WSLCB - Board Caucus
(March 19, 2024) - WSIPP Report Review

2024-03-19 - WSLCB - Board Caucus - WSIPP Report Review - Takeaways
[ Event Details ]

Board members learned about the benefit-cost analysis of legal cannabis conducted by WSIPP researchers, including how agency staff could better collaborate on topics of mutual interest.

Here are some observations from the Tuesday March 19th Washington State Liquor and Cannabis Board (WSLCB) Board Caucus.

My top 5 takeaways:

  • WSIPP Senior Research Associate Amani Rashid discussed trends her team had noticed when looking at outcomes correlated with drive time to cannabis retail access, noting how staff “do our best to try to isolate the correlation" (audio - 12m, video).
    • Rashid told board members that when looking at Washington "relative to similar states," WSIPP officials “don't detect evidence of significant changes in reported cannabis use.” But she stipulated that research on retail access—“obviously a large part of” I-502 and common in legal cannabis states—“found…it's the retail component that tends to predict changes in outcomes more so than just legalization itself.” Although the first stores opened in July 2014, Rashid recognized it wasn’t a fully formed commercial market, and that it took years to roll out around the state. “I point this out because” that made early evaluation more difficult “to really speak to the true relationship between retail and outcomes in Washington state with this single point in time analysis,” she said. Rashid felt “given that different parts of Washington have different access even to this day we consider that the impact of retail on cannabis use may differ depending on that accessibility.”
    • Retail access was defined based on WSLCB licensee addresses and dates of first and last sale, Rashid stated. This information formed the foundation of their analysis, she added, since “the idea here is we want to see…where are retailers and roughly when did they start operations, and when did they close if they're no longer operational?” Using this, they developed an average drive time to reach a retailer “for the average resident in a geographic unit,” Rashid explained, using zip code and census data in “trying to find neighborhood or local effects.”
    • She then said that their report didn’t include tribal retail stores inside Washington, nor retail stores in neighboring Oregon which might be in comparable driving distances, acknowledging “those omissions could…lead, and likely do lead, to us underestimating what access actually is a little bit.” She remarked that “moving forward we'll see how much more we can take into account.” Her intent had been to look at changes in cannabis access in different communities over time and any “coinciding changes in, in cannabis use behavior and other related outcomes.”
      • In remarks to legislators in March 2023, Rashid also agreed with lawmakers there were other data limitations, like respondents’ willingness to truthfully report cannabis use when it was illegal, or how limited sample sizes made analysis of subgroups within a survey less reliable, leading researchers to forego dividing up some results.
      • In a conversation with Foster, Rashid acknowledged the report also didn’t account for local and online sales of hemp cannabinoid consumables.
    • Moreover, Rashid highlighted that "where retailers locate is not random,” and reflected a confluence of factors based on decisions of individual licensees.
      • While licensees have a number of factors to consider when siting their stores, Rashid didn’t mention the many state laws and rules, plus local ordinances that limit where stores can operate, or jurisdictions that ban them entirely.
    • She described looking “at demographic and socioeconomic characteristics” in zip codes with “an operational retailer within their boundaries and zip codes that don't, and even across these comparisons we do see some differences:”
      • “Zip codes with operational retailers versus those without tend to have a smaller percent population White, but a larger percent population Asian and Black…but they are quite small in magnitude, just a couple percentage points.”
      • “We also find that where there's an operational retailer, those zip codes have a higher percent population with a bachelor's degree, a higher percent renting as opposed to owning homes, and a higher median household income.”
      • Decisions around retail access weren’t random, and “retailers are systematically choosing where they are,” meaning she couldn’t “establish true cause-and-effect relationships. Our study doesn't say a retailer opening near residents caused increases in cannabis use. What we are able to do is we account for all of these differences, and we do our best to try to isolate…relationship between access [and] outcomes.”
    • Having established the constraints of the data used to define retail locations, Rashid went on to share the conclusions they had reached around retail access based on calculated average drive time.
      • She indicated, “our first look was just to look at how access relate[d] to cannabis use within the states” with survey data, “and we find that a shorter drive time to a licensed retailer…among legal age adults actually does relate to a greater percentage of reported cannabis use.” She offered the example that “if you reduce drive time to a retailer by 50% that corresponds to an 8.6% higher probability of reporting…heavy past month cannabis use…about 20 out of 30 days.” Rashid called the relationship between access and past month heavy use “small but significant nonetheless.”
    • Working from the assumption that “greater access does predict higher rates of use, we then look at other related outcomes,” such as fatal car crashes in the same area and substance use disorder diagnoses.
      • Rashid revealed, “and sure enough we do find that a shorter drive time to a licensed retailer relates to a modest increase in the number of drivers involved in a fatal crash…about six percent more drivers.” More importantly, Rashid asserted “about 40% of drivers in a fatal crash receive a blood test, and among blood tested drivers we found that a shorter drive time predicted a higher prevalence of drivers who test positive for THC [tetrahydrocannabinol].”
      • Looking at substance use disorder diagnoses in relation to retail access “among Medicaid claimants,” among that segment—”roughly 30% of Washington [was] on Medicaid”---they’d also found “shorter drive times to relate to a higher probability of cannabis use disorder, alcohol use disorder, and opioid use disorder…about a 50% reduction in drive time” showed a “2.3% higher likelihood of a cannabis use disorder diagnosis.” 
      • Besides drive time, Rashid said the report also evaluated retail density, “this would work a little different…the more and more retailers nearby the more competitive the market is,” potentially leading to “more variety in product and more advertising, and it's through those channels we might see that greater density could also relate to differences in cannabis use behaviors.” Looking at density as a measure, “we do see that that does predict higher…probabilities of cannabis use disorder diagnosis,” she reported.
      • In “one of the few cases where we were able to look at underage groups within the data…we did find that a shorter drive time did relate to a higher probability of cannabis use disorder diagnosis among adolescents aged 12 to 17,” said Rashid.
  • Rashid delved into findings around retail access and adolescent cannabis trends, comparing how proximity to stores related with academic outcomes and teenage behavior involving cannabis (audio - 4m, video).
    • Suggesting that cannabis access for adults had “more of a spillover" impact on youth behavior, particular among high schoolers, Rashid stressed “we're not thinking here that the adolescents are driving to the retailers and buying from retailers….there's no evidence that….that's happening.” However, she did believe it related to how likely a legal-aged family member or friend might, and she assumed “more retail means their parents [or peers] might have more in the house.” Rashid also speculated that indirect effects like seeing more cannabis advertising could be impacting underage utilization, “and last thing is just it being around, seeing more adults use it could possibly normalize cannabis use and lower perceptions of risk.”
    • Using Healthy Youth Survey (HYS) data, Rashid indicated that her team compared “students who attend schools near retailers to students who attend schools not near retailers,” and that “when you compare…10th and 12th grade, students that attended school nearby an operational retailer are more likely to report frequent past month use.”
    • Rashid said some comparisons used the entire high school population, with “students who attend a school nearby a retailer have a seven percent higher number of average monthly unexcused absences,” their “most robust finding.” She felt it was possible retail access related to “truant behavior, and we did find that students attending a school near a retailer have a 2.5% lower likelihood of graduating high school in four years.” However, this difference was “driven completely by rural region comparisons, so in urban regions, this doesn't hold,” Rashid added.
    • Rashid commented, “when we looked at the national landscape we didn't see much going on but zooming into the state looking at retail, there do seem to be some patterns between access and adverse outcomes.”
  • Board members posed a few questions to Rashid around partnering on future studies, policy impacts of WSIPP findings, and how to account for socioeconomic differences when evaluating outcomes.
    • Vollendroff was happy to learn Rashid had already connected with Okey, and wanted to better understand “how do things get on your radar for those” additional reports on cannabis. She answered the “sniff test” for their cannabis work was how it related to the required evaluations required by I-502, with health care and labor being key topics “called out” in the initiative. They also looked for subjects that would be used in WSIPP benefit-cost modeling, as well as what data they could find, “anything I don't have data for I can't study.” Timing and how “ambitious” studies were was also dependent on the overall budget and staffing situation at WSIPP, Rashid noted (audio - 3m, video).
      • Board Chair David Postman followed up to find out whether the WSIPP board authorized interim studies. Rashid indicated they were kept “in the loop” but did not exercise authority to prescribe researcher goals (audio - 1m, video).
    • Vollendroff then asked about the best way to use WSIPP reports in WSLCB policymaking. Rashid felt the burgeoning cooperation between her team and the Research Unit would help lead to a “thought partnership” where WSIPP could help parse their data, and WSLCB could give her “a lot of…knowledge about the data” her team gathered (audio - 1m, video).
    • Postman brought up the analysis of HYS responses, curious about accounting for socioeconomic differences, or “if there was a test area that didn't have a retail outlet but had the exact same…socioeconomic” conditions. Rashid explained that they looked to adjust for “measurable differences” like unemployment rates, “other socioeconomics indicators, population, things like that for” high school data, as well as “school characteristics themselves. So like the population breakdown of the school, or percent low income at the school, or teacher-to-student ratio.” She was aware a number of “economic indicators predict different substance and cannabis use behaviors. So we're accounting for a lot of it,” but “there's only so much we can do about that, but we do account for a lot of those relevant indicators” (audio - 2m, video).

Information Set