Gaming in the new market environment

List of Tables3.2 Regression results, UK Saturday Lotto sales millions 653.3 UK Saturday Lotto sales regression 673.4 Spending on Lotto and other models of gambling 724.1 Estimated param

Gaming in the New Market Environment

Edited by Matti Viren

Chapters © their authors 2008

All rights reserved No reproduction, copy or transmission of this

publication may be made without written permission

No paragraph of this publication may be reproduced, copied or transmitted save with written permission or in accordance with the provisions of the Copyright, Designs and Patents Act 1988, or under the terms of any licence permitting limited copying issued by the Copyright Licensing Agency, 90 Tottenham Court Road, London W1T 4LP

Any person who does any unauthorized act in relation to this publicationmay be liable to criminal prosecution and civil claims for damages

The authors have asserted their rights to be identified as the authors

of this work in accordance with the Copyright, Designs and

Patents Act 1988

First published 2008 by

PALGRAVE MACMILLAN

Houndmills, Basingstoke, Hampshire RG21 6XS and

175 Fifth Avenue, New York, N.Y 10010

Companies and representatives throughout the world

PALGRAVE MACMILLAN is the global academic imprint of the Palgrave Macmillan division of St Martin’s Press, LLC and of Palgrave Macmillan Ltd Macmillan® is a registered trademark in the United States, United Kingdomand other countries Palgrave is a registered trademark in the European Unionand other countries

This book is printed on paper suitable for recycling and made from fully managed and sustained forest sources Logging, pulping and manufacturingprocesses are expected to conform to the environmental regulations of the country of origin

A catalogue record for this book is available from the British Library.Library of Congress Cataloging-in-Publication Data

Gaming in the new market environment / edited by Matti Viren

p cm

Includes bibliographical references and index

1 Gambling industry 2 Gambling 3 Lotteries I Viren, Matti.HV6710.G38 2008

List of Tables

3.2 Regression results, UK Saturday Lotto sales (millions) 653.3 UK Saturday Lotto sales regression 673.4 Spending on Lotto and other models of gambling 724.1 Estimated parameters of log sales equations 89

5.2 Gambling revenues (£ million) and gross tax rates, UK 1025.3 Excise rates – UK National Lottery, alcohol, tobacco 103

and petrol products

5.4 Tax revenues, with and without the lottery 1045.5 Share of total income spent on ‘vices’, by income decile 1105.6 Share of spending on ‘vices’, by income decile 1125.7 Correlation coefficients of expenditures 1125.8 Why do people play the National Lottery? 1215.9 What is ‘an excellent way to spend’ lottery money? 1226.1 A summary of the most common forms of offline 128

dependency studies

8.4 Gambling prices for various products in Denmark 199

vi

Lotto draw

4.1 Lotto’s peculiar economies of scale 824.2 Expected value in rollover and regular draws 82

4.5 Effects of varying n on sales and revenue 914.6 Effects of varying τ on sales and revenue 925.1 Tax Lorenz curves – lottery, gambling, tobacco and alcohol 1115.2 Good causes and lottery spending 1197.1 Optimal payout ratio with different demand elasticities 162

and marginal costs

7.2 Operating costs in Europe 1995–2005 1657.3 Personnel costs in Europe 1995–2005 1667.4 Operating costs in the US in 2005 1677.5 Factors affecting operating costs in the US 1687.6 Two alternative marginal costs curves 1747.7 Marginal costs for US companies 1992–2005 1757.8 Effect of prizes on sales in the US 1787.9 Operating costs/sales from European and US lotteries 1797.10 Operating costs for certain private gaming companies, 2005 1808.1 Demand for gambling by a normal and an addicted 187

gambler

8.2 Demand for gambling by an addicted gambler 1888.3 The effect of a tax increase on normal gamblers 1898.4 A stylised theory of the gambling problem 1948.5 Adding the possibility of excess production 1968.6 The distributions of unit costs for the typical market 202

firm and SOE

vii

The preparation of this book would not have been possible without thesupport of the Finnish Veikkaus The CEO, Risto Nieminen, and director,Jari Vähänen, deserve special thanks for supporting the original initiativeand encouraging work at latter stages Ms Paivi Ahtola has also helped alot in all practical matters, especially in organising the authors’ workshopalong with the EL congress in Budapest in May 2007

Marja Heikkinen has kindly checked the English of several chaptersand Päivi Nietosvaara has edited the final version of the manuscript.Several persons have provided useful comments and other help duringthe process of writing different chapters of the book Some of thesepeople are mentioned in the context of individual chapters but theauthors would like to express their gratitude to all of them Also all theEuropean lotteries that responded to a survey on the consequences ofthe changing market environment deserve warm thanks

Finally, the assistance of Alec Dubber of Palgrave and Shirley Tan ofEXPO Holdings Sdn Bhd was paramount for the rapid publication ofthe book

Helsinki, September 2007

Matti Viren

viii

Mark Griffiths, International Gaming Research Unit, Psychology,

Nottingham Trent University, Nottingham, UK

Martin Paldam, Department of Economics, University of Aarhus,

Although we do not exactly know the volume of cross-border sales,

we can use the estimates of a recent survey conducted across Europeanlotteries.2 According to this survey, cross-border sales constitute anaverage of 7.4 per cent of the gaming companies’ total sales It is not astrikingly big figure but when we look at the individual games, we arefaced with certain very high percentages For example, while cross-border sales in lotto are around 1 per cent, their share in sports bettingamounts to up to 39 per cent From the point of view of existing com-panies and governments, the situation is alarming already because ofmajor losses of revenue Governments are also concerned about taxrevenue: private bookmakers do not pay taxes for their cross-border sales,which in turn further deteriorates the governments’ fiscal position.State-owned Lotto companies have legal monopolies in almost allEuropean countries This applies to lotto all over Europe and, in manycountries, to all gambling activities (including slot machines) Alongwith the state companies, there are a number of private bookmakers,which obviously would like to expand their activities to markets nowgoverned by public monopolies For this purpose they keep challenging

the markets by selling their products via the internet and by questioning

the justification of exclusive rights in courts of law In fact, dozens ofcourt cases are currently pending Thus far, the exclusive rights havenot been found illegal Yet, the current system has been found to be atvariance with other legislation and principles of competition in anumber of countries (cf., e.g., Littler and Fijnaut, 2006) More impor-tantly, the European Commission has shown strong interest towardsthe opening of the gambling sector to competition Consequently, ithas made several inquiries on the functioning of the gambling markets

in Europe, the most notable of which was made by the Swiss Institute

of Comparative Law (the so-called Swiss Institute Report, 2006)

Thus, it is quite obvious that cross-border sales will continue to grow

in the future unless strict regulations, such as the restrictions imposed

in the US on the transfer of money via the banking system, are

enforced Moreover, there is growing pressure on the current tional set-up dealing with sport-betting, in particular

institu-This does not mean that the only thing that could happen is thesimple liberalisation of gambling markets It is possible that the taxsystem changes so that a heavy excise tax (along with the VAT) isimposed on all types of gambling, which would effectively take out theprofits from gambling At present, tax rates vary considerably from onecountry to another (among the EL members, taxes may equal the lot-teries’ net profits) This is not important in the current state monopolysystem because the distinction between tax revenue and companyprofits is somewhat artificial from the government point of view In amore competitive system, this would no more be true and familiarproblems of tax competition and tax harmonisation would become relevant Because lottery services can be produced in basically anycountry, there would be large incentives for tax competition, whichwould limit the governments’ chances of compensating the reducedprofits from state lottery companies by taxes Another complicated taxissue deals with the taxation of internet gaming which is right now agrey area in the European Union (Jeanneret-Druckman, 2004) Alongwith the lines of VAT taxation cross-border internet gambling ought

to be taxed in the customer country, yet taxes are not collected in practice

Whatever will happen, it is clear that we are moving towards morecompetitive markets and prices will fall (accordingly, prizes and the pay-out ratios will increase).Thus, while the average payout ratio of theEuropean companies is now at 53.3 per cent as compared to 48.5 percent ten years ago, all companies expect it to increase substantially inthe next ten years If all restrictions were abolished, the payout ratios

could reach 70 per cent according to the current European state lotteries.3

This would obviously increase the volume of sales Using the subjectiveestimate of the companies as a benchmark we could say that such achange in the payout ratios would increase the overall demand byalmost one-third

Quite clearly, such changes would show in all indicators: production,employment, advertising and, maybe most importantly, problem gam-bling What would be the nature and magnitude of these effects is asubject to the analyses and reviews that are presented in this book.Even though we are focusing strongly on the future, this book is not

a forecast Rather than trying to say what the situation is in 2017 wewish to put forward alternatives which are available to decision-makersand consequences of these alternatives, including the alternative that

no significant changes in the current institutional system would takeplace

The nature of gaming will change even without institutional change.This can be clearly seen in the increase of multi-jurisdictional lotteries.The Nordic countries (Denmark, Finland, Iceland, Norway, Sweden andalso Estonia) have the Viking lotto, several European lotteries (Austria,Belgium, France, Ireland, Luxembourg, Portugal, Spain, Switzerlandand the UK with a total population base of 208 million people) areinvolved in the EuroMillions, and in the US, as many as 29 states par-ticipate in Powerball The motivation for this trend is clear: consumersprefer very large jackpots and they are only achievable by puttingtogether the clientele of several small lotteries The idea that we mighthave a fully global lotto one day does not sound terribly unrealisticany more In fact, some suggestions have already been put forward.Along with globalisation, the structure and nature of games keep chan-ging To illustrate previous developments, we could take the example ofFinland In 1940, when the Finnish lottery Veikkaus was founded, itstarted with just one game (Football Pools) In 2006, there were already

12 lotto, keno and sports betting games and, in addition, 12 instant games.Over the years, at least the same number of games have come and goneand many of the remaining games have changed their form – frequentlyinto games playable over the internet (Veikkaus’ Annual Report, 2006).When considering these alternatives, the deep dividing questionconcerns the role of government regulations and the role of statemonopolies in the production of gambling services Thus, we could askwhether there is a case for curtailing the demand for lotto and gam-bling A partly related question concerns the role of state monopolies:should we preserve the monopoly or not? In the latter case of no state

monopoly, we have to consider whether we should have a licencesystem (one or several licences), or whether we should implement acompletely unregulated system.

These questions are difficult because they concern the role of ment and the scope to which governments can go in securing theirfiscal needs and protecting consumers without jeopardising free entre-preneurship and free competition

govern-To illustrate the difficulty of deciding what form the lottery marketsshould take in the future, let us look at the history of lotteries in theUnited States In the United States lotteries have a long history startingbefore the independence Initially, there were both public and privatelotteries with different licensing systems in different states In the early

19thcentury lotteries had their ‘golden age’, but since 1830 growingcriticism and popular opposition led to the curtailment of the lotteries’operations and sales After the Civil War, the situation culminated

in that the government (different states with the help of the FederalGovernment) abolished gambling altogether for almost a century (fordetails, see Clotfelter and Cook, 1991 and Ezell, 1960) Basically, the pro-hibition could be explained by the general view that gambling createsmore harm than benefit for consumers To some extent at least, thispresumption was caused by misdeeds of criminal-minded lotteries thatreceived wide publicity Finally, in 1964, New Hampshire legalised lotteryactivities again, and was gradually followed by most US States Cur-rently, only eight US states prohibit legalised lotteries.4In Europe, theattitudes have not been as strongly divided, although they have variedacross different countries (cf., e.g., Douglas, 1995) It is only in the case ofcasinos that there has been extensive resistance in almost all Europeancountries In the rest of the world, gambling has a somewhat shorter his-tory and it has only recently become a major industry that has benefitedfrom relatively liberal government policies

Right now, the general attitude towards gambling is fairly positive.Relatively few consumers and organisations would be ready to prohibitall sorts of gambling This, in turn, can be interpreted as an indicationthat the favourable welfare effects of gaming proceeds are generallywell recognised Even so, the measurement and role of the welfareeffects are still worth thorough analysis Through demand analysis, wecan derive a large amount of useful and objective information on theseeffects

It is interesting to notice that the acceptance of gambling manifestsitself in long cycles This can be seen especially in the US, where theattitudes have moved from one extreme to another It is hard to say

how much of this is due to religion or, more generally, a certain way ofthinking Sauer (2001) argues that fiscal factors are decisive in this respect.When fiscal needs have been high – due to wars or otherwise increasedgovernment expenditures there has been more willingness to allow gam-bling to alleviate the pressure of taxation In Sauer’s view, the recentgrowth of lotteries can be attributed to the growth of government expend-iture worldwide (e.g., due to the construction of the welfare state) Fiscalpressures have paved a way for new sources of revenue, and gambling is

an obvious possibility for new income source It is difficult to say exactlyhow important fiscal matters are but it is clear that we cannot isolate ‘atti-tudes’ and ‘ways of thinking’ from the development of the economy Cer-tainly, there exist differences between e.g., federalist and non-federalistcountries because state-level authorities have had less alternatives in gen-erating additional revenues This may show up in attitudes towards casinos,age-limits and so on Naturally, the demand for gambling has its owndynamics: when gambling grows extensively, the dark side becomes moreprominent and opposition emerges By contrast, in a heavily regulatedsystem, virtually no problems seem to exist and there is no demand for

an anti-gambling movement

In the 1950s and 1960s, state gaming monopolies seemed the obviousway of arranging gaming activities.5Governments could delegate theproduction of services to the monopoly company and curtail the poss-ible harmful effects by monopolistic pricing At that time, the harmfuleffects probably mattered much less than the fiscal effects As fiscal agen-cies, government monopolies were ideal There was no need to design

a tax system for various products Instead, it was left to the monopolycompany to find the optimal pricing and production scheme

Revenues from gaming monopoly (monopolies) are definitely tant for governments, although the figures do not seem very high.Thus, in the US, lottery revenue represented roughly 2.3 per cent of allown-source general revenue or lottery states (Hansen, 2004) According

impor-to Eurostat, taxes on gambling, lotteries and betting correspond impor-to about0.4 per cent of all taxes and social security contributions to generalgovernment We have to be careful with these numbers because theyare not comparable and because the quality of the European figures atleast is dubious.6Moreover, we have to keep in mind that a possiblesqueeze of lottery revenues would not hurt all levels of government

on the way Let us take one example In Finland, the share of taxes ongambling, lotteries and betting out of the total taxes and social securitycontributions to general government was 1.3 per cent in 2006 How-ever, if we compare these revenues to central government income tax

revenues, the share is as high as 7.6 per cent Thus, from the centralgovernment’s point of view, lottery revenues are important because ifthey have to be compensated with the main tax instruments (incometax or VAT), the average rates should be increased by more than onepercentage point Although all lottery revenue does not go to govern-ment (literally to the account of the treasury) the revenue is equivalent

to other government revenue from the government’s point of view inmost cases Thus, if state lottery proceeds were not available, the gov-ernments should probably provide similar levels of funding to thelottery beneficiaries (when the government is not the beneficiarydirectly) This would require higher taxes, basically entailing a trade-offbetween the ‘lottery tax’ and, say, the income tax Comparing thesetwo alternatives is not easy because it is not completely clear to whatextent ‘lottery tax’ is comparable with income tax (rather it is com-parable to some commodity taxes) The only thing that is clear is thatthe two ‘taxes’ have quite different distributional effects: the ‘lotterytax’ is known to be highly regressive, whereas the income tax is usuallyprogressive Thus, moving away from the ‘lottery tax’ would be goodfor distributional reasons at first sight Yet, we have to keep in mind that the ‘lottery tax’ does not necessarily disappear if we abolished state monopolies but rather it changes its form to (private) lotterycompanies’ profits

At any rate, this trade-off is a real problem because taxation is notonly an issue of distribution From the economics point of view, thedeadweight losses are probably more important In the case of incometaxation, we have long thought to be relatively unimportant (assumingthat labour supply is almost inelastic) However, the general views arequite different at present Thus, in public finances we have experienced

a ‘flight from high income taxes’ to alternative sources of finance.Unfortunately, or fortunately, in the modern world, we have very fewoptions for alternative sources of finance: tax competition sets strictlimits to capital taxation and commodity taxation On top of that, weshould keep in mind that at least in income taxation, both tradeunions and employer’s associations have great influence on decision-making There will never be equally powerful gamblers’ associations tolobby in London or Brussels

Needless to say, the fiscal problem varies from country to countrydepending on the earmarking system and budgeting practice In thecase where the lottery money goes to sports and arts, it looks like there

is no direct obligation on the government to compensate for the loss ofmoney We should remember, however, that sports organisations and

the arts lobby have always been powerful and skilful lobbyists for ernment expenditures and that they would probably receive as much

gov-as now from the government if there were no lottery funds available.Although the revenues from gambling constitute an important frac-tion of governments’ resources there is a big policy problem with thissource of funds: the importance of lottery revenue does not justifymonopolies as such The fact that a government monopoly is highlyprofitable in a certain market does not provide the government with

a justification to set up a monopoly and prevent competition The ernment must have a proper justification for preventing market out-come Basically, this is the position adopted by the European Court ofJustice and several national courts

gov-If the fiscal necessity does not constitute the case for governmentlottery, what could then be the proper justification? The primary expla-nation appears to be the need to curtail the harmful side effects ofgambling, by which we usually mean problem gambling The idea issimple: with monopolistic pricing, the volume of sales stays relativelysmall and if problem gambling is related to the level of sales, problemgambling should also remain well under control

The previous argument makes sense; yet it can be challenged by atleast two arguments First, problem gambling does not appear to be reallywidespread and second, the reduction of demand could be achieved bymeans other than state monopoly, i.e., by taxes and legal restrictions

In this respect, government monopoly could only be better if it behaved

in a way different from that of private operators, e.g., by showing greaterself-restraint in creating new games, restricting sales to high risk-groupsand using less (effective) advertisement to stimulate demand

These are issues which probably vary enormously from country tocountry depending on the role of fiscal authorities (Ministry of Finance)

in the administration of the companies, the system followed in ments (does the board consist of politicians, business people, academics,etc.), and what is the general business culture in public enterprises like Inthe US, state lotteries seem to clearly resemble each other, but in Europethere is no uniform model for them Take for instance, Finland and Italyand compare their market structure, licensing policy, and legislation!The role of problem gambling is quite delicate as it is suggested later

appoint-in this book It is generally argued that problem gamblappoint-ing has beenwell under control thus far, and that it only represents a problem for asmall minority, say 1 per cent of the adult population, which appears

a typical estimate in most jurisdictions But we have to be careful inquantifying the significance of problem gambling Small numbers of

participation may not necessarily mean small economic impact Theconsequences of problem gambling can be experienced by a much largerset of people than the problem gamblers themselves, and the damage

on the lives of the sufferers and their families can be much larger thanour numbers would suggest We could imagine that problem gamblingcould go to extreme forms, such as theft and fraud Then, its trueincidence would no more be marginal

From the point of view of the present analysis, the key issue is notreally the current state of problem gambling but the connectionbetween market change and the growth of problem gambling The factthat problem gambling has been under control does not necessarilymean that problems will not occur shall markets change Increase insales and, in particular, the development of more intensive gamescould change the problem gambling environment entirely.7We shouldalso notice that even if problem gamblers might not lose money, they(and their social environment) would still suffer from problem gam-bling With less binding budget constraints, they could devote theirentire lives to gambling leaving aside schooling, social contacts andwork, which sounds problematic even though the gambler’s utilitywould be maximised Thus, even if increased competition loweredprices, cutting the existing problem gamblers’ monetary losses, assum-ing that their demand is relatively inelastic, problem gambling wouldnot automatically become less problematic as a whole and may evenbecome worse

Besides the desire to limit problem gambling, we have basically two further rationalisations by which the gambling monopoly can

be justified: (1) the gambling markets may not stay competitive and (2) there are several public policy reasons for monopoly

If the gambling market does not become competitive, consumers willnot benefit from lower prices (increased consumer surplus) It is thisfact that is often used as to motivate government intervention in publiceconomics If consumers were not awarded greater surplus, publicmonopoly would not be any worse than the market solution

Our argument stems basically from the fact that costs are dominantly fixed in gambling, especially in lotteries Thus, marginalcosts are small and apparently falling.8 Such a cost structure clearlyfavours monopolies (creating an environment for what could be called

pre-a npre-aturpre-al monopoly) A npre-aturpre-al monopoly would hpre-ave pre-a permpre-anentcost advantage in comparison with smaller firms, driving smaller com-panies out of the market and stopping new companies from enteringthe market Thus, we would not end up with a competitive market

solution in the long run, but rather with a single monopoly or polistic competition depending on the role of costs when sales becomevery large.

mono-Some caveats should, however, be considered here First of all, weknow relatively little about the cost conditions of gaming companies.Moreover, all empirical analyses make use of data for relatively smallcompanies In fact, in the US, the biggest company (New York StateLottery) represents just 12 per cent of the total sales in US, whereas thebiggest European company (Lottomatica) represents 15 per cent of theEuropean sales Hence, we know little about the cost conditions of com-panies that would cover the total US and/or European markets Ano-ther caveat deals with the nature of games The cost conditions forsetting up new games or selling them may be quite different in thefuture, which obviously affects the shapes of relevant cost curves This

is obvious if we compare, e.g., the costs incurred by internet sales withtraditional retailer commissions

It might be argued that a natural monopoly would not be a problembecause the government could control it by various regulations, as hasbeen the case with various energy, water and transportation mono-polies customarily (for a comprehensive literature review on the regu-lation of natural monopolies, see e.g., Depoorter, 2000) Controlling anatural monopoly is somewhat controversial since it is the most effi-cient way of production Yet, the adverse distributional effects usuallymake governments set upper limits to prices Correct pricing is a verydelicate matter because the government should basically give subsidies

to the company to obtain the most efficient production structure Inthe case of full-scale market liberalisation a more serious institutionalproblem would, however, arise because all the companies would beinevitably global Consequently, we would need a global organisation tocontrol the supply side Right now, there is no obvious organisationalsolution for the problem although regional institutions like the EUCommission, might try to do that.9

Turning to public policy problems, there are several issues to be sidered The first issue concerns fraud and other misbehaviour Suchphenomena would most probably be relevant in the case of a seem-ingly competitive system with many small providers of gambling ser-vices Those having difficulties in surviving would be tempted to useillegal, or morally dubious means of surviving in the market They mightalso collaborate with the criminal-minded producers of false results.10

con-Small firms might be founded solely for the purpose of crime, moneylaundering, in particular Money laundering is a big issue in modern

societies, since it is lies behind all economic crime and misbehaviour.

On the other hand, gambling is ideal for money laundering because ofits big volumes and frequent transactions Even now, payout ratios aresufficiently high to exceed the critical levels required to support moneylaundering (see, e.g., Fabre G., 2003) The problem is generally recog-nised and various legal and administrative steps have been done toprevent it but very few cases have, after all, been discovered and brought

While the above-mentioned issues may provide justification for ernment gambling monopolies, there are also arguments that favourthe abolishment of those monopolies On the top of the list is probablythe need to increase competition, decrease government interventionand, in addition, the desire to achieve a more efficient production oflottery services

gov-A more competitive system would, ceteris paribus, lower the ‘lottery

tax’ and increase the demand for lotto It would enhance welfare even

if it increased problem gambling If markets did not stay competitivebut turned into private monopolies, we would not gain very muchfrom liberalising the markets Rather, we would be dealing with certainadverse distributional issues A relevant question would then be: whowill gain from the government’s monopoly lottery versus who willgain from the private lottery monopoly

Public vs private ownership would probably – irrespective of marketstructure – affect the efficiency of production There are two possiblereasons for this First, even if we had a monopoly for reasons otherthan legal considerations in private production, there would always be

a contestable market option Unlike the case of legal monopoly, aprivate monopolist would have to be prepared for the entry of a poss-ible new competitor Thus, it could not set its prices too high even ifthere were no apparent competition Secondly, a private monopoly has

to be responsive to its shareholders and therefore try to increase itsprofits by, e.g., reducing costs Basically, public monopolies have thesame requirement but they have been typically less keen on reducing

costs and increasing efficiency What is said about reducing costs cally applies to innovations and investments, as well If managers arenot under constant pressure to increase profits, less effort is devoted toimprovements.

basi-Naturally, it is generally assumed that the ‘public ownership slack’also shows in public lottery companies Unfortunately, there is noempirical analysis which would allow us to conclude whether the presumption is right or wrong Existing data facilities allow only veryrough comparisons Even though these comparisons are not very informative casual evidence from lottery companies suggests that the differences are not very large and at least some improvement in theefficiency at public lottery companies has taken place Thus, in spite

of a rapid growth of sales, employment has not increased but ratherdecreased There can be various reasons for this but it seems likelyobvious that increased interest in the public companies’ efficiencycould be one reason behind the change We also have to pay attention

to the fact that state monopolies, although they have a legal

mono-poly, are de facto competing with private gambling companies Thus,

gambling markets in Europe are more competitive than contestable in

an abstract sense But for efficiency comparisons, we would definitelyneed empirical analysis to answer the question of whether importantefficiency differences exist, both between state monopolies and betweenprivate and public companies

The problem is that even the efficiency issues have two sides: a more efficient company can sell more, which, from the problem gambling point of view, could be detrimental This reflects the currentschizophrenic situation in the European gambling markets Increasedcross-border trade has made markets competitive and state lottery mono-polies face difficult choices If they start competing (in the sense ofdefending their market share) by lowering prices, increasing adver-tising and introducing more games then private competitors and poss-ibly even the regulators would accuse the monopolies for behaving likecompetitive firms and ignoring the harmful side-effects of gambling

In such a case, their existence would be hard to justify If, by trast, state monopolies do not react in any way to cross-border com-petition, they lose their market shares, and the markets will start functioning like any other market Again it would be difficult to jus-tify the state monopoly if it has no role in market behaviour and no fiscal importance

con-How should the companies, or their regulators, behave in such an ronment? Right now, nobody seems to know Obviously, some solution

envi-has to be reached, and transparent and enforceable rules have to beimposed on the market.

It is also evident that, in the future, we will have to reconsider fiscalissues from various viewpoints It is probable that some tax harmon-isation will take place It would be highly timely for Europe where boththe rates and the institutions vary considerably As we have pointedout earlier, some European countries have a formal lottery tax whileothers rely on direct transfers of profits from state lottery companies.Some (mainly Eastern European) countries also tax winnings, whereasthey are tax free in most other countries The taxation of winningsfrom other countries (private companies & internet) is even more of anopen issue

In the future, one might expect that the European countries wouldmove to a system of formal gambling taxation The taxation would, ofcourse, be realised as source taxes so that winnings would be tax free.But what type of taxes should be imposed? This issue is handled inseveral articles in this volume (especially in Chapter 5) The problem isthat ‘optimal’ tax considerations seem to indicate that the taxes shouldnot be excessively high Yet, there are political economy considerationsthat might lead to relatively high rates Governments would probably beeager to impose very high tax rates on all gambling, not only to reduceproblem gambling but to discourage private gaming companies fromentering the market Gambling is an area where various special interestgroups and lobbies have a lot of influence We think it is unrealistic toexpect that taxation would be organised on a purely theoretical basis.Yet, bearing in mind the above issues, certain basic principles ofwelfare economics should be acknowledged

Here we will not answer the question of how the gambling marketsshould be arranged, even though we provide a large amount of materialwhich helps to take a stance on this matter Thus, individual chapters

of the book focus on the key issues of gambling and provide necessaryinformation to answer the questions: how much will gambling (lotteries)increase if competition increases?; what is the contribution of the presentgames and changes in the existing games?; what are the welfare and fiscaleffects of these changes?, what happens to the market structure if variousrestrictions on cross-border gaming are abolished?; how will problemgaming respond to changes in the sales volume and gaming menu?; and,what public policy concerns the change will bring?

We are well aware that all relevant issues are not covered Take, forinstance, the interplay between lottery companies, governments andvarious beneficiary organisations in the area of public policy, or the

problems of crime and fraud Despite the partial coverage, we believethat the key issues are relatively well represented here and they willfacilitate more intelligent discussion of the choices that could, or have

to be made in the future

What follows next is a short summary of the individual chapters

At the end, some concluding remarks are presented

1.2 Summary of different sections

Just to give an idea of what the book is about, we try to summarisehere the individual chapters Needless to say, the summaries are nosubstitutes to the articles

In the first article (Chapter 2) ‘When Welfare Economics and bling Studies Collide’, Lisa Farrell starts with a basic issue: how doesgambling affect welfare To start, she notes that the betting and gamingsector attracts a lot of interest from outside parties Those within thesector view their products as leisure activities competing in a toughmarket place, whilst groups outside of the industry question the (ethical)nature of these products and their contribution to society There are fewindustries (tobacco and alcohol included) that not only have to face thedaily struggle to earn a profit but also have fight to justify their very exist-ence It is not always easy to unravel some of these often emotive issuesand examine the link between gambling and welfare

Gam-Lisa Farrell shows that in a rapidly expanding market place it isimportant to be able to assess the impact of the gambling industry onsociety in order to inform the policy debate concerning the relaxation

of betting and gaming legislation and the licensing of new products.Her chapter explores the theoretical foundations of gambling impactstudies Such studies aim to weigh the costs associated with the bettingand gaming industry against the benefits in order to provide a figurewhich reflects the net monetary effect of the industry on society Moststudies find that the net impact is negative, i.e., that the costs aregreater than the benefits However, it can be shown that this result isdue to the fact that most studies ignore the direct consumption valuederived from the purchase of these products (as well as a number ofother theoretically inconsistent assumptions)

Farrell defines the total community costs/benefits of any activity as:Total benefits = private benefits + social benefits and Total costs = privatecosts + social costs This can be rearranged as: Net impact = total benefits– total costs = (Private Benefits – Private costs) + (Social Benefit – SocialCosts) In the literature to date it is common to include private costs but

to neglect to include the private benefits associated with the tion value of the product Other studies only look at social costs andbenefits and therefore totally disregard consumer preferences In essence,most studies are methodologically flawed and this renders them entirelynon-comparable.

consump-To address these shortcomings Farrell’s chapter reverts back to fare economics and looks at the theoretical definitions of private costs/benefits and social costs/benefits in order to aid applied researchers.Returning to first principles provides a useful insight into the existingstudies and allows us to suggest improvements to the current methodo-logies utilised by applied economists when conducting impact studies.However, even with firm theoretical foundations difficulties still exist

wel-in trywel-ing to apply theoretical models and concepts to complexity at realworld situations

In the second article ‘Demand Issues in the Market for Lotto andSimilar Games’ (Chapter 3), David Forrest seeks to review attempts byeconomists and statisticians to model the demand for online lotterygames such as lotto This idea behind this article is the notion thatknowing the demand patterns allows us both to make predictionsabout future developments and to assess the current market structurebehaviour of supplies of gambling services

The article starts with an overview of published research and recentresults obtained in a number of current lottery-related projects Thepurpose is to draw out the extent to which the modelling exercisesreported can illuminate the debate on the major trends and issues inthe world of lotteries

The dominating issue in demand studies is the sensitivity of demandwith respect to prices Forrest illustrates the importance of the concept

of price elasticity in assessing the correctness of the pricing policy byexisting lotteries He focuses not only on the price effects but also looks

at the effect of other products on the demand for individual games.This phenomenon – usually called cannibalisation – is important inthe modern gaming environment where new games appear contin-uously and the prices of existing games are manipulated even moreoften If the cross-price effects are ignored we are faced with clearlymisleading results in terms of the overall price sensitivity of lotteryproducts

Forrest points out that lotteries operate in a fast changing andincreasingly competitive gaming sector Worldwide, potential players

enjoy new opportunities to gamble via the internet, whether on lotteries,

betting or forms of gaming (such as poker) to which they may not

previously have had any access at all The availability of cheaperversions of their own products, and of new potential substitutes, poses a threat to the established operators in Europe and elsewhere and makes the maintenance of their traditionally high take-out ratesproblematic.

The problem of cannibalisation may not only be related to differentlottery products but also to other gambling activities, most notablycasinos Forrest summarises some recent evidence from Australia, whichseems to suggest that casinos do not exert great impact on the overalllevel of lottery expenditure except where the network of venues giveseasy access to the most of population

Most demand analyses have thus far been done with traditionaldemand equations which focus on the expected prices (by ‘prices’ werefer to the amounts that individuals pay for lotto – in the aggregate, it

is the inverse of the share of prizes out of total sales) Forrest reviewssome second generation demand equations taking into account thespecific risk features of the lotteries He concludes that it is not onlythe expected price that is important but also the distribution of prizes.Thus far, there has been relatively little evidence on these features Yet,Forrest convincingly argues that the new models provide a better basisfor analysing demand patterns, at the same time offering the suppliers

of lottery services new guidance on how to set the prizes

Ian Walker’s article ‘Lottery Design Lessons from Dismal Science’(Chapter 4) also deals with demand issues Basically, it explores the use

of estimates on how lottery sales vary, in order to understand howchanges in design might affect sales The essence of Walker’s analysis isbased on exploiting exogenous variation in the shape of the prize dis-tribution to explain sales variation over time In particular, he uses therollover-driven variation in the prize distribution as a source of exo-genous variation to test the hypothesis that players are motivated, inpart, by the skewness of the prize distribution While this idea has a longhistory, it has rarely been tested convincingly in the existing literature.Lotto games have strongly left skewed prize distributions – typically mostplayers lose their stakes and only a tiny proportion receive large prizes.However, when a rollover occurs, the size of the largest prizes on offerusually rise and this causes the degree of skewness in the prize distri-bution to change While the dataset that is used in Walker’s analysesdoes not feature any changes in the design parameters of the game, wecan exploit the fact that there is a deterministic relationship betweenthe mean, variance and skewness of the prize distribution (the distri-bution’s first three moments), as well as both the design parameters

and the level of the prize pool rolled over from the previous draw.Since we know how the moments depend on both of these factors, andsince we can estimate how sales depend on the moments, we can com-bine these to infer how sales depend on the design parameters evenwhen there has been no variation in those parameters across the sample.This modelling strategy complements that pursued by other researcherswho attempt to analyse the proximate determinants of sales (see, forexample, Forrest (2008) and references therein) Both strands of relatedliterature make arbitrary functional form assumptions In one case thefocus is on the role of the prize distribution and the assumption is thatthis can be captured through a ‘linear’ function of the moments of theprize distribution One advantage of this more structural approach thatemphasises the effect of the moments of the prize distribution on sales

is that it is explicitly couched within a theoretical framework that hasbeen widely used, in other markets, to analyse brand choices It iscommon for lottery markets to feature a portfolio of games each withparticular characteristics so that each game can be thought of as abrand This theory helps us to understand the relationship betweensales of different games in the same market and, in particular, enables

us to make inferences about how to design games

Sarah Smiths’s article ‘Lotteries as a Source of Revenue’ (Chapter 5)focuses on the highly controversial fiscal aspects of lotteries Smithnotices that lotteries run by states around the world are typically char-acterised by two features The first is a high tax rate While a lottery isnot itself a tax, as it is sometimes called, the fact is that almost theentire take-out rate (the price of a ticket minus the expected value ofthe prize) goes to the government as revenue

Smith shows that the tax rates on most lotteries are high – abovethose on other forms of gambling and those on alcohol and tobacco

In the new competitive gaming environment there is likely to be sure on governments to reduce lottery taxes Assuming that the currenttax rates are at the revenue-maximising level, which the evidencesuggest that they broadly are, cutting the tax rate is the appropriateresponse if demand becomes more price-sensitive, in order to continue

pres-to meet the goal of revenue-maximisation

There is no obvious economic efficiency argument for the fact thattaxes on lotteries are so high Taxes on vices, such as smoking or drink-ing are typically justified by these activities’ big negative externalities.But the evidence on problem gambling suggests that the negativeexternality argument is unlikely to justify such a high tax rate in thecase of lotteries Moreover, the fact that demand for lotteries is fairly

sensitive to price suggests that the consumer welfare loss associatedwith taxing the lottery is likely to be fairly high, and the so-calledRamsey rule would imply that the overall distortionary cost could bereduced by raising taxes on goods with relatively less sensitive demand,subject to distributional considerations.

Reducing lottery taxes would also be likely to improve the overallfairness of the tax system, although the impact would be very smallsince spending on lotteries is a small part of total income

Smith also looks at the issue of hypothecation, and asks whatreduced earmarked revenues might mean for the good causes and forthe lotteries themselves Smith points out that if lottery taxes were cut

it would mean a loss of earmarked revenue for the good causes The ence suggests that this would have little or no effect on the demand forthe lottery – for most people it is the jackpot that matters, not the goodcauses However, the loss would obviously be strongly opposed by thegood causes themselves Most studies (from the US) find that the goodcauses do benefit from earmarked revenues, albeit not by the full amount.While the good causes may have been seen as a useful bit of windowdressing when lotteries were introduced, they are likely to turn into apolitical headache if lottery taxes – and the good causes funding – have

evid-to be reduced

What is perhaps most interesting in Smith’s analysis is the ison of welfare and fiscal effects of the different taxes, including lotterytax If we take government expenditures as given we may ask howgood or bad lottery taxation really is Although we cannot necessarilyproduce definite numbers, we can establish a proper framework forcomparison and predictions

compar-Mark Griffiths’ article ‘Problem Gambling and European Lotteries’(Chapter 6) takes us a bit outside of the economic jargon and focuses

on the complex issue of problem gambling mainly from the point ofview of psychology Although most people gamble occasionally for funand pleasure, Mark Griffiths highlights that gambling brings with itinherent risks of personal and social harm in the form of problem gam-bling Problem gambling can compromise, disrupt or damage family,employment, personal or recreational pursuits More specifically, it cannegatively affect significant areas of a person’s life, including theirphysical and mental health, employment, finances and interpersonalrelationships His chapter examines the extent to which lotteries cancause or exacerbate problem gambling In doing this, Griffiths over-views problem gambling by examining the definitions of problem gam-bling and the social context in which people gamble He also examines

in detail the pathological features of problem gambling; briefly considerssome of the main psychometric screening tools that have been (andare currently) used in the field; and discusses the consequences and co-morbidities of problem gambling.

Griffiths argues that the type of gambling also impacts on the ment of gambling problems This has particular relevance for the lotterysector It also has consequences for understanding the risk factorsinvolved in the disorder, as well as the demographic profile of thoseindividuals who are most susceptible For instance, certain features ofgames are strongly associated with problem gambling These includegames that have a high event frequency (i.e., that are fast and allow forcontinual staking), that involve an element of skill or perceived skill,and that create ‘near misses’ – i.e., the illusion of having almost won(Griffiths, 1999) Size of jackpot and stakes, probability of winning (orperceived probability of winning), and the possibility of using credit toplay are also associated with higher levels of problematic play Gamesthat meet these criteria include electronic gaming machines (EGMs)and casino table games In general, lottery products are not associatedwith problem gambling with the exception of those that have thepotential for continuous gambling (e.g., video lottery terminals (VLTs),scratchcards, and some instant win games on the internet)

develop-Further, Griffiths overviews other specific areas in relation to lotteryproducts, including youth gambling and remote gambling (with par-ticular reference to internet gambling) The main form of problem gam-bling among adolescents has been the playing of slot machines although,

as Griffiths’ chapter highlights, a number of studies have also indicatedthat scratchcards can be problematic to a small minority

Griffiths goes on to argue that gambling is a multifaceted rather than

a single phenomenon Consequently, many factors may come into play

in various ways and at different levels of analysis (e.g., biological, social

or psychological) Central to the latest thinking is that no single level

of analysis is considered sufficient to explain either the aetiology or themaintenance of gambling behaviour Moreover, this view asserts thatall research is context-bound and should be analysed from a combined,

or biopsychosocial, perspective Griffiths also argues that the ational and structural characteristics of the gambling activity are impor-tant in the development of problem gambling and that many lotteryproducts have structural characteristics that are unlikely to facilitateproblem gambling behaviour

situ-Griffiths concludes his chapter by exploring some of the policyimplications for lotteries, including the type of research that needs to

be carried out, legislative policy implications and policy initiatives inthe areas of education, prevention and treatment.

Matti Viren’s article ‘The Economies of Scale and Scope in the LotteryIndustry’ (Chapter 7) takes us to the supply of gambling More concretely,Viren focuses on empirical evidence on scale economies in the lotteryindustry The analyses deal with both European and US state lotteries forthe period of the past 15 years The article is strongly empirical, althoughcertain conceptual and measurement issues are also discussed Most ofthe article is devoted to the estimation of the cost functions from dif-ferent panel data sets Thus, the record of individual countries is largelybypassed The analyses, which make use of a large number of data, more than 500 observations, show quite clearly that there is evidence of both economies of scale and scope In other words, average costs decreasealong with the volume production and the number of products Prac-tically the same result applies to Europe and the US

The paper has potentially important policy implications because theexistence of economics of scale (and scope) indicates that a competitivemarket solution would not probably prevail in the long run, at least in thecase where all legal restrictions to market entry and market operationswere abolished Because the biggest company would always have the low-est unit costs it could take over the whole market The existence of eco-nomics of scope suggests that, in the lottery industry, multi-productsuppliers would have an advantage over providers of a single lotteryproduct Thus, it would not be completely unrealistic to expect that wewould end up with a very large global multi-product gambling monopoly.Although the results of the empirical analysis seem quite clear, there

is still some uncertainty in terms of predictions of costs for very large(global) gambling companies Thus, it is possible that the long-termsolution would not literally be a natural monopoly, but rather mono-polistic competition between relatively few large market players Eventhen, prices (and consumer welfare) could be anything but an idealand perfect competition case Whether regulation could help is anissue briefly touched upon in the article

Viren’s article also presents some evidence on demand behaviour atthe aggregate (country/company) level The results seem to indicate thatthe aggregate demand is relatively sensitive to the output menu More-over, the sensitivity of demand seems to be related to the size of themarket in the sense that sensitiveness increases along with market size.Martin Paldam’s article ‘The Political Economy of Regulated Gam-bling’ (Chapter 8) deals with the public policy issues that are related togambling The idea is to apply some basic concepts and ideas from the

public policy research agenda to the analysis of gambling Paldampoints out that the gambling sector is regulated in order to collecttaxes and control the gambling problem of addiction The regulationmay be done without or through a state-owned gambling enterprisewith monopoly over some or all of the range of gambling products.Paldam introduces a macro model to identify the problem as excess netgambling expenditure and a dead weight loss.

The model is calibrated using Danish data, to give estimates of theaggregate cost to society of the gambling problem Paldam notices thatthe rents from public gambling monopolies are often distributed in away that creates stakeholders The role of these stakeholders representsitself an interesting research issue Paldam’s view is that those afflicted

by the gambling problem are few and unorganised Politicians are underlots of pressure to spend, and the stakeholders are normally powerfulgroups Consequently, the goal of tax collection easily dominates overthe goal of reducing the gambling problem

Paldam also considers efficiency issues: what are the implications ofownership structure for efficiency of production? From policy point ofview this issue is, of course, highly significant Yet, unfortunately, wehave very few hard empirical facts from the gambling world on thisspecific point

1.3 Final remarks

The present book does not cover all areas of the economics of gambling.Especially the area of public policy is so wide and includes so many differ-ent issues that it simply cannot be covered in one volume Because some

of these issues have not been analysed in depth in the existing literature,

it is even more difficult to make sensible assessments Issues like bling and crime’ have been subject to some research but the outcome hastypically been limited to the observation that gambling and crime areconnected somehow The exact nature of this connection has not beenidentified, nor has it been established empirically

‘gam-We have deliberately bypassed all the legal issues of gambling This hasbeen done even if we are well aware that these issues are both importantand timely We have wanted to concentrate on a broad economic per-spective avoiding, at the same time, the highly controversial legal andpolitical questions, which will require different sort of legal and politicalexpertise

Whilst dealing with the recent advances of research, this book alsocovers certain deficiencies in the analysis and data We have had par-

ticular problems with data For the demand analysis, there are tively few data sets, especially outside the UK and the US There is littleinformation usable for public policy type analyses, such as the fiscal issuesand cost conditions For instance, interesting issues of efficiency (effi-ciency comparisons) cannot be tackled with the existing information.The lack of data effectively discourages all research attempts on the part

rela-of the academic community, thus creating an obstacle for useful tion between researchers, on the one hand, and providers and regulators

coopera-of gambling services, on the other hand Whether for lack coopera-of data orother reasons, there has been almost no research on gambling in con-tinental Europe This is lamentable, since our market structure and insti-tutions are drastically different from those of the Anglo-Saxon countries,where most research results come from

The book has touched several controversial issues of gambling most

of which are related to the role of government in regulation, taxationand production of gambling services In this respect, this book doesnot arrive at a given simple result or conclusion Nor does it provideany policy recommendation Rather, it shows the level of complexity

of the issues we are dealing with We cannot, for instance, simply say

‘just look at the elasticities, stupid’ Rather we have to control a widerange of factors and issues before we can make affirmative conclusions

in any direction Needless to say, this requires a lot of additional research

It would also be valuable that both the general public and the makers would know much more about ‘the economics and gambling’and the issues behind the current controversies

2 The survey was conducted by the Finnish National Lottery Veikkaus inApril 2006 The questionnaire was sent to all of the EL (European StateLotteries and Toto Association) member companies of which nearly 50 percent responded.

3 These figures are obviously subject to the possible sampling errors in the

EL questionnaire The average payout rates may not be very informativebecause they vary considerably between games even now (and differentcompanies have different gaming menus) This was reflected in the above-mentioned survey: the payout ratios ranged from 30 to 95 per cent

4 Between 1894 and 1964, no legal government sponsored lotteries operated

in the United States Even other forms of gambling were severely restricted.Thus, only three states permitted horse racing in 1900 In 1931 Nevada legal-ised gambling casinos, slot machines and horse racing (but not lotteries,although now it seems likely that the constitutional ban for a state lotterywill be abolished)

5 Some lotteries are even older than that For instance, the Swedish lottery(Penninglotteriet) was founded in 1897, whereas the Finnish Lottery (Veik-kaus) was founded in 1940 In Denmark, the history of the Danish StateLottery goes back to 1753

6 Unfortunately, we do not have complete data on governments’ gamblingrevenues Although institutions like the Eurostat compile data on govern-ment revenues, the numbers are generally not comparable Sometimes thenumbers may only include direct lottery taxes and leave out the profitswhich are contributed to government (see the GOV_A_TAX_AG Table inthe public finances statistics, Eurostat, 2007) In the case of state lotteries’profits it is sometimes hard to say how to treat the direct contributions to

good causes that the lottery companies make If the government de facto

determines the beneficiaries, the contributions are hardly distinguishablefrom government transfers Finally, one has to keep in mind that, in somecountries, winnings constitute taxable income, whereas in most countriesthey are tax-free (on the average, the taxation of winnings only constitutes

a couple of basis points of the total government revenue)

7 When the respondents in the EL survey were asked to assess the elasticity ofproblem gambling to sales the mean value was roughly one

8 This could be seen in the survey to European lotteries Up to 70 per cent ofthe respondents said that if sales increased, say by 10 per cent, costs wouldincrease less Similarly, as we asked how much the retailers costs wouldincrease as a response to an eventual increase in sales, two-thirds of therespondents said that they would increase less than the sales

9 Legal cross-Atlantic disputes in the Microsoft case suggest that would notrun smoothly, either

10 Bribe might be paid both to fix the results or to encourage the use of illegalmeans to achieve better results (e.g., doping)

11 These are issues which, at least in the history of American lotteries, wereimportant and turned the public opinion against all types of gamblingduring the second half of the 19thcentury Various cases are quite elo-quently described in Clotfelter and Cook 1991 and Ezell (1960)

2

When Welfare Economics and

Gambling Studies Collide

Lisa Farrell

2.1 Introduction

The betting and gaming sector attracts a lot of interest from outsideparties Those within the sector view their products as leisure activitiescompeting in a tough market place, whilst groups outside of the in-dustry question the (ethical) nature of these products and their contri-bution to society There are few industries (tobacco and alcohol included)which not only have to face the daily struggle to earn a profit but alsohave to fight to justify their very existence In this chapter we will try tounravel some of these often emotive issues and examine the link betweengambling and welfare We will critically appraise different measures ofwelfare and existing studies that attempt to measure welfare effects Wewill consider the consequences/impacts of the expanding gaming market

on individual and social welfare It is not however, our intention to vide a full appraisal of the benefits and costs of the betting and gamingsector, the focus here is to provide a better understanding of themethodologies that practitioners use to address welfare issues, so that

pro-we can more clearly evaluate the results of such studies in order to makesensible, evidence-based, policy decisions It is hoped that the chapterprovides a valuable introduction to the important theoretical concepts

on which the existing empirical studies are based We shall presentonly first principles of welfare economics in order to make the materialaccessible to a board audience

The betting and gaming industry consists of a wide range of very ent products each with its’ own set of consumers, although some inter-section across these sets exists (gambling activities are not mutuallyexclusive and the difference characteristics of each game results in variation in the groups of consumers who participate – reflecting the

differ-heterogeneity in consumer tastes and preferences) The market sectorcomprises the following commercial activities: broadly these are i) wager-ing (both on and off course), ii) gaming in the form of casinos andelectronic gaming machines (EGM’s) and iii) numbers games such askeno, lotto and scratch cards To undertake a complete review of thissector is not the focus of this chapter, however there some notableattempts to undertake such an activity See, for example, the studyconducted by the Australian Productivity Commission (1999) whichprovides an extensive and complete picture of an entire countriesbetting and gaming market This review was in part possible due to thevastly superior data collected on the sector within Australia.1Here wewill predominantly refer to academic papers and independent reports

on the welfare effects of gambling in order to present objectiveresearch There are numerous reports published by interest groups butthe difficulty of determining the independence of these studies meansthat most have not been included in this review

There are some general trends across countries in terms of thebetting and gaming industry that are important to mention Mostcountries have undergone considerable deregulation in the last decadeand all countries face the challenge of trying to ‘police’ the growingonline gaming market The betting and gaming industry is at the fore-front of technology and is characterised by extensive product inno-vation which means it often operates within (and exploits) legislativegaps Historically, most countries imposed a stringent regulatory frame-work on the industry but a more liberal society and a realisation thatvaluable and significant tax revenues can be extracted from gamingactivities has lead to an increasing pattern of deregulation Trends inthe literature follow this pattern of deregulation As the limits of opera-tion within the market have changed, researchers have naturally askedwhat the effects of these changes are and from this perspective the literature makes a lot of sense For example, the issue of the merits ofIndian Reverse gaming is unique to the US market2and the resurgence

of literature on numbers games such as lotto by UK academics3 is adirect result of the launch of the new National Lottery game there

So the literature focuses on questions of national importance in thecontext of the domestic market, after all academic research, in part,evolves to meet the needs of policy-makers Simply to look at the liter-ature in its entirety and ask if the body or evidence mostly in favour

of against the existence of the industry is hennas crime which textualises the studies from their historical and cultural setting Eachcountry has its own social, political and cultural environment meshed

decon-with unique legislative and historical backgrounds, that means at bestthe question of the industries net welfare effects has to be asked foreach country in turn and even within countries there are numerousstate variations that can not be ignored.4That said, it is not uncom-mon to find government and consultancy reports that do exactly this.Whilst a shortage of studies (often explained by a paucity of data)means that often we do have to look to international studies to under-stand the impacts of legislative changes the warning still remains, thatwhilst the signs of the important coefficients may be indicative andhelpful, let the policy-maker beware of recommendations based in thesize of estimated qualitative effects based on international experience.Here, we will be proactive and ask, ‘What are the limitations of ourmethodologies and the available data that gives rise to this general lack

of evidence for policy-makers?’ We will look at the methods researchershave utilised in order to address questions relating to the welfare effects

of the gambling industry and consider how these theories sit withinthe theoretical framework of welfare economics It is our intention toprovide a structure within which to think about welfare effects in order

to be better able to evaluate existing and future studies Whilst we willnot provide an exhaustive review of all the literature we will look attypical examples In essence, we will consider what we can learn fromexisting studies and what are the limitations of these studies whichinhibit their generalisation

2.2 Individual welfare effects

Welfare economics has a long tradition in economics In measuringwelfare the distinction made is between individual welfare and socialwelfare Individual welfare considers the welfare of individuals (as opposed

to communities) under the belief that individuals are the best judge oftheir welfare Two essential assumptions are made when measuringindividual welfare and these are: firstly, individuals prefer more to lessand secondly, that individuals preferences are revealed through theirobserved behaviour This implies that if there is a demand for a goodthen the provision of this good must be welfare enhancing If this wasnot the case then individuals would simply switch their resources tothe consumption of commodities which afford them greater utility(satisfaction) In the context of the gambling market this implies that

so long as there is a demand for gambles then the provision of ducts to fulfil this demand generates welfare Moreover, it also impliesthat markets should expand to the point where there is no demand for

pro-new products The introduction of a pro-new gambling product will increaseconsumer welfare because utility-maximising individuals voluntarilypurchase the good That is, assuming income remains constant, indi-viduals must be substituting expenditure away from goods which yieldlower utility The introduction of a tax on the good will erode some

of this benefit, but the overall effect is still welfare improving for theplayer Whether this is socially desirable is another question and onewhich we will address in due course, but from the individuals perspec-tive there must be a welfare gain or there would be no demand for theproduct

Consumer surplus is a measure of the aggregate welfare of the group

of individuals who consume a given commodity It measures the ence between what consumers are willing to pay for a good and themarket price of the good that is what they actually pay It can be calcu-lated by looking at the area bounded by the demand curve and theequilibrium price in the market (the market price) This corresponds toarea PAB in Figure 2.1 Where P is the market price and Q is the quan-tity demanded

differ-Gambling products are usually taxed so it is important to considerthe impact the tax has on consumer surplus Let’s assume that someproportion of the tax t, is passed on to the consumer through higherprices It is evident from Figure 2.1 that this erodes away a proportion

of the consumer surplus – decreasing individual welfare The consumersurplus is now represented by the area (P + t)CA In the extreme casewhen all the tax is absorbed by the producer then the price remainsunchanged and the introduction of the tax has had a neutral impact

on the level of consumer surplus Brinner and Clotfelter (1975) andCook and Clotfelter (1987) both make this point in the context oflottery taxes One further point to note is that the consumer surplusstory is one of partial equilibrium, with more than one market andother methods of taxation It is welfare improving to raise revenue bytaxes that introduce the smallest efficiency loss, which may or may not

be taxes on gambling We know of no optimal taxation studies to sider gambling taxes However, Farrell (1997) does look at the depen-dence of lottery taxes, income tax and unearned income tax Thecomparative statics analysis within suggests that there is considerableinterdependence between direct taxation and the revenue that can beraised from gambling taxation Changes in the tax regime have asignificant impact on risk seeking behaviour and so gambling taxesshould not be perceived as lucrative new sources of revenue The mostrecent literature on the taxation of gambling centres on the issue of

con-what should be taxed, stakes (a commodity tax) or expenditure (an ad valorem tax), see Smith (2000) Paton et al (2001) show that an ad valorem tax is superior in terms of individual welfare (an ad valorem tax

results in lower prices and greater turnover) and efficiency (morerevenue is generated for a given level of price and quantity) However,

it is sufficient for our purposes to simply appreciate that the tion of a tax results in a fall in consumer surplus and so decreases indi-vidual welfare It is important to note that the welfare impacts oftaxation are further complicated by the redistributional effects of thetax and this point will be discussed in detail later

introduc-There have been few empirical studies that have attempted tomeasure the size of the consumer surplus for gambling products In thecase of lotteries Brinner and Clotfelter (1975) and Cook and Clotfelter(1987) both acknowledge that a welfare gain exists, but they concen-trate their analysis on the tax incidence (issues relating to the distribu-tion of the tax burden across the population will be discussed inSection 2.3.2) There are, however, two key studies that we will focus

on here The first is Farrell and Walker (1999) who illustrate themethodology for the case of lotto and the second is the ProductivityCommission (1999) who take this methodology and rise to the chal-lenge of estimating the consumer surplus associated with the entireAustralian gambling market

Farrell and Walker (1999) directly measure the welfare effects fromthe availability of the market for lottery tickets following the launch

of the UK National Lottery by measuring the associated consumersurplus.5This was a natural research question at the time the study wasconducted given that there was no lottery market in existence in the

UK prior to the National Lottery’s launch (with the exception of smallcharity raffles/lotteries) The analysis is slightly complicated by the factthat lottery tickets can only be purchased in integer amounts and thatthe real price of a lottery ticket is not the face value of the ticket butthe expected value of the ticket calculated as, one minus the expectedreturn We shall call this the effective price However, despite theseempirical complications the basic principle remains the same In thecase of lottery tickets the supply curve is perfectly elastic – a horizontalline crossing the x-axis at the effective price of a ticket for any givendraw Note that in the case of lottery tickets the effective price is lower

in a rollover week as the expected return is higher and this essentialprice variation allows us to measure the price elasticity of demand andhence to determine the slope of the demand curve The results of the analysis suggest that in a typical year (with an average number ofrollovers) the welfare gains from the game are in the order of £1 billion(equivalent to a 0.5 per cent reduction in income tax) suggesting con-siderable welfare gains! The relative size of the welfare gain is a reflec-tion of the popularity of the game, 63 per cent of the population played

on a regular basis at the time of the analysis (rising to 72 per cent in arollover draw) although they purchased on average only 1.5 tickets perdraw (rising to 2.3 tickets in a rollover draw).6

In principle it would be possible to conduct a similar analysis for allthe products offered by the betting and gaming sector in order to findthe total welfare impact However, in order to perform this analysis weneed to know the response of sales to changes in price (the price elas-ticity of demand) of all products in order to estimate the demand curves(given that consumer surplus is a measure of the area under the demandcurve and the equilibrium price) A general paucity of data makes thisexercise very difficult in practice To calculate price elasticities we need

to observe price variation and there is the added complication of standing what constitutes the price of a gamble It has become stan-dard practice to define price as the amount paid out by the consumerminus the takeout (player losses), where the takeout is the proportion

under-of any bet (sales revenues) that is not returned as winnings (the out covers supplier costs, profits and taxes) This definition of price iscalled the effective price in the Farrell and Walker example just dis-

take-cussed Under this definition price changes usually only occur whenthere are changes in the tax rates for the product which result in achange in the takeout (given that, the minimum amount that must bereturned to players as prize money is either defined under legislation or

an item in the license contract issued to the operator) Since tax ratechanges are not common events in practice (and tend to be non-random events) there are few studies which have estimated price elas-ticities Without knowledge of price elasticities we cannot determine theslope of the demand curve and it is therefore impossible to measure con-sumer surplus.7Lotteries are an exception where the price changes occurnaturally and more importantly randomly through the design of thegame, namely the occurrence of rollovers Indeed, the vast majority ofthe studies that have estimated price elasticities in gambling marketshave looked at lotteries

There are a few studies that have taken on the challenge of ing the consumer surplus associated with the entire gambling industryand one of these was conducted by the Australian Productivity Com-mission (1999) They suggest that the net benefit to the Australian con-sumer of the betting and gaming market is in the order of 4.4 to

estimat-6 billion Australian dollars It should, however, be noted that in order

to conduct the analysis a number of necessary modelling assumptionshave been made The most important of these is the assumption thatelasticities for gambling activities are constant across national bound-aries, across types of gambling products and through time

To obtain estimates of the required elasticities they look at all the

estimates currently available in the literature and select preferred

esti-mates The problem with this approach is that the elasticities vary bytype of product yet the Productivity Commission assumes them to beconstant across products and further they vary by country For exam-ple, estimates for the price elasticity of lotto are much higher for theUSA than they are for the UK and New Zealand Cloflelter and Cook(1990) report an elasticity of –2.55 for the USA, whereas Farrell andWalker (1998) report estimates of –1.55 to –2.6 for the UK and, BERL(1997) find an elasticity of –1.054 for New Zealand It is hard to knowhow these findings can be informative of the elasticity for Australiawhere there are no estimates in the literature Furthermore, some ofthe studies which estimate price elasticities for gambling products datefrom the 1970’s when the range of products available to the gamblerwas very different from what it is today (following the recent trend inthe liberalisation and deregulation) and this will have impacted onprice sensitivity This exercise is further complicated by the issue of

short run versus long run elasticities Initial responses to price changesmay be different from long run equilibrium responses However ourunderstanding of the dynamics of gambling participation is limited bythe fact that there is a paucity of panel data sets available from which

we can analyse behaviour through time

A further complication in the measurement of consumer surplus arisesdue to the presence of pathological gamblers in the market place Basiceconomic principals tell us that addictive goods have higher price elas-ticities and in the context of the betting and gaming industry, this sug-gests that addicted consumers will be less sensitive to price changesthan non-addicted consumers; implying that there is potentially morethan one demand curve for each gambling product For recreationalgamblers there is a whole market of leisure activities they can engage

in and so they are sensitive to price changes (resulting in a less steepdemand curve relative to pathological gamblers), problem gamblers, onthe other hand, face a smaller market to satisfy their demand for gam-bles and therefore are less price sensitive (resulting in a steeper demandcurve) In terms of measuring consumer surplus this means that we have

to sum the surplus for recreational gamblers and pathological gamblerstogether Figure 2.2 illustrates the associated consumer surpluses Theconsumer surplus (in the absence of any tax) for recreational gamblers

is the area PAB and the consumer surplus for problem gamblers is thearea PEF Again it is noted that the presence of tax reduces the con-sumer surpluses

FG