The Impact of Cryptocurrency Exposure on Corporate Tax Avoidance Among US Listed Companies

Abstract

This study examined the association between corporate cryptocurrency activities and tax avoidance outcomes, utilizing data from US public firms covering the period from 2015 to 2023. Financial data were sourced from Compustat, while details regarding cryptocurrency activities were manually extracted from 10-K and 10-Q filings. Our analysis employed a fixed-effects regression model to examine the impact of these activities on cash effective tax rates (ETR). The findings indicate that firms engaged in cryptocurrency activities tend to have a lower ETR compared with those without such involvement. Notably, this effect was predominantly observed in companies directly engaged in cryptocurrency activities, such as accepting cryptocurrency as a payment method or actively trading cryptocurrency on an exchange platform. In contrast, firms involved in crypto mining or initial coin offerings did not exhibit a similar association. Our findings offer significant regulatory insights for governance bodies concerned with the implications of corporate cryptocurrency activities on tax strategies.

1. Introduction

The cryptocurrency market has experienced significant growth over the past decade, with a total market capitalization reaching approximately USD 2.17 trillion as of 7 October 2024 (CoinMarketCap 2024). As digital assets such as Bitcoin and Ether continue to gain popularity, an increasing number of companies are entering the crypto asset space, either directly or indirectly. The relationship between involvement in cryptocurrency and corporate tax avoidance has emerged as a growing area of interest globally (Solodan 2019; Cong et al. 2023; Yussof and Al-Harthy 2018).

This study explored how corporate cryptocurrency activities may influence tax avoidance strategies. Specifically, it investigated whether and in what ways engagement in cryptocurrencies affects a company’s tax avoidance outcomes. The study first examined whether cryptocurrency holdings correlate with general tax avoidance, measured by the cash effective tax rate. It then delved into the types of cryptocurrency exposure that might have the most significant influence, testing whether activities such as mining, initial coin offerings (ICOs), active trading, or accepting digital currencies as payment can impact a company’s tax liabilities.

Cryptocurrency, a form of digital or virtual asset protected by cryptography, operates on decentralized networks, typically utilizing blockchain technology (Nakamoto 2008).1 Defining and classifying crypto assets is a complex task, particularly for regulatory agencies such as the IRS and SEC (Bakradze 2021). For example, the IRS broadly defines cryptocurrency as virtual currency that utilizes a distributed ledger, such as blockchain. However, it has faced criticism for its lack of clear guidance on the taxation of virtual currencies (Farrington 2021; McDonald 2021). This criticism arises not only from definitional ambiguities but also from challenges related to measurement, valuation, and classification, given the intricate nature of these digital assets. The White House has adopted the term “digital assets” (White House 2022), while the Financial Accounting Standards Board refers to them as “crypto assets” in their authoritative standards update, ASU 2023-08 (FASB 2023a).2

In this study, we did not delve deeply into the specifics of blockchain technology, as explored by authors such as Miglo (2021), Zheng et al. (2023), and Kapengut and Mizrach (2023). Nor did we address the various industry applications of this technology, such as in supply chain management (Uddin et al. 2023), environmental sustainability (Sapra et al. 2023), energy (Khezami et al. 2022), or healthcare (Wenhua et al. 2023). Existing studies have contributed to developing a more representative definition and classification of digital assets. For instance, Watters (2023) stated that cryptocurrencies are a subset of decentralized digital assets (DDAs) encompassing a wide variety of formats. DDAs can be classified into three broad categories: currencies, commodities, and securities. However, classifying certain crypto assets into these categories can be challenging, leading to ongoing litigation regarding the distinction between their currency functionality and security components.3 This challenge also applies to various tokens, including utility tokens, security tokens, governance tokens, and Decentralized Finance (DeFi) tokens, all of which are experiencing increasing regulatory scrutiny.

This innovative financial tool has streamlined business transactions by eliminating the need for intermediaries, such as traditional financial institutions, thereby reducing transaction costs and enhancing overall efficiency. The rapid advancement of cryptocurrencies has brought significant transformations across various sectors of the global economy (Ilham et al. 2019), particularly in the financial and corporate landscapes, reshaping how businesses operate and manage their resources.4

There is a growing body of literature focused on the taxation of cryptocurrency, highlighting its unique characteristics and the varying tax treatments across different jurisdictions. For instance, Bakradze (2021) examined how the United States addresses cryptocurrencies, discussing the current challenges and criticisms of the IRS’s tax policies. Solodan (2019) provided an overview of cryptocurrency taxation in several European countries, summarizing the key legislative features. Huang et al. (2023) explored the implications of viewing cryptocurrency as property from both accounting and taxation perspectives. Additionally, Liotta (2019) highlighted the various challenges cryptocurrencies pose for tax regulation. A key takeaway from this body of research is the need for a taxation regime specifically tailored for cryptocurrencies.

The second line of literature on cryptocurrency and taxation discusses its potential as a tool for tax avoidance, primarily from a theoretical perspective. Marian (2013) highlighted that the unique characteristics of crypto assets—specifically their anonymity and independence from financial intermediaries—allow them to effectively replace traditional tax evasion strategies, such as creating foreign bank accounts in tax havens. Sanchez (2017) argued that cryptocurrencies are increasingly used as tools for money laundering and are viewed as new tax havens due to their decentralized and anonymous nature, thereby calling for stronger regulatory frameworks to combat these illegal activities. Atiles (2022) explicitly investigated how Puerto Rico’s unique tax incentives align with the underlying blockchain technology of cryptocurrencies, positioning the region as a potential tax haven. Nguyen (2022) focused on the specific tax framework for non-fungible tokens (NFTs), explaining how this type of crypto asset can be particularly useful for tax evaders. Cernușca et al. (2020) examined various conceptual and practical issues related to taxing earnings from crypto asset exchanges and transfers.

More recently, studies have explored how cryptocurrency can serve as a tool for tax avoidance using real-life cases and factual evidence. According to Meling et al. (2024), many investors in Norway are not in compliance with crypto tax regulations. This issue persists even among those required to disclose their trading transaction data from exchanges to the tax authorities. A common feature among the existing studies is their examination of the relationship between cryptocurrency and tax avoidance from a theoretical perspective (Bakradze 2021; Solodan 2019; Liotta 2019). While some studies, such as those by Cernușca et al. (2020), have focused on practical applications, their findings are primarily based on descriptive statistics or survey data. To our knowledge, there is no existing research that directly examines the association between cryptocurrency and tax using archival empirical data. Our study aimed to address this gap in the literature by providing empirical evidence to support the need for more detailed regulation of cryptocurrency taxation in the United States.

Despite the growing involvement of companies in cryptocurrency-related activities—such as mining, trading, and accepting cryptocurrencies as payment—the impact of these activities on operational performance and financial outcomes remains underexplored. While these activities provide opportunities for growth and innovation, they also introduce unique challenges, particularly in financial reporting and taxation. The cryptocurrency space currently lacks a standardized accounting and tax framework, which leaves room for potential manipulation of transactions. This absence of clear, authoritative guidelines complicates the consistent measurement and evaluation of performance metrics across different digital assets and activities.

The Internal Revenue Service (IRS) treats cryptocurrencies as property rather than currency for tax purposes. This classification means that transactions involving the sale or exchange of cryptocurrency are subject to capital gains tax rules. When a taxpayer sells or exchanges cryptocurrency, they must report any resulting gains or losses. These are calculated on the basis of the difference between the cryptocurrency’s fair market value at the time of the transaction and its adjusted basis, which includes the original purchase price and any associated costs. Short-term capital gains from assets held for one year or less are taxed at the taxpayer’s ordinary income tax rate. Long-term capital gains from assets held for more than one year benefit from lower tax rates. Additionally, cryptocurrency received as payment for goods or services is considered to be ordinary income, and its fair market value at the time of receipt must be reported.

However, tax regulations for crypto assets are still in the early stages, with significant efforts and detailed regulations yet to be established (Anderson et al. 2024; Schwanke 2017). For instance, as of 2024, the tax return form includes only a basic digital asset question: “At any time during 2023, did you: (a) receive (as a reward, award, or payment for property or services); or (b) sell, exchange, or otherwise dispose of a digital asset (or a financial interest in a digital asset)?” This simple yes or no question falls short of adequately addressing the complexities of crypto-related transactions. We are aligned with the market’s need for comprehensive accounting and tax regulations for cryptocurrencies and hope that our study will contribute meaningfully to this evolving field.

A key question in the context of involvement in cryptocurrency is how it impacts the actual cash taxes paid to authorities. Our study is the first to examine whether and how cryptocurrency exposure influences a fundamental measure of corporate tax outcomes: the cash effective tax rate (ETR). The ETR, widely regarded as the most representative measure of the average corporate tax rate, reflects the cumulative effect of various factors, including the nature of a company’s operations, financial performance, location, regulatory environment, and tax planning strategies. The integration of cryptocurrency into business operations introduces a new dimension to this equation, raising questions about how these digital assets and associated activities might affect corporate tax liabilities.

One significant reason for the lack of research on this topic at the firm level is the issue of data availability. Currently, there are no publicly accessible data on cryptocurrency holdings at the firm level, as there are no established accounting standards for cryptocurrency. Despite the ongoing requests from market participants for authoritative accounting standards, it was not until December 2023 that the FASB released its first accounting standard on crypto assets (ASU 2023-08), which will take effect for fiscal years beginning after 15 December 2024. Consequently, firms are not required to disclose their cryptocurrency exposure until the end of 2024.

In this study, we investigated the relationship between a firm’s cryptocurrency exposure and tax avoidance outcomes. To navigate the data challenge mentioned earlier, we creatively compiled annual and quarterly reports of US public firms that included keywords such as “Bitcoin” and “blockchain” from the SEC website. We manually reviewed these reports to verify the companies’ involvement in cryptocurrency, resulting in a collection of 1300 firm-year observations. However, due to the absence of a linking table between the SEC Edgar database and the WRDS Compustat database—which contains the fundamental characteristics of US firms—we lost approximately one-third of our observations. Additionally, due to missing dependent, independent, and control variables, we ultimately retained only 113 firm-year observations for our analysis.

Using hand-collected data on cryptocurrency holdings from 2013 to 2023, we present preliminary evidence of this relationship. Our findings suggest that firms with cryptocurrency exposure tend to have lower cash effective tax rates, not only in the current year (current-year ETR) but also over a three-year horizon (three-year ETR). In further analysis, our subsample examination, based on specific types of involvement in cryptocurrency, revealed that the prior association is primarily driven by firms that can actively and frequently exchange cryptocurrency, such as those accepting cryptocurrency as a payment method or actively trading it on exchange platforms. In contrast, firms involved in cryptocurrency mining or ICOs do not show a similarly significant association. This suggests that firms may reduce their tax burden through cryptocurrency exposure only if they actively trade these assets. For instance, firms that mine popular cryptocurrencies, such as Bitcoin, and hold their Bitcoin for the long term do not exhibit the same tax benefits as those that engage in more active trading.

By analyzing these dynamics, this research aimed to provide valuable insights into the broader implications of cryptocurrency adoption in the corporate world, particularly in relation to tax policy. Our findings provide direct evidence that, for public firms in the US, cryptocurrency exposure is significantly associated with a lower effective tax rate, not only over a one-year period but also across a three-year window. We further document that this association is primarily driven by firms actively trading cryptocurrency on exchanges. In contrast, we did not observe a similar association for firms engaged in passive cryptocurrency exposure, such as crypto mining or initial coin offerings. The findings of this study will be of significant interest to corporate executives and tax professionals navigating the increasingly complex landscape of digital assets. Our study also carries significant policy implications for various regulatory agencies, including Congress, the IRS, the SEC, and accounting-related regulatory organizations, aligning with existing literature that calls for more comprehensive and detailed tax regulations for crypto assets. One limitation of our research is the relatively small sample size, with 216 firm-year observations in our propensity score-matched sample. However, this limitation is common across concurrent studies in this area. For example, Luo and Yu (2024) examined only 40 global firms, while Anderson et al. (2024) analyzed 758 firm-quarter observations, representing 128 unique firms.

The remainder of the article is organized as follows. Section 2 reviews the relevant prior literature and proposes our testable hypotheses. Section 3 describes the research design and defines the variables. Section 4 presents the sample selection, summary statistics, and empirical results. Finally, Section 5 concludes the article.

2. Literature Review and Hypothesis Development

The regulatory efforts surrounding cryptocurrencies and digital assets are still in their early stages, with significant gaps between regulation and practice (Anderson et al. 2024). Despite calls for clear accounting standards (Bloomberg 2021), the FASB did not issue formal guidance for crypto assets until December 2023 (Deloitte 2018; Ernst & Young 2018). Until then, practitioners relied on non-binding advice, such as the American Institute of Certified Public Accountants (AICPA) guidelines, which recommended treating crypto assets as long-term intangible assets subject to impairment losses (AICPA 2019). However, since these guidelines were not part of the official FASB codification, they were merely suggestions rather than requirements.

In December 2023, recognizing the unique nature of crypto assets, the FASB issued Accounting Standards Update (ASU) 2023-08, introducing the fair value model for crypto assets (FASB 2023a, 2023b). This shift was driven by extensive feedback from a wide range of stakeholders, including investors, practitioners, and regulators. However, this change is just the beginning, and much remains uncertain in how these standards will be applied in practice.

Meanwhile, the Securities and Exchange Commission (SEC) and the IRS have been more proactive in their regulation of the crypto market, though challenges remain. The SEC began regulating crypto assets in 2017, starting with its DAO report, which classified certain blockchain tokens as securities subject to federal laws (SEC 2017).5 This report formed the basis of SEC regulations in crypto space. The SEC also moved to regulate ICOs around the same time, charging several issuers for failing to register their offerings. More recently, the SEC has turned its attention to non-fungible tokens (NFTs), suggesting that some might also fall under securities regulations.

In 2023, the SEC significantly ramped up its enforcement actions, issuing 46 cases related to crypto, a 50% increase from the previous year (Cornerstone Research 2023). These actions focused on fraudulent transactions, improper administrative practices, and market manipulation. The approval of the first spot Bitcoin exchange-traded funds (ETFs) in January 2024 marked another significant regulatory step, allowing investors to trade crypto assets through regulated funds.

The IRS began regulating cryptocurrencies in 2014, treating them as property for tax purposes (IRS 2014a, 2014b). This means that gains from selling crypto are subject to capital gains tax, and companies accepting crypto as payment must recognize it as income at fair market value on the date of receipt (Ankier 2019). The IRS has also stated that income from crypto mining is taxable. However, the guidance remains broad, giving firms considerable flexibility in reporting and managing income from crypto activities.

Overall, while regulatory bodies such as the FASB, SEC, and IRS are making efforts to regulate the rapidly growing crypto market, the framework is still developing. This gap between regulation and practice creates opportunities for firms to exploit these ambiguities, particularly in areas such as tax avoidance. As the market continues to evolve, further regulatory developments are needed to close these gaps and ensure that crypto activities are adequately monitored and taxed.

Cryptocurrencies, characterized by their borderless nature, decentralization, and relative anonymity, present firms with unique opportunities to exploit tax advantages within the existing legal frameworks. The tax avoidance literature highlights that firms often leverage complex financial instruments to minimize tax liabilities, especially in areas where regulation is still evolving (Desai and Dharmapala 2006; Hanlon and Heitzman 2010). Cryptocurrencies fit this paradigm by offering various tax planning strategies due to their unique features (Nakamoto 2008; Marian 2013). The borderless nature of cryptocurrencies allows firms to conduct transactions across jurisdictions with lower tax rates, effectively reducing their taxable income in higher-tax regions (Marian 2013). Certain jurisdictions may offer preferential tax treatments for digital assets or may not have fully integrated cryptocurrency transactions into their tax codes (Uzougbo et al. 2024). This allows companies to capitalize on favorable tax conditions without the need to repatriate cryptocurrency into higher-tax jurisdictions (Slemrod and Wilson 2009).

The decentralized structure of cryptocurrencies further supports tax planning by allowing firms to hold digital assets offshore, deferring tax payments or benefiting from lower tax rates. The tax avoidance literature extensively discusses the use of offshore financial centers and tax havens to reduce tax burdens (Hines 2010; Gravelle 2015). In the context of cryptocurrencies, firms can establish foreign trusts or shell entities that own crypto assets, thereby shielding these assets from domestic tax authorities. The anonymous nature of many cryptocurrencies complicates enforcement by tax authorities, who struggle to track ownership and attribute income. This aligns with traditional tax avoidance strategies, where firms exploit regulatory opacity to lower their tax burdens (Dharmapala 2008; Zucman 2014).

Valuation challenges add another layer of complexity to tax planning with cryptocurrencies. Tax avoidance research shows that firms often take advantage of ambiguities in asset valuation to manipulate reported income (Scholes et al. 2014). The high volatility of cryptocurrency prices enables firms to time the realization of gains and losses, optimizing their taxable income (Hanlon and Shevlin 2002; Gallemore et al. 2014). Furthermore, the lack of active trading for many cryptocurrencies complicates the determination of a fair market value, allowing firms to choose valuation methods—such as recent trading prices or historical costs—that align with their tax planning objectives (Armstrong et al. 2012). The strategic use of fair value accounting for volatile digital assets thus permits firms to reduce their effective tax rates by managing the timing of recognized income and deductions (Frank et al. 2009).

The combination of these factors suggests that firms engaged in cryptocurrency activities are better positioned to reduce their cash ETR through strategic tax management. By navigating the complexities of cryptocurrency regulation, valuation, and cross-border transactions, these firms can achieve significant tax savings, thereby lowering their overall tax burden. This leads to the following hypothesis.

Hypothesis 1 (H1). 

Firms involved in cryptocurrency activities are associated with a lower cash effective tax rate.

The impact of cryptocurrency activities on a company’s cash ETR can vary, depending on the nature of the activities. As discussed earlier, due to the absence of official guidance from the FASB until December 2023, and the fact that firms only voluntarily disclose limited information in their SEC filings, we had to classify cryptocurrency activities into three broad categories: crypto mining, trading, and accepting cryptocurrency as payment. These categories are more frequently mentioned in the firm’s annual or quarterly reports. To gain a more comprehensive understanding of how these activities influence tax avoidance, we analyzed their effects through empirical regression analysis.

For firms engaged in active cryptocurrency trading, the primary focus often lies in capitalizing on short-term gains from price fluctuations. The high volatility associated with cryptocurrency prices allows firms to strategically time their trades, using realized losses to offset taxable gains and reduce taxable income (Hanlon and Heitzman 2010; Gallemore et al. 2014). This ability to manage the recognition of gains and losses aligns with the literature on earnings management and tax planning, where firms exploit opportunities to optimize their tax positions (Frank et al. 2009). However, regulations concerning capital gains on trading are relatively mature, with IRS guidance dating back to 2014. This regulatory framework may limit the flexibility of firms to reduce their tax obligations through trading.

Accepting cryptocurrency as a payment method brings distinct financial and tax considerations. Firms can time the recognition of crypto assets and choose valuation methods that may lower taxable income, especially given the price volatility of cryptocurrencies. The flexibility in recognizing gains or losses creates opportunities for firms to manage reported earnings, potentially resulting in lower cash ETRs. Additionally, the anonymity of cryptocurrency transactions and the ability to accept payments through foreign subsidiaries may complicate the reporting process, making it easier for firms to defer taxable income or take advantage of preferential tax treatments in certain jurisdictions (Lisowsky 2010; Blouin 2014). However, the IRS has established relatively mature guidance on this topic, with the initial standards dating back to 2014. According to the IRS’s updated announcements, cryptocurrency’s value at the time of the transaction is typically recognized as revenue, similar to cash transactions. This established transparency may limit firms’ ability to engage in aggressive tax avoidance strategies.

Crypto mining involves significant initial investments in specialized equipment and ongoing operational costs, such as electricity and maintenance. The cost basis for mined cryptocurrency can be challenging to determine, which might lead to underreporting or deferral of income. Given the increasing costs and declining rewards associated with Bitcoin mining, revenue from mining has become less predictable, raising questions about the tax planning potential of mined cryptocurrencies (Hanlon et al. 2005). Moreover, prior literature suggests that firms may exploit ambiguities in the accounting rules to reduce tax liabilities, especially in industries with rapidly changing technology and valuation challenges.

Overall, the regulations governing cryptocurrency mining, trading, and other activities are still developing. Therefore, the impact of different types of involvement in cryptocurrency on corporate tax avoidance strategies remains uncertain. The relationship between cryptocurrency activities and ETR is not yet clear and warrants further empirical analysis.

Hypothesis 2 (H2). 

The type of cryptocurrency involvement—whether accepting cryptocurrency as a payment option, active trading, crypto mining, or engaging in ICOs—is expected to be either significantly negatively associated with a firm’s cash ETR or insignificantly related, but it is not expected to have a significant positive association.

3. Sample Selection and Research Design

We obtained financial statement information from the Compustat database, while data on firms’ cryptocurrency holdings were manually collected from 10-K and 10-Q reports filed on the SEC website. In our research, we began by screening all US firms with disclosed crypto holdings in their regulatory filings (i.e., 10-K and 10-Q).

Our keyword search included terms such as “Bitcoin”, “blockchain”, “crypto”, “cryptocurrency”, “digital currency”, “digital asset”, “virtual currency”, “initial coin offering”, “ICO”, “digital token”, and “Ethereum”, among others, in both singular and plural forms. This approach enabled us to identify firms that explicitly mentioned their involvement in crypto-related activities.

We downloaded all annual and quarterly reports containing these keywords and manually verified the information. On the basis of the disclosures in these reports, we determined whether a firm was truly involved in cryptocurrency activities. Following this initial identification, we further categorized these activities into specific types, including mining, trading, and accepting cryptocurrency as payment. A firm was classified as involved in cryptocurrency (with our key independent variable, Crypto, set to 1) if it engaged in activities such as accepting cryptocurrency as a payment option, actively trading cryptocurrencies on platforms such as Coinbase or Binance, mining cryptocurrencies, or offering digital tokens through ICOs.

For example, General Enterprise Ventures, Inc. (GEVI) provided the following information in their annual report for the year ending 31 December 2021, which was filed on 12 April 2022 (General Enterprise Ventures 2022). The report included three key terms—cryptocurrency, digital currency, and digital asset—highlighted in yellow, as mentioned earlier.

Revenue

Our Company generated $38,919 and $0 revenue from digital currency mining for the years ended 31 December 2021 and 2020, respectively. The Company commenced the mining of Cryptocurrency in November 2021.

Cost of Revenue

The cost of Cryptocurrency mining revenue was $18,613 and $0 for the years ended 31 December 2021 and 2020, respectively. Cost of revenue consisted of electricity expenses of $8877 and amortization of digital asset machines of $9736.

On the basis of this information, we cautiously concluded that the firm is involved in cryptocurrency, with a specific focus on cryptocurrency mining. Therefore, we categorized the two key independent variables as follows: “Crypto” was coded as one, and “Crypto_Mining” was also coded as one.

Due to the unique and varied nature of these activities across firms, a consistent and standardized definition of the indicator variable was not feasible. Our approach required creativity in defining this variable, given the limited existing studies on this topic.

We employed the propensity score matching (PSM) method for our main regression model. The rationale for this approach is twofold.

Currently, there is no authoritative accounting standard regulating the reporting and presentation of cryptocurrency transactions by firms. The newly issued accounting standard, AUS 2023-08, will only become effective for fiscal years starting after December 15, 2024. Therefore, during our sample period, firms were not required to disclose involvement in cryptocurrency in their SEC filings. As a result, the data we collected on firms’ cryptocurrency activities were largely based on voluntary disclosure.

Although our initial hand-collected dataset included approximately 1300 firm-year observations for cryptocurrency-involved firms, after excluding cases with missing dependent, independent, and control variables, our final sample was limited to 113 firm-year observations. To address the issue of data imbalance between cryptocurrency firms (113 observations) and non-cryptocurrency firms (44,670 observations) and to ensure the reliability of our statistical results, we applied the propensity score matching method. This allowed us to focus on a more balanced and smaller sample for presenting our main regression findings.

Our regression model was constructed following the established literature, incorporating control variables that have been extensively examined in prior research (McGuire et al. 2012; Armstrong et al. 2015). We also controlled for industry- and year-fixed effects, as recommended in studies by Chen et al. (2010).

To test the first hypothesis, which posits that firms involved in cryptocurrency activities are associated with a lower cash ETR, we performed the following regression analysis

$$\begin{array}{cc}\hfill {\mathrm{C}\mathrm{a}\mathrm{s}\mathrm{h}\text{\_}\mathrm{E}\mathrm{T}\mathrm{R}}_t& ={\alpha }_0+{\alpha }_1{\mathrm{C}\mathrm{r}\mathrm{y}\mathrm{p}\mathrm{t}\mathrm{o}}_{t-1}+{\alpha }_2{\mathrm{L}\mathrm{e}\mathrm{v}\mathrm{e}\mathrm{r}\mathrm{a}\mathrm{g}\mathrm{e}}_{t-1}+{\alpha }_3{\mathrm{F}\mathrm{o}\mathrm{r}\mathrm{e}\mathrm{i}\mathrm{g}\mathrm{n}}_{t-1} +{\alpha }_4{\mathrm{S}\mathrm{I}\mathrm{Z}\mathrm{E}}_{t-1 }+\hfill \\ & {\alpha }_5{\mathrm{A}\mathrm{d}\mathrm{v}\text{\_}\mathrm{e}\mathrm{x}\mathrm{p}}_{t-1}{ +\alpha }_6{\mathrm{I}\mathrm{n}\mathrm{t}\mathrm{a}\mathrm{n}\mathrm{g}\mathrm{i}\mathrm{b}\mathrm{l}\mathrm{e}}_{t-1} +{\alpha }_7{\mathrm{N}\mathrm{O}{\mathrm{L}}_{\mathrm{d}\mathrm{e}\mathrm{c}}}_{t-1 }+{\alpha }_8{\mathrm{M}\mathrm{V}\mathrm{E}}_{t-1}+\hfill \\ & {\alpha }_9{\mathrm{M}\mathrm{T}\mathrm{B}\text{\_}\mathrm{R}\mathrm{a}\mathrm{t}\mathrm{i}\mathrm{o}}_{\mathrm{t}-1}+{\alpha }_{10}{\mathrm{C}\mathrm{a}\mathrm{s}\mathrm{h}\text{\_}\mathrm{R}\mathrm{a}\mathrm{t}\mathrm{i}\mathrm{o}}_{t-1}+{\alpha }_{11}{\mathrm{R}\mathrm{O}\mathrm{A}}_{t-1} +{\alpha }_{12}{\mathrm{P}\mathrm{P}\mathrm{E}\mathrm{N}\mathrm{T}}_{t-1}+\hfill \\ & Industry and Year Effects+{\epsilon }_t\hfill \end{array}$$

(1)

where Cash_ETR_t represents the effective tax rate for a firm in year t, Crypto_t is a binary variable equal to 1 if a firm is involved in cryptocurrency activities during year t and 0 otherwise, and α₁ is expected to be negative if involvement in cryptocurrency is associated with a lower ETR.

In this study, we incorporated several control variables based on prior literature to account for factors that may influence corporate tax behavior. Leverage was included to reflect the tax shield benefits of debt, as firms with higher debt levels typically require less additional tax planning (Mackie-Mason 1990). Advertising expenses (Adv_exp) represented a firm’s marketing expenditures, which may correlate with its overall financial strategy, including tax-related decisions. To account for a firm’s incentives to avoid taxes, we included cash holdings (Cash_Ratio), as suggested by DeFond et al. (2024). Capital intensity (PPENT) captures differences in book and tax reporting, following the methodology of Dyreng et al. (2010).

We also included a variable for foreign income (Foreign) to control for income generated from foreign operations, which helped account for differences in international tax planning opportunities and varying tax rates across jurisdictions (Rego 2003). Intangible assets (Intangible) were included, as firms with significant intangible assets, such as patents and trademarks, may adopt different tax strategies due to their ability to shift profits (Dyreng et al. 2010). Additionally, we accounted for net operating loss deductions (NOL_dec), which reflected the carryforward of prior tax losses that can reduce taxable income in the current period.

Furthermore, we included non-tax control variables that captured fundamental firm characteristics, as recommended by prior research on effective tax rates (e.g., Armstrong et al. 2012; Dyreng et al. 2008, 2010). Specifically, we controlled for firm size to reflect complexity and economic scale, the market-to-book ratio (MTB_Ratio) to account for growth opportunities, and the market value of equity (MVE) to reflect the firm’s market capitalization. Lastly, return on assets (ROA) was included to control for the firm’s underlying economic activity and profitability.

To test the second hypothesis, which examines the relationship between cryptocurrency involvement and ETR, we employed a modified version of the initial regression model. In this analysis, we replaced the summary indicator variable for involvement in cryptocurrency with subcategory indicators, such as payment acceptance, trading, crypto mining, and ICO activities. This allows us to investigate whether different types of cryptocurrency activities exert distinct influences on a firm’s ETR.

4. Empirical Analysis

Table 1. Summary statistics.

	N	Mean	STD	Q1	Median	Q3
Cash_ETR	226	0.220	0.189	0.104	0.185	0.282
GAAP_ETR	226	0.200	0.134	0.135	0.206	0.259
Cash_ETR3	214	0.201	0.165	0.096	0.177	0.271
GAAP_ETR3	214	0.170	0.278	0.144	0.203	0.257
Cash_ETR5	213	0.200	0.159	0.115	0.182	0.267
GAAP_ETR3	213	0.171	0.241	0.147	0.208	0.253
Crypto	226	0.504	0.501	0.000	1.000	1.000
Payment	226	0.097	0.297	0.000	0.000	0.000
Trading	226	0.106	0.309	0.000	0.000	0.000
Crypto_Mining	226	0.013	0.115	0.000	0.000	0.000
Crypto_ICO	226	0.018	0.132	0.000	0.000	0.000
Adv_exp	226	0.014	0.033	0.000	0.000	0.015
Leverage	226	0.175	0.170	0.038	0.125	0.270
Foreign	226	0.681	0.467	0.000	1.000	1.000
Size	226	8.833	2.564	7.180	8.840	11.024
Intangible	226	0.201	0.229	0.010	0.116	0.352
NOL_dec	226	0.381	0.487	0.000	0.000	1.000
MVE	226	8.614	2.538	6.899	8.526	10.666
MTB_Ratio	226	5.903	8.886	1.567	2.821	5.490
Cash_Ratio	226	0.276	0.367	0.056	0.167	0.375
ROA	226	0.108	0.176	0.018	0.062	0.140
PPENT	226	0.104	0.165	0.013	0.043	0.115

This table presents the descriptive statistics of the regression variables for our propensity score matched sample (N = 226). All variables are defined in Appendix A.

presents the descriptive statistics of the variables used in the baseline regression. To mitigate the influence of extreme values, we winsorized each continuous variable at the top and bottom 1% of its distribution. To address the issue of imbalanced data, we employed the propensity score matching (PSM) method to construct a smaller, balanced sample for the regression analysis.

In our propensity score matched sample, the mean (median) of the cash effective tax rate (Cash_ETR) was 21.95% (18.54%), with three-year averages of approximately 20.13% (17.68%). The distribution of the effective tax rate is consistent with previous research (Dyreng et al. 2019; McGuire et al. 2012). The mean debt-to-assets ratio (Leverage) was 17.5%, while the mean cash-to-assets ratio was 27.56%. The mean firm size (Size), market-to-book ratio (MTB_Ratio), and return on assets (ROA) were USD 6.86 billion (log value of 8.833), 5.903, and 24.6%, respectively.

The decision to disclose cryptocurrency information is not random and may be influenced by firm-specific characteristics. To address potential endogeneity, we employed a propensity score matching (PSM) approach for hypothesis testing. We estimated the likelihood that each firm would be involved in cryptocurrency activities and matched each treatment firm with cryptocurrency exposure to a control firm without such exposure, using one-to-one matching based on the closest propensity score. After matching, any differences between the treatment and control firms could be attributed to cryptocurrency exposure rather than other firm characteristics. The PSM sample included 216 firm-year observations: 103 observations for cryptocurrency-holding firms and 113 observations for non-cryptocurrency firms.

Table 2. Pane A. First-stage propensity score estimation. Panel B. Propensity score matching covariate balance test.

Pane A
Dependent Variable = Crypto
	Coefficient		Z-statistics	p-Value
Adv_exp	−0.819		−0.77	0.441
Leverage	−0.190		−0.92	0.360
Foreign	0.328	***	4.22	0.000
Size	−0.002		−0.07	0.941
Intangible	−0.192		−1.11	0.266
NOL_dec	0.106		1.53	0.125
MVE	0.102	***	3.08	0.002
MTB_Ratio	0.000		0.25	0.802
Cash_Ratio	0.206	**	2.51	0.012
ROA	−0.086		−0.72	0.471
PPENT	−1.446	***	−5.48	0.000
constant	−3.508	***	−25.28	0.000
N	44,783
Pseudo R2	0.0921
Panel B
	Crypto (N = 113)	Non-crypto (N = 113)	Difference	t-Statistics
Adv_exp	0.012	0.016	−0.003	−0.720
Leverage	0.193	0.156	0.037	1.650
Foreign	0.702	0.658	0.044	0.710
Size	8.800	8.892	−0.091	−0.270
Intangible	0.220	0.182	0.038	1.250
NOL_dec	0.386	0.386	0.000	0.000
MVE	8.528	8.715	−0.187	−0.560
MTB_Ratio	6.949	4.794	2.155	1.850
Cash_Ratio	0.265	0.283	−0.018	−0.370
ROA	0.105	0.110	−0.005	−0.210
PPENT	0.096	0.110	−0.013	−0.610

Panel A, Table 2 reports the results for the first-stage regression to estimate the propensity score of holding cryptocurrency. We used a probit regression to estimate the propensity score. *** and ** indicate statistical significance at the 1% and 5%, respectively. Panel B, Table 2 reports the mean differences of determinant variables of holding cryptocurrency after propensity score matching. All variables are defined in Appendix A.

, Panel A, presents the results of the first-stage regression estimating the propensity scores, where we substituted the dependent variable with our key independent variable, Crypto, from the regression model (1). This logit regression model identifies the key determinants of cryptocurrency exposure.

The results indicate that cryptocurrency involvement is negatively associated with a firm’s investment in long-lived assets, such as property, plant, and equipment (PPENT), with a coefficient of −1.446 and a p-value of 0.000. This suggests that firms heavily investing in long-term assets are less likely to engage in cryptocurrency activities. Additionally, there is a positive association between cryptocurrency exposure and the presence of foreign operations (coefficient = 0.328; p-value = 0.000). This finding implies that firms with foreign subsidiaries are more likely to participate in cryptocurrency activities, possibly because cryptocurrencies facilitate international transactions without the need for intermediary banks. We also found that the cash ratio is positively related to cryptocurrency exposure (coefficient = 0.206; p-value = 0.012), indicating that firms with excess cash are more likely to invest in cryptocurrency, potentially seeking higher returns.

Table 2, Panel B, compares the mean values of covariates in the PSM model between firms with and without cryptocurrency exposure. After matching, none of the means of our explanatory variables were significantly different between the two subsamples. Therefore, we concluded that our PSM process effectively minimized the potential impact of endogeneity related to the selection of cryptocurrency exposure.

Table 3. Cryptocurrency holding and tax avoidance.

	(1)	(2)
	Cash_ETR	Cash_ETR3
Crypto	−0.119 ***	−0.092 *
	[0.009]	[0.058]
Adv_exp	−0.196	−0.962 *
	[0.789]	[0.055]
Leverage	−0.124	−0.081
	[0.198]	[0.347]
Foreign	0.049	0.091 **
	[0.290]	[0.015]
Size	−0.002	−0.024
	[0.940]	[0.235]
Intangible	−0.036	−0.094
	[0.646]	[0.205]
NOL_dec	−0.014	−0.017
	[0.648]	[0.554]
MVE	−0.006	0.010
	[0.794]	[0.610]
MTB_Ratio	0.003	0.002
	[0.252]	[0.494]
Cash_Ratio	−0.097 ***	−0.044
	[0.004]	[0.586]
ROA	−0.100	−0.046
	[0.109]	[0.532]
PPENT	−0.052	−0.229
	[0.707]	[0.100]
Constant	0.427 ***	0.602 ***
	[0.000]	[0.000]
Observations	226	214
R-squared	0.309	0.228

This table reports the results for the baseline regression relating cryptocurrency exposure to tax avoidance. The p-values in parentheses are robust to heteroskedasticity and firm-year clustering. ***, **, and * indicate statistical significance at the 1%, 5%, and 10% levels, respectively. All variables are defined in Appendix A.

examines the relationship between cryptocurrency exposure and corporate tax avoidance, addressing our first hypothesis. The table presents the results of the baseline regression (Equation (1)). In Column (1), the coefficient on involvement in cryptocurrency is negative and highly significant (coefficient = −0.119, p-value = 0.009), indicating that firms engaged in cryptocurrency activities pay significantly lower cash taxes. This supports our key conjecture that firms involved in cryptocurrency are associated with lower effective tax rates. To ensure robustness, we also analyzed cash tax rates over longer periods (three-year averages). The results in Column (2) (coefficient = −0.092, p-value = 0.058) continue to support our first hypothesis, showing that firms with cryptocurrency involvement have a lower tax burden, although the effect is slightly smaller over multi-year averages.

These findings contribute to the existing literature on cryptocurrency as a tool for reducing tax burdens (Marian 2013; Sanchez 2017; Atiles 2022; Cernușca et al. 2020) by providing direct empirical evidence. As discussed earlier, most prior studies have focused on anecdotes, theoretical perspectives, or survey-based empirical analysis. Our study provides the first comprehensive empirical archival evidence for all US companies, advancing the understanding of cryptocurrency’s impact on corporate tax strategies.

To address concerns regarding the smaller sample size, we conducted a Cohen’s f² analysis, which helped quantify the magnitude of the relationships observed in our regression models (Cohen 2013). In our analysis, the effect size for Model 1 (Cash_ETR) was calculated to be 0.447, which was classified as large according to Cohen’s benchmarks (with small = 0.02, medium = 0.15, and large = 0.35). The effect size for Model 2 (Cash_ETR3) was calculated to be 0.295, which fell within the medium to large range. This suggests that cryptocurrency exposure explains a substantial portion of the variance in firms’ effective tax rates, indicating a strong relationship between involvement in cryptocurrency and tax avoidance. Thus, despite the limitations of a smaller sample, the strength of the observed effects supports the reliability of our findings.

To test our second hypothesis regarding the differential impact of various types of cryptocurrency engagements on cash ETR, we replaced the general involvement in cryptocurrency variable with specific subcategory indicators—such as payment acceptance, trading, mining, and ICOs. The subsample analysis, presented in

Table 4. Cryptocurrency exposure types and tax avoidance.

	(1)	(2)	(3)	(4)
	Cash_ETR	Cash_ETR	Cash_ETR	Cash_ETR
Payment	−0.118 **
	[0.033]
Trading		−0.087 **
		[0.031]
Crypto_Mining			−0.035
			[0.635]
Crypto_ICO				0.018
				[0.784]
Adv_exp	0.016	−0.024	−0.049	−0.042
	[0.983]	[0.975]	[0.948]	[0.955]
Leverage	−0.113	−0.153	−0.110	−0.107
	[0.221]	[0.102]	[0.244]	[0.262]
Foreign	0.058	0.054	0.051	0.053
	[0.223]	[0.262]	[0.298]	[0.276]
Size	−0.009	0.006	−0.001	−0.000
	[0.694]	[0.774]	[0.975]	[0.988]
Intangible	−0.043	−0.038	−0.043	−0.043
	[0.592]	[0.632]	[0.603]	[0.596]
NOL_dec	−0.003	−0.006	−0.008	−0.007
	[0.934]	[0.848]	[0.793]	[0.810]
MVE	0.005	−0.008	−0.004	−0.004
	[0.831]	[0.722]	[0.873]	[0.860]
MTB_Ratio	0.002	0.003	0.003	0.003
	[0.448]	[0.296]	[0.315]	[0.312]
Cash_Ratio	−0.087 **	−0.093 ***	−0.086 **	−0.086 **
	[0.011]	[0.007]	[0.013]	[0.013]
ROA	−0.104	−0.122 *	−0.097	−0.101
	[0.118]	[0.086]	[0.139]	[0.132]
PPENT	−0.057	−0.067	−0.065	−0.064
	[0.691]	[0.608]	[0.643]	[0.650]
Constant	0.383 ***	0.385 ***	0.399 ***	0.397 ***
	[0.002]	[0.002]	[0.002]	[0.002]
Observations	226	226	226	226
R-squared	0.292	0.285	0.273	0.273

This table reports the results for the additional regression that related the type of cryptocurrency exposure to tax avoidance. The p-values in parentheses are robust to heteroskedasticity and firm-year clustering. *** and ** indicate statistical significance at the 1% and 5%, respectively. All variables are defined in Appendix A.

, Columns (1) through (4), demonstrated that the negative association between cryptocurrency involvement and effective tax rate is concentrated among firms that either accept cryptocurrency as a payment method or engage in cryptocurrency trading.

In Column (1), the coefficient for Payment is −0.118 (p-value = 0.033), indicating that firms accepting cryptocurrency as a payment option experience an 11.8% reduction in their cash ETR compared with firms that do not offer this option. Similarly, Column (2) shows that Trading is significantly associated with a lower cash ETR (coefficient = −0.087, p-value = 0.031). This finding supports our hypothesis that firms actively trading cryptocurrency can exert greater discretion over the timing and magnitude of realized gains from their crypto investments, thereby reducing their overall effective tax rates.

In contrast, no significant tax reduction effect was observed for firms engaged exclusively in cryptocurrency mining or ICO activities. As shown in Columns (3) and (4), the results for these subcategories are not statistically significant.

5. Summary and Concluding Remarks

This study investigated the influence of cryptocurrency activities on corporate tax avoidance strategies, with a focus on how different types of involvement in crypto affect a firm’s cash ETR. Our findings indicate that companies engaged in cryptocurrency activities tend to exhibit lower cash ETRs, driven by the distinctive financial dynamics and tax planning opportunities inherent to digital assets (Desai and Dharmapala 2006; Hanlon and Heitzman 2010). While certain activities such as crypto trading and mining may have a limited or negligible impact on ETR, primarily due to established tax treatments and the balancing effect of gains and losses (Hanlon and Shevlin 2002), the acceptance of cryptocurrency as a payment method presents significant tax optimization opportunities. This form of involvement, which entails complex valuation and reporting challenges, enables firms to defer income recognition and capitalize on favorable tax treatments, ultimately reducing their overall tax liability (Lisowsky 2010; Blouin 2014).

Our findings suggest that firms with cryptocurrency exposure tend to have lower cash effective tax rates (ETR), not only in the current year but also over a three-year horizon. This indicates that the tax benefits associated with involvement in cryptocurrency are not merely short-term but can extend across multiple years. The results align with prior literature on tax avoidance strategies, indicating that firms exploit the complexities of financial instruments and emerging assets to minimize tax liabilities over time (Desai and Dharmapala 2006; Hanlon and Heitzman 2010).

The subsample analysis provided further insights, revealing that the reduction in cash ETR is primarily driven by firms that can actively and frequently exchange cryptocurrency, such as those accepting cryptocurrency as a payment method or actively trading on exchange platforms. These firms can strategically manage the recognition of gains and losses, aligning with the literature on earnings management and the use of volatile assets for tax planning (Frank et al. 2009).

This research adds to the expanding literature on corporate taxation and emerging technologies, offering critical insights for policymakers and business leaders. As cryptocurrency continues to integrate into the corporate landscape, future research should delve deeper into the specific mechanisms through which various crypto activities impact tax outcomes and explore the long-term ramifications for corporate financial strategies. Such inquiries will not only enhance our understanding of the financial implications of cryptocurrencies but also inform the development of more effective tax policies.

Our study highlights significant policy implications for various regulatory agencies, emphasizing the urgent need for comprehensive and detailed tax regulations regarding crypto assets. As the landscape of digital currencies rapidly evolves, regulatory bodies must establish clear frameworks to address the complexities and unique characteristics of these assets. A precise definition of crypto assets is essential, and tailored tax treatments are necessary due to the challenges in classifying decentralized digital assets. Moreover, it is critical to develop detailed guidelines that assist taxpayers in complying with the regulations and to provide educational resources that enhance their understanding of crypto asset regulations. Given the empirical evidence presented in this study, which indicates that cryptocurrency exposure among public firms in the US is significantly associated with a lower effective tax rate, increased scrutiny of crypto asset transactions is warranted.