Poker machine apocalypse?

There seems little doubt the fate of poker machines post 2023 will be an issue that will get plenty of coverage during the upcoming State election campaign.

Depite the government assuring the Joint Select Committee into Future Gaming Markets it would be open to recommendations based on evidence, Premier Stansfield and Chief Strategist McQuestin had their minds made up well before the Committee reported.  The government arranged for a Dorothy Dixer on the 11^th of September to allow the Treasurer to set out the government’s policy on gaming. If pokies are banned from communities it would have a “devastating effect on pubs and clubs” he told parliament.

Rather than blithely following the dictates of his political overlords the Treasurer should have read a paper prepared by his own department for the Committee.

A close analysis of Treasury’s modelling reveals in the case of regional areas beyond the 50 km reach of casinos where 40% ($42.9 million) of player losses occur, player losses will reduce by 75% ($31.4 million). There will be some migration to casinos and substitution with Keno, in total about 25%. Pubs and clubs will be worse off by $6.4 million, but the $25 million that previously flowed out of towns to Federal Hotels’ Network Gaming will be free to circulate within regional towns.

Currently on average, only 20% of player losses remain with pubs and clubs via net commissions from Network Gaming. The rest flows out of town to Network Gaming. It is well beyond the wit of this writer to figure why the government is so willing to accept the dumb proposition that stopping the haemorrhaging of regional towns will have a devastating effect on those communities. It’s going to be very difficult to convince an electorate growing weary of political lies. 

Regional pubs and clubs will be worse off by only $6.4 million. This is before any changes to their business models to attract some of the extra $25 million circulating in the community rather than being hijacked by Network Gaming.

In order to predict the effects on State taxation receipts, Treasury used different assumptions about player responses within 50 kms of the two casinos and in areas further away. It is a simple exercise to reconfigure the models to calculate  changes in player losses, changes in venue income and the extra income that will accrue to communities, in each of those two  broad areas, before having a revealing look at a few of the larger regional towns/cities.

This paper follows on from the writer’s last blog on the Treasury model and a subsequent request by Anglicare as to what conclusions can be drawn from the Treasury study about the effects in regional areas, and how the Treasury study reconciles with the Anglicare commissioned study by Prof John Mangan titled Removing poker machines from hotels and clubs in Tasmania: Economic considerations, presented as evidence to the Joint Select Committee. The remainder of this blog is the response provided to Anglicare’s request.

The Treasury paper titled: “What is the estimated economic impact to State tax receipts if electronic gaming machines had a $1 bet limit in the casinos and were removed from hotels and clubs?”  looked at changes in player expenditure and taxation receipts using three scenarios, each containing two stages.

The first stage was the removal of EGMs from pubs and clubs. Stage two was the imposition of a $1 bet limit on casino EGMs. The most significant changes occur in Stage 1.

The closing down of EGMs in pubs and clubs will result in a ‘migration’ effect as some players travel to casinos to continue their addiction. Three scenarios termed low, medium and high impact were modelled with different levels of migration.  The lowest level of migration was termed the high impact model where overall EGM player expenditures were most affected.

Furthermore, of the player spending that didn’t migrate to casinos, each of the three scenarios assumed different levels of substituting previous EGM spending with additional Keno spending. The high impact model had the least substitution. In other words players ended up spending less on gaming.

Overall the high impact model showed the largest fall in player expenditure/losses.

Let’s first have a look at the low impact model where there is a high level of migration to casinos and substitution with Keno. The table has been reconfigured to make it more reader friendly with ‘before’ and ‘after’ calculations and the inclusion of the Community Support Levy (CSL) because it is a de facto tax.

LOW IMPACT MODEL
$ million	Tax rate	Before	After	Change
Player losses
EGMs total		190.1	137.3	-52.8
Keno total		36.5	84.1	47.6
Total player losses		226.6	221.4	-5.2
Taxes
Tax EGM	25.88%	49.2	35.5	-13.7
Tax Keno	5.88%	2.1	4.9	2.8
Comm Support Levy	4.00%	4.6	0.0	-4.6
Total tax		55.9	40.5	-15.4

The ‘before’ situation is self explanatory. It relates to player expenditure for 2015/16. EGM losses were $190.1 million ($114.2 million in pubs and clubs and $75.9 million in casinos), whilst Keno losses, almost all in pubs and clubs were $36.5 million. Existing taxes were $55.9 million including CSL.

The ‘after’ situation shows changes in player losses are minimal, a reduction of only $5.2 million.

Taxes on the other hand fall by $15.4 million.  For two reasons, the substitute activity Keno is lowly taxed and CSL does not apply to casinos.

The high impact model shows less migration to casinos, and less substitution of Keno, so that overall player losses are reduced. Gaming taxes fall by $21.7 million with this model.

HIGH IMPACT MODEL
$ million	Tax rate	Before	After	Change
Player losses
EGMs total		190.1	119.8	-70.3
Keno total		36.5	54.1	17.6
Total player losses		226.6	173.9	-52.7
Taxes
Tax EGM	25.88%	49.2	31.0	-18.2
Tax Keno	5.88%	2.1	3.2	1.0
Comm Support Levy	4.00%	4.6	0.0	-4.6
Total tax		55.9	34.2	-21.7

If a $1 bet limit is imposed on EGMs once all have been confined to casinos there will be further reductions in player losses and government revenue.

$ million	Tax rate existing (1)	Existing situation (2)	High impact model		Changes +/-
			Casinos only         (3)	Casinos only & $1 limit           (4)	Casinos only       (5)	Introduce $1 limit    (6)	Total       (7)
Player losses
EGMs total		190.1	119.8	97.6	-70.3	-22.2	-92.5
Keno total		36.5	54.1	59.6	17.6	5.5	23.1
Total losses		226.6	173.9	157.2	-52.7	-16.7	-69.4
Taxes
Tax EGM	25.88%	49.2	31.0	25.3	-18.2	-5.7	-23.9
Tax Keno	5.88%	2.1	3.2	3.5	1.0	0.3	1.4
CSL	4.00%	4.6	0.0	0.0	-4.6	0.0	-4.6
Total tax		55.9	34.2	28.8	-21.7	-5.4	-27.2

EGM player losses are further reduced by $22.2 million to $97.6 million, a reduction of almost 50% overall. Tax revenue will take a further $5.4 million hit if tax rates remain the same.

To demonstrate how the fall in tax receipts is easily rectified let’s return to the high impact model without the $1 bet limit changes. Committee members all agreed the CSL should apply to EGMs in casinos. So let’s do that. Also let’s see what would happen if the Keno tax rate was increased by 15%, leaving it still below the existing EGM rate, and EGM taxes were increased by 10%. This is what model looks like:

HIGH IMPACT MODEL WITH INCREASED TAXES
$ million	Tax rate		Before	After	Change
	Old	New
Player losses
EGMs total			190.1	119.8	-70.3
Keno total			36.5	54.1	17.6
Total player losses			226.6	173.9	-52.7
Taxes
Tax EGM	25.88%	35.88%	49.2	43.0	-6.2
Tax Keno	5.88%	20.88%	2.1	11.3	9.1
Com Support Levy	4.00%	4.00%	4.6	4.8	0.2
Total tax			55.9	59.1	3.2

Needless to say the reduction in player losses is the same at $52.7 million but government revenue actually increases by $3.2 million. This demonstrates tax losses are easily fixed if required.

The migration and substitution assumptions in the high impact model mean the $114.2 million of EGM spending in pubs and clubs before a change will, after the change, lead to players saving $52.7 million (46%) with $61.5 million (54%) diverted back into gaming, either EGMs in casinos or Keno in pubs and clubs. These are the average statewide impacts. Impacts will vary, as discussed below, between regional areas and those areas with readier access to the two casinos.

Professor John Mangan’s detailed study to the Committee modelled three scenarios if EGMs were removed from the community. Scenario number 3 assumed 50% of player spending would migrate to casinos. This is similar to Treasury’s high impact model which shows about 54% migrating back to gaming.

Prof Mangan’s findings were that spending on activities other than gaming would be better for the economy.  Under scenario 3 he found there would be;

·       $21 million extra annually in net additions to gross state product

·       $11 million extra annually in wages, profits and dividends 

·       183 extra FTE jobs across the economy.

Prof Mangan commented on his scenario 3:

“The results, though still positive, are smaller. The lesson here is that the more spending is retained in some form of gambling, the lower the benefits to the economy from redirecting poker machine spending from hotels and clubs. “

The Federal Group in correspondence to the Committee in September 2017 was critical of Prof Mangan’s approach. It argued:

“ ..there is no evidence that a reduction in gaming machine expenditure by removing EGMs from pubs and clubs would result in spending distributed throughout the whole economy.”

The Federal Group provided no evidence for this silly assertion. Presumably the unspent amounts will remain under a bed somewhere?

The Federal Group also rejected a central tenet of Prof Mangan’s study, by claiming:

“.... the idea that gambling displaces other activity is not supported by any evidence...”

That is absolute nonsense. Of course gambling displaces other activities. Money spent on gambling would otherwise be spent elsewhere.

And finally the Federal Group stated there was no evidence FTEs involved in EGM activities in pubs and clubs was 200 as stated by Prof Mangan. This figure was obtained by this writer from a case study presented to the Committee by the Dixon Hotel Group, from a submission by the Federal Group to a 1993 parliamentary inquiry, both crosschecked with the writer’s own experience. Prof Mangan quoted this figure. This was all presented to the Committee in evidence and was included as part of Appendix D, page 198 of the Committee’s final report. No industry representative has explained why the figure may be incorrect. It is important to realise that the 200 figure relates to FTE EGM jobs in pubs and clubs. To the extent that Keno is substituted, fewer gaming jobs will be lost. Also if gaming migrates to casinos so will some FTEs be required to serve the extra customers.

The Federal Group has had many opportunities to specify the exact number of FTEs directly involved with EGMs but have always ducked the question. Worse still they deliberately misled the Committee on the matter. Rather than breaking up employee functions across gaming, accommodation, bars, bottles, food, management, support  etc, if a venue falls within the Australian Bureau of Statistics (ABS) statistical definition of gaming, then all employees at that venue are claimed by Federal Group to be gaming employees. It’s blatant deception. Federal Group in its letter to the Committee on 23^rd September 2017 when the final report was being finalised corrected a previous claim by saying it had 792.4 FTEs direct gaming jobs in Tasmania. Absolute nonsense. Not if gaming is defined by ordinary usage.

The Federal Group is deliberately overstating the number of gaming FTEs, presumably to bolster its position of importance in the economy so the effects of any changes can be exaggerated.

Policy makers are concerned about the impact on regional communities. To assist in this regard it is possible to rejig the high impact Treasury model to split the effects within 50 kms of casinos from regional areas further away. The high impact model assumes the migration to casinos is 60% for players living within 50km of a casino, plus 2.5% for those living further away. For the remainder of player spending that hasn’t migrated to casinos, 25% is assumed to be spent in pubs and clubs on Keno. Reworking the high impact model without a $1 limit gives the following:

EGMs PUBS & CLUBS: HIGH IMPACT MODEL BY REGION
$ million	Within 50 kms	More than 50 kms	Total
Existing EGM losses	71.3	42.9	114.2
Migration to casinos	42.8	1.1	43.9
Substitution with Keno	7.1	10.5	17.6
Reduction in player losses	21.4	31.4	52.8
Net commission before	15.1	8.5	23.6
Net commission after	1.4	2.1	3.5
Net loss to pubs and clubs	13.7	6.4	20.1

The reduction in player losses is split $21.4 million within 50kms of the two casinos and $31.4 million for the regional areas further away. These are reductions in player losses not reductions in venue income.

It is crucial to understand player losses and venue losses are not the same.

The losses to venues are set out in the last three lines of the table. Venues only receive a small portion (30%) of the losses via commissions from the Federal Group. It’s even less when payments back to Federal Group for EGM hire are made. The majority of player losses are lost to local communities.

Venues’ EGM commissions will fall but they will receive a small increase in Keno commissions, although it must be pointed out the current Keno commission is lower than the current EGM commission payable by Federal Group. The savings from EGM hire fees no longer payable will help mitigate the lost commissions. Overall net losses to venues will only be $20.1 million. And that includes pubs owned by Federal Group.

In regional areas outside 50kms from casinos the venue losses are only $6.4 million. The lost wages from those venues will be approximately $2.2 million.

To reiterate, the Treasury high impact model shows player losses in regional areas will reduce by $31.4 million which implies a reduction in gross income to EGM venues of $6.4 million. Regional venues will only be worse off by only $6.4 million. That may result in lost wages in regional EGM venues of $2.2 million. With FTE wages at say $40k pa that’s a loss of 55 jobs in regional gaming venues if they don’t change their business model.

Let’s not forget communities will be better off because more dollars will remain in regional areas. Despite what the Federal Group may believe, the $31.4 million will be spent elsewhere in the community and is likely to produce outcomes along the lines that Prof Mangan has suggested.

For a specific view of what may happen in a particular town with the high impact model assumptions, let’s look at Burnie:

EGMs PUBS : HIGH IMPACT MODEL BURNIE
$ million
	Before	After	Change
Player losses
Venues	1.7	0.4	1.3
Town outflows	5.6	1.6	4.0
Total losses	7.3	2.0	5.3

The table summarises the $7.3 million currently lost with EGMs in Burnie each year. Venues retain $1.7 million. The remainder flows out of town. Player losses in Burnie will reduce by $5.3 million annually. There will be some migration to casinos ($200k) and extra gambled on Keno ($1.8 million) but overall losses are $5.3 million less.  If players are better off by $5.3 million, then someone must be worse off by the same amount. In this instance the four Burnie pubs will lose $1.3 million in gross income after regaining some commission from the extra Keno spending. The amount currently leaving town courtesy of Federal Group’s Network Gaming siphon will fall by $4 million. The pubs’ losses will be less if they change their business models to attract some of the extra funds circulating in the community.

The pattern will be repeated throughout the North West, West Coast and other regional areas beyond the 50 km reach of a casino. A list of towns, with venue numbers and EGM number is included in the Appendix A. Appendix B summarises the effects of the high impact changes on larger regional towns and cities.

If EGMs are confined to casinos it is highly likely that EGM numbers in casinos will increase. Currently there are 1,185 EGMs in the two casinos attracting losses averaging $60k per EGM. Under the high impact model, EGM turnover in casinos will almost double meaning EGMs will each cater for $100k of losses assuming EGM numbers remain the same. With the more unlikely low impact model, EGM turnover will be even higher at $120k per EGM annually. This may be physically difficult to achieve with current casino EGM numbers and floor space.

The predictions of doom would be more credible if a few non-gaming businesses in the community had argued that existing EGM gaming arrangements are needed to preserve the economic health of communities? Alas none have come forward.

The prospect of an Armageddon is a ridiculous exaggeration. Nowhere was this more evident than in a radio interview with the Lord Mayor of Hobart on 16^th October 2017 parroting the agreed government message on the occasion of her preselection as a Liberal candidate for the forthcoming election. If pokies were removed from communities she predicted thousands of job losses, not just direct jobs but all the indirect ones, from the milkman delivering milk for cappuccinos, to the detergent salesman supplying products to clean the floors, the supplier of uniforms, advertising, electricity, all unemployed. It was fear mongering on an unprecedented scale.

A closer examination of the figures tells a completely different story. Of course there will be job losses if venues don’t change, just as there were job losses in non gaming businesses when gaming was introduced. But the changes are unlikely to be anything like the dire predictions from blinkered self interests. The opposite is more likely, higher gross state product, higher wages and more jobs.

APPENDIX A: EGMs OUTSIDE 50KM RADIUS FROM CASINOS
Town	No. Venues	No. EGMs
Bicheno	1	20
Bridport	1	15
Burnie	4	110
Campbell Town	1	20
Devonport	8	230
Dover	1	12
George Town	2	55
Latrobe	1	30
Orford	1	15
Penguin	1	25
Queenstown	2	25
Rosebery	1	10
Scottsdale	2	30
Shearwater	1	20
Sheffield	1	15
Smithton	2	50
Somerset	2	60
St Helens	2	45
Strahan	1	10
Ulverstone	4	110
Wynyard	2	50
Zeehan	1	20
TOTAL	42	977

Note: The Treasury paper did not specify exactly which EGMs fell outside a 50 km radius from casinos. The above table is estimated from the Tasmanian Liquor and Gaming Commission website.

APPENDIX B: LOW IMPACT MODEL -- LARGER REGIONAL TOWNS
	Before $ million			After $ million			Changes $ million
Town	Player Losses	Net Com	Out flow	Player Losses	Net Com	Out flow	Player savings	Pub losses	Extra cash
Burnie	7.3	1.7	5.6	2.0	0.4	1.6	5.3	1.3	4.0
Devonport	11.1	2.3	8.8	3.0	0.6	2.4	8.1	1.7	6.4
George Town	2.0	0.4	1.6	0.5	0.1	0.4	1.5	0.2	1.2
Latrobe	2.0	0.5	1.5	0.5	0.1	0.4	1.5	0.4	1.1
Smithton	2.0	0.4	1.6	0.5	0.1	0.4	1.5	0.3	1.2
Somerset	3.0	0.6	2.4	0.8	0.2	0.6	2.2	0.5	1.7
Ulverstone	6.1	1.3	4.8	1.6	0.3	1.3	4.5	1.0	3.5
Wynyard	2.7	0.6	2.1	0.7	0.1	0.6	2.0	0.4	1.5
Totals	36.2	7.7	28.5	9.7	1.9	7.8	26.5	5.8	20.7

This table show ‘before’ and ‘after’ effects for the larger regional towns. In the case of Burnie and Devonport the player losses are directly taken from Tasmanian Liquor and Gaming Commission statistics for 2015/16 based on LGA areas since there are no venues in the LGA outside town boundaries. Where towns form only part of a LGA player losses are based on an estimate.

The current situation is:

·       Player losses $36.2 million (NB Player losses in regional areas total $42.9 million as per the Treasury model).

·       Net commission after EGM hire fees to venues $7.7 million

·       Outflow to Network Gaming therefore $28.5 million

After the removal of EGMs from communities as per high impact model:

·       Player losses reduced to $9.7 million being migration to casinos and substitution with Keno.

·       Net commission of $1.9 being the extras commission re Keno received by venues.

·       Outflow is the balance of player losses which flow out of towns.

The changes are:

·       Overall player savings of $26.5 million

·       Reduction in venue net commission of $5.8 million plus reduction in outflow from towns of $20.7 million, which together equal the player savings.