt to spread the rebuilding costs over the longer term. This is bog standard public finance. Paul Krugman and Steven Landsburg agree. I walked through the argument slowly a while back.

What do we get if the government prints earthquake bonds and RBNZ prints money to buy them? Here's Nolan:

Now you may believe we should fund the rebuild with a one-off tax – that’s fine, in that case get the government to put a tax in place directly (or to directly cut spending from other place).  However, taxation by stealth of this sort is likely to be worse in multiple ways:

1. We have betrayed RBNZ independence for virtually no reason … understandably a sneak tax by the RBNZ would make people less likely to believe them in the future about holding to their inflation mandate.  As a result, we run into the time-consistency issue in monetary policy again, and it will become more painful for economy when the RBNZ tries to commit to its inflation mandate again.
2. We have a relatively rough redistribution of resources due to this.  By putting in our sneak tax through QE, we transfer resources to those with assets, those doing the rebuild, and those who can easily adjust prices/wages – while hurting those on fixed income, and those who have saved.  It is an inflation tax – pure and simple – and as a result, it will initially transfer resources from those who can’t protect themselves (generally the poor) to those who can (generally the rich).  If we introduce the tax through fiscal policy instead we can sort out these distributional issues a little better.
3. A country that is willing to introduce QE as a clear fiscal transfer – when there is no monetary policy reason – will destroy its credibility with international lenders.  People will scoff at this, but such a policy will increase the level of “inflation insurance” lenders ask for – increasing the cost of credit in New Zealand.

These are obvious and true costs, that have been seen from similar policies around the world for hundreds of years.  QE really isn’t anything new, and if we want a fiscal transfer of this sort just say it (as the Greens previously have to be fair), and do it through fiscal policy – it has nothing to do with the RBNZ.

I suspect Matt's "that's fine" at the start is more recognition that the policy at least is honest rather than that it's a good idea. Borrowing for major one-off capital expenditures is far better than a one-off tax.

Matt's also a little harsher about those peddling crank economics than he's usually willing to go:

The constant banging on about the exchange rate and the RBNZ shows a fundamental misunderstanding of the “issues” NZ faces.

The Greens, and Ganesh Nana, are wrong in stating that the RBNZ has failed.  Distinctly and totally wrong.

I approve.

Bernard Hickey was calling on Twitter last week for a review of the desirability of central bank independence. Matt and I got a bit testy. This is one of those issues where re-politicising whether the central bank should be independent has almost the same effect as removing independence: if RBNZ believes that its independence is conditional not on meeting the PTA but rather on making a mob of monetary cranks happy, then it might be tempted to start skewing its policy. And even if it doesn't, if people think it might be, that also starts wrecking expectations. There's reasonable latent public demand for bad monetary policy. Feeding and encouraging the trolls is a remarkably dangerous game. It's fine to debate what optimal monetary policy is. But suggesting that because RBNZ isn't following your preferred one, we ought reconsider central bank independence... that's intellectual vandalism on par with saying we should do away with Pharmac because you didn't like the Herceptin decision. In sum:

• Good: trying to convince RBNZ of your views on monetary policy.
• Bad: whipping up the hooples to break central bank independence.

100% Romney

Here in NZ, this is a Romney.

So why are there no Kiwi entrepreneurs selling 100% Romney wool woollen mittens to the American market this election season? You could sell them sincerely to earnest and cold Midwest Republicans; you could sell them ironically to coastal hipsters. Have "100% Romney" knitted into the back of the mitten. The slogan works for both markets.

About the Romney:

The Romney is suited to high rainfall and heavy soils, and has the highest resistance to footrot of any breed in New Zealand. It grows a heavy fleece used in carpets, furnishings and knitting yarns. The Romney was the single most popular breed in New Zealand through the 20th century. It currently makes up about 68% of the national flock – over 25 million sheep.

Crumbly Cookie Problems

Economists usually frame equity-efficiency issues in terms of the "leaky bucket". Sesame Street suggests that this could better be modelled as the crumbly cookie problem.

Recall that in the leaky bucket problem, you're presented the following situation. A rich person is on the left. A poor person is on the right. You can scoop up some of the rich person's stuff in a bucket, walk across a bridge, and pour the stuff out for the poor person. But, your bucket is leaky and some of the stuff is lost as you cross the bridge - pure waste. How leaky does the bucket have to be before you say it's no longer worth attempting the transfer? There's no "right" answer to the question; it's more of a value judgment about how bad waste is relative to income differences.

In the little clip below, Ernie and Cookie Monster have one cookie between them. When Ernie has it, Cookie Monster is sad. When Cookie Monster has it, Ernie is sad. They try to share it. I take it as a cautionary tale. But I would, wouldn't I?


I don't think the deadweight costs of redistributive taxation are really 100% [best estimates have it more around 30%]. But I do like the example. How crumbly does the cookie need to be before we either say that Ernie should keep it or that it should be given to the cookie utility monster?*

Update: And here Cookie Monster takes on the problems of political economy. In the state of nature, we can find no neutral arbiter of disputes and none of us can trust ourselves to be a fair judge when our own interests are at stake. And so we give up the power of judicial arbitration to the State. Unfortunately, the deal often isn't as great as we'd have hoped. Cookie Monster here warns that Locke was too optimistic: a Panglossean best-case thinker.

 

* I'm pretty sure that I've seen Cookie Monster hit satiation before, so he can't really be a utility monster.

Package art

The supermarket aisle is a kind of an art gallery. Every product's packaging is designed to be eye-catching; some reach the level of artwork. I love how Warhol's interpretation of the Campbell's Soup Can is now featuring as a version of the Campbell's packaging; if the limited edition were up for sale in New Zealand, I'd be buying it. The packaging can add value directly: it can be beautiful. That adds very real value, no matter what the no-logo types claim.

I've worried about plain packaging pushes on tobacco likely being extended to alcohol. Recall that there's a reasonable case that Inuit print art began because indigenous artists were curious about how the guy who painted the Player's Light packages managed to get the little sailor to look the same on every pack. And alcohol packaging can be rather pretty: variations in bottle style, label images - those are worth keeping.

NBR reports that Beck's has announced a run of bottles here in New Zealand that will feature commissioned NZ art. I've shamelessly scraped their image below, but you should subscribe.

NBR writes (paid, subscribe!) that NZ artist John Reynolds, who created the labels:

...pulled lyrics from record label Arch Hill's extensive song catalogue - featuring artists such as Family Cactus, The Clean and Street Chant - as inspiration to create the 24 multi-layered pieces of artwork. 

The article notes Beck's has been doing this kind of thing since 1987.

How long 'till we see taxpayer-supported propaganda like this directed at alcohol?

Heresthetics and the alcohol purchase age

I lost about $300 on iPredict a little over a month ago. I had really strong expectations that the Government's bill proposing a split alcohol purchase age was going to pass. Standard rules of order seemed to ensure that the government's bill would pass. The amendment to keep it 18 would fail as all those preferring Split or 20 would vote it down; the same would hold in reverse for the amendment to increase the age to 20 across the board.

As we all know, Parliament decided at the last minute to change the normal procedure for this bill and this bill only. Instead of the normal rules of order, Parliament voted for the three options on first preferences. Then, the one with the fewest votes was dropped. So the Split option was dropped and we were left with the head-to-head contest between 18 and 20, won by 18. It's remarkable. Split looked like it was a Condorcet winner. And, had the "increase it to 20" group had their act together, the outcome still would have been Split: they'd have counted noses and all voted strategically for Split over 20. 

I have nothing but admiration for whatever heresthetician came up with this move. You, Sir or Madam, are deserving of some kind of award. I don't know what happened, but somebody managed to convince enough people in Parliament that the proposed rule change was actually more democratic for this kind of choice.

I've held off on blogging this one - I set it as a midterm exam question in my undergrad Public Choice class. Then things got busy. But here's the fun part.

Green MP Kevin Hague supported Keep It 18, as did I. But here's what he wrote after the vote:

In the wake of the parliamentary vote on alcohol purchase age there are some people (the ones I have seen are people who favoured 20) who are suggesting that the voting procedure used led to a perverse outcome.

I disagree completely. This is not a post about the merits or otherwise of the various options, but an argument that while the voting procedure was highly unusual, it produced a more democratic outcome than otherwise would have occurred, and should be considered more often. [emphasis added]

The first thing that I tell my public choice students about social choice is that if there's a Condorcet Winner, we generally consider it to be desirable that a social choice mechanism choose the Condorcet Winner. Here's Hague on what would have happened had we used the standard rules of order:

So recall that in this case the option that was in the Bill was the split age option, but that there were proposed amendments both to retain a uniform age of 18, or to raise it to a uniform 20. The ordinary House process would have seen first a vote on 18. If that had passed then the 20 proposal would have been ruled out of order. If it had failed we would have voted on the 20 option. If that had failed then the split option would have remained in the Bill.

...

Under the traditional voting method, what would have happened is this: the proposal for 18 would be put and fail. We would then have moved on to the 20 proposal, and this also would fail. This would leave the split age in the Bill – the option that was actually the least favoured by MPs.

I agree completely with Hague on the above. That's why I had $300 riding on it.

Instead first preferences were canvassed and the split age option, as the least favoured excluded. Then effectively those who had favoured the split age added their support to their second preference. Hence 68 votes for 18, 53 for 20.

The voting method used ensured that the result could command a majority in the House whereas the traditional method would not have done (except by excluding this whole part of the Bill I guess).

This is truly bizarre. Hague is saying:

1. There was a Condorcet winner: Split would have won against both 18 and 20.* 
2. The system we used, which ruled out the Condorcet winner on the first round, was more democratic.

Deciding what is more or less democratic is a bit of a mess because of all of the various problems in the social choice literature: we're usually trading off horrible features of various systems against each other. But it's awfully strange to prefer, on democratic grounds, that a Condorcet winner be suppressed!

I really really want to read the inside account of how they pulled this off. Think about everything that here was required:

1. Flip from the normal rules of order and convince everybody that you're doing it for fairness reasons rather than to change the outcome.
2. Do it late enough in the game that all the "Increase to 20" group don't have time to think through the game theory. If you do it early, somebody will point out to the ones too dim to figure it out on their own that they cut their own throats by supporting 20 over split at the first round, if their true preference ordering has 18 last. 20 then gets dropped in the first round and Split wins in the second. 
3. Do it despite the government seeming to prefer the split age option, which lets them to be seen to be doing something on alcohol, where there's been a lot of public pressure. 

A proper account of this could make up a chapter in a revised version of William Riker's truly excellent little book, Heresthetics: The Art of Political Manipulation. The youth wings of some of the parties in the Keep It 18 movement seem to have completely gazzumuped their elders. Heck, I'd never even considered it possible that the rules of order could be a choice variable. This was a truly entrepreneurial act in the proper sense of the word. I have nothing but admiration for these kids. A beautiful application of heresthetics. 

Even worse, I doubt that anybody responsible for it would want to take credit for it as doing so might hinder future entrepreneurial efforts. Didn't the CIA award medals anonymously to operatives? Can we do that here?

David Farrar has all the data up on who voted which way in each round. I'll have to think of a fun honours project around this.

* You can build weird cases for non-single-peaked preferences if you think a lot of 20 supporters preferred 18 over Split, maybe so that, in their view, things would get so bad that we'd be able to increase the drinking age to 20 across the board (or a similar case for those who most preferred 18). And maybe then preferences could have cycled. But I don't buy it at all. And, more importantly, Hague doesn't buy it: he's saying there was a Condorcet winner.

Contagion risk

Why bail out banks and financial companies? Capitalism is about profit and loss; if you bail out the losers, there's no end to the cost. Well, unless there's substantial risk that one bank's falling over wrecks the balance sheets of a pile of other banks, inducing them all to fall over too.

U Southern Carolina's Jean Helwege presented at Canterbury this past week. She and Gaiyan Zhang painstakingly went back through the balance sheets to examine the real contagion risk of Lehman's failure. Financial theory tells us that no firm will want to have too much of its balance sheet in any one asset. That's one of the main points of the Capital-Asset Pricing Model: you diversify away from non-systemic risks. And, regulations on banks and financial services mandate that they not have all of their investments in one place. So it wasn't all that surprising then that there really weren't any firms at strong risk of going bankrupt if their Lehman paper stopped being worth anything.

They open with Bernanke's analogy from a CBS interview:

Interviewer: “Mr. Chairman, there are so many people…who say, ‘To hell with them. They made bad bets. The wages of failure on Wall Street should be failure.” 

Bernanke: “Let me give you an analogy. If you have a neighbor who smokes in bed…If suppose he sets fire to his house, and you might say to yourself, you know ‘I’m not going to call the fire department. Let his house burn down. It’s fine with me.’ But then of course,…what if your house is made of wood? And it’s right next door to his house? What if the whole town is made of wood?...What needs to be done to make sure this doesn’t happen in the future? How can we fireproof our houses? That’s where we are now. We have a fire going on.”  

Turns out, there were already pretty big fire breaks and sprinkler systems in place; we seem to have paid for expensive water-bombers to put out fires on worthless garbage barges out at sea.

They conclude:

We find that counterparty contagion is significant but limited in magnitude. This owes to the fact that most of the counterparties have rather small exposures to the bankrupt companies. The largest exposures, revealed in the list of the largest unsecured creditors in the bankruptcy petition, often belong to the trustees of publicly traded bonds and since these bonds are widely held, the exposure of a single financial firm is substantially smaller. Outside of the trustees, we find that many of the creditors are financial firms, which supports the description of financial institutions as highly interconnected. However, financial creditors are rarely at risk of failing as a result of another firm’s troubles because these  firms hold diversified portfolios, as is often required by regulation. We find stronger counterparty contagion effects for Lehman Brothers and AIG but even these extreme events did not involve sufficient counterparty contagion to cause numerous cascading bankruptcies in the financial system.

We could laugh at the Americans, but we did have our own particularly ridiculous bailouts here in New Zealand.

In related news, I probably have the only kids on the planet who can be encouraged to eat the less tasty parts of their dinner on the promise of getting to watch this one for the n= (large and increasing)^th time:



It is ridiculously cute when a 2.5 year old girl pleads for "Keynes and Hayek!"

Best is when she demands the video with breakfast; no better way to start the day, really. "Eye of the Tiger" for economists.

Posting will remain light while I remain mired in grading hell.