Notes on The Impact of Liquor Restrictions on Serious Assaults across Queensland, Australia

Abstract

• Aims: Assess the relationship between a 00:00 liquor restriction (introduced 1 July 2016) and alcohol-related harm by examining its impact on serious assaults during high-alcohol hours (HAH: 20:00–6:00 Friday and Saturday nights) from 1 January 2009 to 30 June 2018.
• Location types: Two policy-exposed categories identified where liquor restriction could be isolated from other measures:
  • Safe Night Precincts (SNPs): designated high-activity NEPs with venues; some SNPs had no venues trading past 3:00 before July 2016, allowing isolation of liquor restriction effects from trading hours.
  • Local Government Areas (LGAs): areas with venues affected by liquor restrictions but no SNP; trading hours changes did not apply in these LGAs.
• Methods: Time series analysis using autoregressive integrated moving average (ARIMA) to estimate the impact on monthly police-recorded serious assaults during the two years following policy introduction, separately for SNPs and LGAs.
• Key findings: Contrary to expectations, monthly serous assaults did not significantly change within SNPs or LGAs after the liquor restriction was implemented.
• Conclusions: The Queensland liquor restriction did not produce a clear, unique reduction in serious assaults. Need to analyze liquor restrictions in conjunction with other policy changes and consider public perception and cumulative effects.
• Keywords: nightlife; night-time economy; alcohol drinking; policy; alcoholic beverages; assault.

Introduction

• Alcohol-related harm in night-time entertainment precincts (NEPs) is a major preventable community burden.
• NEPs: high-risk areas with clusters of on-licence venues (pubs, bars, nightclubs) often hotspots for violence due to high intoxication and an atmosphere of permissiveness/ anonymity.
• Harm drivers: risky alcohol consumption patterns, especially among heavy drinkers; rapid intoxication from high-alcohol-content liquor.
• Interventions often target high-risk liquors (e.g., shots, mixes with >30 mL alcohol, ready-to-drink beverages with >5% alcohol) or implemented as part of a broader policy package; evidence of effectiveness is limited.
• Shot/rapid-consumption bans have gained policy traction for simplicity and perceived validity, despite limited empirical support.
• Queensland policy context: Tackling Alcohol-Fuelled Violence Legislation Act (TAFV), effective 1 July 2016.
  • Ban on sale of liquor designed for rapid consumption or high-alcohol-content after 00:00; exemptions exist (venues primarily selling high-value spirits, seating ≤60 patrons).
  • Part of a broader package including trading hours restrictions.
  • SNPs designated to have a 3:00 trading hour close (reduced from 5:00); the rest of the state faced a 2:00 close (reduced from 3:00).
  • February 2017: restrictions on extended trading permits; July 2017: mandatory identification scanners in SNPs.
• The policy complexity (liquor restrictions + trading hours) complicates estimation of the liquor restriction’s unique effect.
• Rationale for isolated analysis: some SNPs had no venues trading past 3:00, enabling examination of liquor restrictions without trading-hour confounds; LGAs with no SNP also allow assessment of liquor restrictions as a solitary measure.
• Hypotheses (two parts):
  1) Police-recorded serious assaults per month during HAH will significantly decrease after liquor restrictions in SNPs where trading hours did not change venues past midnight.
  2) Police-recorded serious assaults per month during HAH will significantly decrease after liquor restrictions in LGAs with no SNP where trading hours did not alter venues’ hours.

Method

2.1 Data

• Data source: Queensland Alcohol-related violence and Night Time Economy Monitoring project (QUANTEM).
• Ethics: Approved by Human Research Ethics Committees of Deakin University, University of Queensland, and James Cook University.
• Licensing data: Queensland Government Office of Liquor Gaming and Regulation provided venue-level data (licence numbers, licence type, hours, extended permits, addresses).
• Police data: De-identified records of serious assault resulting in bodily harm (including aggravated non-sexual assault, bodily-fluid involvement, grievous harm, etc.).
• Night-time assaults: Used as a robust indicator of alcohol-related harm; other police incidents may reflect policing/enforcement bias.
• Inclusion criteria: Assaults within SNPs and LGAs during HAH from 1 Jan 2009–30 Jun 2018; exclude residential-property assaults.

2.2 Setting

• Licensing data (as of 30 Jun 2016) identified SNPs with no venues affected by the 3:00 trading-hour restriction; SNP details and numbers of venues serving past midnight (high-risk venues) are summarized in Table 1.
• Table 1: SNPs with closing times after 00:00 and not after 3:00 before July 2016
  • Bundaberg CBD: N Licences 26; High-Risk types: Bar 0, Hotel 7, Nightclub 0; Venues affected by liquor restriction: 7
  • Caloundra: N Licences 49; Bar 0, Hotel 3, Nightclub 0; Venues affected: 3
  • Ipswich CBD: N Licences 26; Bar 0, Hotel 12, Nightclub 0; Venues affected: 1
  • Maroochydore: N Licences 39; Bar 1, Hotel 3, Nightclub 2; Venues affected: 6
  • Mooloolaba: N Licences 55; Bar 0, Hotel 3, Nightclub 0; Venues affected: 4
  • Note: For SNP totals, some low-risk venues were also affected (e.g., a community club, restaurant).
• 15 LGAs identified where no venues were affected by the 2:00 trading-hour restriction and at least one venue affected by the 00:00 liquor restriction. Table 2 details follow.
• Table 2: LGAs with closing after 00:00 and not after 2:00 before July 2016
  • Barcaldine (N Licences 19): High-Risk: Hotel 10; Venues affected: 1
  • Charters Towers (35): Hotel 17; Venues affected: 2
  • Cloncurry (23): Hotel 8; Venues affected: 4
  • Diamantina (4): Hotel 2; Venues affected: 1
  • Flinders (8): Hotel 4; Venues affected: 1
  • Goondiwindi (35): Hotel 10; Venues affected: 4
  • Isaac (63): Hotel 11; Venues affected: 2
  • Lockyer Valley (47): Hotel 17; Venues affected: 3
  • Longreach (24): Hotel 8; Venues affected: 2
  • Scenic Rim (122): Hotel 21; Venues affected: 3
  • South Burnett (92): Hotel 24; Venues affected: 4
  • Southern Downs (159): Bar 1, Hotel 22; Venues affected: 3
  • Tablelands (60): Hotel 16; Venues affected: 3
  • Weipa (8): Hotel 1; Venues affected: 1
  • Winton (10): Hotel 5; Venues affected: 1
  • Note: Includes low-risk venues in counts.

2.3 Analysis

• Data aggregation: SNPs and LGAs categorized into two groups; each group examined for liquor restriction impact in isolation (i.e., no trading-hour effect).
• Statistical method: ARIMA time-series analysis to estimate policy impact on monthly serious assaults during the post-intervention period, capturing acute and gradual effects on trends.
• Model selection: Standard ARIMA approach; differencing used to achieve stationarity when needed; seasonality checked via autocorrelation and partial autocorrelation plots; seasonal models used if periodic trends detected.
• Transfer function: Cross-correlograms used to identify best-fitting lag for the intervention variable; liquor restriction coded as a step function (abrupt initiation and continuation).
• Software: Analyses conducted in Stata 15.0.
• Sensitivity: Additional analysis considered the 00:00 prohibition window in a broader context (e.g., 20:00–6:00 range) to test robustness; inclusion of an extra step function for mandatory scanners in SNPs (July 2017) to control for this concurrent intervention.

Results

3.1 SNPs with Venues Only Affected by Liquor Restrictions

• No seasonality detected in the SNPs model.
• Findings: No significant change in monthly serious assaults during HAH after liquor restrictions in SNPs that were not affected by trading-hour restrictions (ARIMA(0,0,1), Q = 30.93, p = 0.85).
• Outlier check: June 2018 showed a large rise in assaults; removing this outlier did not change the non-significant result.
• Figure 1: Plots serious assaults in SNPs with venues only affected by liquor restrictions during HAH (no significant downward trend observed).

3.2 LGAs with Venues Only Affected by Liquor Restrictions

• No seasonality detected in the LGA model.
• Findings: No significant change in monthly serious assaults during HAH after liquor restrictions in LGAs with no SNPs (ARIMA(0,1,1), Q = 29.84, p = 0.88).
• Figure 2: Plots serious assaults in LGAs with venues only affected by liquor restrictions during HAH (no significant downward trend).

3.3 Sensitivity Analyses

• HAH window extended analyses acknowledged that liquor restrictions act after 00:00; additional analyses across all sites tested whether serious assaults decreased between midnight and 6:00.
• ARIMA specifications for the sensitivity models matched the main HAH models; no seasonality detected in either; results showed no significant changes in monthly serious assault trends for both SNPs and LGAs (p > 0.05).
• Additional control: Mandatory identification scanners (July 2017) coded as a separate step function did not meaningfully alter results for full HAH or 00:00–6:00 models.

Discussion

4.1 Summary of Findings

• The study evaluated the unique impact of midnight liquor restrictions on serious assaults in Queensland nightlife spaces, separating them from the broader trading-hours policy.
• Two analysis domains (SNPs and LGAs) were used to isolate liquor restrictions from trading-hour changes.
• Hypotheses predicting reductions in assaults after liquor restrictions were not supported in either SNPs or LGAs.
• The lack of detectable impact persisted even after controlling for the additional intervention (scanners) in SNPs.
• Implication: Liquor restrictions, as a standalone measure within NEPs, may not reliably reduce serious assaults when implemented alongside trading-hour restrictions.

4.2 Liquor Restrictions in SNPs

• The first hypothesis anticipated a decline in serious assaults post-00:00 liquor restrictions within SNPs, especially after midnight.
• Result: No significant post-intervention reductions in HAH or post-midnight periods.
• Possible explanations:
  • Many SNPs were dominated by hotels; effects may vary by venue type, but data could not distinguish assaults by venue type due to venue proximity;
  • The SNPs analyzed represented 3.8 km^2 with 130.53 serious assaults per km^2, versus LGAs without SNPs at 134,463.7 km^2 with <0.01 assaults per km^2, indicating high local clustering that might dilute broader effects.
• The policy’s intended mechanism (reducing rapid, high-risk consumption after midnight) may fail if most harm is driven by patrons who are already intoxicated earlier in the night or who stockpile high-risk liquor before midnight.

4.3 Liquor Restrictions in LGAs without SNPs

• The second hypothesis anticipated reductions in assaults in LGAs without an SNP after liquor restrictions, particularly after 00:00.
• Result: No significant post-intervention reductions; liquor restrictions alone did not appear to meaningfully affect HAH assaults in these smaller-nightlife areas.
• Considerations:
  • LGAs were largely small towns with limited nightlife spaces; high-risk liquor policies may have smaller absolute impact where on-licence consumption is less concentrated.
  • Even when restrictions targeted high-risk liquor types, the overall nocturnal assault pattern may be driven by factors beyond late-night on-licence drinking (e.g., pre-night drinking, street-level disorder, other social factors).

4.4 Implications for Future Interventions

• Theoretical explanations for null findings:
  • Liquor restrictions operate only after midnight, potentially missing heavily intoxicated individuals who are already drinking before midnight.
  • The “shot ban” style approach may be arbitrary with respect to timing; industry exemptions (e.g., venues seating ≤60 patrons) may allow continued access to high-risk beverages post-midnight.
  • Patrons who pre-drink heavily or adjust behavior in anticipation of restrictions may undermine policy effectiveness.
  • Liquor restrictions may need to be paired with broader price-based or availability-based measures to achieve measurable harm reductions.
• Practical considerations:
  • The study’s SNPs largely comprised hotels; a different venue mix might yield different outcomes.
  • The sampling frame excluded the largest NEPs where the interaction between liquor and trading hours is strongest; results may not generalize to those precincts.
  • The study could not parse individual-level factors (e.g., offender/victim characteristics, consumption patterns) that would illuminate mechanisms.

4.5 Policy and Ethical Implications

• Liquor restrictions alone, without broader supports or consistent enforcement, may have limited impact on reducing serious assaults.
• Policymakers should consider cumulative effects and public perception of restrictions, evaluating liquor controls as part of a broader, integrated strategy rather than in isolation.

Limitations

• Concurrent policy: Trading hours restrictions were implemented statewide along with liquor restrictions, complicating isolation of the liquor restriction’s effect; only SNPs where trading hours did not change pre/post were usable for isolation.
• Venues: In SNPs, close venue proximity prevented attribution of assaults to a specific venue; heterogeneity in venue types (hotels, bars, nightclubs) may have influenced outcomes.
• Geography: SNPs had much higher assault density per area than non-SNP LGAs, which could affect detectability of intervention effects.
• Data limitations: The analysis relied on police-recorded assaults; lacked covariates on individuals’ demographics or drinking patterns; could not assess potential substitution or stockpiling behaviors.
• Sensitivity: While scanners were introduced in SNPs (July 2017) and treated as a separate step function, these controls did not alter results meaningfully.

Conclusions

• The study is among the first to isolate liquor restrictions from a package of alcohol-control measures to evaluate their standalone impact on alcohol-related harm in NEPs.
• Across SNPs and LGAs, the 00:00 liquor restriction did not produce a significant reduction in police-recorded serious assaults during high-alcohol hours in the two years following implementation.
• Given the limitations, definitive conclusions about effectiveness in different jurisdictions should be drawn cautiously; future work should consider controlled designs (with appropriate controls) and examine liquor restrictions alongside other policy changes and public perception effects.
• Overall finding: Stand-alone liquor restrictions in nightlife settings are unlikely to be associated with large reductions in serious assaults, though context, venue mix, enforcement, and public attitudes may modulate outcomes in specific locales.

Tables and Figures (as described in the text)

• Table 1: SNPs with closing times after 00:00 and not after 3:00 before July 2016
  • Bundaberg CBD: N Licences 26; High-Risk types Bar 0, Hotel 7, Nightclub 0; Venues affected by liquor restriction: 7
  • Caloundra: N Licences 49; Bar 0, Hotel 3, Nightclub 0; Venues affected: 3
  • Ipswich CBD: N Licences 26; Bar 0, Hotel 12, Nightclub 0; Venues affected: 1
  • Maroochydore: N Licences 39; Bar 1, Hotel 3, Nightclub 2; Venues affected: 6
  • Mooloolaba: N Licences 55; Bar 0, Hotel 3, Nightclub 0; Venues affected: 4
  • Note: totals include low-risk venues.
• Table 2: LGAs with closing after 00:00 and not after 2:00 before July 2016
  • List includes Barcaldine, Charters Towers, Cloncurry, Diamantina, Flinders, Goondiwindi, Isaac, Lockyer Valley, Longreach, Scenic Rim, South Burnett, Southern Downs, Tablelands, Weipa, Winton with their N Licences, High-Risk types, and Venues Affected.
• Figure 1: Serious assaults in SNPs with venues only affected by liquor restrictions during HAH (no significant downward trend).
• Figure 2: Serious assaults in LGAs with venues only affected by liquor restrictions during HAH (no significant downward trend).

References (selected topics mentioned)

• TA F V (TAFV) policy and related legislation in Queensland (2016).
• QUANTEM project protocol and final reports.
• Previous literature on the effects of trading hours restrictions and alcohol policies on assaults (e.g., Newcastle, Sydney studies).
• Methodological references for ARIMA time-series analysis, differencing, seasonality checks, and transfer-function modelling.
• Related debates on the effectiveness of high-risk liquor restrictions and the broader night-time economy policy.