Event studies have been a long-trusted tool for measuring how significant events can influence the price of a company’s stock. For decades, they have underpinned damage quantification in disputes related to traditional securities. However, the financial landscape is evolving. Digital assets, spanning cryptocurrencies like Bitcoin, stablecoins, and non-fungible tokens (NFTs), are recognised in the global markets. Bitcoin alone commands a market capitalisation of approximately USD 1.8 trillion, with more than 700 million individuals worldwide owning cryptocurrency, and 71% of institutional investors allocated to digital assets as of mid-2025.^[1] This explosive growth has been mirrored by a sharp rise in crypto-related litigation, with U.S. court filings increasing more than tenfold in the past decade.^[2]

Against this backdrop, a critical question emerges for valuation experts and litigators alike: can the event study methodology, with its historic applications so heavily enmeshed with the traditional stock market, also be generalised to disputes involving digital assets?

This article argues that event studies can still be an effective tool for causal analysis in the digital asset space, but only if the practitioner makes the necessary accommodations for the event study’s underlying assumptions to hold in that context.

Event Studies in the Stock Market

Event studies are widely used in securities-related legal disputes, particularly in class action litigation cases involving alleged manipulations or securities fraud. Their purpose is to estimate how a specific event impacts returns (i.e., the price movements of a stock). This is conducted in three broad steps:

1. Estimate the counterfactual: this is the return that would have prevailed on the event date if the event had not occurred. It is also called the ‘expected return.’
2. Calculate the impact: this is the difference between the actual returns observed on the event date and the expected return. This difference is called the ‘abnormal return.’
3. Conduct statistical inference: This step assesses whether abnormal returns are indicative of a ‘real’ impact or simply arose due to chance.

While event studies have proven to be effective in securities litigation, their application within that context presents several well-documented challenges.^[3]

Event Studies with Digital Assets

The overarching procedure for estimating price impacts for digital assets is broadly similar to that used for securities. Furthermore, many of the statistical limitations that emerge when applying event studies to securities also hold for digital assets. However, the fundamental differences between the stock market and the digital asset market are such that the underlying framework of an event study may need to be tailored to suit the market in question for results to be accurate. We explore a subset of these divergences and their implications for event studies below.

Valuation of the underlying asset

A stock is ultimately a claim on a company’s future cash flows. This means that the value of a stock can be anchored to valuation fundamentals, such as earnings, dividends, and the risk profile of the business. While event studies do not estimate the intrinsic value of a stock, they are derived from fundamentals to create a credible baseline for what the expected return of a stock should be.

Digital assets, by contrast, do not have stable cash‑flow ‘anchors’: prices coevolve with network adoption, protocol incentives, and security (e.g., hash rate). These factors make it more difficult for equities-based event studies to establish an accurate ‘expected’ return, which may distort impact estimates.^[4] Furthermore, digital asset prices are generally much more volatile than equities prices, which may make it difficult for traditional event studies to distinguish meaningful price changes from ‘noise.’^[5]

Market centralisation

Equity markets are organised around centralised exchanges that operate under uniform trading rules, disclosure standards, and regulatory oversight. Companies listed on these exchanges largely share exposure to common macroeconomic conditions, such as interest rates and monetary policy. These factors foster a broad co-movement in stock prices, which is important for both the conceptual and empirical validity of an event study. Conceptually, this aligns with the financial theory that an even study is rooted in. Empirically, event studies leverage this through using broad market indices (e.g., the S&P 500) and/or sector-specific indices (e.g., the NASDAQ-100 Technology Index) to produce credible estimates of expected returns.

By contrast, digital assets have fragmented platforms, trade globally, and operate under unique technical protocols that define each blockchain network. As a result, price movements may be driven by asset specific or exchange-specific factors rather than by broad market movements.^[6] Consequently, event studies calibrated with broad indices may fail to accurately and precisely estimate expected returns.

Trading Continuity and Market Composition

Equities markets typically operate within fixed trading hours that define the trading day, and are dominated by large, sophisticated, and well-informed institutional investors. Event studies leverage this structure to identify impacts by assuming that:

• events occur within discrete, well-defined windows of time;^[7]
• investors behave rationally and optimally.

The structure of equities markets increases the likelihood that these conditions are met, such that event studies can produce valid estimates.

However, digital asset markets operate on a 24/7 basis; therefore, event studies may struggle to estimate event impacts due to the event window being poorly defined.^[8] Additionally, the investor pool is skewed towards retail investors who tend to be smaller, more sentiment-driven, and less well informed than their institutional counterparts. This market composition may invalidate some of the event study’s assumptions, while amplifying some of the aforementioned issues.^[9]

Potential mitigations

The points outlined above may cause event studies to produce inaccurate and imprecise estimates. This point is crucial as it may have large implications for the quantum of damages claimed. Nevertheless, practitioners can potentially mitigate some of these issues through implementing the following strategies:

1. Expand the market model: enriching market models by explicitly incorporating the specific factors that determine the returns of the digital asset in question can improve the accuracy and precision of expected returns estimates. Practitioners can combine detailed domain knowledge with quantitative techniques, such as cross-validation, to enhance conceptual validity and empirical performance.
2. Use more flexible estimation approaches: market models typically rely on linear regressions, which may be too restrictive to capture the often volatile and complex price patterns of digital assets. Using more flexible empirical approaches may improve estimation of causal impacts.^[10] However, practitioners must strike a balance between pure performance and ease of explanation when selecting their models, particularly within the dispute space.
3. Applying rolling/sliding event windows: where the event window is not clear cut, rolling windows can be used to assess how the event’s impact evolved over time and isolate ‘where’ the event is most likely to have peaked.^[11] However, practitioners must bear in mind the potential ramifications that rolling windows can have on the interpretation of reported impacts and statistical inference.

Overall, event studies hold the potential to be a valuable tool for causal discovery in the digital asset space. However, the correct accommodation, such as those outlined above, must be made with care and communicated clearly when they are necessary. Failing this, estimated impacts and damages may vastly diverge from the truth.

Why HKA?

At HKA, our team of economic experts has built a strong reputation for developing rigorous, innovative solutions for clients navigating digital asset disputes and other high‑value economic challenges. Our experts combine deep technical capability with extensive sector insight including a sophisticated understanding of the digital asset ecosystem and advanced expertise in econometric modelling and statistical analysis. This blend of specialist knowledge allows us to deliver clear, evidence‑based assessments on matters involving digital assets, market dynamics, valuation, and complex economic questions.

For more information or to discuss how we can assist, please contact:.

---

^[1] Cryptocurrency Adoption by Institutional Investors Statistics 2025 • CoinLaw

^[2] Crypto Litigation: Parent Companies and Industry Segments at a Glance.

^[3] These limitations are well documented and include: (i) limited ability to account for confounding events; (ii) incorrect modelling of counterfactual price; (iii) low statistical power to determine true price impacts; and (iv) incorrect inference when dealing with multiple events. See Event Studies in Securities Litigation: Low Power, Confounding Effects, and Bias 93 Washington University Law Review 2015-2016

^[4] This is also known as bias, whereby estimated impacts differ from their true value.

^[5] This is known as low statistical power, whereby the practitioner is more likely to dismiss an impact as statistically meaningless despite the impact being real. Academic evidence suggests event studies applied to equities may also suffer from low statistical power, but this may be exacerbated when applied to digital assets.

^[6] Some examples include regional legislation, protocol-specific developments (e.g., upgrades or outages) and sentiment shocks amplified by social media.

^[7] This is validated by anchoring events to market openings and closings, and calendar days.

^[8] For instance, price reactions to the same event may be staggered across geographies due to varying speeds of news diffusion and when the event occurs in local time, creating a blurry event window.

^[9] For instance, retail investors’ inattention, irrationality (e.g., excess ‘hype’ resulting from social media) or limited capacity to determine the underlying value of digital assets may increase price volatility. See ’The inefficiency of Bitcoin’ for a detailed discussion of some of these points. However, market composition will vary by asset and coin. For example, retail investors would be a smaller share of Bitcoin holders but could be the sole holders of other coins, such as ‘meme coins.’ Furthermore, the composition of the overall digital asset market is likely to shift in the long run as more institutional investors enter the market, thereby reducing the hurdle of market composition in empirical analysis.

^[10] Potential alternatives include Bayesian methods, Synthetic Control Methods, Machine Learning methods, and advanced Time Series techniques that account for excess volatility (e.g. GARCH models).

^[11] For example, assessing the cumulative abnormal over incrementally growing windows (e.g. hours 0-2 after the event, then hours 0-4, etc.) can determine the speed and saturation of the impact. Estimating the abnormal returns over mutually exclusive bins (e.g. hours 0-2, 2-4, etc.) allows for cleaner attribution of the event’s impacts.

---