Effects of a Law allowing Sunday Package Liquor Sales
On Alcohol-Related Crash Death and Injury in New Mexico

Draft, November 21, 1999

Steven Flint
Santa Fe, NM

Abstract

Historically, sales of package liquor in New Mexico on Sundays have been barred by state law. In 1995, the liquor industry actively solicited the State Legislature and new Governor Gary Johnson to amend the law to allow package liquor sales on Sunday, with success. Analysis of statewide traffic crash records show that 10.77% of alcohol-related crashes occurred in the Sunday PM/Monday before 7am period in the 36 months after the June 1995 effective date of the change, compared to 8.80% for the 36 months prior to the change, a large change corresponding to 93 extra crashes, 7 extra deaths, and 45 extra injuries annually. No significant change was observed for non-alcohol-related crashes. Results strongly support the environmental health tenet that increased alcohol availability can be a major factor in increased death and injury from alcohol abuse.

Introduction

Historically, sales of package liquor in New Mexico on Sundays have been barred by state law. In 1995, the liquor industry actively solicited the State Legislature and new Governor Gary Johnson to amend the law to allow package liquor sales on Sunday, with success. To minimize political opposition from religious groups, the law provided for liquor sales only after noon on Sundays. The new law took effect in mid-June, 1995, and outlets began offering liquor on Sunday immediately.

New Mexico has long ranked among the highest states each year for alcohol-related crashes per capita, and political leaders have recognized strong public support for reducing driving while impaired by alcohol. To help market the bill to legislators and provide political cover for those who supported it when they later faced constituents, advocates for the bill argued that in fact Sunday sales would reduce alcohol-related crash death and injury because

  1. it would reduce "bootleg" Sunday sales by unlicensed sellers who ignored Sunday sales laws and presumably ignored other prohibitions against liquor sales to minors and intoxicated persons, etc.; and
  2. it would eliminate the need of "honest citizens" to drive home impaired because the ban on package sales forced them to buy liquor on Sunday by the drink at bars, for on-site consumption rather than buying for transport to be consumed at home.

Public health advocates opposed the measure to expand liquor availability on the basis of the environmental health argument that increased availability would translate to increased DWI and increased alcohol abuse and secondary effects of that abuse more generally. They recognized, however, the complexity of potential effects, such as reduction in the consumption of alcohol in hair spray by extreme alcoholics with no access to package sales on Sunday, or the reduction of alcoholics consuming on Saturday night the Sunday stash they brought in preparation for the dry Sunday to come.

In the wake of the change in availability of liquor law, police records on crashes in New Mexico offer a means to measure effects of the change on alcohol-related crashes. Rarely in US history have record systems allowed clear measurement of secondary effects of alcohol availability changes, so this change in law and availability is nearly unique in offering a means to test the environmental health hypothesis that increased alcohol availability will translate to increased impaired driving and therefore to increased death and injury on the highways.

Method

New Mexico police traffic crash records capture police reports on all crashes on New Mexico public roads that involve death, injury, or $500 or more property damage. Records are available in consistent form since 1982, and they include a wealth of information about each crash.

A change to allow sales on Sunday after noon should have an effect on Sunday consumption after noon and on Mondays prior to 7am, when liquor stores reopen, but it should have no effect on crashes on other days of the week or on Sundays before noon and Mondays after 7am. New Mexico's other initiatives against DWI may also had some progressive effects during the time period involved, but none should have produced an abrupt "step" effect in June, 1995, and none should have had a differential effect on Sunday afternoons and nights but not other times of the week.

To measure effects, then, I obtained counts of crashes of alcohol-involved crashes and alcohol-involved fatal-injury crashes for Sunday noon-Monday 7am (SP=sales period) and Monday 7am-Sunday noon (CP=control period), by 12-month period before and after the June, 1995, change. To remove effects of unrelated anti-DWI measures in the period, the percentage (A%) of SP alcohol-related crashes of (SP+CP) alcohol-related crashes was calculated. To consider any potential changes in total crash distribution between Sundays and other days of the week, the percentage (N%) of SP non-alcohol-related crashes of (SP+CP) non-alcohol-related crashes was also studied.

Results

Table 1 shows alcohol-related and non-alcohol-related SP and CP counts and percentages for three year periods before and after the change in availability.

Table 1 -- All Crash Counts and Percentages in 36-month intervals Before and After Legalizing Sunday Package Sales





Period


Sunday
Period Alcohol-Related


Control
Period Alcohol-Related



SP % Alcohol-Related

Sunday
Period Not Alcohol-Related

Control
Period Not Alcohol-Related


SP %
Not Alcohol-
Related

Before

1,536

15,916

8.80%

9,558

134,700

6.62%

After

1,548

12,824

10.77%

10,428

142,498

6.82%

The change in the SP% Alcohol-related percentage is statistically significant at the 0.001 level, and indicates that 283 additional alcohol-related crashes happened on Sundays in the After period than would have happened had the prior 8.80% level continued. The change in the SP% Not Alcohol-Related is much smaller in magnitude (0.2%) and is barely statistically significant (p<.05).

Police indicate whether alcohol is related to a crash based on the citation of any driver in the crash for DWI or clear indication of alcohol impairment for any driver or pedestrian in the crash. Some studies have suggested the alcohol impairment is better reported by police for fatal and injury crashes, since officers may find the burden of processing an arrest or answering questions about why they made no arrest excessive so they do not check for or record indicators of alcohol involvement. If the increase in % SP alcohol-related crashes is due to some change in reporting propensity for Sundays occurred because officers know of the availability change, one would expect that change to be less prevalent for crashes where reporting was already good, ie, fatal and injury crashes. Table 2 shows alcohol-related and non-alcohol-related SP and CP fatal and injury crash counts and percentages for three year periods before and after the change in availability.

Table 2 -- Fatal and Injury Crash Counts and Percentages in 36-month intervals Before and After Legalizing Sunday Package Sales





Period


Sunday
Period Alcohol-Related


Control
Period Alcohol-Related



SP % Alcohol-Related

Sunday
Period Not Alcohol-Related

Control
Period Not Alcohol-Related


SP %
Not Alcohol-
Related

Before

854

8,646

8.99%

3,441

47,795

6.72%

After

917

7,318

11.13%

3,898

51,657

7.02%

The change in the SP% Fatal and Injury Alcohol-related percentage is statistically significant at the 0.001 level, and indicates that 176 additional fatal and injury alcohol-related crashes happened on Sundays in the After period than would have happened had the prior 8.99% level continued. The change in the SP% Fatal and Injury Not Alcohol-Related is much smaller in magnitude (0.3%) and is not statistically significant (p>.05).

The hypothesis that the change in SP% alcohol-related crashes is related to the availability of Sunday sales implies a step change in the SP% value as opposed to a gradual change. To measure this, Table 3 presents alcohol-related and non-alcohol-related SP and CP crash counts and percentages for the three twelve month periods before and after the June 1995 change in availability.

Table 2 -- All Crash Counts and Percentages in 3 12-month intervals Before and After Legalizing Sunday Package Sales





Period


Sunday
Period Alcohol-Related


Control
Period Alcohol-Related



SP % Alcohol-Related

Sunday
Period Not Alcohol-Related

Control
Period Not Alcohol-Related


SP %
Not Alcohol-
Related

25-36 mo before

557

5,487

9.22

3,270

44,855

6.79

13-24 mo before

485

5,233

8.48

3,127

43,408

6.72

0-12 mo before

494

5,196

8.68

3,161

46,437

6.37

0-12 mo after

599

4,801

11.09

3,509

48,127

6.80

13-24 mo after

501

4,188

10.68

3,542

48,428

6.82

25-36 mo after

448

3,835

10.46

3,377

45,943

6.85

The change in the SP% Alcohol-Related value is clearly an abrupt step, not a ramp change, while little change is apparent in the SP% Not Alcohol-Related Value. The smallest change between any two before-after pairs in the SP% Alcohol-related percentage is statistically significant (25-36 months prior compared to 25-36 months after) at the <0.05 level. The change between the smallest SP% Not Alcohol-Related and the After values is much smaller in magnitude (0.48%) and is statistically significant (p<.01), but the Changes between the other Before values and the After values are not (p>.05).

Figure 1 graphs the percentage values in the above table, showing the Step very clearly.

Figure 1: %SP Crashes by Year

Discussion

The data strongly support the hypothesis that 1995 legislation that allowed sales of package liquor on Sunday afternoons and evenings starting in June 1995 produced a statistically significant increase in alcohol-related crashes in the hours from Sunday at noon to Monday at 7am of each week. While broader alcohol-related crash trends produced dramatic gradual decreases in alcohol-related crashes over the period, Sunday sales period alcohol-related crashes showed a statistically significant increase. The increase of 23% compared to Sunday proportions prior to the change corresponds to 93 extra alcohol-related crashes annually, corresponding to about 7 extra deaths and 45 extra persons injured annually according to severity proportions typical of New Mexico alcohol-related crashes.

The data provide strong support for the environmental health tenet that alcohol availability is one major factor in risk of injury caused by alcohol abuse.