Analysis of industry claims about the effects of higher gambling taxes

In July and August 2025, Gordon Brown, the Social Market Foundation (SMF) and the Institute for Public Policy Research (IPPR) all called for higher taxes on gambling. The SMF responded to the Treasury consultation on gambling duties with a proposal that would raise £2.0 billion per year. IPPR went further and proposed increases totalling £3.2 billion, enough to pay for the abolition of the two-child limit and benefit-cap in the social security system in 2026/27.

There has been an inevitable backlash from the gambling industry. The latest form of this is a mini report published by the Betting & Gaming Council (BGC), commissioned from Ernst & Young, purporting to show how negative the effects of higher gambling duties will be. This note considers the Ernst & Young arguments in detail as many don’t stand up to scrutiny.

Ernst & Young report

The intention of the Ernst & Young mini report is to show how many jobs will be lost and how much tax revenue foregone if gambling taxes are raised to show that raising these taxes is not worth doing.

The Betting and Gaming Council's press release of the report said that IPPR’s proposals would cost 40,000 jobs and cost the UK economy £3 billion a year. These estimates are seriously flawed for a number of reasons, each of which is discussed in more detail below.

1. Elasticities lack credibility: They do not use independent analysis to calculate elasticities but choose their own “following discussions with the BGC”.
2. Lack of microeconomic analysis of firm behaviour: They utilise their (overstated) elasticities without considering the microeconomics of firm responses to consumer elasticities.
3. Overstated job loss estimates: Their job loss estimates don’t take into account that online gaming employs fewer people than other forms of gambling.
4. No account of wider economic effects: They take no account of the redistribution of consumer expenditure to other forms of economic activity, which would generate additional jobs.
5. Fail to acknowledge high growth in online gaming revenues: Even where they have a point – that recent regulatory changes may reduce online gaming volumes to some degree – underlying growth in online gaming is very strong.

Elasticities lack credibility

In 2014, His Majesty’s Revenue and Customs published a report commissioned from Frontier Economics examining elasticities in the gambling industry. This report sought to quantify the extent to which consumers respond to changes in the ‘price’ of gambling (defined as the amount of stakes not paid out in winnings) with changes in how much they bet.

The Ernst & Young report uses the Frontier Economics elasticities for its ‘central elasticity’ scenario but all of the Betting and Gaming Council publicity is based on Ernst & Young’s ‘high elasticity’ scenario where the elasticities “were developed in discussion with the BGC”. This includes the claim that the IPPR’s proposals will cost 40,000 jobs. Discussion with an interested party is not an appropriate basis for credible economic modelling.

Lack of microeconomic analysis of firm behaviour

Ernst & Young assumes that any tax rise will be passed on in full to consumers: ie that any tax rise will result in an equivalent increase in the price of gambling, which they say is the standard assumption used in Treasury policy costings. 

What they ignore, however, is that gambling companies can observe the responsiveness of their customers to price changes that the degree of responsiveness varies according to the type of gambling, and this is likely to change how gambling companies respond.

Microeconomics of remote betting and machine gaming duties

According to the 2014 Frontier Economics report mentioned above, consumer elasticities (responsiveness to price changes) for Remote Betting (e.g. online sports betting) and Machine Gaming (for example, slot machines) are -0.5 and -0.6 respectively. 

If elasticities were at -1, there would be a pound-for-pound reduction in stakes for each pound of additional taxation. However, at -0.5 or -0.6, this means there’s a degree of stickiness to demand for these forms of gambling: despite the price increase, some demand will continue.

Frontier Economics found that win-rates for remote betting are around 94 per cent; ie that 6 per cent of stakes are not paid out in winnings by gambling companies and so are retained by gambling companies as their ‘gross gambling yield’. With a duty rate of 15 per cent, this means that government duty revenues are 0.9 per cent of stakes while 5.1 per cent of stakes are retained by the gambling company after paying out winnings and gambling duties.

Suppose that, in response to IPPR’s proposed duty increase from 15 per cent to 25 per cent, gambling companies increase their gross gambling yield from 6 per cent to 10 per cent by reducing payouts. (This is what Ernst & Young call 100 per cent pass through.) This means the win-rate falls from 94 per cent to 90 per cent and, as Ernst & Young say, the elasticities mean that stakes will fall. Our analysis is that an elasticity of -0.5 and a price increase from 6 per cent to 10 per cent means a reduction in stakes of 23 per cent.

So, stakes are now 77 per cent of their previous level, but the gambling company is taking 10 per cent of this lower stake: their gross gambling yield is now 7.7 per cent of the previous level of stakes. With gambling duty rates up to 25 per cent, the payment to the government is 1.925 per cent of previous stakes, while the gambling company retains 5.775 per cent of previous stakes: higher than the previous surplus of 5.1 per cent. Note also that, even though duties have only increased from 15 per cent to 25 per cent, duty revenues have more than doubled from 0.9 per cent to 1.925 per cent.

What’s going on is that, because of the stickiness of demand for remote betting, even if gambling companies increase the price, the reduction in stakes will be more than offset by the increase in the proportion retained as gross gambling yield. This creates a situation in which the volume of gambling reduces, company surpluses increase, and government revenue increases.

Conveniently, in their ‘high elasticity’ scenario, Ernst & Young wish away effects such as this by assuming all elasticities are at -1 or below, but without any evidence to do so.

Remote gaming

The situation is quite different for remote gaming (such as online slots and casino), where Frontier Economics estimated the elasticity to be -1.5, which means consumers are highly responsive to price changes.

Here, even if duty rates increase, it is simply not in the interest of gambling companies to increase their prices. If they do, consumers’ responsiveness means their gross gambling yield will fall even further. If this elasticity is correct, the way for firms to increase their gross gambling yield is to increase win-rates – ie reduce their prices. This should attract more stakes, increasing gross gambling yield, despite the increase in the proportion of stakes paid out as winnings.

To illustrate, Frontier Economics estimated win-rates for online gaming at 96 per cent of stakes – ie gross gambling yield is 4 per cent of stakes – while duty rates are around 20 per cent. (They are actually at 21 per cent but the maths in this illustration is easier if we assume 20 per cent for now.) In this case, government duty revenue is 0.8 per cent of total stakes and, after paying out winnings and gambling duty, the company surplus is 3.2 per cent of stakes. 

Now suppose duty rates rise from 20 per cent to 50 per cent (as in the IPPR/SMF proposal). At first, government revenue jumps to 2.0 per cent of total stakes and the company surplus (after winnings and duty) drops to 2.0 per cent. 

If the gambling company responds by raising prices to 10 per cent (ie Ernst & Young’s 100 per cent pass-through), win-rates reduce to 90 per cent. Our analysis shows that the -1.5 elasticity means this will reduce stakes by as much as 75 per cent. So, the gambling company’s gross gambling yield will be 10 per cent of 25 per cent of previous stakes: so 2.5 per cent of previous stakes. Gambling duties will amount to 1.25 per cent of previous stakes and the company’s surplus after paying winnings and duties is now 1.25 per cent of previous stakes.

What this shows is that the act of increasing the price in response to the duty change makes the company’s surplus after paying winnings and duty fall from 2.0 per cent of previous stakes to 1.25 per cent. Ernst & Young fail to explain why a company would voluntarily choose to do this.

In fact, the rational response for the gambling company is to raise its win-rate, increasing gambling volumes (and increasing government revenue). Our analysis shows it is unlikely the company will manage to restore its surplus to the level before the duty increase. Therefore, in contrast to the remote betting and machine gaming case, gambling companies will feel some ‘pain’ from the duty increase.

A constraint that might apply to some companies is that they cannot increase win rates because they are already at the maximum possible given their costs. In this situation, they will have to bear the full costs of the duty increase. However, average win-rates have reduced over the past five years (from 96.4 per cent to 95.9 per cent), so there is some scope for gambling companies to reverse this process.

However, it is worth noting that the Frontier Economics study was carried out more than 10 years ago, when online gaming was a tiny fraction of its current level. In the 5 years to 2023/24, average win rates in online gaming were reducing without the fall in volume of gambling that elasticities of –1.5 would imply. This suggests that there might be a difference between firm-specific elasticities – where qualitative evidence suggests that gamblers are highly sensitive to price in choosing whether to switch between companies – and industry-wide elasticities, where there is more stickiness in demand. 

If it is the case that industry-wide elasticities are actually between 0 and –1, then remote gaming is in the same situation as remote betting and machine gaming. In this case, it would be in the interest of gambling companies to increase the price (worsen odds) and this will lead to reduced gambling volumes, higher company surpluses and higher government revenue. 

The fact is we have only one study of gambling elasticities to rely on, published in 2014 before the large increase in online gaming over the last decade. However, if gambling companies do observe high elasticities (less than –1), they are not likely to reduce win-rates and government revenues won't fall. On the other hand, if they observe a degree of stickiness in demand, they can reduce win-rates and protect both their and the government’s revenue.

Overstated job losses

Previous reports by Ernst & Young for the Betting & Gaming Council have estimated total employment in the gambling industry as 109,000 people. However, these reports show a smaller proportion working on online gaming than the share of total revenues from that source. It’s not surprising that a form of gambling that is driven by online algorithms employs fewer people than in-person betting at the racecourse or on the high street.

Yet Ernst & Young assume that job losses will be in proportion to the reduction in total stakes. Combined with their deliberately high elasticities and ignoring of the microeconomic effects set out above, they estimate dramatic reductions in the volume of online gaming and then say the job losses will be in proportion to this reduction. Even if we accepted their high elasticity scenario, and we ignored the microeconomic issues, their own reports show that job losses from reductions in online gaming will be lower because that form of gambling employs fewer people.

No account of wider economic effects

In a note to one of the tables in the Ernst & Young report, they acknowledge that their report does not take account of redistribution of economic activity to other sectors of the economy. Yet basic economic theory suggests that if gambling becomes more expensive (ie win rates reduce) leading to a reduction in consumption, much of the money saved from lower expenditure on gambling will be spent by consumers on other products, creating jobs in those sectors.

The true cost to the whole economy of a reduction in gambling volumes is not the reduction in gambling but the extent to which this leads to a reduction in total economic activity: ie the amount that doesn’t get spent on other products and services. The failure to estimate the likely extent of this redistribution of economic activity means their estimates of the total effect on the UK economy are incomplete. 

Fail to acknowledge high growth in online gaming revenues

One point they make that potentially carries some weight is that legislation is already in place for tightening of rules around online gambling. The government’s impact assessment of these measures suggested that, across the industry, this could reduce gross gambling yield by between 8.0 and 13.9 per cent of online gross gambling yield. Taking into account this hit to revenue, their estimate – using the midpoint of the government’s impact assessment – is that the net increase in remote gaming gross gambling yield will be 4 per cent between 2023 and 2026. 

While this presents a picture of online gaming growing modestly, official Gambling Commission data shows that gross gambling yield for online gaming grew by an average of 7.0 per cent a year between 2018/19 and 2023/24. Less comprehensive data covering 70 per cent of the market shows that turnover – total stakes – for online gambling grew at an average of 12.7 per cent a year in the 15 months to April to June 2025 (the latest available data). 

Even with a hit to online gaming gross gambling yield of 11 per cent, if this long-term growth trend continues, we can expect growth rates to cover the cost of the increase in regulation in less than two years.

For full details of our microeconomic analysis, see our technical note which you can download from this page.