WEBVTT 1 00:00:03.260 --> 00:00:14.010 Pingchuan Ma: Welcome to the Tobacco Online Policy Seminar Tops. Thank you for joining us today. I'm Pingchuan Ma, a researcher in Dr. Shang's lab at The Ohio State University Center for Tobacco Research. 2 00:00:14.010 --> 00:00:29.919 Pingchuan Ma: TOPS is organized by Michael Pesko at University of Missouri, Ce Shang at The Ohio State University, Michael Darden at Johns Hopkins University, Jamie Hartmann-Boyce at the University of Massachusetts Amherst, and Justin White at Boston University. 3 00:00:30.180 --> 00:00:43.589 Pingchuan Ma: The seminar will be 1 hour, with questions from the moderator and discussant. The audience may post questions and comments in the Q&A panel, and the moderator will draw from these questions and comments in conversation with the presenter. 4 00:00:43.990 --> 00:00:48.849 Pingchuan Ma: Please review the guidelines on TobaccoPolicy.org for acceptable questions. 5 00:00:49.350 --> 00:00:53.810 Pingchuan Ma: Please keep the questions professional and related to the research being discussed. 6 00:00:54.020 --> 00:01:01.269 Pingchuan Ma: Questions that meet the Seminar series guidelines will be shared with presenter afterwards, even if they're not read aloud. 7 00:01:01.580 --> 00:01:04.180 Pingchuan Ma: Your questions are very much appreciated. 8 00:01:04.709 --> 00:01:14.499 Pingchuan Ma: This presentation is being video recorded and will be made available, along with the presentation slides, on the TOPS website at TobaccoPolicy.org. 9 00:01:14.940 --> 00:01:21.719 Pingchuan Ma: I will turn the presentation over to today's moderator, Ce Shang from The Ohio State University, to introduce our speaker. 10 00:01:22.370 --> 00:01:23.749 Ce Shang: Thank you, Pingchuan. 11 00:01:23.970 --> 00:01:35.939 Ce Shang: Today, we continue our summer 2026 season with a single paper presentation by Brad Davis, entitled, Restricting Sales of Flavored Nicotine Vaping Products. 12 00:01:36.190 --> 00:01:46.710 Ce Shang: effects on nicotine vaping product and cigarette sales in Canada. This presentation was selected via a competitive review process by submission through the TOPS website. 13 00:01:47.110 --> 00:01:55.640 Ce Shang: Dr. Brad Davis is a postdoctoral fellow in health economist at the University of Missouri's Social Impact Lab. 14 00:01:55.830 --> 00:02:06.670 Ce Shang: His research examines the effects of public policy on health behaviors and outcomes, with recent work focusing on tobacco regulation and cancer screening. 15 00:02:06.740 --> 00:02:20.170 Ce Shang: He uses quasi-experimental methods and large administrative and retail data sets to evaluate how policy changes affect consumer behavior, healthcare utilization, and health outcomes. 16 00:02:20.420 --> 00:02:31.850 Ce Shang: He serves as an administrator for the Tobacco Online Policy Seminar TOPS, and earned his PhD in economics from Florida State University in May 2024. 17 00:02:31.940 --> 00:02:43.459 Ce Shang: Dr. Mike Pesko, a professor at the University of Missouri, is a co-author of the study, and will answer select questions in the Q&A. Dr. Davis, thank you for presenting for us today. 18 00:03:07.140 --> 00:03:09.220 Brad Davis: Alright, 19 00:03:09.380 --> 00:03:18.200 Brad Davis: First off, thank you to everyone for being here, and thank you to the, POPS Executive Board for having me back to present. 20 00:03:18.830 --> 00:03:32.030 Brad Davis: So today I'll be presenting my work on restricting sales of flavored nicotine vaping products, and the effect that that has on vaping product sales and cigarette sales within Canada. 21 00:03:32.280 --> 00:03:36.949 Brad Davis: This is joint work with, Mike Pesko and also, Abigail Friedman. 22 00:03:38.480 --> 00:03:41.720 Brad Davis: So first off, I'll start with the disclosure. 23 00:03:41.830 --> 00:03:48.740 Brad Davis: This work is supported by National Institutes of Health, the FDA, and also Health Canada. 24 00:03:48.880 --> 00:03:55.809 Brad Davis: And, the views expressed in this presentation are the views of the authors, and not necessarily the views of the funders. 25 00:03:56.810 --> 00:04:06.179 Brad Davis: In addition, I nor my co-authors have received any tobacco or pharmaceutical-related funding in the previous 10 years. 26 00:04:07.280 --> 00:04:10.239 Brad Davis: Or I should say, from tobacco and pharmaceutical companies. 27 00:04:10.360 --> 00:04:23.920 Brad Davis: And then, also we want to thank Health Canada for providing some of the data we use in these analyses, and then also the University of Missouri Social Impact Lab for programmatic support. 28 00:04:26.070 --> 00:04:35.309 Brad Davis: So in recent years, we've seen a number of government entities restrict nicotine vaping products. 29 00:04:36.270 --> 00:04:41.990 Brad Davis: Over 40 countries have outright banned nicotine vaping product sales. 30 00:04:42.210 --> 00:04:48.330 Brad Davis: 15 countries have flavor restrictions on nicotine vaping products. 31 00:04:48.540 --> 00:05:01.900 Brad Davis: And then within the United States, 7 states, have, restrictions on… on, nicotine vaping product flavors, as well as, over… around 200, local governments. 32 00:05:02.630 --> 00:05:17.969 Brad Davis: Now, the stated rationale for these policies is often to reduce youth appeal to vaping who might be attracted to it due to the fruity flavors that these vaping products often come in. 33 00:05:18.790 --> 00:05:34.330 Brad Davis: Now, we think about the… how consumers might respond to a flavor restriction on, vaping products. There's a number of different responses. One, they could reduce… this could reduce initiation into the… the vaping product market. 34 00:05:35.270 --> 00:05:40.709 Brad Davis: If… if… for consumers who might be, tempted to initiate into it due to the flavors. 35 00:05:40.890 --> 00:05:50.129 Brad Davis: For consumers that are already in the nicotine vaping product market, they could exit the tobacco market entirely. 36 00:05:51.430 --> 00:06:00.820 Brad Davis: They could substitute to a flavor that's not prohibited, such as maybe an unflavored vape product or tobacco-flavored vape product. 37 00:06:01.960 --> 00:06:14.670 Brad Davis: Or they could substitute to a tobacco product that is, not within the, the, the vaping category, so… including, combustible tobacco. 38 00:06:15.640 --> 00:06:22.090 Brad Davis: Now this fourth possibility is a particular A key public health concern. 39 00:06:22.190 --> 00:06:27.979 Brad Davis: Because while vaping is harmful, cigarettes are far more deadly. 40 00:06:28.840 --> 00:06:47.690 Brad Davis: So in this paper, it's estimated that the… that vaping has 37% the quality adjusted life years harm of that of smoking, and we know smoking, can have a… an impact on life expectancy, estimated on average to be 4.3 years. 41 00:06:49.440 --> 00:07:00.789 Brad Davis: So a key policy concern is, or a key question is, do, vaping product flavor restrictions cause substitution to, combustible cigarettes? 42 00:07:01.610 --> 00:07:10.419 Brad Davis: Within the U.S, we have some compelling evidence of this. We have numerous studies that show that these vaping products 43 00:07:10.680 --> 00:07:17.539 Brad Davis: Or these vaping flavor restrictions lead to increases in, cigarette smoking. 44 00:07:18.540 --> 00:07:24.169 Brad Davis: In addition, Friedman et al. estimates that for one fewer, 45 00:07:25.160 --> 00:07:34.110 Brad Davis: vaping cartridge that's sold. Due to a flavor restriction. 12 additional cigarettes are purchased. 46 00:07:35.270 --> 00:07:46.859 Brad Davis: So we… so we already see evidence of substitution, within the U.S. market, and then the… the, BAGE systematic review finds that out of 21 population studies. 47 00:07:47.040 --> 00:07:56.369 Brad Davis: 15 of these studies report an inverse relationship between vaping product availability and youth cigarette smoking. 48 00:07:57.340 --> 00:08:06.690 Brad Davis: So again, there's some… some evidence of here of the substitutability between vaping products and… and cigarettes, even among youth. 49 00:08:09.640 --> 00:08:25.190 Brad Davis: So what do we do in this paper? We leverage, provincial variation in the, in flavor… MVP flavor restrictions within Canada to estimate the effect on vaping product sales and cigarette sales. 50 00:08:25.680 --> 00:08:35.540 Brad Davis: This is some of the first evidence, causal evidence, outside the U.S. on how vaping product flavor restrictions affect the tobacco product market. 51 00:08:36.740 --> 00:08:39.299 Brad Davis: So why do we care about 52 00:08:39.460 --> 00:08:43.560 Brad Davis: Canada's regulatory… or about the, 53 00:08:44.230 --> 00:08:53.289 Brad Davis: about this study in Canada in particular. Well, one thing that's of note is that Canada has a stricter tobacco regulatory environment. 54 00:08:54.230 --> 00:09:08.020 Brad Davis: And so, so with that, some examples of what that looks like, Canada has a national menthol cigarette ban. They have more restrictions on flavored cigars, mandatory plain packaging. 55 00:09:08.210 --> 00:09:16.410 Brad Davis: And a nicotine concentration limit within vaping products of 20 milligrams per milliliter. 56 00:09:17.470 --> 00:09:29.659 Brad Davis: And so… so these different policies could… can affect the substitutability between vaping products and cigarettes. So let's just take the first one, for example, the national menthol cigarette ban. 57 00:09:29.930 --> 00:09:33.619 Brad Davis: Let's… let's say we have a consumer who 58 00:09:33.760 --> 00:09:42.080 Brad Davis: Who prefers the flavored vaping products, and maybe their next preferred product might be a menthol cigarette. 59 00:09:43.190 --> 00:09:55.949 Brad Davis: In the… in the presence of… and so let's say there's a flavor restriction, maybe they would substitute to that menthol cigarette. But within Canada, where the menthol cigarette is banned. 60 00:09:56.250 --> 00:10:08.530 Brad Davis: There's fewer… that option is no longer available, and so maybe they might decide to substitute to a, maybe a tobacco-flavored vape, or maybe they might exit the tobacco market entirely. 61 00:10:09.220 --> 00:10:16.770 Brad Davis: So in other words, with fewer combustible alternatives, there might be less substitutability between these two products. 62 00:10:17.460 --> 00:10:20.920 Brad Davis: But if we find evidence of substitution. 63 00:10:21.870 --> 00:10:38.080 Brad Davis: in Canada, despite its stricter tobacco control environment, it suggests that the substitutability between vaping products and cigarettes may be generalizable across countries, even countries with different tobacco control environments. 64 00:10:40.040 --> 00:10:51.750 Brad Davis: So this table here gives an overview of the implementation dates of these different policies across provinces. 65 00:10:52.630 --> 00:10:59.420 Brad Davis: So during the time of this study, there are 6 provinces that adopt a vaping product flavor restriction. 66 00:10:59.690 --> 00:11:19.079 Brad Davis: Now, there is some variation in the scope of these, of these restrictions, so some of these provinces, such as Nova Scotia, they outright ban the sale of flavored vaping products at all locations, while others limit it to specialty vape shops. 67 00:11:20.310 --> 00:11:26.020 Brad Davis: In addition, there are differences in which flavors are exempt from the policy. 68 00:11:26.070 --> 00:11:41.930 Brad Davis: So all six of these provinces, they allow sales of unflavored and tobacco-flavored vaping products, but Ontario and Saskatchewan, allow for mint or menthol-flavored sales. 69 00:11:43.630 --> 00:12:02.470 Brad Davis: And then there's four provinces that don't adopt a flavor restriction during the time of this study. I should note that Quebec has since adopted a nicotine vaping product flavor restriction, but it was after the time period of this study. I think they… I believe they adopted it in 2023. 70 00:12:05.440 --> 00:12:12.039 Brad Davis: So for our data, all of our data is from July 2018 to June 2023. 71 00:12:12.950 --> 00:12:18.010 Brad Davis: We get our… vaping product sales from Nilsen IQ. 72 00:12:18.190 --> 00:12:23.600 Brad Davis: Our sales data is specific to the gas and convenience store retail sector. 73 00:12:24.240 --> 00:12:35.209 Brad Davis: We observe sales in 6,000 stores, which makes up about 20, 24% of the gas and convenience store retail sector. 74 00:12:36.100 --> 00:12:42.529 Brad Davis: And this data is by UPC, and province in 4-week time period. 75 00:12:43.040 --> 00:12:53.860 Brad Davis: And so UPC is just the barcode that is on the product, and so with the UPC also comes a lot of detailed data on the product characteristics. 76 00:12:54.010 --> 00:12:59.319 Brad Davis: And we're able to, categorize these sales into different flavor categories. 77 00:13:00.780 --> 00:13:07.689 Brad Davis: Now, there's 9 provinces in this sample, but some of the 78 00:13:07.800 --> 00:13:14.559 Brad Davis: Provinces are aggregated to date, or some of the provinces are aggregated within the data. 79 00:13:14.740 --> 00:13:17.410 Brad Davis: So there's only 6 panels in total. 80 00:13:18.520 --> 00:13:22.619 Brad Davis: So, Alberta and Quebec are our, control groups. 81 00:13:22.900 --> 00:13:34.260 Brad Davis: And then we have British Columbia and Ontario, and then two province groups. So, the Maritimes are aggregated together, which includes New Brunswick, Nova Scotia, and Prince Edward Island. 82 00:13:34.600 --> 00:13:44.860 Brad Davis: Now, this is just the way the data came, so there's no way for us to disaggregate across those three provinces. I wish we could, but unfortunately we cannot. 83 00:13:45.840 --> 00:14:03.600 Brad Davis: And then also another province group made up of Manitoba, which, which did not adopt flavor restriction, and then Saskatchewan, which did. So, I'll describe this more in my analysis, but in our model, this group will receive a partial treatment. 84 00:14:06.470 --> 00:14:18.999 Brad Davis: And then our cigarette sales data comes from Health Canada, and this is, this data is on shipments from manufacturers and importers to retailers and wholesalers at the province-by-month year level. 85 00:14:19.800 --> 00:14:37.820 Brad Davis: This is high-quality data because it's, mandatory that this be reported under the Tobacco reporting regulation laws within Canada, so it contains the… the universe of, cigarette sales within… with… or of legal… I should say legal cigarette sales within Canada. 86 00:14:37.820 --> 00:14:39.969 Brad Davis: And it's available at the province… 87 00:14:40.270 --> 00:14:44.340 Brad Davis: Or, yeah, sorry, sorry, it contains all provinces and territories. 88 00:14:44.550 --> 00:14:57.499 Brad Davis: For our main analysis, we restrict it to the 9 largest provinces based on recommendations from Health Canada, but we also do include a sensitivity check where we use all provinces and territories. 89 00:15:00.110 --> 00:15:11.999 Brad Davis: And then we also utilize data from Google Trends, and this is going to provide a measure, or… yeah, this data provides a measure of Google search intensity for two different outcomes. So we look at 90 00:15:12.060 --> 00:15:24.160 Brad Davis: Google searches containing the phrase, vape shop, and then also Google searches that are categorized by Google as being related to NVP or e-cigarette shopping. 91 00:15:24.970 --> 00:15:35.679 Brad Davis: And so, while this is not a measure of sales in other retail categories, beyond 92 00:15:35.680 --> 00:15:44.990 Brad Davis: gas and convenience store. We do think it is a… it is informative of consumer interest within these other categories. 93 00:15:46.190 --> 00:15:50.150 Brad Davis: And is, therefore useful for our analyses. 94 00:15:52.440 --> 00:15:58.240 Brad Davis: So I'll start by describing a basic two-way fixed effect model. 95 00:15:59.480 --> 00:16:06.649 Brad Davis: Where our outcome is the… is the log of the… or the natural log of the… 96 00:16:06.870 --> 00:16:10.060 Brad Davis: A vaping product unit sales per capita. 97 00:16:10.370 --> 00:16:14.350 Brad Davis: Or the natural log of cigarette stick sales per capita. 98 00:16:15.410 --> 00:16:19.640 Brad Davis: Where flavor restriction is our, variable of interest. 99 00:16:19.770 --> 00:16:22.219 Brad Davis: Which is proportional 100 00:16:22.380 --> 00:16:32.539 Brad Davis: to, or is equal to the proportion of the province or panel with a flavor restriction in month, year T. So for most of the data, that's a 0-1 binary. 101 00:16:33.470 --> 00:16:44.080 Brad Davis: For those, for those panels that are grouped together, it's equal to the proportion of the population with a flavor restriction in place. 102 00:16:45.430 --> 00:16:59.640 Brad Davis: And then we also include a vector of covariates that are broadly under two categories, one being other tobacco policies, and then also policies related to the COVID-19 pandemic. 103 00:17:01.120 --> 00:17:06.900 Brad Davis: So for tobacco controls, or tobacco covariates, we use, cigarette taxes. 104 00:17:07.130 --> 00:17:13.380 Brad Davis: Vaping product taxes with separate variables for ad valorem rates and also excise rates. 105 00:17:13.869 --> 00:17:18.569 Brad Davis: E-cigarette, nicotine, concentration limits. 106 00:17:19.190 --> 00:17:26.950 Brad Davis: And then for COVID covariates, we… we use… include, COVID-19 deaths per capita, 107 00:17:27.310 --> 00:17:41.920 Brad Davis: a measure of non-essential business closures and internal movement restrictions, and then also a binary indicator for the Atlantic bubble, which was a COVID-related policy within the Atlantic region of Canada. 108 00:17:43.120 --> 00:18:00.419 Brad Davis: And then our model also includes province-level fixed effects, which are going to control for the time-invariant characteristics of each province, and then month-year fixed effects, which control for national trends or national changes in tobacco sales within Canada. 109 00:18:02.250 --> 00:18:12.859 Brad Davis: Now, there are recent advances in the difference-in-difference literature, which highlight that when you have an environment with a staggered adoption policy. 110 00:18:13.030 --> 00:18:22.409 Brad Davis: It can bias the estimates due to comparisons between newly treated provinces and provinces that have already been treated. 111 00:18:23.060 --> 00:18:29.230 Brad Davis: Now, there's a number of different strategies to try to mitigate this concern. 112 00:18:29.710 --> 00:18:33.390 Brad Davis: The one that we go with is called the Stacked Difference in Difference Model. 113 00:18:34.400 --> 00:18:37.739 Brad Davis: Essentially what this does is it reconstructs the… 114 00:18:37.940 --> 00:18:43.169 Brad Davis: The data into, event time, rather than calendar time. 115 00:18:43.540 --> 00:18:49.669 Brad Davis: So for every province that gets treated, we take, 21 116 00:18:50.190 --> 00:19:00.620 Brad Davis: Months of data before and after that policy gets adopted for that… for that province, and as well as the controls, and that creates one stack of data. 117 00:19:02.170 --> 00:19:11.819 Brad Davis: And then we do that for… for every province that adopts, and then you stack that data on top of each other, and it's now in event time. 118 00:19:12.250 --> 00:19:16.409 Brad Davis: So that's kind of a brief one-minute overview. 119 00:19:16.970 --> 00:19:24.080 Brad Davis: Now 21 months was chosen just because that's the longest that could be chosen based on the data that we have. 120 00:19:27.090 --> 00:19:35.309 Brad Davis: And then, in the stack difference-in-difference model, you also include the stack by unit fixed effects and stack by event time fixed effects. 121 00:19:35.890 --> 00:19:42.709 Brad Davis: Now there's… there has been a recent update to the model from the Wing, Friedman and Hollandsworth paper. 122 00:19:43.350 --> 00:19:57.240 Brad Davis: Essentially showing, or this, this paper introduces some, some weights that, help properly, properly balance the model, and so that we include those weights in our analyses. 123 00:19:59.320 --> 00:20:03.909 Brad Davis: Now we do, a number of robustness and sensitivity checks. 124 00:20:04.290 --> 00:20:06.759 Brad Davis: For our sensitivity checks. 125 00:20:07.280 --> 00:20:24.900 Brad Davis: We… we do… we use… we do a model where we use volume as the sales outcome. In our main specification, we use units, because volume is missing in about 1% of our sales observations. 126 00:20:24.990 --> 00:20:28.869 Brad Davis: And so, and so as a sensitivity, we… we… 127 00:20:28.990 --> 00:20:45.860 Brad Davis: use volume as one of the outcome, or as the outcome. For another specification, we exclude March through August 2020, just to ensure that our results are robust to anything that might be happening around the start of the COVID pandemic. 128 00:20:46.810 --> 00:20:55.550 Brad Davis: And then we also use a model, or have a model using population weights rather than the stack difference in difference weights. 129 00:20:56.950 --> 00:21:01.320 Brad Davis: And then for, for some identification and inference checks. 130 00:21:01.470 --> 00:21:14.900 Brad Davis: There is one of our models where we were a little bit concerned about the pre-policy trends, and so we also used the synthetic difference-in-difference model as a robustness check. 131 00:21:15.390 --> 00:21:32.210 Brad Davis: What this model does is it re-weights the model, such that the pre-policy trends between the treatment and control group better align, and therefore is more robust to violations of the parallel trends assumption. 132 00:21:33.990 --> 00:21:35.869 Brad Davis: And then… 133 00:21:36.060 --> 00:21:43.179 Brad Davis: One of the limitations of our data is that we have a small number of panels, and so therefore a small number of clusters. 134 00:21:43.470 --> 00:21:52.329 Brad Davis: And when you have a small number of clusters, this can lead to standard errors that over-reject the null hypothesis. 135 00:21:53.390 --> 00:22:03.359 Brad Davis: There's a number of ways that can be used to… to address this. We do two sensitivity checks, one using the wild cluster bootstrap. 136 00:22:03.660 --> 00:22:10.230 Brad Davis: p-values. And then another using a, doing a placebo test. 137 00:22:10.420 --> 00:22:17.100 Brad Davis: And plotting the coefficients of the placebo test and creating p-values based on that distribution. 138 00:22:17.490 --> 00:22:30.659 Brad Davis: I don't have these results in my main presentation, but I do have it in the appendix if we wanted to look at that, and then, of course, it's also in the paper, and will be on the presentation slides on the website, too. 139 00:22:32.920 --> 00:22:38.029 Brad Davis: Let's go ahead and pause here for any questions. 140 00:22:42.140 --> 00:22:51.550 Ce Shang: Thank you, Brad. Our discussion today is Dr. Madeline Dove, an Associate Adjunct professor in the Public Health Sciences Department at UC Davis. 141 00:22:52.060 --> 00:22:55.179 Ce Shang: Melanie, do you have any questions for Brett? Thank you. 142 00:22:57.990 --> 00:23:02.380 Melanie Dove: Yes, thank you for this opportunity to discuss. 143 00:23:02.580 --> 00:23:07.690 Melanie Dove: This is a really interesting study, and you explain it very clearly. 144 00:23:07.840 --> 00:23:13.530 Melanie Dove: You know, like you mentioned, there has been, you know, several studies in the U.S. looking at the impact of 145 00:23:13.730 --> 00:23:21.820 Melanie Dove: Flavor restrictions on sales, tobacco sales and use, and this is, you know, kind of one of the first in Canada. 146 00:23:21.900 --> 00:23:34.669 Melanie Dove: And Canada has a really interesting environment, you know, where they've banned menthol cigarettes, and so the results, I think, can, you know, provide something very new and interesting. 147 00:23:35.300 --> 00:23:43.720 Melanie Dove: Let's see… I was wondering, if you could describe the tobacco purchasing? 148 00:23:43.990 --> 00:23:58.139 Melanie Dove: patterns in Cal… California… in Canada? You know, where are most vapes sold? Is it convenience stores and gas stations? You know, are the patterns similar to the U.S? 149 00:24:01.140 --> 00:24:05.790 Brad Davis: So I can speak to the, to the… To the vaping market. 150 00:24:06.180 --> 00:24:14.969 Brad Davis: About a third of sales happen at gas and convenience stores, and then the other two-thirds is from vape shops and online. 151 00:24:15.500 --> 00:24:22.740 Brad Davis: So… so we only observe about 10% of the overall, vaping product market sales. 152 00:24:23.170 --> 00:24:27.949 Brad Davis: And then what was the first part of your question about… about cigarette sales? 153 00:24:29.130 --> 00:24:40.709 Melanie Dove: Let's see, well, I guess I was mainly interested in the vape sales, because the data you have is just from gas and convenience stores, so, yeah, thank you for that. 154 00:24:41.010 --> 00:24:42.910 Melanie Dove: Explanation, yeah. 155 00:24:43.470 --> 00:24:50.739 Melanie Dove: Is online sales… do, are online vape and cigarette online sales legal? 156 00:24:50.990 --> 00:24:52.240 Melanie Dove: In Canada. 157 00:24:52.450 --> 00:24:55.260 Brad Davis: Yeah, yes, yes, they are. Yep. 158 00:24:56.210 --> 00:24:56.670 Melanie Dove: Okay. 159 00:24:56.670 --> 00:24:58.340 Brad Davis: But, but I mean, 160 00:24:59.190 --> 00:25:12.990 Brad Davis: You know, the flavor restriction, some of them could apply to the vape shops and online sales as well, but vape shop and online sales are generally permitted for, for vaping products. 161 00:25:14.180 --> 00:25:15.780 Melanie Dove: Hmm. Okay. 162 00:25:18.380 --> 00:25:29.099 Melanie Dove: Okay, and it looks like you use separate data sources, for the vape sales and cigarette sales. Why did you use separate data sources? 163 00:25:30.960 --> 00:25:37.360 Brad Davis: Really, it came down to, availability. I mean, the… the cigarette sales are, 164 00:25:38.480 --> 00:25:49.090 Brad Davis: are… are very high quality and contain the universe of legal sales, so, so that… that's obviously a very good data source to have. 165 00:25:49.570 --> 00:26:00.169 Brad Davis: I believe our Nielsen IQ data did not contain cigarette sales, as far as I'm aware. I think Mike could, could verify that. 166 00:26:00.360 --> 00:26:04.609 Brad Davis: And… Yeah, yeah, so that… 167 00:26:04.730 --> 00:26:09.330 Brad Davis: That's the main reason for the differences in the data sources. 168 00:26:09.880 --> 00:26:12.079 Brad Davis: And, of course, a limitation 169 00:26:12.220 --> 00:26:29.460 Brad Davis: of the Nielsen IQ is that it is restricted to gas and convenience stores, so we want to make sure that we're clear that… that that is the sector that we're… we're measuring effects for. And then also why we try to pull in that Google Trends data 170 00:26:29.590 --> 00:26:30.640 Brad Davis: Which 171 00:26:31.140 --> 00:26:39.409 Brad Davis: You know, it's not… it's not a substitute for sales data, but might give us some indication of consumer interest in those other categories. 172 00:26:40.460 --> 00:26:51.530 Melanie Dove: Right, right. Yeah, it sounds like you have great data sources, you know, yeah, especially with the Google, with the, cigarette sales is very comprehensive, and then 173 00:26:51.650 --> 00:26:53.549 Melanie Dove: Adding in the Google sales. 174 00:26:53.740 --> 00:26:56.090 Melanie Dove: Google search data as well. 175 00:26:56.500 --> 00:27:13.110 Melanie Dove: Okay, one… I have one more question. I recently, you know, I worked on a study, and we used the stacked difference-in-difference approach, and I was just wondering if you had considered other approaches to, kind of deal with the 176 00:27:13.160 --> 00:27:17.149 Melanie Dove: The potential bias from staggered policy implementation. 177 00:27:21.510 --> 00:27:37.459 Brad Davis: We may have very early on, I… I can't remember the details exactly. I… I like the stacked because it creates a balanced panel for all the treated provinces, and so for that reason, I… I like it. 178 00:27:39.250 --> 00:27:52.249 Brad Davis: you know, we… that's something we could consider doing, is… is showing results for some of the other models. I will say, we do… we do show results for all of our outcomes using synthetic difference-in-difference. 179 00:27:52.790 --> 00:28:00.319 Brad Davis: Which does not have that staggered adoption, problem, because it does the estimates one at a time for each 180 00:28:00.450 --> 00:28:01.930 Brad Davis: Treat… treatment group. 181 00:28:02.490 --> 00:28:12.440 Brad Davis: But, you know, we could consider incorporating maybe another model or two from the, Saggered adoption literature. 182 00:28:14.810 --> 00:28:18.139 Melanie Dove: Alright, thank you. Excited to see the results. 183 00:28:19.080 --> 00:28:19.890 Brad Davis: Thanks. 184 00:28:20.020 --> 00:28:34.150 Ce Shang: Thank you. I think I.B. Friedman added some comments. she says the Canadian government data is more comprehensive, but only includes cigarettes, so that answers 185 00:28:34.440 --> 00:28:40.589 Ce Shang: Why, you know, you guys use the government data for measuring cigarette sales. 186 00:28:40.740 --> 00:28:53.280 Ce Shang: So I see Mike is answering some of the questions, but there are some open ones. One is from Chris Saenz. Should we think of Google search interest as a measure of consumer demand? 187 00:28:53.490 --> 00:29:01.330 Ce Shang: Google search interest tracks actual sales or consumption, a reflection of consumer preferences. 188 00:29:01.520 --> 00:29:06.370 Ce Shang: Or what consumers are thinking about in their purchasing decisions. 189 00:29:06.720 --> 00:29:14.660 Ce Shang: He's wondering what the overlap of this outcome is with your other data outcomes, and how it's different. 190 00:29:15.710 --> 00:29:16.360 Ce Shang: Yeah. 191 00:29:17.660 --> 00:29:24.779 Brad Davis: That would be very interesting to see, the overlap between the Google search data and an actual, 192 00:29:25.100 --> 00:29:27.210 Brad Davis: sales… sales data? 193 00:29:27.430 --> 00:29:32.579 Brad Davis: And would maybe… Tell us a little bit more about 194 00:29:32.760 --> 00:29:47.570 Brad Davis: what the Google search data could tell us. I don't want to… I want to be very cautious and make sure we don't overstate what the Google Trends data tells us, because I… I don't think it's, you know, it's not a proxy for sales. 195 00:29:47.630 --> 00:29:54.000 Brad Davis: The main reason we're using it is because we want to make sure that there… we want to… 196 00:29:54.550 --> 00:30:01.120 Brad Davis: Look for evidence of substitution to the vape shop or online market. 197 00:30:01.740 --> 00:30:12.330 Brad Davis: And that… and that's really one of the main reasons we're… we're using it, because if… if there was… there were increases in Google searches related to vape shops. 198 00:30:12.520 --> 00:30:22.609 Brad Davis: or, online retailers. There might be evidence that consumers are substituting away from the gas and convenience sector to these 199 00:30:22.670 --> 00:30:35.190 Brad Davis: another retail sector. I'll go ahead and preview and just say that that's not what our results show. And so, I think of it as, currently, I think of it as… 200 00:30:35.510 --> 00:30:42.260 Brad Davis: As a… As an indicator of consumer interest, and maybe you could say preferences. 201 00:30:42.700 --> 00:30:43.150 Ce Shang: Yeah. 202 00:30:43.150 --> 00:30:50.299 Brad Davis: But I don't want to overstate it and say that it's, you know, a perfect proxy for sales, although, like I said, that would be a very interesting study. 203 00:30:50.530 --> 00:31:02.649 Ce Shang: Yeah, Chris also mentioned that, he thinks it would be useful to add the citation of the correlation before… between Google search for nicotine terms and the survey-based measures of use or sales. 204 00:31:02.650 --> 00:31:03.260 Brad Davis: Okay. 205 00:31:03.260 --> 00:31:11.769 Ce Shang: So, I don't know if there is any study like this, but maybe someone has already done something looking into the correlations. 206 00:31:12.340 --> 00:31:14.239 Brad Davis: Yep, yep, no, that's a good thought. 207 00:31:14.240 --> 00:31:14.920 Ce Shang: Yep. 208 00:31:15.030 --> 00:31:21.390 Ce Shang: So, I don't see any other open questions. Please go ahead and continue the presentation. Thank you. 209 00:31:39.170 --> 00:31:41.919 Brad Davis: Alright, so I'll go ahead and, 210 00:31:42.140 --> 00:31:46.730 Brad Davis: Getting into the results, we'll start with just looking at the descriptive evidence. 211 00:31:47.680 --> 00:31:56.249 Brad Davis: So these are… this is showing, trends in the population-weighted unit sales in the treatment and control provinces. 212 00:31:57.510 --> 00:32:15.460 Brad Davis: So we can see prior to adoption of any flavor restrictions, we see that sales were increasing quite rapidly in both treatment and control groups. But then once these flavor restrictions start going into place, we see, sales in the treatment group begin to level out. 213 00:32:16.440 --> 00:32:30.000 Brad Davis: Whereas the sales in the control group continued growing, such that after all the flavor restrictions have been adopted, there's now a big gap between the treatment and control 214 00:32:30.240 --> 00:32:35.129 Brad Davis: Control groups, where sales are roughly double that. 215 00:32:35.630 --> 00:32:40.300 Brad Davis: The sales in the control group are roughly double that of the treatment provinces. 216 00:32:42.560 --> 00:32:51.830 Brad Davis: Looking at the cigarette sales trends, we see that in the treatment group, sales were a little bit, le… or were a little bit lower. 217 00:32:52.100 --> 00:32:57.159 Brad Davis: Then, in the control groups before the flavor restrictions went into effect. 218 00:32:57.720 --> 00:33:08.799 Brad Davis: And then after the policy, after the flavor restrictions go into effect, we see that cigarette sales are roughly on the same level, but actually a little bit, 219 00:33:08.940 --> 00:33:11.730 Brad Davis: A little bit higher than the control group. 220 00:33:11.940 --> 00:33:14.860 Brad Davis: Right. Something else I think is, you know. 221 00:33:15.080 --> 00:33:28.419 Brad Davis: Interest that's important to point out is that cigarette sales are declining over time for both groups, which I think is going to help us at the end when it comes to interpretation of our results. 222 00:33:31.290 --> 00:33:37.569 Brad Davis: Alright, so looking at our, formal modeling results, the results of our, of our models, 223 00:33:37.790 --> 00:33:40.800 Brad Davis: Looking at the gas and convenience store. 224 00:33:41.150 --> 00:33:45.259 Brad Davis: Total MVP sales, unit sales? 225 00:33:45.490 --> 00:34:02.619 Brad Davis: We see a 29.5% decrease. Now, it's not, statistically significant, but it is economically meaningful. It's a very, very large but imprecise reduction in sales. 226 00:34:03.600 --> 00:34:09.260 Brad Davis: Now, breaking that up into, specific flavor categories. 227 00:34:09.730 --> 00:34:15.449 Brad Davis: So looking at the… the flavored… Nicotine vaping product sales. 228 00:34:15.750 --> 00:34:20.390 Brad Davis: We estimate a 99.9% reduction 229 00:34:20.550 --> 00:34:25.509 Brad Davis: In the flavored cells, right, this is consistent with what we would expect to see. 230 00:34:25.969 --> 00:34:28.690 Brad Davis: For the mint and menthol. 231 00:34:29.350 --> 00:34:36.209 Brad Davis: sales category. We also estimate a very large decline, about a 90… 97%. 232 00:34:36.370 --> 00:34:39.120 Brad Davis: So not as large as the flavored category, but… 233 00:34:39.239 --> 00:34:48.010 Brad Davis: That's what we would expect, because there were some provinces that allowed… that continue to allow the mint and menthol sales. 234 00:34:50.400 --> 00:34:56.699 Brad Davis: And then looking at, tobacco and unflavored, Sales? 235 00:34:57.180 --> 00:35:07.779 Brad Davis: So we… we joined Tobacco and unflavored into one… one category. We estimate very large increases in sales in this category, over… over a 100% increase. 236 00:35:08.810 --> 00:35:12.790 Brad Davis: So here we do find evidence of, of substitution from… 237 00:35:12.920 --> 00:35:18.829 Brad Davis: The… the categories that are prohibited to categories that are not prohibited. 238 00:35:23.310 --> 00:35:27.399 Brad Davis: So here's our, the results from the Google Trends analysis. 239 00:35:27.760 --> 00:35:34.110 Brad Davis: So here we have searches related to vaping product shopping. 240 00:35:34.330 --> 00:35:40.109 Brad Davis: We estimate about a 22% reduction in those… those Google searches. 241 00:35:41.560 --> 00:35:44.170 Brad Davis: And then for, for vape shops. 242 00:35:44.660 --> 00:35:58.360 Brad Davis: for our Google searches containing the phrase vape shop. Similarly, we estimate a… about a 17% decrease, although only statistically significant at about the 11% level, so it's just beyond, 243 00:35:58.640 --> 00:36:03.610 Brad Davis: would be considered statistically significant. But… 244 00:36:04.410 --> 00:36:15.729 Brad Davis: What these results tell us is that these results are not consistent with there being a lot… with there being substitution to other retail, sectors. 245 00:36:19.700 --> 00:36:29.770 Brad Davis: And then, looking at our cigarette sales data, we estimate about a 9.6% increase in cigarette sales. 246 00:36:29.880 --> 00:36:39.879 Brad Davis: Now, we were a little bit concerned when we… when we first looked at this result, just because some of the… the pre-policy, coefficients are… 247 00:36:40.330 --> 00:36:48.410 Brad Davis: Statistically different than zero. And… and so that's why we pulled in the synthetic difference-in-difference model. 248 00:36:49.380 --> 00:37:00.739 Brad Davis: Which, again, re-weights the model to better align the pre-policy trends, and is therefore more robust to violations of the parallel trends assumption. 249 00:37:00.830 --> 00:37:12.690 Brad Davis: And so here, we do… we do see better parallel trends in the pre-policy period, and then, and then afterwards, we… we also estimate a, positive 250 00:37:12.860 --> 00:37:22.180 Brad Davis: increase in cigarette sales. A little bit larger, are about about a 21% increase. 251 00:37:22.380 --> 00:37:28.529 Brad Davis: But nevertheless, this is still, you know, this is directionally consistent with our, with our 252 00:37:29.030 --> 00:37:33.890 Brad Davis: Early results and boost our confidence of that result. 253 00:37:35.980 --> 00:37:42.240 Brad Davis: So, as previously mentioned, we do a number of sensitivity checks. 254 00:37:42.430 --> 00:37:53.870 Brad Davis: This is looking at… at the total NVP sales, starting with the… this is our baseline model. Standardized is using, volume as the outcome rather than units. 255 00:37:54.260 --> 00:38:09.799 Brad Davis: Then we… then we drop March through August 2020, and then a population-weighted model. All these… these coefficients are, you know, very consistent, and do not meaningfully change the conclusions of our 256 00:38:10.010 --> 00:38:11.590 Brad Davis: Of our analyses. 257 00:38:12.560 --> 00:38:18.689 Brad Davis: Same thing with the cigarette sales. All these coefficients are fairly, consistent. 258 00:38:21.370 --> 00:38:34.510 Brad Davis: Looking at these sensitivity checks within flavor-specific categories, again, all these results, are… are showing very, very fairly consistent, point estimates. 259 00:38:34.900 --> 00:38:38.159 Brad Davis: And, do not change the… our conclusions. 260 00:38:41.450 --> 00:38:45.730 Brad Davis: And then, if we look over at this left-hand column here. 261 00:38:46.000 --> 00:38:56.829 Brad Davis: These are different models that we… that we use, or alternative models. So, starting with the baseline model, it's the stacked difference in difference using a longed outcome. 262 00:38:57.040 --> 00:38:59.429 Brad Davis: And then we show that a… 263 00:38:59.930 --> 00:39:03.260 Brad Davis: Conventional to a fixed effects model with a logged outcome. 264 00:39:03.720 --> 00:39:13.090 Brad Davis: the synthetic DID with the logged outcome, and then stacked difference and difference using… excuse me… using levels instead of logs. 265 00:39:14.490 --> 00:39:27.129 Brad Davis: All of these results are, across the different models, are directionally consistent, and there is a little bit of variation in the precision of the estimates, but, 266 00:39:27.130 --> 00:39:35.269 Brad Davis: the conclusions are… are the same. So if we just look at this first column, looking at… at the unit sales per capita, 267 00:39:35.350 --> 00:39:48.209 Brad Davis: All of these estimates are fairly similar in their magnitude, and actually, in two of these… of these four models, we actually get statistically significant estimates. 268 00:39:48.630 --> 00:39:51.999 Brad Davis: going to… The far right column. 269 00:39:52.310 --> 00:39:55.829 Brad Davis: Looking at cigarette sticks per capita. 270 00:39:56.220 --> 00:40:07.339 Brad Davis: Three of these models estimate about a 10% increase in cigarette sales, and they're one of them that the, synthetic difference-in-difference model is a… is a, 271 00:40:07.560 --> 00:40:09.580 Brad Davis: About a 20% increase. 272 00:40:10.490 --> 00:40:14.470 Brad Davis: So results are… are very consistent across the modeling. 273 00:40:14.650 --> 00:40:21.110 Brad Davis: Now, I do want to mention a few limitations of our data, or of the analyses here. 274 00:40:21.760 --> 00:40:29.119 Brad Davis: One, we're using sales data, so… and so we can't say anything 275 00:40:29.460 --> 00:40:33.160 Brad Davis: Or we don't… we don't know very much about the… 276 00:40:33.260 --> 00:40:43.039 Brad Davis: The margins across where these… these sales are changing, and so we can't say anything about… about youth, so, we can't distinguish between youth or adults. 277 00:40:43.050 --> 00:40:53.629 Brad Davis: We can't say anything about initiation or cessation. And then, participation versus the, intensity of use. 278 00:40:54.840 --> 00:40:57.900 Brad Davis: Another limitation of our data? 279 00:40:58.130 --> 00:40:59.580 Brad Davis: again, we… 280 00:40:59.800 --> 00:41:13.110 Brad Davis: Our data for vaping product sales is restricted to the gas and convenience store sector, and so the results may not generalize to sales in other retail channels. 281 00:41:13.600 --> 00:41:20.339 Brad Davis: So we tried to mitigate that a little bit with the Google Trends data, but still, it's a limitation of the paper. 282 00:41:21.280 --> 00:41:34.929 Brad Davis: And then also, these flavor restrictions do vary a little bit across province in terms of the scope of where sales are permitted, and then… and then what products are, exempt. And so… 283 00:41:35.770 --> 00:41:44.339 Brad Davis: Our estimates should be interpreted as the estimated effect as of the average of these policies, and not any one specific, 284 00:41:44.590 --> 00:41:49.199 Brad Davis: Of these, not any one specific flavor restriction. 285 00:41:51.240 --> 00:41:53.880 Brad Davis: So in summary, we find that 286 00:41:54.620 --> 00:42:11.860 Brad Davis: the nicotine vaping product flavor restrictions within Canada nearly eliminated sales within the flavored category, and then also eliminated sales within the mint and menthol category in those provinces that prohibited it. 287 00:42:12.230 --> 00:42:21.530 Brad Davis: And then we also find evidence of substitution, with large increases in the tobacco and unflavored 288 00:42:22.150 --> 00:42:30.800 Brad Davis: vaping product category. And then we also see evidence of substitution to combustible cigarette sales. 289 00:42:31.840 --> 00:42:51.179 Brad Davis: Now, if we remember looking back at the trends, trends were declining in both groups, and so when we say, cigarette sale, we estimate that cigarette sales increased by 9.6%. That should be understood or interpreted as it slowed down the decrease of, cigarette sales. 290 00:42:52.210 --> 00:42:54.420 Brad Davis: This… 291 00:42:55.060 --> 00:43:10.530 Brad Davis: Our findings in this study are consistent with what we see in the U.S. So this Friedman et al. paper looking at sales within the U.S. similarly finds that the vaping product flavor restrictions decreased 292 00:43:10.960 --> 00:43:14.169 Brad Davis: Vaping product sales and increased cigarette sales. 293 00:43:14.750 --> 00:43:19.050 Brad Davis: And so because of our… the findings of our paper, 294 00:43:19.310 --> 00:43:31.930 Brad Davis: It tells us that this pattern of substitution between vaping products and combustible cigarettes persists in Canada despite its stricter tobacco control environment, and therefore, 295 00:43:32.480 --> 00:43:38.100 Brad Davis: This pattern of substitutability, m-might, 296 00:43:39.500 --> 00:43:44.130 Brad Davis: Might extend to… to other countries as, of, as well, of varying, 297 00:43:44.260 --> 00:43:49.179 Brad Davis: Tobacco… ovarian Tobacco control environments. 298 00:43:51.060 --> 00:43:54.760 Brad Davis: So for policy implications, 299 00:43:55.880 --> 00:43:59.579 Brad Davis: Right, we present evidence that these flavor restrictions can 300 00:43:59.950 --> 00:44:10.299 Brad Davis: Shift consumption towards combustible cigarettes, which are more harmful, and could plausibly reduce welfare and harm public health. 301 00:44:10.920 --> 00:44:15.510 Brad Davis: Now there are… there's… Tucson, there's another… 302 00:44:15.790 --> 00:44:31.109 Brad Davis: two sides to the, to the public health equation. To understand the total impact or the net impact on welfare and public health, we would also need to know something about what is happening, what impacts they have on youth vaping behavior. 303 00:44:32.050 --> 00:44:48.559 Brad Davis: Youth are less likely to internalize the full cost of tobacco consumption, and therefore have more internalities associated with them, and more potential for reductions in welfare. 304 00:44:49.020 --> 00:44:51.210 Brad Davis: And so, 305 00:44:51.420 --> 00:45:03.090 Brad Davis: And so, again, if the full… the net impact on public… on welfare and public health, it will depend on, the impacts they have on youth vaping behavior, which… which we can't speak to in this study. 306 00:45:04.450 --> 00:45:19.950 Brad Davis: Now there are some studies, that, that have, that are looking at this, or have looked at this. So this first paper here was recently, was a… is… was recently published. It's a public health paper looking at, Quebec's flavor restriction. 307 00:45:20.300 --> 00:45:26.330 Brad Davis: They find that… that it was not associated with changes in youth vaping behavior. 308 00:45:27.370 --> 00:45:47.059 Brad Davis: There's some mixed results within the U.S, so Cotty et al. is a U.S.-based paper that, that finds that flavor restrictions reduce the vaping intensity of adolescents, but did not have impacts on… on overall vaping rates. 309 00:45:47.140 --> 00:45:55.309 Brad Davis: And then, Dr. Dove, our discussant, she also has, some work in this category, looking at, 310 00:45:55.610 --> 00:46:10.599 Brad Davis: flavored Tobacco restrictions and its impact on adolescents, with some mixed results there. She presented this paper a few years ago, and then also, a couple months ago, presented a more recent paper. 311 00:46:12.400 --> 00:46:24.299 Brad Davis: That… the recording from that presentation is not up right now, but it should be up at some point in the fall, so if you're interested in that, be on the lookout for that. 312 00:46:25.000 --> 00:46:32.750 Brad Davis: But, in conclusion, given the, the, the evidence here that 313 00:46:32.860 --> 00:46:48.760 Brad Davis: that these policies can, increase, sales of, combustible cigarettes. Policy makers should… that are considering adopting these policies should bring that into the… to the, 314 00:46:48.920 --> 00:46:55.900 Brad Davis: To the… to the thought process and… and the calculation of… of the potential impacts of these policies. 315 00:46:57.530 --> 00:47:00.009 Brad Davis: And that's all I have. Thank you, guys. 316 00:47:02.120 --> 00:47:03.220 Ce Shang: Thank you, Brett. 317 00:47:03.430 --> 00:47:06.359 Ce Shang: Let's turn to our discussant today. 318 00:47:06.540 --> 00:47:10.529 Ce Shang: Dr. Dove to see whether she has any comments. Thank you. 319 00:47:12.110 --> 00:47:19.609 Melanie Dove: Thank you. These are, really interesting results, and thank you for including our work in your discussion. 320 00:47:19.950 --> 00:47:23.830 Melanie Dove: Let's see, I just have a few questions. 321 00:47:25.780 --> 00:47:33.619 Melanie Dove: Let's see, it does seem like the increase in the cigarette sales is a little bit smaller than the… 322 00:47:34.060 --> 00:47:38.009 Melanie Dove: You know, the decrease in the flavored vape product sales? 323 00:47:38.400 --> 00:47:49.080 Melanie Dove: Is this something that should be emphasized or taken into account, and is… is this something that, you know, the U.S. studies have found as well? 324 00:47:51.740 --> 00:47:57.940 Brad Davis: Hmm. So, I mean, one thing we do on our paper is we… we estimate the… 325 00:47:58.370 --> 00:48:09.420 Brad Davis: Whether the nicotine content from the change in vaping sales is equivalent to the nicotine content from cigarette sales. 326 00:48:09.850 --> 00:48:23.379 Brad Davis: We estimate that it's roughly equivalent. Now, there's a lot of assumptions built in, and so, you know, it's hard to say for certain, but based on that, and that's in the paper, so you can read that if you're 327 00:48:23.590 --> 00:48:27.919 Brad Davis: Read about that if you're more interested in it. 328 00:48:28.150 --> 00:48:31.740 Brad Davis: So I think that's… that's… it's roughly consistent now. 329 00:48:31.940 --> 00:48:39.079 Brad Davis: As far as in the US, I… I'm not… I'm not sure. I think… 330 00:48:39.200 --> 00:48:43.190 Brad Davis: So… so, Dr. Friedman, who, 331 00:48:43.670 --> 00:48:51.119 Brad Davis: was one of the co-authors on that paper. She might be here, she might be able to comment on that, but I'm not certain offhand. 332 00:48:53.670 --> 00:48:55.500 Melanie Dove: Okay, thank you, that, 333 00:48:55.640 --> 00:49:02.110 Melanie Dove: Yeah, the calculation that you did for the, yeah, the nicotine content, that sounds interesting. 334 00:49:04.570 --> 00:49:15.519 Melanie Dove: Let's see, and then, I was just wondering, you mentioned that some of the provinces allowed, still allowed, mint and menthol vapes? 335 00:49:15.670 --> 00:49:23.519 Melanie Dove: And I was wondering if, yeah, if there was a different result for the cigarette sales among those 336 00:49:23.640 --> 00:49:29.100 Melanie Dove: Provinces that allowed, mentor mental vibes. 337 00:49:30.080 --> 00:49:35.810 Brad Davis: Yep, so we do… we do a… a one-province at a time analysis. 338 00:49:36.030 --> 00:49:37.270 Brad Davis: And… 339 00:49:37.650 --> 00:49:55.979 Brad Davis: You know, we thought that we might see, if they… if they exempted the mint category, that we'd see, less… a less substitution to combustible cigarettes. It wasn't really clear to us that there was any pattern. Now, across the… across all 340 00:49:56.170 --> 00:50:06.789 Brad Davis: provinces that adopted a flavor restriction, the… the coefficient was positive for cigarette sales. But again, it… the magnitudes did not really align with 341 00:50:07.090 --> 00:50:10.450 Brad Davis: With any one specific policy type. 342 00:50:12.330 --> 00:50:13.120 Melanie Dove: Okay. 343 00:50:13.350 --> 00:50:14.969 Brad Davis: So, I don't… yeah. 344 00:50:20.280 --> 00:50:20.880 Melanie Dove: Okay. 345 00:50:21.100 --> 00:50:27.589 Melanie Dove: I think I saw… did Dr. Friedman, I think… Did she respond? 346 00:50:27.590 --> 00:50:40.550 Ce Shang: Yes, she said, it's… to your prior question, this is hard to parse with sales data, since we cannot distinguish intensive versus extensive margin changes. 347 00:50:43.250 --> 00:50:43.870 Melanie Dove: Okay. 348 00:50:47.110 --> 00:50:52.289 Melanie Dove: I mean… Okay, I just had one more question, 349 00:50:53.010 --> 00:50:58.699 Melanie Dove: So you, so there's this… your results show that, 350 00:50:58.820 --> 00:51:03.369 Melanie Dove: You know, there's this… this, you found an increase in cigarette sales. 351 00:51:03.620 --> 00:51:06.230 Melanie Dove: I was just wondering, you know. 352 00:51:06.450 --> 00:51:15.229 Melanie Dove: Can you recommend, like, how would you modify the policy, or can you recommend a different policy to 353 00:51:16.070 --> 00:51:21.299 Melanie Dove: That… that might be more effective in reducing to Tobacco use or sales. 354 00:51:24.430 --> 00:51:41.160 Brad Davis: Hmm, I don't… I don't know, that's a really, a really challenging question, just because, you know, I… I think we also need… need evidence, more evidence on the impacts on… on youth and their vaping behavior. I mean, I think given the… 355 00:51:41.220 --> 00:51:45.460 Brad Davis: the increases in combustible sales, I… 356 00:51:45.540 --> 00:52:03.949 Brad Davis: I would be cautious, more cautious about adopting a… a more comprehensive policy that… that bans it across locations. Like, I would be more in favor of some exemptions for mint or mintball categories, and just maybe be a little bit more hesitant about, 357 00:52:05.650 --> 00:52:20.079 Brad Davis: about more comprehensive policies, but again, I think we also need more evidence on… on the impacts on youth vaping behavior, to really be able to understand the net impacts of the policy. 358 00:52:23.250 --> 00:52:23.890 Melanie Dove: Okay. 359 00:52:24.100 --> 00:52:27.680 Melanie Dove: Is there something in particular that you think would help? 360 00:52:27.800 --> 00:52:31.760 Melanie Dove: For the evidence on youth vaping behavior. 361 00:52:32.370 --> 00:52:48.360 Brad Davis: Well, so, I mean, there is some, right, we have done some studies in the U.S. on this work. I think, and particularly, you know, there's not much within Canada, so I think that would be valuable, but also just to see, you know. 362 00:52:48.580 --> 00:52:57.829 Brad Davis: initiation into the… to the vaping product market among youth. Right? If we see that these policies are really 363 00:52:58.130 --> 00:53:01.490 Brad Davis: Are really effective at reducing initiation into the vaping 364 00:53:01.590 --> 00:53:05.119 Brad Davis: The vaping market, then, 365 00:53:05.570 --> 00:53:20.870 Brad Davis: you know, then that might add to some of the public health and welfare benefits of the policy, and at that point, you know, you're just… you're measuring, you know, the trade-off in the two policies. 366 00:53:22.030 --> 00:53:25.930 Brad Davis: So, you know, and I… I think… 367 00:53:27.200 --> 00:53:46.529 Brad Davis: you know, something I recently thought of that, you know, you asked about maybe other changes to make. Something specific to Canada, and I don't know how much this would make a difference, but Canada has fairly restrictive policies on… on, nicotine pouches. 368 00:53:46.780 --> 00:53:58.890 Brad Davis: Right? And, during… during the time of this study, nicotine pouches were not available at all, at least not to be sold legally within Canada, and they only recently became legal. 369 00:53:59.900 --> 00:54:03.640 Brad Davis: You know, in the presence… 370 00:54:03.930 --> 00:54:21.280 Brad Davis: of nicotine pouches, consumers… some consumers might substitute from… from vaping to pouches rather than to combustible cigarettes. So that, you know, that's also a policy dimension as well. Now, in the U.S, 371 00:54:21.500 --> 00:54:26.940 Brad Davis: You know, nicotine pouches are much more broadly available. 372 00:54:27.720 --> 00:54:32.120 Brad Davis: And we see that substitution pattern as well. 373 00:54:32.740 --> 00:54:38.839 Brad Davis: So… Yeah, I mean, so that's… that's another policy dimension to… to consider. 374 00:54:40.110 --> 00:54:42.330 Melanie Dove: Yeah, yeah, that's really interesting. 375 00:54:42.560 --> 00:54:55.980 Brad Davis: Yeah. And then again, you know, it presents another trade-off, because nicotine pouches also create a new opportunity for more people to initiate into the nicotine and tobacco market, so… 376 00:54:57.810 --> 00:55:00.090 Melanie Dove: Right, right. 377 00:55:01.150 --> 00:55:05.040 Melanie Dove: Yeah, thank you for that, and… 378 00:55:05.530 --> 00:55:09.699 Melanie Dove: Yeah, those are all of the questions I had. 379 00:55:09.820 --> 00:55:11.030 Melanie Dove: Pretty steady. 380 00:55:11.290 --> 00:55:12.379 Brad Davis: Yep, thank you. 381 00:55:13.280 --> 00:55:18.200 Ce Shang: Thank you. So, I see that there is some discussion in the Q&A, but I think Mike 382 00:55:18.490 --> 00:55:21.740 Ce Shang: Had most questions answered. 383 00:55:21.950 --> 00:55:28.079 Ce Shang: Okay, so I see another question from Chris. Can we still make a statement, though? 384 00:55:28.120 --> 00:55:44.729 Ce Shang: Using, Alcot's estimate of the relative harm of vaping versus smoking about the net public health impact of this policy, using net changes in cigarette versus vape sales that you estimate, or does this require too many assumptions? 385 00:55:46.470 --> 00:55:49.539 Brad Davis: I'm sorry, could you, could you read that question again? 386 00:55:49.540 --> 00:55:53.979 Ce Shang: It's about the calculation of the net public health. 387 00:55:54.260 --> 00:56:04.019 Ce Shang: Eugene Hunt, I'm not probably pronouncing his name wrong. Alcott A-L-L-C-O-T-T, do you remember the paper that, 388 00:56:04.020 --> 00:56:04.690 Brad Davis: Yes, yes. 389 00:56:04.690 --> 00:56:05.170 Ce Shang: doing the. 390 00:56:05.170 --> 00:56:05.510 Brad Davis: Oh. 391 00:56:05.510 --> 00:56:09.079 Ce Shang: harms of vaping versus smoking? Yeah. 392 00:56:09.650 --> 00:56:15.889 Brad Davis: Yup. Yeah, I mean, I do think, you know, that there are some… 393 00:56:16.040 --> 00:56:21.990 Brad Davis: potential for public harm due to the increases in combustible, cigarettes. Now, 394 00:56:22.200 --> 00:56:27.030 Brad Davis: It also depends on where the reduction in vaping is taking place. 395 00:56:27.190 --> 00:56:28.470 Brad Davis: So… 396 00:56:29.030 --> 00:56:39.689 Brad Davis: you know, as I've mentioned, if the reduction in vaping is taking place among adolescents, which, you know, could be, that some of it is. 397 00:56:39.920 --> 00:56:44.320 Brad Davis: Right, there could be some public health benefits from that as well. 398 00:56:44.810 --> 00:56:52.129 Brad Davis: Right, doing this exact calculation is, is, probably beyond the scope of the paper. 399 00:56:52.840 --> 00:56:55.710 Brad Davis: But, you know, perhaps something to think about a little bit more. 400 00:56:57.630 --> 00:57:07.409 Ce Shang: Sure, and Doug Levy has shared a recent paper on the potential mortality impacts of vaping versus smoking initiated during adolescence. 401 00:57:07.560 --> 00:57:15.870 Ce Shang: So that could also help understanding, the findings in the context. So I think we are about time. 402 00:57:16.160 --> 00:57:19.709 Ce Shang: So I'll let our MC, Pingchuan, to take us out. 403 00:57:20.120 --> 00:57:23.509 Ce Shang: Thank you, Brad, for the great presentation. Thanks. 404 00:57:26.400 --> 00:57:45.450 Pingchuan Ma: Thank you, Ce and Brad. We are out of time, so if you still have burning questions or thoughts for Brad Davis, you can join us for Top of the Tops, an interactive group discussion. To join, please copy the Zoom meeting room URL posted in the chat, and switch rooms with us once this event concludes. 405 00:57:45.450 --> 00:57:52.109 Pingchuan Ma: We'll leave the webinar room open for an extra minute after the end to give everyone a chance to copy the URL. 406 00:57:52.280 --> 00:58:08.169 Pingchuan Ma: Which is beat.ly slash chopsmeeting, all in lowercase. Thank you to our presenter, moderator, and discussant. And finally, thank you to the audience of 140 people for your participation. Have a chops-notch weekend.