Title:  Government response to the consultation on proposals for changes to Gaming Machines and Social Responsibility Measures     IA No: ​      RPC Reference No: RPC-4137(2)-DCMS Lead department or agency:   Department for Digital, Culture, Media and Sport (DCMS)    Other departments or agencies:      N/A Impact Assessment (IA) Date: 17/05/18 Stage: Final Source of intervention: Domestic Type of measure: Regulatory Contact for enquiries: Ronald.whittington@culture.gov.uk​ / 020 7211 2283 Summary: Intervention and Options RPC Opinion: GREEN Cost of Preferred (or more likely) Option Total Net Present Value Business Net Present Value Net cost to business per year (​EANDCB in One-In, Three-Out Business Impact Target Status 2014 prices) -4,660 -4,660 450 In 2270 What is the problem under consideration? Why is government intervention necessary? Gambling-related harm produces several negative externalities including but not limited to: increased healthcare costs, welfare costs, and other costs to individuals associated with problem gamblers (e.g. family, friends and employers). Therefore, the marginal social cost is higher than the marginal private cost to the gambler. Government intervention is necessary to ensure we strike the right balance ​between socially responsible growth and the protection of consumers and wider communities​. Unfortunately due to the nature of gambling-related harm (inelastic demand), it is difficult to dis-incentivise consumption of this good and therefore the burden of reducing harm must be placed on the suppliers of the good. This impact assessment primarily covers gambling on gaming machines. ​Under the Gambling Act 2005, gaming machines are defined by categories depending on the maximum stake and prize available. B2 gaming machines have a maximum stake of £100, a maximum prize of £500 and are primarily located in licensed betting offices (LBOs). 1 What are the policy objectives and the intended effects? The government’s policy objective is to look across the industry and determine what, if any, changes are needed to strike the right balance between socially responsible growth and the protection of consumers and wider communities. Underlying this objective is the government’s aim to reduce gambling-related harm. What policy options have been considered, including any alternatives to regulation? Please justify preferred option. We consulted on a range of measures ​covering gaming machine stakes and prizes, allocations, and other levers that might alleviate the risk of harm. ​In terms of regulatory options, we consulted on industry proposals (via a call for evidence), as well as a reduction in the maximum stake for B2 gaming machines from £100 to somewhere between £50 and £2. The main alternative to regulation that the government has considered and assessed is the voluntary arrangements through industry and others to pilot and evaluate social responsibility measures aimed at reducing harm. It is not clear that the measures have been as effective as hoped and this option has been discounted. We have also discounted maximum stakes on B2 gaming machines over £30 (more details on this are provided in the consultation response and below). - Option 1 - Retain status quo - Option 2 - Reduce maximum stake on B2 gaming machines from £100 to £30. Status quo across all other gaming machine stakes and prizes and allocations. - Option 3 - Reduce maximum stake on B2 gaming machines from £100 to £20 (non-slots), £2 slots. Status quo across all other gaming machine stakes and prizes and allocations. - Option 4 - Reduce maximum stake on B2 gaming machines from £100 to £2, increase in prize gaming stake from £1 to £2 and prize increase from £70 to £100 (£500 to £1,000 aggregate). ​Status quo across all other gaming machine stakes and prizes and allocations. The government's preferred option is Option 4, to reduce the maximum stake on B2 gaming machines from £100 to £2 and to maintain the status quo on all other gaming machines with the exception of prize gaming. We are clear that we want to see socially responsible growth, i.e. growth that is not at the cost to those most vulnerable. While Option 4 will have a business impact on the betting sector, it targets those most vulnerable to gambling-related harm by reducing high level session losses, while not preventing growth across the gambling industry as a whole (which we continue to see year on year). In addition, in their advice to government, the Gambling Commission are clear that they support a precautionary reduction in stake on B2 gaming machines, but also make clear that setting a revised maximum is a matter of judgement for government. Will the policy be reviewed? ​Yes​. If applicable, set review date: 3 years after implementation Does implementation go beyond minimum EU requirements? Are any of these organisations in scope? Micro yes What is the CO​2​ equivalent change in greenhouse gas emissions? (Million tonnes CO​2​ equivalent) No Small yes Traded:    N/A   Medium yes Large yes Non-traded:  N/A     I have read the Impact Assessment and I am satisfied that, given the available evidence, it represents a reasonable view of the likely costs, benefits and impact of the leading options. Signed by the responsible Minister: Date: 17/05/18 2 Summary: Analysis & Evidence Option 2 Description: ​ Option 2: Reduce maximum stake on B2 gaming machines from £100 to £30. Status quo across all other gaming machine stakes and prizes and allocations.   FULL ECONOMIC ASSESSMENT Price Base Year 2017     PV Base Year ​   2019   Time Period Years 10 ​     Net Benefit (Present Value (PV)) (£m) Low:​ -2150 COSTS (£m) Total Transition (​ Constant Price) Years High: -1190 Best Estimate: -1490 Average Annual Total Cost (excl. Transition) (Constant Price) (Present Value) 140 1190 Low 2.1 High 5.4 250 2150 Best Estimate 3.8 170 1490     Description and scale of key monetised costs by ‘main affected groups’ The most significant monetised annual cost will be to gambling businesses and supply chains affected by the lowering of maximum stakes on B2 gaming machines from £100 to £30 in reduced GGY (Gross Gambling Yield). Transition costs are established to comprise of one-off IT costs, training costs and marketing costs (see paragraphs 14-16) Other key non-monetised costs by ‘main affected groups’ Increase in session duration resulting from a stake reduction on B2 machines may reduce the availability of machines at peak times. This may impact B2 GGY if players cannot exercise their preference for machine play due to inadequate supply. There is also a cost associated with the dis-benefit reducing stake size has on non-problematic gamblers. This dis-benefit is due to a reduction in the utility these gamblers get from gambling at the new stake if their preference for staking behaviour is higher than the new cap. BENEFITS (£m) Total Transition (​ Constant Price) Years Low N/A     Average Annual Total Benefit (excl. Transition) (Constant Price) (Present Value) N/A 0 3 High Best Estimate N/A N/A 0 0     0 0 Description and scale of key monetised benefits by ‘main affected groups’ Other key non-monetised benefits by ‘main affected groups’ The most significant non-monetised benefits are the benefits to society of reduced gambling-related harm. Also, as a result of any decrease in harm, there could be some benefit to government from reduced expenditure on government programmes linked to gambling-related harm such as healthcare (specifically psychological therapy), welfare costs, housing costs and criminal justice costs. During consultation, we received a report produced by the Centre for Economics and Business Research (CEBR) which provided estimates of the cost to government of problem gambling related to B2 gaming machines (see paragraphs 24-28 for our consideration of CEBR report).1 Key assumptions/sensitivities/risks Discount rate (%) ● Take up (proportion of players who remain playing under the new staking regime), session duration increase (how long each person plays for on average), channel shift (players move to other forms of gambling) and attrition (players stop gambling). ● The central assumptions are a cause of potential under or over estimation of costs. If take up, attrition or channel shift are higher or lower this can drastically change the outcome of the model. ● There is also an important implicit assumption that reduced spend on gambling will result in a reduction in gambling-related harm. ● There is the key risk associated with the fact it is difficult to quantify gambling harm and inherently problematic to conduct cost-benefit analysis on policies to reduce it.   3.5%    BUSINESS ASSESSMENT (Option 2) Direct impact on business (Equivalent Annual) £m: Costs: 140 Benefits:0 Score for Business Impact Target (qualifying provisions only) £m: 720 Net:-140       1 ​http://www.coin-opcommunity.co.uk/wp-content/uploads/2018/01/CEBR-Report-final-Jan-2018.pdf 4 Summary: Analysis & Evidence Option 3 Description: ​ Option 3: Reduce the maximum stake on B2 gaming machines from £100 to £20 (with £2 on slots). Status quo across all other gaming machine stakes and prizes and allocations. FULL ECONOMIC ASSESSMENT Price Base Year 2017     PV Base Year ​   2019   Time Period Years 10 ​     Net Benefit (Present Value (PV)) (£m) Low: -4180 COSTS (£m) Total Transition (​ Constant Price) Years High: -1700 Best Estimate: :-2450 Average Annual Total Cost (excl. Transition) (Constant Price) (Present Value) 200 1700 Low 2.1 High 5.4 480 4180 Best Estimate 3.8 280 2450     Description and scale of key monetised costs by ‘main affected groups’ The most significant monetised costs will be to gambling businesses and supply chains affected by the lowering of maximum stakes on B2 gaming machines from £100 to £20 (with £2 on slots). Transition costs are established to comprise of one-off IT costs, training costs and marketing costs (see paragraphs 14-16) Other key non-monetised costs by ‘main affected groups’ Increase in session duration resulting from a stake reduction on B2 machines may reduce the availability of machines at peak times. This may impact B2 GGY if players cannot exercise their preference for machine play due to inadequate supply. There is also a cost associated with the dis-benefit reducing stake size has on non-problematic gamblers. This dis-benefit is due to a reduction in the utility these gamblers get from gambling at the new stake if their preference for staking behaviour is higher than the new cap. BENEFITS (£m) Total Transition (​ Constant Price) Years Low N/A     Average Annual Total Benefit (excl. Transition) (Constant Price) (Present Value) 0 0 5 High Best Estimate N/A 0 0    0   0 0 Description and scale of key monetised benefits by ‘main affected groups’ Other key non-monetised benefits by ‘main affected groups’ The most significant non-monetised benefits are the benefits to society of reduced gambling-related harm. Also, as a result of any decrease in harm, there could be some benefit to government from reduced expenditure on government programmes linked to gambling-related harm such as healthcare (specifically psychological therapy), welfare costs, housing costs and criminal justice costs. During consultation, we received a report produced by the Centre for Economics and Business Research (CEBR) which provided estimates of the cost to government of problem gambling related to B2 gaming machines (see paragraphs 24-28 for our consideration of the CEBR report).2 Key assumptions/sensitivities/risks Discount rate (%) ● Take up (proportion of players who remain playing under the new staking regime), session duration increase (how long each person plays for on average), channel shift (players move to other forms of gambling) and attrition (players stop gambling). ● The central assumptions are a cause of potential under or over estimation of costs. If take up, attrition or channel shift are higher or lower this can drastically change the outcome of the model. ● There is also an important implicit assumption that reduced spend on gambling will result in a reduction in gambling-related harm. ● There is the key risk associated with the fact it is difficult to quantify gambling harm and inherently problematic to conduct cost-benefit analysis on policies to reduce it.   3.5%    BUSINESS ASSESSMENT (Option 3) Direct impact on business (Equivalent Annual) £m: Costs: 240 2 Benefits: ​0 Net: -240 Score for Business Impact Target (qualifying provisions only) £m: 1190 ​http://www.coin-opcommunity.co.uk/wp-content/uploads/2018/01/CEBR-Report-final-Jan-2018.pdf 6 Summary: Analysis & Evidence Option 4 (Final proposal) Description: ​ Option 4: Reduce the maximum stake on B2 gaming machines from £100 to £2, Increase in prize bingo stake from £1 to £2 and prize increase from £70 to £100 (£500 to £1,000 aggregate), Status quo across all other gaming machine stakes and prizes and allocations. FULL ECONOMIC ASSESSMENT Price Base Year 2017     PV Base Year ​   2019   Time Period Years 10 ​     Net Benefit (Present Value (PV)) (£m) Low: -7750 COSTS (£m) Total Transition (​ Constant Price) Years High: -3100 Best Estimate: :-4660 Average Annual Total Cost (excl. Transition) (Constant Price) (Present Value) 360 3100 Low 2.1 High 5.4 900 7750 Best Estimate 3.8 540 4660     Description and scale of key monetised costs by ‘main affected groups’ The most significant monetised costs will be to gambling businesses and supply chains affected by the lowering of maximum stakes on B2 gaming machines from £100 to £2. Transition costs are established to comprise of one-off IT costs, training costs and marketing costs (see paragraphs 14-16) Other key non-monetised costs by ‘main affected groups’ Increase in session duration resulting from a stake reduction on B2 machines may reduce the availability of machines at peak times. This may impact B2 GGY if players cannot exercise their preference for machine play due to inadequate supply. There is also a cost associated with the dis-benefit reducing stake size has on non-problematic gamblers. This dis-benefit is due to a reduction in the utility these gamblers get from gambling at the new stake if their preference for staking behaviour is higher than the new cap. BENEFITS (£m) Total Transition (​ Constant Price) Years Low N/A     Average Annual Total Benefit (excl. Transition) (Constant Price) (Present Value) 0 0 7 High Best Estimate N/A 0 0    0   0 0 Description and scale of key monetised benefits by ‘main affected groups’ Other key non-monetised benefits by ‘main affected groups’ The most significant non-monetised benefits are the benefits to society of reduced gambling related harm. Also, as a result of any decrease in harm, there could be some benefit to government from reduced expenditure on government programmes linked to gambling-related harm such as healthcare (specifically psychological therapy), welfare costs, housing costs and criminal justice costs. During consultation, we received a report produced by the Centre for Economics and Business Research (CEBR) which provided estimates of the cost to government of problem gambling related to B2 gaming machines (see paragraphs 24-28 for our consideration of the CEBR report).3 There is likely to be some benefits to the gambling industry of the increases in stakes and prize on prize gaming but we have insufficient data to provide estimates of this. Key assumptions/sensitivities/risks Discount rate (%) ● Take up (proportion of players who remain playing under the new staking regime), session duration increase (how long each person plays for on average), channel shift (players move to other forms of gambling) and attrition (players stop gambling). ● The central assumptions are a cause of potential under or over estimation of costs. If take up, attrition or channel shift are higher or lower this can drastically change the outcome of the model. ● There is also an important implicit assumption that reduced spend on gambling will result in a reduction in gambling-related harm. ● There is the key risk associated with the fact it is difficult to quantify gambling harm and inherently problematic to conduct cost-benefit analysis on policies to reduce it.   3.5%    BUSINESS ASSESSMENT (Option 4) Direct impact on business (Equivalent Annual) £m: Costs: 450 3 Benefits: ​0 Net: -450 Score for Business Impact Target (qualifying provisions only) £m: 2270 ​http://www.coin-opcommunity.co.uk/wp-content/uploads/2018/01/CEBR-Report-final-Jan-2018.pdf 8 Evidence Base​ ​(for summary sheets) 1. Introduction Background All commercial gambling in Great Britain (with the exception of spread betting and the National Lottery) is regulated through the Gambling Act 2005 (‘the Act’), including gaming machines and all types of venues licensed to offer gaming machines. Latest Gambling Commission statistics show that the total Gross Gambling Yield for the gambling industry (including lotteries) in Great Britain, is £13.8bn4 ​(Apr 2016 – Mar 2017) (1.8% increase from Apr 2015 - Mar 2016).​ Under the Act gaming machines are defined by categories depending on the maximum stake and prize available: Table 1 4 Machine category Maximum stake Maximum prize Allowed premises B1 £5 £10,000 (£20,000 linked progressive jackpot on a premises basis) Casinos B2 £100 £500 Betting premises and tracks occupied by pool betting and all of the above B3 £2 £500 Bingo premises, Adult Gaming Centre and all of the above B3A £2 £500 Members’ club, commercial club or Miners’ welfare institute only B4 £2 £400 Members’ club or Miners’ welfare club, commercial club and all of the above. C £1 £100 FEC, Qualifying alcohol licensed premises and all of the above. D (money prize) 10p £5 Travelling fairs, unlicensed (permit) Family Entertainment Centre and all of the above ​http://live-gamblecom.cloud.contensis.com/PDF/survey-data/Gambling-industry-statistics.pdf 9 D non-money prize (other than crane grab machine) 30p £8 All of the above D non-money prize (crane grab machine) £1 £50 All of the above D combined money and non-money prize (other than coin pusher or penny falls machines) 10p £8 (of which no more than £5 may be a money prize) All of the above D combined money and non-money prize (coin pusher or penny falls machine) 20p £20 (of which no more than £10 may be a money prize) All of the above The Gambling Commission’s Industry Statistics5 2011 – 2016 (published in November 2017) in regards to gaming machines are summarised below. This data does not include gaming machines in pubs as these premises are licensed by local authorities rather than the Gambling Commission. Table 2: Gaming Machines: Average Gaming Machine Numbers Machine category Apr 2011-Mar 2012 Apr 2012-Mar 2013 Apr 2013-Mar 2014 Apr 2014-Mar 2015 April 2015-Mar 2016 April 2016-Mar 2017 B1 2788 2675 2657 2630 2780 2979 B2 33350 33467 34549 34949 34653 33611 B3 13496 15653 17303 18599 22317 24008 B4 256 232 219 311 248 168 C 46377 49835 61930 73005 76595 81713 D 65021 65751 53982 43298 43048 40437 161287 167614 170640 172793 179641 182916 Total Gross Gambling Yield (GGY) represents the amount retained by an operator after prizes are paid. GGY from gaming machines has increased since 2008-09. Category B2 5 http://www.gamblingcommission.gov.uk/news-action-and-statistics/Statistics-and-research/Statistics/I ndustry-statistics.asp​x 10 gaming machines contributed over £1.8bn in GGY to the betting industry in 2016/17. (However, there is approximately 30% of B3 play included in the industry statistics as B2.)6 Table 3 shows gaming machine GGY in the gambling sector over a five-year period. Table 3: Gaming Machines across whole gambling sector (minus pubs): Gaming Machine GGY (Gross Gambling Yield)(£m) Machine category Apr 2011-Mar 2012 Apr 2012-Mar 2013 Apr 2013-Mar 2014 Apr 2014-Mar 2015 April 2015-Mar 2016 B1 126.27 130.11 145.88 157.50 179.07 192.27 B2 1458.45 1547.83 1577.27 1692.59 1759.59 1815.34 B3 192.10 262.56 303.23 326.76 359.39 390.43 B4 1.68 1.43 0.94 1.01 1.04 1.17 173.08 229.93 220.90 224.74 235.48 233.50 95.17 105.42 101.93 103.55 103.26 99.53 187.01 57.53 48.84 38.88 34.05 9.40 2233.77 2334.81 2398.99 2544.53 2671.89 2741.65 C D Aggregated categories Total April 2016-Mar 2017 Rationale for Intervention Context Prior to the Act, a “Triennial Review” of the maximum stake and prize limits across different types of gaming machines became established, with the Gaming Board of Great Britain advising Government on proposals from the gambling industry. The review mainly considered whether increases were required to keep stake and prize levels in line with inflation. Since the Act there have been periodic reviews of stakes and prizes across all gaming machines defined under the Act. This review also includes a review of gaming machine allocations permitted under the Act. In 2013, as part of its drive to create the conditions for growth in the gambling industry, the then government implemented an approach to stake and prize regulation, based on the previous triennial review system. As part of this new approach, it asked industry and other stakeholders to put forward proposals for changes to stake and prize limits for consideration. 6 Due to the way in which Gambling Commission data is collected, the £1.8bn figure overstates GGY attributable to B2 content with approx 30% of the £1.8bn being attributable to B3 content. 11 The 2013 review7 concluded that whilst it was clear that reducing stakes on B2 machines would have an adverse economic impact on the betting industry, it was not clear how great an impact a reduction would have on problem gambling. The Gambling Commission and the Responsible Gambling Strategy Board (RGSB), in their advice to the then government, made clear that the industry must lead in making data available on player behaviour in order to support robust, independent research into B2 gaming machines. It also set the industry the challenge of making progress on developing harm mitigation measures. In response, the then government made clear that they would keep this machine category under review. Since this point, a number of studies have been published on B2 gaming machines using industry data and industry, including the Association of British Bookmakers (ABB), have introduced a range of player protection measures under their social responsibility codes. In October 2016 the government published a call for evidence8 which invited evidence-based proposals from all interested parties on changes to maximum stakes and prizes and allocations across all gaming machines under the objective of balancing socially responsible growth with the protection of consumers and wider communities. The call for evidence also invited proposals in regards to social responsibility measures and advertising which will not be covered in this assessment. In October 2017, we published a consultation document which outlined the limitations of self regulation and our continued concerns related to gambling-related harm on B2 gaming machines. We set out the evidence we had received on this issue at the call for evidence stage: ● consistently high rates of prevalence of problem gamblers among machine players in betting shops (our best proxy for B2 gaming machines) (11.5% of players are problem gamblers and a further 32% are considered at risk of harm);9 ● a high proportion of gross expenditure on machines in betting shops is attributed to problem gamblers;10 ● a high proportion of the number of problem gamblers who receive treatment identify machines in betting shops as their main form of gambling;11 ● the high-staking nature of B2 machines that offer a maximum stake of up to £100 can lead to significant losses in a short space of time. In comparison to other gaming machines, B2 machines generate a greater proportion and volume of large-scale losses (for example, more than £500 in a session);12 7 https://www.gov.uk/government/uploads/system/uploads/attachment_data/file/249311/Government_ Response_to_Consultation_on_Gaming_Machine_Stake_and_Prize_Limits_FINAL.docx.pdf 8 https://www.gov.uk/government/uploads/system/uploads/attachment_data/file/562122/Call_for_evide nce_-_Review_of_Gaming_Machines_and_Social_Responsibility_Measures.pdf 9 Health survey for England and Scotland 2012 showed that problem gambling rate was 7.2% rate amongst machine players in LBOs (of which B2s are the predominant machine). NatCen data for England, Scotland and Wales for 2015 showed that this figure had increased to 11.5% though this change was not considered statistically significant. http://www.gamblingcommission.gov.uk/PDF/survey-data/Gambling-behaviour-in-Great-Britain-2015.pdf 10 ​http://about.gambleaware.org/media/1259/natcen-secondary-analysis-of-loyalty-card-survey-final.pdf​ p.6 ​http://www.gamcare.org.uk/publications/annual-reviews-and-statistics http://www.gamblingcommission.gov.uk/news-action-and-statistics/news/2017/New-data-to-inform-g overnment-g ambling-review.aspx 11 12 12 ● ● ● losses are larger and sessions longer for those who bet at the current maximum stake (£100) than those who play at a lower level. The amount of money lost in a session and length of sessions are good proxies for gambling-related harm, and such losses might be harmful even to those who would not be defined by a survey screen as problem gamblers;13 problem gamblers are disproportionately found at higher stakes and are more frequent users of the maximum stake;14 the above factors are amplified by the concentration of betting shops (and therefore B2 machines) in areas of high deprivation.15 At consultation stage we set out proposals to reduce the maximum stake on B2 gaming machines from £100 to somewhere between £50 and £2, and set out four illustrative options: £50, £30, £20 (and £2 on slots) and £2. However, we acknowledged that depending on what evidence we received during the consultation, the final proposal may end up anywhere between £50 and £2. We also said that a stake reduction could be accompanied by additional measures to reduce harm, including: changes to the speed of play, a way of tracking and monitoring play on B2 gaming machines, and measures to improve player control. We also acknowledged that there is significant uncertainty about what players will do in response to a stake reduction and consequences of displacement are unknown. However, we are clear that the potential for displacement should not prevent taking action in regards to specific gambling products if there is a body of evidence to suggest they are associated with more harm than others. We are clear that if other forms of gambling are harmful to an unacceptable extent, then action should be taken on them as well. Nor does it necessarily follow that displaced gamblers will be exposed to the same or higher levels of harm. And the reason we have focused on B2 gaming machines is because of the potential for more harm on this product over others. However, noting both Gambling Commission and RGSB advice, we will carefully monitor the impact of changes to the maximum stake so that we can better understand displacement and the potential risks around this. Problem under consideration Gambling-related harm produces several negative externalities including but not limited to: increased healthcare costs, welfare costs, and other costs to individuals associated with problem gamblers (e.g. family, friends and employers). Therefore, the marginal social cost is higher than the marginal private cost to the gambler. Government intervention is necessary to ensure we strike the right balance between socially responsible growth and the protection of consumers and wider communities. Unfortunately due to the nature of gambling-related harm (extremely inelastic demand), it is difficult to dis-incentivise consumption of this good therefore the burden of reducing harm must be placed on the suppliers of the good. 13 Ibid ​http://about.gambleaware.org/media/1259/natcen-secondary-analysis-of-loyalty-card-survey-final.pdf 15 Contextualising machine gambling characteristics by location - final report - A spatial investigation of machines in bookmakers using industry data, Geofutures, 2015 14 13 Price elasticity of demand is a measure used in economics to show the responsiveness of the quantity demanded of a good or service to a change in its price (everything else remaining equal). ​Table 4 illustrates estimates of the elasticity of demand for gaming machines by academia as summarised in: The UK betting and gaming market: estimating price elasticities of demand and understanding the use of promotions16, a report prepared for HM Revenue and Customs in June 2014 by Frontier Economics Limited. Evidence from the UK is scarce. Frontier Economics’ preferred estimate for own-price elasticity of demand for gaming machines was established to be own price inelastic (-0.6). We assume that the price elasticity of demand associated with ​problem gambling​ on B2 gaming machines is extremely inelastic, due to the intrinsic nature of gambling addiction. ​Whilst this information is included for illustrative purposes, we develop assumptions from alternative sources to model the impacts of the proposed measures, which are laid out below in paragraph 1-16 in Section 2: Cost & Benefits. Table 4: Estimates of the Price Elasticity of Demand for Gambling Machines Gambling-related harm is wider than problem gambling, but based on the latest ​Health Survey data for England (HSE) 2015 and the Scottish Health Survey (SHeS) 2015, combined with Gambling Commission data for Wales17, problem gambling rates were 0.8% of the adult population in 2015. The same data found that problem gambling rates amongst players of ‘gaming machines in bookmakers’ (our closest proxy for B2 gaming machines) is 11.5%. Although not possible to quantify accurately due to lack of data, a recent study by the Institute for Public Policy and Research (IPPR) estimated that the current population of 16 17 ​https://www.gov.uk/government/uploads/system/uploads/attachment_data/file/322845/report313.pdf ​http://natcen.ac.uk/media/1464625/gambling-behaviour-in-great-britain-2015.pd​f 14 problem gamblers are associated with between £260 million and £1.2 billion a year of extra cost to government. These costs were spread across health (primary care, mental health services, secondary mental health services and hospital inpatient services), welfare and employment (Job seeker allowance claimants and lost labour tax receipts), housing (statutory homelessness applications), and criminal justice (incarcerations)18. The IPPR caveat these estimates with the following: “Due to limitations in the available data, these findings should not be taken as the excess fiscal cost caused by problem gambling. Instead, they should be taken as an illustrative estimate for the excess fiscal costs incurred by people who are problem gamblers, beyond those that are incurred by otherwise similar members of the population. And due to variations in the quality of data for different areas of interaction, the methods for estimating excess incidence and unit costs are not directly comparable across different interactions.” In addition, it is important to clarify that these are the costs associated with people who are classified as problem gamblers, not costs directly attributable or caused by their gambling habits. And the report makes no attempt to estimate how much of those costs might be reduced if those individuals were not problem gamblers. Building on this analysis, ​a report produced by the Centre for Economics and Business Research (CEBR) provided estimates of the cost to government of problem gambling specifically related to B2 gaming machines.19 This analysis estimates that problem gambling linked to B2 machines could be associated with an excess fiscal cost in the region of £210 million and in welfare terms, the population of B2 problem gamblers could be imposing a cost of £1.5 billion on themselves, their families and their wider social networks. As this analysis is based on the IPPR report, the same caveats as above apply. Despite our difficulty in quantifying, there are obvious social costs attributed to high stake game play on B2 machines, enabled by the propensity for a problem gambler to lose up to £100 per spin and thus large sums of money over a short space of time. The ability to lose large sums of money quickly is a proxy for harm, and a stake reduction is likely to reduce session losses and therefore the potential for harm and the associated costs. There is a programme of work led by the Responsible Gambling Strategy Board (RGSB) to better define and measure gambling-related harm which may improve the evidence base in the future. Alternatives to regulation The main alternative to regulation which we have assessed is voluntary regulation by industry and previous measures introduced on B2 gaming machines in 2015. A number of respondents to the call for evidence and consultation highlighted the perceived inadequacies of previous and current industry codes on social responsibility (self regulation), specifically on B2 gaming machines, primarily citing the lack of evidence of impact and effect of the measures. Where any evaluation has taken place, primarily of the measures introduced by the bookmakers on B2 gaming machines, it is not clear that the measures have been as effective as hoped. 18 19 ​https://www.ippr.org/publications/cards-on-the-table ​http://www.coin-opcommunity.co.uk/wp-content/uploads/2018/01/CEBR-Report-final-Jan-2018.pdf 15 The evaluation of the ABB’s code on social responsibility, of which the headline measure related to the introduction of measures that allowed players to set the time and total expenditure they wished to spend on B2 gaming machines, was published in May 2015 and concluded that only 0.5% of machine sessions in the first month after implementation included the use of time and spend limits.20 But they could not establish if this was because players did not want to use the function, or did not know about it. Due to the small proportion of sessions that included a voluntarily set threshold they were unable to draw any conclusions on the impact of this tool on player’s behaviour. In addition, the evaluation of the Player Awareness System (PAS) rolled out by ABB members on B2 gaming machines was published in October 2016.21 It found that although this measure had potential, there was a considerable way to go before it could be considered successful. At consultation we recognised the effort and resource being put into responsible gambling activities across the industry as a whole, including but not limited to the measures set out above. However, we said that there is a need for considerable improvement in the methods of identifying harmful play on gaming machines, and the development of interventions to help players who might be suffering harm, including across other categories of gaming machine. The recent RGSB publication analysing industry progress echoes this, concluding that “there is still much to do if the Strategy is to make visible progress towards its objectives”, with a need to increase the pace of delivery over the next 12 months.22 In addition to these voluntary measures, in 2015 the then government introduced a new requirement that those accessing higher stakes (over £50) on B2 gaming machines must load cash via staff interaction or use account-based play (i.e. set up an account with the operator which required the use of an account card to play the machine). The evaluation of these measures, published in January 2016, found that there was a drop in stakes above £50, and a corresponding increase in stakes between £40-50 resulting in an overall small reduction in amount lost by those gambling on B2 machines.23 The evaluation found that the measure may also have increased the level of control exercised by those gambling on B2s but was not able to provide an analysis on the impact on harm. Objectives The government’s objective is to look across the industry and determine what, if any, changes are needed to strike the right balance between socially responsible growth and the protection of consumers and wider communities. Underlying this objective is the government’s aim to reduce gambling related harm. ​https://about.gambleaware.org/media/1167/abb-early-impact-report-final-report.pdf ​https://about.gambleaware.org/media/1335/pas-evaluation_final-report_13102016.pdf 22 ​http://www.rgsb.org.uk/PDF/Strategy-progress-report-2016-2017.pdf 23 https://www.gov.uk/government/uploads/system/uploads/attachment_data/file/493714/Evaluation_of _Gaming_Machine__Circumstances_of_Use___Amendment__Regulations_2015.pdf 20 21 16 Option Considerations Following analysis of responses to the call for evidence the government consulted on a number of illustrative options, making clear that the resulting package of measures may ultimately differ from those presented in the consultation stage impact assessment. The illustrative options presented in the consultation impact assessment compared to what has been taken forward for the purposes of this final impact assessment can be summarised as follows: Table 5 - Illustrative options set out in the consultation stage impact assessment v.s. options brought forward for final consideration Option Consultation Stage Impact Assessment Final Impact Assessment stage Option 1 Status Quo Taken forward (counterfactual) Option 2 Industry Proposals Discounted Option 3 Illustrative option including: Reducing maximum stake on B2 gaming machines to £50 Discounted Option 4 Illustrative option including: Reducing maximum stake on B2 gaming machines to £30 Taken forward for consideration in final impact assessment (Re-classified as ​Option 2​) Option 5 Option 6 Illustrative option including: Reducing maximum stake on B2 gaming machines to £20 (non slots) and £2 (slots) Taken forward for consideration in final impact assessment. 24 (Re-classified as ​Option 3​) Illustrative option including: Reducing maximum stake on B2 gaming machines to £2 Final proposal (Re-classified as ​Option 4​) 24 B2 gaming machines offer different gaming content, the most popular of which is roulette. By non-slots we are referring to virtual casino or sports games, such as roulette. ​By slot game we are referring to: a game which is mechanical or virtual in nature; uses spinning reels, discs or other representation of moving or changing symbols (fruits, numbers or other graphics) where the object is to align those symbols across one or more lines to win a prize and/or advantage within the game; and where this can feature as a primary or secondary game component e.g. feature games which use spinning reels as a secondary part of the game. 17 Industry Proposals - Rationale for discounting During the call for evidence industry proposed a number of changes to the current regulatory regime. These proposals represent maintaining the status quo for some parts of the industry and relaxing the limits for other sectors of the industry and were set out in more detail in the consultation and consultation impact assessment published in October 2017. At consultation stage we received no further relevant data or new evidence of the likelihood of success of the voluntary proposals. Based on estimates that we received from industry during the call for evidence, this package of measures would create an annual economic benefit of £432 million to the gambling industry (rounded to the nearest £1 million), primarily affecting the casino sector, with some additional benefits to the arcade and pub sectors. This estimate is based on assessments commissioned by industry and provided to the department at the call for evidence stage. We note that there are also potential supply chain benefits from increased investment in new machines and software and there are also potential consumer demand benefits due to higher prizes, but due to the lack of data surrounding consumer preferences and industry investment plans these are non-quantifiable. However, the likely increase in consumption following stake and prize increases and increased availability of gaming machines, may lead to increased levels of gambling-related harm, and therefore ​negative externalities including but not limited to: increased healthcare costs, welfare costs, and other costs to individuals associated with problem gamblers (e.g. family, friends and employers). Following consultation, government maintains that this option is unlikely to meet the primary objective of the review of balancing socially responsible growth with the protection of consumers, and will not be taking it forward. Reducing maximum stake on B2 gaming machines to £50 - Rationale for discounting At consultation, we put forward an illustrative option to reduce the maximum stake to £50. We estimate that this package of measures will have an annual cost of £27 million25 (rounded to the nearest £1 million), primarily affecting the betting sector, with subsequent impacts on gaming machine manufacturers and supply chain for which we cannot provide a monetised impact. In their advice to the department the Gambling Commission stated that they agreed with the Responsible Gambling Strategy Board’s (RGSB) advice that there is a precautionary case for a stake cut below £50. £50 is the effective limit for most players with only a tiny proportion of play above this level of stake. The Gambling Commission state that they think the maximum stake should be reduced to between £2 and £30 if it is to have a significant effect on the potential for players to lose large amounts of money in a short space of time. 25 This figure has been updated with changes in assumptions for purposes of the final impact assessment. The figure estimated in the consultation impact assessment for this option was £35 million (rounded to the nearest £1 million) 18 We agree with the primary point that £50 equates to the de facto status quo and a stake cut to this level would therefore have little to no effect on reducing harm to the player. For these reasons, we are discounting this option. Final Options Option One: Retain status quo Option one proposes that no changes are made to stake and prize limits for any gaming machine category or to any allocations of gaming machines currently permitted under the Act. In the absence of any changes to stakes and prizes, it is reasonable to assume growth of the machine gambling sector to continue along the current trajectory (machine GGY has increased by 4.3% a year on average since 2013 (last review)). Following consultation, the government maintains that option one will not meet the primary objective of the review - to strike the right balance between socially responsible growth and the protection of consumers and wider communities. Option Two: Reducing maximum stake on B2 gaming machines to £30 At consultation, we put forward an illustrative option to reduce the maximum stake to £30. We estimate that this package of measures will have an annual cost of £170 million, primarily affecting the betting sector, with possible impacts on gaming machine manufacturers and supply chain. In their advice to the department, the Gambling Commission stated that the maximum stake should be reduced to between £2 and £30 if it is to have a significant effect on the potential for players to lose large amounts of money in a short space of time. We have concerns that at £30, players are still able to lose large sums of money quickly, with industry data showing that players can lose in excess of £1000 in a single session with average stakes up to £30 and of those sessions in which players lost more than £500, around 20% involved average stakes up to £30.26 In addition, we have concerns that there appear to remain a high proportion of problem gamblers at this staking level (42% at £30 or more compared to 19% at £2 or less)​.27 ​We think that the social cost at this staking level would be magnified by the fact that ​players of B2s are more likely to be unemployed/low earners than those who opt for different forms of gambling and tend to live in areas with greater levels of income deprivation.28 26 http://www.gamblingcommission.gov.uk/news-action-and-statistics/news/2017/New-data-to-inform-g overnment-gambling-review.aspx 27 http://www.rgsb.org.uk/PDF/Advice-in-relation-to-the-DCMS-review-of-gaming-machines-and-socialresponsibility-measures.pd​f 28 Contextualising machine gambling characteristics by location - final report - A spatial investigation of machines in bookmakers using industry data, Geofutures, 2015 & https://about.gambleaware.org/media/1259/natcen-secondary-analysis-of-loyalty-card-survey-final.pdf 19 Option Three: Reducing maximum stake on B2 gaming machines to £20 (£2 on slots) At consultation, we put forward an illustrative option to reduce the maximum stake to £20, with a further reduction to £2 on slots. We estimate that this package of measures will have an annual cost of £280 million, primarily affecting the betting sector, with subsequent impacts on gaming machine manufacturers and supply chain for which we cannot provide a monetised impact. We have concerns that at £20, players are still able to lose large sums of money quickly, with industry data showing that players can lose in excess of £1000 in a single session with average stakes up to £20 and of those sessions in which players lost more than £500, around 7% involved average stakes upto £20.29 In addition to the Gambling Commission advice set out above, based on the data we have access to, we have similar concerns with this option that there appear to remain a high proportion of problem gamblers at this staking level (42% at £20 or more compared with 19% at £2 or less)​.30 ​We think that the social cost at this staking level would be magnified by the fact that ​players of B2s are more likely to be unemployed/low earners than those who opt for different forms of gambling and tend to live in areas with greater levels of income deprivation31. In regards to B2 slots, we note Gambling Commission advice which states that they "think there is a case for a £2 maximum stake for B2 slots" reflecting risks associated with this content including lower return to players and less opportunities for players to manage their own risks through the way they play. These characteristics are reflected in the outcomes that we see in gaming machine data for slots play in which it is apparent that they give rise to a greater proportion of significant losses to the player than other B2 content. We support this position in regards to slots which we have set out under option 4 below. Option Four: (Final proposal) Reducing maximum stake on B2 gaming machines to £2 We noted in the consultation impact assessment that a majority of respondents to the call for evidence who raised the issue of stake on B2 gaming machines highlighted the discrepancy between the current maximum stake of £100 and the maximum stake on all other accessible gaming machines on the high street of £2. Respondents also highlighted evidence suggesting that £2 would reduce harm to the player and to wider communities. At consultation stage, a majority of respondents also supported this illustrative option. A detailed summary of responses is set out in the consultation response. The government's preferred option is to reduce the maximum stake on B2 gaming machines from £100 to £2 and to maintain the status quo on all other gaming machine stakes and prizes and allocations, with the exception of prize gaming. We estimate that this package of 29 http://www.gamblingcommission.gov.uk/news-action-and-statistics/news/2017/New-data-to-inform-g overnment-gambling-review.aspx 30 http://www.rgsb.org.uk/PDF/Advice-in-relation-to-the-DCMS-review-of-gaming-machines-and-socialresponsibility-measures.pdf 31 Contextualising machine gambling characteristics by location - final report - A spatial investigation of machines in bookmakers using industry data, Geofutures, 2015 & https://about.gambleaware.org/media/1259/natcen-secondary-analysis-of-loyalty-card-survey-final.pdf 20 measures will have an annual cost to business of £540 million, primarily affecting the betting sector, with possible impacts on gaming machine manufacturers and supply chain. Taking into account both the Responsible Gambling Strategy Board’s (RGSB) and Gambling Commission’s advice, we are clear that a reduction in maximum stake on B2 gaming machines is justified in order to reduce gambling-related harm and therefore protect the consumer. At a revised level of £2, we think that the volume of high level session losses would be significantly reduced - current industry data shows that there are no sessions with losses over £1000 with average stakes up to £2 and that there are less than 0.0% of sessions with losses over £500 with average stakes up to £2. Even at stakes as low as £10, players can lose in excess of £1000 (0.2% of sessions that result in losses between £1000.01 and £5000 involve average stakes at £10 or below), and there are a greater proportion and overall volume of losses greater than £500 (5794 sessions - 0.9% - of sessions that result in losses between £500.01 and £1000 involve average stakes at £10 or below). And we think that a £2 level is more likely to target the greatest proportion of problem gamblers and to protect the most vulnerable players, such as those who are unemployed and those in more deprived areas. As a result of any decrease in harm, there are likely to be some benefit to the taxpayer from reduced expenditure on government programmes linked to gambling-related harm such as healthcare (specifically psychological therapy), welfare costs, housing costs and criminal justice costs. The CEBR report received during consultation provided estimates of the societal costs of problem gambling related to B2 gaming machines. The estimates they provide for potential reductions in gambling-related harm following a maximum stake cap of £2 on B2 gaming machines of between £430 million and £1.3 billion per year could possibly weigh positively against our cost estimates (although we do not adopt them for the purposes of our impact assessment, see paragraph 24-28 for consideration of the CEBR report)32. This option would also see an increase stake from £1 to £2 and prize increase to £100 (£1,000 aggregate) on prize gaming and no changes to stake and prizes and allocation for other categories of gaming machines. ​There is likely to be some benefits to the gambling industry of the increases in stakes and prize on prize gaming but we have insufficient data to provide estimates of this - industry could not provide estimates of economic benefits and we do not have data to model this. ​We have considered all consultation responses and are content that an uplift to stake and prize on prize gaming is in keeping with the objective of this review because activities associated with prize gaming are low risk. It provides for a more elderly clientele a longer, more sociable opportunity, akin to bingo, but at low stake and prize levels in a more convenient location. The Gambling Commission will be asked to monitor any potential risks that arise as an outcome of these changes as part of our Monitoring & Evaluation strategy. 32 ​http://www.coin-opcommunity.co.uk/wp-content/uploads/2018/01/CEBR-Report-final-Jan-2018.pdf 21 2. Cost & Benefits Methodology 1. This section of the impact assessment assesses the likely benefits and costs that will accrue to different groups as a result of implementing the government’s policy proposals. Before proceeding to present an analysis of the benefits and costs, it is important to be clear about the methodological basis for appraisal. This needs to take account structure of analysis, proportionality, and technical parameters. Key areas of impact and the structure of analysis 2. The government’s objective is to look across the industry and determine what, if any, changes are needed to strike the right balance between socially responsible growth and the protection of consumers and wider communities. It is anticipated that, if players are unable to stake as much within the same duration of their gambling play, there is likely to be a reduction in overall GGY on B2 machines and a corresponding impact on industry GGY. For the purpose of the impact assessment, these changes must be considered from both economic and social perspectives. Proportionality 3. These impacts should be assessed to a level of analytical detail that is proportionate to the intervention being made. There are different factors to take into account when considering proportionality. Firstly, the policy is not irreversible because stake and prize limits for all categories of gaming machine are subject to regular review. Secondly, the regulatory intervention and associated distribution of impacts seeks to strike a balance between allowing normal leisure gamblers use of a product while protecting those who may be at risk. The impact assessment uses existing evidence gathered during a call for evidence and consultation in combination with contributions from industry and sector specialists. Presentation of the status quo and other technical issues 4. There are a number of presentational and technical points that apply across different policy options. The status quo option represents no change and therefore for the purposes of appraisal does not introduce any new benefits or costs that might affect the existing baseline. All monetised impacts are presented in present value terms unless otherwise stated, discounted at the Green Book determined rate of 3.50% per annum. All prices and monetised impacts are presented with a ​2017​ price base year​, 2019 present value base year​ unless otherwise stated. Framework for analysis 5. To be able to appraise the policy proposal regarding maximum stake size on B2 gaming machines there needs to be a mechanism for relating the proposed regulatory intervention to changes in industry GGY. The mechanism through which it operates is broadly as follows: 22 a. Percentage of High Staking Players Reducing Stakes to the New Cap: ​This is the primary mechanism through which the impacts on industry revenues will be realised. The model assumes that there will be some players who will stake down to the new maximum stake and a proportion of players who will either stop gambling or gamble through other channels. On this logic, it follows that reduced spend per session is likely to lead to a corresponding decrease in machine GGY. The extent to which people adapt their gambling behaviour as they play at different staking levels is uncertain. b. Level of Session Adjustment:​ ​GGY losses might be dampened if some players choose to gamble at a lower stake limit for a longer period. However, those who do not adjust their sessions in this way will on average spend less per session than they were before leading to a negative impact on GGY. c. Attrition from Gambling: ​GGY is lost from players who choose not to gamble at the new cap and choose not to migrate their spend elsewhere. For these players reducing the stake cap on B2 machines takes this spend out of the gambling industry as they choose to not gamble this money. d. Gambling through other Channels: ​GGY losses are mitigated when players who choose not to play at the new stake limit migrate their spend to other forms of gambling. It is assumed that for those who gamble through multiple channels currently, migrating spend from one to another will be far more likely than for those players who solely engage in B2 gambling. The assumptions made under paragraphs 5 (a-d) are informed by industry estimates and data provided by the Gambling Commission but have been analysed and modified by DCMS where appropriate. This comes with the caveat that for some areas, there is a lack of data to better inform how players will respond to a stake reduction and are ultimately based on assumptions about changes in player behaviour. Changes to these assumptions can have significant impacts on the estimated impact on industry and proportionate sensitivity analysis has been conducted to ensure assumptions are robustly tested. Estimating the Baseline 6. Industry statistics, collected by the Gambling Commission, show that B2 gaming machines in licensed betting offices (LBOs) generated £1.8bn in GGY (16/17). However, players can access other types of gaming machine content on these terminals (primarily B3) so we know that of this £1.8bn, a proportion (around 30%) is attributable to other gaming machine content. This is because industry data is categorised under the highest category of game available on that particular terminal. For example, if a terminal offers category B2 and B3 games, operators attribute all GGY to the B2 machine category. Using Gambling Commission data will therefore overestimate machine GGY on B2 gaming machines. 23 7. We therefore requested more granular data from the gaming machine suppliers (SG and Inspired Gaming) which provides a breakdown of GGY by content as well as more detail on staking patterns which we have used to develop a more robust baseline. GGY is derived for roulette games and slot games so that separate impacts on GGY can be observed with varying maximum stake caps for slot games as opposed to roulette games. The raw data is broken down into staking bands of tens of pounds which we use to structure our assumptions regarding behavioural change. Due to lack of time series data for roulette and slot games on B2 gaming machines, due to lack of historic data by banding, and because the gambling industry is a dynamic landscape with fast moving technological innovation and changing consumer preferences this baseline is assumed to be flatlined across the 10 year appraisal period. Table 6: Stake Bands Calculating the impact on GGY: Assumptions 8. The cost of the different stake cap options is calculated through obtaining the net impact of B2 gaming machine GGY lost directly from imposing the stake cap on both slot and roulette games, both through reduced GGY at the lower stake cap and through attrition, countered against recaptured GGY from displacement to other forms of gambling such as over the counter products, B3 displacement and online play to estimate net GGY losses. 9. We assume player behaviour varies depending on the extent to which a stake cap would force the player to change their staking behaviour. For the B2 model we separate this into three groups; those staking 3 bands (bands of £10s) or higher than the stake cap, those staking 2 bands above the proposed cap and those staking in the band above the proposed cap. Below is a series of tables detailing the differing assumptions for each group. 24 Table 7 - Summary of assumptions underpinning B2 model Assumptions of Model - 3 Bands+ above cap Central High Low Take up of lower stake 40% 80% 20% Level of session adjustment 28% 50% 0% Attrition 10% 40% 0% Assumptions of Model - 2 Bands above cap Central High Low Take up of lower stake 51% 80% 40% Level of session adjustment 28% 50% 0% Attrition 10% 30% 0% Assumptions of Model - 1 Band above cap Central High Low Take up of lower stake 68% 90% 50% Level of session adjustment 28% 50% 0% Attrition 10% 20% 0% 10. Take up of lower stake​: This assumption was informed by a report by KPMG on behalf of the Association of British Bookmakers (ABB). Through industry consultation they estimated that, dependent on current stake level, a proportion of customers have a general preference for maintaining machine play at the new stake. We tested this assumption with sector specialists in the Gambling Commission. The extent to which people adapt their gambling behaviour as they play at different staking levels is uncertain. On this point, during consultation we received alternative analyses by the Centre for Business and Economic Research (CEBR)33 and from NERA economic consulting specifically in regards to the 33 ​http://www.coin-opcommunity.co.uk/wp-content/uploads/2018/01/CEBR-Report-final-Jan-2018.pdf 25 modelling of the impact of a stake cut on B2 gaming machines. In summary, both reports estimated that the impact to the industry would be less substantial than what was set out in the consultation impact assessment. The returns stipulate that there’s likely to be greater displacement to other forms of gambling, therefore reducing overall impact on that sector. Both reports acknowledge that player response is difficult to predict. We acknowledge the uncertainty involved in accounting for player response to a stake reduction and in assessing consultation responses, including these publications, as well as previous evaluations of gaming machine measures we have amended our assumption regarding Take up of Lower Stake. We have reduced this assumption down by 10% across the board, based on information that suggested that there may be a greater shift of B2 users displaced onto other substitutes - particularly online play - as a result of the stake cap. Take up of lower stake is an independent variable which impacts the estimate for displacement. 11. Level of session adjustment:​ The central estimate of level of session adjustment is derived from gaming machine data provided to the department to inform the evaluation of the £50 regulations published in January 201634. We haven’t received any information following the consultation which would inform a change in this assumption. 12. Attrition​: The level of attrition, that is, the people who will no longer take part in any gambling activity as a result of a stake cap on B2 machines is extremely difficult to predict. A conservative estimate of 10% was established for the central scenario, informed by KPMG’s report for the ABB. This assumption was agreed with the industry and was informed by their analysis of the effect of implementation of the Gaming Machine (Circumstances of Use) (Amendment) regulations on 6th April 2015. This amendment required that those accessing higher stakes (over £50) load cash via staff interaction or use account based play. Due to inherent uncertainty in behavioural change, we’ve conducted sensitivity analysis testing different scenarios for attrition (see below). Attrition, in conjunction with Take up of lower stake is an independent variable which informs displacement. 13. Displacement to other forms of gambling​: Displacement is the dependent variable in the calculation to measure potential GGY change, calculated after taking the proportion of take up of lower stake (with session increase) on B2 gaming machines and attrition into account. What is left is the proportion displaced into spending on other forms of gambling, thus recouping GGY for betting companies. Table 8 shows the proportions calculated to be displaced to other forms of gambling once other assumptions have been accounted for. It is assumed that the equivalent amount of GGY of this proportion previously spent on B2 machines, is alternatively displaced into other forms of gambling. 34 https://www.gov.uk/government/uploads/system/uploads/attachment_data/file/493714/Evaluation_of _Gaming_Machine__Circumstances_of_Use___Amendment__Regulations_2015.pdf 26 Table 8: Percentage calculated to be displaced to other forms of gambling for each band Figures rounded to the nearest 10% Table 9: Model Outputs: Net impact (per annum) - Central Estimates Option 2 Option 3 Option 4 Roulette - GGY loss from cap -£380 million -£540 million - £930 million Slots - GGY loss from cap - £20 million - £50 million - £50 million GGY Recouped from displacement £220 million Net Impact on GGY - £170 million £310 million £440 million - £280 million - £540 million Figures rounded to the nearest £10 million, totals may not sum due to rounding Transition Costs 14. During the consultation we did not receive any publishable data regarding potential transition costs for industry. 15. In the absence of obtaining publishable costs from the consultation, on 6th April 2015 the previous government implemented the Gaming Machine (Circumstances of Use) (Amendment) regulations 201535 which implemented regulatory changes to B2 gaming machines, requiring programming updates not dissimilar to what is being proposed in this impact assessment. ​This amendment required that those accessing higher stakes (over £50) load cash via staff interaction or use account based play.​ In that instance, the three largest operators indicated one-off IT costs from approximately £0.4 million to approximately £1.0 million as a result of the regulation. Estimates from these operators for training costs were between £0.7 million and £1.0 million. Major operators indicated marketing costs from approximately £1.0 million to £3.4 million, the caveat in the evaluation was that it wasn’t clear how much of this cost can be attributed as an unavoidable cost due to the regulations. 35 https://assets.publishing.service.gov.uk/government/uploads/system/uploads/attachment_data/file/4 93714/Evaluation_of_Gaming_Machine__Circumstances_of_Use___Amendment__Regulations_201 5.pdf 27 16. Combining the lower estimates for the three categories of costs suggests a lower estimate for the one-off cost to major operators of approximately £2.1m. Combining the upper estimates for the three categories of one-off costs gives an upper estimate of £5.4m. This gives a range of £2.1 million to £5.4 million. The mid-point was taken at £3.8 million for purposes of obtaining a suitable central estimate. Because option 2 - 4 would involve similar alterations to B2 gaming machines, we assume that these transition costs are applicable across all options. Costs Option 1 17. Option 1 is the counterfactual which we assess our options against. There are no costs associated with maintaining the status quo. The baseline of GGY from B2 gaming machines is established in paragraphs 6-7. Option 2 18. B2 stake limited to £30: Central estimate of cost to industry is approximately £170m per year. Low Central £140 million High £170 million £250 million Figures rounded to the nearest £10 million Option 3 19. B2 stake limited to £20 (£2 on slots): Central estimate of cost to industry is approximately £280m per year. Low Central £200 million High £280 million £480 million Figures rounded to the nearest £10 million Option 4 (Final Proposal) 20. B2 stake limited to £2: Central estimate of cost to industry is approximately £540m per year. 28 Low Central £360 million High £540 million £900 million Figures rounded to the nearest £10 million 21. Pursuing Option 4 creates a central estimated cost in GGY of £540m per year, which represents a cost of £4,660m in net present value terms over a 10 year appraisal period. This is a slightly lower estimate than what was provided during the consultation stage impact assessment, which has changed in line with amended assumptions regarding reduced take up of lower stake but more moving to online play. Non-monetised costs: Options 2-4 22. There may a disbenefit associated with restricting gamblers ability to gamble in their preferred way. This is because non-problem gamblers gain utility from staking at the level they wish to, and any restriction of their ability to stake at the level they wish may lead to a loss of utility derived from gambling. Benefits 23. We expect there to be significant benefits to society as a result of a reduction in stake on B2 gaming machines which we haven’t been able to monetise. These benefits accrue via reduced gambling-related harm. It is impossible to accurately quantify these benefits given the data available but given the social costs of gambling-related harm we expect any action taken to limit the prevalence or harm from addiction to have significant positive social benefits. 24. The Centre for Economics and Business Research (CEBR) have attempted to quantify some of these social benefits36. 25. To produce indicative estimates of the possible societal impacts of a reduction in B2 maximum stakes, CEBR build on the evidence presented in the 2016 IPPR report titled “Cards on the table - The cost to Government associated with people who are problem gamblers in Britain”37. They used this to establish the share of the excess fiscal costs identified by IPPR that might reasonably be associated with problem gambling specifically linked to B2 gaming machines. 26. Separately to this, they estimate welfare impacts by drawing on the Housing Association’s Charitable Trust (HACT) and Simetrica “Community and investment and homelessness values from the Social Value Bank” database (see www.socialvaluebank.org). This resource provides data that can capture the monetary value of impacts that typically occur ‘outside the market’ (or, 36 37 ​http://www.coin-opcommunity.co.uk/wp-content/uploads/2018/01/CEBR-Report-final-Jan-2018.pdf ​https://www.ippr.org/publications/cards-on-the-table 29 equivalently, outside of the domain of macroeconomic indicators like GDP and employment). The Social Value Bank is used to obtain the value that the average person attributes to aspects such as having a job, being financially stable and having good health and we assume that these are valued in the same way by B2 problem gamblers as by any other member of society. 27. The CEBR analysis reveals that problem gambling linked to B2 machines could be associated with an excess fiscal cost in the region of £210 million. From this they derive an corresponding estimate of £1,723 per B2 problem gambler (their definition thereof). They stipulate that, in welfare terms, the population of B2 problem gamblers could be imposing a cost of £1.5 billion on themselves, their families and their wider social networks. They establish that this equates to about £13,780 per B2 problem gambler. 28. CEBR then apply percentage reductions of 25%, 50% and 75% to show possible magnitudinal reductions in the estimated costs of gambling-related harm following the policy change. The benefits drawn from implementing the £2 stake cap are estimated to be between £430 million and £1.3 billion per year. This range of benefits would suggest the costs estimated for options 2-4 would fall within the range of producing a positive cost-benefit ratios, and positive NPVs. However, we’ve included this information ​for illustrative purposes only​ to show what the benefits might be. As we’re unable to replicate CEBR’s analysis, or critique and test all the assumptions that underpin the calculations and outputs, we’re not in a position to adopt it for the purposes of our impact assessment. It is also impossible to predict what magnitude of change will be realised following a £2 maximum stake cap and speculative to suggest what is achievable. 30 Sensitivity Analysis 29. Due to the complex nature of predicting behavioural change, it is proportionate to conduct sensitivity analysis on key assumptions where possible. The main assumptions which demonstrate degrees of uncertainty, and, exhibit the greatest sensitivity to the outputs are ​Take Up of Lower Stake​ and ​Attrition​. Take Up of Lower Stake measures the proportion of players that “stake down” and continue on B2 gaming machines at the new stake cap. The less take up of lower stake, consequently the more displacement there is to other forms of gambling which lessens the cost to industry of a maximum stake cap. 31 Take up of lower stake Table 10: Central Assumptions Bands above cap Take up of lower stake 3 or more 40% 2 1 51% 68% 30. In table 11 and 12, 10% deviation from the central assumption in take up of the lower stake has been considered. Table 11: Take up of Lower Stake 10% less than central assumption Bands above stake cap Take up 3 or more 30% 2 41% 1 58% Central Estimate Option 2 Central Estimate Option 3 Central Estimate Option 4 £150 million £240 million £440 million Figures rounded to the nearest £10 million,attrition and session adjustment assumed to be fixed in line with central assumptions, 10% and 28% respectively Table 12: Take up of Lower Stake 10% more than central assumption Bands above stake cap Take up 3 or more 50% 2 61% 1 78% Central Estimate Option 2 Central Estimate Option 3 Central Estimate Option 4 £200 million £330 million £640 million Figures rounded to the nearest £10 million, attrition and session adjustment assumed to be fixed in line with central assumptions, 10% and 28% respectively 31. The higher the take up of lower stake, the less people there are being assumed to displace their B2 GGY to other forms of gambling. This increases the cost of the proposals as less GGY is created from B2s at the lower stake. Attrition 32. Attrition represents the estimated proportion of players that will no longer stake on B2s, nor alternatively spend GGY previously spent on B2 machines on other forms of gambling. This isn’t to say that they don’t already partake in other forms of gambling, but that the money they’re currently willing to pay on B2 machines is displaced into other goods and services (or invested). In the central estimate, we 32 assume attrition to be 10%. This is the product of consultation with sector specialists at the Gambling Commission and established in KPMG analysis which was informed by industry. However, if players spend a greater proportion of money previously spent on B2 machines on other forms of gambling (thus displacement effect is greater) the impact on the cost of the proposals for the betting industry is reduced. To address this, table 13 shows the impact of a 5% rate of attrition, half the central assumption to illustrate the order of magnitude of potential percentage change. Table 13: Attrition rate of 5% Bands above stake cap Take up Attrition 3 or more 40% 5% 2 51% 5% 1 68% 5% Central Estimate Option 2 Central Estimate Option 3 Central Estimate Option 4 £140 million £240 million £490 million Figures rounded to the nearest £10 million Extreme Scenario 33. Player behaviour is inherently difficult to predict and because of this we additionally consider a more drastic scenario combining different assumption changes. In the following scenario, we treat high stakes gambling (in this case, average stakes above £30) as particularly habitual and assume the plethora of other forms of gambling offer near-perfect alternatives to B2 machines. This scenario commands a reduced Take up of the lower stake as consumers continue to chase greater rewards at higher stakes, and, very low attrition as gambling alternatives to B2s offer equal utility. Therefore, in a set of assumptions for the extreme scenario central estimate, we assume take up of lower stake to be 30% lower than that of our current central assumptions, and, attrition of 2% across all bands. Table 14 :Extreme scenario assumptions Bands above cap Take up of lower stake Attrition 3 or more 2 1 10% 21% 38% 2% 2% 2% Level of Session Adjustment remains consistent with central assumption at 28% 34. These assumptions inform the proportion transferring their previous B2 spending onto other forms of gambling in equivalent amounts of GGY. 33 Table 15: Displacement to other forms of gambling in extreme scenario Bands above cap Total % moving to other forms of gambling 3 or more 2 88% 1 77% 60% 35. This combination of assumptions for the extreme scenario produces the following costs to industry. Table 16: Extreme Scenario outputs - Annual Cost to Industry Annual Cost to Industry Central Estimate Option 2 Central Estimate Option 3 Central Estimate Option 4 £50 million £80 million £160 million Figures rounded to the nearest £10 million, Level of session adjustment remains fixed in line with central assumption at 28%. Potential additions to options 36. We said in the consultation impact assessment that there may be some expected additional costs to industry if they are required to increase player protection measures on gaming machines (e.g. tracked play or ‘nudge’ type measures such as time and spend limits as referenced in the consultation document) and requested further data in the consultation document. The Gambling Commission will be taking this work forward with industry so this will not be covered here. Industry adjustment costs, on-going costs and supply chain impacts 37. See paras 14-16 for more detail. Assessment of impact on competition 38. The changes to industry GGY need to be assessed in terms of their impact upon competition. This can be considered through the lens of substitution i.e. players choosing to gamble on products other than B2 gaming machines. 39. The maximum stake for B2 gaming machines makes this category of machine relatively unique in comparison to other forms of gaming machines available on the high street or other accessible locations. Casinos offer roulette for high stakes, via both live tables and electronic touch bet terminals. The spin speed is significantly slower than for B2 machines. The cost of travelling to a casino and the difference in the environment may be weighed against the additional utility that could be derived from staking above the cap. 40. There are also potential gains for online providers of casino games if high staking B2 roulette players wish to continue staking at similar levels. Players who gamble 34 online as well as on B2 machines at present, or those who move their gambling online, may provide some benefit to online competitors. However, as with casinos, the environment is different and some machine players may not wish or may not be able to switch to account-based online play. 41. Of the possible substitutions which exist for B2 gaming in betting shops, it is important to note that the fundamental nature of B2 gaming may be altered by the proposed regulatory intervention. Whilst we have assumed that all customers previously staking below the cap will continue to play, exact behavioural changes are sensitive to the assumptions in the model. As a result, the proposed regulatory intervention has the potential to result in significant substitution to other gambling products inside or outside betting shops. With the scope for significant amounts of substitution, it is possible that this regulatory change could alter the competitive landscape. Impact on the Exchequer 42. Changes in industry GGY will have an impact on Exchequer revenues. However, estimates of Exchequer revenues are a matter for HM Treasury and HM Revenue and Customs and are not estimated as part of this Impact Assessment as per the rules within the Green Book. Impact on reducing regulation - One-in, Three-Out 43. The method of assessing whether new legislative proposals add to, or subtract from, the current stock of business regulation is known as “One-in-Three-out” (OITO). Proposals that impose direct costs are described as “IN”, while proposals that lead directly to benefits are described as “OUT”. Proposals that lead to indirect costs and benefits only are classified as “Zero net cost”. 44. The “One-In-Three-Out” (OITO) status of this policy proposal is defined as “IN” because the proposals are regulatory. Given that the proposed regulatory intervention is likely to directly affect player behaviour and therefore industry GGY, the impacts will be classified as direct, included within the OITO framework and scored accordingly as a net cost to businesses. Macroeconomic impacts 45. The framework for analysis only looks at gambling markets through the impact on machine revenues. It does take into account some isolated economic effects in other gambling markets however, it does not attempt to look at the effects on the rest of the economy (general equilibrium) because doing so would be disproportionate relative to the size of the policy impacts expected. Impacts on individuals and society 46. The Health Survey for England (HSE) 2015 and the Scottish Health Survey (SHeS) 2015 provide information about gambling behaviour in England and Scotland. Further analysis of this data, accompanied by data for Wales, undertaken by NatCen provides in-depth analysis of gambling and problem 35 gambling levels and examines the associations with problem and at-risk gambling. Latest data was published in August in 201738. 47. During consultation, we also received a report produced by the Centre for Economics and Business Research (CEBR) which provided estimates of the cost to government of problem gambling related to B2 gaming machines (see paragraph 24-28 for our consideration of the CEBR report)39. SaMBA 48. There is likely to be some small and micro businesses affected by this policy. The final option primarily affects betting shops (with a very small proportion of B2 gaming machines offered in some casinos – 191 machines out of a total of 33,611 - the remainder offered in betting shops). Therefore, this SaMBA analyses costs and benefits to provide a figure for the final option. 49. There are approximately 8,677 betting premises in the UK owned by 253 separate LBOs, of which 7,509 are operated by the largest licensed betting office chains (William Hill, Ladbrokes Coral and BetFred) . The share of operator costs is assumed to be proportionate for all bookmakers and evenly distributed, thus, the cost to SaM businesses is expected to be proportionate. 50. Information on LBOs was provided by the Gambling commission, including a breakdown of the number of premises per LBO and the number of employees per LBO for 2017. The caveat the Gambling Commission provided for their return was that any active operators who returned a null amount for their premises numbers is likely to have one shop and therefore have 6 employees. Without more detailed information, and, to ensure a precautionary approach is being taken, all small and micro LBOs are assumed to run B2 gaming machines. 51. The UK definition for a small business is a business with less than 50 employees. The definition of micro business is a business with less than 10 employees. From the data return, it’s established that there are 8,677 active premises. Of this, a total number of 113 premises can be considered to be owned by small businesses (licensed to companies employing between 11 and 50 people) and 196 premises can be considered to be owned by micro businesses (licensed to companies employing 10 or less people). Therefore the proportion of premises considered to be run by small businesses amounts to approximately 1.3% (113 divided by 8,677). The proportion owned by micro business is approximately 2.3% (196 divided by 8,677). When considering the cost of the policy options, it is assumed it’s evenly distributed across the betting industry, and, without more implicit evidence, it’s reasonable to assume small businesses may be burdened with 1.3% of the total cost and micro businesses 2.3% of the total cost. The impact to small and micro businesses imposed by the central estimate of each option is therefore estimated below in table 20. 38 39 ​http://natcen.ac.uk/media/1464625/gambling-behaviour-in-great-britain-2015.pd​f ​http://www.coin-opcommunity.co.uk/wp-content/uploads/2018/01/CEBR-Report-final-Jan-2018.pdf 36 Table 20: Impact of options on SaMBA businesses: Costs Option 1 Option 2 Option 3 Option 4 (Final Proposal) Small businesses £0.0 million £2.2 million £3.7 million £7.0 million Micro businesses £0.0 million £3.9 million £6.4 million £12.2 million Total impact for small and micro businesses £0.0 million £6.2 million £10.1 million £19.3 million Figures rounded to the nearest £0.1 million, totals may not sum due to rounding 52. We received 5 responses from independent betting shops during the consultation. One stated it could close its only store as a result of the £2 cap on B2 machines, whilst another said it could have to close 4 out of 6 of its stores. The other three respondents did predict industry impacts, but did not specify the potential impact to their own operations. This sample is too small to draw any meaningful assumptions to better estimate the business impact for the independent shops operating in the UK. 53. We acknowledge there could be a possibility that small businesses may be disproportionately affected by this policy, compared to larger businesses who may be less reliant on gaming machine GGY or have a greater online presence to offset the impact, but we have not been able to access data to support this assertion. In addition, it does not necessarily follow that those operators with an online presence would utilise this part of their business to support unprofitable premises in the land-based sector. In addition, the gambling industry, including the betting sector, is dynamic and will innovate in response to these changes, including diversifying their offering to less-harmful forms of gambling and it is something we will monitor closely. 54. Following further engagement with independent bookmakers at the consultation stage, we have explored a number of options to mitigate any disproportionate impact on small and micro-businesses and will be taking forward the following: a. Information - we are working across government to understand how this can be applied and where additional information provided to small and micro businesses can help mitigate the impact of this policy upon them. b. Extended transition period - this poses challenges around implementation with legislation coming into force that would apply to all premises. In addition, it would raise similar risks as set out under para 55a - i.e. the high likelihood of substitution of problem gamblers. However, following the publication of the consultation response, we will engage with industry further on an appropriate implementation period, which is initially expected 37 to be 9-12 months, based on consultation responses we received from gaming machine suppliers. 55. We have also considered the following mitigation options which we will not be taking forward. a. Full exemption - This is not considered appropriate. Allowing some operators (across betting and casino sectors) to continue to offer a maximum stake on B2 gaming machines that government believes may be harmful would be detrimental to the aim of the policy of reducing harm, from the point of implementation. The policy is aimed at those most vulnerable to harm, including problem gamblers. Approximately 4% of LBO premises (309 out of 8,677), on the high street would still be able to offer this product if a full exemption was considered, which isn’t insignificant. There is a high likelihood that there is going to be significant substitution because there is no difference between B2 gaming machines offered by large operators than those in small or micro businesses. In addition, those who are most vulnerable to harm are also more likely to displace to premises that still offer the higher maximum stake demonstrated by price inelasticity of gaming machine players40 and because problem gamblers disproportionately gamble at higher stakes and are more frequent users of the maximum stake.41 As a result, exempting small and micro businesses may perpetuate the harms associated with problem gambling for the individual as well as for the communities where these operators are based. Furthermore a full exemption may also lead to unintended consequences such as creating barriers to growth for small and micro businesses as they may limit the number of people they employ to remain in scope for the exemption. In addition, maximum stake and prizes on categories of gaming machines are set out in the Gambling Act 2005 alongside the allocations of gaming machines by premises type. An exemption for small and micro businesses would create a two tier system of bookmaker premises, creating an inconsistent regulatory regime, which would pose challenges for the Gambling Commission to regulate and create confusion for consumers. b. Partial exemption - This is not a relevant option for this policy, as any form of exemption would have the same consequences as those set out above c. Temporary exemption - as above d. Different requirements by firms size - as above e. Financial aid - not applicable f. Opt-in and voluntary solutions - We’ve previously considered industry voluntary options and evidence shows that these have not had the impact we’d hoped in protecting consumers. Therefore, we don’t believe further consideration of voluntary solutions is viable. 40 https://www.gov.uk/government/uploads/system/uploads/attachment_data/file/322845/report313.pdf 41 ​ ttp://about.gambleaware.org/media/1259/natcen-secondary-analysis-of-loyalty-card-survey-final.pdf h 38 56. In conclusion we will seek to mitigate the impact of the policy on small and micro businesses by ensuring information is shared and an appropriate transition period is fully considered. On balance, other possible mitigation options have not been considered suitable when weighed against the policy objectives and the intention to protect the most vulnerable in society from gambling-related harm posed by B2 gaming machines. Conclusions 57. The impact assessment has considered the impact to industry GGY, adjustment and on-going costs and societal impacts. To achieve the policy objectives, option 4 is put forward as the preferred option to strike a proportionate balance between socially responsible growth for the industry whilst protecting consumers. 58. We are working closely with the Gambling Commission to develop a monitoring and evaluation strategy to accompany this measure. This will include in its scope the impact on: business, consumers (specifically in relation to harm) and wider society. In doing so, we will be looking at the following evidence and data sources; gaming machine data in bookmakers specifically in relation to session losses on B2 machines, staking patterns and problem gambling rates overall and by product. This will also require further engagement with industry to secure access to data so that we can fully understand the impact of this policy change. More detail on the scope of the evaluation will be set out following the publication of the consultation response. Table 17: Impact of options in Net Present Value terms, over a 10 year appraisal period NPV (£m) IA Metrics 2017 Prices Central Low High Option 1 0 0 0 Option 2 1490 1190 2150 Option 3 2450 1700 4180 Option 4 4660 3100 7750 Figures rounded to the nearest £10 million 39 Table 18: Business impact of options EANDCB (£m) Central Estimate (2014 prices; 2015 present value) Option 1 0 Option 2 140 Option 3 240 Option 4 450 Figures rounded to the nearest £10 million Table 21: Conclusion of Option Appraisal Option Option Description Conclusion 1 Maintain Status Quo The government maintains that option one will not meet the primary objective of the review - to strike the right balance between socially responsible growth and the protection of consumers and wider communities. 2 B2 Gaming Machine Stake Cap from £100 to £30 In their advice to the department the Gambling Commission stated that the maximum stake should be reduced to between £2 and £30 if it is to have a significant effect on the potential for players to lose large amounts of money in a short space of time. In addition to this advice, we have concerns that, based on the data we have access to, there appear to remain a high proportion of problem gamblers at this staking level (compared to lower staking levels) and that gaming machine data suggests players are still experiencing high session losses, our best proxy for harm, with average stakes up to £30. On this basis, we are discounting this option. 3 B2 Gaming Machine Stake Cap from £100 to £20, £2 on slots In addition to the Gambling Commission advice as above, we also have concerns with this option that, based on the data we have access to, there appear to remain a high proportion of problem gamblers at this staking level (compared to lower staking levels) and that gaming machine data suggests players are still experiencing high 40 session losses, our best proxy for harm, with average stakes up to £20. In regards to B2 slots, we note Gambling Commission advice which states that they "think there is a case for a £2 maximum stake for B2 slots" reflecting risks associated with this content including lower return to players and less opportunities for players to manage their own risks through the way they play. These characteristics are reflected in the outcomes that we see in gaming machine data for slots play in which it is apparent that they give rise to a greater proportion of significant losses to the player than other B2 content. 4 B2 Gaming Machine Stake Cap from £100 to £2 This is the preferred option because in line with the policy objectives, this option will target those most vulnerable and reduce high level session losses and therefore the potential for harm. In addition, in their advice to Government, the Gambling Commission are clear that they support a precautionary reduction in stake on B2 gaming machines, but also make clear that setting a revised maximum is a matter of judgement for Government. Annex A - Central Outputs from B2 Gaming Machines Model Maximum stake cap on B2 Gaming machines Annual Cost to industry Annual Cost to industry (with £2 cap on Slots) £50 £30m £90m £30 £170m £210m £20 £260m £280m £2 £540m £540m Rounded to the nearest £10 million 41