Ban damage

Quantifying the damage of smoking bans to the pub industry has never been easy. As well as a long-term (post 1960s) trend towards people drinking at home, there are clearly other factors affecting the licensed trade, including higher rents and taxes. Few people seriously argue that the smoking ban has not had a negative effect on Britain's pub trade. The question is how much?

In England and Wales, the picture has been muddied by the economic downturn which began in the first year of the ban, leading to recession in year two. Our ability to assess ban damage has also not been helped by the difficulty in finding solid and specific data on pub closures. The British Beer and Pub Association has issued figures showing 39 pub closures a week, more recently rising to 52 pub closures. Measured against 2 or 4 pub closures a week before the ban, one would have to be rather blinkered to pretend that the smoking ban has not been a major factor (although some have tried), but they are still annual figures for the whole of Britain. We have not had the benefit of seeing quarterly figures for each part of the UK.

But now we do. As the Morning Advertiser reported on Friday, and Taking Liberties reports today, data from the CGA Database give us a much clearer idea of what has been happening. Analysed by CR Consulting for the Save Our Pubs and Clubs campaign, the report concludes:

The results demonstrate a very close statistical relationship between the introduction of the smoking bans and the acceleration of the decline of the British and Irish pubs. This relationship is considerably stronger than those that could be attributed to a change in the beer duty or the recession.

The strength of this report is that it looks at four countries which implemented bans at different times (Ireland: 2004, Scotland: 2006, England & Wales: 2007). Pub closures accelerated in all four countries from the middle of 2008, which obviously indicates an effect from the recession. Since there are multiple, combinative factors leading to pub closures, it may be that the recession was the final nail in the coffin for pubs already weakened by the smoking ban. Or it may not. The best way to tell would be to see how many pub closed in previous recessions. My understanding is that pubs are traditionally seen as recession-proof, but I have yet to find reliable data for this.

But if the smoking ban's impact on pubs in England and Wales is muddied by the following year's recession, the same cannot be said of Ireland and Scotland, whose respective bans were enforced four and a half years and two and a half years before the recession began.

What is striking about the figures from the CGA Database is that there is the same pattern of pub closures, regardless of implementation date and regardless of the wider economy. Large-scale pub closures began in Ireland, then in Scotland and finally in England and Wales, perfectly following the chronology of the smokefree legislation.

Furthermore, the rate of closures is remarkably similar across the four countries. Closures in year two of the smoking bans are within the narrow range of 4.1% and 5.6%. In year three, they range from 5.5% to 7.6%. And in year four, both Scotland and Ireland saw the closure rate rise to 11% (all % are based on percentage change since the smoking ban.

In the first year of a smoking ban, the damage is less pronounced. Pubs are unlikely to close straight away—most can at least make it through the first winter. The exception here is Ireland which saw a large number of closures straight away (7.3%). Since the Irish economy was positively booming in 2004/05, this cannot be attributed to the wider economy. It would be interesting to see data for previous years.

What this report shows is that England's rate of pub closures in 2008/09 (4.9%) was very similar Scotland's rate of closures in 2007/08 (4.6%). Likewise, Ireland's rate of closures in 2007/08 (11%) was virtually the same as Scotland's rate of closures in 2009/10 (11.1%).

In other words, each country is experiencing the same phenomenon to almost exactly the same degree but in different years. The only plausible explanation for this time-lag is the implementation of smoking bans. When you look at closure rates in the context of when each smoking ban was introduced, the data fits like a glove.

It also fits all the other evidence. It fits what publicans have been saying:

The readers' poll showed 77% of licensees think that trade has suffered as a result of the ban. Almost two thirds (63%) say business is worse than expected, and 72% predict a "challenging" or "very challenging" outlook for their business. Three out of five licensees said they had let staff go or reduced their hours. In addition, 73% want the ban lifted.

It fits what market analysts have been saying:

Pubs have sold 175 million fewer pints in the past year as a direct result of the smoking ban, according to market analysts AC Nielsen.

It fits what pubgoers have been saying:

There's no need for any fancy statistical analysis of trends over time. Just ask the customers.

It fits what the share prices of the Pubcos have been telling us:

You'll notice that the collapse of the share price began almost on the dot of July 1st 2007. Recession? No—that didn't start until October 2008, by which time the company had lost 75% of its value. Supermarket booze? 'Twas ever thus. Bad management? Perhaps, but the story is the same for all the pub companies.

It fits what economic theory predicts will happen when a externality is imposed on a business; it fits what the pub industry did predict would happen; it fits what has happened in other countries, in other states and in other cities.

The only thing it doesn't fit is the rhetoric of anti-smoking groups like ASH:

"Smoke-free polices are not only good for health, they are good for business. Evidence shows that in countries where smoke-free laws have been introduced, trade has generally increased."

Amanda Sandford, ASH, 2003