Access to Heritage
Table of Contents

5. Entry fees: economic efficiency arguments

`[The duties of the state are] first... that of protecting the society from the violence and invasion of other independent societies;... second... that of protecting, as far as possible, every member of the society from the injustice or oppression of every other member of it;... third... that of erecting and maintaining those publick institutions and those publick works which, though they may be in the highest degree advantageous to a great society, are of such a nature, that the profit could never repay the expence to any individual or small number of individuals.'

Adam Smith, The Wealth of Nations, 1776

5.1 The group of arguments discussed in chapter 4 are mostly put without reference to any particular groups in society - they concern proposed rights of all (to have free access to public property) or proposed duties of all (to pay for private benefits). The arguments that follow bring in concepts of economic efficiency and of equity and distributional justice. Note that the most common argument about economic efficiency (`fees discourage potential visitors') is answered most commonly not by direct rebuttal in terms of economic efficiency, but by an argument about equity (`free entry is a subsidy by poorer nongoers to richer goers') which happens to have opposite implications. This allows the possibility that both statements may be true: a tradeoff between efficiency and equity may be necessary.

5.2 We use `maximising economic efficiency' conventionally to mean the same as `maximising community welfare/social welfare'. Conventionally, we assume that total community welfare is the sum of the welfares (satisfactions, utilities) of all individuals, and that it is improved by any change in which the winners' gains outweigh the losers' losses (and vice versa). [1] In the present context, this means that if the lost satisfaction of a would-be visitor discouraged by an entry fee is not compensated by a gain somewhere else in the economy, it is a loss to `total community welfare'. This conceptual scheme disregards issues to do with distributional justice: these are considered in chapter 6.

`Fees discourage potential visitors'

5.3 Many submissions argued that user charges would discourage potential visitors, particularly poorer people. The proposition is presented as a problem because all accepted that wide public use of these public facilities is a fundamental good (this contrasts with the position of private enterprise, where `discouraging potential visitors' might well be the rational course if a business finds it can maximise profit by niche-marketing an expensive service to a few people). [2] The argument was made more strongly in relation to museums and galleries than national parks, probably because of the greater underlying concern in museum circles about the need for affirmative action to broaden their appeal.

`The [Western Australian] Museum is strongly opposed to the imposition of a general admission charge... “on philosophical, management and economic grounds'. The Museum believes that the small increases in revenue would be more than offset by increased costs, decreased attendance and other negative impacts on the museum's aims and functions.' (Government of Western Australia, submission 53 p464)

5.4 Some put the argument in context of general opposition to user charges, others put it merely to press for suitable concessions. These submissions were mostly in the form of general fears or warnings. Almost all who supported user charges said that they should not be so great as to discourage visitation (or `significantly' discourage visitation), and most who supported user charges claimed (particularly in relation to national parks) that in fact charges do not have much effect in discouraging visitation.

This depends on case by case facts about elasticity of demand

5.5 The basic question is, if a certain fee is charged, how much will visitation drop? In other words, what is the `price elasticity of demand'? And what are the effects broken down by socio-economic grouping? The first question is about economic efficiency: it is relevant for predicting the deadweight loss of community welfare resulting from imposing user charges - to be explained just below (it is also relevant to some significant non-economic questions, such as the likely loss of community esteem and support if visitation drops significantly). The second question is relevant to the equity problem: we wish to know not only how many would-be visitors are being discouraged by the fee, but also who they are - for example, whether they are disproportionately poorer people; whether they are occasional visitors who now no longer visit at all or more committed visitors curtailing their repeat visits (equity considerations are picked up in chapter 6).

5.6 Before considering the evidence we will briefly explain some key concepts.

Theory: willingness to pay, consumer surplus & elasticity of demand

5.7 A consumer's surplus is an amount that a buyer would be willing to pay for a good in excess of the market price. If the newspaper costs $1, but I would pay up to $1.50 for it, my willingness to pay is $1.50 and my consumer surplus is 50c. Different people would be willing to pay different amounts varying from some high figure (someone who needs that day's edition for an important personal notice, for example) down to zero. Their consumer surplus varies correspondingly from some high figure down to zero in the case of a person who values the paper at just $1 (those who value it at less than $1 do not buy).

5.8 If the price increases to $1.10, buyers who value it between $1.00 and $1.10 will drop out. `Price elasticity of demand' is the measure of what proportion of buyers drop out consequent on a certain price increase (or are attracted by a decrease). [3] If few drop out, elasticity is said to be low. A high willingness to pay relative to actual price suggests a low elasticity of demand. [4]

5.9 A `demand curve' is a graph on which price is plotted against quantity demanded. The slope of the curve at each price point shows how quickly demand is dropping with incremental increases in price. The shape of the curve (that is, the distribution of people's willingness to pay through the various price points) will be distinctive for each good in each market. In the case of museums/galleries and national parks, it may well be different from one place to another. For examples, see the demand curves tabulated in APPENDIX 6 (Wet Tropics World Heritage Area; Durham cathedral).

5.10 In this chapter we assume conventionally that willingness to pay (observed or estimated) is the measure of the value (`satisfaction', `utility') to individuals of the benefits that they are (or would be) paying for. [5] Accordingly, in the case of a non-rival public good free at the point of use, users' total willingnesss to pay for use represents the addition to `total community welfare' arising from their use. [6]

Estimating willingness to pay

5.11 In private markets willingness to pay is easily tested by observing the results of changes in price. In the case of public goods such as museums/galleries and national parks we may wish to estimate willingness to pay by other means. [7] Willingness to pay for community/ non-use benefits (see paragraph 4.37) through general taxation can be estimated by survey methods of which the best known is the contingent valuation method (which questions people on how they would behave in hypothetical markets). Such studies are now common in relation to environmental goods such as nature conservation, still rare in relation to cultural goods such as subsidy-assisted arts or cultural institutions. Willingness to pay for direct use may be estimated by various methods which, again, have been used commonly for recreational use of natural areas but (to the best of our knowledge) rarely for arts and culture. Some examples of studies of willingness to pay are noted in APPENDIX 6.

5.12 A popular method for estimating willingness to pay for a national park visit is the `travel cost method'. This relies on the observation that as travel cost increases demand falls off (people make more short journeys than long journeys); it assumes that behaviour in relation to entry fees is the same as in relation to travel costs (people treat entry fees as simply part of the travel package); and it uses survey data on people's revealed willingness to pay travel costs to predict their willingness to pay entry fees. [8] There are some theoretical and practical problems involved (chiefly, how to value the many components of travel, including the opportunity cost of time), so results must be considered approximate. [9]

Possible different attitudes to different types of expense

5.13 In practice, reported willingness to pay (elicited by survey questions such as `How much would you be willing to pay to enter the national park?') may be much smaller than `revealed consumer preference' estimated by observing people's actual behaviour in response to travel costs. [10] The main reason usually put forward for this is `strategic bias': when questioned people may nominate a figure less than their true willingness to pay in the hope that this will keep the price down. [11] Another possibility is that people may perceive different types of expense differently. Perhaps some people, whether for principled or emotional reasons, regard a $10 payment to the government for a national park - which they see as a public facility - as somehow more of an impost than a $100 payment to a motel-keeper on the road. [12]

`There are many limits to both consumer sovereignty and rationality... people, just like institutions, employ quite different ways of reasoning in different components of their lives - the outcomes will be ambiguous, inconsistent and even contradictory.... [13]

5.14 This possibility, if true, would rebut the basic assumption of the travel cost method that people's behaviour in response to entry fees is the same as in response to travel costs. [14] The point is relevant here because it relates to the possibility raised in paragraph 4.52 that even a small user charge, when applied to something traditionally seen as a `worthy' community facility, may arouse irrational objections and drive away potential visitors to an unexpected extent.

The `deadweight loss' from charging for non-rival public goods

5.15 User charges levied on a publicly supplied non-rival good (as defined in paragraph 4.29) in effect transfer income from the consumer surplus of individual users to the government revenue of community as a whole, where it can be applied for other beneficial community purposes. The direct users' loss is the community's gain, but the total benefit (the addition to `total community welfare' arising from use of the good) is unchanged - providing total use continues unchanged (that is, providing demand at the chosen price is completely inelastic). It is analogous to raising the price of the newspaper from $1 to $1.10, supposing that all users are actually willing to pay something more than $1.10. The price rise leaves the buyers poorer and the publisher correspondingly richer; but since both are part of the community as a whole, the net position of the community as a whole is unchanged.

5.16 On the other hand, if some would-be users are discouraged (the charge is greater than their willingness to pay) they do not get the benefit of the good any more. But a marginal reduction in use does not reduce the cost of providing the good (because it is a non-rival public good provided in one lump to all: the cost of accommodating an extra user is zero. Qualifications are discussed from paragraph 5.22). From the point of view of the community as a whole some people's loss of benefit, not compensated by any economy in public running costs, is a deadweight loss. [15]

5.17 The deadweight loss of welfare is an unavoidable detriment of any public policy decision to shift part of the cost burden of a non-rival public good from general taxpayers to direct users. It is obviously important to know how big it is - the larger the loss, the more it suggests that for the sake of maximising total welfare user charges should not be imposed. [16] The question depends on the facts about elasticity of demand: how much does a certain charge drive away potential users? The evidence on this is picked up from paragraph 5.49.

Principle of marginal cost pricing of public goods

5.18 A common statement of conventional economic theory is that to maximise total community welfare the price of a public good should be set at the marginal cost of production. [17] In brief, to charge less creates a situation where the consumer of the last unit of output values it (is willing to pay) less than the marginal cost of producing it: in other words, the last unit costs more to produce than the value of the benefits it creates: the resources used to produce it could be employed to more benefit elsewhere in the economy. To charge more than marginal cost creates the opposite situation, where the last discouraged consumer would value it at more than the marginal cost of production. This is the position described in the previous section: the benefits forgone by discouraged consumers are the deadweight loss. The result is fairly obvious where the marginal cost is zero; the marginal cost pricing rule generalises to situations where the marginal cost is not zero. [18] (Some further textbook explanations of this are in APPENDIX 3, `marginal cost pricing') [19]

5.19 The marginal cost pricing rule assumes (as is usually true) that through all price levels demand will be elastic to some degree: there is no such thing as a `modest' charge which does not affect demand at all. If in fact demand is completely inelastic, even at some price higher than marginal cost, no-one drops out so there is no deadweight loss. [20]

5.20 In the case of non-rival public goods provided in one lump to all (for example, a bridge), the marginal cost of serving an extra user (allowing an extra person to cross the bridge) is zero, and accordingly the most efficient (welfare maximising) charge is zero. [21] This is simply another way of saying what we have already noted in paragraph 5.16: to bar the extra user causes a deadweight loss of welfare for that person without any compensating saving to public running costs.

5.21 Arguably a museum or gallery is a non-rival good analogous to the bridge. Once the community, for public interest reasons, has decided to maintain a collection open at certain hours, the costs of this may be regarded as fixed; the marginal cost is the cost of admitting a visitor during those hours. This cost is practically zero, and so should be the charge. Similarly in principle in national parks; though in that case marginal costs associated with visitation (rubbish collection, track maintenance, environmental remediation...) are more obvious.

Problem of defining `marginal' costs

5.22 We now digress to pre-empt a possible criticism of the above statement. The criticism is that our assumption about where the boundary lies between fixed costs and marginal costs is arguable. In the case of a museum, it could equally be argued that the relevant fixed cost is `maintaining the collection' and the relevant marginal cost associated with visitors is `opening the doors'. In that case the marginal cost becomes not `practically zero', but rather `paying attendants' - and the marginal cost pricing rule suggests that this cost should be charged to visitors. Or it could be argued that the relevant marginal cost includes the considerable cost of housing the collection in an accessible city building for display purposes, rather than a cheap suburban warehouse. In each case, the supposedly `marginal' cost, translated into a charge averaged over the relevant users as a group, increases. At the limit, it could be argued that the very existence of the museum is simply a marginal element of `the state's cultural services', and accordingly the whole cost of it should be charged out to its various users.

5.23 Which costs are rightly called `marginal'? In the Tasmanian Wilderness World Heritage Area, the question is how much general track management costs (including backlog maintenance) should be charged to today's walkers:

`A permit fee of $5 per person per night would generate income of $200,000 per year, only slightly more than the sum required to fund the development and annual operation of the permit system. A fee of $10 per night would fund the system and contribute approximately $200,00 towards continuing track maintenance at existing levels. If permit fees were the only track management funding available, a fee of $35 per night would be required to fund the [Walking Track Management] Strategy in its entirety.' [22]

5.24 A similar issue arose in the 1996 controversy over increases to the Great Barrier Reef Environmental Management Charge - the issue there being, whether the `research, education and management projects' which the charge funds [23] should, in principle be regarded as `core services' and budget funded. [24]

5.25 In the economic theory of private firms these questions do not arise because there is no concept of `core services' carried on for community benefit reasons. The frame of reference is simply a period of time: the marginal cost of producing an extra widget next week is one thing; the marginal cost of producing it next year is another, and these calculations determine how the firm expands or contracts production over time. [25] But in the case of `merit goods' such as museums and art galleries `period of time' is not a useful criterion, since we wish to maintain them in some relatively fixed form indefinitely.

5.26 Textbook statements about public goods (`price should be set at marginal cost') tend to overlook this problem: they implicitly envisage the textbook public good (by analogy with the output of widgets) as a homogenous mass of which an extra unit may be produced. In fact museums/galleries and national parks are carrying on a package of activities in different regional markets with different elements of public good, community benefit and opportunities for charging. Their problem is not how to charge for an extra unit of a homogenous service, but how to charge for a stable output of a mixture of different services.

5.27 A possible solution to this problem is simply to say this: the relevant marginal cost, for the purpose of applying the marginal cost pricing rule, is the avoidable cost of a service which, if it cannot recover costs at marginal cost pricing (not enough people are willing to pay), management would discontinue without hesitation. If management does hesitate, it is a clear indication that it regards the service as `core' (that is, providing community/ non-use benefits) to some degree. [26]

5.28 Once we have thus defined the relevant `marginal cost', the rule that the price should be set at marginal cost can be invoked. If a museum regards its core business as maintaining a certain collection open to the public at certain minimum politically acceptable hours, the cost of doing this should be regarded as fixed; the relevant marginal cost, and the efficient charge, is the cost of admitting another visitor during those hours (practically zero). [27] At the opposite extreme, if the community regards the very existence of the museum as a dispensible item at the margin of `the state's cultural expenditure', then logically its whole costs should be charged to visitors, and if not enough people are willing to pay the museum should be closed. [28] This type of exercise would soon flush out people's real feelings on which services are `core'. [29]

5.29 The committee suggests that the former is the more common situation, and accordingly the statement at the end of the previous section may stand: the efficient marginal charge for admission to the museum is practically zero.

Marginal cost pricing versus `charge for private benefits'

5.30 The principle of marginal cost pricing for public goods conflicts with the proposition put by supporters of entry fees that users should pay for the private benefits from use. The fact that public goods are supplied communally (because of non-rivalry or non-excludability) has no particular bearing on who gets what benefits from them. The marginal cost of providing a service may be zero, yet the benefits may be mostly private (admitting a traveller to a non-congested bridge, for example). [30] In this case marginal cost pricing suggests a price of zero; charge for private benefit suggests some charge, possibly a high one given that all or most of the benefits are private. [31]

5.31 The statement that `users should be charged for private benefits' is based on an analogy with private goods. It is not necessarily appropriate when applied to public goods. Whether it is appropriate depends entirely on the facts about willingness to pay, elasticity of demand and the deadweight loss on a case by case basis. Where demand is inelastic and few visitors are discouraged by a charge, the deadweight loss is small: the charge acts merely to transfer benefits from the consumer surplus of direct users to the government treasury: it is not objectionable on grounds of overall economic welfare (though it may still be objectionable for emotional and symbolic `citizenship' reasons (chapter 4), equity reasons (chapter 6), or for creating pressure for inappropriate developments in order to maximise income (chapter 7)). Where demand is significantly elastic (assuming marginal costs are low), the charge may well be objectionable because of the deadweight loss. From paragraph 5.49 we consider the inquiry's evidence on these points.

Charging for private benefits: a hypothetical case

5.32 A community of 1 million people maintains a museum which costs $1million a year to run. Everyone contributes $1 a year towards the museum through taxation. $1million is the fixed cost of maintaining the collection and opening the doors at hours approved by the community; the marginal (avoidable) cost of admitting a visitor during those hours is zero.

5.33 The museum receives 100,000 visits per year (assume for simplicity that this consists of 100,000 people each making one visit). If the whole cost of the museum were charged to them it would be $10 per visit. Free entry is subsidised by the visitor's own tax payment $1 and other people's tax payments $9. The 900,000 non-visitors subsidise the 100,000 visitors. The community maintains the museum primarily because of perceived community/non-use benefits: the subsidy to private benefits of direct users is a windfall gain to them which the community tolerates because of the perceived community benefits of wide access.

5.34 Attitudes change: one day the community decides that it wishes to charge visitors for the private benefit of their visit. A survey estimates that average willingness to pay for direct use is $2. (that is, total willingness to pay for direct use is $200,000; this means that community/ non-use benefits must be worth at least $800,000, assuming the community is acting rationally in funding the museum with $1million). A charge of $2 is introduced. (Alternatively, coming from the other direction: the community decides that community/ non-use benefits are really worth only $800,000: on the face of it visitors must be charged to make up the $200,000 deficit.)

5.35 However [this is the crux] $2 is only the average willingness to pay over the group of visitors. Half the visitors value a visit at more than $2, half at less. When a charge of $2 is introduced, half the visitors drop out. Revenue is not $200,000, as hoped, but $100,000.

5.36 Remaining visitors, being more committed, are willing to pay, on average, something more than $2 (alternatively: we still have a deficit of $100,000 to make up). We repeat the process until at some point revenue is maximised (revenue can never reach $200,000 because at any price some visitors are lost and their willingness to pay is not captured).

5.37 Suppose revenue is maximised with 30,000 visitors at $5 each = $150,000 (any further increase in price is outweighed by the further loss of visitors). The cost of the museum to the public purse is reduced from $1million to $850,000. The average cost per visit is now $33.33, consisting of entry fee $5, visitor's own tax payment $0.85, other people's tax payments $27.48. The subsidy to each visitor by others has decreased from 90 per cent to 82 per cent of fully distributed costs; in absolute terms it has increased from $9 to $27.48. The 70,000 discouraged visitors represent a deadweight loss of welfare: that is, they could gain satisfaction from being admitted without imposing any extra cost on the museum.

5.38 Some options now are:
1. Regard this as optimal for satisfying the philosophical position that visitors should pay for private benefits from use. At $850,000 the subsidy has been minimised in total, and the subsidy to each visitor by others, as a percentage of average costs per visit, has also decreased.
2. Alternatively: the community has decided that the community/ non-use benefits of the museum's existence are worthy only $800,000 a year; it still has a `deficit' of $50,000 which it cannot cover through entry fees (not enough people are willing to pay); the costs are greater than the benefits; the museum should be closed so that the money it costs may be reallocated to more benefit elsewhere in the economy.
3. Revert to free entry. This accepts that a windfall gain to direct users is unavoidable if it is desired to maintain wide access to the museum for community benefit reasons.
4. Adopt some compromise position (50,000 visitors at $2, for example).

Comment

5.39 1. The figures are chosen to make the point. No resemblance to any real museum is intended.

5.40 2. The point is that a optimal entry fee (one that is a politically satisfactory balance between revenue-raising and loss of visitors) must be decided with reference to the actual distribution of willingness to pay (the shape of the demand curve), not by aiming for a ratio of cost recovery suggested by analogy with some estimate of the ratio between total use benefits and total non-use benefits.

5.41 3. In the example median willingness to pay (the charge at which half the visitors drop out) equalled average willingness to pay ($2). In reality the median will usually be lower, since the average is pulled up by a few high bids, and the median is pulled down by many low bids. This worsens the problem.

5.42 4. The problem is worse, the greater the disproportion between median willingness to pay for direct use and full costs ($2 and $10 in the example). The greater the disproportion, the more it suggests that the place should either be maintained with free entry or closed.

5.43 5. The problem is worse if the community perceives not only a `fixed' community benefit from the existence of the museum, but also a `variable' community benefit depending on the amount of use (that is, the community regards the museum as a `merit good', the use of which creates spillover benefits for the community). On this basis, if the community values the 100,000-visit museum at $800,000, it will value the 30,000-visit museum at something less, increasing the `deficit' which the entry fee is asked to make up.

5.44 6. The example deals with equity between users and non-users (which is the implicit foundation of the view that `users should be charged for private benefits'); it does not deal with equity between rich and poor, more educated and less educated etc. In practice a free-entry museum with 100,000 visitors may well be a different sort of place from a charging museum with 30,000 visitors, both in itself and in community perceptions of it.

5.45 7. With the charge, the total subsidy to visitors by others is decreased, while the subsidy per visit is increased. Whether this is a better position in terms of equity between users and non-users is a matter of opinion.

5.46 8. In strict economic terms the lost welfare of discouraged visitors may be small, if they valued a visit at (were willing to pay) close to zero. In political terms the number of discouraged visitors is obviously very important, especially if there is a `merit good' argument that the visits of less motivated people should be particularly encouraged.

5.47 9. Short of closure, the 30,000-visit museum at $5 is clearly optimal for minimising public subsidy. The 100,000-visit free entry museum is clearly optimal for encouraging wide access. If a compromise is desired these conflicting aims must be weighed up, and this is fundamentally a matter of political judgment, not economic analysis (though there may be cases where a best compromise is suggested by the shape of the demand curve).

5.48 10. The example speaks of visitation. We should not forget other functions of museums, such as research. In principle their costs and benefits can be analysed similarly. In practice allocating the fixed costs of the museum among its various functions may be problematic.

Evidence on elasticity of demand

Elasticity of demand for national parks

5.49 The Committee's evidence mostly suggested that in general elasticity of demand for national parks is low at present prices.

`The scenarios where fees are modelled show that with modest fees of around $2-5 per visitor-day, visitor growth [to the Wet Tropics World Heritage Area] will decline only marginally.' [32]

`The [1991] ESD report pointed out that there was no evidence to suggest that the (then) current level of [national park] charges was a disincentive to visitation.' (Treasury, submission 20 p174)

`In New South Wales, a `park use' fee applies to less than 20 per cent of all parks in the State and then primarily only at weekends. The charge, when introduced saw a drop off in admissions of about 10 per cent. Over time the numbers grew back to previous levels... Extensive surveys undertaken in NSW Parks (1993/4) show a substantial willingness by a majority of visitors to contribute to the upkeep of parks.' (M Donovan, submission 6 p48,49)

`But by and large, over small increases, you are not getting a drop-off in demand... Let us talk about the famous Reef Tax increase from $1 to $6. That would not have made any difference in terms of visitation... Fees of the magnitude I am talking about are on that part of the demand curve where it is fairly inelastic.' (Prof. T Hundloe, evidence 7 August 1997 p378)

5.50 On the other hand, a recent report for the New South Wales National Parks and Wildlife Service has more mixed conclusions:

`This empirical evidence [recent studies of Minnamurra, Dorrigo, Gibraltar Range & Kosciuszko national parks and Hartley historic site] seems to support the notion that visitation to national parks with high levels of infrastructure development and that offer outstanding visitor experiences for which there are few if any substitutes, eg Kosciuszko, Dorrigo and Minnamurra national parks, is price inelastic. While visitation to those parks that offer basic facilities and for which there may be a number of substitutes, such as Gibraltar Range national park, is price elastic.... Economic theory, empirical evidence and revenue data all tend to suggest that demand for many national parks may be price elastic and therefore the appropriate strategy for increasing the Service's revenue is to decrease entrance fees in a number of national parks.' [33]

5.51 More details on this and some other studies of willingness to pay for particular national parks are in APPENDIX 6. Given the variety of national parks, and the variety of uses (daytrip, camping weekend, holiday of a lifetime...), we should be cautious of generalising from studies of particular places. The safest comment is that price elasticity of demand probably varies a lot from one park to another depending on the circumstances, particularly whether the park offers a rare or unique experience with a high drawing power. A modest entry fee might have no noticeable effect on visits to Kakadu or Uluru; but it might have a marked effect on visits to a near-city national park where the socio-economic profile of visitors is probably lower, the fee is a greater proportion of visitors' total costs, and there are more alternative recreations in the area. For example:

`...the introduction of entrance fees at the Port Arthur Historic Site in Tasmania was connected to a 25 per cent decrease in visitation and a shift from recreationists to tourists with relatively high socio-economic profiles. “With charging came a change in perception by local Tasmanians, Port Arthur was no longer a pleasant place to have a picnic under the famous oak trees but a `tourist attraction' or at best a museum”.' [34]

Elasticity of demand for museums/galleries

5.52 The Committee has heard of little research on willingness to pay for visiting particular museums and galleries comparable to that mentioned above in relation to national parks. The uncontested position that museum/gallery visitors are disproportionately drawn from the educated professional classes suggests, on the face of it, that elasticity of demand should be low: the intuitive suggestion is that these people are committed visitors, and have money, and so should not be discouraged by `modest' charges.

5.53 On the other hand, there are anecdotal views that suggest the opposite - such as the common statement that people see museums as a good place to bring the kids on a wet Sunday afternoon. [35] The `wet Sunday afternoon' visitors, coming to the museum because it is free, will be much more likely to drop out if it is no longer free. We note also a 1991 Australia Council study of South Australian art gallery visitors: 71 per cent of surveyed visitors thought that entry should be free, and 44 per cent said that an admission charge would discourage their attendance. The likelihood of `strategic bias' in these answers should be remembered, as explained in paragraph 5.13 (visitors would have an incentive to overstate their opposition to a charge); but the fact that 44 per cent say they would be discouraged by entry fees is still suggestive, and tends to contradict the stereotyped view of art gallery visitors as predominantly a well-to-do elite of committed goers. [36]

5.54 The Committee received disturbing evidence that contradicts the intuitive suggestion of low elasticity of demand, and supports the anecdotal views suggesting higher elasticity - evidence of museums and galleries which have introduced entry fees in recent years and suffered unexpectedly severe drops in visitation, or which have taken them away and enjoyed large increases. This may be reconciled with the intuitive suggestion about committed visitors by postulating that the dropouts are drawn disproportionately from the occasional/uncommitted visitors who are a larger proportion of the whole population (see paragraph 6.23). [37]

5.55 We quote at length because of the importance of this matter:

`Historically, state-funded museums in Australia allowed free public entry. This began to change in the early `90s. In New South Wales, the Powerhouse Museum was the first to charge (September 1991) followed by the Australian Museum (February 1992). The Museum of Victoria, a pioneer in this area, had introduced a general entry charge in November 1990. The data available after seven years of general entry fee levy is sufficient to allow us to make some observations. There is a clear trend, the introduction of entry fees results in a marked reduction in museum visits. The degree of visitor collapse is variable, with falls of between 30 per cent and 80 per cent observed.... This effect cannot be dismissed as a temporary phenomenon...
`The Western Australian Maritime Museum introduced an entry fee to capitalise on the tourism generated by the America Cup races when they were held off that city in 1987. Staff estimate that despite a special exhibition program held at the time, about 40 per cent of potential visitors turned away at the door rather than pay for admittance.
`In 1987 the Otago Museum [Dunedin, New Zealand] introduced an entry fee and experienced a 75 per cent collapse in visitation. Income generated at the door was less than had been previously collected through donations. In the light of this experience, the entry fee was removed in 1988 and visitation rapidly recovered.
`Similarly, the introduction of entry fees at the Old Parliament House Museum in Adelaide in January 1988 precipitated a 58 per cent fall in visitor numbers (average annual visitation prior to charging 88,700 and after the introduction of fees 36,900). Visitation has shown no sign of recovery in over ten years. It is worth noting that there has been no increase in income.' (P Filmer-Sankey, Newcastle Regional Museum, submission 56 p522-3)

`The weight of New Zealand and overseas experience shows that after admission charges are introduced to an existing free entry institution, visitor numbers decline and generally remain lower than before. This is also true where a new museum introduces charges after an initial period of free entry. A point of entry charge may be expected to reduce estimated total visitor numbers by 15 to 40 per cent. Auckland Institute and Museum anticipated a 50 per cent reduction. Australian museums experienced falls of 50 per cent, in the Australian Museum [Sydney], to 80 per cent in the Museum of Victoria. Canadian modelling exercises for revenue generation in national museums assumed around a 20 per cent drop in visitor numbers. National museums in Great Britain experienced an average 43 per cent drop when admission charges were introduced in 1974...' (Museum of New Zealand Te Papa Tongarewa, Admission Charges - the issues, Wellington 1994, p32)

`In 1996 the National Gallery of Victoria abolished its entry fees. Attendance close to doubled on 1995 figures. In 1989 the Powerhouse Museum [Sydney] attendance was 1,500,000. In the year that admission charges were levied, 1991, this fell to less than 700,000.... The Science Museum in London saw visitor numbers fall by 55 per cent after introducing a £2.50 fee. The Victoria and Albert experienced a 35 per cent drop after a smaller charge [£2.00] was introduced. Clearly the impact of admission charges is a critical issue.' (Australia Council, submission 35 p285)

`...there are numerous instances where this pattern has been repeated, with variations in terms of the percentage impact. However, in Birmingham, the museum dropped charges for all of its external sites and the visitor rates soared by an average of 126 per cent. For example at Birmingham Museum's Nature Centre attendances jumped from 65,000 to 360,000 - a 300 per cent increase.' (Museums Australia Inc., submission 25 p214)

5.56 The evidence is not all on one side. The Museum of New Zealand Te Papa Tongarewa, source of the second extract above, also cautioned:

`Evidence on the relationship of charging to attendances is ambivalent, however... Higher prices may not reduce visitor demand, contrary to traditional economic theory [Cooper-Martin]. In 1993 many of the charging UK national museums had visitor increases, while the free British Museum and National Gallery both had fewer visitors.' (Museum of New Zealand Te Papa Tongarewa, Admission Charges - the issues, Wellington 1994, p33)

5.57 The Queensland Museum (which has free entry in Brisbane, and entry fees at country branches) considers that `...the application of modest admission charges to cover some of the public costs of maintaining our natural and cultural heritage assets does not seem to deter visitors and is not seen as having a significant adverse effect on access and equity.' [38] The Australian Museum (Sydney), which has an entry fee, comments:

`There appears however to be no clear relationship between general admission charges and visitor numbers and profile... the Powerhouse Museum in Sydney increased general admission charges and made an additions charge for the special exhibition showing at the time, and the numbers of visitors increased...' (Australian Museum, submission 30 p241)

5.58 In the case of special exhibitions, resistance to paying may depend partly on the type of exhibition concerned:

`...The attendances at the Australian charged exhibitions are lower than those at the international charged exhibitions... It would be very difficult to attempt to charge for contemporary art exhibitions.' (Dr C Turner, Queensland Art Gallery, evidence 21 May 1997 p121)

5.59 Some question the accuracy of visitor counts:

`When anyone gives you figures on attendances where they do not charge, treat those figures with great suspicion. They will not be accurate. There is no way they can be accurate - everyone will report that attendances drop; they will naturally do that because for the first time they have an accurate record when they do charge.' (G Morris, Museum of Victoria, evidence 15 September 1997 p407)

5.60 - but others deny that this is a problem:

`The fact is that most [free-entry] institutions made very serious attempts to estimate their visitor numbers.... To imply that the pre- and post-charge visitors level differences are a result of improperly applied sampling is to suggest that the museum administrations, prior to the introduction of fees, effectively practised fraud on a sustained and massive scale. No honest mistake could overcount by factors of between two and five.' (P Filmer-Sankey, Newcastle Regional Museum, submission 56 p529)

5.61 In analysing visitation figures we must also be alert for possible biases which complicate trends over time or complicate comparison of charging and non-charging institutions. These include the fluctuations caused by special exhibitions or marketing campaigns; seasonal factors; the general state of the economy and its effect on recreational spending; and competition from other attractions (which naturally differs from one city to another).

`I just caution anyone who says that they drop really dramatically: do not rely on those. Look for secondary indicators, such as shop revenue and restaurant revenue.' (G Morris, Museum of Victoria, evidence 15 September 1997 p408)

`The opening of new museums and the development of competing attractions including theme parks and large-screen theatres has increased competition among museums. In a city like Sydney the competition is fierce.' (Australian Museum, submission 30 p242)

5.62 As well, there is the possibility of a honeymoon effect where charges are removed or a corresponding bounce-back effect where they are imposed. On this opinions were mixed:

`As you know, the National Gallery of Victoria took its charge off [on 1 July 1996] and reported a lot of people going there. I do not doubt for a moment that that is true. [But] Why they went there has not been tested.... What did they do there? Did they just go to enjoy the free entry for a while? Was is a novelty? Will it drop off?' (G Morris, Museum of Victoria, evidence 15 September 1997 p409)

Senator Hogg: `You see, anecdotally in some areas when you levy a user-pays charge, you find that the usage by the public drops off as a result...'
Mr Dawes: `Without being able to quote the exact figures, that has been the experience in a number of museums that I am aware of. There is an initial drop-off which is quite marked and then there is some recovery, but never a full recovery to previous numbers.' (M Dawes, Australian War Memorial, evidence 6 August 1997 p254)

5.63 On the last point, a supporter of free entry comments:

`There is still some debate about the long term effect of entry fee introduction. Some institutions express optimism that the acknowledged depression in visitation is showing signs of lifting (Australian Museum 1995/96 annual report). This view is sometimes also shared by funding agencies. It is almost invariably the case that temporary rises in visitor numbers are the response of the public, who are now beginning to treat museums as just another discretionary entertainment option, to some large temporary exhibition... [In the case of the Australian Museum, Powerhouse Museum and Museum of Victoria] `... visitor numbers have crashed to around one third of the pre-entry fees levels, more significantly, they have stayed down. There has been no effective recovery, and despite the odd optimistic phrase in annual reports unsupported by data, there is no reason to expect recovery...
[conversely] `The Newcastle Regional Museum is now well into its second year of free entry... the very significant increase in visitation commencing in July 1996 gives every indication of being stable. Attempts to detect a “honeymoon effect” have failed...' (P Filmer-Sankey, Newcastle Regional Museum, submission 56 p525, 588)

5.64 More generally, there is the possibility of some general trend in society that is turning people away from museums (the `complex phenomenon' hypothesis). To this supporters of free entry answer:

`We have seen, however, that the correlation between a sudden collapse in visitation and the introduction of entry fees is very sharp indeed. We have also seen almost identical examples from the northern and southern hemispheres spread over a period of more than five years. There are no significant common factors other than the introduction of entry fees... Let us remember that no such fall has been observed in institutions such as the Museum of South Australia, Western Australia and the War Memorial that maintained free entry... If some common factor or factors other than the entry fee is responsible for the observed visitor decline then the onus is on the apologists for entry fees to identify it or them.' (P Filmer-Sankey, Newcastle Regional Museum, submission 56 pp584,588)

5.65 A point of particular anomalies in the evidence was `free days'. Some, as part of an argument against entry fees, pointed out the popularity of free days:

`The Australian Museum enjoys up to 10,000 visitors on its annual free open day while the National Maritime Museum hosts up to 20,000 on its equivalent free day. Some institutions offer regular free entry days (Powerhouse - one a month) or the National Gallery of Victoria (free Mondays prior to 1/7/96) [when general entry fees were removed]. Attendances on such days were and are very significantly higher than charged entry days.' (P Filmer-Sankey, Newcastle Regional Museum, submission 56 p526) [39]

5.66 But the Queensland Museum reported contrary examples:

`At the commencement of operations of the first three Branches in provincial Queensland, in Gympie, Townsville and Toowoomba, one free day a month was advertised... there was no definite indication that local residents took advantage of the free day, which provided an amusing bonus to tourist visitors who had the good fortune to turn up on the first Monday of the month. The free day has now been discontinued at all Branches.' (Queensland Museum, submission 31 p249)

5.67 A possible explanation for the anomaly is that different museums/galleries may advertise their free days with different effectiveness. If it becomes fixed in people's minds that the museum charges, it is a chancy matter whether they find out that occasionally it doesn't.

5.68 All the above evidence, unfortunately, gives practically no information on how big were the charges that caused the stated reactions. The point is relevant for assessing the rule proposed by most supporter of entry fees - namely, that fees should be `modest' so as not to discourage visitation. This proposed rule makes an assumption about elasticity of demand: it assumes that at low prices elasticity is low (few people are discouraged), and at higher prices elasticity is higher. If in practice even modest fees do significantly discourage visitation, the proposed rule, and the case for entry fees, is severely undermined.

5.69 In evidence to the inquiry actual dollars were mentioned occasionally, anecdotally, but not so much as would allow the Committee to draw firm conclusions on this point. On balance the Committee sees no reason to be confident of the proposed rule. It may well be that even `modest' fees do significantly discourage visitation. For example, it is sometimes said that if you are going to charge an entry fee, you might as well make it $10 rather than $2. By implication, even at $2 significant damage to visitation is done. The following comment from the USA supports this idea:

`J Carter Brown [in panel discussion] said that when the Museum of Fine Arts first instituted an admissions fee, attendance fell sharply. Feldstein wondered if there could be a big effect of moving from free admittance to paid admittance but a small effect of marginal changes in the fee once it exists.' (Feldstein M ed., The Economics of Art Museums, University of Chicago, 1991, p60)

5.70 The anecdotal evidence also spoke of `buyer resistance' when charges reach a certain high point. [40] These points together suggest a typical demand curve in which demand falls off relatively steeply at the change from zero to low prices, more slowly through the middle range of prices, then more steeply again at some higher price as the fewer remaining visitors reach the limit of their willingness to pay.

5.71 Making allowance for the uncertainties, on balance there seems to be strong evidence that on the whole entry fees in Australian museums and galleries (probably even `modest' fees) do significantly discourage visitors. [41]

5.72 The Committee acknowledges that all this discussion of visitor numbers tends to neglect the point that the effectiveness of a museum or gallery should be gauged not only by the number of visitors but also by the quality of their experience.

`...increased focus on visitor needs... is reflect in an increased focus on `customer service'. The effects of this have been obscured by a lack of understanding of what education is, and a tendency to equate `value' with the number of visitors to a museum, which is invalid.' (Australian Museum, submission 30 p240)

5.73 The implication is that a less visited, better quality museum experience may be just as valuable to society as a more mass-marketed experience. The Committee agrees. But witnesses did not suggest that the loss of visitors reported above has actually been compensated by improved experience of the remainder. Given the common statement that user charges are forced upon reluctant managers by cuts to budget funding, the Committee considers that outcome most unlikely; so the problem remains.

`Price signals promote efficient allocation of resources'

Allocation of resources between sectors of the economy

5.74 We now mention an argument that was not much raised in submissions to this inquiry, though it is common enough in literature about subsidies to the arts and excludable public goods generally. The argument runs: user charges signal to people that the resource is not without cost, thus dampening demand. Without these `price signals' at the point of use, demand might be excessive: that is, it might lead to an oversupply of the `free' good (given that for public goods demand translates into supply not through market forces but through political decisions which hold no guarantee of reaching an economically optimal outcome). `Oversupply' is here defined in terms of economic efficiency: it has no moral connotations to do with overindulgence, but simply means `supplied beyond the point where marginal costs [which the user does not have to face] outrun marginal benefits, suggesting that the resources used could be reallocated to more benefit elsewhere in the economy' (and vice versa for `undersupply'). This is line with the marginal cost pricing rule - see paragraph 5.18.

5.75 The argument has analogies in other areas - for example, debate over whether medical services free at the point of use encourage overservicing. The Committee comments:

5.76 1. Firstly, if levying the charge changes the character of the good, a comparison of economic efficiency before and after is invalid because it is not comparing like with like. We have suggested in chapter 4 that charging entry fees for museums, galleries and national parks does indeed change their character from community facilities which we visit as citizens to businesses which we visit as customers. The change is a significant one.

5.77 2. Secondly, in proportion as the main benefits of the public good are community/ non-use benefits (cultural enrichment, maintaining the place for future generations, `intrinsic value' of nature conservation etc.), the `price signals' argument has less application, since signals obtained about willingness to pay for direct use say nothing about the community/ non-use benefits. In the limiting case, if we could imagine a public good that conveys all community/ non-use benefits and no private benefits in use, an attempt to charge for direct use would reveal a willingness to pay of zero; [42] but this would not mean that the good should be withdrawn. Indeed it would have no bearing at all on the question of whether the community/non-use benefits are worth the public (taxation/ subsidy) costs: this is a separate question which must be answered in other ways, whether by political consensus or by survey techniques such as the contingent valuation method mentioned in paragraph 5.11. The evidence suggests that the community does maintain museums, galleries and national parks primarily for the sake of community/non-use benefits. [43]

5.78 3. Thirdly, the `price signals' argument does not support any charge greater than that already supported by the marginal cost pricing rule. We have argued from paragraph 5.22 that for museums and galleries the relevant marginal cost should be regarded as the cost of admitting another visitor during the politically approved opening hours - that is, practically zero. Similarly in principle in national parks, though in that case marginal costs associated with increments of visitation are more obvious. Where the marginal cost is zero, in terms of the `price signals' argument any amount of demand for direct use can and should be accommodated freely.

5.79 4. Fourthly, there are countervailing theoretical considerations. The collective political choice about how much of a public good to supply involves not only the opinions of direct users, who might value it a lot, but also the opinions of non-users, who might value it little. If the non-users are more numerous they are likely to prevail, however great the value of the good is to users, causing `undersupply' (that is, supply below the level which maximises total community welfare). Technically:

`...the preferences of the median voter will dominate the outcome, under certain well-known assumptions (Black 1948)... the mean [average] preference is an economic efficiency benchmark... whilst the median preference is a political benchmark... [If the median preference is less than the average preference] the difference could be seen as an empirical measure of `collective failure' in provision of public goods [ie provision below the economically efficient level]' [44]

5.80 The economic jargon simply elaborates the commonsense proposition that governments who perceive political advantage from keeping the lid on taxes are not easily talked into setting up new public institutions which will require ongoing subsidy.

5.81 A contrary hypothesis is that powerful élites (such as cultural élites or environmental groups) have influence out of proportion to their numbers. But in the case of the arts generally such research evidence as there is, though admittedly patchy, suggests that in fact the median voter does prevail and subsidised arts are indeed undersupplied. Throsby and Withers' 1982 survey found an average willingness to pay for the arts through taxes between two and three times as much as government support for the arts. [45]

`In previous work one of the present authors (Withers 1979) has shown that the use of median rather than mean measurement... provides a better explanation of variations in state and local arts assistance levels in Australia. In the case of the present [1982 Australian] survey, the median preferences obtained are substantially smaller than the mean levels reported... This result suggests a wide disparity between economically efficient (mean-based) levels of provision of the public good and democratically determined (median-based) levels of provision. The bias in this instance is toward underprovision.' [46]

5.82 Throsby, Withers and Johnston's 1994 survey extended to the full gamut of public expenditure, allowing respondents to suggest tradeoffs between different areas of expenditure. Respondents' average willingness to pay for `arts and public broadcasting' through taxes was 7 per cent higher than their actual tax liability. [47] More details of this and some other research from around the world is in APPENDIX 6. None suggests an oversupply of subsidised arts.

5.83 For national parks the Committee has no comparable information (that is, setting surveyed willingness to pay taxes against actual public subsidy). But we note that in Throsby, Withers & Johnston's study just mentioned, average willingness to pay for `environment' was over twice as great as actual government expenditure - the highest disparity among all the 16 categories of public expenditure surveyed. [48] See APPENDIX 6.

5.84 On balance, the Committee thinks it unlikely that museums, art galleries and national parks are oversupplied. It should also be noted that the value of national parks and potential park land is likely to increase over time, since suitable land is a finite and increasingly scarce resource under pressure of population growth, while higher standards of living allow a continuing shift of community preferences to outdoor recreational activities and environmental goods. [49]

5.85 Finally, the Committee comments that if `price signals' are still thought desirable to counteract pressure to oversupply, they are not without cost - the cost being the deadweight loss of welfare of discouraged visitors. The benefit must be weighed against the cost after assessing how great is the risk of oversupply.

Allocation of resources within the institution

5.86 A related argument is that being forced to rely on a proportion of cost recovery from user charges will focus the minds of managers on improved `customer focus':

`ABARE [Charging Users of the Great Barrier Reef Marine Park: report to Great Barrier Reef Marine Park Authority, 1991] has stated that, if users are charged for entry into national parks, they will develop an interest in the way the park spends its money. Seeking value for money, this consumer interest will drive the park authorities to supply more cost effective services than previously.' (Treasury, submission 20 p174)

5.87 This has two aspects:

5.88 1. Pressure to improve productivity: that is, to do the same things with less money. The inquiry took no evidence on whether cost recovery targets have in fact caused or are likely to cause improved productivity in museums, galleries and national parks. The Committee comments generally:
a) Given that museums, galleries and national parks have public service missions the outputs of which may be hard to quantify or hard to allocate among fixed costs (`education'; `cultural enrichment'; `nature conservation'), measuring productivity improvements may be difficult. In particular, the operation's effectiveness concerns not only the number of visitors per costs, but also the quality of their experience.
b) Pressure to improve productivity has no necessary connection with user charges as such: it arises from the imbalance between total income and total costs. To create pressure to improve productivity user charges are neither necessary (governments can squeeze budget funding whenever they like anyway) nor sufficient if, demand being inelastic, the fee income simply allows managers to offload the `deficit' problem onto visitors without changing their own behaviour.

5.89 2. Pressure to do things differently in order to attract visitors or provide more services for visitors. The Committee comments: this pressure is not necessarily an unalloyed good: it may conflict with fundamental conservation or education charters. We take this point up in chapter 7 (from paragraph 7.63).

Comparison of museums/galleries and subsidised performing arts

5.90 We digress to mention a related argument which may seem to support the view that people should pay for the private benefits of a museum/gallery visit. This argument relies on the analogy with performing arts: since user charges for subsidised performing arts are non-controversial, why shouldn't museums and galleries be treated the same way?

5.91 However, there are significant differences:

5.92 1. Performing arts are more often rival (congested, sold out) than museum entry is. Price is then a way of rationing, and the `price signals' argument has application. [50] The permanent exhibits of Australian museums and galleries are rarely congested. [51]

5.93 2. In the case of museums and galleries, with their high fixed costs of maintaining collections, it is more strongly arguable that the marginal cost pricing rule suggests an efficient entry fee of practically zero (see paragraph 5.18). In performing arts the marginal costs of the marginal performance are much more obvious.

5.94 3. Arguably satisfying option demand (that is, maintaining something for possible future use) is a relatively more important function of museum subsidies than of performing arts subsidies. Performing arts companies may come and go and come again, depending on current demand; a museum collection, once lost, is lost forever. [52] Satisfying option demand is a pure public good. [53]

5.95 4. In the case of performing arts the existence of similar private goods (that is, unsubsidised `popular' culture) is more obvious. In the case of museums and galleries similar unsubsidised private operations are rare. [54]

5.96 5. Percentage cost recovery from user charges is typically much greater in subsidised performing arts than in museums and galleries. [55]

5.97 6. Perceptions of the `merit good' element may be different in performing arts (more entertainment?) from museums and galleries (more education?).

5.98 The last three points are related by history: subsidised performing arts have evolved from similar private goods, as governments have taken over the role that wealthy patrons filled in past centuries; while museums and galleries by and large have from the start been established communally for public benefit reasons to do with education and social improvement. Hence `user pays' arguments are typically framed oppositely in the two cases: the argument on performing arts being whether and how much governments should subsidise them; the argument on museums and galleries being whether and how much users should pay. Most of the points above suggest free entry for museums and galleries more strongly than for performing arts, and this conclusion is supported by the fact that in museums and galleries free entry is common and entry fees controversial, while in performing arts free entry is uncommon and entry fees uncontroversial. At the very least, we conclude that a user charges policy for performing arts is not an appropriate template for a policy on museums and galleries.

`Entry fees may be better than reduced services'

5.99 Notwithstanding all the theory of the deadweight loss, it can be argued that entry fees, though undesirable, may be less undesirable than reduced service:

`As long as an extra visitor to the museum imposes no additional costs on the museum or on other visitors, the ideal admission policy is to have no charge at all... [however] The case against charging admission fees and publication royalties is fully persuasive only when there is enough support from private contributions and government funds to maintain the appropriate level of museum activities. When these other sources of funds are inadequate, higher admission fees may be better than lower levels of spending.' (Feldstein M ed., The Economics of Art Museums, University of Chicago, 1991, p4-5)

5.100 This is the counterpart of the argument that `user charges revenue may allow improved service'. We consider that argument in chapter 7 (from paragraph 7.36) in its more political dimension (the basic question being, will governments actually allow site managers to keep the money?); we consider the economic efficiency dimension here.

5.101 The argument acknowledges that entry fees (marginal cost of provision being zero) cause a deadweight loss of welfare; nevertheless `higher admission fees may be better than lower levels of spending' if the revenue can be used to create benefits that outweigh the deadweight loss (for example, by improving the quality of experience of the remaining visitors).

5.102 The Committee comments:

5.103 1. Although fee may be better than no fee, if it forestalls the need for a detrimental truncation of service, it is still a second-best position. The deadweight loss is still a loss. The `appropriate' level of public funding is still best from the point of view of total welfare. [56] The question is whether society accepts the deadweight loss as the price of refusing to subsidise visitors' private benefits, or whether society is willing to tolerate the subsidy to private benefits as the price of maintaining wide access and maximising total welfare.

5.104 2. In practice evaluating higher levels of activity with fee versus lower levels of activity without fee with economic rigour will be problematic. The costs of a fee in diminishing the museum's ability to attract volunteerism, sponsorship, public esteem and public subsidy in the longer term must be factored in.

5.105 3. We should not overstate the importance of such strictly economic calculations. Less quantifiable political considerations (such as the emotional and symbolic value of free public institutions discussed in chapter 4, versus the user pays philosophy of `charge for private benefits') are probably more important in practice.

Conclusions on economic efficiency arguments

5.106 If demand is inelastic, entry fees do not change anyone's behaviour: they simply appropriate part of visitors' consumer surplus for the public treasury. They do not change total community benefits created by the museum/ gallery/ national park, and accordingly they cannot be faulted on grounds of economic efficiency. They may in theory improve economic efficiency (through the argument that price signals abate excessive demand), but in the case of museums, galleries and national parks the Committee does not think this is a significant consideration in practice. Of course entry fees may still be objectionable for reasons discussed in other chapters.

5.107 If demand is significantly elastic, entry fees discourage use, causing a deadweight loss of welfare to the discouraged visitors, who are barred from the benefits of use without any compensating saving of public running costs. This assumes that the marginal costs associated with a visit are low. In the case of museums, galleries and (to a lesser extent) national parks the Committee thinks this a reasonable assumption.

5.108 The appropriate response to these considerations depends on the facts about demand on a case by case basis. In particular:

5.109 1. It is not appropriate simply to apply a general philosophy that `users should pay for private benefits' without regard to the circumstances of the case. This position is based on considerations of equity between users and non-users. It relies on analogy with the benefits traded in private markets; but the analogy is not apt. These are public goods, and the benefits may be greatest when fully publicly funded. The cost of implementing `user pays', in terms of the lost welfare of discouraged visitors, must be considered.

5.110 2. A rational charge (one that is a politically satisfactory compromise between revenue raising and the deadweight loss of welfare) must be decided with reference to the distribution of willingness to pay, case by case. It is not rational to aim for a ratio of cost recovery modelled on an estimate of the ratio of total user benefits to total non-use benefits, even within the individual place; even less within the museum/ national park estate as a whole.

5.111 3. Where demand is significantly elastic, marginal costs being low, entry should be free or fees held at the necessary low level so as not to depress demand. In the latter case the question arises whether the fee is worth charging at all, in light of the damage that it may do to public esteem for the place as a community facility.

5.112 In the case of national parks elasticity of demand may vary greatly from one to another: a national park with rare attractions and strong demand may support high fees, and vice versa.

5.113 In the case of museums and galleries the Committee's evidence was that entry fees commonly cause a significant drop in visitation - a deadweight loss of welfare which the Committee regards as regrettable. This loss of welfare has no particular compensations except to satisfy the philosophical position that direct users should pay for private benefits - a position which is not necessarily sound in the case of public goods. It would be better for the community to accept the subsidy to private benefits of direct users for the sake of maintaining wide access to these public facilities.

5.114 The Committee acknowledges that most of the evidence on these matters was anecdotal. In particular, predictive studies of willingness to pay (such as the studies of particular national parks mentioned in APPENDIX 6) are not matched by retrospective studies showing the actual effects on visitation of charges actually implemented. There is a case for more wide-ranging research of a statistically sound sample of places, which would plot the history of user charges against the history of visitation to see what cause-effect relationship appears. Of course other factors that make visitation fluctuate, such as those mentioned in paragraph 5.61, would have to be accounted for, which greatly complicates the task. [57]

RECOMMENDATION 3

5.115 The Committee recommends that the Department of Communications and the Arts, in consultation with State/Territory authorities and relevant peak bodies, should sponsor research on the relationship between user charges and visitation in cultural institutions.

5.116 The Committee recommends that Environment Australia, in consultation with State/Territory authorities and relevant peak bodies, should sponsor research on the relationship between user charges and visitation in national parks.

5.117 Further conclusions amalgamating efficiency and equity considerations are in chapter 6 (from paragraph 6.37).

 

Footnotes

[1] The change is called a `potential Pareto improvement' after economist Vilfredo Pareto (1848-1923). Note that the simple `Pareto criterion' treats `the community' as simply the sum of individuals acting in rational self-interest: it disregards things like envy and altruism, and it does not admit the existence of the community as a body corporate, a whole which may be different from the sum of the parts. It explicitly disregards concerns about distributional justice: a highly unequal distribution of income may be `Pareto superior' to a more equal distribution. How this conceptual scheme should be finessed to make it more relevant to the real world is a matter of debate among welfare economists.

[2] Fees as a way of rationing use of fragile resources is considered from paragraph 7.107.

[3] Demand may also be influenced by variables other than price - such as opening hours in the case of museums. All the references here are to elasticity depending on price.

[4] Where the starting price is zero - as with free-entry museums or national parks - consumer surplus equals willingness to pay; as a result discussions of user charges, assuming free entry as the starting point, sometimes loosely use the two terms interchangeably.

[5] The assumption is conventional in economic analysis but possibly arguable on non-economic grounds. See Appendix 6, introduction.

[6] This assumes that the cost of providing the good is regarded as sunk, which is reasonably arguable for (as here) merit goods provided by community decision primarily for the sake of community/non-use benefits. The main point here is that, marginal costs being zero for existing non-rival goods, increments of use create increments of `total community welfare'.

[7] Willingness to pay for direct use of excludable public goods through user charges can of course be tested by experiment, as with market goods. We may still want an estimate by other means because - 1. taking a charge from nothing to something is less a marginal change than a quantum leap, which requires political commitment and may have unpredictable results; 2. the politically acceptable limit to a charge may be less than full willingness to pay by an amount which it would be interesting to know; 3. these goods (particular national parks and museums) are to a large extent one of a kind in regional markets, so analogies with past experience elsewhere may be unreliable; 4. the `merit good' aspect means that we do not wish to discourage visitation by accidentally charging too much.

[8] One reason for the predominance of natural areas in travel cost studies is that the method is less reliable in urban areas because it is harder to obtain the spread of different travel costs from different points of origin that is necessary to estimate the demand curve.

[9] Driml S M, Sustainable Tourism in Protected Areas?: an ecological economics case study of the Wet Tropics World Heritage Area, PhD thesis, Australian National University, Canberra 1996, p198ff, 229ff

[10] For example, see Driml S M, Sustainable Tourism in Protected Areas?: an ecological economics case study of the Wet Tropics World Heritage Area, PhD thesis, Australian National University, Canberra 1996, p193,224ff. This used a travel cost survey of Australian tourists and a questionnaire for overseas tourists.

[11] Strategic bias may mean understating willingness to pay, if respondents fear that their response may influence a decision to increase charges, but do not think that their response will affect provision of the good; or overstating willingness to pay in the opposite conditions.

[12] In Driml's study, among the overseas tourists who objected to paying for entry, reasons of principle (such as `access to nature should be free for everyone') were much more prominent than reasons do with expense or budget. Driml S M, Sustainable Tourism in Protected Areas?: an ecological economics case study of the Wet Tropics World Heritage Area, PhD thesis, Australian National University, Canberra 1996, p195. See paragraph 5.53 footnote 36 for evidence on attitudes to entry fees in art galleries.

[13] Hamilton-Smith E., `Challenging the Willingness To Pay model,' Valuing Natural Areas: applications and problems of the contingent valuation method, ed. Lockwood M. & De Lacy T., conference papers Charles Sturt University June 1992, p149

[14] Some other obvious possibilities are people may regard visible definite expenses (such as entry fees) more negatively than unnoticed or uncertain expenses (such as wear and tear on the car) people may regard many small expenses more critically than a few larger expenses (the person who settles on the $100 motel without a second thought may greatly resent paying $2 for a bad cup of coffee).

[15] For the moment we disregard considerations such as whether cost recovery targets encourage management to be more productive. See paragraph 5.86.

[16] If the decision to shift part of the cost burden to direct users is motivated by equity/distributional concerns (such as `unfair for poorer nongoers to subsidise richer goers' - see paragraph 6.1), additional questions are suggested, such as How much equity (redistribution) is worth how much inefficiency (loss of total community welfare)? Will the change in fact achieve equity goals, or be swamped by side-effects? (see paragraph 6.37). It raises the alternative possibility of selective taxes (for example, a levy on local property owners to finance local civic improvements). `Since museums provide a largely local service, using general local tax revenues to cover their deficits is consistent with the intent of the two-part tariff: those paying the taxes are at least potential users of the subsidized service.' Heilbrun J & Gray C M, The Economics of Art & Culture: an American Perspective, Cambridge University Press 1993, p214

[17] A marginal (variable, avoidable) cost is the extra cost associated with an extra unit of production. If it costs $10 to produce 10 widgets, and $15 to produced 11 widgets (perhaps because there is some economy in producing them in batches of 10), the average cost of producing the first 10 is $1 each, and the marginal cost of producing the 11th is $5. To take a non-rival public good: a lighthouse produces `lighthouse services'; when a new ship comes into view an extra unit of service is at once produced and consumed, at a marginal cost of zero.

[18] A theoretical alternative is `perfectly discriminatory pricing', in which each person's willingness to pay is somehow discovered, and each person is charged exactly that much. In this case the whole consumer surplus is appropriated for the public treasury, but again no-one drops out so there is no deadweight loss. Of course this is usually impossible in practice.

[19] In the case of non-excludable (or not actually excluded) goods the principle has no application: the possibility of charging does not arise. In the case of non-rival goods, its application is trivial: the marginal cost is zero and the price should be zero, as is obvious intuitively anyway from the above discussion of the deadweight loss. The principle is most relevant to cases such as public utilities where the service is provided by using a large non-rival infrastructure to produce units of output which also have a marginal cost. The units of output are rival and excludable, but are still produced as public goods because of their dependence on the non-rival infrastructure.

[20] This is the rationale for `two-part tariffs' for public utilities such as water supply - a flat fee for connection to the system to cover the high fixed costs, taking advantage of the fact that the demand for connection is very inelastic, and marginal cost pricing for marginal units of water, for which the demand is more elastic.

[21] When the bridge becomes congested it is no longer non-rival: admitting an extra user imposes a congestion cost on all users. In Australian museums, galleries and national parks congestion is rarely an issue, and for simplicity we disregard it from now on.

[22] Tasmanian Parks & Wildlife Service, User Regulation in the Tasmanian Wilderness World Heritage Area, 1996. The paper proposes a permit system for overnight walkers in the World Heritage Area. The matter is still under consideration.

[23] I Burston (Great Barrier Reef Marine Park Authority), evidence 20 May 1997 p44

[24] Tourism Council Australia, submission 57 p598

[25] The longer the period, the greater the opportunity to adjust inputs so as to produce the marginal widget more efficiently. See APPENDIX 3, `marginal cost'.

[26] We disregard subtleties such as whether management might wish to maintain a certain suite of services, perhaps with cross-subsidy, for the sake of its overall public image or market niche.

[27] It should not be forgotten that museums have functions other than admitting visitors, such as research. The principle can be applied to them, so that the museum's various costs are charged out to its various beneficiaries according to judgments about which services are core and value-added. In practice, assigning the fixed costs of the museum among its various activities may be problematic.

[28] The extreme view has little to recommend it even theoretically. In the world of economics textbooks, if the weakest firm in a market fails its customers may transfer to alternative suppliers of similar goods, who are strengthened thereby, the process repeating itself as necessary until an equilibrium is reached. By contrast, each museum/gallery is to a large extent one of a kind, either by virtue of its distinctive collection, or in a regional market (since, by contrast with the widgets produced by the textbook private firm, `consumption' must usually take place at the site of production: the people must go to the museum). If a museum closes its customers may not have `alternative suppliers' and other museums may not be strengthened. Thus if each museum was treated as a marginal element in the state's total cultural expenditure, and required to pay its full costs from user charges, probably all would close. In other words, to regard `the state's cultural expenditure', or even `the state's museum expenditure', as a homogenous mass of service which can be marginally truncated at will is wrong. Similarly for national parks.

[29] `Core' and `value-added' services are discussed further in chapter 7, where we conclude that what is here called `dispensible' should be equated to `value-added'. See paragraph 7.20.

[30] The marginal social cost includes external costs of the activity: that is, costs suffered by persons other than the user/beneficiary - such as noise and pollution in the case of road traffic. This is the main reason for proposing road use charges. We disregard this here for simplicity.

[31] Submissions that `users should be charged for private benefits' rarely gave clear suggestions about how much they should be charged: most stopped with the statement that charges should be `modest' so as not to discourage use. They did not argue explicitly that a cost recovery target should be modelled on the ratio of use benefits to non-use benefits. The idea is an intuitively appealing one and may sometimes have been implicit (though it conflicts with `modest charges so as not to discourage use'). It is present, for example, in the Department of Finance's 1989 paper What Price Heritage: `...questions can be raised at the margin as to whether the current balance between public and private financing accurately reflects the balance between public and private benefits...' The proposition, though intuitively appealing, is wrong. A optimal charge (one that represents a politically satisfactory balance between revenue raising and the deadweight loss of welfare) depends entirely on the distribution of willingness to pay, case by case: it has no necessary connection to the ratio of total use benefits to total non-use benefits, not within a single place, even less averaged over some larger sector. See paragraph 5.32: `Charging for private benefits: a hypothetical case'. Department of Finance, What Price Heritage: the museums review and the measurement of museum performance, Canberra 1989, p29.

[32] Driml S M, Sustainable Tourism in Protected Areas?: an ecological economics case study of the Wet Tropics World Heritage Area, PhD thesis, Australian National University, Canberra 1996, p257

[33] Worboys G & Gillespie R, Improving Revenue Return to the NSW NPWS, September 1996, attachment 3. [NB the main report is not publicly available]

[34] Hall C M & McArthur S, Heritage Management in Australia and New Zealand: the human dimension, OUP, Melbourne 1996, p43, quoting Wesley R, `An historic site as a tourist attraction - case study Port Arthur', Heritage Managment, Parks, Heritage & Tourism, conference proceedings, Royal Australian Institute of Parks & Recreation, Hobart 1992, p41-4

[35] G Marginson (National Campaign for the Arts Australia Ltd), evidence 6 August 1997 p298

[36] Bennett T & Frow J, Art Galleries: Who Goes?: a study of visitors to three Australian art galleries, with international comparisons, Australia Council 1991, p15. `The view that admission should be free, however, was relatively constant across all the demographic variables.' This tends to support the proposition that reasons for objecting to entry fees are more to do with principled attitudes than affordability: otherwise one would expect poorer visitors to object to fees more vigorously than richer visitors. See paragraph 5.13 footnote 12 for similar evidence in national parks.

[37] In such equity arguments the relationship between frequent visits by a small section of society and occasional visits by a larger section may be significant. Note that the statement `museum visitors are disproportionately drawn from the educated classes' does not necessarily mean that most museum visits are made by such people. Depending on the proportions, frequent visits by a few people may be outweighed by occasional visits by the many.

[38] Queensland Museum, submission 31 p249

[39] See also paragraph 6.30.

[40] For example, G Wellard (Queensland Dept of Environment), evidence 21 May 1997 p98,100

[41] We say `Australian' advisedly. Overseas evidence is not necessary analogous. The great museums of Europe, for example, may support high charges: in their character as monopoly suppliers of rare and highly prized experiences they are more akin to the great national parks of Australia than to Australian museums.

[42] We disregard altruism.

[43] The most obvious evidence of this is that the community does actually consent to maintain them primarily by public subsidy. See Throsby C D & Withers G A, The Economics of the Performing Arts, Melbourne 1979 p192-3, quoted in paragraph 6.16 below; and the studies reported in APPENDIX 6.

[44] Throsby C D & Withers G A, `Bias and demand for public goods', Journal of Public Economics 31, 1986, p312; Black D, `On the rationale of group decision-making', Journal of Political Economy 56, 1948, pp23-34

[45] Throsby D & Withers G, What Price Culture, Australia Council 1984, p18. Its topic was arts funded by the Australia Council. We assume a relevant analogy to museums and galleries, but there are some obvious differences - see paragraph 5.90.

[46] Throsby C D & Withers G A, `Bias and demand for public goods', Journal of Public Economics 31, 1986, p321. Withers G, `Private demand for public subsidies: an econometric study of cultural support in Australia,' Journal of Cultural Economics 3, 1979, p53-61

[47] Withers G, Throsby D & Johnston K, Public Expenditure in Australia, Economic Planning Advisory Commission, commission paper no. 3, 1994, p33

[48] Withers G, Throsby D & Johnston K, Public Expenditure in Australia, Economic Planning Advisory Commission, commission paper no. 3, 1994, p33. Note also that `The Resource Assessment Commission inquiry (RAC 1991) into the forest and timber industry undertook both a travel cost study and contingent valuation study of forest of south eastern Australia. The results of the studies indicated `that the willingness to pay or consumers' surplus per person per year for the preservation values were approximately three times the willingness to pay per person per year for recreation values... The RAC identified that this is a common outcome when the two methods are applied simultaneously.' Gillespie R, Economic Value and Regional Economic Impact: Minnamurra Rainforest Centre, Budderoo National Park, NSW National Parks & Wildlife Service, 1997, p27, quoting Bennett J et al., The Economic Value and Regional Economic Impact of National Parks, Ecological Economics Conference papers, the Australian and New Zealand Society for Ecological Economics, in association with the Center for Agricultural and Resource Economics, 1995, p133. Resource Assessment Commission, Forest and Timber Inquiry, Canberra 1991.

[49] Gillespie R, Economic Value and Regional Economic Impact: Minnamurra Rainforest Centre, Budderoo National Park, NSW National Parks & Wildlife Service, 1997, p27. Visits to Victorian national parks and other parks, for example, have grown from 3.4 million per year to 10.2 million per year from 1977 to 1993. Hall C M & McArthur S, Heritage Management in Australia and New Zealand: the human dimension, OUP, Melbourne 1996, p133

[50] `If there were no prices charged... there would be mile-long queues from Covent Garden to the river. In such circumstances it seem more sensible to ration by price rather than by capacity to sleep out on the pavements so as to be at the head of the queue... But the position of the national museums and galleries is radically different... the correct analogy is not with the public subsidised opera or theatre but rather with the parks and libraries...' Robbins L, `Unsettled questions in the political economy of the arts', in Blaug M ed., The Economics of the Arts, London 1976, p178

[51] This also means that the high charges successfully levied by many congested great museums of Europe, with their massive tourist trades, have little relevance as a model for Australia.

[52] This argument should not be pressed too far. Cultural continuity is important, and a cultural tradition once lost may not be easily resurrected.

[53] Satisfying the option demand for any good is itself a pure public good, since conserving something to satisfy one person's option demand also satisfies other people's option demand necessarily (non-excludable) and at no extra cost (non-rival). Therefore private markets will not efficiently satisfy option demand.

[54] Note that `similar' does not mean `the same.' If a publicly supplied or subsidised good was really identical to a privately supplied good, we might well ask what makes the subsidised version a `public good'. On closer inspection it is usually seen to be not identical. For example, one difference between the subsidised `serious' theatre company and the unsubsidised cinema is simply that in the case of the theatre society perceives non-use or `merit good' benefits; in the case of the cinema it doesn't. The non-use benefits (being non-rival and non-excludable) make the theatre a public good.

[55] This does not necessarily mean that the ratio of use to non-use benefits differs correspondingly between performing arts and museums/ galleries. We have noted that an efficient user charge is found by reference to the distribution of willingness to pay for use, not by reference to some estimate of the ratio of use benefits to non-use benefits (see paragraph 5.32ff). Subsidised performing arts typically achieve higher cost recovery than museums/galleries simply because, the market being what it is, they can. This phenomenon has no particular bearing on the ratio of use benefits to non-use benefits. We could imagine some cultural activity which achieves full cost recovery through charges, yet still creates equally large non-use benefits.

[56] `The appropriate level of museum activities is easier to define conceptually than to assess in practice. The appropriate level of spending ... is the highest level of spending at which any increase in spending does not produce current and future benefits... that are as large as the additional costs.' Feldstein M ed., The Economics of Art Museums, University of Chicago, 1991, p4

[57] The Committee wrote to state museums seeking information on entry fees vis-à-vis visitation which might be the start of such a study. We acknowledge their responses with thanks, but we do not reprint the information here because to do so, without a study of the other factors that cause visitation to fluctuate, would be too misleading. Some figures on visitation to larger museums from 1990, drawn from Australia Council research papers, are in APPENDIX 5.