How not to lie with pub closure statistics

My report about pub closures was generally well received when it was published yesterday, with the notable exception of the All-Party Parliamentary Save The Pub Group, which sent out a bizarre and vitriolic press release. The press release was too full of holes to get any mainstream media attention but it can be read here. The APPG, which has the Campaign for Real Ale as its secretariat, makes the baffling assertion that I am a ‘tobacco lobbyist’ and claims that my report argues that ‘pub closures are all down to the smoking ban not the disastrous pubco model, when the evidence clearly shows otherwise’. None of this is true. The report identifies the smoking ban as one of three factors – alongside the recession and alcohol duty – which have led to excessive pub closures since 2006. I also say that general social change is responsible for about 40 per cent of the recent decline. And I am not, nor have I ever been, a lobbyist.

Evidently the APPG is upset that I rejected their arguments about the PubCo model being the major cause of pub closures. I did so for the simple reason that the rate of pub closures is very similar in both the PubCo sector and the independent sector. Figures from CGA Strategy show that 16.5 per cent of pubs in the non-managed (mostly PubCo) sector closed between December 2005 and March 2013, whereas there was a closure rate of 14.6 per cent in the independent sector. Clearly both parts of the pub trade have suffered to a similar extent. I put this down to a fundamental lack of demand for pubs as they currently exist.

The correct way to look at the data is to calculate the number of net closures in each sector by taking the number of closures and adding the number of openings. The APPG does not do this. Instead, they take the number of net closures and add or substract the number of transfers between sectors. So, if a PubCo sells 3,000 pubs to the independent sector, the APPG counts this as 3,000 closures in the PubCo sector and 3,000 new openings in the independent sector.

This is obviously wrong. A pub that is sold from one sector to another does not represent a closure. It has simply changed hands. In the report, I wrote: ‘The All-Party Parliamentary Save the Pub Group, amongst others, has been guilty of misrepresenting transfers as closures in an effort to support their claim that the “the non-managed (largely leased/tenanted) sector has seen many more net closures than those of independent freehouses” and that “PubCo pubs are being sold off for alternative use and bulldozed in their thousands”.’

What the APPG lacks in evidence it makes up for in strident repetition and so it is unsurprising to find the group repeating its misleading claims once again in its press release. It reads…

The reality of pub closures figures, as previously published by the Save the Pub Group is as follows: –

· CGA Strategy figures show that there has been a much greater drop in the number of leased/tenanted pubs compared to freehouses between December 2005 and March 2013 – the number of ‘non-managed’ (tenanted/leased, mostly tied) pubs fell by 5,117 compared with a fall of only 2,131 in the number of ‘free trade’ pubs. (CGA Net Closure figures December 2005 – March 2013 – Source: CGA, sent the Save the Pub Group 2013 November 2013)

· The pubco’s trade association, the British Beer and Pub Association’s own figures show that over ten years non-managed pubs decreased by over 8,000 whilst the free trade sector actually expanded by 1,600. (Source: BBPA Statistical Handbook 2013)

· Enterprise Inns and Punch Taverns collectively disposed of over 5,000 pubs between 2008 and 2012, a THIRD of all of their pubs in just 4 years. No other model had anything like this disposal rate. (Figures taken from Enterprise Inns and Punch Taverns annual reports available online, 2008 Ent Inns 7763 and Punch 7560 by 2012 Ent Inns 5720 and Punch 4529)

The third point answers the second. The total number of pubs in the independent sector has risen because the PubCos have had a high ‘disposal rate’ in the last decade. They are selling pubs to the independent sector, but transfers within the industry do not tell us anything about closures because they are not closures. To see the closure rate we need to look at the data from CGA Strategy mentioned in the first point above. Here it is

The APPG says ‘the number of “non-managed” (tenanted/leased, mostly tied) pubs fell by 5,117 compared with a fall of only 2,131 in the number of “free trade” pubs’. They do not mention the fact that the non-managed sector was more than 40 per cent larger than the independent sector in 2005 and therefore had more pubs to lose. They also ignore the net closure figures which are clearly displayed above in favour of a comparison between total numbers after transfers have been taken into account.

To see the proportion of net pub closures in each sector, you simply need to divide the number of net closures by the number of pubs that existed at the start of the series and then multiply by 100. For the independent (‘free’) sector, this is 3,258/22,324 x 100 = 14.59 per cent. For the non-managed sector (most of which is made up of tied PubCo pubs), the figure is 5,304/32,177 x 100 = 16.51 per cent. The difference between these figures is trivial. Pubs have clearly been closing at a similar rate in both sectors.*

The stubborn refusal of the APPG to do this basic mathematics does not reflect well on them, nor does their repeated use of misleading statistics in parliament. The government has looked at this issue on numerous occasions and come to the same conclusion. In May of this year, for example, the Department for Business, Innovation and Skills noted that ‘between March 2010 and September 2012 the closure rate was very slightly lower in “non managed” pubs, 4.3%, than in ‘free’ pubs, 4.5%. The ‘net closure’ is the more appropriate statistic to use as it takes into account ‘churn’, where pubs close for a short period then reopen. However, if one uses gross closure figures then proportionately even more “free” pubs are closing, 3.4% versus 5%.’

The Campaign for Real Ale has its reasons for wanting to be rid of the PubCos, just as it had its reasons for wanting to be rid of the beer tie in 1989 (CAMRA lobbied for the Beer Orders which directly led to the creation of PubCos – CAMRA now admits that it was ‘naive’). Like many special interest groups, it has been effective in using an All-Party Group to pursue its objectives in parliament and there are many PubCo tenants who hope that the government will give them a free lunch by effectively tearing up their contracts. That is a whole other story, but it is a shame that the APPG cannot make their case without resorting to skewed figures and smear tactics.

*As an aside, the APPG’s press release says that ‘Mr Snowdon’s report contains the myth about tied versus free of tie pub closures, when CGA Strategy themselves have confirmed that they do not collect figures for tied versus free of tie pubs!’ It is true that CGA Strategy only has figures for ‘non-managed’ pubs. The great majority of non-managed pubs are PubCo pubs, but some are not. I acknowledged this in the text of my report. Note, however, that despite making this criticism, the APPG proceeds to use CGA Strategy’s figures in its response, albeit incorrectly.

Head of Lifestyle Economics, IEA

Christopher Snowdon is the Head of Lifestyle Economics at the IEA. He is the author of The Art of Suppression, The Spirit Level Delusion and Velvet Glove; Iron Fist. His work focuses on pleasure, prohibition and dodgy statistics. He has authored a number of papers, including "Sock Puppets", "Euro Puppets", "The Proof of the Pudding", "The Crack Cocaine of Gambling" and "Free Market Solutions in Health".

2 thoughts on “How not to lie with pub closure statistics”

  1. Posted 11/12/2014 at 15:40 | Permalink

    What a very unpleasant and over-the-top response from the APPG. However at least Mr Mulholland didn’t call Chris a “disingenuous, manipulative, illiberal little sh*t”, which is what he called a constituent in an infamous tweet.
    Tied pubs are surely just a particular form of franchise which is freely entered into. As a tenant, your ability to buy beer is constrained, just as MacDonald’s franchisees have to buy their burgers and buns through the company. The benefit is a relatively low rent. Untied pubs which pay a higher rent have a business model which implies greater risk, as fixed costs constitute a larger part of their expected revenue.

  2. Posted 11/12/2014 at 21:46 | Permalink

    The model is potentially very important for people with limited access to capital. Economists who obsess with market failure type models often talk about difficulty of accessing capital markets and when the market solves the problem they complain about too

Comments are closed.