Advisory Board for Safer Gambling (ABSG) advice on proposals for a voluntary code on deposit thresholds

Introduction

11 April 2024 (revised 11 July 2024)

The Gambling Commission asked for ABSG’s advice on the introduction of an industry measure using net deposit thresholds. The advice requested was for the following:

1. The potential impact on consumers and/or opportunities it may create within the regulatory regime.
2. Insights from the academic literature in the UK and internationally.

Background

What has ABSG said to date on the management of financial risks?

Advisory Board for Safer Gambling (ABSG) has advised the Gambling Commission through its Progress Reports (2021, 2022) and advice on the Gambling Act Review (GAR) (2022) that mandated financial checks on net losses would be the most appropriate regulatory intervention to take forward as one of a number of interventions (opens in new tab) required to prevent harm. In our GAR advice we observed operators themselves concluded that operators would not act unless all are mandated to act and that more stringent mandated checks (opens in new tab) were therefore required to create leverage for improving customer interaction and reducing harm.

In its advice to the Commission Board (9 April 2021) ABSG made the following specific points:

1. Net losses are a better metric to use, as deposits can be used as savings accounts, seen elsewhere as in utilities allowing customers to accrue direct debit payments.
2. £500 per month net loss over several months would pick up the 1 percent who are classified as high risk gamblers (Forrest and McHale 2020, opens in new tab), but it would not pick up the wider group who have high likelihood of being harmed - the 25 percent identified in Muggleton and others 2021 paper (opens in new tab). This is why the communications around this needs to emphasise that there is no 'one size fits all' regulatory solution and that actions announced in the short term are a step towards further regulatory measures to improve protection and reducing harms.
3. Over the longer term, develop a formula for calculating affordability that includes not only expenditure, but also time spent gambling, type of product and age - 3 additional variables that together would give us a much more targeted and comprehensive measure of harm than a £500 net loss per month alone. That formula could be mandated by the Commission - all operators would have to apply it, and the Commission could test and develop it within the single customer view project.
4. In summary:
   1. focus first on curbing excessive (binge) gambling and agree a net loss trigger of £500 per month for customer interaction
   2. commission work on an algorithm for calculating affordability based on expenditure, time, product design and age
   3. signal that this is the direction of travel and that there will be further requirements on the industry.

What has the Commission done to date?

Over the past 3 years, the Commission has had ongoing engagement with its stakeholders on improving financial risk checks in order to improve protection for consumers. The analysis of the Commission’s November 2020 consultation found that 75 percent of respondents agreed that more should be done to protect vulnerable consumers.

Following the publication of the White Paper, the Commission introduced new proposals and published a consultation on financial vulnerability and financial risk checks. The Commission’s proposed financial risk thresholds for intervention are £125 net loss per rolling 30 days and £500 net loss over 365 days. The Commission consultation proposals on financial vulnerability and financial risk are summarised in Table 2 of the Appendix in this guidance.

ABSG comments on the proposed net deposit threshold set by operators

The comments from Advisory Board for Safer Gambling (ABSG) are as follows:

1. The scale and type of spending limit options include loss limits, deposit limits, maximum bet limits and time limits. We are unclear as to why net deposit thresholds are being suggested by the industry as an interim voluntary measure.
2. Deposit limit use by consumers is not universal. Industry sources estimate that approximately 400,000 accounts in the UK set voluntary deposit limits every month (opens in new tab). Forrest and McHale’s study found that a fifth of customers set limits (opens in new tab), and Auer and others¹ found only 8.3 percent of their sample of 175,818 players set voluntary limits (opens in new tab).
3. Weekly deposit limits are mandated in some jurisdictions and vary considerably (for example, limits are £214 across all operators in Germany, £685 in Austria and £397 in Norway).
4. A recent and widely referenced randomised controlled trial in Finland (opens in new tab) concluded that prompting online gamblers to set voluntary deposit limits did not affect subsequent net losses compared to gamblers who were unprompted. Customers most at risk of a limit chose either not to set it, or chose an ineffective limit size, or removed it altogether.
5. Behavioural Insights Team (BIT) research (opens in new tab) on the use of deposit limits as a ‘safer gambling’ tool also found that only a fifth of account holders used deposit limits, 1 in 3 set limits of £50,000 or more, and a free text approach rather than drop down menu with high anchors was the most effective method of limiting spend. Despite these findings, operators did not implement the recommendations on free text.
6. The international lower risk gambling guidelines (opens in new tab) developed by a cross jurisdictional group of experts using all the evidence available make no reference to net deposit thresholds, concentrating advice on gambling spend proportionate to household income, time spent gambling and type of gambling (Table 3 within the Appendix). Further evidence from these guidelines is shown in Table 4 in the Appendix on the relationship between increases in expenditure and harms. The White Paper also makes clear that consideration of a customer’s financial context rather than fixed thresholds alone should form the basis of effective customer interaction. The Financial Conduct Authority's (FCA’s) Financial Lives Survey (opens in new tab) suggests 24 percent of UK adults display low financial resilience, and 47 percent display one or more characteristic of vulnerability which may impact on their ability to manage their financial transactions.
7. Independent studies conclude that mandatory gambling management tools are the most effective harm reduction and minimisation strategies, as voluntary tools, whether consumer or operator initiated, are not consistently applied and do not impact on those most vulnerable to harm.
8. There is insufficient evidence from research upon which to base an initiative using industry threshold deposits, and a clear imperative for further Gambling Commission-led research along the lines outlined in our advice on 9 April 2021.

Specific questions on the proposal:

9. There is no reference as to how this measure would be evaluated (Timeframe? How many customers? What type of interaction and when? How are operators satisfied that there is no risk of harm?)
10. The voluntary code allows for a ‘suite of potential interaction types’ (paragraph 4), giving too high a level of discretion on how operators might act. What criteria would be used to determine operator actions?²
11. There is no detail in this document as to what ‘recognition of the code’ in the Commission’s assessment work means in practice.
12. It would be helpful to have clarity on what the evidence base for these thresholds was and how many operators were signalling a commitment to its delivery prior to implementation.
13. Is there a risk that operator resources required to implement this voluntary code could impact on resources required for mandatory changes following the outcome of the Commission consultation?

Table 1: Summary of key international research on thresholds for lower risk gambling

Summary of key international research on thresholds for lower risk gambling

Source	Gamble no more than this percentage of annual income	Losses	Frequency of play	Spend per month	Type of gambling	Deposit limits
Canadian Lower risk gambling guidelines 2021 (opens in new tab)	1%	Not included	No more than 4 days per month	Not included	No more than 2 types	Not included
Dowling and others 2021 (opens in new tab)	0.83% to 1.68%	Not included	2.5 to 3 times per month	£15 to £25	No more than 2 types	Not included
Louderback and others 2021 (opens in new tab)	6.7%	£22 per month	Not included	£143	Not included	Not included
Zendle and Newall 2024 (opens in new tab)	2.3%	£955 per year	Not included	£80	Not included	Not included
Currie and others 2017 (opens in new tab)	1.7%	Not included	8 times per month	£44	Not included	Not included
Hodgkins and others 2021 (opens in new tab)	1% to 3%	Not included	5 to 8 times per month	£70	3 to 4 types	Not included

International studies suggest safe spending thresholds of 1 to 6.7 percent of gross annual income (mean equals 2.2 percent, median equals 1.6 percent). The average income of UK adults is £32,500. The proposed £5,000 net deposit threshold trigger for intervention per account is considerably higher than any evidence would support, and far exceeds any financial risk threshold set out by the Commission or in the White Paper. Furthermore, the Commission’s own evidence suggests that online gamblers hold 3 accounts and use 1.5 on a monthly basis. A fifth of 18 to 34 year old online gamblers hold 5 or more accounts and could, under these proposals therefore hold £25,000 or more on deposit before any operator intervention is activated.

Further evidence is provided by Zendle and Newall's (2024) data fusion approach (opens in new tab). This work recruited participants who completed the Problem Gambling Severity Index (PGSI) and then provided authorisation to share their electronic gambling expenditure via an open banking application. Therefore, like the Muggleton and others (2021) paper which looks at objective electronic gambling expenditure, this paper extends this by looking at expenditure levels at different levels of harm. It used Williams and Volberg (2014) PGSI categories (opens in new tab), where 0 equals no-risk, 1 to 4 equals at-risk and 5 or more equals high-risk. Additionally, this work used net-expenditure (deposits minus withdrawals), as its key measure of gambling activity, where Muggleton and others (2021) used only outgoing deposits.

The work found that no-risk gamblers lost an average of £196.95 a year, or £16.41 a month. At-risk gamblers lost an average of £955.21 a year, or £79.60 a month and high-risk gamblers lost an average of £2,506.91 a year, or £208.91 a month. Note that these are cross-operator figures, which for comparability to the proposed intervention values would require a fully-operational ‘single customer view’ mechanism (Newall and Swanton 2024) (opens in new tab), as recommended in ABSG’s advice to the Commission. The Commission’s recent work with Warwick University has highlighted the importance of looking at gambling expenditure across as well as within operator accounts in order to understand impact.

Another way of looking at this is to consider the 'at-risk' expenditure levels as being at the top end of affordable levels. The Zendle and Newall study also included data on estimated income, which enables some comparisons with existing international literature on lower-risk guidelines (Table 1). The at-risk sample in this study had a mean estimated income of £41,157.80, which meant losses of 2.3 percent of their income on gambling. By comparison, the high-risk group lost 5.9 percent of their income on gambling.

Finally, follow up calculations on the dataset used in this study looked specifically at what impact a £5,000 net deposit threshold with a single gambling operator on a rolling monthly basis would have (Figure 1). This calculation shows that only 2.6 percent of those at-risk of harm (with PGSI scores greater than 0) would be subject to an operator risk assessment. This represents a false negative rate of 97 percent, suggesting that the Betting and Gaming Council (BGC) threshold is set too high for these checks to have a meaningful impact in preventing gambling-related harm for the majority of consumers who are at risk of harm.

All of these figures suggest that the proposed intervention levels are too high. We suggest that a more meaningful way forward in the longer term is to work towards assessing direct measures of spend with each operator over each month to create a false negative curve which showed the relationship between spend and PGSI scores. This could be achieved through integrating Gambling Survey for Great Britain (GSGB) survey results with regular operator returns to the Commission, and would show where the threshold would have the most impact. Integration into GSGB would provide a better avenue for evidence-based regulation, as it would allow for representative sampling.

---

¹ First author is owner of neccton ltd which sells the product mentor.

² Jonsson and others (opens in a new tab) showed that a phone call was most effective in reducing losses and increasing the uptake of gambling management tools amongst at risk gamblers and yet Patterns of Play found less than 0.13 percent of customers received a phone call. Forrest and McHale concluded that the threshold for customer interaction across 139,000 accounts was too high.

Our recommendation

Based on the empirical evidence, we do not support this proposal, for the following reasons:

• it would allow too much discretion in how operators would act
• it would not achieve the level of interaction required to minimise harm amongst at risk consumers
• it would be activated at too high a threshold for both age groups
• it would not achieve what the single customer view proposals set out to address such as triggering interactions based on customer level of spend and loss and patterns of play across multiple accounts.

Figure 1: False positive and false negative rates at financial risk check thresholds up to £300

Figure 2: False positive and false negative rates at financial risk check thresholds up to £5,000

Figures 1 and 2 show the relationship between false positive rates and false negative rates at different deposit thresholds. The graphs are identical except for the limit on their x-axes (£300 versus £5,000). At £5,000 within-month threshold, the analysis revealed a false negative rate of 0.9743590 (that is around 97 percent of 'at risk' gamblers in the sample were flagged 0 times during the past 12 months).

Appendix

Table 2: Summary of Gambling Commission proposals on financial risk checks (July 2023)

Summary of Gambling Commission proposals on financial risk checks (July 2023)

Key financial risk for consumers	Category of assessment	Proposed thresholds for consultation – 25 years old and over	Proposed thresholds for consultation – under 25 years old	Proposed net loss definition	Notes – other key aspects of proposals
Significant financial vulnerability for example bankruptcy	Light touch check using public data, and some aggregated data	£125 net loss per rolling 30 days or £500 per rolling 365 days	£125 net loss per rolling 30 days or £500 per rolling 365 days	Loss of deposited monies with an operator, not counting restaked winnings or bonus funds	Check need not be repeated within 12 months Gambling and deposits may continue while check taking place General data protection considerations apply
Binge gambling	Enhanced assessment	£1000 net loss per rolling 24-hour period	£500 net loss per rolling 24-hour period	As with significant financial vulnerability In addition, positive net position in preceding 7 days may be taken into account	Check need not be repeated within 6 months Gambling may continue, further deposits halted Particular requirements for data protection considerations
Significant losses over time	Enhanced assessment	£2,000 net loss in rolling 90 days	£1,000 net loss in rolling 90 days	As with significant financial vulnerability In addition, positive net position in preceding 7 days may be taken into account	Check need not be repeated within 6 months Gambling may continue, further deposits halted Particular requirements for data protection considerations

Table 3: Canadian lower risk gambling guidelines 2021

View the original Lower-Risk Gambling Guidelines poster (opens in new tab).

Guideline 1: Gamble no more than 1 percent of household income

Do not bet more than 1 percent of your household income before tax per month. For example, someone with a household income of $70,000 before tax should gamble no more than $58 per month.

The following table shows how much you can gamble each month to follow this guideline.

How much you can gamble each month to follow guideline 1

Yearly household income	Maximum monthly amount
$10,000	$8
$20,000	$17
$40,000	$33
$70,000	$58
$80,000	$67
$100,000	$83
$120,000	$100
$130,000	$108

Guideline 2: Gamble no more than 4 days per month

The guidelines suggest you do not gamble more than 4 times a month. This works out to roughly once a week.

For the Lower-Risk Gambling Guidelines to help individuals lower their risk of experiencing harms from gambling, they must follow all 3 guidelines at the same time and not be selective about which guidelines to follow when gambling.

Guideline 3: Avoid regularly gambling at more than 2 types of games

If you gamble on a regular basis, do not play at more than 2 types of games.

Familiar types of gambling games include:

• lotteries
• sports betting
• table games (such as roulette, poker blackjack, craps)
• slot machines
• scratch tickets
• horse racing
• bingo
• Video Lottery Terminals (VLTs)
• online forms of gambling like poker or slot machines.

Special risk populations and contextual factors

If you have experienced any of the following risk factors, then these limits may not be suitable for you:

• problems from alcohol, cannabis or other drug use
• problems with anxiety or depression
• personal or family history of gambling problems or substance use disorders.

In these instances, you should consider gambling less than these guidelines recommend or not at all.

Gambling Type

The type of gambling games you play makes a difference.

Fast-paced games that involve frequent betting can more quickly lead to problems. With slot machines, electronic gaming machines, poker and many online forms of gambling, people can spend a lot of money in a short time.

Table 4: Change in risk from reference group occurring when gambling expenditure per month predicts financial, relationship, emotional and psychological, and health harms

Total sample size is 59,099.

Change in risk from reference group occurring when gambling expenditure per month predicts financial, relationship, emotional and psychological, and health harms

Type of harm and the risk associated	Gambling expenditure (percentage) per month
	Less than or equal to 0.10%	0.11 to 0.50%	0.51 to 1.0%	1.1 to 2.0%	2.1 to 3.0%	3.1 to 4.0%	4.1 to 5.0%	5.1% and higher
Sample size in category	17,634	15,926	7,708	6,250	2,988	1,700	1,082	5,811
Financial harm
Sample reporting harm	335	535	429	507	337	221	144	1,510
Percentage reporting harm	1.9%	3.4%	5.6%	8.1%	11.3%	13.0%	13.3%	26.0%
Relative risk	N/A	1.8	3.0	4.3	5.9	6.8	7.0	13.7
Relationship harm
Sample reporting harm	173	249	207	287	178	129	99	1,045
Percentage reporting harm	1.0%	1.6%	2.7%	4.6%	6.0%	7.6%	9.1%	18.0%
Relative risk	N/A	1.6	2.7	4.7	6.1	7.7	9.3	18.3
Emotional and/or psychological harm
Sample reporting harm	441	638	460	616	374	250	178	1,551
Percentage reporting harm	2.5%	4.0%	6.0%	9.9%	12.5%	14.7%	16.5%	26.7%
Relative risk	N/A	1.6	2.4	3.9	5.0	5.9	6.6	10.7
Health harm
Sample reporting harm	142	221	157	219	133	87	76	776
Percentage reporting harm	0.8%	1.4%	2.0%	3.5	4.5%	5.1%	7.0%	13.4%
Relative risk	N/A	1.6	2.5	4.4	5.5	6.4	8.7	16.6