Using the R software and the zip codes of active licensed testing labs and distributors, we generate a map that shows the geographical location of licensed labs and distributors in mid-2019 . We have no information about which labs served which distributors. However, we expect that labs are better able to compete for nearby distributors because they would have lower transportation time and cost and may be more likely to have closer business relationships. In order to estimate average transport costs from distributors to labs, we randomly assigned distributors located within a 160 mi radius to each lab. Based on 2019 data, this was the longest travel distance from a distributor to the nearest lab. This travel distance radius ensures that each distributor in the sample is covered by at least one laboratory. Based on the annual number of samples that we estimate each lab is able to test , we estimate the share of total testing done by small labs, medium labs, and large labs. We then estimate the number of distributors per lab. In each of our 1,000 simulations, 70% of the 49 licensees with specific locations were randomly chosen to represent small-scale labs, 20% were randomly chosen to represent medium-scale labs, and 10% were randomly chosen to represent large-scale labs.The minimum capital investment in testing equipment needed to satisfy regulations is substantial. We estimate that in small labs , capital investment in equipment is about $1.1 million; in the medium-sized labs , capital investment in equipment is about $1.8 million; and in large-scale labs , capital investment in equipment is about $2.8 million. These capital costs, amortized over a 10-year time span with a 7.5% rate of depreciation and interest, vertical growing system represent less than 15% of total annual expenses. Annual costs of operating range from $1.4 to $2.2 million for small labs, $2.7 to $3.7 million for medium-sized labs, and $6.2 to $8.1 million for large labs. Consumables are the largest share of total annual costs in large-scale labs, whereas labor is the largest share of costs in small-scale labs. In medium-scale labs, consumables and labor have about equal shares of annual costs. Different-sized labs differ in their capacity and efficiency.

Large-scale labs test about four times the amount of cannabis per hour than medium labs, and more than 10 times what small labs test. The cost advantage of large testing labs comes from a more efficient use of inputs such as lab space, equipment, and labor. Table 6 summarizes the average of estimated testing capacities, annual costs, and testing cost per sample for each of the three lab size categories. Cost of collection, handling and transport also vary by lab size. As of April 2018, the longest distance between a lab and a distributor in California was about 156 miles. Fig 3 shows the cost of collection, handling, and transportation per sample for distances between labs and distributors, of less than 156 miles. As expected, the longer the distance, the higher the sampling cost. Large labs have relatively low sampling costs even at long distances. The highest possible sampling cost we assume for small labs is about $35 per sample if the distributor is located 156 miles away . On average, costs of collection, handling, and transportation represents a small share of total lab costs per sample. Fig 4 shows the distribution of full testing cost per sample from 1,000 Monte Carlo simulations assuming 49 labs. Variability of the cost per sample within small labs is high, with the highest and lowest cost within that group differing by $463. The difference between the highest and lowest costs in large labs is $88, with a lowest cost per sample of about $273. The average full cost per sample tested is about $313 for large labs, $537 for medium labs, and about $778 for small labs . Large cost differences per test and per batch document the large-scale economies and differences in operational efficiencies across labs of difference sizes. The aggregate amount of cannabis flowing through licensed labs in 2019 remains relatively small relative to the anticipated amounts expected in the future. That means labs that may anticipate growth, operate well below capacity.

Substantial scale economies suggest that, as the market settles, the smallest labs must either expand to use their capital investment more fully, leave the industry, or provide some specialized services to distributors that are not accounted for in the analysis presented here. Simply put, the average cost differences shown in Table 6 or the simulated ranges displayed in Fig 4 should not be understood as a long run equilibrium in the cannabis testing laboratory industry.In 2018, the first year of mandatory testing enforcement, according to official data published by the California Bureau of Cannabis Control and posted publicly on its website, failure rates in California averaged about 5.6% . Failure rates for the first seven months of 2019, the second year of the testing regime, have averaged 4.1% . We assume a 4% failure rate for the current market in California. By comparison, in Washington State, in 2017, the second year after the testing began, 8% of the total samples failed one or more tests. The Colorado Marijuana Enforcement Division reported that during the first six months of 2018, 8.9% of batches of adult-use cannabis failed testing, with infused edibles and microbial tests for flower accounting for the most failures. Batch size significantly affects the per-pound testing cost of cannabis marketed, especially when batch size is smaller than 10 pounds. Fig 5 shows the costs of one pound of cannabis marketed coming from different sizes of batch flowers using 0%, 4%, and 8% rejection rates. As rejection rates increase, the differences between the costs per pound of testing different batch sizes decreases. For example, given a 0% rejection rate, the cost of testing per pound of cannabis marketed from a one-pound batch is about 27 times higher than the cost of a 48-pound batch; on the other hand, given an 8% rejection rate, the cost of testing per pound of cannabis marketed from a one-pound batch size is only seven times higher than the cost from a 48-pound batch size.In this paper, we use a simulation model to estimate the costs per pound of mandatory cannabis testing in California.

To do this, we make assumptions about the cost structure and estimated the testing capabilities of labs in three different size categories, based on information collected from market participants across the supply chain. For each lab, we estimate testing cost per sample and its share, based on testing capacity, of California’s overall testing supply. We then estimate a weighted average of the cost per sample and translate that value into the cost per pound of cannabis that reaches the market.We use data-based assumptions about expected rejection rates in the first and second round of testing, pre-testing, and the remediation or processing of samples that fail testing. Our simulations rely on information collected from several sources, including direct information from testing labs in California, price quotes from companies that supply testing equipment, interviews with cannabis testing experts, data on testing outcomes for cannabis and other agricultural products from California and other states, data on pesticide detection in California crops, and data on average wholesale cannabis batch sizes. Costs needed to start a testing lab that meets California regulations depend on the scale of the lab. As lab scale rises, testing capacity rises faster than do input costs, so average costs fall with scale. We find that a large lab has four times the total costs of a small lab but 10 times the testing capacity, in part because large labs are able to use their resources more efficiently. Testing cost per pound of cannabis marketed is particularly sensitive to batch size, how to dry cannabis especially for batch sizes under 10 pounds. Testing labs report that batch size varies widely. The maximum batch size allowed in California is 50 pounds, but many batches are smaller than 15 pounds. We assume an eight-pound average batch size in the 2019 California market, but we expect that the average batch size will increase in the future as cultivators become larger and more efficient and take advantage of the opportunity to save on testing costs .Testing itself is costly, but losses inflicted by destroying cannabis that fails testing is a major component of overall costs. Low or zero tolerance levels for pesticide residues are the most demanding requirement, and result in the greatest share of safety compliance testing failures. Cannabis standards are very tight compared to those for food products in California. A significant share of tested samples from California crops have pesticide residues that would be over the tolerance levels established for California cannabis . Some foods that meet pesticide tolerance established by California EPA may be combined with dried cannabis flowers to generate processed cannabis products . Pesticide residues coming from the food inputs may generate detection levels of pesticide over the tolerance levels set by cannabis law and regulation, even if they are otherwise compliant as food products. Tobacco has no pesticide tolerance limits because it is considered to be an inedible crop used for recreational purposes. Cannabis has multiple pathways of intake, such as edibles, inhalable, patches, etc., and also may be prescribed for people with a health condition, searching for alternatives to traditional medicine. Some labs report that when samples barely fail one test, they have a policy of re-testing that sample to reduce the probability of false positives.

Some labs have reported up to 10% in variation in test results from the same sample. Some labs indicate that about 25% of samples need to be re-tested to be sure that results are accurate. Such concerns have been widely reported. In July 2018, some producers voluntarily recalled cannabis products after receiving inconsistent results of contaminant residues from different laboratories; and some California labs have also been sanctioned by the Bureau of Cannabis Control for failing state audits on pesticide residue tests. A major issue for legal, taxed and licensed cannabis market is competition with cannabis marketed through untaxed and unlicensed segment. Higher testing costs translate into higher prices in the licensed segment. Safety regulations and testing may improve the perceived safety and quality of cannabis in the licensed segment, thus adding value for some consumers. However, price-sensitive consumers move to the unlicensed segment when licensed cannabis gets too expensive. A useful avenue for further research is to investigate cannabis testing regulations and standards across states to assess implications for consumer and community well being and competition with unlicensed cannabis. Compared with other agricultural and food industries, the licensed cannabis industry in California has relatively little data. Banking is still done in cash, and sources of government financial data are less available for cannabis than they are for other industries. As the licensed cannabis segment develops, we expect that increased access to data on the market for testing services, including on prices, quantities, and batch sizes. Data from tax authorities, the track and-trace system, and the licensing system will then help clarify the costs and implications of mandatory cannabis testing.Industrial hemp ., has been an agronomically important crop since 2700 BC in China. Today, it serves a purpose in a variety of different industries, such as pharmaceuticals, nutraceuticals, textiles, composite materials, bio-fuels, foods, cosmetics, and hygiene products . Hemp is one of humanity’s earliest domesticated plants going back to the Neolithic times in parts of East Asia . Hemp is the non-psychoactive form of Cannabis, differentiated from marijuana only by having less than 0.3% tetrahydrocannabinol concentration in dry mass . In 1970, the Controlled Substances Act was passed in the United States, which stated that all Cannabis sativa, psychoactive or not, was a Schedule 1 drug with “high abuse potential with no accepted medical use; medications within this schedule may not be prescribed, dispensed, or administered” . The passage of the Controlled Substances Act forbade any individual from researching or growing Cannabis in any form, including hemp, and it was not until forty-eight years later with the passage of the 2018 Farm Bill that researchers and growers could again study and grow hemp. With 48 years of absence from the scientific literature, the renewed interest in hemp as a crop with high agronomic value has stimulated significant research activity.