Purpose of the Analysis

Category Descriptions

State Regulations

The Data

Numbers of Active Licenses

Numbers and Types of Licenses by Licensee

Conclusions

 

Purpose of the Analysis

US states that have legalized cannabis activity issue cannabis licenses to individuals or businesses for different types of activities, each of which is performed at a different stage of the supply chain: Grow, Process, Distribute, and Retail. These activity licenses may be held by individuals or organizations either singly or in combinations to achieve different configurations of operations: 

• Single-license businesses (e.g., one Retail operation), 
• Horizontally integrated businesses (e.g., multiple Retail operations), and/or 
• Vertically integrated businesses (e.g., joint Grow and Retail operations).

The ways in which cannabis companies choose to configure operations are determined by two primary factors:

• Regulations, including the numbers of licenses issued for each activity and regulations about how activities may, must, or cannot be performed, and
• Economic forces– supply (cost) issues and demand issues.

The configurations companies choose to undertake determine how companies compete with each other in the industry, which affects the eventual accessibility of cannabis for customers, that is, the variety in products, forms of use, and prices. 

This analysis examines cannabis licenses issued to industry suppliers in an attempt to answer questions such as:

1. Do states differ in their relative quantities of companies that Grow, Process, and Retail?
2. How do companies choose to organize? Are they independents? Horizontally integrated? Vertically integrated?
3. In companies that are integrated, which activities are combined? Do these patterns of integration differ across states?

Category Descriptions

Four Types of Suppliers

At its most basic level, the cannabis supply chain consists of four distinct sets of suppliers (see Figure 1):

• Growers (Cultivators): Grow and harvest cannabis plants.
• Processors (Extractors or Manufacturers): Extract cannabis oil from cannabis plants and/or process cannabis plant material and oil into end use products.
• Distributors (Wholesalers or Transporters): Transporters transport cannabis products between facilities in the cannabis supply chain. Wholesalers and Distributors aggregate cannabis products across suppliers, store products, and/or transport them to Retailers.
• Retailers (Dispensaries): Sell cannabis products to Customers (Cannabis Users). Increasingly, states are also issuing licenses for Delivery of cannabis products to Customers.

Figure 1

Since cannabis is a Schedule I drug, the legal status of cannabis within a state is wholly determined by the state. States that have legalized cannabis use require all cannabis activity to take place within state borders. That is, no cannabis raw materials, working products, or end use products may cross state lines.

The skills and capital equipment necessary to successfully manage each activity (Grow, Process, Distribute, Retail) within the supply chain are very different. A priori, then, the natural economic configuration for the industry would seem to be that each activity is provided independently of the others. At the same time, certain types of market environments may lead companies to combine certain types of activities, so as to benefit from improved efficiency and/or profitability. The most prominent factors the lead companies to combine operations are the following. 

Economies of Scale: To the extent that there are economies of scale — that is, it’s cheaper or more efficient to increase the volume of operations — then companies will integrate horizontally (i.e., within activities, across locations). Economies of scale exist when there are large fixed costs of operations. In the case of cannabis, there are usually large costs (money and/or resources) associated with acquiring licenses, property, and equipment and complying with regulations. In this case, cannabis businesses may choose to integrate horizontally to scale operations and thus decrease average cost per unit of revenue.

Economies of Scope:  To the extent that there are economies of scope — that is, it’s cheaper or more efficient to combine two successive activities together, such as growing and extracting cannabis, within the same operation — then companies will integrate vertically (i.e., across activities). Economies of scope tend to occur when the production process for separate activities require a common input, such as a common piece of equipment or a common skillset. Alternatively, to the extent that it’s costly to transport products between successive stages of the supply chain, then companies will either co-locate and/or vertically integrate. 

Supply Chain Uncertainty: To the extent that there is supply chain uncertainty, that is, there may be shortages at a particular stage of the supply chain, then companies will integrate into performing that activity themselves, so as to minimize supply chain disruptions. In cannabis, if supply of cannabis product is uncertain, Retailers may integrate backward into Growing to ensure supply. If store shelf space (Retail) is in short supply, Growers may integrate forward into Retail. If there are too few Processors (e.g., extractors) to process cannabis, Growers may integrate forward into Processing. If there are too few Transporters to move product between activities/locations, companies may integrate into Transportation/Distribution.

Four Types of Configurations

Configuration Definitions

A company that wants to Grow, Process, Distribute, or Retail cannabis products within a state that has legalized these activities must possess a license to do so. Different states have different regulations regarding the extent to which a single company may or must acquire multiple licenses for a particular activity. For purposes of the analysis, I refer to the license configurations of companies as falling into one of four classes:

• Single-License Business: A company with one license performs one activity along the supply chain; that is, it either Grows or Processes or Distributes or Retails cannabis. An industry in which all suppliers are Single-License Businesses is unintegrated, that is, any company along the supply chain may buy from any company preceding it in the supply chain or sell to any company following it in the supply chain (see Figure 2).

Figure 2

• Horizontally Integrated (Multiple Licenses, One Activity): A company is horizontally integrated if it performs a single activity along the supply chain but at multiple locations, such as the Growers or the Processors in Figure 3. Chain stores, such 7-Eleven or McDonald’s, are examples of horizontally integrated Retailers. 

Figure 3

• Vertical Integrated (Multiple Licenses, Multiple Types of Activities): A company is vertically integrated if it performs multiple activities along successive stages of the supply chain. The last company in Figure 4 is fully vertically integrated, that it, is Grows, Processes, Distributes and Retails only products created by the company itself. 

Figure 4

• Both Horizontally and Vertically Integrated: A company is both horizontally and vertically integrated if it performs multiple activities along successive stages of the supply chain, and it also contains more than one operation for at least one of its activities.

Implications for Competition

Company configurations are important because they determine the nature of competition in the market. 

If market environments (the nature of regulations and economic forces) lead to smaller, unintegrated companies, there tends to be more competition, more variety, and lower prices. There also tends to be greater availability, i.e., a larger number of businesses. However, since unintegrated companies are less formally coordinated, there also tends to be more uncertainty, that is, more supply shortages and gluts at different points in the supply chain, potentially leading to disruption in supply of end use products.

If, on the other hand, market environments lead to larger, more integrated companies, then there tends to be less competition, so variety and new product innovation tend to suffer, customer service tends to suffer, accessibility tends to be lower (i.e., you have to travel farther to get to a dispensary), and prices tend to be higher. There also tends to be less availability (i.e., there are fewer products to choose from). On the other hand, economies (efficiencies = lower costs) realized by larger companies (integrated either horizontally or vertically) may possibly lead to greater accessibility and/or lower prices. Furthermore, horizontal integration provides leverage (vis-à-vis suppliers and customers), and vertical integration provides coordination and control. Greater leverage and/or coordination make it easier to ensure consistency of supply (in the case of product shortages) and/or access to shelf space (in the case of shortages in access to retail) or transportation (in the case of shortages in transportation or distribution).

For a more in-depth discussion on how company configurations shape competition, see my previous blog, How Regulations Shape the Cannabis Industry.

State Regulations

There are certain regulations on cannabis supply chain activities that are common across states. For example,

• Growers, Manufacturers, Retailers, Distributors (and Testing Labs) must all be licensed by Local (City or County) and State governments.
• Licenses to Grow cannabis limit the number of plants a licensee may grow.
• Any transport of cannabis samples or products between suppliers in the supply chain must be conducted by a Licensed Transporter or Distributor.
• Before being shipped to Retailers, samples from all cannabis products must be tested at a licensed Testing Lab for potency and contamination. In many states, Testing Labs may hold multiple licenses to conduct operations in different locations. However, in every state Testing Labs are prohibited from engaging in other cannabis-related activities. That is, Testing Labs are not able to Grow, Process, or Retail cannabis.
• All cannabis products must be tracked from seed to sale using state-approved tracking software, such as METRC.
• Medical cannabis customers must have a medical card issued by the state to purchase cannabis from Medical Retailers. Recreational customers only need an ID showing they are at least 21 years old.

Of course, state-specific regulations will play large roles in shaping differences across states in how cannabis companies choose to combine activities across the supply chain. It’s important, then, to understand how regulations regarding specific supply chain operations compare and contrast across states. While I have some familiarity with state regulations, a more complete investigation of differences between states in regulations that might help to explain any differences found in the analysis reported in this study are beyond the scope of the analysis. Nonetheless, there are a few tidbits I’ve discovered that provide flavor to how regulations may differ across states. 

• In 2017 California proposed a new license type, Processor, for processors separately located from Growers.^[1]Presumably, then, companies with a grow license can also process cannabis onsite. To the extent that other states require Growers to have separate licenses to process their own cannabis onsite, California may have fewer Process licenses than other states.
• When adult use cannabis was legalized in Colorado, the state initially required Retailers to grow themselves 70% of what they sold. The state used this requirement to limit cannabis activity by commercial growers for the first nine months of AU activity, thereby slowing initial growth.^[2]This regulation forced initial Retailers to vertically integrate backward into Growing. Since initial configurations tend to persist, Colorado may have more integration between Grow and Retail than other states without similar requirements.
• In Michigan, Growers that are co-located with Processors or Retailers may transfer cannabis product across facilities without the need for using a licensed Distributor.^[3]To the extent that transportation of cannabis products by licensed Transporters is costly, this regulation will encourage co-location, if not vertical integration, between Growers and Processors, Growers and Retailers, and Processors and Retailers. 

The Data

State cannabis agencies in the states listed in Figure 5 provide licensee information for each license issued.^[4]The variables defined in Figure 5 are:

• Med Legal: Year in which the state legalized medical cannabis activity
• AU Legal: Year in which the state legalized adult use cannabis activity
• Med vs. Adult Use: Do the state license data distinguish medical from adult use activity?
• Retail: Does the state provide data by licensee for retail activity?
• Grow: Does the state provide data by licensee for grow (cultivation) activity?
• Process: Does the state provide data by licensee for process (manufacturing) activity?
• Distribution/Wholesale: Does the state provide data by licensee for distribution and/or wholesaling activity?
• Transportation: Does the state provide data by licensee for transportation (B2B) activity?
• Delivery: Does the state provide data by licensee for delivery (B2C) activity?
• Testing: Does the state provide data by licensee for lab testing activity?
• Tiered Grows: Does the state provide data by licensee for different sizes of grow operations?

I concatenated these data across activity types and aggregated them to the company level. This aggregation of the raw license data provided information for unique companies on license counts by type of activity. 

Note 1: The list of current license holders in each state is a moving target, as new licenses are constantly being issued and old licenses expire or are revoked. The data I have are from mid-to-late 2023. 

Note 2: The analysis uses Grow license counts, which is problematic, since grow operations vary tremendously within and across states. Some states provide tiered licenses, that is, different license types depending on venue (indoor or outdoor) and/or size of operations (either plant counts or canopy size in square feet). But in the licensing data I used, most states do not distinguish size of operations – and the ones that do use different distinctions – so I simply counted total number of licenses issued. Also, the lack of sufficient breakout between Medical and Adult Use licenses precluded an analysis of licenses and licensees based on this distinction. 

Figure 5

Numbers of Active Licenses

Total Numbers of Licenses

States differ in the kinds of cannabis activities they choose to license. The most basic types of activity are: 

• Retail
• Grow
• Process
• Transport/Delivery
• Testing Lab

Other types of activity include:

• Distinguishing Medical Use from Adult Use
• Microbusiness License: Small business licenses to Grow, Process, and Retail in limited volumes
• Research Organization License: License permitting cannabis research 
• Integrated Activity License: License permitting some combination of activities, such as Grow and Process, which vary across states that issue these licenses
• Consumption Lounge, Hospitality, or Event: License to operate a facility with on-site cannabis consumption 

Figure 6 provides total counts of licenses for all cannabis activity partitioned by licenses for Retail (R), Grow (G), and Process (P) activities, Transport (T)/Delivery (D) activities, and all other activities. The vast majority of licenses in all states are issued for Retail, Grow, and Process activities. California and Massachusetts have nontrivial numbers of Transport/Delivery licenses. California has also issued several hundred licenses for Microbusinesses. The takeaway from Figure 6 is that by focusing the analysis primarily on Retail, Grow, and Process activities, we’re capturing the vast majority of cannabis activity.

(* Note that Oklahoma is a Medical Only cannabis state, while the other states are all Adult Use states)

Figure 6

Grow vs. Process vs. Retail

My first question was: How do total counts and proportions of licenses issued for each of the three basic activities (R, G, P) differ by state? Figure 7 displays license counts for each of these three activities across states.

Figure 7

Of course, there is huge state-to-state variation in population, so a more informative comparison considers per-capita license counts, as seen in Figures 8A and 8B. Note that license counts for Montana and Washington were only avail for Retail, but not Grow or Process. 

Figure 8A

Figure 8B

The distributions in Figures 7, 8A, and 8B provide the following insights:

• Large Total Absolute Numbers of Licenses

The states with the largest absolute numbers of licenses are Oklahoma and California. Oklahoma has perhaps the least stringent license requirements in the nation: any resident can obtain a license for $2,500.^[5] As for California, historically, the state has had a large legacy cannabis history, due to its naturally fertile grow environment. After 20 years of legalized medical cannabis in California, “there were an estimated 50,000 to 60,000 cannabis growers in California, ranging from 15-plant mom-and-pop grows in the back yard to thousand-plant-or-more grows in the mountains of the Emerald Triangle, Santa Cruz, and the foothills of the Sierra. And there were thousands of indoor grows in warehouses in all the major metropolitan areas.”^[6] It’s no surprise, then that California has issued so many Grow licenses.

• Relatively Large Total per-Capita Numbers of Licenses 

Vermont and Alaska especially, but also Montana and Maine have low absolute numbers of licenses but relatively high per-capita numbers. This is an unexpected mystery!

• Largest Total Per-Capita Licenses

On a per-capita basis, Oklahoma still stands out as the state with the most licenses. However, New Mexico rises to second place. New Mexico has no residency requirements to obtain a license and the same low license fees as Oklahoma of $2,500.^[7]

• Small Total Absolute Numbers of Licenses

States with small numbers of licenses have capped the number of licenses issued at low levels (New Jersey, Maryland, Illinois, Arizona, Connecticut). These states also (i) require vertical integration (businesses Grow, Process, and Sell) (New Jersey, Arizona) and/or (ii) charge relatively high license fees (Maryland, Illinois, Connecticut). ^[8]

• Proportions of Licenses

The states with greater numbers of per-capita licenses tend to have relatively more Growers and fewer Retailers, while states with fewer per-capita licenses tend to have relatively more Retailers and fewer Growers. At the same time, Figure 8A makes it clear that states vary widely on the per-capita numbers of both Retail and Grow licenses issued, which is perhaps the big takeaway from Figures 7, 8A, and 8B. 

Correlations

Figures 9A and 9B provide correlations across the different variables, where

• Lic/Pop (M) is the total per-capita number of licenses issued
• 2023 Pop (M) is the state population in millions
• Med Legal is the year in which medical use was legalized
• AU Legal is the year in which adult use was legalized
• Home Grow = 1 if consumers are allowed to grow cannabis at home; otherwise Home Grow = 0
• 2022 Sales ($B) is total estimated cannabis sales in billions of dollars 
• Total Licenses is total licenses issued across all categories
• Retail, Grow, and Process are license counts for each of the respective activities

Figure 9A

Figure 9B

The correlations in Figures 9A and 9B provide the following insights:

• States with larger populations tend to have greater cannabis sales, more cannabis licenses, but fewer licenses per-capita, that is, licenses scale less than proportionately with population. 
• The first states to legalize medical cannabis were also the first to legalize adult use. These states are also more likely to allow home grow, to issue more licenses and to have greater cannabis sales. 

The fact that states that were first to legalize also have greater sales makes sense, since presumably states with higher propensities to use cannabis legalized cannabis activity earlier than states with lower propensities. 

The fact that states that have more recently legalized cannabis use have lower sales and have issued fewer licenses is likely due to a combination of factors that are all likely to be interrelated: (i) it takes time to ramp up operations in newly legalized states, (ii) newly legalized states tend to limit license grants (at least initially) to control ramp up (iii) newly legalized states have stricter license requirements, i.e., either require vertical integration and/or charge high license fees. 

Ratio Comparisons

Let’s now see how states compare on cannabis sales and access to Retailers. Figure 10 compares 2022 per-capita cannabis sales and 2022 cannabis sales per Retailer across states. Figure 11 compares per-capita numbers of Retailers and numbers of Retailers per Grower across states. The figures are sorted from left to right on per-capita total Retail + Grow + Process licenses.

Figure 10

 

Figure 11

 

Figure 10 provides the following insights:

• Per-capita sales of cannabis vary substantially across states with Missouri and New Jersey at the lower end and Alaska at the higher end, followed by Colorado, Nevada, Montana, and Massachusetts. The first states to legalize adult use are Washington, Colorado, Oregon, and Alaska, which I would expect to have high per-capita sales, and, indeed, Colorado and Alaska top the list. However, the large per-capita sales in Montana are surprising, given the state legalized adult use cannabis only relatively late, in 2020.
• Cannabis sales per retailer vary substantially across states, with Arizona, Illinois, and New Jersey at the high end and Oklahoma, New Mexico, Montana, and Maine at the low end. 

Figure 11 provides the following insights:

• Per-capita counts of cannabis Retailers vary substantially across states, where Oklahoma, New Mexico, and Montana have the most Retailers per person, while New Jersey, Rhode Island, Maryland, and Illinois all have the least.
• The number of Retailers per Grower also differ substantially across states, but states tend to have either relatively low numbers of Retailers per Grower or relatively high numbers.
• The correlation between Retail/Pop and Retail/Grow is (0.37).

Figure 11 shows that Illinois and New Jersey had among the fewest retailers per capita, while Oklahoma and New Mexico had the most. The outcome displayed in Figure 10 is thus a reflection of the fact that when there are fewer Retailers, sales are more concentrated within those few, and when there are more Retailers, sales are spread out across those larger numbers of retailers (correlation = (0.27)) (see Figure 12). In other words, while states with higher sales tend to have more retailers, the number of retailers scale more quickly than sales do. 

Figure 12

In Figure 11, Retail/Pop is a measure of retailer accessibility for consumers. Having fewer per-capita retailers decreases accessibility for consumers, which decreases legal market sales (correlation = 0.41) (see Figure 13). So perhaps the surprisingly large sales in Montana are enabled by relatively high accessibility of retailers. Notable are the relatively low per-capita sales, together with low per-capita retailers in California. These two factors go hand-in-hand. More specifically, if California were to increase accessibility to legal market cannabis sales by enabling more retailers (as of Feb 2022, almost two-thirds (62%)  of California cities prohibit cannabis retailers^[9]), legal market sales would increase substantially.

Figure 13

In Figure 11, Retail/Grow is a measure of cannabis product accessibility, that is, the ability of cannabis product providers to be able to access shelf space in dispensaries. When Retail/Grow is high, product suppliers are more easily able to access shelf space, and thus consumers. This gives product suppliers relatively more leverage over retailers. Conversely, when Retail/Grow is low, product suppliers have trouble getting their products into dispensaries and thus being able to gain access to consumers. In these cases, shelf space is often rationed, such as by charging product suppliers for access (“slotting fees”), as is happening in California.^[10]

The negative correlation between the two data series in Figure 11, Retail/Pop and Retail/Grow, suggests that states with greater retail access also tend to be states with relatively fewer Retailers than Growers. I think this is more coincidence than reflective of some deeper insight. 

Numbers and Types of Licenses by Licensee

Now that we’ve compared total cannabis license counts by activity across states, we can move on to the licensee level data. Within each states, I aggregated individual licensees by business to see how businesses choose to combine different cannabis activities within a single organization.

Number of Licenses per Licensee

Figure 14 shows license counts per licensee (unique business) for both total licenses and  RGPT (Retail, Grow, Process, Transportation/Distribution) licenses.

Figure 14

The data displayed in Figure 14 indicate that the average number of licenses per unique business varies across states, with Maryland, Illinois, and Vermont having the lowest numbers and Nevada and Arizona having the highest. Note that Arizona encourages growers to be vertically integrated, which leads to inflated license counts per licensee relative to those in the other states. Since Nevada does not explicitly encourage integration, the high license counts per licensee in that state are notable.

The information in Figure 14 provide some sense of the extent of concentration of licenses. However, to get a better sense of the distribution, let’s look at licensees at the low end of the distribution (i.e., licensees with a single license) and those at the high end (i.e., licensees with the most licenses).

Number of Licensees with a Single License

Figures 15A and 15B looks at the total number of licenses issued for all cannabis activity in each state and asks: which portion of total licenses were issued to businesses that have a single license, to either Retail, Grow, Process, or Transport/Distribute cannabis? Figure 15A provides the portion of total licenses issued to a single licensee, while Figure 15B provides the portion of total licensees with a single license Figure 15A.

Figure 15A

Figure 15B

Figures 15A and 15B provide the following insights:

• Based on the display of license counts per licensee in Figure 14, we would expect Maryland, Illinois, Vermont, Alaska, and Oregon to have the largest portions of licensees with a single license. Indeed, they do. The vast majority of licensees in Vermont (over 80%) and two-thirds of licensees in Oregon and Maryland have single-license businesses. 
• At the other end of the spectrum, as expected, Nevada has the smallest portion of single-license businesses of the states examined. Consistent with the fact that Nevada stood out as having more licenses per licensee on average, it also has, by far, the least portion of single-license businesses.
• Of the single license holders, there are more Growers than Retailers in Vermont, Oregon, Alaska, and California, while there are more Retailers than Growers in Massachusetts and Maryland. As discussed earlier, these states might have greater imbalances than other states with respect to access. That is, Retailers in Maryland may have more problems keeping their shelves stocked with product, while Vermont, Oregon, Alaska, and California Growers may have more problems gaining access to shelf space in dispensaries.
• Single-license holders are relatively balanced between Growers and Retailers in Colorado, Connecticut, Illinois, and Maine. 
• Illinois is notable in having the largest portion of single-license businesses engaged in Transport. In particular, Illinois has a disproportionately large portion of Transporters relative to Growers and Retailers.

Generally speaking, states with more single license holders are likely to be more competitive, that is, provide greater product variety at lower prices than states in which licenses are more concentrated. At the same time, however, a market can have a lot of small business but still not be competitive, if a few small organizations are concentrated enough to dominate the market. 

Concentration of Licenses across Licensees

So now let’s look at the other end of the spectrum from the license holders with the fewest numbers of licenses (i.e., a single license): license holders with the most licenses. Figure 16A displays the portion of total licenses held by the top 5 businesses with largest numbers of licenses (Retail, Grow, Process, or Transport/Distribute), and Figure 16B displays the portion of total licenses held by the top 10 businesses. Each figure displays both the portion of licenses held by the top license holders, as well as the portion of total licensees these top businesses comprise. 

For example, in California: 

• There are 10,367 total licenses (for all license types)
• I have identified 4,293 unique businesses^[11](Retail, Grow, Process, or Transport/Distribute)
• The top five businesses hold a total of 700 licenses (Retail, Grow, Process, or Transport/Distribute)
• So then the top 5 businesses:
  • Comprise 5/4,293 = 0.1% of total businesses
  • Hold 700/10,367 = 6.8% of total licenses

Figure 16A

Figure 16B

Figures 16A and 16B provide the following insights:

• We saw previously that Connecticut has the fewest numbers of licensees, while California has the most. This explains the relatively large percentage accounted for by the top 5 and top 10 licensees in Connecticut and the relatively small percentage in California.
• A comparison of the two figures indicates that the top 5 licensees in each state have substantially more licenses than the next 5 largest licensees.
• While the top licensees in California each have large number of licenses, because the state has issued so many licenses in total, the largest licensees still have a relatively small portion of total licenses in the state. On the other hand, the top 5 licensees in Connecticut hold one-third of all licenses in the state and the top 10 licenses hold almost half, making the largest licensees in Connecticut very large fish in their small pond.
• Given the patterns in licenses per licensee for Nevada in Figures 14 and 15, I would have expected Nevada to be more concentrated at the top, but that’s not the case. In other words, in Nevada, most licensees simply have multiple licenses without too much concentration at the top – a much more balanced distribution – as compared with, say, California, which has a good number of single-license businesses at the bottom and also a couple of licensees at the top with disproportionately large holdings. 

An important caveat for the information presented in Figures 15 and 16 is that a tallying up of license counts does not account for differences in the size of operations. Most markets tend to be comprised of large numbers of small business with a few larger businesses that dominate the market. So just because there are a lot of licensees with a single license, the few licensees with many licenses can still account for the majority of activity. In other words, a more informative analysis would examine market shares by licensees. Unfortunately that information is not easy to come by.

As an aside, a surprising finding in the California market is that the top license holders, either alone or in combination with other individuals hold shocking numbers of licenses. In particular, the top 10 licensees hold a total of 862 Small Outdoor Cultivation licenses, a whopping 40% of all Small Outdoor Cultivation licenses issued in California, and these Top 10 licensees jointly account for 10% of all licenses issued in California, where the remaining 90% of licenses are held by over 4,000 other much smaller licensees.

Configurations of Licenses across Licensees

The last set of examinations in this analysis consider the configurations of licenses across licensees and seeks to understand which combinations of licenses are most prevalent. For purposes of tractability, the analysis is limited to Retail, Grow, and Process activities and does not include Transport.

Figures 17A and 17B compare counts of licensees for the different combinations of licenses held by licensees, where Figure 17B is a zoom-in of Figure 17A.

Figure 17A

Figure 17B

Figures 17A and 17B show that in most states, the majority of licenses are held by licensees engaged in a single type of activity, either Retail or Grow or Process.

Figure 18 compares proportions of licenses, where proportions are taken relative to only Retail, Grow, and Process licensees. Figure 18 sorts states from left to right based on total proportions of single-sector licenses (Retail Only + Grow Only + Process Only).

Figure 18

Figure 18 shows that over 80% of licensees are engaged in a single type of activity in 7 of the states, Illinois, Oregon, Vermont, California, Maryland, Connecticut, and Alaska. At the other end of the spectrum, in Nevada less than half of licensees engage in only one type of activity, and in Massachusetts just slightly more than half do.

Of the licensees who are engaged in more than one type of activity, that is, who are vertically integrated to some extent, Grow and Process, seems to be the most prevalent combination, though substantial portions of licensees engaged in more than one activity do both Grow and Retail. Finally, a nontrivial portion of licensees in Arizona, Massachusetts, Nevada, Missouri, and Maine are integrated into all three activities, Retail, Grow, and Process. Note that vertically integrated companies may also be horizontally integrated.

The last issue to consider is the extent of 

• Single-License Businesses: (R Only and 1 License) or (G Only and 1 License) or (P Only and 1 License), versus,
• Horizontally Integrated Businesses: (R Only and > 1 License) or (G Only and > 1 License) or (P Only and > 1 License), versus
• Vertically Integrated Businesses: (R & G) or (R & P) or (G & P) or (R & G & P).

This information is displayed in Figure 19, where states are sorted from left to right based on proportions of single-license businesses. 

Figure 19

The key insights gained from Figure 19 are:

• There is large state-by-state variation in cannabis company configurations. 
• There does not seem to be any clear correlation between per-capita sales and company configurations based on license counts. 

Conclusions

An analysis of cannabis activity licenses issued indicates that states vary significantly in the relative numbers of cannabis licenses issued for Grow, Retail, and Process activities and in the extent of integration of business within and across activities. The numbers of licenses issued for the different types of activities are determined solely by the state regulatory commissions, so the differences in issued licenses are regulatory in nature. However, the configurations of businesses are shaped by both regulatory and economic environments. The large differences in businesses configurations across states thus reflect large difference in both regulatory and economic environments.

Of the variables analyzed, the best predictor of cannabis sales is the total size of the population. If we assume the desired outcome is to maximize per-capita cannabis sales, then we could ask which combinations of licenses are best predictors of per-capita cannabis sales? Unfortunately, none of the variables examined are predictive of per-capita cannabis sales. This is probably due to the fact that what matters in not the numbers of licenses, but the activity levels per license (e.g., sales generated per dispensary or pounds of cannabis grown per grow licenses) and the distribution of that activity across licensees.

Since we don’t have market share information, the best we can do is look at how businesses choose to organize and assume that the existing configurations are efficient – or at least moving towards efficient configurations as the different markets continue to mature. 

 

 References

^[1] Proposed Cultivation Regulations: New License Type "Processor." Law Offices of Omar Figueroa. Retrieved from https://www.omarfigueroa.com/proposed-cultivation-regulations-new-license-type-processor/

^[2] Coulter, L. (2014, Oct 2). Colorado Rule Change Opens the Door for Wholesale Marijuana Growers. GGS Structures. Retrieved from ggs-greenhouse.com/blog/colorado-rule-change-opens-door-wholesale-marijuana-growers

^[3] Michigan Marijuana Manufacturing License. (n.d.). Michigan Cannabis Information. Retrieved from https://michigancannabis.org/licensing/manufacturing

^[4] My thanks to Keegan Skeate of Cannlytics for making these data available.

^[5] State-by-State Marijuana Policies. (n.d.). National Cannabis Industry Association. Retrieved from https://thecannabisindustry.org/ncia-news-resources/state-by-state-policies/

^[6] Nikki and Swami (2019, May 16). California’s Cannabis Industry Is Royally Messed Up — Just Look at the Numbers. Merry Jane. Retrieved from https://merryjane.com/culture/californias-cannabis-industry-is-royally-messed-up-just-look-at-the-numbers

^[7] State-by-State Marijuana Policies. (n.d.). National Cannabis Industry Association. Retrieved from https://thecannabisindustry.org/ncia-news-resources/state-by-state-policies/

^[8] State-by-State Marijuana Policies. (n.d.). National Cannabis Industry Association. Retrieved from https://thecannabisindustry.org/ncia-news-resources/state-by-state-policies/

^[9] California Department of Cannabis Control debuts data tool showcasing access areas for cannabis business. (2022, May 26). California Department of Cannabis Control. https://cannabis.ca.gov/2022/05/california-department-of-cannabis-control-debuts-data-tool-showcasing-access-areas-for-cannabis-business/

^[10] Schroyer, J (2022, Jul 26). California marijuana retailers charge product suppliers for shelf space, creating friction in the industry. MJBizDaily. https://mjbizdaily.com/california-marijuana-retailers-charge-product-suppliers-for-shelf-space-creating-friction-in-the-industry/

^[11] I identified unique business manually using business names, DBAs, and owner names. There is surely some errors in my method, but the analysis is accurate enough to provide some indication of business concentrations.