Hkjc betting station square

Support for hypothesis two would arise if low BP horses were under-bet in the first three races at HV. This is predicted by hypothesis two because a low BP should afford a considerably greater advantage at HV than at the previous ST meeting. In addition, less under-betting on low BP horses should occur in the last three races at HV, as bettors learn from the earlier race outcomes at HV and place less emphasis on the outcomes of races at the previous ST meeting.

Similarly, hypothesis two would be supported if there was evidence of less over-betting on low BP horses at ST in the last three races c. As discussed above, the configurations of both ST and HV suggest that low BP confers an advantage since horses running close to the inside rail travel less distance. However, other factors can affect the BP advantage. In particular, it is possible that the condition of the surface on certain sections of track perhaps as a result of weather conditions may lead to certain BPs offering an advantage at a particular time on any given day.

However, since individuals often anchor on the most immediate information, it is possible that bettors place too much emphasis on the most recent race result when forming their subsequent probability judgments. Late bettors are confirmed as relatively knowledgeable bettors, who achieve higher returns than other bettors, by a number of studies e.

The odds observed five minutes before the race starts late odds are therefore regarded 13 as being influenced by individuals with moderate expertise. It is well known that professional bettors operate in the Hong Kong market and that they spend considerable resources in analyzing results of previous races in order to make their selections.

They tend to bet heavily in the last couple of minutes before the race Benter, Since odds in a pari-mutuel market are determined by the relative amount of money bet on each horse see Eq. However, late bettors have the opportunity to bet in a manner to exploit any biases they observe in earlier betting patterns.

Consequently, final odds are assumed to be those most influenced by bettors with the greatest expertise. Clearly, the final odds not only incorporate the wagers of those with greatest expertise, but also those with moderate and least expertise. Similarly, the odds five minutes before the race incorporate the wagers of those with moderate expertise and of those with least expertise. The distinctions between the betting of those distinguished by their level of expertise may therefore be masked to some extent by comparing the odds at these different points in time.

This problem would be overcome if we could compare early odds with those determined by the amount of money wagered on horses a from the commencement of betting on the day of the race up to 5 minutes before the start of the race and b from 5 minutes before the race until the race starts.

This would enable the marginal odds created by the three groups distinguished by their levels of expertise to be more clearly defined. Unfortunately, this would require knowledge of the amount bet in the market in these different time periods, which we could not obtain.

Consequently, in analyzing the results it should be bourn in mind that any distinctions we do observe in the early, late and final odds are likely to under-state the true distinctions between the betting behavior of these those with different levels of expertise. Each of the models used to test hypotheses are developed on the basis of a early, b late, and c final odds and these are used to assess differences in the degree and the nature of anchoring effects between bettors with different increasing levels of expertise.

We focus, at this stage, on the results associated with the final odds in the market, since these represent the combined probability judgments of all the market participants. Taken alone, these results suggest that Hong Kong bettors do not anchor their judgments on the clear advantage afforded to those horses with low BP; this advantage being easily discerned from the configuration of the racetracks, from previous race results and from media comment.

Consequently, one might conclude that anchoring on BP advantage is absent. This finding is clearly at odds with the main conclusions of the anchoring literature e. However, as demonstrated below, this inconsistency arises because the conclusions based on the Hong Kong market as a whole are misleading.

In fact, combining results from the two racetracks leads to model mis-specification since the effect of BP on winning probabilities and the manner in which bettors handle BP information at the two racetracks appears to be different. This is confirmed by comparing the sum of the information contained in the separate CL models for the two racetracks with the information contained in the CL model developed for the Hong Kong market as a whole. It is clear, therefore, that the results for the Hong Kong market as a whole mask the manner in which bettors at the two tracks use information.

In sum, the evidence presented above appears, in relation to bettors at ST, to support hypothesis one, namely, that bettors anchor their probability judgments on BP advantage, whereas the evidence drawn from HV appears to reject the hypothesis.

Consequently, there is no clear evidence that bettors simply anchor on their pre-conceived view of BP advantage discerned from the configuration of the racetrack or from the results of previous races. Consequently, the results displayed in Table 2 and discussed above may point to anchoring by bettors on the results from the previous race meeting held at the other venue earlier in the week in support of hypothesis two. The results displayed in Tables 1 and 2 confirm the over- and under-betting of low BP horses in the manner predicted by hypothesis two.

HV meetings following ST meetings, and vice-versa also supports hypothesis two see Table 3. These results imply that bettors over-estimate the advantage of low BP at ST and underestimate it at HV at different track meetings. Consequently, taken together, these results offer strong support for hypothesis two.

In particular, there appears to be anchoring on the results from the previous race meeting for the different track meetings; BP being under-bet at HV following a ST meeting and over-bet at ST following a HV meeting. It appears that bettors are able to learn to appropriately account for BP at a particular track, provided this learning is not disrupted by an intervening race meeting at the alternative venue, where a different BP bias prevails. In the latter case, bettors appear to anchor on the results from the previous meeting, even when these are not relevant to races at the current track.

Further evidence to test hypothesis two is presented in Tables 4 and 5. At HV the BP variable has a coefficient which is negative and significant in both sets of races, but the level of significance is greater in the last three races than in the first three races; suggesting that a low BP offers more advantage in the later races.

This may arise because the race types and distances of races can vary between early and late races, and these factors may influence BP advantage. In addition, the inside rail is often moved for later races, allowing horses drawn on the inside to run on fresh ground which has not been damaged by previous runners.

The precise manner in which these factors combine to influence the effect of BP on race results is not clear, but the key issue in respect of the current research is that at HV the advantage of a low BP increases in later races. This suggests that low BP confers some advantage in the first three races but no advantage in the later races. As indicated above, many factors could contribute to the BP advantage changing between the early and late races e.

However, the reasons for this are not the central focus of this enquiry. Rather, we are concerned with the extent to which bettors anchor on BP, whatever level of advantage it confers at different times of the betting day. Consequently, when exploring to what extent BP is accounted for in odds, we do not control for different factors which could cause changes in the BP advantage throughout the betting day.

Once again, we first explore the results associated with the odds which represent the combined probability judgments of all the market participants; that is, the final odds. This suggests that, as the coefficient in the model incorporating only BP is also negative for the first three races at HV, bettors fail to take full account of the advantage of low BP in the early races.

A different picture emerges in relation to the last three races at both tracks. However, by the end of the HV race meeting there are typically ten races at ST and eight at HV they have fully adjusted their subjective probability estimates to account for the clear advantage enjoyed by horses with low BPs. Similar results are obtained when separating races run at HV immediately following a race meeting at ST. Consequently, bettors at HV appropriately account for BP in these later races, but they probably do not achieve this by simply relying on information from the first few race results at HV, since the advantage of low BP is much greater in the last three races.

The ability of bettors to develop such sophisticated mental models is confirmed by Ceci and Liker The influence of anchoring from the previous race meeting is further confirmed when examining same track meetings at HV separately. These results together provide strong evidence of anchoring on experience from the previous race meeting at ST. Further support for hypothesis two is provided by an analysis of early and late races at ST. It may be that bettors anchor on the results of races run earlier that afternoon at ST.

In addition, the lack of low BP advantage displayed in the results of later races at ST may well act as a considerable anchor on the judgments of bettors in the subsequent early races at HV; and the results 19 reported above support this view.

However, the analysis of early races at ST offers a challenge to hypothesis two: in the first three races of a ST meeting, where low BP confers a moderate advantage, bettors fully account for this in their betting decisions. However, because the advantage of a low BP at ST in these early races is considerably lower than that experienced in the last three races at HV, hypothesis two would predict that ST bettors would anchor on the later results at HV and, consequently, would over-estimate the benefit of a low BP at ST.

Separating races run at same and different track meetings does not shed additional light on the reasons for bettors at ST correctly accounting for BP advantage in the first three races. No BP advantage is discerned in the first three races at ST in either same or different track meetings; the BP coefficients in CL models simply incorporating BP are not significantly different to zero same track meetings: Taken together, the results relating to the manner in which bettors at ST and HV account for BP bias in early and late races offer some support for the view that bettors anchor their probability judgments on race results they encounter at the previous race meeting.

However, the simple relationship suggested by hypothesis two is not confirmed. Whilst this is difficult to confirm, we aim to further our understanding of the extent to which bettors anchor on previous race outcomes by testing hypothesis three, namely, that bettors anchor their probability judgments on the BP of the winning horse in the previous race. One way in which bettors can discern such changes is to observe the result of the preceding race.

In addition, at both racetracks these coefficients are more negative and more significant than those for CL models based on all races at these meetings see Table 1. Following a winner from a high BP, the coefficient of BP remains negative but is of reduced significance at HV, and at ST it is no longer significant at any conventional level.

In this case these coefficients are both less negative and less significant than for those for CL models based on all races at these meetings see Table 1. Our main concern is to examine to what extent bettors base their probability judgments on the BP advantage information contained in the most recent race outcome.

The results are displayed in Table 7, and again we focus on the results associated with the final odds, which represent the combined probability judgments of all the market participants. When the previous race was won by a horse from a low BP, the BP coefficients are not significant at any conventional level in models developed for ST and HV which incorporate both BP and odds implied probabilities.

These results suggest that bettors at both tracks fully account for BP advantage in race following a low BP winner. In addition, when the previous race was won by a horse with a high BP, the BP coefficients are not significant at any conventional level in the model developed for HV which incorporates both BP and odds implied probabilities. However, following a low-BP winner they fully account for low BP advantage in the next race. This suggests that bettors use previous race outcome information to improve their calibration.

There are potentially two reasons for bettors correctly accounting for BP advantage in this subsequent race. In other words, the results are also consistent with bettors ignoring the previous race result when a high BP horse wins because it does not accord with their mental model of the likely BP advantage.

Previous research suggests that individuals discount evidence which runs counter to existing perceptions; the inconsistent evidence may be discounted deliberately to reduce dissonance and effort Harries et al. Bettors certainly do not appear to unduly anchor on a high BP victory a low probability event according to their mental model in the previous race. A somewhat similar effect has been found in a laboratory gambling task Carlson, However, following a winner from a low BP they correctly account for the enhanced low BP advantage in the next race.

Whether this arises from appropriate use of information contained in the result of the previous race, or simply because the advantage of a low BP is now more in line with bettors pre-conceived view of this advantage, can not be determined. However, when a horse from a high BP wins a race the low BP advantage in the subsequent race becomes insignificant see Table 6. Bettors do not appear to fully account for the information contained in a high BP victory, since the results show that they continue to over-estimate the advantage of a low BP in the subsequent race see Table 7.

It is again possible that they discount the high BP winner since it is a low probability event according to their mental model of BP advantage Carlson, However, some of the evidence presented above suggests that bettors may fail to take full account of previous race outcomes which fall outside their mental model of BP advantage; this result is in line with some previous research e.

Harries et al. Consequently, it is suggested that the manner in which bettors use information, and hence their degree of anchoring, may depend upon a complex relationship between the strength of their existing mental model, the degree to which the potential anchor accords with this mental model, and the nature of the anchor information. Expertise and Anchoring Finally, we turn to testing hypothesis four, which is based on previous laboratory findings that decision makers with greater expertise are less prone to anchoring effects e.

These effects are considerably more pronounced when the bets of those considered as having less expertise are examined. For example, the BP coefficients in the CL models incorporating odds and estimated using HV data are considerably more significant for models based on early odds and late odds than in models based on final odds see Table 2. These results are confirmed when races are separated into same and different track meetings.

At same track meetings the difference 23 between models based on early and final odds is less pronounced as might be expected, since less bias is introduced by anchoring on an inappropriate BP advantage observed at the alternative venue. However, final odds still appear to correctly account for BP to a greater extent. Taken together, these results suggest that those with less expertise are more likely to anchor their judgments on race outcomes from the previous race meeting.

In practice this phenomenon is more pronounced than is demonstrated by the results displayed in Table 2. This arises because the odds are determined by the relative amounts of money on each horse see Eq. Consequently the final odds, which we have associated with those with the greatest expertise, will be determined by the money wagered by these individuals and by the earlier bets of those with less expertise. Examination of the degree to which BP advantage is accounted for in odds in early and late races at HV and ST suggests that those with less expertise are more likely to anchor on the results at the previous race meeting.

These coefficients are increasingly significant for models based on odds formed by bettors with less expertise; suggesting that less experienced bettors under-value the advantage of low BP to a greater extent than those with more expertise. These results are confirmed when specifically exploring different track meetings.

For example, in the first three races at HV the BP coefficient is negative and significant at the. Consequently, bettors with less expertise appear more prone to anchor on the results from the race meeting at ST held prior to the HV meeting. None of the coefficients for BP in the CL models estimated on the three sets of odds final, late and early odds for the first three races at ST are significant see Table 5; this result also holds for models based on both same and different track meetings , although they are all positive and the Z value is greatest for the model based on odds formed by bettors with least expertise.

These results provide weak evidence in favor of more anchoring on the results from the previous HV meeting by those bettors with less expertise. In the CL models estimated for the last three races at ST, based on final and late odds early odds are not included in the comparison since these were formed by individuals without the benefit of observing the earlier races at ST , the BP coefficients are positive and significant see Table 5.

This suggests that, despite the reduced advantage of low BP in these late races, bettors continue to anchor on the results from earlier in the day when low BP had a more distinct advantage. However, the degree of anchoring on earlier results is slightly less pronounced for those bettors with greater expertise.

Neither of the coefficients for BP in CL models for the last three races at HV, estimated on the late and final odds, is significant, although their relative size provides some weak evidence that bettors with greater expertise use information from earlier results more appropriately to adjust their probability estimates in later races.

Taken together, the results discussed above suggest that bettors with greater expertise are less prone to anchoring effects. As indicated in section IV, our method for distinguishing the bets of those with different levels of expertise somewhat understates the degree to which they differ. Consequently, our results point strongly to expertise reducing the degree to which anchoring is employed.

This finding is this is in line with the results of experimental studies. For example, Shelton found that more-experienced auditors audit managers and partners were less influenced by irrelevant information encountered during audit judgments than auditors with less experience audit seniors.

The final set of results exploring the relationship between expertise and anchoring do not accord with conclusions reached in previous research: A comparison of CL models based on final and late odds, estimated for races run at HV and ST, reveal no appreciable differences in 25 the manner which the immediately preceding race outcomes are used by bettors with varying levels of expertise see Tables 6 and 7.

It appears that bettors with varying levels of expertise are able to appropriately employ information contained in the most recent race result. However, as indicated above, those bettors with less expertise are more prone to anchoring on an earlier series of race outcomes. Overall, the results suggest that experts are less prone to anchoring.

However, it could be argued that to some extent it is surprising that the experts in the Hong Kong market are subject to any anchoring at all. It has been widely reported e. Benter, that large betting syndicates, using sophisticated computer models operate in the Hong Kong market.

They are attracted to these markets due to the high betting volumes, which provide them with the opportunity to recoup the significant set up and operating costs of such operations. These are designed to systematically account for factors likely to influence the probability of a horse winning.

Anecdotal evidence suggests that these syndicates make substantial profits and there is evidence that their betting activities may eliminate some biases which are observed in markets where these syndicates do not operate e.

However, whilst these syndicates may employ computer models and be successful, the models are developed by human beings and there is no evidence that they bet optimally. In fact, whilst their bets do reduce the degree of anchoring displayed in the odds in this market, some anchoring remains, suggesting that even highly sophisticated experts judgments remain subject to the bias caused by this heuristic.

Conclusion A number of conclusions emerge from the research: First, that caution must be exercised when examining anchoring effects in a real world environment, since mis-specification of the problem can easily occur by combining data inappropriately. In the current study, no anchoring was observed when data for the Hong Kong betting market as a whole was analyzed.

However, this false conclusion arose because bettors, when making their probability 26 judgments, under-valued the advantage of a low BP at one track and over-valued it at another. It is suggested, and demonstrated, that both of these behaviors arise through anchoring and yet their effects were masked when examining the market as a whole. Second, the paper demonstrates that anchoring in real world environments is a complex phenomenon.

Bettors do not, for example, simply anchor on the configuration of a racetrack or on media analysis which suggests that particular BPs will have an advantage. Rather, bettors appear subject to complex anchoring effects. In particular, the research suggests that bettors can anchor on a series of race outcomes from a race meeting held earlier that week.

However, this behavior appears to be moderated by other factors, such as the current racetrack configuration and their own mental model of BP advantage. In fact, bettors do not anchor on the result of the most recent race or series of races run on the same day when they are forming probability judgments. On the contrary, they appear to appropriately use this information when forming their probability judgments later that day, particularly, when the previous race outcomes are in line with their mental model of BP advantage.

However, there is some evidence that when race outcomes are contrary to their mental model, bettors fail to take full account of the information. For example, following a high BP victory, bettors do not sufficiently adjust their subjective probability estimates of a horse with a low BP winning the next race. In summary, the results suggest that anchoring effects in the horse race betting market are subtle and are influenced by a range of factors, including racetrack configuration and previous race outcomes; these factors appear to interact to produce a complex anchoring phenomenon.

Third, the results suggest that basic anchoring effects in a real world environment can be robust and relatively long lived. This result is in sharp contrast to previous studies conducted in the laboratory, which suggest that because basic anchoring does not involve a direct comparison process between anchors and targets, basic anchoring effects are fragile and can easily disappear e.

Brewer and Chapman, Further work is needed to explore the nature of the factors experienced in a real world environment which appear to increase the impact of basic anchoring effects. Fourth, the results suggest that, in general, decision makers with greater expertise use information more appropriately and are less prone to anchoring effects. However, whilst this is true for anchoring based on a recent series of previous races, either at the current, or at another racetrack, no differences were detected between those with greater and less expertise in terms of the degree to which they anchor on the results of the last race encountered.

This, to some extent, accords with expectations in betting markets, since expert bettors are widely 27 regarded as being distinguished from those with less expertise by the considerable resources they spend in overcoming the complexity of analyzing and developing probability estimates based on the outcomes of previous series of races Benter, In summary, the paper presents evidence of subtle, complex anchoring effects in a real world environment.

Most previous anchoring studies have been conducted in the laboratory on individuals. An important feature of the current study is that it examines behavior in a financial market. In many ways one might expect markets to eliminate individual anchoring effects, as some individuals seek to exploit the biases of others.

The odds, reflecting the betting behaviors of those subject to the biases and the behaviors of those who seek to capitalize on the biases, should, in theory, be driven to a point where no bias exists. Remarkably, we find that these markets still exhibit the effects of anchoring. Clearly, the anchoring effects we observe are complex and it is possible that the behavior of individuals in a betting market, who are faced by a negative expected return, may be different to that of individuals facing other situations involving risk.

In addition, race specific factors e. Consequently, more research is needed to identify the factors which influence the degree and nature of anchoring in betting markets and in other real world contexts. Endnotes 1. All odds in HK are rounded to the nearest lower 10 cents. Consequently, by knowing this and the level of deductions d, we can use Equation 1 to work back from the final odds to provide an estimate of pijs. Real and laboratory gambling, sensation-seeking, and arousal.

British Journal of Psychology; 75, Asch, P. Market efficiency in racetrack betting. Journal of Business, 57, Efficiency and profitability in exotic bets, Economica, 59, Ashton, A. Sequential belief revision in auditing. The Accounting Review, 63, Ashton, R. Journal of Behavioral Decision making, 15, Benter, W.

Computer based horse race handicapping and wagering systems: A report. Hausch, V. Ziemba Eds. London: Academic Press. Beyer, A. Bonner, S. Experience effects in auditing: The role of task-specific knowledge. The Accounting Review, 65, Brecher, S. Beating the races with a computer. Bruce, A. The analysis of risky decision making: Methodological issues and implications. Market ecology and decision behavior in state contingent claims markets.

Journal of Economic Behavior and Organization, 56, Carlson, B. Anchoring and adjustment in judgments under risk. Caverni, J. Gonzalez Eds. New York: North-Holland. Ceci, S. A day at the races: A study of IQ, expertise and cognitive complexity. Journal of Experimental Psychology: General, , Cervone, D. Anchoring, efficacy, and action: The influence of judgment heuristics on self-efficacy judgments and behavior. Journal of Personality and Social Psychology, 50, The limits of anchoring.

Journal of Behavioral Decision Making, 7, Chapman, G. Incorporating the irrelevant: Anchors in judgments of belief and value. Gilovich, D. Kahneman Eds. Cambridge, U. Cohen, M. Three paradigms for viewing decision biases.

Klein, J. Orasanu, R. Zsambok Eds. Cotton, M. Value betting. Oswestry, U. Davies, M. Belief persistence after evidential discrediting: The impact of generated versus provided explanations on the likelihood of discredited outcomes.

Journal of Experimental Social Psychology, 33, Dowie, J. On the efficiency and equity of betting markets, Economica, 43, Covariance decompositions and betting markets-early insights using data from French trotting. Vaughan Williams Eds. Epley, N. Putting adjustment back in the anchoring and adjustment heuristic: Differential processing of self-generated and experimenter-provided anchors.

Psychological Science, 12, Estrada, C. Positive affect facilitates integration of information and decreases anchoring in reasoning among physicians. Organizational Behavior and Human Decision Processes, 72, Figlewski, S. Subjective information and market efficiency in a betting model. Journal of Political Economy, 87, Goodman, J. The interactive effects of task and external feedback on practice performance and learning.

Organizational Behavior and Human Decision Processes, 76, Harries, C. Combining advice: The weight of a dissenting opinion in the consensus. Journal of Behavioral Decision Making, 17, Hausch, D. Transactions costs, market inefficiencies and entries in a racetrack betting model. These results relate to all gamblers rather than excessive ones.

It is possible to speculate that the illegal gambling results are driven by the high proportion of excessive gamblers who are also illegal gamblers, though further work would be necessary to establish this. In the interim, the results provide some tentative suggestion that illegal gamblers in general may be more susceptible to a range of personal and financial issues than their legal counterparts. These issues, in turn, lead to considerations of how illegal gamblers finance their habit.

In common with many other jurisdictions, Hong Kong bans offering credit in its legal gambling institutions, though of course no such rules apply to the illegal sector. In order to gauge the relative attractiveness of credit, the survey suggested a hypothetical scenario of legally offering credit in gambling.

As indicated in Table 3 , very few illegal gamblers indicated that they had borrowed from unofficial money lenders or obtained money to gamble illegally. Given the small numbers involved, no definitive evidence can be drawn from the survey. However, Wong et al. Likewise, while the illegal gamblers surveyed did not admit to committing crimes to finance their habit, Lahn and Grabosky found that one in four Australian pathological gamblers and one in ten problem gamblers had undertaken gambling-related illegal activity at some point.

This work still leaves open the possibility that those committing crimes were more likely to gamble. Given the overrepresentation of excessive gamblers amongst illegal gamblers, these figures are sobering. His work implies that excessive gamblers who run up debts with illegal operators may be particularly likely to engage in crimes such as fraud.

A separate issue concerns the fact that illegal gambling is controlled by criminal syndicates. The survey underlying this paper concerned illegal gambling patrons rather than operators. As illegal bookmakers do not pay fees or taxes, the activity can be highly profitable. Further, to the extent that funds are laundered and sent offshore, they do not contribute to the financial well-being of Hong Kong. The fact that illegal operators, by definition, operate outside the conventional justice system also compels them towards violence, since the legal system cannot be used for collection of unpaid debts and indeed the collection of any gambling debt is illegal in mainland China Wang and Antonopoulos The impact of reducing illegal gambling in Hong Kong The survey results and international literature discussed above raises the issue of reform.

Three aspects are relevant: The social implications of liberalising current gambling arrangements. The financial implications of gambling liberalisation. The economic implications of gambling liberalisation. The evidence that a large proportion of illegal gamblers also appear to be excessive gamblers is noted above. Excessive gambling may, of course, occur among both legal and illegal gamblers. For example, those gambling through the HKJC are exposed to initiatives such as: Physical, online and TV programs aimed at promoting responsible gambling.

A ban on credit in gambling. Exclusion or self-exclusion. Online self-assessment tools and finance plans to manage gambling. Information about excessive gambling counselling and treatment centres. Controls on underage gambling. In addition, the legal gambling sector channels money into initiatives designed to help excessive gamblers both legal and illegal and society in general such as the Integrated Centre on Addiction Prevention and Treatment ICAPT , contributions to the Ping Wo fund dealing with excessive gamblers and general contributions to taxation and charity.

The HKJC contributes some three quarters of its revenue to tax and charitable causes. In contrast, the illegal gambling market offers few such safeguards and makes no known direct contribution to initiatives aimed at addressing gambling addiction or general taxation. While evidence of behaviours in illegal gambling establishments is limited, past international work has pointed to the fact that illegal operators play upon human frailties, encouraging feelings of power and omnipotence amongst winners and presenting them with gifts e.

In Hong Kong, Chinese male gamblers are likewise known to enjoy acts of bravado, bringing followers on trips to Macau and distributing large winnings in order to demonstrate their prowess Chan and Ohtsuka The fact that illegal operators play upon such vulnerabilities, rather than seeking to ameliorate them, may increase the risk of gambling addiction amongst illegal gamblers.

Accordingly, expanding the range of legal gambling options may help reduce exposure to illegal gambling operators, the harmful practices which they could encourage among patrons and the attendant social ills. A second perspective is the financial one. As a part of the abovementioned survey work, both legal and illegal gamblers were asked a series of questions about how their gambling spend would change in the event of a series of independent regulatory changes on gambling.

This allowed for modelling of the impacts of these regulatory changes by —, including both changes in legal gambling spending and the diversion of spending from illegal gambling markets, as illegal gamblers are attracted to the expanded range of legal betting options. Full details of the modelling approach are provided in Oxford Economics The modelling represents only an initial estimate based on the survey results and further confirmatory research is required.

Further, no judgement is made on the desirability of the modelled options. Conversely, construction of a Hong Kong casino was found to displace little illegal gambling revenue. This is probably due to the fact that the illegal casino market is itself small—because such gamblers already have a legal outlet through casino gambling in Macau.

These options are also associated with additional tax revenues. Tax is not an economic benefit, but a transfer between members of society. However, the proposition that such funds are better distributed back into Hong Kong society through taxes, rather than into the hands of illegal gambling operators is one that few would disagree with.

These considerations lead to the question of the economic effects of illegal gambling. Empirical studies on the long term economic effects of crime and the application of criminal proceeds rival those of illegal gambling in terms of their scarcity. While it is difficult to be certain in the case of Hong Kong, past work has suggested that criminal investment patterns tend to be sub-optimal with a particularly large over-emphasis on the real estate market Kruisbergen et al.

While some funds may be invested back into the local housing market, given the need to avoid detection and strong mainland links, it is possible that Hong Kong illegal gambling operators send substantial amounts of money to offshore real estate. Discussion Empirical work on illegal gambling is rare. This study is intended as a first step rather than as a last word. Much more needs to be done in undertaking research on the characteristics of illegal gamblers and the policy implications of such activity.

However, in order to frame this initial analysis, we set out to test several hypotheses on the nature of illegal gamblers in Hong Kong. In terms of H 1 , we find evidence that illegal gambling tends to attract individuals with a particularly avid interest in gambling. Hong Kong illegal gamblers bet more frequently and bet larger amounts than their legal counterparts. They are more attracted by good odds and betting flexibility than legal gamblers.

This is consistent with the suggestions of de Bruin et al. In particular, it may be that, as suggested by Binde, the legal market simply is not sufficient to cater for the diverse tastes of illegal gamblers. We also find support for H2.

Consistent with the international work of de Bruin et al. The causal link between excessive gambling and illegal gambling remains unclear. This is consistent with the suggestion of authors such as Binde However further research will need to be done to establish this more clearly. The evidence for H3 is more nuanced. A greater proportion of illegal gamblers appear to have suffered from at least one of the identified social problems than is the case for legal gamblers.

However, the relevant sample sizes are relatively small, and results are less clear for individual issues. This is not surprising in itself. The difficulty of finding an adequate sample to investigate the social problems due to gambling even in international work with very large initial sample sizes has been noted above. Reticence to self-identify and the issue of causality add additional layers of complexity though it should be noted that the relevant survey questions explicitly asked illegal gamblers if these conditions were due to their gambling.

Nonetheless, noting these caveats, the survey results provide some tentative evidence of a greater potential susceptibility of illegal gamblers to social problems which might be expected due to the relatively high proportion of excessive gamblers in their ranks. In particular, susceptibility to conditions such as relationship breakdown, emotional stress of partners or children and debt accumulation are consistent with past literature Productivity Commission ; Kunzi et al.

Further research would be necessary to expand on these tentative results and on the findings of this exploratory survey in general. One approach to undertaking such work would be a fully random sample of the population which would pick up illegal gamblers. As an alternative, a population survey of middle aged males—the group known to be the most likely to gamble illegally—could be conducted.

While this might also require a relatively large sample size, such an approach would be more efficient and practical than a purely random approach and allow for a further exploration of the study themes discussed above. Both past research and the current study could be also used to help inform future policy decisions. The lack of constraints in illegal gambling environments, as alluded to in the work of Bensimon et al. This is in contrast to the direct and indirect ways in which legal gambling establishments attempt to address the issue of excessive gambling.

Previous studies have also alluded to the social harms associated with excessive gambling and both the current study and past work have indicted the overrepresentation of excessive gambles within the ranks of illegal gamblers. Liberalisation may offer one approach to ameliorating some of these issues. Expanding the range of legal gambling options would draw illegal gamblers away from the damaging influence of illegal operators into the legal word, where controls help restrain excessive and damaging gambling behaviour.

Moreover, while more work needs to be done to confirm these initial results, some liberalisation measures in areas such as football betting could provide financial benefits and keep money out of the hands of illegal operators to be redistributed back to society through taxation.

Conclusion This study provides some initial insights into the nature of illegal gambling in Hong Kong, including its social and financial implications. Understanding the nature of illegal gambling is a first step in determining how to tackle the problem. Further research is required to confirm and extend its results, which will aid in helping to make important policy decisions regarding the regulation of Hong Kong gambling into the future.

An exception is made when a past study clearly differentiates between sub-sets such as pathological gamblers. Studies such as Wong et al. While the characteristics of illegal gamblers are examined in that report, the number of identified illegal gamblers equated to 6 out of respondents, limiting what conclusions could be drawn. As indicated in that survey, one difficulty is reticence to truthfully report on such gambling activities, which could contribute to underreporting.

Broad population surveys are often hampered by such problems and the difficulties of adequately sampling such sub-groups: a parallel issue arises with attempts to study excessive gambling through such surveys.

The survey we undertook for this paper asked respondents which gambling channels they had used whether through the HKJC or otherwise and avoided directly questioning whether they had participated in illegal gambling. This included a major national survey of gambling habits in the Australian general population involving regular gamblers, non-regular gamblers and non-gamblers. In addition, a separate survey involved clients in counselling centres.

Both surveys included various questions asking respondents if a variety of social, financial and relationship issues were caused by their gambling. This was supplemented by expert testimony on the links between a range of economic social and financial issues and excessive gambling. Results differed across the range of psychiatric conditions and overall, However note that this result relates to pathological gamblers rather than the broader group of excessive gamblers.

For the purposes of this study, an illegal gambler is defined as someone within Hong Kong who places a bet with an operator other than the HKJC. Illegal gambling channels may include unlicensed operators in Hong Kong, offshore operators outside of Hong Kong excluding Macau casinos and online operators both inside and outside Hong Kong.

This is consistent with the international literature that finds illegal gamblers straddle both the legal and illegal modes of gambling. While the survey provider noted that the use of convenience sampling may introduce some sampling bias in respect of illegal gamblers, they also noted that it is uncertain how material this is, given that a study of the underlying population of illegal gamblers does not exist.

Further research would assist in clarifying this issue. As discussed below, however, it is notable that the profile of gamblers developed through this study is consistent with that established by international work e. A Mann—Whitney U test was used in respect of differences in gambling spend. Two versions of the legal gambler results were developed: a population-weighted version, using past gambling population data and an unweighted version.

Since there was little prior indication of the actual population distribution of illegal gamblers, the unweighted legal survey results were compared to the unweighted illegal gambler results. Discussions with the survey provider also indicated that weighting did not make a major difference to the legal gambler results.

All capex investing assured

All persons entering the racecourses and all catering premises must have received vaccination in accordance with the prevailing Vaccine Pass requirement. Any person who a has been tested positive for COVID by COVID nucleic acid test or by rapid antigen test, b is currently under home quarantine for being a close contact of a confirmed case or a household member of a close contact, c has travelled outside Hong Kong and is currently subject to a compulsory quarantine order, or d is required to undergo COVID testing pursuant to a compulsory testing notice or direction except for compulsory testing required after the end of compulsory quarantine or self-monitoring for persons returning from places outside Hong Kong and has not been confirmed negative will not be admitted.

Tickets are available for sale online or at off-course betting branches. No person under 18 is allowed to bet or enter premises where bets are accepted. Special Measures The Club has implemented a series of special measures to ensure the health and safety of customers at the racecourse. The new generation digital wagering experience HKJC has been acknowledged as a world leader in wagering technology.

Under the Smart Platform strategy, the Club is moving to the next generation of its digital transformation of wagering experience by implementing a Smart Platform, it has incorporating new innovations and state-of-the-art technologies into products and services to meet customer needs. All of these steps can now be taken online. Customers are no longer required to visit the betting branches for application and are therefore not limited by branch service hours.

Experience Booking Online booking of racecourse catering venues has never been easier! A variety of venues are available to choose from, including those suited to a younger crowd and ones that cater to more seasoned racing fans.