Gaming inquiry update

The parliamentary committee investigating Future Gaming Markets has received written submissions and has held five days of hearings.  Only a few witnesses ventured into accounting and economic aspects of gaming and this happened on the 7^th and 8^th February. The questioning by the Committee was pretty low key. It seemed they were struggling with the issues, not surprising given the enormity of the task in an area unfamiliar to all of them. This note was written in response to the Committee’s offer to accept comments from me that may assist in their deliberations following my brief appearance on the 8^th February.

The submissions and appearances of interest  (submissions and transcripts can can be found here ) were from:

·       Australian Leisure and Hospitality Group, the largest pokie operator in Australia associated with Woolworths

·       Dixon Hotel Group, a local group with 35 hotels (not all pokie pubs)

·       Tasmanian Hospitality Association

·       Tourism Industry Council Tasmania

·       Federal Hotels

One  area where no progress was made was trying to understand the costs associated with pokies at the venue level and at the network level. Each face different costs some fixed and some variable. It is crucial to understand the differences if pokies are allowed to survive outside casinos,and a more equitable split is to be recommended between the network operator, the venue, players and government. The note concludes with an Econ 101 presentation setting out costs and revenue for network and venue operator(s) before and after possible changes.

Australian Leisure and Hospitality Group

ALH paid almost half a million in payroll taxes in 2015/16. This implies a taxable payroll of approx $8 million. With 216 employees this implies average wages of just under $40k. This is a useful metric to bear in mind.

ALH claimed to have paid $6.9 million in gaming tax in the latest year. Strictly speaking they didn’t pay anything because it was paid by Federal Hotels. They probably meant it was paid by Federal Hotels in respect of gaming in their five establishments.  The amount of Keno tax couldn’t possibly be more than $100k as only $2 million was collected state-wide from 160 venues including two casinos. That leaves, say, $6.8 million for EGM tax and CSL. That implies player losses of over $22 million or $150k per EGM. This is most unlikely. The average in Glenorchy is $79k per EGM.

ALH stated their five venues were in the middle of the road on a turnover basis. This is untrue. Three are in the top 12 and the remaining two well above average. Network Gaming produces weekly rankings of all pubs and clubs should the committee want the details.

ALH argued the financial benefits from gaming need to be “fairly and more broadly distributed” and referring to the owner operator model said “the Government should reap the benefit of the additional tax rather than have someone clip the ticket on the way through.” That begs the question what is the quantum of the benefits being referred to? How much is Federal Hotels clipping on the way through? And precisely what exactly are the estimated network costs that an owner operator would face? ALH are large operators. They would have some idea. Get answers to those questions and it will be a large piece in the jigsaw puzzle. For the committee to consider the owner operator model such answers are a prerequisite.

As a corollary it would not be unreasonable to ask ALH for a $ figure or at least a % breakup of receipts between accommodation bars bottleshop food gaming and other plus a $ figure for overall wages, totals across all five pubs. This info is readily available by all operators and it would allow the committee to see the contribution of each area to the overall entertainment mix.

ALH said “a key metric for our group is a return on investment”. What is that rate of return? What discount rate (ie implied rate of return) would ALH use to value a perpetual license were it to be offered in Tasmania?  

Dixon Hotel Group

It is important at the outset to recognise that Dixons run a different model than say ALH. With 35 pubs, they only operate three. The rest are leased. Any comments need to be seen in this context.

Taking over distressed assets, recapitalising and enhancing them and then selling the leasehold interest to a lessee operator and collecting a rental stream henceforth is the basic modus operandi.

 As an aside they probably don’t pay payroll tax as that liability lies with each of the lessees whose payrolls are likely to be below the $1.25 million threshold. ALH (see above) as owner/operators admitted to payroll tax remittances of $500k and Federal Hotels probably pay about $6 million annually.

Given the size of the operation it would not be unreasonable to get on the record the number of Dixon owned pubs with EGMs and the number of machines in each. For each of the other multiple owners who are also operators, the info is provided by Tasmanian Liquor and Gaming Commission.

There is a tendency for the industry to thrust the plight of small owner operators into the spotlight as a stalking horse for the larger multiple site owners.

There is much data in the Dixon submission that will be useful to the committee.  The EGM profitability case study in Annexure 2 reproduced below probably needs further explanation because there are no accompanying notes
So here goes:

·       The pub has EGM turnover of $200k per week leading to losses of $20k per week or $1.04 million per year.

·       The machine rental is $90k which suggests it’s probably a 20 EGM venue.

·       The implied EGM losses are therefore $52k per EGM per year which puts it about average across the state. It is representative of Dixon pubs. It needs to be remembered however that average losses for Glenorchy EGMs for instance are 60% higher. Top performers attract even larger losses.

·       Wages are listed at $75k. This is 7% of overall player losses (equivalent to 22% of commission received by pubs which are calculated at 30% of player losses). This is a significant piece in the jigsaw. Just for the record, gaming wages are 7% of player losses. It confirms the figures in my submission.

·       Power is $25k which together with wages of $75k gives variable costs at the venue of $100k or 10% of losses.

·       Rates etc of $15k plus licence fee of $5k together with machine rental of $90k gives total fixed costs of $110k or $5.5k per EGM

·       Rental/interest of $60k is mostly the return to Dixons as landlord. To reflect the return to the venue, to the landlord and tenant combined, to enable comparisons with owner operator pub owners, the $60k needs to be added to the net profit of $43k to give a venue net profit from 20 EGMs of $103k.

With fixed costs of $5.5k per EGM plus variable costs of 10% of player losses it’s possible to establish break even points and profitability of EGMs in pubs and clubs. This is set out in the attachment at the end of this submission and backs up the numbers provided in my submission.

What happens at the venue level is reasonably clear.  From an overall viewpoint, with wages at 7% of losses, state-wide losses of $114 million for 2015/16 imply a wage component for EGMs in pubs and clubs of $8 million. Applying the average wage from ALH of $40k this implies gaming FTEs in pub for EGM operations of 200. There may be many more licensed persons but FTE numbers of gaming employees at any time may only be 200.

The next question then becomes is the $103k venue return reasonable, bearing in mind it’s only an average venue?  Dixons made the point with the case study in their submission that the net profit to the venue operator of $43k was “exceptionally poor” and “pathetic in terms of return to venue when compared to other States.” (See page 4 of their submission). To repeat the point already made, this is the amount left to the lessee after the rent in respect of the gaming room is paid to the landlord. If the lessee isn’t receiving enough then perhaps the landlord is grabbing too much? Maybe it’s not only Federal Hotels getting too much of the pie, perhaps the landlord under the leasing model as adopted by the Dixons is also getting too much? Is the lessee being left in such a destitute position without adequate returns? Is government the only saviour? This situation throws up many questions which are relevant.

·       Dixons said (on page 39 of the 8^th Feb transcript) that “hotels leaseholds sell on a 40% return which is two-and-a-half times the net profit”. Dixons acquire hotels, fix them up and sell a leasehold interest to tenants. They recoup some of their initial outlay and lock into a future stream of rental returns.

·       Net profit for a tenant will be after payment of rent. How is this determined? My submission on page 21 talks about the rental value of gaming being 25% of net commission income, as compared to say bottleshops where rents are closer to 5% of bottleshop turnover.  Gaming is far more profitable. The higher the net revenue the higher the rental figure, which leads to a higher value of the leasehold and a higher rental stream for the landlord. That’s how the lease model works. Not unlike franchise arrangement that are much in the news (Dominos, 7- Eleven etc).

·       Leases usually run for a long term. What arrangements are in place where lease terms extend beyond 2023?  How will lease agreements deal with changes in tax rates or total removal of EGMs from pubs and clubs if they occur after 2023?

·       Are rents tied to movements in gaming revenue?

·       There’s a feedback loop at work here. The greater the gaming profits the higher the value of the leasehold and the higher the rents payable. Lessees may be getting a pathetic return but that could be due to rents payable to landlords?

·       In a leasehold situation tangible gaming assets owned by tenants probably only amount to a few bar stools. The EGMs are hired and the premises owned by the landlord who receives a rental return. The tenant owns a lease, only part of which relates to gaming. It is not immediately clear why a $43k return from 20 EGMs with average player losses is pathetic? If the tenant expects a 40% return wouldn’t he be happy with a $43k return if he had paid $100k for the lease. And that’s without returns from other areas of hotel operations such as bars, bottles and food. Maybe the leases had a higher sale value? Maybe if the returns from gaming comprise the bulk of returns to the venue the business model is overloaded with risk and inherently flawed? Maybe the tenant is being used as a stalking horse for the landlord? Maybe the landlord is trying to lock in rental rates of return underwritten by punters?

·       Buying a leasehold business carries a much higher risk, that’s why they sell on a 40% return. What is the return the Dixons would expect if perpetual licences for EGMs were created at the venue level assuming the landlord would acquire the license? 40%? A lesser figure? What is the lower bound? How would an upfront price for a licence be reflected in the value of a leasehold interest? And in the amount of rent charged? Even if selling perpetual licenses was a good idea, the community will lose control? Who will benefit most from higher leasehold values and higher rents? Will communities with smaller leasehold operations be better off? And if 40% was the discount rate used to value licences is it worth it from the government’s viewpoint? The higher the expected return (or discount rate) the lower the receipt to government via the up front tender price.

The leasehold model crystallises the issues with perpetual licenses. To put it simply, perpetual licenses are a way of capitalising the value of future player losses in the same way as leases do. Certainty for the license and lease holders necessarily means certainty of future losses for players.

As with ALH it would not be unreasonable to ask Dixons for a $ figure or at least a % breakup of receipts between accommodation bars bottleshop food gaming and other plus a $ figure for overall wages for its EGM pubs, not figures for each pub, but totals across all venues. And similar figures for non EGM pubs. This would allow the committee to see the contribution of each area to the overall entertainment mix and the difference between EGM and non EGM pubs.

The constant thread running through all industry submissions favouring a different network model are all based on the view that Federal Hotels are making too much as the middle man.  What is an estimate of network costs, say per EGM, that Dixons think is reasonable?

Tasmanian Hospitality Association THA

The general manager of Shoreline gave evidence that his pub, which “sits comfortably in the top 10”  employs  “17 you would say are directly involved in gaming...... which is about 13 full-time equivalents directly involved in the hotel on the gaming side”. We know Glenorchy venues with 30 EGMs, not all of which are Top 10 have average player losses of $2.4 million (my submission page 20). The Shoreline probably has more losses than the Glenorchy average, let’s say, $2.7 million of losses. This means its EGM commissions are $810k (30% of losses). Now if 13 FTEs are employed in gaming then applying the wage per FTE of $40k (from ALH above), gaming wages are $520k. Out of total gaming commissions of $810k? That is quite ridiculous. And that is before machine hire of about $140k .

Hence when it is stated that 13 FTEs are directly involved on the gaming side it doesn’t mean that gaming is all they did. They must have had other duties, probably bar duties. If we apply the Dixon wage percentage of 7% of player losses, gaming wages would amount to $189k which is equivalent to just under five FTEs in gaming. The impression given was that gaming employment was 2 1/2 times greater.

The industry misuse statistics. Sometimes the number of licensed persons is given as if it reflects the wage component of gaming. Other times the number of persons directly employed is given hoping the reader will jump to the conclusion that being directly employed in gaming means they don’t have other duties.

On the question of gaming employment the THA put forward in its submission, and it was noted at the hearing, that employment in EGM venues was twice that of non EGM venues. However, correlation doesn’t imply causation. EGM pubs employed more than non EGM pubs before EGMs were installed. They were selected for EGMs because they were better performing, had more customers and hence more staff. It is not in the least bit surprising that EGM venues maintain or even increase their relative position vis-a-vis non EGM venues given the advantages bestowed on them via the implicit subsidy inherent in the current EGM arrangements. Because they are bigger and employ more they should be given further encouragement? Is that the argument? That seems a perverse view of the role of public policy. The answer maybe is to remove excess profits from EGM operations by adjusting machine parameters and tax rates?

On the matter of employment the committee was concerned about the employment effects in regional areas if player losses are reduced. Firm answers were hard to find. The primary effects are easy to ascertain. Direct gaming employment is 7% of player losses. The Tasmanian Liquour and Gaming Commission publish losses per local government area in its annual reports. Where the number of venues in a LGA are two or fewer the losses are aggregated but it would a simple matter to obtain the disaggregated figures from TLGC .

Two hotel operators appeared as part of the THA delegation. It is not unreasonable to ask the % breakup of sales between accommodation bars bottleshop food gaming and other plus a $ figure for overall wages so the committee can understand the revenue mix in what are two quite different operations. In the case of the Shoreline the evidence was that $4.5 million was spent on capex in 2001. How much was spent on the gaming room? And what returns have been generated as a result? Operators are asking the government to assist them. At some point there must be an obligation to reveal their current returns. It would be mandatory if asking a bank for a loan or asking the Department of State Growth for a grant for instance? What’s the difference?

Again we have the contention that Federal Hotels are making too much as the middle man. The industry must therefore have a view as to what the network costs would be under a different network model. What is THA’s estimate for network costs, say per EGM?

A confusing potpourri of figures presented by all participants inevitably makes it difficult to see the wood from the trees. On the matter of EGMs in pubs we know player losses were $114 million in 2016.  That’s a big loss to punters but pubs and clubs other than those owned by Federal Hotels only earned about $26.3 million in commission. After machine hire and wages only $11.8 million remained. (These figures are from page 23 of my submission). Less a few other variable costs such as electricity and cleaning and the 86 pubs and clubs probably only made a total of $10 million of gross profit. How does this compare with total gross profits in the hospitality industry? It will be a small fraction. THA will or should know. The industry tales of widespread woe may be over hyped.  Anyone with losses greater than EGM averages can easily take a haircut. The possibility of say reducing the house percentage from 10% to 5% might be one way of easing the burden of player losses combined with stepped tax rates.

Tourism Industry Council Tasmania TICT

TICT’s submission stated “there is no current market research on the association between gaming activity and visitation to Tasmania.....there is no evidence to suggest the presence of these facilities in Tasmania has any influence on our appeal as a visitor destination.”

One quarter of EGM pubs have tourism accreditation and they are important contributors to regional economies. But that leaves three quarters of pubs with EGMs which are more focussed on the local trade? Should they benefit via the freeloader effect from a measure that may be designed to assist regional communities?

That begs the question, what are the determinants of sound public policy for gaming from an economic and financial viewpoint? For instance

·       Is it a revenue raiser, bearing in mind gambling taxes are regressive taxes and aren’t as crucial to State coffers as popularly believed?

·       Is it to ensure players aren’t exploited and operators only make normal rates of return by adjusting machine parameters licence conditions and tax rates?

·       Is it to structure tax rates to assist lower turnover venues such as clubs and regional pubs? Will the freeloader effect mean others will benefit more?

Federal Hotels

One of the main arguments that underpin the rationale for another favourable deal is the contribution Federal Hotels makes to the Tasmanian economy of $343 million in 2016. This is based on a Deloitte report which is included in the written submission. The report deliberately overstated the contribution of gaming ($274 million) by including all income from the two casinos including accommodations receipts from 450 rooms, bars, food, seminars and conferences. The breakup of receipts would have been readily available and would have provided a better view of the contribution of the various segments and would have provided the committee with a clearer view of Federal Hotels’ operations.

In any event conclusions drawn from what economists call input-output studies are of dubious worth.  For a publicly available recital of the shortcomings of this approach, see Gretton, P. (2013), On input-output tables: uses and abuses, Staff Research Note, Productivity Commission, Canberra. (http://www.pc.gov.au/research/supporting/input-output-tables/input-output-tables.pdf

The following is an extract:

“Abuse primarily relates to overstating the economic importance of specific sectoral or regional activities. It is likely that if all such analyses were to be aggregated, they would sum to much more than the total for the Australian economy. Claims that jobs ‘gained’ directly from the cause being promoted will lead to cascading gains in the wider economy often fail to give any consideration to the restrictive nature of the assumptions required for input-output multiplier exercises to be valid. In particular, these applications fail to consider the opportunity cost of both spending measures and alternate uses of resources, and may misinform policy-makers.”  

If not spent on gaming, money will be spent elsewhere with consequent different employment effects. There is no evidence that gaming produces the best outcome even ignoring social harms.

Another of the restrictive assumptions of input-output models is that interstate leakage doesn’t occur. Dividends for example may fall into this category?

What is interesting about the Macq 1 development is, of the $45 million project, Federal Hotels’ share is only $10 million for the costs of the fit out. The balance of $35 million is presumably funded by the owner of the building believed, according to media reports, to be Vos Constructions.  The same arrangement applies at the nearby Henry Jones Art Hotel?

·       Why has Federal Hotels chosen a policy of large dividend payments rather than retaining earnings, reducing debt and developing and owning properties instead of leasing?

·       Wouldn’t this have made Federal Hotels less reliant on securing a favourable gaming arrangement post 2023?  Maybe even allowing greater returns to players? Or to the government?

·       In broad terms where have the shareholders spent the $200 million of dividends received since 2003?

·       Do any of Federal Hotels’ associates have any interest, either freehold or leasehold in any other tourism assets or in entities that have such an interest?

·       Why did Federal Hotels abandon its regional tourism strategy by selling Freycinet, Cradle Mountain and Strahan properties? Who bought the assets? RACT bought the business assets but who bought the real estate? Do any associates of Federal Hotels have an interest?

·       Why shouldn’t the sale of the regional properties be seen as reneging on a condition of the 2003 license deal? (The 2003 parliamentary inquiry clearly confirmed Federal Hotel’s placed the regional strategy on the table as a quid pro quo during the 2003 license extension negotiations with the government).

·       Given everyday commercial risks it seems unlikely that Vos Constructions as a builder would also be the owner of Macq 01 (and Henry Jones), so who is the owner? The reality is landlords have a much larger interest in leased tourism assets than tenants.  Given that Federal Hotels is seeking government assistance it’s not unreasonable to ask who its partners are. And what about the proposed Port Arthur project? Are there others involved?  We have seen the same pattern with the regional tourism properties sold by Federal Hotels. RACT bought the businesses but not the real estate? Who did? Perhaps not a specific question for this inquiry but it would add to a better understanding of the evolving structure of the tourism industry.

Given Federal Hotels’ ability to attract strategic partners as well as having a strong balance sheet weakened only by a relentless pattern of dividends, it is not clear why gaming should be structured to allow it to earn excess profits.  Its competitors don’t receive assistance.

There’s a need to ascertain the level of profits in Network Gaming. Every industry participant has indicated a desire for a larger share of EGM losses. But what exactly are the costs of networking? The fixed costs?  Any variable costs? Whatever system is recommended for pubs and clubs, an understanding of network costs is essential. Even if EGMs are confined to casinos the level of network costs will affect the recommended tax rates.

If the committee is interested in the level of excess profits in EGM pubs they should ask Federal Hotels how much the net profits from gaming contributed to the purchase price for Mackey’s Royal Hotel, Latrobe and Furner’s Hotel, Ulverstone, bought for a combined price of $33 million in 2010, and the acquisition of the Newstead Hotel in 2015 for $8.5 million

Attachment: Revenue and costs per EGM at venue and network level

It’s important to understand the costs facing both venues and the network provider(s) if any change to the system, including changes to taxation, is contemplated.

Only EGMs in pubs and clubs are considered. What follows is a graphical presentation of the numbers in my submission.

At the venue level this is the situation:

Revenue and costs are shown on a per EGM basis.

·       Revenue (red line) has a 30% slope due to the fact that revenue at the pub level comprises 30% of player losses.

·       Costs (blue line) has a 10% slope due to variable costs of 10% (wages = 7% of losses plus electricity, cleaning etc). Fixed costs of $10k per EGM is also included. These include machine hire, rates etc plus a return on capital invested in facilities.(NB It does not include a rate of return on the cost of any licences acquired by a new owner from an existing operator after 1997)

·       The yellow shaded area shows a deficit situation where costs exceed revenue.

·       The green shaded area shows surpluses where revenue exceeds costs.

·       Because fixed costs include a return on capital employed in the gaming area the surplus amount represents excess profits.

Let’s look at the situation facing the network provider.

The situation shown here is before taxes and CSL.

·       Revenue ( red line) has a 61% slope indicating the share of player losses to the network provider (NB network share of 61% + pub share of 30% plus GST share 9% = 100%).

·       Fixed costs of $10k per EGM are shown. This is an estimate which includes operating costs and a return on capital. There doesn’t appear to be any variable costs. Network costs will be the same regardless of EGM player losses.

·       Deficits and surpluses are shown with the yellow and green shaded areas.

·       The surplus amount represents excess profits.

Now let’s introduce tax and CSL of 30%:

·       The cost line (blue) now has a slope of 30% equal to the rate of tax and CSL. Some of the surplus has been eliminated.

The combined situation for the venue and the network operator (a combination of Charts 1 & 3) is shown here:

·       The revenue (red) line has a slope of 91% (player losses less 9% GST).

·       The cost (blue) line has a slope of 40% (tax & CSL of 30% plus variable costs at the venue level of 10%)

·       Fixed costs are $20k per EGM.

·       Every additional $1 of player losses results in a surplus of 51 cents or 51 cents.

·       As the fixed costs include a rate of return on capital, break even is occurring below the average turnover per EGM of $48k. Above the break even excess profits are generated.

If we apply a system of stepped tax rates as proposed in my submission the combined situation will look like this:

·       The revenue line still has a 91% slope

·       The surplus amount is smaller.

·       The cost line is now a dog leg with greater slope and directional changes when losses per EGM  losses reach $5k and $40k.This has the effect of removing some of the surpluses.

·       The aim of public policy arguably should be to set the steps and the rates so that operators don’t earn usurious rates of return. Even so, the area of surplus (green) indicates some excess profits remain.

It must be stressed the stepped tax proposal is predicated on the basis of removing excess profits annually via taxes rather than via a lump sum up front tender process.

The final step, dividing the surplus between the network operator (assuming there is one) and the venue, is not included. My submission pencilled in a 50/50 split but this was without a firm idea of network costs.

The previous chart suggests small turnover venues (clubs for instance) will struggle to reach break even. To date the discussion has assumed similar costs across venues leading to one breakeven point common to all venues. But costs will vary between venues, not only operating costs but fixed costs including a return on capital. Small pubs and clubs probably have lower fixed costs, with lower returns on investment required with less extravagant gaming room fit-outs.

 If it were considered desirable public policy to assist smaller venues the tax threshold could be set at the break even point.  Tax would remove excess profits above the breakeven point thus allowing normal rates of return. The following chart sets out the scenario:

With knowledge of costs it is not difficult to structure tax steps to achieve any public policy outcome.

 If there are to be one or more network operators selected via a tender, it could be specified that, say, maximum network prices are set every five years by OTTER (Office of the Tasmanian Economic Regulator).  In other words the government sets the machine parameters and tax rates and OTTER sets the network fees. Normal profits would be the outcome for any operator who achieves break even.