SEARCH

SEARCH BY CITATION

Abstract

  1. Top of page
  2. Abstract
  3. Introduction
  4. Casino Catchments and the Distribution of Gambling Impacts
  5. Estimating the Extent of Casinos Social and Economic Impacts
  6. Study Area Description
  7. Methods
  8. Results
  9. Discussion
  10. References

Casinos have proliferated throughout Australia and in many other parts of the world since the late twentieth century. An emerging body of research has started to explicitly consider the social and economic impacts of casinos in different settings. Many of the potential impacts of casinos are spatially patterned and relate to the connectivity of patrons and venues. In this paper, we use a predictive trade-area analysis technique, the Huff model, to estimate the spatial extent of casino catchments in Australia and compare these outputs to travel data from National Visitors Survey. Many casinos draw patrons from regional areas and from other states, a set of cross-jurisdictional patterns that pose regulatory challenges in terms of managing economic benefits and the distribution of harms arising from casino gambling. Avenues for logical extensions of the approach are discussed as well as alternative methods of sourcing data for validating predictive outputs.


Introduction

  1. Top of page
  2. Abstract
  3. Introduction
  4. Casino Catchments and the Distribution of Gambling Impacts
  5. Estimating the Extent of Casinos Social and Economic Impacts
  6. Study Area Description
  7. Methods
  8. Results
  9. Discussion
  10. References

Casinos have proliferated throughout Australia and in many other parts of the world over the last 20 years as part of the global liberalisation of the gambling industries. Casino development in Australia took place in two waves (McMillen 1990). Four small casino-hotels were built in Tasmania and Northern Territory between 1973 and 1984, followed by the establishment of larger urban casinos in every Australian state capital by 1996 (Lynch and Veal 1998). A similar expansion at a much larger scale occurred in the U.S., where the number of states authorising casino gambling rose from 2 to 38 in a little over two decades (American Gaming Association 2012; Eadington 1998a). Paralleling the expansion of casinos in Australia was the diffusion of electronic gaming machines (EGMs) into “community” venues outside the casino system, such as hotels and returned servicepersons and sporting clubs (Marshall 2007). While the liberalisation of gambling has arguably resulted in economic development (Eadington 1999; Williams, Rehm, and Stevens 2011; but see Walker and Jackson 2007), it has also been associated with a suite of negative social impacts including pathological gambling and criminality (Productivity Commission 1999; Williams, Rehm, and Stevens 2011).

All casino impacts, positive or negative, are spatially patterned. For example, casinos that target tourist markets may experience increased local benefits and diminished local harms compared with those that rely predominantly on a local market (Calcagno, Walker, and Jackson 2010; Dombrink and Thompson 1990). In particular, if the points of origin of casino visitors are beyond the boundaries of the host jurisdiction, many negative externalities associated with casino gambling such as taxes, crime or problem gambling are exported. Large, tourist-centric casino-resorts such as Las Vegas and Macau are archetypical of this spatially patterned export of harms, although American Indian casinos may also fit this category insofar as they are not frequented by their tribal owners (Cornell 2008).

Consequently, estimation of the spatial distribution of casino patrons may be useful for local casino regulation. Policy makers may wish to treat gambling venues that cater to local markets quite differently to those serving international tourists (e.g., The Australian Institute for Gambling Research 2000). However, despite the widespread practice among casino managers of collecting detailed data about their patrons' demographics and gambling behaviour (Watson and Kale 2003), little is publically known about the spatio-demographics of casino patrons. This is problematic as a knowledge of the residential locations of casino patrons at the regional and national levels is important to assessing the distribution of casino-related economic and social impacts (Grinols and Omorov 1996). Unfortunately, while the local economic and social impacts of casino development are routinely assessed, few studies have specifically investigated the spatial extent of these impacts (Williams, Rehm, and Stevens 2011). To this end, the current paper demonstrates the application of a normative spatial model for the characterisation of casino trading areas in Australia. The projected catchments are compared with reported casino visitation in a national domestic tourism survey. The results are discussed in the context of regulatory implications and extensions of the modelling approach.

Casino Catchments and the Distribution of Gambling Impacts

  1. Top of page
  2. Abstract
  3. Introduction
  4. Casino Catchments and the Distribution of Gambling Impacts
  5. Estimating the Extent of Casinos Social and Economic Impacts
  6. Study Area Description
  7. Methods
  8. Results
  9. Discussion
  10. References

While many of the benefits of casino gambling accrue to the location and jurisdiction in which the casino is located, social harms may occur elsewhere. Excepting the consumer surplus, the positive impacts of casino development can all reasonably be expected to be greatest in the areas proximate to a casino development. However, many of the harmful impacts of casino gambling, including taxation and gambling-related harm, are in large part spatially coincident with the residential location of the gamblers themselves (Eadington 1986). Therefore, when gamblers travel to a casino venue, a spatial disjunction between casinos' costs and benefits may occur. From a community and regulatory perspective, this potential disjunction is intensified when gamblers cross the jurisdictional borders where the collection of gambling tax revenues and the responsibility for dealing with gambling harm are located.

The cross-border distribution of consumption and its attendant impacts are not, of course, confined to casino gambling. Indeed, numerous studies have found that for gasoline, lotteries, tobacco and alcohol, cross-border shopping is driven by differentials in availability and prices between jurisdictions (for recent reviews, see Leal, López-Laborda, and Rodrigo 2010) in addition to touristic motivations (Timothy and Butler 1995). A vast body of research has demonstrated that cross-border purchasing results in substantial displaced taxation revenue (e.g., Asplund, Friberg, and Wilander 2007; Beatty, Larsen, and Sommervoll 2009; Chiou and Muehlegger 2008; Garrett and Marsh 2002), although fewer studies have examined the export of consumption-related harms from the shopping destination (e.g., Beatty, Larsen, and Sommervoll 2009; Hanewinkel and Isensee 2007; Hyland et al. 2005; Svensson 2009). In these studies, the spatial extent of cross-border shopping is largely determined by the price differentials between jurisdictions relative to the cost of transport (Leal, López-Laborda, and Rodrigo 2010). What differentiates casino gambling from other cross-border trade are its explicit ties to tourism and the locational constraints placed on casino development. In consequence, the geography of casino demand does not necessarily match that of supply. For many consumers, the most attractive or closest casino may be located in another jurisdiction, increasing the potential for cross-border consumption even without price differentials.

The potential for the cross-border distribution of the costs and benefits associated with casinos has not escaped notice, least of all by state governments. A calculus of jurisdictional competition, whereby casinos are developed in part to attract interstate patrons, prompting neighbouring jurisdictions to develop their own casinos to keep gambling revenue at home (Felsenstein and Freeman 2001), has been an important factor in the spread of casinos over the last three decades (Calcagno, Walker, and Jackson 2010). In the view of one state governor supporting casino liberalisation, “the most important thing is to be able to stop losing billions of dollars to surrounding states” (quoted in Donahue 1997: 76–77). This may result in an increased level of casino proliferation at a speed greater than would be the case in the absence of redistributionary pressures (Donahue 1997; but see Wenz 2008 who finds little evidence for this at the county level).

Projected casino catchments are therefore crucial to development and planning. If it is the local authority's objective to maximise local benefits while minimising local harms, then they will promote casinos that cater to an interstate or tourist market, rather than those earning income from “locals.” In this context, Eadington (1998b) produced a typology of casino types by catchment size, identifying casino-resorts with international catchments at one end of the spectrum, urban casinos with town-level catchments in the middle of the spectrum and dispersed gaming devices (i.e., EGMs) with presumably highly local catchments at the other. This typology was later extended to include riverboat casinos and Indian casinos, which have regional catchments (Eadington 1998a). This typology is pertinent to policy makers who may wish to tailor their regulations to encourage venues with large catchments and thereby receive positive casino impacts while exporting casino-related harms.

Knowledge of catchments is likely to have other policy implications. Social support provision may need to be targeted to areas experiencing disproportionate levels of casino-related social harm. To that end, the spatial location of gamblers needs to be ascertained, especially if gamblers are leaving their home jurisdiction to gamble (Hing and Breen 1996). Similarly, governments that wish to redistribute gambling-related revenue in an equitable manner (such as the Queensland Government, which operates several regional “community benefit funds”) would benefit from knowledge of casino catchment dynamics.

Estimating the Extent of Casinos Social and Economic Impacts

  1. Top of page
  2. Abstract
  3. Introduction
  4. Casino Catchments and the Distribution of Gambling Impacts
  5. Estimating the Extent of Casinos Social and Economic Impacts
  6. Study Area Description
  7. Methods
  8. Results
  9. Discussion
  10. References

Despite their policy relevance, little is known about the extent of casino catchments except that 9 percent of visits to Australian casinos in the 2006–2007 financial year were from interstate patrons (Australasian Casino Association 2008). In consequence, this paper investigates the catchments of Australia's 13 casinos. The following specific questions are posed:

  1. What is the spatial extent of the catchments for all Australian casinos?
  2. Which catchments span multiple jurisdictions?
  3. Can casinos be categorised with respect to their spatial catchment characteristics?

To this end, the Huff model, a well-established predictive trade-area tool (Ault and Johnson 1973; Greene and Pick 2012; Huff 1964; Wang 2006), was used to predict casino trading areas at the national level. The Huff model predictions were compared with available tourism survey data, the only publicly accessible source of gambling visitation data in Australia. The results are discussed in the context of regulatory implications and potential extensions of the modelling approach.

Study Area Description

  1. Top of page
  2. Abstract
  3. Introduction
  4. Casino Catchments and the Distribution of Gambling Impacts
  5. Estimating the Extent of Casinos Social and Economic Impacts
  6. Study Area Description
  7. Methods
  8. Results
  9. Discussion
  10. References

Australia's 13 casinos are all located in urban areas. Each of the eight state capitals are provisioned with a casino and five additional casinos are located in smaller towns with a significant tourism industry (see Figure 1). The casinos located in the larger cities tend to receive higher gambling revenues. While gambling revenue statistics are only available at the state level, five states contain a single casino, providing precise revenue data for those venues. Expenditure is highly concentrated within the population of casinos, with Crown Casino (VIC), Star City Casino (NSW) and Burswood Casino (WA) accounting for over 70 percent of national casino gambling expenditure (see Table 1).

figure

Figure 1. Australia's Population Distribution and Casinos. Population Distribution is Visualised Using a Kernel Density Surface with a Smoothing Bandwidth of 125 km.

Download figure to PowerPoint

Table 1. Casino Gambling Expenditure and EGMs by State
StateNumber of casinosCasino gambling expenditure ($m)Casino gambling expenditure (% of state gambling expenditure)Casino EGMs (n)Per cent of state's EGMs in casinos
Source: Productivity Commission (2010).
NSW174810%1,5002%
VIC11,21824%2,5009%
QLD458017%3,5028%
SA112911%9467%
WA153547%1,750100%
TAS211427%1,28035%
NT212224%82841%
ACT1198%00%

Methods

  1. Top of page
  2. Abstract
  3. Introduction
  4. Casino Catchments and the Distribution of Gambling Impacts
  5. Estimating the Extent of Casinos Social and Economic Impacts
  6. Study Area Description
  7. Methods
  8. Results
  9. Discussion
  10. References

Trade-area model

A gravity model approach was used to predict the spatial catchments of casinos at the national scale. Gravity modelling, also known as spatial interaction modelling, is a widely used method of estimating flows between origins such as residential zones and various destinations. While gravity models can be used to estimate numerous kinds of flows in a region (Clarke et al. 2001; Greene and Pick 2012; Roy 2004; Wang 2006), they have rarely been applied in gambling studies (Doran and Young 2010). Robitaille and Herjean (2008) used a gravity model to estimate the accessibility of video lottery terminals in Montreal, Canada, while Doran and Young (2010) applied gravity modelling techniques to investigate neighbourhood-level gambling vulnerability in Darwin, Australia. In this paper, gravity modelling methods were used to predict the spatial catchments of Australian casinos at the national scale.

Key elements of gravity models are points of origin that represent the locations of consumers and destinations (Golledge and Stimson 1997). Origin points in our gravity model were defined at the centroids of Mesh Blocks, a micro-level geographical unit at which the Australian Bureau of Statistics (ABS) releases population counts from the quinquennial census (Harper 2005). Mesh Blocks are delineated in order to capture variation in local population distribution, with an intended size of 20–50 dwellings except in non-residential areas (Harper 2005). Mesh Blocks were chosen as the origin points in this study as their relatively small extents minimise the effect of the modifiable areal unit problem. Only Mesh Blocks on the Australian mainland or in Tasmania were used in this study (n = 304,758).

The probability of residents at each Mesh Block centroid visiting each casino in Australia was calculated using the following form of the Huff model (Huff 1964):

  • display math

where Pij is the probability of residents at Mesh Block centroid i visiting casino j, Sj is the attractiveness of casino j, Dij is a measure of distance between Mesh Block centroid i and casino j, and β is a distance decay parameter. The attractiveness of various casinos was derived from estimated casino-gambling expenditure. Casino gambling expenditure data is available at the state level and was estimated in states containing multiple casinos by allocating total state expenditure to each casino in proportion to the number of EGMs they contain (see Table 2). Expenditure data is considered best means to predict the spatial distribution of known market demand (Cooper and Nakanishi 1988). Although expenditure by international visitors may bias this demand distribution, international tourists only make 4 percent of visits to Australian casinos (Australasian Casino Association 2008) and thus are not considered in this model. Network distance between each Mesh Block centroid and casino was calculated using the national road network from Geoscience Australia's 1:250,000 scale topographic maps (Geoscience Australia 2006). An additional road network link, following the path of the ferry between Tasmania and the Australian mainland, was added to allow for connectivity between the two islands, with a penalty applied to the network distance to account for the extra time and inconvenience required to catch a ferry. A distance decay parameter of −1.57 was used, taken from the parameter estimate of a calibrated Australian national-scale gravity model (Bell et al. 2002). A power function was used to approximate distance decay, as it is recommended for inter-city flows (Fotheringham and O'Kelly 1989) and consistent with other recent Australian national-scale models (e.g., Bell et al. 2002). The location of casinos was manually geocoded using Google Maps.

Table 2. Estimated Gambling Expenditure by Casino
CasinoStateEGMsProportion of state's casino EGMsState casino gambling expenditure (million AUD$)aEstimated gambling expenditure (million AUD$)
  1. Note: All EGM counts are for 30 June 2009. Expenditure data is for the 2008–2009 financial year. Casino revenue estimates may not sum to state total due to rounding.

  2. Sources: 

  3. a Productivity Commission (2010). b Data obtained from relevant state government authorities by the authors.

Star City CasinoNSW1,500a100%748748
Crown CasinoVIC2,500a100%1,2181,218
Treasury CasinoQLD1,332b37%580214
Jupiters Gold CoastQLD1,383b38%580223
Reef Hotel CasinoQLD530b15%58085
Jupiters TownsvilleQLD358b10%58058
SKYCITY AdelaideSA946a100%129129
Burswood ComplexWA1,750a100%535535
Wrest Point CasinoTAS745b58%11466
Country Club TasmaniaTAS535b42%11448
SKYCITY DarwinNT635b72%12287
Lasseters Hotel & CasinoNT251b28%12235
Casino CanberraACT0a 1919

Predicted catchments for individual casinos were mapped. Following observations from research into retail catchments (Dennis, Marsland, and Cockett 2002; Martin 1982), the spatial extent of the catchment area of a casino is defined as the area that has a greater than 10 percent probability of interaction with it.

Observed catchments

Observed catchments were derived for each casino from Australia's largest travel survey, the National Visitors Survey (NVS; Barry 1999). The NVS is a general population telephone survey of the travel patterns of Australian residents. Respondents are asked to report destinations to which they have travelled for day trips over 40 km in the last 7 days, and destinations in which they stayed overnight in the last 28 days. Respondent residential location and travel destinations are coded at the Statistical Local Area (SLA) level, a meso-scale geographical classification produced by the ABS.

In addition to eliciting information on movements, respondents are asked about activities they undertook at the travel destination. “Visiting casinos” is listed as 1 of 45 specifically coded activities that respondents are able to report. However, because respondents are not explicitly asked if they visited casinos but are instead asked “what (leisure activities) did you do during this trip?” and then probed for more specific answers, it is likely that actual rates of casino visitation by respondents are under-reported (Mick O'Halloran, Tourism Research Australia 2011, pers. comm.). The format of this question exacerbates existing biases, in which gambling in Australia is generally under-reported in general population surveys (Doughney 2009). Because the NVS does not ask respondents whether visiting a casino was the primary purpose of their visit, we were unable to untangle planned casino visitation from opportunistic visitation. Despite these shortcomings, the NVS was selected as, to the authors' knowledge, it is the only publically accessible Australian data source that provides geocoded casino visitation information.

A total of 1,344,841 NVS unit records were temporally aggregated across the period 1998–2011 in order to maximise the number of responses and analysed. Responses were aggregated to the SLA level after excluding respondents whose residential or destination SLAs were uncodable (6.7 percent, n = 90,668). For each SLA, the number of visits by residents to each town or city hosting a casino was calculated separately for day trips and overnight stays. The number of these trips in which respondents reported visiting a casino was then calculated. Day trip visits to casinos were weighted by a factor of four, in order to match the 28-day temporal period of the overnight-visit NVS question. For each origin SLA, the number of casino visiting trips was divided by the number of overall survey responses, to arrive at a figure representing the proportion of respondents to have reported visiting a particular casino in the last 28 days.

In order to reduce the substantial variance present in the data due to differing numbers of responses between SLAs, and to take advantage of spatial dependence, the proportion of respondents visiting a casino was smoothed using a local empirical Bayes rate estimator (Marshall 1991). The neighbour list calculated using SLA centroid-based k-nearest neighbours, with k = 4. Smoothing and neighbour calculations were carried out using the spdep package in R (Bivand, Pebesma, and Gómez-Rubio 2008). The resultant proportion was then mapped for each destination casino.

Results

  1. Top of page
  2. Abstract
  3. Introduction
  4. Casino Catchments and the Distribution of Gambling Impacts
  5. Estimating the Extent of Casinos Social and Economic Impacts
  6. Study Area Description
  7. Methods
  8. Results
  9. Discussion
  10. References

Trade-area model

Maps of the trading areas predicted by the Huff model showed diverse results, with casinos showing markedly different catchment characteristics, especially with respect to the number of jurisdictions spanned (see Table 3). Crown Casino and Star City Casino, for example, are predicted to have national catchments extending into every state (e.g., Figure 2). Catchments spanning moderately large areas across one or two states are more typical, with casinos in Queensland, South Australia, Western Australia and the Northern Territory predicted to exhibit this pattern (e.g., Figures 3 and 4). Only three casinos, the Wrest Point Casino, Country Club Tasmania and Canberra Casino were predicted to have highly localised catchments of less than 100,000 km2, completely confined within state borders (e.g., Figure 5).

figure

Figure 2. Crown Casino's Predicted National Catchment.

Download figure to PowerPoint

figure

Figure 3. Burswood Complex's Predicted Regional Catchment.

Download figure to PowerPoint

figure

Figure 4. Skycity Adelaide's Predicted Regional Catchment.

Download figure to PowerPoint

figure

Figure 5. Wrest Point Casino's Predicted Local Catchment.

Download figure to PowerPoint

Table 3. Predicted Catchment Classification
CasinoStatePredicted catchment typePredicted number of states spannedPredicted catchment size (km2)
Star City CasinoNSWNational86,290,000
Crown CasinoVICNational87,195,000
Treasury CasinoQLDRegional2818,000
Jupiters Gold CoastQLDRegional2770,000
Reef Hotel CasinoQLDRegional1521,000
Jupiters TownsvilleQLDRegional1322,000
SKYCITY AdelaideSARegional3546,000
Burswood Entertainment ComplexWARegional33,927,000
Wrest Point CasinoTASLocal160,000
Country Club TasmaniaTASLocal157,000
SKYCITY DarwinNTRegional2727,000
Lasseters Hotel and CasinoNTRegional1560,000
Casino CanberraACTLocal212,000

Observed catchments

Despite the very large sample size of the NVS, only 12,640 visits to casinos were reported, occurring on 2.9 percent of trips (n = 443,493). Visits to casinos were more likely to be reported for overnight trips (5.2 percent) than day trips (1.2 percent). The number of visits to casinos reported in the NVS concords closely with estimated casino revenues used in the predictive model (see Table 4 below). Crown Casino, Jupiters Gold Coast, SKYCITY Adelaide and Wrest Point Casino had a greater number of visits reported in NVS per million AUD$ estimated expenditure than the national average (3.6), while Star City Casino, the Reef Hotel Casino and Burswood Entertainment Complex had a lower than expected ratio.

Table 4. Number of Reported Visits by Casino
CasinoEstimated gambling expenditure (million AUD$)Number of reported visits in NVSNumber of visits reported in NVS per million AUD$ estimated expenditure
  1. Pearson's product-moment correlation, r = 0.92, n = 13, p < 0.001.

Star City Casino7481,5342.1
Crown Casino1,2185,0294.1
Treasury Casino2138103.8
Jupiters Gold Coast2171,2926.0
Reef Hotel Casino86851.0
Jupiters Townsville641973.1
SKYCITY Adelaide1297585.9
Burswood Complex5351,2432.3
Wrest Point Casino6686313.1
Country Club Tasmania482064.3
SKYCITY Darwin913343.7
Lasseters Hotel and Casino311585.1
Casino Canberra191316.9
All casinos3,46512,6403.6

The observed catchments reported in the NVS data follow similar patterns to those predicted by the trade-area model, but with some deviations. Crown Casino's catchment (see Figure 6) does indeed span the entire continent as predicted, but coverage is highly patchy. The Burswood Complex's catchment (see Figure 7) is not as extensive as predicted and is largely coincident with the boundaries of its home state, Western Australia. SKYCITY Adelaide's catchment (see Figure 8) is larger than predicted, extending into Victoria, New South Wales and the Northern Territory. Wrest Point Casino's catchment (see Figure 9) is very intense within regional Tasmania, but is largely confined to the island state, with very few mainland respondents reporting visiting the casino.

figure

Figure 6. Crown Casino's Observed Catchment.

Download figure to PowerPoint

figure

Figure 7. Burswood Complex's Observed Catchment.

Download figure to PowerPoint

figure

Figure 8. Skycity Adelaide's Observed Catchment.

Download figure to PowerPoint

figure

Figure 9. Wrest Point Casino's Observed Catchment.

Download figure to PowerPoint

Discussion

  1. Top of page
  2. Abstract
  3. Introduction
  4. Casino Catchments and the Distribution of Gambling Impacts
  5. Estimating the Extent of Casinos Social and Economic Impacts
  6. Study Area Description
  7. Methods
  8. Results
  9. Discussion
  10. References

Main results

The predicted and observed catchment maps demonstrate considerable variation in the spatial extent of Australian casinos. The predictive trade-area outputs from the gravity modelling suggest that there are three general types of casino catchments in Australia, supporting Eadington's (1998a, 1998b) notion that casinos fall within categories with respect to spatial attributes. The first catchment type is highly localised, drawing patrons at a local scale (e.g., Canberra and Hobart casinos). These casinos may have a similar impact to large clubs that have comparable numbers of EGMs (e.g., Doran, Marshall, and McMillen 2007). A second group of casinos had larger regional catchments that were more extensive but still mostly contained within the state they were located in. Examples of such catchments are the catchments for the Treasury and Reef Casinos in Queensland, the SKYCITY casino in Adelaide and Lasseters Casino in the Northern Territory. A third group of casinos had catchments that extended beyond the boundaries of the states they were located in and were predicted to draw a substantial quantity of patrons from surrounding states. Examples of casinos with these characteristics were the Star City Casino in Sydney and Crown Casino in Melbourne. A substantial quantity of cross-border casino visitation was evident for these national casinos.

The validity of these general trends is supported by the observed NVS casino visitation maps. The Crown Casino, Australia's largest, does indeed have the largest observed catchment, stretching across all states, including much of Australia's rural and remote hinterlands. However, a substantial degree of spatial competition was evident: Few visitors to Crown Casino reside in the area between Sydney and Brisbane, potentially due to casino gambling opportunities closer to home. The Burswood Complex's catchment was substantially smaller than predicted and was limited almost entirely to its host state, Western Australia. This is likely to reflect Australia's urban structure as a federation of state-specific primate systems, which is especially pronounced in Western Australia, South Australia and the Northern Territory (Johnston 1969). The observed spatial extent of the SKYCITY Casino in Adelaide is similar to that predicted, although a greater predicted connectivity with the Northern Territory was evident. Finally, the trade area of Wrest Point Casino very closely matches that predicted, with a limited catchment evident that is almost exclusively confined to the island state. Overall, while the model appears to generally match the patterns of casino visitation reported in the NVS, the trade-area model appears overly optimistic for casinos with a predicted national catchment, estimating a greater quantity visits from rural areas nationally than is supported by the visitation data.

The use of NVS response maps to validate predicted trade areas is subject to several limitations. Firstly, relatively few respondents per SLA report visiting casinos (M [SD] = 0.7 [3.77] respondents per SLA-casino pair), especially in rural areas where population density is lower. Secondly, the raw number of respondents varies substantially between SLAs, further inflating the variance in the proportion of visitation. Even after the application of a spatial smoother, this results in patchwork-like catchment maps, where individual responses can have a disproportionate visual representation, especially in large rural and remote SLAs with low population density. Thirdly, the scale of the maps presented here obscures the interactions between high-density, small area cities from which many casino visits originate. Finally, neither the NVS response maps nor the trade-area model consider international travel, although overseas residents make around 4 percent of Australian casino visits (Australasian Casino Association 2008).

Implications

The catchments of Australia's casinos are highly variable. Both the Huff model and the NVS observations demonstrate that Australia's largest casinos are associated with substantial interstate visitation. In contrast, Australia's smallest casinos are little visited by interstate travellers. While this result is hardly surprising when the relative revenues of these casinos are considered (see Table 2), the regulatory implications are important. Firstly, with the exception of Canberra casino, which does not contain EGMs, the casinos that are least successful in attracting interstate visitors are those located in tourist destinations. In particular, the Reef Hotel Casino, Jupiters Townsville, Wrest Point Casino and Country Club Tasmania have catchments that rarely cross state borders. These casinos have been largely unsuccessful in extracting revenue from neighbouring states. Unless these casinos attract a substantial number of international visitors, they should be regulated as though their markets were entirely local (Shoemaker and Zemke 2005).

The national catchments of Star City Casino and Crown Casino demand different considerations. While they are successful at attracting interstate visitors, these casinos are likely to also be competing with non-casino community venues for the EGM gambling market. These casinos are highly accessible at the local level, located in the largest cities and often positioned at easily reachable locations within the cities themselves (Productivity Commission 2010). In the context of gambling-related harm, this means that casinos can be considered to have dual catchments, both at the national level where they compete with other casinos and within their host cities where they compete with community venues.

For casinos with regional or national catchments, one immediate observation that can be made here is that some economic and social impacts are exported to other states, which have different regulatory structures. The level of such impacts would not be quantified by state-level modelling or problem-gambling prevalence surveys, as they would not capture those casino patrons who are from other states. Furthermore, although many residents of rural and remote areas visit casinos, there is little recognition in Australia of the potential for gambling venues to extend their social and economic impacts regionally, interstate and even nationally. It is therefore significant that the observed visitation patterns described in this paper demonstrate that many casinos draw patrons from regional areas within a state.

In general, the fluidity and cross-jurisdictional catchment patterns suggest that there is a pressing need to better understand the accessibility and visitation patterns of casino patrons within Australia. Current prevalence surveys and techniques are unlikely to capture these trends. Products such as the Geocoded National Address Files could be used to conduct spatially targeted surveys to address this issue. Specific gambling-related surveys would best be conducted at venues themselves, in order to capture the interactions between, for example, distance travelled to gambling venue and social impacts. Surveys such as these could provide governments with evidence of the national distribution of gambling's social impacts, in order to implement appropriate harm-minimisation strategies and redistribute government gambling revenue in a more equitable manner. However, as the destination-style gambling hypothesis suggests, unless casino gambling is controlled at the national level, there is little incentive for regulators to reduce the cross-border redistribution of casino gambling's impacts.

Finally, the Huff model appears to be well suited to predicting casino trade areas at the national level, and should be considered for application in other areas. The use of estimated expenditure as a casino attractiveness measure in the trade-area model is validated by the close correlation between estimated expenditure and number of reported visits. Future research should empirically assess and calibrate model fit, including the strength of competition between venues and might usefully investigate elasticity of demand in relation to increasing accessibility using a partially constrained model (Ottensmann 1997). Travel time or cost should be considered as a distance metric in future research, as it allows for the incorporation of air travel, which is important in the national and regional travel context. Future data collection might focus on understanding the scale and intensity of gambling's social and economic impacts in casinos' peripheries, beyond the estimation of spatial catchments presented here. The Huff model holds considerable potential for researchers and policy makers in understanding the geographic scope of casino impacts, especially given the continued trend towards the proliferation of regional casinos (e.g., in Ohio and Pennsylvania).

References

  1. Top of page
  2. Abstract
  3. Introduction
  4. Casino Catchments and the Distribution of Gambling Impacts
  5. Estimating the Extent of Casinos Social and Economic Impacts
  6. Study Area Description
  7. Methods
  8. Results
  9. Discussion
  10. References
  • American Gaming Association. 2012. The State of the States: The AGA Survey of Casino Entertainment, 2012th ed. Washington: American Gaming Association. http://www.americangaming.org/sites/default/files/uploads/docs/sos/aga_sos_2012_web.pdf archived at http://www.webcitation.org/67ZhE4um1 (accessed May 2012).
  • Asplund, M., R. Friberg, and F. Wilander. 2007. Demand and distance: Evidence on cross-border shopping. Journal of Public Economics 91(1–2): 141157.
  • Ault, D., and T. Johnson. 1973. Probabilistic models applied to hospital planning. Growth and Change 4(2): 713.
  • Australasian Casino Association. 2008. The Australian casino industry: Economic Report 2006–2007. Sydney, NSW. http://www.auscasinos.com/assets/files/pdf/TheAustralianCasinoIndustry-EconomicReport-0607.pdf archived at http://www.webcitation.org/6AFRZIfUd (accessed August 2012).
  • Barry, T. 1999. The national visitor survey. Journal of the Bureau of Tourism Research 1(1): 18.
  • Beatty, T., E.R. Larsen, and D.E. Sommervoll. 2009. Driven to drink: Sin taxes near a border. Journal of Health Economics 28(6): 11751184.
  • Bell, M., M. Blake, P. Boyle, O. Duke-Williams, P. Rees, J. Stillwell, and G. Hugo. 2002. Cross-national comparison of internal migration: Issues and measures. Journal of the Royal Statistical Society. Series A, (Statistics in Society) 165(3): 435464.
  • Bivand, R., E. Pebesma, and V. Gómez-Rubio. 2008. Applied Spatial Data Analysis with R. New York: Springer.
  • Calcagno, P., D. Walker, and J. Jackson. 2010. Determinants of the probability and timing of commercial casino legalization in the United States. Public Choice 142(1): 6990.
  • Chiou, L., and E. Muehlegger. 2008. Crossing the line: Direct estimation of cross-border cigarette sales and the effect on tax revenue. The B.E. Journal of Economic Analysis & Policy 8(1): 139.
  • Clarke, G. , M. Madden , G. Clarke , and M. Clarke , eds. 2001. Applied spatial interaction modelling. In Regional Science in Business, 137158. Berlin: Springer.
  • Cooper, L., and M. Nakanishi. 1988. Market-Share Analysis?: Evaluating Competitive Marketing Effectiveness. Boston: Kluwer Academic Publishers.
  • Cornell, S. 2008. The political economy of American Indian gaming. Annual Review of Law and Social Science 4(1): 6382.
  • Dennis, C., D. Marsland, and T. Cockett. 2002. Central place practice: Shopping centre attractiveness measures, hinterland boundaries and the UK retail hierarchy. Journal of Retailing and Consumer Services 9(4): 185199.
  • Dombrink, J., and W.N. Thompson. 1990. The Last Resort: Success and Failure in Campaigns for Casinos. Reno: University of Nevada Press.
  • Donahue, J. 1997. Tiebout? or not tiebout? The market metaphor and America's devolution debate. The Journal of Economic Perspectives 11(4): 7381.
  • Doran, B., and M. Young. 2010. Predicting the spatial distribution of gambling vulnerability: An application of gravity modeling using ABS mesh blocks. Applied Geography 30(1): 141152.
  • Doran, B., D. Marshall, and J. McMillen. 2007. A GIS-based investigation of gaming venue catchments. Transactions in GIS 11(4): 575595.
  • Doughney, J. 2009. Lies, damned lies and “problem gambling” prevalence rates: The example of Victoria, Australia. Journal of Business Systems, Governance and Ethics 2(1): 4154.
  • Eadington, W. 1986. Impact of casino gambling on the community comment on Pizam and Poleka. Annals of Tourism Research 13(2): 279282.
  • Eadington, W. 1998a. Casino gaming—Origins, trends and impacts. In Casino Gambling in America: Origins, Trends and Impacts, ed. K. Meyer-Arendt and R. Hartmann , 315. New York: Cognizant Communication Corporation.
  • Eadington, W. 1998b. Contributions of casino-style gambling to local economies. The Annals of the American Academy of Political and Social Science 556(1): 5365.
  • Eadington, W. 1999. The economics of casino gambling. The Journal of Economic Perspectives 13(3): 173192.
  • Felsenstein, D., and D. Freeman. 2001. Estimating the impacts of crossborder competition: The case of gambling in Israel and Egypt. Tourism Management 22(5): 511521.
  • Fotheringham, S., and M. O'Kelly. 1989. Spatial Interaction Models: Formulations and Applications. Studies in Operational Regional Science. Dordrecht, Boston: Kluwer Academic Publishers.
  • Garrett, T., and T. Marsh. 2002. The revenue impacts of cross-border lottery shopping in the presence of spatial autocorrelation. Regional Science and Urban Economics 32(4): 501519.
  • Geoscience Australia. 2006. GEODATA TOPO 250K series 3: User guide. Canberra: Geoscience Australia.
  • Golledge, R., and R. Stimson. 1997. Spatial Behavior: A Geographic Perspective. New York: Guilford Press.
  • Greene, R., and J. Pick. 2012. Exploring the Urban Community: A GIS Approach, 2nd ed. Upper Saddle River, NJ: Pearson Prentice Hall.
  • Grinols, E., and J.D. Omorov. 1996. Development or dreamfield delusions: Assessing casino gambling's costs and benefits. Journal of Law and Commerce 16: 4987.
  • Hanewinkel, R., and B. Isensee. 2007. Access to cheaper cross-border cigarettes may decrease smoking cessation intentions in Germany. Tobacco Control 16(1): 7071.
  • Harper, P. 2005. Information paper: Draft mesh blocks, Australia. Catalogue no. 1209.0.55.001. Canberra: Australian Bureau of Statistics. http://www.abs.gov.au/ausstats/abs@.nsf/mf/1209.0.55.001 archived at http://www.webcitation.org/6AFRwtbEs (accessed August 2012).
  • Hing, N., and H. Breen. 1996. Day tripping in the tweed: An empirical investigation of “Pokie Tours” from Southern Queensland to Northern New South Wales clubs. Australian Journal of Hospitality Management 3(2): 113.
  • Huff, D. 1964. Defining and estimating a trading area. The Journal of Marketing 28(3): 3438.
  • Hyland, A., J. Bauer, Q. Li, S. Abrams, C. Higbee, L. Peppone, and M. Cummings. 2005. Higher cigarette prices influence cigarette purchase patterns. Tobacco Control 14(2): 8692.
  • Johnston, R. 1969. Population changes in Australian small towns, 1961–1966. Rural Sociology 34(2): 212217.
  • Leal, A., J. López-Laborda, and F. Rodrigo. 2010. Cross-border shopping: A survey. International Advances in Economic Research 16(2): 135148.
  • Lynch, R., and A. Veal. 1998. The casino in the post-industrial city: The social and economic impact of the Sydney casino. In Leisure Management: Issues and Applications, ed. M. Collins , and I. Cooper , 181196. Oxon, UK: CAB International.
  • Marshall, D. 2002. The evolving market structures of gambling: Case studies modelling the socioeconomic assignment of gaming machines in Melbourne and Sydney, Australia. Journal of Gambling Studies 18(3): 273291.
  • Marshall, R. 1991. Mapping disease and mortality rates using empirical bayes estimators. Journal of the Royal Statistical Society. Series C, Applied Statistics 40(2): 283294.
  • Martin, P. 1982. Shopping Centre Management. London: E. & F.N. Spon.
  • McMillen, J. 1990. Casino policies: Have Australians had a fair deal? Journal of Gambling Studies 6(1): 329.
  • Ottensmann, J. 1997. Partially constrained gravity models for predicting spatial interactions with elastic demand. Environment and Planning A 29(6): 975988.
  • Productivity Commission. 1999. Australia's Gambling Industries. Canberra: AusInfo.
  • Productivity Commission. 2010. Gambling. Canberra: Productivity Commission.
  • Robitaille, E., and P. Herjean. 2008. An analysis of the accessibility of video lottery terminals: The case of montreal. International Journal of Health Geographics 7(1): 115.
  • Roy, J. 2004. Spatial Interaction Modelling: A Regional Science Context. Advances in Spatial Science. Berlin, New York: Springer.
  • Shoemaker, S., and D.M.V. Zemke. 2005. The “locals” market: An emerging gaming segment. Journal of Gambling Studies 21(4): 379410.
  • Svensson, J. 2009. “Travellers” alcohol imports to Sweden at the beginning of the 21st century. Nordic Studies on Alcohol and Drugs 26(2): 193203.
  • The Australian Institute for Gambling Research. 2000. The Impact of the Expansion of Gaming on the Tourism, Entertainment and Leisure Industries. Melbourne, Vic: Victorian Casino and Gaming Authority.
  • Timothy, D., and R. Butler. 1995. Cross-boder shopping: A North American perspective. Annals of Tourism Research 22(1): 1634.
  • Walker, D., and J. Jackson. 2007. Do casinos cause economic growth? American Journal of Economics and Sociology 66(3): 593607.
  • Wang, F. 2006. Quantitative Methods and Applications in GIS. Boca Raton, FL: Taylor & Francis.
  • Watson, L., and S. Kale. 2003. Know when to hold them: Applying the customer lifetime value concept to casino table gaming. International Gambling Studies 3(1): 89101.
  • Wenz, M. 2008. The spatial evolution of casino gambling. Cityscape: A Journal of Policy Development and Research 10(3): 203228.
  • Williams, R., J. Rehm, and R. Stevens. 2011. The social and economic impacts of gambling. http://hdl.handle.net/10133/1286 archived at http://www.webcitation.org/6AFSDWenk (accessed August 2012).