The Great Tobacco Debate

If you’ve been living under a rock, you may have missed the debate that’s been raging over the past couple of weeks over tobacco consumption.

The chief protagonists of this debate are, on the one side, economist Stephen Koukoulas and Media Watch; and, on the other side, a number of journalists with the Australian newspaper (Sloan, Ergas, Kerr, Creighton, even Jack the Insider and Cut'n'Paste got in on the action).

The debate kicked off with claims and counterclaims about whether tobacco consumption has increased of decreased since plain packaging laws were fully enforced from December 2012. While there has been a range of peripheral issues brought into the mix (including the credentials and leanings of each side, other measures available through health departments, and views on state paternalism) the central argument has been over the interpretation of the ABS’s chain volume measures (found in the National Accounts) and whether they show that tobacco consumption really has decreased since plain packaging was introduced.

It's interesting that the debate seems to be fuelled by (if not always focused on) the the tension between freedom and state paternalism, particularly given Australia's long history of gradually tightening control of the tobacco industry and our acceptance of gradually removing private freedoms with regard to smoking (see below). Just to clarify my position on this: I have no sympathy for a business that makes money by selling an addictive poison. Nevertheless, the information below is factual and based on the numbers.

I’m satisfied, from what I’ve read from each side (hindered by the Australian’s paywall, since I don’t pay for substandard content), that chain volume measures are a reliable indicator of tobacco consumption (see Mr Koukoulas's comments here about ABS data, as well as corroborating evidence from Treasury presented here). So let’s take that as a given. The graph below shows chain volume measures for household expenditure on tobacco since December 2012 (when plain packaging was fully enforced). Although there was a brief increase, overall, it has dropped. This is largely due to a big dip in the last quarter, and, just as one quarter of negative growth does not a recession make¹, one quarterly dip in tobacco consumption does not confirm an ongoing trend.

Household expenditure on cigarettes and tobacco, chain volume measures (seasonally adjusted, trend and quarterly increase)

Note: although plain packs were enforced from December 2012, the first quarter to include changes was March 2013, which covers January, February and March 2014.
So, overall, tobacco consumption is down since plain packaging was introduced.

What I haven’t seen canvassed is whether the decrease since 2012 is less, greater or the same as the long-term downward trend in tobacco consumption. I notice that Mr Koukoulas has been careful not to claim causation in his arguments; simply that consumption has decreased. This is what he had to say on 18 June 2014 (last Thursday):

... I have not defended the plain packaging laws. While I think most policies aimed at reducing smoking are worthwhile, I merely noted that the ABS data on the household consumption of tobacco highlighted, in no uncertain terms, the embarrassing errors in Christian Kerr's story in The Australian of 6 June 2014. The volume of tobacco consumed is falling. I did note the plain packaging laws, plus the excise increase, as factors that may account for the obvious fall in consumption.

Surely the crux of this debate should whether plain packaging is effective. Otherwise, who really cares if consumption goes north or south? To get any kind of sense of whether plain packs have been effective, we need to compare the before and after rates.

If we zoom out a little (using the same chain volume measure that has been the basis of much of the debate), we can see that the recent trend is a continuation of an existing one. If anything, it looks like the steady decrease in tobacco consumption slowed at the same time plain packaging was enforced.

Household expenditure on cigarettes and tobacco, chain volume measures (seasonally adjusted, trend and quarterly increase)

If we zoom out even further, as far back as this data set takes us, we can see that tobacco consumption has been falling quite streadily since the start of the 1980s .

The Federal Department of Health conveniently posted information on its website on Friday showing corroborating statistics of tobacco sales, tobacco consumption (as measured by Treasury) and smoking prevalence rates, as well as some key reasons for the decreases (some of which are marked on the below graph). The graph below shows the chain volume measures (quarterly rates, seasonally adjusted and trend; and quarterly per cent increase or decrease).

Household expenditure on cigarettes and tobacco, chain volume measures (seasonally adjusted, trend and quarterly increase)

Over the 32 year period between June 1982, when consumption began to tumble, up until the full enforcement of plain packaging (the December quarter of 212), consumption fell by an average of 0.615% per quarter. Since plain packaging was introduced, it has been 0.600% per quarter.

On the basis of that, you would have to say that plain packaging has had no (or a negligible) effect, unless the most recent quarter becomes the start of a deeper downward trend. It's important to note that, in addition to plain packaging, the tobacco excise was hiked 12.5 per cent in December 2013 (just in time for the big drop in the March quarter), which muddies the waters of causation even further. I'm not expert on marketing or market signals, but I would hazard a guess that a 12.5 per cent increase in the cost of cigarettes is a stronger influence on smoking behaivour than a boring package. Excises have had a dramatic impact in the past. For example, following pressure from health groups, the government hiked the tobacco excise and customs duty for successive years from 1992 to 1995, with moderate increases through the late 1990s. From 1999 through to 2010 there was no real increase to the excise or duty (only adjustments for CPI)². Have a look at the above graph. See the big dive in the first half of the 1990s, and the long flat spot after it? Those are the periods that the excise and duty rose and then levelled out.

These numbers are surprising. Surely packaging is an important factor in consumer behaviour; otherwise, companies would not spend millions on branding. Of course, tobacco is not like other markets - its consumers are addicts, and there are complex psychological processes at work that are not as significant in other markets. Perhaps this lends itself to the conclusion that it will take time to see the real effect of plain packaging, as it is more likely to divert new smokers, rather than change the behaviour of existing tobacco addicts. Although, the increase in the number of calls to Quitline following the introduction of plain packaging provides some cause for optimism for existing smokers.

It may well be that the steady decine in smoking since the early 1980s has been achieved through sustained efforts to intervene in this market (bans on advertising, designation of smoke-free areas, excise hikes, advertising campaigns) and that plain packaging is the most recent in a long line of strategies to keep tobacco consumption on the down.

The conclusion? As the Federal Department of Health concludes on the matter, "The effect of tobacco plain packaging, as part of Australia's comprehensive package of tobacco control measures, will be seen over the longer term."

_____
¹The most common criterion for a recession is three quarters of negative growth. Surely we would want to see three quarters of below-trend (or should that be above-trend?) negative growhth in tobacco consumption?
²Tobacco in Australia: Tobacco Taxes in Australia.