In the Weeds: State Taxation of Cannabis

Introduction

The Center on Budget and Policy Priorities estimates that the COVID-19 pandemic will lead to a $500 billion collective state budgetary shortfall through fiscal 2022.¹ As it stands, rainy day funds and federal aid will only cover a fraction of this gap, leaving a $360 billion target for creative state revenue generation. Some politicians previously opposed to cannabis legalization are now warming to the possibility as a means of chipping away at this funding gap.² As a Pennsylvania state senator put it, “Everything is on the table.”³

It is perhaps unsurprising that states facing budgetary shortfalls are weighing the possibility of recreational cannabis taxation as a revenue source. Most, if not all, of the 11 states that legalized cannabis in the last decade cited revenue as a leading factor supporting legalization. Maybe more surprising, however, is the degree to which states vary in how they tax marijuana — from the type of tax levied, to the timing of the tax in the supply chain, to the ultimate rate borne by consumers. Eight years of state experimentation have yielded no uniform approach to cannabis taxation, nor a settled understanding of its optimal design. The same is true of how states choose to allocate the resulting revenues.

What can these eight years of lessons teach us about the viability of marijuana legalization and taxation as a means to addressing COVID-related budget shortfalls? Our analysis proceeds in three parts. Part I describes existing state cannabis taxation schemes, comparing the ways in which state legislators sold cannabis legalization to the voting public with the reality of how cannabis is now sold to the buying public. Part II attempts to define an optimal cannabis tax rate, considering two competing definitions between which states must choose: a revenue-maximizing rate, and a socially optimal rate that minimizes the externalities associated with cannabis use. Finally, Part III discusses both how a cannabis tax should be administered and how the proceeds should be spent. We conclude that, while cannabis is unlikely to be the revenue boon that legalization proponents once touted it to be, states can maximize its impact by implementing a price-based tax regime and allocating revenues between targeted funds and the localities in which sales are made.

I. A Descriptive Account of State Cannabis Taxation

A. How It Was Sold to the Voting Public

Since 2012, 11 states and the District of Columbia have legalized recreational cannabis.⁴ Colorado and Washington went first, touting high potential tax revenues as a key selling point,⁵ and every state since has cited revenue as a main factor supporting legalization. Importantly, however, states also typically cite a number of other factors — we’ll call them “plus factors” — in favor of legalization, primarily the public health and safety benefits from eliminating illegal sales; curbing discriminatory enforcement of marijuana-related crimes; and, perhaps implicitly, the growing perception that cannabis and alcohol pose similar public health risks and thus warrant similar treatment and that criminalization is not an effective deterrent for use and abuse.

Accordingly, most state legalization referenda included conditions on how tax revenues from cannabis sales could be spent, earmarking funds toward spending programs congruent with these plus factors. In Colorado, for example, the ballot initiative specified that “the first $40 million in revenue raised annually by [a marijuana] tax [would] be credited to the public school capital construction assistance fund.”⁶ Oregon allocates 40 percent of its cannabis revenues to schools and 20 percent to its “Mental Health, Alcoholism and Drug Services fund.”⁷ In Alaska, 50 percent goes to a “recidivism reduction fund,” while another 25 percent is allocated to a “marijuana education and treatment fund.”⁸ And Illinois, the most recent state to legalize marijuana, will spread revenues across a number of targeted funds for, among other things, processing cannabis expungement petitions and administering drug treatment, mental health, and legal aid and re-entry programs.⁹

There is good evidence that such spending conditions were smart politics. Research indicates — and the above discussion illustrates — that “voters will react more favorably to a tax on an item or activity when revenues are earmarked for an expenditure that benefits consumers of the item or participants in the activity, or one that remedies a harm that the item or activity causes.”¹⁰ Since disagreement exists over whether to legalize, earmarking toward plus factors may act as a political chip to placate the skeptical and secure the necessary votes. But, as we discuss in Part III, earmarking may make sense as a matter of policy as well as politics. To see why, we begin by examining the on-the-ground reality of marijuana taxation.

B. How It Is Sold to the Buying Public

The choice to bring marijuana into the legal market was predominantly sold to the voting public as a way to raise revenues to support various social programs. In practice, however, configuring a successful taxation scheme has proved difficult, and for many states, cannabis is not the cash cow it was expected to be.

States face downward pressure on marijuana tax rates from multiple sources. First, more recent states to have legalized cannabis have generally opted for lower rates from the get-go.¹¹ As such, first movers like Colorado and Washington have faced a decision: either drop rates to compete or lose demand to cross-border substitution.¹² Other substitution effects — such as legal-to-illegal and recreational-to-medicinal substitution — make assessing demand difficult and have contributed to inaccurate revenue forecasts. The first year of legal cannabis sales in California, for example, generated just a third of the revenue projected by analysts.¹³ Conversely, Nevada’s inaugural year of sales outpaced forecasts by 40 percent.¹⁴ Revenue from “sin taxes” is “notoriously volatile and difficult to predict,”¹⁵ and cannabis is no exception.

Combatting substitution from the legal to the illegal market is of particular concern for states. For one, curbing illegal sales was a common plus factor touted by legalization advocates: Large illegal markets pose significant public health concerns. And second, legal-to-illegal substitution may generate huge revenue losses for states. A 2019 study found that “illegal sellers outnumber legal and regulated businesses almost 3-to-1” in California.¹⁶ As such, 80 percent of marijuana sales in California take place on the illegal market, resulting in revenues four times those of its legal counterpart.¹⁷ Even in Colorado, one of the nation’s oldest legal marijuana markets, “state officials estimate that only 60 percent of the marijuana consumed . . . is sold legally.”¹⁸ Michigan, for its part, has adopted a lower tax rate than most states, citing illegal-market competition as the driving force.¹⁹ But while legal-to-illegal substitution is a primary concern for states, it only goes so far: Massachusetts lawmakers increased the 3.75 percent retail excise tax approved by voters — designed to compete with the illegal market — to 10.75 percent because they believed revenue projections lagged too far behind other states.²⁰ Crafting an optimal tax scheme is difficult when states face competing forces pushing rates in opposite directions.

Recreational-to-medicinal-market substitution further complicates the task of assessing demand and generating accurate revenue projections. State approval of medical cannabis consumption typically predates recreational legalization.²¹ As a result, new recreational markets must compete with existing medicinal markets, and medicinal-market purchases are universally taxed at lower rates — if they’re taxed at all.²² Because of this, making it too easy for recreational consumers to obtain medical marijuana could significantly decrease a state’s cannabis tax revenues. Despite the lower (or zero) tax rate, many commentators predicted that recreational markets would “cannibalize medical marijuana sales,” possibly because states have erected significant hurdles (and higher up-front costs) for would-be consumers to overcome before obtaining medical marijuana.²³ The results, at first, were mixed.²⁴ But now, legalization efforts result in almost immediate declines in medicinal consumption.²⁵

One mechanism for mitigating this substitution effect is increasing the costs associated with obtaining medicinal marijuana recommendations, or “cards.” In Oregon, medicinal consumption fell 65 percent from 2015 to 2019, and some attribute part of this freefall to the $200 annual cost associated with obtaining a medical card.²⁶ Less drastic declines occurred in Nevada and Colorado, where the comparable costs are $50 and $25, respectively.²⁷ But these upfront costs can only go so far to deter medicinal-market substitution. After Massachusetts closed recreational marijuana stores as part of its stay-at-home order amid the COVID-19 pandemic, the number of medical marijuana patients increased by roughly 14 percent.²⁸ The responsiveness of medicinal-market sales to shortfalls in recreational-market supply illustrates the fungibility of these two products: Consumers can switch from recreational to medicinal cannabis (and back) based on price and supply rather than medical necessity.

With these substitution effects in mind, Table 1 depicts state taxation schemes as of May 2020. Many of these regimes look different than those originally sold to the voting public. There is not necessarily a one-to-one match between the tax policies in the ballot initiatives legalizing cannabis and those in the subsequent legislation that implements legalization. Voters in Oregon, for example, passed a “harvest tax,” which state lawmakers later replaced with a 17 percent retail sales tax.²⁹ Other states have changed their tax schemes over time in response to concerns about ease of administration.

The revenue generated from these tax schemes makes up a miniscule percentage of overall state budgets, as depicted in Table 2. For localities empowered to share in the marijuana revenue stream, however, the revenue impact may be more profound. For instance, in Edgewater, Colorado, cannabis revenue reached $1 million in 2016, or one-sixth of the city’s total annual budget.³⁰ That’s roughly 18 times the percentage cannabis generates in revenue for Colorado’s total annual budget, and 46 times the average across all states with legal markets. And Cortez, Colorado, brings in roughly $850,000 a year in cannabis revenues for a more modest 6.2 percent of the city’s annual budget.³¹ Of course, not all cities will display such significant numbers.³² Still, we posit that optional local taxes can be particularly impactful and that a local focus might play an important role in the successful implementation of a legal market.³³

Table 1. Current State Taxation Schemes for Adult-Use Cannabis

Legalization Date	2019	2018	2016	2016	2016	2016	2014	2014	2012	2012
Type of Tax	IL	MI	CA	MA	ME	NV	AK	OR	CO	WAb
Wholesale excise tax	7%		Xa		Xa	15%	Xa		15%
Retail excise tax	10-25%a	10%	15%	10.75%		10%				37%
Retail sales tax		6%	7.3%	6.25%	10%	6.85%		17%	15%	6.5%
+ optional local tax	X		X	X (3% cap)		X	X	X (3% cap)	X
aTax rate or amount varies by type of marijuana product. bWashington also institutes a 0.484 percent gross receipts tax. Note: Vermont has legalized cannabis but has no tax system in place.

Table 2. 2019 Cannabis Tax Revenue Breakdown

State	Gross Revenue (in millions)	Percentage of Budget
Alaska	$19	0.16%
California	$629	0.29%
Colorado	$302	0.93%
Massachusettsa	$71	0.17%
Nevada	$99	0.67%
Oregon	$118	0.27%
Washington	$396	1.08%
Average	$234	0.36%
Note: Illinois and Michigan have not yet taxed cannabis for a full year, Maine started tracking tax revenues after January 2019, and Vermont has no tax system in place. aEstimate based on sales.

This is not to say, of course, that states should always leave it to localities (or that the federal government should always leave it to the states) to tax small-dollar items while reserving for themselves only larger revenue streams that significantly affect their budgets. Instead, we suggest that localizing the benefits of taxation makes sense when the product being taxed generates externalities that are borne largely by the locality in which consumption takes place. State and federal gasoline and alcohol excise taxes make sense because states and the nation bear many of the externalities associated with those products thanks to the interstate highway system, which carries pollution — and drunk drivers — between states. We should ask whether the same is true of marijuana, and we do so in the next section.

II. Finding an Optimal Tax Rate

To find an optimal tax rate, states must first decide which optimal rate to pursue: (1) the revenue-maximizing tax rate, or (2) the socially optimal tax rate. The answers to these two questions point in different directions. We first address the considerations underlying each of these two optimal rates, and then address how states should pick between the two.

A. The Revenue-Maximizing Rate

As a general rule, to maximize tax revenues from cannabis sales, states would like as large a share of purchases as possible to take place in the legal market rather than in the illegal or medicinal markets³⁴ and in their state rather than in competing states. As discussed in Part I.B., these substitution effects point toward lower overall tax rates: Higher rates will cause more consumers to substitute into either illegal or medicinal sales or out-of-state sales in states with lower rates. And this latter substitution effect will only grow stronger as more states legalize. One study estimated that legal sales in the portion of Washington along the Oregon border fell by 36 percent after Oregon legalized at a lower rate.³⁵ These effects may be a large part of the reason why the more recent states to legalize have tended to set lower rates than first movers, and why some first movers have debated lowering their rates in response.³⁶

On top of these cannabis-specific factors pushing rates down, a more general substitution effect is also in play. To the extent that there are substitutes for marijuana as a leisure activity that are not subject to additional vice taxes, a higher cannabis tax means people will simply buy more of other products that are taxed less, thus further reducing revenues. On this view, marijuana is just one leisure product among many. Too aggressive a tax rate will push consumers elsewhere, and revenues will fall.

Lastly, and on a more macro level, it is possible that increased cannabis consumption distorts a user’s labor-leisure calculus, decreasing labor supply in the process. This would decrease taxable wages and perhaps slow economic growth, putting a dent in the state’s tax revenues. Such an effect would point toward a higher tax rate to deter marijuana consumption. But it is far from clear that these distortions happen in practice. One study of older individuals found no effect of the legalization of medical marijuana on extensive labor supply (the decision whether to work), and significant increases in intensive labor supply (the hours worked by existing labor force participants).³⁷ Another study found that marijuana decriminalization resulted in significant increases in weekly earnings for Black adults, while leaving earnings for white adults unchanged.³⁸ Also, legal cannabis is a huge job creator: One study estimated that the industry would create more jobs than the entire manufacturing sector by 2020.³⁹ We could not locate any research on labor force participation among recreational users, but these data points, while imperfect, suggest that any labor-leisure distortion among new recreational users will likely be outweighed by the benefits of decriminalization and the economic growth driven by recreational sales.

On balance, the revenue-optimal cannabis tax rate must be low enough to keep substitution into substitute goods or illegal, medicinal, and out-of-state cannabis at bay, but not so low as to make revenue generation impossible. There’s a floor: The revenue-optimal rate must, of course, be greater than zero.⁴⁰ But if revenue maximization is the goal, states enjoy a good deal of flexibility: They can continue to adjust the tax rate up or down in search of ever-higher revenues.

B. The Socially Optimal Rate

A socially optimal rate, on the other hand, puts revenue generation on the back burner, and instead seeks to minimize the social costs created by marijuana usage.⁴¹ Many of the same considerations that pushed the revenue-maximizing rate lower will do the same here, with one critical difference: Because revenue isn’t the goal, there’s no longer any theoretical floor on the “optimal” rate. In theory, the socially optimal tax rate could be zero, or it might even be a subsidy.

Here, again, a low rate is necessary to undercut the illegal market. But in the socially optimal scenario, the impetus for curbing illegal sales is not the collection of incremental revenues. It is instead driven by the elimination of the harms associated with illicit marijuana consumption, such as cannabis abuse, drug-related violence, and the harmful effects of discriminatory criminal enforcement. As discussed in Part I.A., states frequently cited the possibility of eliminating these harms as a plus factor in support of legalization.

While the need to curb illegal sales pushes the socially optimal rate down, evidence about the public health impacts of cannabis legalization is more mixed. On the one hand, after legalization, Colorado saw marijuana-related hospitalizations, emergency room visits, and poison-control calls all nearly double.⁴² But on the other, one study of 19 states found that, in the year after the legalization of medical marijuana, alcohol consumption dropped, and traffic fatalities fell by between 8 and 11 percent.⁴³ Such research is still in its infancy, and states should closely monitor the health effects of cannabis legalization as we learn more over time.

Finally, states might incur more intangible, aesthetic externalities because of the simple fact of open cannabis use. The magnitude of such externalities, however, is likely small and is undoubtedly outweighed by the benefits of ending criminalization and its attendant discriminatory effects.

States thus face a tradeoff: Lower rates will move existing users from the illegal market to safer, legal channels — unequivocally a social good — but may also encourage consumers who were previously deterred by higher rates and would not have otherwise purchased on the illegal market to try marijuana for the first time. Consequently, the socially optimal tax rate must strike a balance between encouraging legal consumption for existing users and deterring new users, who bring along with them new social costs. And understanding the true extent of the social, public health, and aesthetic costs (or benefits!) caused by new users will be crucial for ascertaining the true socially optimal tax rate. The realities of the market will continue to evolve as legalized cannabis becomes more widely available and as more states legalize. As such, any optimal tax rate is likely to vary from state to state and to change in any given state over time. States should not fear altering course as new data regarding their tax choices becomes available.

C. How to Pick Between the Two

Both rate choices come with trade-offs, and, because of the necessary floor in place if revenue is the goal, the revenue-maximizing rate is likely to be higher than the socially optimal one. Too high a rate will keep sales on the illegal market. Too low a rate, however, might result in political blowback — as illustrated in Massachusetts — and could increase overall marijuana consumption. Whether these are acceptable consequences may depend on how the state sold cannabis legalization to the voting public. Was the concern only with maximizing revenue? This might be the case for states considering legalization as a means of recouping budget shortfalls caused by COVID-19. If so, the revenue-maximizing rate seems like the logical choice.

If states do opt for a revenue-maximizing regime, some substitution between the legal and illegal sales is bound to remain, and with that the social costs of an illegal market. To preserve a rate high enough to generate funds beyond the costs of administration, states will be forced to turn to enforcement to curb illegal sales. And on this point, they won’t be able to count on federal authorities. As one DEA agent told The New York Times: “We’ve got our hands full with the opioid epidemic to be honest.”⁴⁴ But in implementing such enforcement measures, states must remain “mindful [of the fact] that criminal enforcement in the past ha[s] disproportionately targeted people of color.”⁴⁵ As noted above, plus factors including curbing discriminatory enforcement of marijuana-related crimes have typically accompanied the revenue promise of legalization. States opting for revenue maximization should be careful not to solve one externality created by that choice (the continued existence of an illegal market) by means that perpetuate a different externality (discriminatory enforcement) that legalization had been intended to resolve.

There’s a compelling argument for tying the legislature’s hands to the promises it makes to voters — that is, for urging states to set a rate that aligns with the stated purpose behind the push to legalize. This position, for one, encourages politicians to initially say what they mean. Preventing a bait-and-switch may force politicians to internalize the costs of their promises. If legalization is enacted for a revenue-generation purpose, lawmakers should acknowledge and take responsibility for the fact that a revenue-maximizing rate will bring with it avoidable social costs that would be eliminated or minimized under the socially optimal rate. It is important to remember, however, that states’ and voters’ goals can change, so it is worth asking (and re-asking) what interests are served by the current rate and whether that status quo is acceptable.

III. Implementation Strategies

We turn now to two questions that states must answer when deciding how to implement such a tax: what sort of tax to levy, and how to allocate the resulting revenues.

A. How to Tax

As Table 1 shows, current state tax schemes vary between pure ad valorem systems, weight-tax regimes, and mixed approaches. Interestingly, most sin taxes are tied to product weight or quantity rather than price: cigarettes are taxed per pack; alcohol by the gallon; and soda by the ounce.⁴⁶ Cannabis taxation thus far has bucked this trend,⁴⁷ but there are exceptions. Maine recently instituted a hybrid system combining a weight-based wholesale excise tax with a price-based retail sales tax.⁴⁸ And some localities in California levy a quasi-weight tax, assessing per-square-foot taxes on growers.⁴⁹

Our optimal tax rate discussion informs the choice of how to tax. Since the revenue-maximizing and socially optimal rates are different, the choice in tax design should leave flexibility to switch between the two as states experiment and learn more about their cannabis markets. Weight-based taxes are more complex than price-based taxes, requiring states to determine both how much tax is appropriate ex ante, where to assess this tax in the supply chain (e.g., at cultivation, processing, or retail), and to establish a system for measuring, reporting, and verifying product weights. These burdens proved too much for Washington, which discarded its weight-based tax regime for a single retail excise tax.⁵⁰ Weight-based taxes also lead to potency distortions. Producers can vary the THC level of marijuana plants without increasing the plant’s weight, and products with greater THC levels typically sell at a higher price. As a result, producers are encouraged to sell more potent products to maximize profit margins under a weight-based regime.⁵¹

Despite these facts, some commentators have recently argued that new states legalizing marijuana should opt for a weight-based tax over a price-based regime.⁵² Underlying their position is that weight-based taxes result in more stable revenues for states in the long run, especially in the face of falling cannabis prices.⁵³ But states can just as easily account for revenue fluctuations in a price-based system: for example, by placing a percentage of cannabis revenues into a trust fund when revenues are high, the state creates a buffer that can otherwise compensate for low revenue periods. Or, even more simply, states can adjust the price-based tax up or down as prices change to keep revenue constant. If states earmark cannabis revenues toward specific social programs, they may be particularly interested in achieving a stable revenue stream. But stability alone is no reason to opt for a weight-based tax regime over an ad valorem system.

To avoid the administrative headaches inherent in a weight-based system, states could impose a price-based tax solely on retail transactions. Under this approach, the tax is fully transparent to the end consumer and does not result in a tax on a tax.⁵⁴ On at least one metric, however, a manufacturer-level excise tax may perform better than a retail tax. Maine levies a weight tax at cultivation to combat illicit sales — the tax establishes a record of when a marijuana plant is first produced, “making it harder for [it] to disappear along the way.”⁵⁵ No such record is formed if the tax is levied only on retail, so such a regime may be less effective at combating illegal sales.

In this sense, perhaps a single excise tax based on price at the wholesale level is ideal. Since the entire supply chain for legal marijuana exists in the state — at least until legalization goes federal — the tax will match the location of consumption. This strategy will require greater administration costs (versus a pure retail tax) to combat vertical-integration distortions.⁵⁶ In response to a cultivation tax (whether based on price or weight), firms might vertically integrate (i.e., operate both as wholesaler and retailer) to avoid a wholesale-retail transaction and, thus, the associated excise tax. After Washington changed its tax from a weight-based cultivation tax to a single retail tax, “the amount of cannabis produced by non-vertically integrated firms increased by 42 percent” — a finding that the Tax Foundation described as prima facie evidence of vertical-integration distortions.⁵⁷ But slightly higher administrative costs may be a small price to pay for striking the balance between preserving flexibility in the system and more effectively curbing illicit sales. Either way, a price-based tax at the retail or manufacturer level is preferable to a weight-based tax to allow for experimentation in the search for optimization.

B. Who Should Tax

With a path charted toward implementing a successful marijuana tax, we must now consider who is the best tax collector — states or localities? The answer to this question lies in understanding which entity bears the brunt of the externalities caused by cannabis usage. We suggest that states and localities share these externalities, and thus propose that states create a mechanism for localities to share in cannabis tax revenues.

Broadly speaking, the externalities of marijuana usage that we have discussed fall into three categories: economic, health-related, and aesthetic. The effects of economic externalities, like the labor-leisure distortion, reached beyond the location of consumption, and suggest that the state is the best tax collector. But as discussed above, in the case of cannabis, these externalities are likely to be small if they exist at all. As to health-related externalities, state and federal governments, rather than localities, likely bear most of the cost, as they fund most publicly provided healthcare. And there is evidence that legal markets place an increased burden on public health infrastructure. State-level tax collection can promote more centralized responses to these larger problems. Nevertheless, at least some health consequences of marijuana usage will fall on the locality in which users reside — for example, cannabis users who get sick may miss school or work, imposing costs on local education systems and businesses. Finally, the aesthetic externalities arising from cannabis use, to the extent that they exist, are borne largely — if not entirely — by the localities in which cannabis is used and sold. Residents that disapprove of public consumption of marijuana may oppose local dispensary permits at the outset or stop frequenting parts of town in which dispensaries are located. At the extreme, they might even move somewhere that does not permit cannabis sales.

Health-related and aesthetic externalities likely outweigh economic effects, and these costs are likely to be borne by states and localities together. The health effects of cannabis are in turn more significant than the aesthetic costs, which suggests that states rather than localities should collect most of the tax. But granting localities their fair share is still important because localities do bear some real externalities, and local participation in the legal market is critical for successfully curbing illegal sales. Combining this reality with the fact that tax revenues from cannabis sales will likely be more impactful at the local level leads to the conclusion that cannabis revenues should be allocated to a combination of (1) targeted funds that counteract these externalities and (2) direct transfers to localities.

Limited localization more closely matches revenues to the externalities associated with marijuana usage and bolsters efforts to combat illegal sales. By granting localities a larger portion of revenues, more localities will likely permit legal sales. As such, legal marijuana becomes more accessible for more people, discouraging a resort to illegal channels. Importantly, widespread licensing could achieve greater illegal-to-legal-market substitution at higher tax rates than more dispersed licensing. If the nearest dispensary is 5 miles rather than 50 miles away, consumers may be willing to pay slightly higher prices on legal sales before defaulting back to the illegal market. This would compound the impact of marijuana taxes on local budgets. And that higher rate might help deter some first-time users, allowing states to capture a larger share of existing marijuana demand without artificially spurring new demand and thus incurring additional health-related externalities.

These proposed transfers to localities can be effectuated in one of two ways. First, states could allow localities to charge their own local tax on marijuana sales.⁵⁸ And second, states could remit part of state-collected revenues back to the localities in which sales are made.⁵⁹ Either method encourages localities to license dispensaries and thus participate in the resulting revenues.⁶⁰ The first option would also allow localities to experiment with different schemes, providing more data in the hunt for an optimal tax rate. On the other hand, allowing localities too much latitude in setting rates might simply exacerbate the race to the bottom, transforming a problem of cross-border substitution into one of cross-county substitution. Perhaps this is why some states like Massachusetts opted for a middle-ground solution — allowing localities to charge optional local taxes but setting upper and lower bounds on those rates.

Of course, if the goal is purely to close state budget shortfalls caused by COVID-19, this localization may seem off base. But as discussed in Part I.B., evidence from states that have legalized thus far indicates that marijuana revenues alone will not close the gap much. The revenue game alone is not worth the (hemp) candle. In this light, transfers to localities might make a good deal of sense. Localities may be better suited than states to direct funds toward people who need them in a crisis, and if revenues help fund local infrastructure or education projects, perhaps these localities will be better prepared for the next pandemic.

Conclusion

The financial repercussions from COVID-19 will be long-lasting. Finding new ways to raise funds is imperative to protect state solvency, and cannabis taxation may be an attractive place to start. Competing goals behind marijuana legalization and political constraints push tax rates in opposite directions, making an optimal rate difficult to determine, to say nothing of the fact that states must decide which optimal rate to pursue: the rate that maximizes revenue, or the rate that minimizes social costs. New states hoping to legalize cannabis considering current societal conditions should learn from existing taxation schemes to determine the best path forward. We suggest that part of this path include taxes based on price rather than weight or potency, and a focus on local control and revenue sharing. If localities bear many of the social costs associated with marijuana usage, they should also receive a piece of the revenue pie to aid in mitigating externalities stemming from legalization. Legal cannabis is a new industry, and states deciding whether and how to legalize must consider myriad competing factors. By diving into the weeds here, we hope to assist them in this task.