Content uploaded by Curtis Hall
Author content
All content in this area was uploaded by Curtis Hall on Dec 23, 2021
Content may be subject to copyright.
THC and the FDIC: Implications of Cannabis Legalization for the
Banking System
James D. Brushwood*
University of Arizona
Curtis M. Hall
Drexel University
Amanda Marino
San Diego State University
Roberto Pedace
Scripps College
Abstract
Using data from U.S. banks’ regulatory filings, this study examines how the legalization of
cannabis at the state level affects deposits and lending activity of banks in legalizing states.
Numerous states have legalized business activities related to the production, distribution,
and use of recreational cannabis, but these activities remain illegal at the U.S. federal level.
This creates significant regulatory uncertainty for businesses located in legalizing states,
particularly federally-insured banks. Using difference-in-difference estimates, we find
evidence of significant increases in deposits and lending after recreational cannabis
legalization. We attribute this finding to either limited concern about the perceived riskiness
of deposits surrounding cannabis legalization, or optimism that policies would be adapted
to limit banks’ risks from regulatory uncertainty.
Keywords: Banking; Cannabis Legalization; Regulatory Uncertainty
JEL Classification Codes: G20; G21; G28; H77; H73; I18
Abbreviations:
CSA: The Controlled Substances Act of 1970
D-in-D: Difference-in-Difference
FDIC: Federal Deposit Insurance Corporation
GDP: Gross Domestic Product
*Corresponding Author: 1271 Campus Delivery, Fort Collins, CO 80523-1271; +1 (979) 777-5161;
james.brushwood@colostate.edu
1
THC and the FDIC: Implications of Cannabis Legalization for the
Banking System
1 Introduction
This study examines the effect of state-level recreational cannabis legalization on
banks’ deposit and lending activity. As of April 2021, recreational cannabis use had been
legalized in seventeen states and the District of Columbia.
1
Social issues aside, the potential
economic benefits to state and local governments, along with private cannabis-related
businesses, present clear motivations for states to pursue legalization. Primarily, legalizing
these previously illegal activities allows governments to tax income from cannabis
businesses and enables the people involved in these businesses, such as employees,
entrepreneurs, and suppliers, to more fully participate in the economy. Placing the potential
magnitude of legalization in perspective, ArcView Group estimates the total economic impact
of legalizing cannabis at $14.8 billion in 2019, and projects a potential magnitude of $46.8
billion by 2025.
2
However, the ability of legal cannabis-related businesses (and the individuals
involved in them) to establish banking relationships, a key element of economic
participation, remains potentially hindered by the conflict between state and federal law.
Even as cannabis legalization is enacted at the state level, the production, distribution, and
1
As of April 2021, Alaska, Arizona, California, Colorado, Illinois, Maine, Massachusetts, Michigan, Montana
Nevada, New Jersey, New Mexico, New York, Oregon, Vermont, Virginia, and Washington had passed initiatives
to legalize recreational cannabis and its related businesses in some form. Our sample includes banks located
in the following states, as they began issuing licenses for legal retail sales of recreational cannabis prior to
January 2016 (effective dates in parentheses): Alaska (February 2015), Colorado (January 2014), Oregon
(October 2015), and Washington (July 2014).
2
From a 2021 executive summary from Arcview Group summarizing its report “The State of Legal Cannabis
Markets: 8th Edition” produced in partnership with BDS Analytics.
2
use of cannabis remains illegal at the federal level. Because banks are insured by the federal
government through the Federal Deposit Insurance Corporation (FDIC), they may be
hesitant to engage in banking relationships with cannabis-related businesses for fear of
compromising their federally insured status or risking criminal prosecution. If that is in fact
the case, this regulatory uncertainty has micro-level implications on banks’ profit by
generating frictions in deposit-loan activity and could subsequently have macro-level
consequences in the form of foregone economic growth.
3
Although the Obama administration provided guidance in the Cole Memorandum on
how banks can accept deposits from cannabis-related businesses through the promise of
non-intervention, banks may resist openly engaging in these relationships as the status and
reliability of the memorandum’s promises are perceived to be uncertain, especially under
changing regulatory environments and subsequent administrations.
4
In fact, Attorney
General Jeff Sessions rescinded the Cole Memorandum on January 4, 2018, so the threat of
federal prosecution remains unclear.4
Using financial data from bank regulatory filings with the FDIC from 2011-2016, we
examine the effect of state-level cannabis legalization on bank deposits and loans. We
estimate our models for deposits and loans across a sample of treatment and control banks
using a difference-in-difference (D-in-D) approach.
3
For example, Guiso, Sapienza, and Zingales (2004) find that local financial conditions and decisions (e.g.,
access to credit) can have a significant impact on economic growth in a highly-integrated economy.
4
Even after issuance of the Cole Memo in February 2014, many cannabis businesses and banking organizations
expressed uncertainty about the reliability of the memo’s assertions in the popular press (for one example, see
Kovaleski 2014).
4 Attorney General Sessions stated that "the previous issuance of guidance undermines the rule of law" (Wilber
and Frosch 2018). While the timing of this decision is outside of our sample period, the move validates the
uncertainty and concerns regarding the reliability of the Cole Memo. A January 4, 2018 article in the New York
Times reporting the Attorney General’s move specifically identifies the banking problem as the most pressing
concern (Savage and Healy 2018).
3
We first examine whether banks operating in states where recreational cannabis was
legalized experienced increases in deposits after legalization. While the legal implications of
accepting deposits from cannabis-related businesses are uncertain, the question of whether
these funds make their way into the banking system is an empirical question as several
possibilities suggest that local bank deposits would be positively affected by legalization.
First, banks may believe that the federal government will keep its promise not to prosecute
banks that do their due diligence, as specified in the Cole Memorandum and related Treasury
Department guidance, allowing cannabis businesses to legally establish banking
relationships (Chemerinsky et al. 2015). Second, bankers and potential depositors may
operate discretely with customers not specifically identifying themselves as cannabis
businesses and banks opting not to probe further.
5
Consistent with these considerations, we
find evidence that banks located in legalizing states exhibited higher levels of deposits after
legalization relative to banks in the same period in non-legalizing states.
Secondly, we examine whether the influx of potential deposits after legalization
resulted in greater lending activity. While prior empirical evidence suggests a positive
relationship between deposits and lending, we examine whether this relationship is present
when the available deposits are subject to regulatory uncertainty. If banks are concerned
that the FDIC will not insure deposits from cannabis-related businesses or that the federal
government may seize these deposits, banks may be more hesitant to lend these funds.
However, if banks are unconcerned about the potential risk associated with accepting
cannabis-related deposits or optimistic about the chances that regulations will adapt to the
5
Discussions with industry professionals familiar with the banking relationships of cannabis-related
businesses provide support for the two mechanisms identified in this study.
4
needs of legalizing states, then banks in legalizing states may show little, if any, hesitancy in
converting deposits to loans. Consistent with the latter explanation, we find that lending
activity increased for banks in legalizing states, and while increases in lending were less
sensitive to deposits compared to non-legalizing states, this difference is economically small.
Our study contributes to the literature by examining how businesses (specifically
banks) make operating decisions under legal uncertainty. In this case, the uncertainty is
rooted in a misalignment between state and federal law. Additionally, our focus on the
banking system is important because the ability to establish a banking relationship is vital
for economic participation (Beck, Demirgüç-Kunt, and Maksimovic 2004; Dupas et al. 2016).
Finally, studies about the economic effects of recreational cannabis legalization are
particularly timely, as the number of states legalizing and considering legalization is growing
rapidly.
Depsite regulatory uncertainty from the inconsistency between state and federal
laws, our results indicate that banking activity (deposits and subsequent loans) increase
considerably in legalizing states relative to non-legalizing states. While uncertainty can
result in overly cautious behavior and hinder economic activity, we do not find evidence of
this with cannabis laws and the banking industry.
6
If agents (banks) are risk averse,
increased uncertainty would require higher expected returns for participation in the risky
activity. Consequently, our results imply that legislative uncertainty pertaining to cannabis
and banking may be outweighed by significantly larger expected profits.
7
6
For example, prior studies have found firms limit investment around regulatory uncertainty (see Kingsley et
al. 2012; Lopez et al. 2017).
7
Romi et al. 2020 find that companies, including banks, charge large preimums to offset the risks of working with
cannabis-related clients.
5
2 Background and Motivation
The Controlled Substances Act (CSA) of 1970 made cannabis an illegal substance at
the federal level in the United States. While prior laws such as the Marijuana Tax Act of 1937
limited the distribution of cannabis, the CSA is the law that formally characterizes cannabis
as a Schedule I narcotic, defined as a drug with a high likelihood of addiction and no safe dose
(see Chemerinsky et al. 2015 for a more complete history). The CSA makes the manufacture,
possession, and distribution of Schedule I narcotics, such as cannabis, illegal and punishable
with sentences up to life in prison for the larger volume manufactures and dealers
(Chemerinsky et al. 2015). Despite numerous attempts to modify cannabis’ criminal status,
cannabis remains illegal under the CSA, and the Supreme Court has upheld Congress's
authority to regulate cannabis – even cannabis grown at home for the personal use of the
grower (Kamin 2014).
While the federal law concerning the legal status of cannabis remained unchanged,
states began easing cannabis laws in the mid 1990’s by passing state legislation legalizing
cannabis for medical use. In 2012, Colorado and Washington became the first U.S. states to
legalize recreational cannabis use by passing “adult-use voter initiatives,” which legalized
the possession of less than one ounce of cannabis. Other states followed suit, and by April of
2021, adult recreational cannabis use was legal in seventeen states and the District of
Columbia.
The move for states to legalize recreational cannabis is motivated by two primary
benefits: (1) revenue gains from the taxation of cannabis sales and (2) cost savings from
ceasing the enforcement of cannabis laws, the burden of which has historically fallen
6
disproportionately on state and local law enforcement (Chemerinsky et al. 2015). In 2017,
not long after legalization, Colorado collected $247 million in sales tax and license fees from
the cannabis industry. That amount accounted for 2.3 percent of Colorado’s general fund
revenue (Felix and Chapman 2018). If the sales revenue generating this type of tax revenue
is also generating deposits and loans, it could provide substantial stimulus in other sectors
of state economies.
Recognizing the potential benefits of recreational cannabis legalization but also the
conflict between state and federal laws, the governors of Colorado and Washington
requested regulatory guidance from the federal government. In August 2013, Deputy
Attorney General James Cole issued a memorandum (referred to as the Cole Memo)
announcing that the U.S. Justice Department would not move to block the implementation of
recreational cannabis laws. The Cole Memo specified the Justice Department’s CSA
enforcement priorities, which are concerned with issues such as the sale to minors and the
movement of cannabis across states, but left the states on their own to achieve adequate
enforcement of these priorities. Essentially, the Cole Memo communicated that states that
wished to regulate and tax cannabis, for either medical or adult recreational use, would be
left alone providing the states demonstrate competency in achieving enforcement of the
federal priorities (Kamin 2013; 2014).
While the Cole Memo left the enforcement of cannabis possession and distribution
laws to the states, cannabis and cannabis-related businesses (located in states where
cannabis is legal) still face high levels of uncertainty for business transactions that require
some oversight from the federal government. Perhaps the most important and most well-
documented of these issues is the ability to establish banking relationships (Blake and
7
Finlaw 2014; Hill 2015). Although banks can choose either a state or federal charter, all
banks are subject to a number of federal regulations, which may make banks hesitant to
engage in relationships with cannabis businesses. While lack of access to banking services
creates a number of problems for the businesses themselves, state governments were
primarily concerned with unbanked businesses’ ability to underreport taxable income. With
this in mind, the legalizing states’ governors sought banking-specific guidance from federal
banking regulators. In February 2014, The U.S. Treasury Department answered this request
by issuing guidelines on how banks could provide services to cannabis-related businesses
while fulfilling their obligations under the Bank Secrecy Act. The Treasury Department
guidance, however, is vague with respect to cost of compliance and threat of prosecution
over time as national political attitudes change, so banks may remain reluctant to provide
services to cannabis-related businesses.
Despite the reluctance created by regulatory uncertainty, we posit two mechanisms
by which banks may accept deposits from cannabis-related businesses. First, banks may
believe that the federal government will keep its promise not to prosecute banks that
conduct their due diligence as specified in previously mentioned federal guidance. If banks
have compliance measures in place or believe the cost of developing new measures does not
exceed the benefit of servicing these businesses, banks may accept deposits from cannabis-
related businesses. Second, bankers and potential depositors may operate discretely with
customers not specifically identifying themselves as cannabis businesses and banks
simultaneously opt to avoid investigating further.
8
8
See Tighe (2015) for additional discussion.
8
These mechanisms may lead to banks accepting deposits from cannabis-related
businesses, but their actual acceptance of these deposits is an empirical question. The
objective of this study is to address this empirical question by examining whether local bank
deposits increased after cannabis legalization and whether banks subsequently used these
funds to make loans. We expect lending decisions to differ from deposit taking decisions
because one potential consequence of accepting cannabis-related deposits is the forfeiture
of these funds in a criminal investigation of the depositor. In the face of significant
uncertainty concerning the permissibility of risky deposits, risk averse banks may seek to
enhance their liquidity position by reducing loan-to-deposit ratios where cannabis-related
businesses are most prevalent (i.e., in legalizing states).
3 Empirical Methodology and Data
3.1 Empirical Design
We examine the effect of cannabis legalization on banking activity by applying a D-in-
D methodology. We estimate the following model across a sample of bank-quarter
observations before and after cannabis legalization across both legalizing and non-legalizing
states:
(1)
where i indexes the bank, j indexes the state, t indexes the calendar quarter, and D indicates
a variable and/or parameter that is specific to (1). Deposits represents the natural log of
9
total deposits. Our primary explanatory variable of interest, PostLegal, is a dummy variable
equal to one for banks located in legalizing states after the state’s legalization effective date,
and zero otherwise. The legalization effective dates for each of our four treatment states are
as follows: Alaska - February 2015, Colorado - January 2014, Oregon - October 2015, and
Washington - July 2014.
9
The bank effects () control for fixed differences between banks
and the states in which they are located, while the quarter fixed effects () control for
economy-wide changes. Therefore, is the D-in-D estimator and our primary coefficient of
interest.
Our methodology is similar to that of Bertrand and Mullainathan (1999, 2003), as our
settings share the commonality of a nationwide study with staggered passage of state-level
laws. In their case, states passed antitakeover laws at different times across their sample
period, with some states never passing antitakeover laws. Our setting is similar as states
legalized recreational cannabis at different times during our sample period, and other states
did not legalize at all. As articulated by Bertrand and Mullainathan (2003), a significant
benefit of this methodology, is that our control group is not restricted to bank-quarters of
banks located in states that never pass a law, but also includes bank-quarters of banks
located in legalizing states for which the law was not yet effective.
As discussed, our D-in-D approach coupled with bank-level fixed effects should
control for fixed differences between banks located in legalizing and non-legalizing states
9
In our analyses, we exclude banks in Maine and the District of Columbia (D.C.). While recreational cannabis
possession and use was legalized in Maine effective November 2016, the state had yet to issue any licenses to
legally produce or sell the product as of December 2017 due to political and regulatory delays. Similarly, D.C.
has legalized the use and possession of cannabis; however, D.C. has not legalized, nor issued licenses for,
recreational retail sales. We also exclude any banks based in states that had voted or otherwise passed
legislation legalizing recreational cannabis, but legalization was not yet effective during our sample period.
These states include California, Massachusetts, and Nevada.
10
across our sample period. The possibility exists, however, that changes in banking activity
manifest differently between banks in legalizing states and non-legalizing states before and
after legalizing for reasons other than cannabis legalization. While this risk is mitigated
partially by our relatively short sample period (2011-2016) and the inclusion of quarter
fixed effects, we also estimate a specification of the model that includes time-varying bank
and state characteristics. In estimating deposit levels (Deposits), the vector in (1)
includes several control variables following Demirguc-Kunt and Huizinga (2004);
specifically, the vector includes bank-level controls for bank risk (Equity), profitability (ROA),
liquidity (Liquid), and banks’ product mixes and quality of service (Overhead). We also
include the state’s gross domestic product (GDP) per capita (GDP/Cap) as a macroeconomic
control.
10
A similar D-in-D approach is applied to estimate our other outcome measure of
banking activity, the level of bank loans (Loans). Although there is some overlap in the
specification of the Deposits and Loans models, they are not identical. Our specification in
(1) is modified to estimate loans as follows:
(2)
where L indicates a variable and/or parameter that is specific to (2) and the other subscripts
retain their previous definitions. Loans represents the natural log of total loans. The
specification in (2) includes the outcome variable (Deposits) from (1) as an explanatory
10
See Table 1 for details regarding variable measurement.
11
variable because deposits have been shown to be positively associated with lending
(Ivashina and Scharfstein 2010; Beatty and Liao 2011). Our interest now turns to both
(the overall effect of PostLegal on loans) and (the interacted or cannabis legalization effect
on the relationship between deposits and lending).
Additionally, the vector in (2) includes several bank-level controls identified by
prior literature (Beatty and Liao 2011 and Kim and Sohn 2017.) These controls are the level
of tier 1 capital (Tier1), banks’ liquidity ratio (Liquid), overhead ratio (Overhead), total assets
(Assets), non-deposit liabilities (MFund), unused commitments (Unused), profitability (ROA),
total non-performing loans (), and the ratio of the bank’s loan loss reserve to total loans
(LLR). We also control for state-level macroeconomic characteristics as in the deposits
model, so unsurprisingly there is some overlap between the vector in (1) and the
vector in (2).
[Insert Table 1]
3.2 Data and Summary Statistics
In examining the impact of recreational cannabis legalization on deposit and lending
activity of the banks in legalizing states, we estimate our models of deposits and loans using
a sample of U.S. banks with available financial data from FDIC-provided Call Reports between
2011 and 2016. Our sample is composed of 150,566 bank-quarter observations from 6,932
unique banks located in 46 different states.
11
11
As discussed in footnote 8, we exclude California, Maine, Massachusetts, and Nevada banks from our sample,
as these states had enacted legislation related to recreational cannabis during our sample period, but the
legislation had not yet become effective or resulted in the issuance of any licenses for cannabis businesses.
12
Table 2 presents summary statistics for our full sample, as well as separately for
banks located in states that legalized cannabis at some point during our sample period and
banks located in states that never legalized during our sample period. For the full sample,
the mean level of deposits is about $154 million and mean loan level is about $107 million.
Additionally, the banks in our overall sample have mean total assets of about $181 million,
return on assets of 0.6 percent, and Tier 1 capital ratio of 10.6 percent.
12
Compared to banks
in non-legalizing states, banks in legalizing states report higher deposits ($202 million vs.
$160 million) and higher levels of loans ($143 million vs. $107 million ). Additionally, among
other differences, banks in legalizing states appear to be larger, use less equity financing,
carry lower liquidity, and incur greater overhead costs.
[Insert Table 2]
4 Results
4.1 Deposits
Table 3 presents the results from the estimation of our deposits model. Our variable
of interest is PostLegal, an indicator variable equal to one for each bank-quarter observation
after the state’s legalization effective date, and zero otherwise. PostLegal is also set equal to
12
In the absence of a particularly meaningful maximum size threshold, we do not impose a maximum size cut-
off for inclusion in our sample. The largest bank-quarter observation in our sample reports $139 billion in
unlogged assets, however, of the 6,932 unique banks in our sample, only 47 banks ever report assets greater
than $10 billion. To the extent that the largest banks are more likely to engage in banking operations beyond
their home state, we expect the inclusion of these banks to bias against finding that state-level regulation
significantly affects the banks in our sample. Nevertheless, our results are not sensitive to the exclusion of
banks reporting assets greater than arbitrary thresholds such as $2, $10, and $25 billion.
13
zero for all bank-quarters of banks located in non-legalizing states and for bank-quarters
associated with banks in states that have not yet legalized. The reported standard errors are
corrected for heteroskedasticity, clustering at the state level, and account for serial
correlation as suggested by Bertrand, Duflo, and Mullainathan (2004).
[Insert Table 3]
Column A reports the results of our base D-in-D model, including only our primary
variable of interest, PostLegal, and the bank and quarter fixed effects. Additonal controls are
added in Columns B-D; Column B adds state-level GDP per capita (GDP/Cap), Column C
includes all controls except for GDP/Cap, and Column D presents the full model (1) with all
control variables. The coefficient on PostLegal is positive and statistically significant in all
four specifications, suggesting that deposits are higher for banks in states after recreational
cannabis legaliziation. Interpreting the coefficients for PostLegal suggests that recreational
cannabis legalization resulted in increased deposits ranging from 3.14 percent in our full
model (Column D) to 4.33 percent in our base model (Column A). The coefficients of our
control variables are also consistent with prior literature (e.g. Demurgic-Kunt and Huizinga
2004). In particular, equity, liquidity, and overhead are negatively associated with deposits,
while profitability is positively associated.
4.2 Loans
Table 4 presents the results of our estimated loans model (2). Column A reports the
base model, including only PostLegal, Deposits, PostLegal*Deposits, and the bank and quarter
14
fixed effects. Similar to our presentation in Table 3, Columns B-D contain estimates with
additional control variables; Column B adds state-level GDP per capita (GDP/Cap), Column C
includes all controls except GDP/Cap, and Column D provides our full model (2) results with
all control variables. The coefficient for PostLegal is positive and statisticaly significant in all
four specifications, indicating that bank lending increased after recreatiaonal cannaibis
legalization. Interpreting the coefficients for PostLegal suggests that recreational cannabis
legalization resulted in increased loans ranging from 8.62 percent in our full model (Column
D) to 6.54 percent in our base model (Column A).
[Insert Table 4]
While the overall effect of recreational cannabis legalization on lending is
noteworthy, we are partucularly interested in whether the increased levels of post-
legalization deposits, documented in Section 4.1 above, resulted in increased lending. We
capture this effect using the interaction term PostLegal*Deposits. Unsurprisingly, the
coefficient on Deposits is positive and statistically significant, capturing the rate at which
deposit dollars affect loans. However, the coefficient on PostLegal*Deposits is negative and
statistically significant in all four specificatons, suggesting that the rate at which deposits
translate into new loans is lower in legalizing states after legalization.
Using the specification in Column D, our estimates indicate that a 10 percent increase
in deposits results in a 3.93 percent (0.3928 x 10) increase in loans for banks in non-
legalizing states. In legalizing states after legalization, the same 10 percent increase in
deposits results in a 3.87 percent ((0.3928 - .0060) x 10) increase in loans. While this
15
differene is statistically signifianct, its economic impact is inconsequential. Additionally,
despite the slightly smaller increase in loans in response to deposits, the PostLegal
coefficient is large and consistent with increased banking activity in legalizing states. For
example, this translates to only about $993 per bank-quarter and an
economic wide effect of $1.84 million for the 1,979 PostLegal bank-quarter observations.
5 Conclusion
We examine the effect of state-level recreational cannabis legalization on deposits
and loans of banks operating in legalizing states. Using a D-in-D design we compare banks
in legalizing states to control banks from non-legalizing states and find that banks in
legalizing states exhibit higher deposits after legalization in comparison to banks in non-
legalizing states. The positive effect of recreational cannabis legalization on bank deposits
is both statistically and economically significant. Moreover, the estimated effect is robust to
changes in model specification. We also find that banks have higher loan levels after their
states legalize. Lending is less sensitive to deposits in legalizing states, but the size of this
effect is small. Although we are unable to specifically identify whether the higher level of
deposits is due to the acceptance of deposits from cannabis-related activities and whether
banks knowingly accept these types of deposits, our findings suggest that the risk from
regulatory uncertainty did not decrease banks’ willingness to accept deposits or make loans.
In fact, legalization resulted in a significant increase in banking activity with no substantial
dampening in deposit to loan conversions.
Altogether our results indicate that deposits and loans increased for banks after
recreational cannabis legalization. This may suggest that banks were either unconcerned
16
about the potential risk associated with accepting cannabis related deposits or optimistic
about the chances that regulations will adapt to the needs of legalizing states. As the Senate
prepares to vote on the SAFE Banking Act, many involved in cannabis-related businesses
think that the act is trying to fix a problem that does not actually exist (Beuerlein, 2020). The
American Bar Association stated that once one or two banks or credit unions begin working
with marijuana businesses within the confines of the FinCEN guidance, other banks and
credit unions begin to follow suit (Bricken, 2017). Despite experiencing a slight dip in 2020,
banking services to cannabis-related businesses have been steadily increasing since 2014,
according to the 2021 first-quarter FinCEN report.
13
Although many have speculated about
the increased legal risks to banks, there is a lack of evidence for instances where banks are
criminally prosecuted or lose their federally insured status. If these negative reprocussions
rarely happen, it makes sense that banks would not respond to the legislative uncertainty.
As more state regulators issue statements in support of banks and credit unions serving the
cannabis industry, the financial institutions can become more optimistic about the chances
that regulations will adapt in their favor with time (Levinson, 2020).
Our study contributes to the literature by examining how organizations, in this case
banks, make decisions under regulatory uncertainty. Overall, our findings suggest that banks
make calculated decisions when addressing regulatory uncertainty around state cannabis
legalization. However, it should be noted that our sample covers a period where bank
managers could have been reasonably optimistic about their future ability to engage with
caanabis related businesses. As discussed earlier, the Obama administration stated in the
13
https://www.fincen.gov/sites/default/files/shared/508_299423_MJ%20Banking%20Update%201st%20QT
R%20FY2021_Public_Final.pdf
17
Cole Memorandum that they would allow states to pursue recreational legalization as long
as they complied with the enforcemet priorities of U.S. Justice Department. Future research
may chose to examine banks’ decisions around cannabis legalization after the change in
presidential administrations.
18
REFERENCES
Beatty, A. and Liao, S., 2011. Do delays in expected loss recognition affect banks' willingness
to lend? Journal of Accounting and Economics, 52(1), 1-20.
Beck, T., Demirgüç-Kunt, A. and Maksimovic, V., 2004. Bank competition and access to
finance: International evidence. Journal of Money, Credit and Banking, 627-648.
Bertrand, M. and Mullainathan, S., 1999. Is there Discretion in Wage Setting? A Test Using
Takeover Legislation. The Rand Journal of Economics, 535-554.
Bertrand, M. and Mullainathan, S., 2003. Enjoying the Quiet Life? Corporate Governance and
Managerial Preferences. Journal of Political Economy, 111(5), 1043-1075.
Bertrand, M., Duflo, E. and Mullainathan, S., 2004. How much should we trust difference-in-
differences estimates? The Quarterly Journal of Economics, 119(1), 249-275.
Beuerlein, T., 2020, Jan 7. Dispelling Banking Myths In The Cannabis Industry. Forbes.
https://www.forbes.com/sites/forbesbusinessdevelopmentcouncil/2020/01/07/di
spelling-banking-myths-in-the-cannabis-industry/?sh=44cddf3c25e2
Blake, D. and Finlaw, J., 2014. Marijuana legalization in Colorado: Learned lessons. Harv. L.
& Pol'y Rev., 8, 359.
Bricken, H., 2017, August 15. Navigating the Hazy Status of Marijuana Banking. American
Bar Association. https://www.americanbar.org/groups/business_law/publications/
blt/2017/08/03_bricken
Chemerinsky, E., Forman, J., Hopper, A. and Kamin, S., 2015. Cooperative federalism and
marijuana regulation. UCLA L. Rev., 62 (74), 50-122.
Demirgüç-Kunt, A. and Huizinga, H., 2004. Market discipline and deposit insurance. Journal
of Monetary Economics, 51(2), 375-399.
Dupas, P., Karlan, D., Robinson, J. and Ubfal, D., 2016. Banking the Unbanked? Evidence from
three countries (No. w22463). National Bureau of Economic Research.
Felix, A. and Chapman, S., 2018. The Economic Effects of the Marijuana Industry in Colorado.
Rocky Mountain Economist, Federal Reserve Bank of Kansas City, April 16, 2018.
Guiso, L., Sapienza, P. and Zingales, L., 2004. Does local financial development matter? The
Quarterly Journal of Economics, 119(3), 929-969.
Gulen, H. and Ion, M., 2016. Policy Uncertainty and Corporate Investment. The Review of
Financial Studies, 29 (3): 523-564.
19
Hill, J.A., 2015. Banks, marijuana, and federalism. Case W. Res. L. Rev., 65, 597-647.
Ivashina, V. and Scharfstein, D., 2010. Bank lending during the financial crisis of
2008. Journal of Financial Economics, 97(3), 319-338.
Kamin, S., 2013. The limits of marijuana legalization in the states. Iowa L. Rev. Bull., 99, 39.
Kamin, S., 2014. Cooperative Federalism and State Marijuana Regulation. U. Colo. L. Rev., 85,
1105-1122.
Kim, D. and Sohn, W., 2017. The effect of bank capital on lending: Does liquidity
matter? Journal of Banking & Finance, 77, 95-107.
Kingsley, A.F., Vanden Bergh, R.G. and Bonardi, J.P., 2012. Political markets and regulatory
uncertainty: Insights and implications for integrated strategy. Academy of
Management Perspectives, 26(3), 52-67.
Kovaleski, S., 2014. U.S. Issues Marijuana Guidelines for Banks. The New York Times
February 14, 2014: A.10.
Levinson, B., 2020, Dec. 1. Now is the Time to Change the Downward FinCEN Cannabis
Banking Trend. JD Supra. https://www.jdsupra.com/legalnews/now-is-the-time-to-
change-the-downward-39274/
Lopez, J.M.R., Sakhel, A. and Busch, T., 2017. Corporate investments and environmental
regulation: The role of regulatory uncertainty, regulation-induced uncertainty, and
investment history. European Management Journal, 35(1), 91-101.
Romi, A.M., Carrasco, H., Camors, C.A., Masselli, J.J., 2019. From the black-market to the gray-
market: Accounting’s role in the budding cannabis industry. Working Paper: Texas
Tech University.
Savage, C. and Healy, J., 2018. Trump Administration Takes Step That Could Threaten
Marijuana Legalization Movement. The New York Times January 4, 2018: A.1.
Tighe, P.A., 2015. Underbanked: Cooperative banking as a potential solution to the
marijuana-banking problem. Mich. L. Rev., 114, 803-832.
Wilber, D.Q., Frosch, D., 2018. U.S. News: Sessions ends Obama easing of marijuana
prosecutions. The Wall Street Journal January 5, 2018: A.3.
20
Table 1
Variable Definitions
Variable
Definition
Deposits
The natural log of total deposits.
Loans
The natural log of total loans.
PostLegal
An indicator variable equal to one for banks in legalizing states and
quarters beginning after the state’s legalization effective date, and zero
otherwise. The legalization effective dates of our four treatment states
are as follows: Alaska – February 2015; Colorado – January 2014; Oregon
– October 2015; Washington – July 2014.
Equity
Total book equity divided by total assets.
Liquid
The ratio of liquid assets to total assets.
Overhead
Total non-interest expenses divided by total assets.
Assets
The natural log of total assets.
NPL
The total loans reported as past due or nonaccrual.
Tier1
The ratio of tier 1 risk-weighted capital to total assets.
MFund
The ratio of non-deposit liabilities to total assets.
Unused
Total amount of unused commitments to total assets.
ROA
Income before taxes divided by total assets.
LLR
The total loan loss reserve divided by total loans.
GDP/Cap
The state’s quarterly real GDP divided by the state’s population.
21
TABLE 2
Summary Statistics
All observations
Legalizing
States
Non-legalizing
States
Mean
Std
Mean
Std
Mean
Std
11.947
1.194
12.214
1.278
11.939***
1.190
11.589
1.280
11.868
1.320
11.580***
1.278
0.112
0.041
0.110
0.030
0.112***
0.042
0.343
0.164
0.324
0.154
0.343***
0.164
0.019
0.011
0.022
0.013
0.019***
0.011
12.109
1.178
12.355
1.258
12.101***
1.175
0.008
0.013
0.012
0.017
0.008***
0.013
0.106
0.032
0.104
0.027
0.106***
0.032
0.045
0.050
0.036
0.044
0.045***
0.050
0.018
0.021
0.020
0.018
0.018***
0.021
0.006
0.006
0.005
0.007
0.006***
0.006
0.016
0.008
0.020
0.010
0.016***
0.008
0.047
0.007
0.052
0.003
0.047***
0.007
N
150,566
4,735
145,831
Notes: Table 2 reports summary statistics across all bank-quarter observations in our sample. In
addition, bank-quarter observations are split between banks located in states that legalized at some
point during our sample period and those located in states that never legalized during our sample
period. *, **, and *** denote statistical significance at 10%, 5%, and 1%, respectively, from two-tailed t-
tests of differences in means between the legalizing and non-legalizing subsamples.
22
TABLE 3
The Effect of Cannabis Legalization on Bank Deposits
Dep Var:
Column A
Column B
Column C
Column D
0.0433***
(0.0044)
0.0377***
(0.0044)
0.0355***
(0.0041)
0.0314***
(0.0041)
-3.4308***
(0.0261)
-3.4216***
(0.0261)
1.0539***
(0.0968)
1.0948***
(0.0967)
-0.3304***
(0.0071)
-0.3267***
(0.0071)
-3.3780***
(0.0928)
-3.3302***
(0.0928)
8.0499***
(0.4074)
5.8358***
(0.3809)
Quarter FEs
Yes
Yes
Yes
Yes
Bank FEs
Yes
Yes
Yes
Yes
N
150,566
150,566
150,566
150,566
0.229
0.231
0.329
0.330
Notes: Robust standard errors corrected for clustering at the bank level are reported in parentheses; *, **, and *** denote statistical
significance at 10%, 5%, and 1%, respectively. See Table 1 for comprehensive variable definitions.
23
TABLE 4
The Effect of Cannabis Legalization on Bank Loans
Dep Var:
Column A
Column B
Column C
Column D
0.0654**
(0.0305)
0.0581***
(0.0304)
0.0889***
(0.0198)
0.0862***
(0.0198)
0.9030***
(0.0020)
0.9010***
(0.0020)
0.3923***
(0.0029)
0.3928***
(0.0029)
-0.0047*
(0.0025)
-0.0045*
(0.0025)
-0.0061***
(0.0016)
-0.0060***
(0.0016)
0.4689***
(0.0192)
0.4689***
(0.0192)
-1.4738***
(0.0038)
-1.4724***
(0.0038)
0.5582***
(0.0030)
0.5567***
(0.0030)
0.2089***
(0.0091)
0.2087***
(0.0091)
0.2409***
(0.0346)
0.2450***
(0.0346)
-0.3551***
(0.0503)
-0.3468***
(0.0503)
-0.2992***
(0.0245)
-0.3029***
(0.0245)
-0.6290***
(0.0245)
-0.6413***
(0.0513)
5.7925***
(0.3040)
2.1991***
(0.1980)
Quarter FEs
Yes
Yes
Yes
Yes
Bank FEs
Yes
Yes
Yes
Yes
N
150,566
150,566
150,566
150,566
0.701
0.702
0.874
0.874
Notes: Robust standard errors corrected for clustering at the bank level are reported in parentheses; *, **, and *** denote statistical
significance at 10%, 5%, and 1%, respectively. See Table 1 for comprehensive variable definitions.
