WEB SITE INFORMATION: VERMONTADVOCATE.COM

Act 58 apparently has a weakness that should be addressed with clarification language specified in H.293 for the 2015-2016 session.

 

This bill can be observed at the following location:

 

http://legislature.vermont.gov/assets/Documents/2016/Docs/BILLS/H-0293/H-0293%20As%20Introduced.pdf

 

Many thanks to the primary sponsors and the 11 co-sponsors for their support of this bill.

 

 

This bill will not change the current meaning or intent of this law – it will only clarify its compliance requirements related to “the prohibition on sales to minors” and define a specific mechanism for responding to DLC failure to satisfy the “as frequently and as comprehensively as necessary” requirement.

The Department of Liquor Control (DLC) must be persuaded that in its effort to enforce the prohibition on sale of tobacco products to minors, it has an obligation to comply with the letter and the intent of the following law, which was part of Act 58, 1997:

 

 “Sec. 13.  COMPLIANCE TESTING; TOBACCO SALES TO MINORS

(a)  The Department of Liquor Control shall conduct or contract for compliance tests of tobacco licensees as frequently and as comprehensively as necessary to assure consistent statewide compliance with the prohibition on sales to minors of at least 90 percent for 17-year old buyers.  An individual under the age of 18 participating in a compliance test shall not be in violation of 7 V.S.A. § 1005.

(b)  Any violation by a tobacco licensee of  7 V.S.A. §§ 1003(a) and 1007 after a first sale violation or during a compliance test conducted within six months of a previous violation shall be considered a multiple violation and shall result in the minimum license suspension in addition to any other penalties available under this title.  Minimum license suspensions for multiple violations shall be assessed as follows:

(1)  Two  violations     One weekday

(2)  Three violations    Two weekdays

(3)  Four violations     Three weekdays

(4)  Five violations     Three weekend days, Friday through Sunday”

 

In December 1994/January 1995, when the DLC conducted its initial tobacco compliance tests, compliance was found to be only 36% - literally off the chart. Eight months later, compliance had risen to a remarkable 74% simply because licensees had learned that compliance testing would be done.

Of all the factors that might affect compliance, testing frequency has the most reliable and strongest correlation and is by far the most cost-effective.

This law will significantly reduce the youth addiction rate if it is actually obeyed by the DLC.

 

The DLC has essentially disregarded the requirements of this law for the entire nineteen years that it has been in effect. The net effect of this disrespect of the law has been and still is more than 300 additional youth addictions to nicotine every year. That’s a loss of more than 5,600 youth to drug addiction so far and the total is continuing to increase at the rate of about one youth per day.

What does Vermont’s tobacco compliance-testing law require?

Analysis by Bruce Cunningham

September 3, 2007 (with some updates for current data)

The following Vermont law prohibits the sale of tobacco products to minors:

TITLE 7, CHAPTER 40, § 1003. Sale of tobacco products; requirements; prohibitions

(a) A person shall not sell or provide tobacco products to any person younger than 18 years of age.

Vermont also has a related law, Section 13 of Act 58, 1997:

COMPLIANCE TESTING; TOBACCO SALES TO MINORS

(a)  The Department of Liquor Control shall conduct or contract for compliance tests of tobacco licensees as frequently and as comprehensively as necessary to assure consistent statewide compliance with the prohibition on sales to minors of at least  90 percent for 17-year old buyers. 

The following chart shows the Department of Liquor Control’s (DLC) data for compliance enforcement testing (a proxy measurement related to the actual requirement of the law) results from 2007 to date. Although 2007-2016 was the best period so far, this apparent improvement in performance happened during a period when the rate of compliance testing (tests per month) was dramatically decreased by more than 50% with a resulting increase in the margin of error from about +/- 2 to about +/- 3. Vermont statute requires “at least 90%” which means the lower end of the margin of error must be above 90%. All of the research nationwide since the Synar law was effected in 1997 indicated that this data is suspect and should not be considered significant unless a rational explanation (or even a reasonable hypothesis) can be provided to explain this anomaly. The DLC has nothing to offer supporting this data   (It should be noted that the divergence of the trend lines from the middle of 2008 to the middle of 2009 is an anomaly that can only be explained at this point by an increase in the difference between the DLC proxy measurement testing results and the actual compliance with the prohibition on sales to minors that the law requires.)

 

It is clear that the Act 58 requirement has never been met and unless something changes, it never will.

 

 

Casual (or careless) reading of Act 58 may result in a misinterpretation that would suggest we have almost achieved the level of protection (90%) from drug addiction for our youth that the legislature intended and the law requires. Unfortunately, we have not. It is important to read the law carefully and consider its specific requirement and its fundamental intent, which is to assure that when a typical 17-year-old anywhere in the state at any time attempts to purchase tobacco products, he or she will be unsuccessful at least 90% of the time.

One of the most significant causes of the continuing failure to comply with Act 58 is the DLC’s apparent persistent failure to understand what the language of Section 13 of Act 58 actually requires. I am asking the Legislature to help resolve this fundamental misunderstanding. The first step toward that goal is taking a close look at the language of the law.

My analysis of the language of Section 13 of Act 58 of 1997 is based on the concept, which is consistent with the way the courts interpret laws, that the words in the law were chosen carefully and deliberately and that all of them have meaning. In regard to this law, I am confident that this concept is valid because I was personally involved in the initial drafting and all of the rewriting that led to the final language and I attended every committee meeting in that process where that meaning was discussed. 

The Act 58 requirement that gives it the power to significantly reduce minor drug addiction is contained in the sentence: The Department of Liquor Control shall conduct or contract for compliance tests of tobacco licensees as frequently and as comprehensively as necessary to assure consistent statewide compliance with the prohibition on sales to minors of at least  90 percent for 17-year old buyers.  I’ve italicized several critical phrases.

First, prohibition on sales to minors”: Act 58 itself does not prohibit such sales – this phrase refers to the prohibition which is in the first law noted above, TITLE 7, CHAPTER 40, § 1003. The most serious misconception is the notion that the 90% requirement is for compliance test results rather than for the prohibition on sales to minors. Compliance tests are not a measurement of actual compliance with § 1003. Many research studies have shown that real-life attempts to purchase are much more likely to be successful than attempts in compliance tests, which are necessarily conducted according to protocols that significantly alter both the circumstances of the tests and the behavior of the buyers. For example, one study showed buying attempts by smokers to be about six times more likely to succeed than those of non-smokers. (The Vermont DLC protocol uses only non-smokers.)

Compliance tests are a “proxy measurement” – i.e. a measurement that is different from but has a positive correlation with the desired objective of greater than 90% real-life compliance with the prohibition. License suspensions for compliance test failures are used to motivate improved licensee compliance with the prohibition on actual sales to minors and assure the positive correlation.

Act 58 requires the use of compliance tests as a tool to drive actual compliance with the prohibition because the strong correlation between compliance-testing frequency and actual real-life compliance was well known before 1997 when the law was written.

Control tests (measurements without the artificial restrictions of the DLC compliance-testing protocol) are needed to determine the actual compliance with the prohibition on sales to minors. This has, so far, not been done in Vermont so we don’t know how far we are from meeting the 90% requirement – but research shows we can be sure that actual compliance is significantly lower than the compliance-test results. (My personal educated guess is that actual compliance with the prohibition on sales is no higher than about 75%.)

Second, at least”: This phrase was included in the law because there are error margins to be considered in both the compliance tests (proxy measurement) and the control tests. To assure the intended protection for our youth, the lower net error limit must be above 90%. This is necessary because there are practical financial limits to sample sizes in both measurements. Consideration of the errors assures the intended drug-abuse prevention results regardless of sample sizes. Competent professional statistical analysis is required to evaluate performance in relation to the requirements of both laws noted above.

To illustrate the effect of the margin of error on the meaning of the reported data, consider an extreme example: performing only one compliance test in an environment where actual overall compliance is about 90%. The one test might be a pass and it might be a fail, so the reported outcome will be either 100% or 0% compliance. If the test is a pass, we can only say that compliance is at least more than 0%. If the test is a fail, we can say only that compliance is between 0% and 100%. It follows that the more testing is done, the more we can reduce the uncertainty of our conclusion. Statistical analysis of current compliance-testing data indicates that the error margin for the reported compliance testing is approximately +/- 2%.

Third, “consistent statewide”: To satisfy the intent that all minors in the state be equally protected all the time, compliance with the prohibition needs to be consistent both geographically and in time. This is, of course, not an absolute. Similar to the “at least” criterion discussed above, common sense needs to be applied in deciding how consistent is consistent enough. What we do not want is any significant segments of the population at risk for any significant periods of time. For example, if one county has real-life compliance below 90% for two months in a row, an increase in testing frequency in that county should be applied until the problem is solved.

Fourth, as frequently and as comprehensively as necessary”: The first draft of this law did not contain this phrase. Instead, it said “… conduct compliance tests of all tobacco licensees in the state at least three times per year…” without requiring 90% compliance. This frequency was specified because prior research had shown that that a positive outcome has a stronger correlation to testing frequency than any other contributing factor, and that three times per year per licensee would probably be enough to produce the desired result of 90% compliance. It was recognized in committee that it might require a higher or lower testing frequency in Vermont to achieve the intended results so the current phrase was inserted instead to make it clear that the DLC should promptly increase testing frequency whenever it failed to meet the “at least 90%” ultimate objective. The Liquor Control Board acknowledged this requirement in March, 2004 but the DLC never honored the board’s commitment made at that time to increase testing frequency by 70%. Refusing to respect this phrase is the primary cause of the DLC’s eight-year failure to comply with this law.

Finally, note that “for 17-year old buyers” does not apply to the compliance testing – it applies only to the actual real-life compliance with the prohibition by tobacco licensees as may be determined with control testing. The DLC is free to use any age buyers in its compliance testing with the understanding that more frequent testing would probably be needed with lower age buyers to achieve the required actual real-life compliance for 17-year-olds as measured in control tests.

 

CONCERN FOR LICENSEES

It should be noted that the retailers (tobacco licensees) will not be hurt if the DLC does what the law requires.

Every licensee who is making a good faith effort to comply with the law wants to know if they are making mistakes so that they can take action to improve their training and procedures. As the DLC gradually increases its compliance testing frequency, the licensees will keep pace with gradual improvement in their compliance with the prohibition on sales to minors due to the more frequent attention given to their performance. They will not be punished more frequently unless they fail to improve their performance to meet the 90% requirement of the law.

 

 

ANALSIS OF COST-EFFECTIVENESS OF COMPLYING WITH SECTION 13 OF ACT 58 COMPARED WITH OTHER PREVENTION STRATEGIES IN THE VERMONT COMPREHENSIVE TOBACCO CONTROL PROGRAM

Updated 9/28/07

ASSUMPTIONS:

  1. The Tobacco Control Program Budget is split roughly 50/50 between prevention and cessation efforts.
  2. Almost all prevention work succeeds before age 19 if it succeeds at all
  3. Most 12th graders are 17 years old when starting and some are 18 by the time they graduate
  4. The TCP comprehensive programs are responsible for the 22% decrease in 12th grade smoking from 1995 to 2005
  5. Current compliance testing produces commercial availability that is higher than the <10% required to significantly reduce youth prevalence

ESTIMATES:

  1. current 12th grade population in Vermont is 9,117
  2. new DLC estimate of cost of compliance testing in Vermont is $53 per test (with insufficient testing to achieve the <10% commercial availability required by law)
  3. 3,000 tests per year will probably produce <10% commercial availability
  4. conservative estimate from Joe DiFranza: if actual commercial availability is maintained at less than 10%, youth prevalence will be reduced by 25%

CALCULATIONS:

50% of $5 million budget = $2.5 million

current preventions = 22% of 9117 = 2006 preventions per year

$2,500,000 / 2006 = $1,515 per prevention for comprehensive TCP prevention

3,000 tests x $53 = $159,000 = cost of compliance testing to comply with the law (this includes the current level of about 1,500 tests per year)

25% reduction in current 12th grade smoking = 25% of 23% of 9117 = 524 additional preventions per year if we comply with the current law

 

$159,000 / 524 = $303 per additional prevention for complying with Act 58 (4.1 times the cost-effectiveness of the average for other prevention measures)

 

CONCLUSION:

Adjusting training/testing spending priority to be closer to all higher performing states would provide more than enough funds within current allocations of MSA revenue to fully comply with the law. All assumptions and estimates were intended to be conservative but even if they are off by a total factor of 4, enough compliance testing to comply with the law is still clearly the most cost-effective strategy for reducing youth addiction to tobacco.

 

I. Short history of compliance testing since Act 58 became law

In 1998, the Department of Liquor Control chose to “contract for compliance tests” as the law allowed. The contract was for 3,000 tests per year because it was known at that time that testing every licensee three to four times per year would produce at least 90% compliance. By the time eight months of the contract had been completed, 2,007 tests had been completed at a cost of $20,034 and compliance for that period was 92%. This eight-month period was in full compliance with Sec. 13 above.

The DLC terminated the contract for compliance testing in September 1998 and started doing the testing with its own personnel in 1999, doing fewer than 1,300 tests per year since that time. Average annual compliance fell below the mandatory 90% in 1999 and until late in 2004 did not rise to the minimum required by law again.

 

II. Interpretation of compliance testing data

The DLC typically points to the raw annual average compliance as an indication of performance even though the law specifically requires “consistent statewide compliance … of at least 90 percent”. The annual average is not a valid standard because significant inconsistency can exist during a year. For example, if compliance is below 90% for part of the year and above for the rest with an average of 90%, this is not complying with the law. What is needed to realistically evaluate performance in comparison to the actual requirements of the law is some basic statistical analysis.

A competent statistician will recognize that calculation of error margins is needed to analyze these data. For example, if you consider the raw annual averages of 82% and 86% for 2001 and 2002 without calculating error margins, it might look like a significant improvement. But the error margins for both years are approximately ±2%. This means that the 82% could be as much as 84% and the 86% could be as little as 84% - i.e. there is no statistical significance to the apparent change.

The solid line on the graph clearly shows that compliance peaked at approximately 86% around the beginning of 2002 and did not increase again until April 2004 when the DLC increased the frequency of compliance testing by 70%.

 

III. The penalty for failure to comply with Act 58

Act 58 specifies no penalty for violation. The negative impact of the DLC’s failure to comply falls on Vermont’s youth. The current rate of Vermont minors becoming addicted to tobacco is approximately four per day. Because of the gateway effect of tobacco addiction, some of these lead to addictions to other drugs such as marijuana, cocaine, and heroin as well.

Solid scientific evidence indicates that full compliance with Act 58 by the DLC could reduce the tobacco addiction rate to about two per day within two years. We may have missed the opportunity to prevent about 700 addictions per year for the past four years. It is past time to put an end to this continuing failure.

The most recent empirical information supporting the relationship between compliance and the addiction rate comes from the 2003 Vermont Youth Risk Behavior Survey (YRBS). One of the charts in the published data on Page 32 is titled “Percent of students who smoked cigarettes on one or more days during the past 30 days” with trends from 1985 to 2003 shown on page 33. These data indicate that smoking actually increased by 3% (from 30% to 33%) for 12th graders during the past two years while smoking decreased for grades 8 through 11. (A similar anomaly is shown for marijuana use.) Vermont compliance has been less than 90% for 12th graders in the past five years while it has been above 90% for the lower grades.

The Maine YRBS shows smoking decreased 4.4% (from 30.9% to 26.5%) for 12th grade and this was similar to decreases in their lower grades. Maine’s compliance has been above the “threshold” of 90% using 17-year-old buyers for the past five years. Maine has therefore been providing more protection for 12th graders by meeting the requirements of Vermont’s law even though Maine is only required to comply with the weaker Synar Amendment requirements of 80% using 15- to 17-year-old buyers.

Suppose that 12th grade smoking would have decreased by about 4% in Vermont just as it did in Maine if we had complied with our own state law. The implied net effect of this would have been approximately one less new addict per day in Vermont for the last two years. It would have cost us no more to do this.

 

IV. The DLC’s excuses for failure to comply with Act 58

The DLC has offered many excuses for not complying with Act 58 since 1998. These excuses are generally part of an overall theme of “we’re doing the best we can with what we’ve got”. The theme and the individual excuses are in conflict with reliable, documented, factual information.

The bulleted items below are from a document submitted by the DLC to the Senate General Affairs Committee in 2002 as testimony against a bill that would have required increasing compliance testing by only 5% for the following month whenever results fell below 90% for the current month. Following the bulleted items are my comments regarding why these are not valid reasons for continuing to violate Act 58.

Ø  DLC claim: “Act 58 of 1997 authorized a 90% compliance rate.”

Act 58 did not “authorize” 90% - it required it.

Ø  DLC claim: “90% compliance is unrealistic in today’s business climate with high employee turnover, a key reason for the fluctuation in percentages.”

90% is not “unrealistic” in terms of achievability. Nine other states had compliance rates higher than 90% in 2002. 

The U.S. Department of Health and Human Services Center for Substance Abuse Prevention publishes annual reports titled “State Synar Non-Compliance Rate Table” showing the non-compliance rates for all states. The DLC report to the Legislature for 2002 indicated compliance of 85.7%. There were 21 states that exceeded that level. Nine of those 21 states have compliance higher than 90%. The top state is South Dakota with compliance 95.5%. South Dakota received a Synar award in 2002 for the quality and consistency of its compliance testing and enforcement program. It is noteworthy that all tobacco compliance testing in South Dakota is under contract with Tom Radecki’s company in Illinois – the very same company that did the testing for Vermont in 1998 when 2,000 tests were done in eight months producing an average compliance for that period of 92%. South Dakota recently received a Synar award for the quality of its testing.

Compliance rate is directly related to the frequency of testing. Mandatory license suspensions for multiple violations make consistent compliance a condition of continuing to sell tobacco products. Testing every licensee approximately three times per year essentially assures at least 90% compliance.

90% compliance is also realistic in terms of available resources. The DLC now receives $319,000 per year from tobacco settlement funds specifically for tobacco enforcement. Three thousand tests per year will produce more than 90% compliance, and this could cost as little as $61,000 per year with a contractor performing the tests.

Half of the tobacco licensees tested in the last five years have never failed a compliance test and they have the same high (approx. 60% per year) employee turnover rate as those licensees who fail the tests.

Ø  DLC claim: “Calendar 2001 compliance was 82%, a 5% increase over the previous year.”

But that 82% was only a 3% increase over the 79% achieved in 1999. Neither change is statistically significant. The year-to-year variations seen so far are within the margin of error for this kind of measurement. The fact is the law requires “at least 90%”, and that is what is needed to accomplish what the law was intended to accomplish – i.e. a significant reduction in the rate of youth addiction to nicotine. Any improvement is still failure until 90% is exceeded.

Ø  DLC claim: “Most Control States like Vermont have not achieved even 80% compliance.”

This is not true. According to the 2002 Synar report, 38 states have achieved compliance levels higher than the 80% required by the federal law. In 2001, 32 states exceeded 80%. Other states do not have a law requiring 90% compliance. The federal “Synar Law” threatens states with a reduction in federal funds for drug abuse programs if they are not making progress toward 80%. The fact is some other states try to do only what the law requires them to do – and some of them fail. The failure of some other states to comply with a weaker federal law is not a valid excuse for our failure to comply with our own law.

Ø  DLC claim: “Nationally, compliance averages less than 80%.”

Even if this were true, it would not excuse our failure. Our job is not to be above average – it is to comply with our law and protect our youth from drug addiction.

Ø  DLC claim: “Despite dramatic decreases in the sales of tobacco to minors, youth usage has decreased only 5% nationally.”

The DLC has apparently misinterpreted the fact that there has been a dramatic increase in compliance to mean that there has been a dramatic decrease in sales to minors. This is not true.

DLC records show that about 50% of licensees tested over the last five years have never failed a compliance test. With an overall compliance average of about 85%, this means that the average compliance rate for the other half of the licensees is only about 70% - and in this group, some are above 70% and some below.

Sales to minors have remained high because the minors know very well which stores will sell them cigarettes and they go to those stores to buy them. It is only when compliance remains consistently at 90% or more for all stores that it becomes difficult enough to buy to become a deterrent to becoming a nicotine addict. This is exactly why the Vermont law requires “at least 90%” and that it be “consistent” and “statewide”. This is also why the law includes mandatory license suspensions for multiple failures – to either force improvement or weed out the worst stores.

Ø  DLC claim: “We must keep the pressure on illegal sales, but also look at the social sources of tobacco as well as the access children have to sources of tobacco.”

There have been many surveys that report on where minors obtain cigarettes. It is very important to understand that the data reported in these surveys is about where current addicts get their cigarettes. That information is not relevant to consideration of the need for 90% compliance because the primary purpose of this law is prevention, not intervention.

What is relevant is where non-smoking minors are likely to get cigarettes in the process of becoming addicts, and it is not possible to conduct a survey to determine this. It is, however, reasonable to assume that non-smokers are less willing than addicts to steal, use false identification, ask others to violate the law, or beg for handouts to obtain cigarettes (the “social sources” indicated in this DLC quote). When all stores comply with the law at least 90% of the time, this makes it difficult enough to obtain cigarettes to discourage about half of the potential smokers from becoming addicted.

Ø  DLC claim: “Department resources are spread thin. DLC currently oversees alcohol, tobacco compliance, education, licensing, enforcement, public outreach, START, etc.”

As noted above, the DLC receives $319,000 per year from tobacco settlement funds for tobacco enforcement. Sec. 13 of Act 58 does not mention education in its requirements. It only requires that the DLC “conduct or contract for compliance tests of tobacco licensees as frequently and as comprehensively as necessary”. Most of the tobacco money (about 85%) is now spent on the DLC education program that existed before the department had any tobacco enforcement responsibilities. While the education program now has some added tobacco content, the effort to comply with the specific compliance testing requirements of Act 58 should be fully funded before these funds are diverted to the education program that is only indirectly related to this critical and cost-effective primary enforcement responsibility.

Ø  DLC claim: “A 90% compliance rate is not sustainable. Contracting out this process has caused problems in the past.”

With tobacco Master Settlement Agreement funds expected to be available indefinitely in sufficient amount to fully fund enough compliance testing to assure at least 90%, there is no rational basis for the claim that this would not be “sustainable”. The DLC tried contracting for the compliance testing for less than one year and experienced some minor problems in this short period that are easily solvable. A similar contract today would cost only about $61,000 to fully comply with the law. The DLC could do sufficient testing for about $80,000. The fact is either option, contracting or DLC testing, is well within the resources provided for the DLC now and in the future.

 

My overall comments regarding these DLC excuses:

If, as I have suggested, the DLC’s comments above are not sufficient reason for failing to comply with the requirements of Act 58, why do we still have this problem? I think the answer to this can be summed up in one very familiar phrase: “It’s a dirty job but somebody has to do it.” Compliance testing certainly is a “dirty job”. That’s why the media, the licensees, the lobbyists, and even the DLC call the tests “stings”, which they are not.

It is very tempting to believe that an education program is a reasonable alternative to compliance testing to achieve the desired result. Unfortunately, that is not the case for two reasons: it is not cost effective and it cannot produce the results required by the law with any reasonable amount of spending. 

The DLC’s tobacco education program became mandatory for everyone selling tobacco on June 21, 2002. The graph above clearly shows that no improvement in compliance is shown since that time – partly because about 75% of sellers were already receiving the required training before that date, and partly because the effectiveness of education compared to compliance testing is naturally limited by several factors. Regarding cost, sufficient compliance testing to fully comply with Act 58 can cost about $60,000 (the contractor producing 92% compliance in South Dakota can do 3,000 tests per year in Vermont for $61,000), while an education program that still fails to comply with the law may cost more than five times that amount. 

The Vermont Grocers Association and the DLC lobbied strongly but unsuccessfully against Sec. 13 of Act 58 in 1997. Then-Commissioner Norrie Hoyt recommended to every committee that considered it that Sec.13 be eliminated altogether. The DLC has striven to achieve the 80% compliance required by federal law to avoid losing federal funds for drug abuse prevention programs – a penalty that would be imposed under the Synar Law. But there is no direct financial “penalty” to the department or the state for failure to meet the requirements of Sec. 13. The only “penalty” for Vermont for failing to comply with Sec. 13 is a continued nicotine addiction rate for minors that is approximately twice what it could be with consistent statewide compliance of at least 90%.

The political reality in Vermont is that a failing education program is more tolerable to tobacco licensees than sufficient compliance testing to comply with the law. The Vermont Grocers Association is adamantly opposed to any increase in compliance testing and many legislators are afraid to take the heat, either in lobbying or votes, that may result when the DLC obeys the law. Those licensees who fail to obey the law prohibiting tobacco sales to minors now complain loudly about the tests that they fail - and that complaining will escalate out of proportion to any increase in testing. But the Legislature decided to do the right thing in 1997, in spite of this political reality, because protecting our young people from drug abuse was, and still is, worth whatever political cost may result. We should have the courage now to respect that decision and respect the law. The complaints will quickly diminish once such respect becomes a normal part of the way we do business in Vermont.

 

V. Request to the Liquor Control Board for compliance with Act 58

On March 26, 2003, I attended the meeting of the Liquor Control Board to ask them to comply with the law (transcript of my testimony attached). Chairman Dan O’Brien responded with a letter to me dated April 8, 2003 (attached) that contained the following statement:

“While the Department of Liquor Control agrees that Act 58 requires a ninety percent compliance rate by tobacco licensees, the Liquor Control Board strongly believes that the overall objective of that legislation is being met. In the Board’s opinion, the overall objective is to increase the rate of compliance among tobacco licensees, thereby limiting the availability of tobacco products to underage customers.”

By this LCB definition of the “overall objective”, any consistent compliance above the 77% reported in 1996 would be satisfactory. It is not meaningful to measure performance against so low and vague a standard. Fortunately, we do not need to interpret a standard because Act 58 defines that very well. The language of Sec. 13 is deliberately specific in requiring “consistent statewide compliance with the prohibition on sales to minors of at least 90 percent” because in 1997 the Legislature decided that anything less than 90% would not accomplish the actual objective, which was to significantly reduce the rate at which Vermont youth are becoming addicted to nicotine. The Legislature recognized that that objective could not be accomplished at any level below 90%.

My response to Mr. O’Brien’s letter (attached) contained the following statement:

“Whatever the Liquor Control Board’s or my opinion is regarding the objective of Sec. 13, this section of the law has specific requirements. If these specific requirements are not met, the law is being violated. It is clear that the Department of Liquor Control has not met these specific requirements for the last four years.”

It is my understanding that no individual, or board, or department has the prerogative to disregard the specific requirements of a law because they think they are satisfying the overall objective of that law in another way.

VI. Skepticism regarding effectiveness of Sec. 13 of Act 58

I understand that there are some people who are skeptical about the expected reduction in the youth addiction rate if the DLC complies with the law. That same skepticism was part of the deliberations that led to this law being enacted in 1997. There are two important things to consider regarding this skepticism. First, there are no guarantees. This is an area of behavioral science. Many factors are involved in individual pathways to addiction. Many aspects of Vermont’s tobacco control program are done on faith that is based on predictions, and the predictions are based on good scientific studies. It is simply not possible to prove that this law will significantly benefit our young people. The 1997 Legislature and the Governor dealt with this same skepticism and decided that making this law was the right thing to do. Now we need to do what the law requires and see what the results are for our youth.

Second, there are many good scientific studies that support the expectation that complying with Sec. 13 will significantly reduce the youth addiction rate. Dr. Leonard Jason at DePaul University, one of the top experts on this subject, has published Preventing Youth Access to Tobacco, a compilation of seven reports of such studies. An e-mail from Dr. Jason to the Vermont Liquor Control Board regarding the importance of 90% compliance is included in attachment 1. Dr. Jason’s research was the foundation of Congressional considerations that led to enactment of the federal Synar Amendment in 1992 requiring compliance testing in every state. Senator Durbin of Illinois and others are now considering proposing legislation that would change the Synar law to require 90% compliance nationwide (attachment 4) for the same reasons that it is important to comply with our law now in Vermont.

VII. Conclusion

Thank you for your attention to and interest in solving this problem. Responsible and effective enforcement of Act 58 by the DLC will significantly reduce the rate of youth addiction to nicotine and other drugs.

Please contact me if you would like to have any more details. I also have contacts with several national experts on this issue and I would be happy to facilitate further communication with them.

Bruce Cunningham

 

 

 

 

Attachments: (the web site can be used with a click of the mouse if you read this on my web site)

1.    My testimony to the Liquor Control Board 3/26/03 http://www.vermontadvocate.com/act58test.htm

2.    LCB response to my testimony 4/8/03 http://www.vermontadvocate.com/llcb.htm

3.    My response to the LCB 4/14/03 (the LCB has not answered this letter) http://www.vermontadvocate.com/lobrien.htm

4.    Excerpt from 1999 Christian Science Monitor article regarding Synar Law http://www.vermontadvocate.com/CSM99.htm

5.    Governor Douglas’s response to my request for his help in solving this problem. My request was in the form of a letter which contained essentially all of the information in the document above. http://www.vermontadvocate.com/lgovresp.htm

6.    Analysis of Vermont’s 2001-2002 addiction rate problem http://www.vermontadvocate.com/12vs11.htm

 

 

THE 2012 DLC TOBACCO COMPLIANCE TEST REPORT

WHAT IS NOT TRUE or HALF TRUE, and WHAT THEY DIDN’T TELL YOU

Analysis by Bruce Cunningham - January 2013

Special note: This is the 10th analysis of the DLC annual compliance test report that I have written and submitted to the DLC. All of the reports since the first one in 2003 have been relatively easy to write because there has been little difference in the reports from year to year. They are pretty much equally flawed.

I keep hoping that telling the truth will make a difference and I keep being disappointed. Maybe this year will be different. (I said the same thing the last four reports with no effect)

Summary:

For the 15th straight year since Act 58 became effective in 1997, the Department of Liquor Control has violated the law by failing to “conduct or contract for compliance tests of tobacco licensees as frequently and as comprehensively as necessary to assure consistent statewide compliance with the prohibition on sales to minors of at least 90 percent for 17-year-old buyers. “

It should be noted that the “prohibition on sales to minors” cited in Act 58 is not about the DLC compliance enforcement test results, which are only a proxy measurement related to sales in real life (commercial availability). Actual compliance with the prohibition of sales to real-life minors who are not involved in a DLC test is significantly lower than the test results noted in this report – probably 30-40% lower.

The DLC’s failure to comply undermines the purpose of the prohibition law and Act 58  – prevention of youth addiction to nicotine. Complying with this law will prevent more than 300 youth addictions per year.  The DLC should comply with the law forthwith. The bills being considered by the Legislature will help make that happen.

The following analysis deals with specific problems with the DLC report for 2012:

What is not true:

 

Page 1  “Vermont’s compliance data is statistically incomparable to that of other states using 15 and 16 year old minors.” Our results are statistically comparable with other states provided we make an adjustment for ages of buyers. All states report compliance testing results by age of buyers used. Analysis of approximately 43,000 SAMHSA (Synar) FFY2006 compliance tests nationwide shows that if a combination of 16- and 17-year-old buyers were used, our compliance testing results would still have been far below the requirements of Act 58. This is a valid use of statistics to compare Vermont results to those in other states. This higher result would have positioned Vermont at 28th in the nation for Synar testing. In addition, of the 14 states with more than 200 tests with 17-year-olds, eight had test results higher than Vermont’s for that age. If the DLC had been complying with the actual requirements of Act 58, Vermont would probably be ranked #1 in Synar performance.

 

Page 5  2001 … DLC complied with Act 58 ….” Average proxy measurement was 82%. The DLC failed to comply again.

Page 5  2002 … DLC complied with Act 58 ….” Average proxy measurement was 86%. The DLC failed to comply again.

Page 6  2003 … DLC complied with Act 58 ….” Average proxy measurement was 85%. The DLC failed to comply again.

Page 6  2004 … DLC complied with Act 58 ….” Average proxy measurement was 89%. The DLC failed to comply again.

Page 6  2005 … DLC complied with Act 58 ….” Average proxy measurement was 87%. The DLC failed to comply again.

Page 7  2006 … DLC complied with Act 58 ….” Average proxy measurement was 88%. The DLC failed to comply again.

Page 7  2007 … DLC complied with Act 58 ….” Average proxy measurement was 87%. The DLC failed to comply again.

Page 7  2008 … DLC complied with Act 58 ….” Average proxy measurement was 89%. The DLC failed to comply again.

Page 8  2009 … DLC complied with Act 58 ….” Average proxy measurement was 89%. The DLC failed to comply again.

Page 8  2010 … DLC complied with Act 58  ….” Average proxy measurement was 88%. The DLC failed to comply again.

Page 9  2011 … DLC complied with Act 58  ….” Average proxy measurement was 91%. The DLC failed to comply with the actual requirements of Act 58 again.

Page 14  2012… TEST RESULTS FOR CALENDAR 2012: The Department is using monies appropriated from the National Master Tobacco Settlement to comply with Act 58.   ….” Average proxy measurement was 90%. The DLC failed to comply with the actual requirements of Act 58 again.

 

Note: “at least 90%” requires that the lower margin of error of the measurement of actual real-life compliance (i.e. less than 10% commercial availability) be above 90%. The margin of error for the average annual proxy test noted above is approximately +/- 2%. This means our raw annual average would have to be about 92% to claim that this proxy measurement is “at least 90%”.  Actual real-life compliance, which the law requires to be at least 90%, would be significantly lower than this proxy average – probably about 55%. We are therefore about 35% below the actual requirement of Act 58.

 

What is half true:

Page 1 “Some national studies indicate that aggressive enforcement programs that have raised compliance rates in the mid-to–high 80% range have had no impact upon teenage tobacco use rates (in part because of a shift to social sources when commercial sources are less plentiful).” These studies actually showed that real-life compliance less than 90% is not sufficient to achieve the desired results. The research done by Leonard Jason of DePaul University, which was the primary basis for the US Congress requiring compliance testing in every state starting in 1992 (Synar Amendment), showed that compliance above 90% reduced youth addiction by 78%. Further recent research by Dr. Joe DiFranza indicates about 2% reduction in youth smoking for every 1% improvement of real-life compliance. Our law requires “consistent statewide compliance with the prohibition on sales to minors of at least 90 percent for 17-year-old buyers.” because that is what is required to significantly reduce drug addiction. Also, the assumption about a “shift to social sources” is based on information gained from current addicts. It also fails to consider the fact that most of the “social sources” obtain cigarettes through illegal sales to minors.  The behavior of non-smoking minors regarding sources is different from those who are addicted to nicotine and effective compliance enforcement dries up the supply in “social sources”. Act 58 is a prevention strategy – not a cessation strategy.

Also, this skepticism regarding the effectiveness of compliance testing in reducing the prevalence of youth smoking is further contradicted by a recent meta-analysis of all research to date on compliance testing by Dr. Joe DiFranza, a widely recognized expert in the field since the early 1990s. One of his recent works, Restricted Access to Tobacco Reduces Smoking Rates Among Youth”, one of the chapters in Focus on Smoking and Health, published in 2005 by Nova Science, explains how such skepticism is based on a lack of understanding of the possible difference between high compliance and low “commercial availability” – the actual ability of real-life youth to purchase tobacco products. Here is Joe’s conclusion from that report:

When evaluating the impact of youth access interventions on youth tobacco use, it is crucial to demonstrate that the intervention has reduced the commercial availability of tobacco. The scientific literature clearly supports a conclusion that interventions that truly restrict commercial access to tobacco reduce the prevalence of smoking among youth. In regard to this point, the literature is not mixed. Among the 270 papers reviewed, there was not a single study that demonstrated a significant reduction in the availability of tobacco without its having had a favorable impact on youth smoking. Any confusion about the effectiveness of preventing children from purchasing tobacco stems not from conflicting evidence, but from the failure of previous reviewers to attend to the crucial issue of whether interventions had reduced the commercial availability of tobacco.

[emphasis added]

This is a very powerful statement, and it is strongly supported by Joe’s meta-analysis of a large body of research on compliance testing.

In addition, Dr. DiFranza and three other researchers have recently done another study titled “The Impact of the Synar Amendment on Adolescent Smoking” that shows an estimated 2% reduction in youth smoking for every 1% improvement in compliance test results. Research during the past three years shows that there are 18 studies showing significant reduction in actual commercial availability with every one showing reductions in youth smoking rates. Joe DiFranza’s statement in his recent study: “Those who harbor personal or political opposition to enforcement programs can no longer hide behind the false argument that the efficacy of these programs is unproven.”

 

Page 14 “TEST RESULTS FOR CALENDAR 2012: … the number of licensees sampled was 968 …” The DLC continued to disregard its March, 2004 commitment to increase compliance testing frequency by 70% even though the compliance testing proxy measurement averaged only 91% for the year and the estimated actual compliance was only about 55%. The number of tests for the year would have had to be at least 1,900 to honor that 70% increase commitment and it would have required even more than that to meet the actual requirements of Act 58.

What they didn’t tell you:

Page 1 “Vermont is one of a few states to use 17-year-olds exclusively for compliance tests.” The actual distribution of ages used in compliance tests in other states is not shown in the DLC report. The actual number of Synar tests by age and gender is available for every state and I have analyzed these results for 2007. There were 13 other states that did more than 200 compliance tests with 17-year olds with a total of about 9,400 tests among the 13. Three of those states showed better than 90% compliance for their testing with 17-year-olds. Using 17-year-olds exclusively does not make “at least 90%” impossible or even impractical. It simply requires more frequent testing. Also, Act 58 does not require “exclusive” use of 17-year-olds. The language of Act 58 allows the DLC to use other ages.

I also analyzed the nationwide 2007 compliance test results by age of buyers and used that to project what Vermont’s results would have been if we had also used 16-year-olds in the same proportion as the rest of the states. This would have improved our compliance test results by only 3% - still not complying with our law.

Page 1 “For licensees educated by the Vermont Department of Liquor Control’s Education Division, rates of tobacco compliance have consistently … exceeded a 90% success rate, with an average of 93%.” The DLC received almost $300,000 of Tobacco Master Settlement Agreement funds for tobacco enforcement. Only about 6% of that money was spent on compliance testing and the rest was spent on training of alcohol and tobacco licensees and their employees. Most of that training is related to sales of alcohol.  With training mandatory, only 46% of tobacco sellers tested were trained by the DLC rather than by their employers.

Recent statistics show 93% proxy test results for DLC-trained clerks, 87% for employer-trained clerks, and 93% for employees with no training at all. Apparently the costly seminars have no significant effect on performance while the relatively low cost on-line training program is not significantly more effective.

Also, it should be considered that the DLC is currently training roughly about 16 times as many people as the law requires at considerable expense while training by the employers is done at no cost to the state

One of the most remarkable and important statements in the RTI report comparing Vermont compliance testing to other more successful states is the following:

Those retailers who are more willing to make serious efforts to comply with youth access laws may also have been more likely to request DLC-provided training for their clerks.

In addition to this correct observation by RTI, the turnover of retail clerks in Vermont is about 60%.That means in a typical store with more than one clerk required to cover the long hours, one of those clerks will have to be replaced every year and some clerks are for temporary periods. These clerks with very short experience are more likely to fail a test and less likely to have had an opportunity to attend a scheduled DLC seminar. They are trained by their employers so they can start working immediately and some of them get tested before they get to a DLC seminar – after which they will have much more experience.

In other words, there might have been no significant difference in the performance of this same set of clerks if they had been trained by their employers. This could account for all of the difference in passing proxy tests. It may also explain why other states get better results than Vermont without spending more money on training than on compliance testing.

Page 5  “Section 12 [sic], Act no. 58 of 1997 directed the Department of Liquor Control (DLC) to carry out “compliance tests of tobacco licensees as frequently and as comprehensively as necessary to assure consistent statewide compliance with the prohibition on sales to minors of at lease 90% for 17-year old buyers.” The language of this law was chosen very carefully and deliberately. Note that it refers to both “compliance tests” and “compliance with the prohibition on sales to minors”. These are not, and were not intended to be, the same thing. The DLC compliance enforcement tests are a proxy measurement of compliance with the prohibition on sales to minors. The first is obtained through the DLC testing protocol and the second is what happens in real life when minors actually try to purchase tobacco products in the normal retail environment. The testing protocol includes about six restrictions on behavior of buyers that are different from real-life transactions. The 90% requirement applies to the real-life conditions – not to the compliance test results. This has always been the case. In addition, the word “consistent” is critically important and it applies to consistency both geographically and in time. The intent is that all Vermont minors shall be protected all the time by consistent commercial availability less than 10%.  (All of this was clearly defined in my status report of April 11, 2005 and documented in all my monthly reports to the TERB since January, 2006.)

The monthly compliance testing results documented by the DLC on pages 20-32 in this report clearly show a lack of the consistency required by the law.

The DLC continues to describe its performance compared to the requirements of the law in terms of the raw annual average of the proxy test results for the whole state. This presentation ignores several critical statistical considerations. First, considering the sample size and the nature of what is being measured, the calculation of the uncertainty (error margin) of the proxy measurement will show that it is great enough to require an average compliance significantly higher than 90% to assure “at least 90%” at any time with a reasonable level of confidence. In addition, the six-month moving average of even these compliance testing results has been hovering around 90% for more than three years and the month-to-month variability has a range of about 10%. Even for time periods with the highest averages, about half the population of the state had compliance averaging about 85%. The consistency requirement has clearly never been met even for this proxy measurement, much less for the actual requirement of the law.

Page 13 “That after conducting tobacco stings (immediately afterward or at least within three days), the Liquor Control Investigator will inform licensees who sold tobacco products to youths. …” The DLC chooses to inform all licensees immediately after a test – even those who passed the test. Tests are supposed to be “unannounced”. When licensees are informed immediately, the integrity of all subsequent tests on that day is compromised because any licensee who has been tested can inform other licensees that tests are being conducted on that day. This is an egregious mistake. Other states specifically prohibit immediate notification because of this risk to the integrity of the data.

Page18 “PLANS FOR CALENDAR YEAR 2013  The Department of Liquor Control will test licensees as frequently as necessary to assure statewide compliance with the prohibition on sales to minors of at least 90% for 17-year-old buyers.” This same sentence was also in the annual reports for 1999-2011. The law actually requires the DLC to “assure consistent statewide compliance …” The elimination of the word “consistent” makes this commitment by the DLC significantly weaker than the requirements of the law. Even with this weaker wording, the DLC has never honored this repeated written commitment. On 3/23/04, the Liquor Control Board said “The Board’s intent is to get to the 90% compliance as required by statute.” They didn’t say when. Unless there is a significant increase in the frequency of testing, this commitment will again not be honored in 2013.

Also on page 18: … “DLC will also continue its mandated education program to licensees and their employees.” The “mandate” specifies DLC training for license holders once every three years. With about 1,000 tobacco licensees, that would be about 333 mandated trainings per year. According to Bill Goggins, the DLC trains approximately 5745  per year in its free seminars and an additional 2631 are trained using the DLC on-line training (with a fee of $10) for a total of 8376 – about 25 times what the mandate requires.

 “The effectiveness of this program can be seen in recent statistics that show a 93% success rate for those licensee’s employees who attend DLC seminars.” As noted on page 4 above, recent statistics actually show 93% proxy test results for DLC-trained clerks, 87% for employer-trained clerks, and 93% for employees with no training at all. Apparently the costly seminars have no significant effect on performance while the relatively low cost on-line training program is not significantly more effective.

“The Department will be contracting with the FDA to do compliance testing. The contract has been approved and the training by the federal government will start in 2013. The data from these tests will be used to meet the requirements of Act 58 and the Federal Synar requirement.”

Since the FDA testing will be done according to the enforcement protocol used by the DLC, the FDA testing will still be a proxy measurement and will therefore still not be valid for estimating the actual commercial availability in Vermont that Act 58 requires. The FDA contract is just an additional source of funding for tobacco compliance testing that could be used by the DLC for enforcement testing to comply with actual requirements of Section 13 of  Act 58, 1997..

 

The DLC report can be viewed at the following web page:

http://liquorcontrol.vermont.gov/sites/liquorcontrol/files/Tobacco Compliance Report 2012.pdf

 

*********************

Bruce Cunningham

301 Seneca Creek Road

Hinesburg, VT     05461

March 22, 2005

Mike Hogan

Commissioner, Department of Liquor Control

Green Mountain Drive, Drawer 20

Montpelier, VT 05620-4501

 

Dear Mike,

I have reviewed your “Report To The General Assembly on Tobacco Compliance Tests Conducted During Calendar Year 2004” and your related “Report To The General Assembly on Act 122 Section 75(a) Use of Education and Enforcement Funds for Fiscal Year 2004”. I would like to have some supplemental information to address four questions that have arisen with initial reading of the reports.

First, on pages 6, 7, and 11 are statements that the “…DLC complied with Act 58…” with the specific compliance testing requirement of Act 58, 1997 noted on page 4. The relevant text of that act is as follows:

The Department of Liquor Control shall conduct or contract for compliance tests of tobacco licensees as frequently and as comprehensively as necessary to assure consistent statewide compliance with the prohibition on sales to minors of at least 90 percent for 17-year old buyers.

 

The statewide annual average compliances noted on pages 8 and 9 in your report for years 1999 through 2004 are 79%, 77%, 82%, 86%, 85%, and 89% respectively. These data clearly show that the DLC has failed to comply with Act 58 for the last six years.

Question 1: How do you propose to resolve this contradiction between your statements and your data?

Second, the phrase “as frequently and as comprehensively as necessary” in the law clearly implies that the frequency of testing should be increased if compliance is not at least 90%.  It is clear from your report that there is no pattern of increased testing frequency in response to consistent unsatisfactory results. The only significant increase noted is the 70% increase in April, 2004. The results by month for 2004 show that of the 9 months with testing frequency increased by 70%, only 5 months were above 90% compliance, and the last two months were back down to the pre-April level. In addition, we have since learned that your reported results for January and February, 2005 are 81.5% and 78.2% respectively. It appears from this data that the 70% increase was insufficient to “…assure consistent statewide compliance with the prohibition on sales to minors of at least 90 percent…” as Act 58 clearly requires.

On page 14, your plans for 2005 indicate that you intend to comply with Act 58 by testing “as frequently as necessary”.

Question 2: What additional increase in compliance testing frequency will you implement this year to meet the full requirements of the law?

Third, as you noted on page 1 of your funding report, Act 122, 2004 contains the following language:

 Sec. 75.  Liquor control - enforcement and licensing …

Tobacco fund    290,000 …

(a)  The above tobacco fund appropriation shall be used for the sole purpose of education and enforcement with priority given to education.  …  The department of liquor control shall continue its obligations under Act 58 of 1997 to conduct, or contract for, compliance tests to assure consistent statewide compliance with the prohibition on sales of tobacco products to minors of at least 90 percent for 17-year-old buyers.

 

The key words in this language are “priority” and “obligations”. Your report on this funding shows that 87% of the tobacco fund appropriation was spent on your education program while only 13% was spent on compliance testing. It appears that the compliance testing share of this appropriation could be significantly increased to meet the “obligations under Act 58” while still maintaining a “priority” (more than 50%) for education.

Your reported results for 1997 (page 8) indicate that 92% compliance was achieved with a testing frequency equivalent to about 3,000 tests per year. This outcome suggests that returning to 3,000 tests per year might result in compliance with Act 58.

Question 3: What percentage of the tobacco fund appropriation would you need to spend on compliance testing to do 3,000 tests per year?

Question 4: If you spent 50% of the appropriation on compliance testing, what testing frequency could you maintain?

Please provide answers to the four questions noted above.

Thank you for your cooperation.

Sincerely,

Bruce Cunningham

 

(as of 2/3/09, I have received no response to this letter.)

 

Status of Prevention of Tobacco Sales to Minors in Vermont

Analysis by Bruce Cunningham, author of Sec. 13, Act 58, 1997

August 10, 2005

Since adults generally do not want young people to become addicted to nicotine, it follows that we should support drug abuse prevention efforts by not selling tobacco products to minors. Vermont has a law that prohibits selling tobacco to minors. We also have a law, Section 13 of Act 58, 1997, which requires the Department of Liquor Control to assure consistent statewide compliance with that prohibition of at least 90%. This is better than the federal Synar law and every other state law in the country, which require no more than 80% compliance. Recognizing that compliance of more than 90% is needed to significantly reduce youth addiction to nicotine, Vermont enacted this law in 1997 to provide better protection for our young people than the relatively ineffective 80% required by the federal government. While we are doing a better job than most other states, we have still not met the requirements of Act 58.

In March, 2004, in an effort to comply with Section 13 of Act 58, 1997, the Vermont Liquor Control Board committed to increasing the frequency of its tobacco compliance testing by 70% - to approximately 156 tests per month. After the increase, the six-month overall state average compliance as measured by the DLC increased significantly for about six months but it has now fallen back to the pre-increase level.

The chart below shows the DLC compliance testing frequency and monthly results for 2004 and 2005 with the six-month moving average for compliance shown by the solid line starting in June, 2004. It is clear that the DLC has failed to honor its commitment to increase the frequency of compliance testing in order to comply with the law.

The fundamental intent of Act 58 is to assure that when a typical 17-year-old anywhere in the state at any time attempts to purchase tobacco products, he or she will be unsuccessful at least 90% of the time. It is important to understand that this intent is about 10% commercial availability rather than 90% compliance testing.

The actual language of Act 58, which was carefully and deliberately chosen to reflect its full intent, is:

Sec. 13.  COMPLIANCE TESTING; TOBACCO SALES TO MINORS

(a)  The Department of Liquor Control shall conduct or contract for compliance tests of tobacco licensees as frequently and as comprehensively as necessary to assure consistent statewide compliance with the prohibition on sales to minors of at least 90 percent for 17-year old buyers. 

To clarify the difference between what this law requires and what we have achieved so far, imagine that Act 58 had instead been written as follows (differences in italics):

The Department of Liquor Control shall conduct or contract for compliance tests of tobacco licensees as frequently and as comprehensively as necessary to assure average annual compliance with the prohibition on sales to minors of at least 90 percent for 17-year old buyers behaving according to the DLC testing protocol. 

These two versions might seem to require the same thing to a casual and uncritical reader, but they in fact do not. The second (imagined) version reflects what seems to be the interpretation of some people who are responsible for implementing the law.

Disregarding the differences between the DLC testing protocol and normal behavior, consider the following two hypothetical scenarios for a given 12-month period:

1.    1000 compliance tests are performed in both the first and second six months. Suppose 875 licensees pass the tests in the first six months and 925 pass in the second six months. Also suppose that all of the test fails occur in the southern half of the state. Overall average compliance for the year is 90%.

2.    2,000 compliance tests are conducted and compliance in every county in every month is between 90% and 94%. Overall average compliance for the year is 92%.

The first scenario provides all minors in the northern half of the state with maximum possible protection (100%) while minors in the southern half have only 75% protection in the first half of the year and 85% protection in the second half of the year. In other words, half of the minors would not be protected to the degree intended by the law. This first scenario therefore fails to satisfy the requirements of the law.

Even considering the margin of error in testing results, the second scenario does protect all minors in the state with at least 90% compliance for the entire year.

Although both of these scenarios satisfy the requirement stated in the second (imagined) version of the law noted above (“average annual” compliance of 90%), only the second scenario satisfies the actual requirements of Act 58.

The first scenario is, of course, unrealistic. It is presented as an extreme to illustrate how much room for failure would be allowed with a requirement as vague as that stated in the imagined version of the law. The more specific actual language of the current law was chosen to avoid such a scenario.

A much more realistic example to consider is what we have today in actual compliance test results for the last six months. The DLC increased the frequency of tobacco compliance testing by 70% in April, 2004. The 6-month raw average compliance measured by the DLC for Feb. through July of 2005 is 86.6%. During those six months, however compliance in nine counties (bold in table below) that together contain 80.8% of the state’s population remained below the 90% level required by the law with an average compliance of 84.7%. Considering the margin of error for these 6-month results, another two counties with 12.1% of the population do not meet the requirement for “at least” 90% compliance.  This is a total of 96.8% of the youth in Vermont who are not being protected to the extent required by Act 58.

Addison

81.4

Bennington

95.5

Caledonia

91.4

Chittenden

85.4

Essex

100.0

Franklin

87.2

Grand Isle

100.0

Lamoille

84.4

Orange

85.7

Orleans

81.4

Rutland

85.9

Washington

88.9

Windham

91.4

Windsor

80.8

To illustrate the effect of the margin of error on the meaning of the reported data, consider an extreme example: performing only one compliance test in an environment where actual overall compliance is about 90%. The one test might be a pass and it might be a fail so the reported outcome will be either 100% or 0% compliance. If the test is a pass, we can only say that compliance is at least more than 0%. If the test is a fail, we can only say that compliance is between 0% and 100%. It follows that the more testing is done, the more we can reduce the uncertainty of our conclusion. Statistical analysis of current compliance testing data indicates that the error margin for the reported compliance is approximately +/- 2%. This means that if we are going to “assure at least 90%”, the measured level will need to be about 92% or higher.

In addition to the mathematical analysis outlined above, we also must consider the fact that the law also does not say “90% as measured by whatever compliance testing protocol and practices the DLC decides to use”. The current DLC protocol imposes several restrictions on buyer behavior that probably bias results toward higher indicated compliance.

What the law actually describes is the result required for real-life attempts under normal conditions. If compliance is going to be “consistent”, it must be at least 90% for real-life sales (10% commercial availability) as well as sales during testing. The DLC has so far conducted no control tests to determine the difference between its testing results and the actual compliance of tobacco licensees for attempted purchases by typical 17-year-olds. Whatever that difference is, it must be applied as a correction factor to the results obtained using the DLC protocol.

After eight years of this law being in effect, we have still not achieved the level of protection for our youth that the law intends and specifically requires. Currently more than 90% of Vermont’s youth are not protected as required. We need to get compliance consistently up to at least 90% in all parts of the state with error margins and testing biases considered. This will require competent statistical analysis of results and control testing to determine the effect of testing bias. It will also require another increase in the frequency of compliance testing.

 

 

*********************

[note: some of the analysis below is out of date because the current testing frequency is now (at the end of 2009) down to about 600 per year – which makes this even worse than it was with 1,500 per year. Each store would now be tested about once every two years]

 

Why did a 70% increase in the frequency of compliance testing not result in a lasting improvement in compliance results? It’s about consequences.

Apparently this increase was not enough to significantly change the consequences of poor performance. To understand why, it would be helpful to consider an example.

Suppose there is a tobacco licensee who is selling cigarettes to 17-year-olds approximately 50% of the time – i.e. this store has individual compliance of 50%.

The current testing frequency is approximately 1,560 tests per year – which is only a 30% increase in testing compared to 2003. There are currently about 1,250 tobacco licensees. That means each store is tested an average of approximately 1.25 times per year or about once every nine months.

The law allows a warning for a first offense and requires a one-day license suspension for a second offense with six months of a previous fail. Licensees who fail are usually “targeted” for a follow-up test within six months if they fail a test. Small fines are also allowed for multiple offenses within a three-year period but these fines have little effect on performance.

Now consider this scenario for the store that is a 50% performer. With the current testing frequency, this store is going to fail compliance tests approximately once every 18 months and receive only a warning. After each fail, there is a .5 probability that the follow-up will also be a failure. This means that approximately three years will pass between one-day license suspensions. Two-day suspensions would be very rare – about once every 6 years. That’s an average of one license suspension every 24 months – not enough to affect performance.

There were about 1,100 compliance tests in 2003. Comparable consequences for that testing rate were one-day suspensions about every four years and an additional two-day suspension every eight years – or approximately one suspension every 32 months.

Now suppose the average time between compliance tests is five months instead of nine. That translates to 3,000 tests per year with the current number of licensees. All test failures would occur within six months of the previous test. Two or three failures in a row would always result in one-day and two-day license suspensions. The licensee would fail half of the tests – or approximately once every 10 months - with about half of those failures being a second or third offense. This means there would be a one-day license suspension every 20 months and an additional two-day suspension every 40 months. That’s an average of one license suspension every 13 months. This frequency of suspensions would probably have a significant positive effect on licensee performance.

The financial benefits of 50% compliance instead of 90% compliance are profits in the neighborhood of 50 cents per illegal sale to a minor plus the long-term customer loyalty of those minors. Also to be considered is the fact that higher compliance is just more difficult to accomplish on a day-to-day basis. How much “pain” is needed to motivate a poor performing licensee to give up these advantages? Apparently losing some sales and aggravating customers once every two years is not enough. Doubling that level of “pain” with 3,000 tests per year is likely to produce a more long-lasting improvement – maybe even enough to comply with Act 58.

It should be noted that with a frequency of 3,000 tests per year, a 90% compliant licensee will fail a test (warning only) only about once every four years. Two fails in a six-month period (one-day suspension) would only happen about once every 40 years.

 

********************************************************************************************

Vermont Compliance Compared To Other States

 

The federal Center for Substance Abuse Prevention (CSAP) nationwide Synar compliance testing results from FFY2004 to FFY2012 are shown in the chart on the next page. CSAP reports “non-compliance” – i.e. percent selling to minors.

This chart shows Vermont’s actual reported results ranked only 31st out of 50 states plus the District of Columbia.

It should be noted that every other state is required by federal or state law to achieve only 80% compliance for a proxy test. Only Vermont has a law requiring “consistent statewide compliance with the prohibition on actual sales to minors of at least 90%”.

If we look at the results and rankings since FFY2004, we find that the same 90.5% compliance in that year would have been ranked 17th overall and the actual results reported in that year had Vermont ranked 41st overall. This means that even though our performance improved in the last 8 years, other states have improved even more than Vermont in the same time period even though our standard requires much more than the Synar law.

The Vermont DLC claims that since all compliance testing in Vermont is done with 17-year-olds while other states are allowed to use lower ages, it is not fair to compare our results to those other states. However, all states are required to break down their reported testing results by age of buyers used and analysis of those results shows that the 17-year-old testing in at least eight states in FFY2003 were higher than 90% while Vermont’s performance was only 85% for the same age level. Using similar data, we can deduce that our proxy test results would only be about 92% while meeting the actual requirements of our law would probably make us number 1 or 2 in the national Synar ranking.

                        Analysis by:     Bruce Cunningham, September, 2013

                                                482-2933

SYNAR RESULTS FOR ALL STATES FOR FFY2004 - FFY2012:

(numbers are for non-compliance – subtract from 100 for compliance)

State Name

2004

2005

2006

2007

2008

2009

2010

2011

2012

2012 rank

notes

2011 rank

Maine

8.8

5.3

8.5

5.8

5.2

5.8

4.6

4.2

1.8

1

 

9

Mississippi

3.9

6.0

3.2

3.2

5.1

3.8

3.8

2.3

1.8

2

 

1

Minnesota

12.1

13.7

13.4

9.5

7.9

8.3

9.0

4.1

2.4

3

 

8

Montana

11.2

14.5

9.5

8.8

4.1

5.5

3.6

3.6

3.0

4

 

5

Nevada

11.4

15.3

14.4

16.4

9.9

5.2

6.3

3.1

3.1

5

 

2

South Dakota

8.5

7.1

6.5

15.6

8.3

9.2

7.1

3.6

4.0

6

 

5

Louisiana

7.4

7.3

6.7

4.7

7.2

7.1

4.3

4.3

4.1

7

 

10

Arizona

8.9

8.3

7.6

14.6

5.7

6.7

4.8

5.9

4.3

8

 

16

Wisconsin

18.5

8.3

7.8

5.5

4.5

7.2

5.7

4.5

4.4

9

 

11

New Mexico

14.9

6.4

7.7

9.3

10.2

6.1

7.4

11.6

5.2

10

 

42

Indiana

16.6

11.9

14.8

9.9

14.7

11.2

9.5

5.7

5.3

11

 

15

Arkansas

16.6

4.2

2.2

4.7

4.2

4.8

4.5

3.4

5.4

12

 

4

California

12.2

14.0

10.2

13.2

10.7

12.6

8.6

7.7

5.6

13

 

25

Kentucky

6.7

5.6

6.3

6.3

6.2

4.7

3.3

5.4

5.7

14

 

13

New York

9.4

9.5

11.7

9.2

6.1

7.0

6.1

5.6

5.8

15

 

14

Hawaii

6.2

5.3

5.6

5.9

8.7

11.2

6.2

4.6

5.9

16

 

12

Alaska

10.2

12.1

20.6

13.6

11.4

15.1

9.0

7.6

6.5

17

 

23

DC

41.9

20.5

12.6

13.9

15.7

9.2

12.5

14.7

6.5

18

 

47

Oklahoma

10.0

13.9

9.6

9.0

12.5

17.9

9.4

11.5

6.8

19

VT if 15-17 used

41

Idaho

14.7

12.3

12

13.2

13.8

11.5

15.3

14.3

6.9

20

 

45

Kansas

22.1

38.0

19.2

19.9

12.9

15.9

8.3

7.8

7.6

21

 

26

Utah

8.9

8.0

8.5

9.9

8.4

8.4

7.0

7.1

7.6

22

 

20

New Hampshire

10.7

13.4

11.4

11.9

10.2

11.7

15.0

8.6

7.7

23

 

29

Delaware

3.9

0.9

3.5

5.0

5.4

2.8

2.2

3.1

8.1

24

VT if 16-17 used

2

Alabama

8.7

13.9

9.3

11.2

10.3

11.7

10.9

7.6

8.2

25

 

23

Iowa

5.2

11.4

8.2

11.8

11.6

10.0

8.2

8.7

8.2

26

 

30

Colorado

10.5

9.0

12.2

7.2

8.5

11.0

7.9

7.0

8.3

27

 

19

Florida

7.1

7.0

7.3

8.1

6.2

7.3

10.2

6.2

8.4

28

 

18

Wyoming

8.0

8.7

7

6.5

7.7

9.0

9.6

7.3

8.7

29

 

22

South Carolina

11.9

11.5

10.9

12.3

12.4

11.6

10.8

7.9

9.5

30

 

27

Vermont

15.9

12.7

13.1

10.7

14.0

12.9

11.2

11.2

9.5

31

 

39

Massachusetts

11.6

15.2

18.2

22.7

10.3

11.6

13.7

7.1

9.6

32

 

20

Pennsylvania

10.8

7.0

8.1

7.0

6.8

5.1

5.5

6.1

9.6

33

17

Rhode Island

16.9

9.5

9.1

8.0

11.1

11.7

9.3

11.1

9.9

34

 

38

Missouri

8.9

11.5

11.5

6.3

5.6

9.6

7.5

10.6

10.2

35

 

36

New Jersey

13.0

12.6

15.6

11.2

12.2

11.9

4.6

8.5

10.2

36

 

28

Georgia

6.2

16.5

16.4

14.2

8.7

12.1

9.5

14.1

10.3

37

 

44

Ohio

13.5

14.3

12

11.6

17.0

15.9

13.5

9.9

10.4

38

 

33

Nebraska

15.5

15.5

10.7

14.5

11.4

12.1

13.5

10.7

10.6

39

 

37

Connecticut

18.9

18.0

10.6

11.4

14.0

13.7

9.7

13.3

11.3

40

 

43

Texas

23.8

15.5

12.4

7.2

13.4

11.3

11.3

9.0

11.3

41

 

31

Washington

10.8

11.7

9.6

5.2

9.4

15.4

8.9

9.9

11.3

42

 

33

North Dakota

7.3

6.9

8.2

10.1

5.8

1.6

5.9

9.7

11.6

43

 

32

Virginia

10.2

13.1

9.7

10.0

9.7

13.9

13.5

11.3

13.0

44

 

40

West Virginia

10.3

13.4

13.5

16.9

14.2

17.9

11.4

14.6

13.4

45

 

46

North Carolina

14.8

16.8

14.8

10.3

11.5

17.4

14.8

10.3

13.9

46

 

35

Michigan

18.7

19.4

15

14.5

15.3

14.3

14.1

18.8

14.9

47

 

50

Illinois

16.8

6.4

6

9.5

5.3

12.5

6.0

4.0

15.5

48

 

7

Tennessee

13.5

10.9

13

16.3

10.9

13.1

14.9

16.1

16.7

49

 

48

Maryland

12.1

8.0

8.9

9.5

15.7

5.1

16.8

17.0

17.3

50

 

49

Oregon

16.3

14.6

17.8

11.0

15.6

18.8

16.7

19.3

17.9

51

 

51

 

 

 

 

 

 

 

 

 

 

 

 

 

VT rank

41

31

39

29

44

39

37

39

31

 

 

 

 

UPDATE: SYNAR results for FY2012 show Vermont ranked 31st out of 50 states plus D.C. for compliance test results Even when correction is applied for ages used in testing, Vermont is still at the median of all states even though our compliance with our law should put us at least in the top three in the country and probably at the top.

 

 

Two red herrings introduced by the DLC:

  1. “attempted” vs. completed compliance tests
    • Inaccurate town license lists is not a new problem.
    • Since DLC strategy is to test all licensees once per year, licenses more than one year old are known to be invalid on the last attempt and can be verified by phone before another attempt.
    • Suppose every license on the lists was a valid tobacco seller, how many more tests would be completed per year? The DLC should provide an estimate.
    • The law says “conduct” – not “attempt”. Whatever the impact of inaccuracies is, until the errors are eliminated, the DLC should attempt as many tests as necessary to comply with the law.
  2. “non-commercial sources”

·         Compliance testing is a prevention tactic.

·         YRBS data regarding non-commercial sources is obtained from current smokers.

·         No such data is available for potential smokers.

·         Laws and enforcement methods are already in place for non-commercial sources.

·         Behavior of non-smokers is likely to be different from nicotine addicts regarding begging and stealing.

·         Research shows significant reduction in new smokers (prevention) when commercial availability is less than 10% as required by Vermont law.

·         Whatever the effect of non-commercial sources, it is not an excuse for the DLC failure to obey the law.

 

Both of these are, of course, legitimate concerns -- but they should not distract us from reducing the commercial availability of tobacco for minors. These red herrings should neither prevent nor delay immediate and continued compliance with the statutory requirements for compliance testing.

 

 

From the description in the minutes of Commissioner Hogan’s comments:

Also, the Youth Risk Behavior Survey shows that most youth who have smoked at least one cigarette in the past 30 days obtain their cigarettes from social sources and not from commercial sources…

From the Commissioner’s written comments regarding the RTI report:

Page 26-More checks (300% Cunningham) increase costs by 200% will not solve the non-commercial source for youth.  That is where the problem lies!

From minutes description of discussion:

  Preventing Vermont’s young people from starting to smoke is one of the highest priorities of the Tobacco Control Program.

Also:

  The Committee in particular wishes to address the non-commercial sources of cigarettes for kids.  Vermont data indicates that the vast majority of minors who smoke get their cigarettes from sources other than retail purchase (e.g., giving someone else money to buy, borrowing, taking [read: stealing]  from a store or family member).  Methods for reducing those non-commercial sources are not well-developed.  

We do have laws in place prohibiting all current non-commercial sources. Are enforcement  methods really not well-developed? Consider the methods currently applied for non-commercial sources for marijuana, methamphetamines, cocaine, heroin, and other illegal drugs which have non-commercial sources similar to those for tobacco products for minors. Our comprehensive program also already includes elements designed to discourage youth from using such sources and our law enforcement agencies are already responsible for applying similar enforcement methods for tobacco.

Consider an analogy: suppose about 20% of convenience stores and grocery stores were selling marijuana. Would that be the top enforcement priority? If so, what level of commercial availability would have to be achieved for non-commercial sources to become more important?

Is the Department of Liquor control also going to become responsible for enforcing the current laws that prohibit providing tobacco products to minors, shoplifting, or stealing from family members? Are they going to start a new education program with DLC seminars to train youth in how not to violate the law? Of course not.

Why, then, is the DLC emphatically pointing out the data that indicates that current nicotine addicts obtain their drug from primarily non-commercial sources and claiming “That is where the problem lies!”? The answer to this question is: Because it’s an effective red herring. This is made clear by the sentence just noted from the Commissioner’s written comments.  This observation about the behavior of addicts is apparently intended to divert attention from the issue of whether the DLC should be held accountable for obeying the law (Act 58) regarding the prevention tactic of compliance testing.

(Also, suggesting that tripling the frequency would triple the cost is misleading. Economy of scale should be considered.)

Introducing this red herring is not a new tactic. I have been trying for years to explain that the YRBS data about non-commercial sources is relevant to the behavior of nicotine addicts but not necessarily to the behavior of the non-smokers whom we are trying to help with Tobacco Control Program prevention tactics. Current smokers are driven by their addiction to “beg, borrow, or steal” to get the drug they need whenever buying it from a tobacco licensee is difficult. This is not true of the non-smokers. If buying is difficult enough (less than 10% commercial availability), some of them don’t become addicts.

One dose of crack cocaine usually results in addiction. Nicotine addiction develops more slowly. Research clearly indicates that the prevention tactic of compliance testing actually has a positive effect on the roughly half of potential smokers for whom addiction appears not to develop unless they can purchase a few packs of cigarettes themselves with relative ease – i.e. higher than 10% commercial availability. The willingness of these young people to pursue non-commercial sources cannot be accurately measured, but it is certainly lower than the willingness of those who are already addicted. Because the YRBS data is obtained from youth who are already highly motivated by their addiction, it does not apply to the compliance testing (prevention tactic) target group.

The level of commercial availability required by current Vermont law (less than 10%) will help prevent nicotine addiction for more kids now. We cannot now accurately predict and we will never be able to accurately measure how much it will help, but we can be confident that it will help. Whatever the Enforcement Committee, and eventually the TERB, wants to do regarding investigating or taking action on any other possible alternatives, we are way beyond the time when the DLC should fully comply with Act 58. Whatever else we do or don’t do, full DLC compliance with the law should begin immediately.

 

Attachments: (the web site can be used with a click of the mouse if you read this on my web site)

 

See my analysis of the TERB 2005 annual report at 2005 TERB Report

See my analysis of the TERB 2006 annual report at 2006 TERB Report

See my analysis of the RTI report on compliance testing at RTI report analysis