Small Brewers Coalition commentary on recent SIBA open letter to the Treasury

SIBA’s recent publication of a comparison of a survey of 43 of its own members’ cost of production with the European Economics report published 15 months before is highly misleading. It knowingly compares two completely different sets of data, prepared quite differently, one open for review, the other unaudited.

It accompanies an open letter written to the Exchequer Secretary to the Treasury, Robert Jenrick MP by Mike Benner, Chief Executive of SIBA, which is also very misleading. Its publication coincides with the launch of the Treasury review and so may have the effect of guiding small brewers as to how they should complete the cost element of the questionnaire in order to support SIBA’s political lobbying.

A detailed commentary on the data comparison follows below, but, in summary:

What can we therefore conclude?

The Treasury used European Union rules in determining that the maximum relief should be 50% of full duty and so set the maximum level of compensation for diseconomies of scale – the Coalition developed an Industry consensus proposal that there should be reform of SBR and it should iron out the worst of the market distortions caused by the adoption of the 50% rate, but not materially disadvantage any individual group of brewers. It was constructed as a pragmatic compromise – many would have liked to have seen a more material change but were prepared to compromise if SIBA were.

SIBA subsequently imposed a simple red line demand that there should be no change below 5,000hl and in so doing sought to block any meaningful change. That demand has been repeated in a public letter to the Minister, Robert Jenrick MP, once the Treasury announced their review. This is a wholly inappropriate demand from a trade body in such circumstances, and it has not received the support of the rest of the sector whose objective has always been to work with the Government to achieve sensible and constructive reform.

SIBA’s own data shows there is significant economies of scale between a 1,000hl and 5,000hl brewer (just over £30/hl which in a highly competitive market is material). This is in line with common sense and broadly supports the Coalition view. The data also suggests the costs of production then go up again between 5,000hl and 20,000hl before starting to fall (as below). The recent survey data is emboldened.

For comparative purposes the previous data SIBA prepared based on 20 respondents is in the line below. That essentially supported the then SIBA argument that there were no appreciable economies of scale below 20,000hl, therefore no need for reform at that level.

Hectolitre bands 0 -1,000 1,000-2,500 2,500 -5,000 5,000 – 10,000 10,000-20,000 20,000 – 30,000
SIBA survey Autumn 2018
Total production costs inc duty
£131.39 £114.13 £100.86 £118.99 £121.83 £116.25
Previous SIBA data 2017 £130.63 £134.95 £126.61 £132.59 £131.43 £117.78

Note The lower band of SIBA data was linked to a report they commissioned from CEBR. The Coalition published a detailed review of the methodological flaws in that report. This is available on the SBDRC website.

The letter from the Chief Executive of SIBA to the Exchequer Secretary states ‘Having read your letter … many small brewers will be left confused, with the view that the Government is already minded to reform SBR by redistributing relief from smaller to larger brewers. This has no economic, political or moral basis. This would have a devastating impact….’

The EE survey shows that many small brewers would welcome a very long overdue review and will not jump to the conclusions SIBA claim – they want to compete on a level playing field and build long-term, sustainable businesses.

We believe there is an economic, political and moral case for reform below 5,000hl.

From an economic perspective it is self-evidently the case that a 5,000hl brewer will be more efficient that a 500hl brewer. It is economically illiterate to suggest the opposite. If a very small brewer is treated as having the same costs of production as a brewer ten times his size this is clearly an unfair system.

Apart from being unfair to the larger of the small brewers, it is deeply unfair to the more than 50% of small brewers, mostly under 1,000hl, who are not members of SIBA. These brewers will also not benefit from the SIBA Beerflex scheme, a key route to market for many SIBA members. The very small brewers have no separate body to represent their specific interests and have therefore not been able to challenge the SIBA argument that a brewer 5-10 times their size should receive exactly the same percentage duty relief, despite it clearly putting them at a relative competitive disadvantage.

From a political point of view, no trade body should seek to undermine an objective review before it is properly underway. Trying to apply pressure to ensure only one acceptable outcome based on a one-sided view that ignores the reality of what is happening in the market – and where there is obviously oversupply leading to price discounting which will lead to brewery closures – is unhelpful at best. All trade bodies have a responsibility to act in good faith when Government is answering a call for a long overdue review.

Finally, from a moral point of view, SIBA has spent over ten years trying to block a Treasury led review, right up to the point it was announced. Many small brewers have suffered under the current system – very small brewers who could not break through and larger brewers who could not compete with the very generous duty reduction allowed to their smaller competitors. Some have scaled back, other have been forced to close or sell. This includes businesses that kept the independent brewing flag flying through the very difficult years from the 1980s through to the introduction of SBR in 2002. There is a clear case to introduce a system that is fair at all levels of brewing and does not disproportionally advantage one group at the expense of smaller and larger competitors.

SIBA have recently issued updated guidelines for completing the Treasury questionnaire, and as well as encouraging brewers to include all costs (including admin, marketing and other costs) they are also suggesting that brewers adjust the salaries if they think they or staff are not being paid enough. There is no other business in the UK where the appropriateness of the scale of government subsidy would be judged on that basis. This interpretation moves SBR well beyond a scheme to encourage growth of small brewers by offsetting the diseconomy of scale of the more efficient ones, to a simple profit support scheme – and that certainly wasn’t the intention of the original scheme.

The Coalition welcome the Treasury review and would encourage all brewers to complete the questionnaire carefully, based on their own data, and to add any comments they may specifically wish to make in the notes at the end or in a covering letter. We would ask that when brewers are completing the questionnaire they move beyond narrow self-interest and consider what structure will be in best to achieve a long-term sustainable small brewer sector.


Detailed commentary on the data comparison:

1. The EE report was prepared to specifically answer the question as to whether there were significant economies of scale amongst small brewers in the UK, and if so, what was the shape of the diseconomy of scale curve and how did it compare with that assumed in Small Brewers Relief.

2. It also sought to refute the specific SIBA assertion that there were no material economies of scale between a brewer at 500hl per annum and one at 5,000hl pa.

3. Earlier attempts had been made by Oxford Economics to look at actual production cost data in the sector, but it had proven of very limited value due to the difficulties of accurate and verifiable data collection. There were a series of data outliers that materially skewed the averages and it became evident that different brewers had very different categorisation of production costs.

The conclusion reached was that unless a very large sample base (over 100 brewers) was collected, and analysed, by a completely independent body using strict selection criteria based on product mix, geographical location and business structure, and reviewed to normalise differences in capital investment, depreciation policy and standardise cost allocation, the results would not be statistically significant. Unfortunately, the latest SIBA survey does not fulfil these criteria.

Indeed, the existence of SBR at current levels means that brewers that would not otherwise be economic can still enter and compete in the market. If the costs of these brewers are assessed as part of an exercise of cost comparison, the current SBR structure becomes, in part at least, self-justifying

4. The EE survey collected responses from over 100 brewers on their views of how SBR was working and there was a widespread view that it should be reformed to remove anomalies and make the relief more effective in encouraging growth. There was a general view that the relief didn’t reflect the real diseconomies of scale.

5. To address the problems of collecting actual costs, listed above, and to establish a more objective and verifiable diseconomy of scale curve, it was decided to attempt to recreate a notionally ‘efficient’ brewery at 12 different volume points ranging from 500hl per annum to 1mhl pa. The key assumption was that at each volume level the brewery was configured with plant and labour to operate at a high level of efficiency. Whilst this was highly unlikely to be the case in the real world where most businesses were either growing or declining so not optimally efficient, it did mean there was a consistent set of open assumptions from which a curve could be drawn plotting cost of production per hl against size. The resulting curve was then overlaid against that reflected in the current SBR system – the position on the y axis of the graph (cost per hl) was not taken into account – only the curve shape.

6. Other important assumptions in the EE work included the type of beer brewed – at all brewery sizes, a 4.2% standard bitter (with relatively low hopping compared to modern IPAs). One recipe had to be chosen to ensure like for like comparisons and this one was chosen because it was the best-selling ale profile in the UK.

7. SIBA appointed Garbutt and Elliot, a Leeds based accountancy firm, to help design the survey but retained control over the analysis and conclusions. SIBA took the EE data and compared it with their own unaudited data from 43 of their members. Most of the brewers who provided SIBA with data were under 10,000hl with a sample of only five brewers between 10,000 and 30,000 and none above that level.

8. SIBA did not compare like for like volume bands. The costs of the nine brewers classed as between 1 and 1,000hl were averaged and then compared with a 1,000hl brewery – if it is assumed that none of the SIBA brewers produced less than 500hl per annum as there are no brewers under that heading to compare with the EE model, it still means that the average volume would be 750 rather than 1,000hl, so ¾ of the size…a material difference. The next SIBA band contains 14 brewers between 1,000 and 2,500hl. In the absence of any other data we must assume the average volume falls halfway, so an average of 1,750hl…again significantly below 2,500hl. At the EE band of 5,000hl SIBA compares an average brewer of 3,750hl pa, at 10,000hl SIBA is using data averaged on 7,500hl, at the EE band level of 20,000 SIBA is comparing brewers at an average of 15,000hl and finally at 30,000hl SIBA is comparing with brewers at an average of 25,000hl.

SIBA banded in such a way that they consistently compared with brewers who were on average between 2/3 and 3/4 of the size of the EE comparator, and so were inevitably likely to have higher costs of production.

[SIBA had the actual volumes for each brewer data available from their members which would have allowed them to cluster in such a way that the average of their cluster matched the EE band they were comparing it with].

EE volume band in hl 500 1,000 2,500 5,000 10,000 20,000 30,000
…compared by SIBA to: none 750 1,750 3,750 7,500 15,000 25,000
n/a 25% less 30% less 25% less 25% less 25% less 17% less

9. Taking each heading in turn: Malt – the EE data is now 15 months old, and there has been a significant increase in malt prices in the last season – this may be reversed in the open market in due course, but SIBA did not correct it in this data, despite being made aware of the changes. The EE costs assumed that at 5,000hl the brewer would have their own mill and begin receiving whole malt in larger bags (at lower cost) and at 10,000hl would be receiving most of their malt in bulk, (a saving of approximately 30% versus bagged pre milled malt) and would have higher yields of wort from malt as a result of more efficient plant.

The capital costs of the silo are incorporated in the depreciation figures. In fact, many maltsters will offer to fund bulk silos and payback on the investment at 10,000hl will be less than three years, so is a wise investment for an efficient business to make to reduce marginal costs. SIBA’s figures suggest none of their brewers have made that investment. The two figures are therefore not comparable. It is interesting to note that at the smaller end of the sample the EE estimated costs are actually slightly higher than SIBA’s own members claim.

10. Hops. It is undoubtedly the case that some of the newer brewers brewing ‘hop forward’ beers might use more hops per hectolitre than those brewing traditional bitters, and that if highly aromatic they are more likely to be more expensive, non-domestic hops. (This may allow the brewer to command a price premium to recover the extra cost.) Smaller brewers buying on the spot market will pay more than brewers buying large quantities on the spot market.

The EE and SIBA reports are therefore not directly comparable as the first is looking at diseconomies caused by scale, not recipe variances. It is worth noting that whilst the bases are different, both sets of figures show costs reducing with scale.

11. Other Whilst it is not clear what costs the SIBA data includes the scale of each is similar and both decline with scale, the SIBA data more markedly.

12. Other Direct Costs- Production labour There are dramatic differences between the two sets of data, and this is almost certainly caused by entirely different assumptions, but the SIBA data cannot be interrogated. It is highly likely that with a small business where individuals are multi-tasking it is difficult to separate the time spent on production tasks from other tasks like admin, sales, distribution, and marketing. As a result, these may all be grouped together. If a brewer is only brewing twice a week it is not appropriate to include 5 days of wages in determining cost of production. The SIBA figures do also demonstrate some internal anomalies – it is difficult to understand why production costs start rising again between 5,000 and 10,000hl, then fall dramatically. This may be a result of a very small number of data points.

13. Energy and water Interestingly, the SIBA and EE data almost match, albeit the comparison of EE data was with breweries about 25% smaller on average (see chart in 8 above), implying that even small-scale differences on a small expenditure item might make a difference. There is an odd variance in costs in SIBA data between 5,000hl and 30,000hl but this may be simply due to the small data sample (20 breweries)

14. Other consumables This data is not separated by SIBA and may have therefore been included in Material costs, other (point 11 above).

15. Rent and rates There are material variances at all levels between the EE and SIBA data of approach double in each case and this may be due to the SIBA data not allocating any rent to non-beer production (i.e. Offices, Distribution, Warehousing or on-site Shop/Bar) whilst EE assumed only 50% of total rent and rates allocated. Clearly, this is not a like for like comparison between two sets of similar data. There is also one outlier figure under SIBA’s 0 -1,000hl band of £18.18 which is an average of 9 brewery figures and is possibly explained by very sub optimal space use at the lowest volume end, incorrect cost allocation or midyear start-ups (none identified in SIBA data).

16. Depreciation (and capital investment) It is not clear why the SIBA data varies as it does but the higher depreciation at higher volumes is almost certainly as a result of choosing shorter periods over which to depreciate plant and higher plant expenditure. It may also be that the SIBA data included plant not directly used for production like fork lifts for warehousing and loading duties, vehicles for distribution etc. The EE report assumed, for the purposes of simplicity, a flat 15-year straight line depreciation on production plant only. In practice most vessels will have a depreciable life of 20 years plus but computer-controlled equipment and those machines requiring regular maintenance will be less, so there will be a wide mix of depreciation rates. The EE capital cost estimate was based only on producing the defined volume per annum, and in practice a brewer will always build in additional capacity, so the capital costs of the EE and SIBA data will not be comparable. The EE model also assumed all bottling and canning would be contracted out initially and only brought in house once an adequate payback on the investment could be achieved. It also assumed that 75% of plant was bought new, 25% used and reconditioned.

17. Repairs and maintenance The disparity between the two sets of figures may be explained by the comments above – the EE report assumed a repairs and maintenance cost of 20% of the depreciation figure of only plant used directly in production, and buildings used for production. In practice, brewing plant generally has low repair costs, packaging is higher, and the buildings require regular maintenance (unless short term lease in which case the landlord may cover), and so it is important to allocate repair and maintenance costs to usage, and for most brewers their systems would not differentiate costs between repairs related to brewing, distribution or office.

18. Title ‘Benchmarking survey’ This survey was ostensibly conducted by SIBA in the Autumn of 2018 to allow brewers to compare their costs with a benchmark, to encourage them to focus on areas where they could improve efficiency and therefore profitability. The contributors were promised free access to the data whilst all other brewers had to pay £100. Unfortunately, the lack of transparency as to how the data was gathered and what is included/excluded makes it very difficult to use for benchmarking purposes. If an individual brewer compares his costs with the survey he needs to know more in order to draw like for like comparisons and determine whether he is indeed behind the benchmark, or ahead. It would have been more useful if a set of notional model breweries was created from the data with a clear explanation of what was and was not included.

The speed with which this data has been rushed out, the dropping of the £100 charge and general publication and the misleading comparison with the EE data all implies this exercise was conducted primarily for political lobbying purposes. This conclusion is supported by the preamble to the survey, first sentence: “In order that SIBA is in a position to protect, enhance and defend SBR it is highly desirable to gather further and more up to date data on members costs of production following analysis of the data gleaned from the SIBA members’ survey”

19. Use of EE data in Treasury submission SIBA incorrectly imply that EE cost of production data misled the Treasury into believing costs were much lower and therefore SBR should be structured differently. All documentation from the Coalition makes it clear that the call is for a fairer system which reflects the variation in the real diseconomies of scale, not the absolute amount. The EE modelling was constructed to try to demonstrate in an open and challengeable format what a real diseconomy of scale curve might look like in the small beer sector, and to illustrate the market distortive effect of the current system.

It was clearly a difficult task because every brewery is different so some reasonable assumptions as to a base case had to be made. At no point did the Coalition suggest this was the real costs of all brewers, and the absolute data was not used when the Coalition was asked to propose an alternative to the current SBR system. Indeed, the Coalition proposal is very far from that which the data suggests might be optimal, and was a compromise proposed in the hope that all parties could congregate around one reasonable and modest solution. SIBA rejected this compromise and it is now open to all brewers to make their own submission to the Treasury and for the Treasury to determine the best solution.

Appendix One

Below are two graphs, duty exclusive and duty inclusive, comparing SIBA’s original data to support their successive claims that: i) there were no economies of scale below 5,000hL (CEBR report); and, ii) that relief at 50% should be extended beyond 5,000hL. For reasons outlined above, the data itself is contentious and cannot be used to compare either of their surveys to that of the EE Report. Indeed, if they compared their own data collection and analysis to itself, two serious concerns arise even before comparing either set with EE data. First, the new data confirms clearly that there are economies of scale below 5,000hL and so undermines their position that no change below 5,000hL is warranted on that basis. Second, it shows such vast contradiction between their data sets, nearly 100% in the 2500-5000hL range, duty excluded, as to call into question the usefulness of either of these sets.