This study examined marijuana dispensary density in relation to neighborhood characteristics and local marijuana policies in Washington State from 2014 to 2016. Consistent with previous research (Amiri et al., 2019; Firth et al., 2020; Shi et al., 2016; Tabb et al., 2018), marijuana dispensaries were disproportionately located in CTs with greater disadvantage. Dispensaries appeared to open in CTs with particularly high levels of disadvantage in 2014 and then disperse across the state while retaining higher concentrations in disadvantaged urban CTs. The association between dispensary density and race depended on urbanicity. Marijuana dispensary density was higher in non-White CTs in urban areas but lower in such neighborhoods in rural settings. State and local marijuana policies emerged as important predictors, underscoring the importance of designing thoughtful and equitable license allocation procedures and policies.

A central finding was that marijuana dispensary density was higher in CTs with greater disadvantage, more people of color, and markers of disorder. This is consistent with previous analyses that concluded that recreational dispensary density was higher in CTs with more deprivation (Amiri et al., 2019), poverty (Tabb et al., 2018), people of color (Shi et al., 2016), and crime (Shi et al., 2016). However, our findings suggest that marijuana dispensaries are not indiscriminately opening in disadvantaged areas. Specifically, dispensary owners seem to avoid locations with violent crime and vacant properties that may deter customer business by making them feel unsafe. In addition, marijuana licensees may seek locations near off-premise spirits outlets, assuming the two businesses will share a customer base. Alternatively, the co-location of marijuana dispensaries and off-premise alcohol outlets may result from local governments using similar zoning and land use approaches to the two businesses (Thomas and Freisthler, 2016). Finally, the association with the presence of a highway suggests licensees prioritize locations with convenient access for customers. While our results describe dispensary locations, stakeholder interviews with licensees would complement these data to better understand selection of available land parcels. This information could inform efforts to equitably spread dispensaries across communities with consideration of business performance.

Several potential explanations for why dispensary density would be higher in disadvantaged CTs. The WSLCB allocation procedure that preferenced locations with more medical marijuana users may have inadvertently assigned more marijuana licenses to deprived CTs if marijuana demand was higher in them (Morrison et al., 2014). At a more local level, NIMBYism (“not in my backyard-ism”) may have also been at play, particularly in 2014 when the first dispensaries opened and associations with disadvantage were the strongest. NIMBYism often occurs when people organize to prevent adult entertainment (e.g., strip clubs, alcohol outlets) or otherwise stigmatized businesses (e.g., methadone treatment centers) from opening in their neighborhoods, often to allay fears about amenity harms and reducing amenity property values (Morrison, 2013). NIMBYism may occur by neighborhood residents protesting new marijuana licensing applications on a case-by-case basis, or it can be formalized via zoning and land use regulations or city/county bans/moratoria on marijuana dispensaries, which may disproportionately occur in locations with more resources (Matthay et al., 2022). As of 2016, 65% of WA residents lived in a city/county with marijuana zoning laws, and 30% lived in cities/counties with a permanent or temporary ban on marijuana sales (Dilley et al., 2017). Available commercial parcels that conform with local and state laws may be more common in deprived neighborhoods, as has been shown with alcohol outlets (Hippensteel et al., 2019). Given the racist history of zoning and land use policies, future research should explore whether common dispensary location policies, such as restricting them to commercial zones and establishing minimum distance requirements, could play a role in the disparate distribution of marijuana dispensaries. However, these policies are likely not the only explanation; an analysis that accounted for such land use and zoning provisions in Portland, Oregon concluded that there would be 73% more dispensaries in two otherwise equal neighborhoods that differed by one unit of disadvantage (Firth et al., 2020). Emerging evidence suggests that dispensary density is associated with marijuana use prevalence and frequency (Ambrose et al., 2021; Everson et al., 2019), including daily/near-daily use of adolescents and young adults (Shih et al., 2021). Considering that most dispensary sales are to daily/near-daily users (Kerr and Ye, 2022; Kilmer et al., 2014), the geographic patterning of marijuana dispensaries may hold public health consequences for those who live nearby.

After our study period, WA’s social equity program considered disadvantage directly when they awarded 44 licenses that were not issued, cancelled, or forfeited. A redress for the disparate burden of the war on drugs shouldered by communities of color, these equity-based licensing programs aim to allocate profits of the new legal cannabis sales to disproportionately impacted areas (Kilmer et al., 2021). WA defined social equity applicants as people who fulfilled at least two of the following three criteria: 1) lived in a disproportionately impacted area (e.g., high poverty, unemployment) for at least five years between 1980 and 2010, 2) was or had a family member arrested/convicted of a cannabis-related offense, and 3) had a household income less than the state median ($82,400). While access into this new industry ought to be equitable, it is unknown whether these programs will result in more dispensaries opening in disproportionately impacted areas, possibly unintentionally compounding existing harms.

As expected, policies shaped the distribution of marijuana dispensaries across WA. The WSCLB’s algorithm that originally allocated licenses to cities and counties granted more licenses to jurisdictions with higher rates of medical marijuana use (Caulkins and Dahlkemper, 2013). This suggests that states’ decisions about how to initially distribute licenses will determine what the marijuana retail environment looks like within their borders; these decisions ought to be considered carefully. City- or county-level thresholds on marijuana dispensary density were also associated with the number of dispensaries in the CT. These policies were associated with fewer marijuana dispensaries during the initial phase of legalization in both urban (2014) and rural (2014 and 2015) CTs. However, these limits were associated with more marijuana dispensaries in urban areas in 2016. While this may seem counterintuitive, this finding may suggest that these policies are in place in the urban areas with the greatest demand and most likely to encounter oversaturation in the future. It is also worth noting that four of the ten cities that passed thresholds on marijuana dispensary density set those thresholds at the number of licenses allocated by the WSLCB, while the other six were more restrictive (Dilley et al., 2017).

This study contains several limitations. First, while CTs are proxies for neighborhoods, they may not fully represent the neighborhood in which people live and dispensaries are located. Second, we only considered retail marijuana density. Future studies should also consider other dimensions of access, such as product and economic availability. Third, we utilized the OFM classification of rural counties. A more granular classification system with RUCA codes is useful for areas with similar characteristics, but the OFM county-level classification systems offered stability over time.

The marijuana industry is relatively new, yet rapidly growing. Currently, 21 states including Washington D.C. have legalized recreational marijuana use, yet the authors are aware of only four other studies on the distribution of recreational marijuana dispensaries in the U.S. to guide policymakers (Amiri et al., 2019; Firth et al., 2020; Shi et al., 2016; Tabb et al., 2018). There are a host of legislative, practical, and social equity concerns regarding recreational marijuana licensing that will determine how dispensaries distribute in communities and who bears the brunt of any harms that may manifest as a result. Marijuana licensing commissions ought to proceed cautiously and evaluate their efforts, such as by monitoring the number of licenses according to neighborhood disadvantage as the recreational market commercializes.