Bitcoin's bubbly behaviors: does it resemble other financial bubbles of the past?

The paper examines whether Bitcoin prices exhibit patterns similar to well-known historical financial bubbles and whether historical regulatory responses to bubbles can inform present-day cryptocurrency regulation. The context is rapid global adoption of cryptocurrencies, high volatility, and speculation-driven demand. The authors articulate two research questions: (1) Does Bitcoin display trends akin to prior bubbles? (2) Can actions taken by authorities after historical bubbles be applied to Bitcoin to prevent future bubbles? The study positions Bitcoin’s expansions within periods of monetary expansion and excess liquidity, and postulates parallels between Bitcoin and speculative episodes such as tulip mania, the Mississippi and South Sea bubbles, the 1929 crash, and the dotcom bubble. The stated hypothesis—that Bitcoin’s price behavior parallels other speculative assets—is confirmed empirically, aiming to guide regulators on investor protection in digital assets.

The literature review outlines the concept and drivers of bubbles—exogenous shocks, technological change, credit conditions, and investor expectations—drawing on Kindleberger and others. Five reference bubbles are profiled: tulip mania (1634–1637), Mississippi (1719–1720), South Sea (1720), the 1929 NYSE crash, and the dotcom bubble (1997–2001), including their contexts, price dynamics, and policy responses (e.g., contract adjustments post-tulips; enforcement and state assumption of debts post-Mississippi; rescues post–South Sea; strengthened regulation post–dotcom). The review then surveys Bitcoin bubble research along two approaches: (A) fundamental value-based (challenging given lack of cash flows) and (B) chartist/statistical, focusing on price patterns and drivers such as media, monetary policy, and market manipulation. Prior comparative works link Bitcoin with tulips, dotcoms, and South Sea phases, noting both similarities and fundamental differences. The paper situates itself within the chartist/statistical tradition, extending Phillips et al. (2013)-style comparisons to include multiple historic bubbles and Bitcoin.

Data sources: Bitcoin (USD, daily) and Nasdaq Composite (USD, daily) from Investing.com; Dow Jones Industrial Average (USD, monthly) from Dow Jones; South Sea Company and East India Company (GBP, daily) and Mississippi Company (livres de tournois, daily) from Frehen et al. (2013); tulip prices (daily reconstructed index) from Thompson (2006). The compiled dataset contains 9,967 records and is provided in an external repository. Study periods include Bitcoin bubble phases (2011, 2013A, 2013B, 2017, 2021A, 2021B) and historical bubbles. Analytical framework: The study draws on elements of the Johansen–Ledoit–Sornette (JLS) model for rational-expectations bubbles with finite-time singularity, considering elapsed time from bubble start to end and deviations from fundamental value. The JLS framework is complemented by: - Volatility assessment via standard deviation of prices during each bubble phase. - Growth assessment via compound annual growth rate (CAGR), using minimum before maximum, maximum, and elapsed days converted to years. - Bubble Index (BI): ratio of current (peak) price to historical average price over a defined window; BI > 1 indicates overvaluation. - Bubble size: percentage increase from bubble start to peak. - Correlation analysis: Pearson correlation coefficients measuring linear relationships between Bitcoin bubble time series and selected historical bubble series. Diagnostic checks include descriptive statistics, VIF for multicollinearity (no issues), and Levene’s test for homogeneity of variances (results discussed; heteroskedasticity not indicated for most cases).

- Across multiple metrics, Bitcoin bubble phases exhibit significant similarities to tulip mania and the Mississippi bubble. - Volatility (standard deviation): Highest in Bitcoin 2021A (σ ≈ 16,569.5), 2021B (≈ 9,256.3), and 2017 (≈ 4,394.7), followed by Mississippi (≈ 2,705.3), suggesting Bitcoin’s volatility profiles resemble Mississippi in certain periods. - Bubble size (percentage increase from start to peak): Bitcoin bubbles are among the largest—BTC 2011 ≈ 31,900%; BTC 2013A ≈ 2,094.49%; BTC 2013B ≈ 1,612.86%; BTC 2017 ≈ 1,994.98%—comparable to Mississippi (≈ 3,689.50%) and tulips (≈ 2,269.66%), and to a lesser extent the South Sea Company (≈ 840.71%). - Bubble Index (BI): BTC 2021A BI ≈ 1.76, close to Mississippi (≈ 1.66) and tulips (≈ 1.73). BTC 2021B BI ≈ 1.51, proximate to Mississippi (≈ 1.66). BTC 2013B BI ≈ 2.66, near South Sea Company (≈ 2.92). - CAGR: Tulip bubble shows an extreme CAGR (≈ 918,252.56) over a short duration, while Bitcoin bubbles show high CAGRs (e.g., BTC 2013A ≈ 5,117.94; BTC 2011 ≈ 647.56), indicating fast run-ups; apart from tulips, Bitcoin’s CAGRs do not closely match others. - Duration parallels: Post-peak declines show similar day-lengths (e.g., BTC 2021A ≈ 85 days vs tulips ≈ 87 days). - Correlations: Significant relationships observed—BTC 2017 with DJIA (r ≈ 0.889); BTC 2021A and Nasdaq (≈ 0.775) and BTC 2021B with Nasdaq (≈ 0.564); Mississippi and tulips (≈ 0.907). Most Bitcoin bubbles except 2011 correlate positively with tulips (≈ 0.370 to 0.507), and BTC 2013A, 2021A, 2021B correlate with Mississippi (reported high values, strongest for BTC 2021A). - Policy relevance: Given digital, decentralized features of Bitcoin, most historical measures used to curb bubbles are unlikely to be directly effective now. Regulation of exchanges, fraud prevention, and environmental considerations are more feasible levers.

The findings affirm that Bitcoin’s price dynamics in multiple episodes mirror hallmark features of classic bubbles, particularly tulip mania and the Mississippi bubble. High volatility, rapid appreciation followed by steep declines, comparable bubble sizes, and positive correlations suggest that Bitcoin often follows a speculative pattern rather than fundamentals-based valuation. This supports the hypothesis that Bitcoin displays bubble-like behavior akin to historic episodes. However, unlike historical bubbles that ended with the asset’s disappearance or market closure, Bitcoin persists with partial recoveries, reflecting its technological nature and ongoing market adoption. The policy implications differ due to decentralization: while direct supply control or contract annulments (as in tulips or Mississippi) are infeasible, modern regulatory approaches—exchange oversight, anti-fraud enforcement, transparency, and investor protection—can mitigate risks. Findings also indicate macro-financial linkages (e.g., correlations with major equity indices), suggesting spillover considerations for regulators.

The study confirms significant parallels between Bitcoin price behavior and major historical bubbles. Bitcoin’s 2013 and 2021 episodes, in particular, align closely with tulip mania and the Mississippi bubble across volatility, bubble size, bubble index, and correlation metrics. The research advances a comparative, multi-bubble framework applying JLS elements alongside volatility, CAGR, BI, bubble size, and correlation analyses to a unified dataset. While cross-era comparisons are intrinsically challenging, the evidence supports that Bitcoin frequently exhibits bubbly behavior. Policy recommendations emphasize regulating intermediaries (exchanges), combating fraud, and enhancing investor protections, noting the limited applicability of historical measures to a decentralized digital asset. Future research should expand the bubble set (e.g., real estate) and include other major cryptocurrencies (e.g., Ether, Ripple, Cardano) to generalize findings and refine detection and policy tools.

- Cross-era comparability: Economic, legal, technological, and social contexts differ substantially across centuries, complicating direct comparisons. - Asset differences: Bitcoin (a decentralized digital asset) differs fundamentally from equities or commodities in cash flows and intrinsic valuation, affecting metric interpretation. - Ongoing market: Bitcoin continues to trade; identified bubble periods may be superseded by future episodes. - Generalizability: Results from Bitcoin may not extrapolate to all cryptocurrencies given heterogeneous dynamics and market structures. - Sample scope: Some important bubble types (e.g., real estate) were omitted, potentially affecting comparative insights and conclusions.