research paper on cryptocurrency

Thank you for visiting nature.com. You are using a browser version with limited support for CSS. To obtain the best experience, we recommend you use a more up to date browser (or turn off compatibility mode in Internet Explorer). In the meantime, to ensure continued support, we are displaying the site without styles and JavaScript.

• View all journals
• Explore content
• About the journal
• Publish with us
• Sign up for alerts
• 04 April 2022

Crypto and digital currencies — nine research priorities

• Andrew Urquhart 0 &
• Brian Lucey 1

Andrew Urquhart is professor of finance and financial technology at ICMA Centre, University of Reading, Henley Business School, Reading, UK.

You can also search for this author in PubMed   Google Scholar

Brian Lucey is professor of international finance and commodities at Trinity Business School, Trinity College, Dublin, Ireland; Institute of Business Research, University of Economics, Ho Chi Minh City, Vietnam; and Institute for Industrial Economics, Jiangxi University of Economics and Finance, Nanchang, China.

Money is at a crossroads. A race is on to decide who creates it, who can access it and how, who controls it, and to what degree and how it is regulated. The outcome could decide whether governments have access to all our financial data, whether criminals can easily launder vast sums unseen, and whether the benefits of finance can be extended to the billions of people globally who lack access to banks.

Access options

Access Nature and 54 other Nature Portfolio journals

Get Nature+, our best-value online-access subscription

24,99 € / 30 days

cancel any time

Subscribe to this journal

Receive 51 print issues and online access

185,98 € per year

only 3,65 € per issue

Rent or buy this article

Prices vary by article type

Prices may be subject to local taxes which are calculated during checkout

Nature 604 , 36-39 (2022)

doi: https://doi.org/10.1038/d41586-022-00927-5

Nakamoto, S. Decentralized Business Rev. 21260 (2008).

Corbet, S., Lucey, B., Urquhart, A. & Yarovaya, L. Int. Rev. Finan. Anal. 62 , 182–199 (2019).

Article   Google Scholar  

Teichmann, F. M. J. & Falker, M.-C. J. Money Laundering Control 24 , 775–788 (2020).

Foley, S., Karlsen, J. R. & Putnins, T. J. Rev. Finan. Stud. 32 , 1798–1853 (2019).

Shen, D., Urquhart, A. & Wang, P. Eur. Finan. Manage. 26 , 1294–1323 (2020).

Corbet, S., Lucey, B., Peat, M. & Vigne, S. Econ. Lett. 172 , 23–27 (2018).

Lucey, B. M., Vigne, S. A., Yarovaya, L. & Wang, Y. Finan. Res. Lett. 45 , 102147 (2022).

Easley, D., O’Hara, M. & Basu, S. J. Finan. Econ. 134 , 91–109 (2019).

Baur, D. G., Hong, K. & Lee, A. D. J. Int. Finan. Markets Inst. Money 54 , 177–189 (2018).

Agur, I., Ari, A. & Dell’Ariccia, G. J. Monet. Econ. 125 , 62–79 (2022).

Download references

Reprints and permissions

Competing Interests

The authors declare no competing interests.

Related Articles

• Information technology
• Mathematics and computing

Carbon pricing reduces emissions

News & Views 24 JUL 24

Megastudy shows that reminders boost vaccination but adding free rides does not

Article 26 JUN 24

Extending the Sustainable Development Goals to 2050 — a road map

Comment 17 JUN 24

Why the next pandemic could come from the Arctic — and what to do about it

World View 04 SEP 24

Waste management won’t solve the plastics problem — we need to cut consumption

News & Views 04 SEP 24

International whaling convention needs boosting, not dismantling

Correspondence 03 SEP 24

Intellectual property and data privacy: the hidden risks of AI

Career Guide 04 SEP 24

Quantum hacking looms — but ultra-secure encryption is ready to deploy

News 13 AUG 24

AI ‘deepfake’ faces detected using astronomy methods

News 22 JUL 24

Faculty Recruitment, Westlake University School of Medicine

Faculty positions are open at four distinct ranks: Assistant Professor, Associate Professor, Full Professor, and Chair Professor.

Hangzhou, Zhejiang, China

Westlake University

Postdoctoral Researcher - Neural Circuits Genetics and Physiology for Learning and Memory

A postdoctoral position is available to study molecular mechanisms, neural circuits and neurophysiology of learning and memory.

Dallas, Texas (US)

The University of Texas Southwestern Medical Center

Assistant/Associate Professor (Tenure Track) - Integrative Biology & Pharmacology

The Department of Integrative Biology and Pharmacology (https://med.uth.edu/ibp/), McGovern Medical School at The University of Texas Health Scienc...

Houston, Texas (US)

UTHealth Houston

Faculty Positions

The Yale Stem Cell Center invites applications for faculty positions at the rank of Assistant, Associate, or full Professor. Rank and tenure will b...

New Haven, Connecticut

Yale Stem Cell Center

Postdoc/PhD opportunity – Pharmacology of Opioids

Join us at MedUni Vienna to explore the pharmacology of circular and stapled peptide therapeutics targetting the κ-opioid receptor in the periphery.

Vienna (AT)

Medical University of Vienna

Sign up for the Nature Briefing newsletter — what matters in science, free to your inbox daily.

Quick links

• Explore articles by subject
• Guide to authors
• Editorial policies

• Search Menu
• Sign in through your institution
• Browse content in A - General Economics and Teaching
• Browse content in A1 - General Economics
• A11 - Role of Economics; Role of Economists; Market for Economists
• A13 - Relation of Economics to Social Values
• A14 - Sociology of Economics
• Browse content in C - Mathematical and Quantitative Methods
• Browse content in C0 - General
• C02 - Mathematical Methods
• Browse content in C1 - Econometric and Statistical Methods and Methodology: General
• C10 - General
• C11 - Bayesian Analysis: General
• C12 - Hypothesis Testing: General
• C13 - Estimation: General
• C14 - Semiparametric and Nonparametric Methods: General
• C15 - Statistical Simulation Methods: General
• C18 - Methodological Issues: General
• Browse content in C2 - Single Equation Models; Single Variables
• C21 - Cross-Sectional Models; Spatial Models; Treatment Effect Models; Quantile Regressions
• C22 - Time-Series Models; Dynamic Quantile Regressions; Dynamic Treatment Effect Models; Diffusion Processes
• C23 - Panel Data Models; Spatio-temporal Models
• Browse content in C3 - Multiple or Simultaneous Equation Models; Multiple Variables
• C32 - Time-Series Models; Dynamic Quantile Regressions; Dynamic Treatment Effect Models; Diffusion Processes; State Space Models
• C38 - Classification Methods; Cluster Analysis; Principal Components; Factor Models
• Browse content in C4 - Econometric and Statistical Methods: Special Topics
• C45 - Neural Networks and Related Topics
• Browse content in C5 - Econometric Modeling
• C50 - General
• C51 - Model Construction and Estimation
• C52 - Model Evaluation, Validation, and Selection
• C53 - Forecasting and Prediction Methods; Simulation Methods
• C55 - Large Data Sets: Modeling and Analysis
• C58 - Financial Econometrics
• Browse content in C6 - Mathematical Methods; Programming Models; Mathematical and Simulation Modeling
• C61 - Optimization Techniques; Programming Models; Dynamic Analysis
• C62 - Existence and Stability Conditions of Equilibrium
• C65 - Miscellaneous Mathematical Tools
• Browse content in C7 - Game Theory and Bargaining Theory
• C70 - General
• C72 - Noncooperative Games
• C73 - Stochastic and Dynamic Games; Evolutionary Games; Repeated Games
• C78 - Bargaining Theory; Matching Theory
• Browse content in C8 - Data Collection and Data Estimation Methodology; Computer Programs
• C81 - Methodology for Collecting, Estimating, and Organizing Microeconomic Data; Data Access
• Browse content in C9 - Design of Experiments
• C91 - Laboratory, Individual Behavior
• C92 - Laboratory, Group Behavior
• C93 - Field Experiments
• Browse content in D - Microeconomics
• Browse content in D0 - General
• D03 - Behavioral Microeconomics: Underlying Principles
• Browse content in D1 - Household Behavior and Family Economics
• D10 - General
• D11 - Consumer Economics: Theory
• D12 - Consumer Economics: Empirical Analysis
• D14 - Household Saving; Personal Finance
• D15 - Intertemporal Household Choice: Life Cycle Models and Saving
• D18 - Consumer Protection
• Browse content in D2 - Production and Organizations
• D20 - General
• D21 - Firm Behavior: Theory
• D22 - Firm Behavior: Empirical Analysis
• D23 - Organizational Behavior; Transaction Costs; Property Rights
• D24 - Production; Cost; Capital; Capital, Total Factor, and Multifactor Productivity; Capacity
• D25 - Intertemporal Firm Choice: Investment, Capacity, and Financing
• Browse content in D3 - Distribution
• D30 - General
• D31 - Personal Income, Wealth, and Their Distributions
• Browse content in D4 - Market Structure, Pricing, and Design
• D40 - General
• D43 - Oligopoly and Other Forms of Market Imperfection
• D44 - Auctions
• D47 - Market Design
• D49 - Other
• Browse content in D5 - General Equilibrium and Disequilibrium
• D50 - General
• D51 - Exchange and Production Economies
• D52 - Incomplete Markets
• D53 - Financial Markets
• Browse content in D6 - Welfare Economics
• D60 - General
• D61 - Allocative Efficiency; Cost-Benefit Analysis
• D62 - Externalities
• Browse content in D7 - Analysis of Collective Decision-Making
• D70 - General
• D71 - Social Choice; Clubs; Committees; Associations
• D72 - Political Processes: Rent-seeking, Lobbying, Elections, Legislatures, and Voting Behavior
• D73 - Bureaucracy; Administrative Processes in Public Organizations; Corruption
• D74 - Conflict; Conflict Resolution; Alliances; Revolutions
• D78 - Positive Analysis of Policy Formulation and Implementation
• Browse content in D8 - Information, Knowledge, and Uncertainty
• D80 - General
• D81 - Criteria for Decision-Making under Risk and Uncertainty
• D82 - Asymmetric and Private Information; Mechanism Design
• D83 - Search; Learning; Information and Knowledge; Communication; Belief; Unawareness
• D84 - Expectations; Speculations
• D85 - Network Formation and Analysis: Theory
• D86 - Economics of Contract: Theory
• D87 - Neuroeconomics
• Browse content in D9 - Micro-Based Behavioral Economics
• D90 - General
• D91 - Role and Effects of Psychological, Emotional, Social, and Cognitive Factors on Decision Making
• D92 - Intertemporal Firm Choice, Investment, Capacity, and Financing
• Browse content in E - Macroeconomics and Monetary Economics
• Browse content in E0 - General
• E00 - General
• E03 - Behavioral Macroeconomics
• Browse content in E1 - General Aggregative Models
• E17 - Forecasting and Simulation: Models and Applications
• Browse content in E2 - Consumption, Saving, Production, Investment, Labor Markets, and Informal Economy
• E20 - General
• E21 - Consumption; Saving; Wealth
• E22 - Investment; Capital; Intangible Capital; Capacity
• E23 - Production
• E24 - Employment; Unemployment; Wages; Intergenerational Income Distribution; Aggregate Human Capital; Aggregate Labor Productivity
• Browse content in E3 - Prices, Business Fluctuations, and Cycles
• E30 - General
• E31 - Price Level; Inflation; Deflation
• E32 - Business Fluctuations; Cycles
• E37 - Forecasting and Simulation: Models and Applications
• Browse content in E4 - Money and Interest Rates
• E40 - General
• E41 - Demand for Money
• E42 - Monetary Systems; Standards; Regimes; Government and the Monetary System; Payment Systems
• E43 - Interest Rates: Determination, Term Structure, and Effects
• E44 - Financial Markets and the Macroeconomy
• E47 - Forecasting and Simulation: Models and Applications
• Browse content in E5 - Monetary Policy, Central Banking, and the Supply of Money and Credit
• E50 - General
• E51 - Money Supply; Credit; Money Multipliers
• E52 - Monetary Policy
• E58 - Central Banks and Their Policies
• Browse content in E6 - Macroeconomic Policy, Macroeconomic Aspects of Public Finance, and General Outlook
• E60 - General
• E61 - Policy Objectives; Policy Designs and Consistency; Policy Coordination
• E62 - Fiscal Policy
• E63 - Comparative or Joint Analysis of Fiscal and Monetary Policy; Stabilization; Treasury Policy
• E64 - Incomes Policy; Price Policy
• E65 - Studies of Particular Policy Episodes
• E66 - General Outlook and Conditions
• Browse content in E7 - Macro-Based Behavioral Economics
• E71 - Role and Effects of Psychological, Emotional, Social, and Cognitive Factors on the Macro Economy
• Browse content in F - International Economics
• Browse content in F0 - General
• F02 - International Economic Order and Integration
• Browse content in F1 - Trade
• F14 - Empirical Studies of Trade
• Browse content in F2 - International Factor Movements and International Business
• F21 - International Investment; Long-Term Capital Movements
• F22 - International Migration
• F23 - Multinational Firms; International Business
• Browse content in F3 - International Finance
• F30 - General
• F31 - Foreign Exchange
• F32 - Current Account Adjustment; Short-Term Capital Movements
• F33 - International Monetary Arrangements and Institutions
• F34 - International Lending and Debt Problems
• F36 - Financial Aspects of Economic Integration
• F37 - International Finance Forecasting and Simulation: Models and Applications
• F38 - International Financial Policy: Financial Transactions Tax; Capital Controls
• Browse content in F4 - Macroeconomic Aspects of International Trade and Finance
• F40 - General
• F41 - Open Economy Macroeconomics
• F42 - International Policy Coordination and Transmission
• F43 - Economic Growth of Open Economies
• F44 - International Business Cycles
• F47 - Forecasting and Simulation: Models and Applications
• Browse content in F5 - International Relations, National Security, and International Political Economy
• F51 - International Conflicts; Negotiations; Sanctions
• Browse content in F6 - Economic Impacts of Globalization
• F63 - Economic Development
• F65 - Finance
• Browse content in G - Financial Economics
• Browse content in G0 - General
• G00 - General
• G01 - Financial Crises
• G02 - Behavioral Finance: Underlying Principles
• Browse content in G1 - General Financial Markets
• G10 - General
• G11 - Portfolio Choice; Investment Decisions
• G12 - Asset Pricing; Trading volume; Bond Interest Rates
• G13 - Contingent Pricing; Futures Pricing
• G14 - Information and Market Efficiency; Event Studies; Insider Trading
• G15 - International Financial Markets
• G17 - Financial Forecasting and Simulation
• G18 - Government Policy and Regulation
• G19 - Other
• Browse content in G2 - Financial Institutions and Services
• G20 - General
• G21 - Banks; Depository Institutions; Micro Finance Institutions; Mortgages
• G22 - Insurance; Insurance Companies; Actuarial Studies
• G23 - Non-bank Financial Institutions; Financial Instruments; Institutional Investors
• G24 - Investment Banking; Venture Capital; Brokerage; Ratings and Ratings Agencies
• G28 - Government Policy and Regulation
• G29 - Other
• Browse content in G3 - Corporate Finance and Governance
• G30 - General
• G31 - Capital Budgeting; Fixed Investment and Inventory Studies; Capacity
• G32 - Financing Policy; Financial Risk and Risk Management; Capital and Ownership Structure; Value of Firms; Goodwill
• G33 - Bankruptcy; Liquidation
• G34 - Mergers; Acquisitions; Restructuring; Corporate Governance
• G35 - Payout Policy
• G38 - Government Policy and Regulation
• G39 - Other
• Browse content in G4 - Behavioral Finance
• G40 - General
• G41 - Role and Effects of Psychological, Emotional, Social, and Cognitive Factors on Decision Making in Financial Markets
• Browse content in G5 - Household Finance
• G50 - General
• G51 - Household Saving, Borrowing, Debt, and Wealth
• G52 - Insurance
• G53 - Financial Literacy
• Browse content in H - Public Economics
• H0 - General
• Browse content in H1 - Structure and Scope of Government
• H11 - Structure, Scope, and Performance of Government
• H19 - Other
• Browse content in H2 - Taxation, Subsidies, and Revenue
• H22 - Incidence
• H24 - Personal Income and Other Nonbusiness Taxes and Subsidies; includes inheritance and gift taxes
• H25 - Business Taxes and Subsidies
• H26 - Tax Evasion and Avoidance
• Browse content in H3 - Fiscal Policies and Behavior of Economic Agents
• H31 - Household
• Browse content in H4 - Publicly Provided Goods
• H40 - General
• H41 - Public Goods
• Browse content in H5 - National Government Expenditures and Related Policies
• H50 - General
• H52 - Government Expenditures and Education
• H53 - Government Expenditures and Welfare Programs
• H54 - Infrastructures; Other Public Investment and Capital Stock
• H55 - Social Security and Public Pensions
• H56 - National Security and War
• H57 - Procurement
• Browse content in H6 - National Budget, Deficit, and Debt
• H63 - Debt; Debt Management; Sovereign Debt
• Browse content in H7 - State and Local Government; Intergovernmental Relations
• H70 - General
• H72 - State and Local Budget and Expenditures
• H74 - State and Local Borrowing
• H75 - State and Local Government: Health; Education; Welfare; Public Pensions
• Browse content in H8 - Miscellaneous Issues
• H81 - Governmental Loans; Loan Guarantees; Credits; Grants; Bailouts
• Browse content in I - Health, Education, and Welfare
• Browse content in I1 - Health
• I11 - Analysis of Health Care Markets
• I12 - Health Behavior
• I13 - Health Insurance, Public and Private
• I14 - Health and Inequality
• I18 - Government Policy; Regulation; Public Health
• Browse content in I2 - Education and Research Institutions
• I22 - Educational Finance; Financial Aid
• I23 - Higher Education; Research Institutions
• I28 - Government Policy
• Browse content in I3 - Welfare, Well-Being, and Poverty
• I30 - General
• I38 - Government Policy; Provision and Effects of Welfare Programs
• Browse content in J - Labor and Demographic Economics
• Browse content in J0 - General
• J00 - General
• Browse content in J1 - Demographic Economics
• J11 - Demographic Trends, Macroeconomic Effects, and Forecasts
• J12 - Marriage; Marital Dissolution; Family Structure; Domestic Abuse
• J13 - Fertility; Family Planning; Child Care; Children; Youth
• J15 - Economics of Minorities, Races, Indigenous Peoples, and Immigrants; Non-labor Discrimination
• J16 - Economics of Gender; Non-labor Discrimination
• J18 - Public Policy
• Browse content in J2 - Demand and Supply of Labor
• J20 - General
• J21 - Labor Force and Employment, Size, and Structure
• J22 - Time Allocation and Labor Supply
• J23 - Labor Demand
• J24 - Human Capital; Skills; Occupational Choice; Labor Productivity
• J26 - Retirement; Retirement Policies
• J28 - Safety; Job Satisfaction; Related Public Policy
• Browse content in J3 - Wages, Compensation, and Labor Costs
• J30 - General
• J31 - Wage Level and Structure; Wage Differentials
• J32 - Nonwage Labor Costs and Benefits; Retirement Plans; Private Pensions
• J33 - Compensation Packages; Payment Methods
• J38 - Public Policy
• Browse content in J4 - Particular Labor Markets
• J41 - Labor Contracts
• J44 - Professional Labor Markets; Occupational Licensing
• J45 - Public Sector Labor Markets
• J46 - Informal Labor Markets
• J49 - Other
• Browse content in J5 - Labor-Management Relations, Trade Unions, and Collective Bargaining
• J51 - Trade Unions: Objectives, Structure, and Effects
• J52 - Dispute Resolution: Strikes, Arbitration, and Mediation; Collective Bargaining
• Browse content in J6 - Mobility, Unemployment, Vacancies, and Immigrant Workers
• J61 - Geographic Labor Mobility; Immigrant Workers
• J62 - Job, Occupational, and Intergenerational Mobility
• J63 - Turnover; Vacancies; Layoffs
• J64 - Unemployment: Models, Duration, Incidence, and Job Search
• J65 - Unemployment Insurance; Severance Pay; Plant Closings
• J68 - Public Policy
• Browse content in J7 - Labor Discrimination
• J71 - Discrimination
• Browse content in J8 - Labor Standards: National and International
• J88 - Public Policy
• Browse content in K - Law and Economics
• Browse content in K1 - Basic Areas of Law
• K12 - Contract Law
• Browse content in K2 - Regulation and Business Law
• K22 - Business and Securities Law
• K23 - Regulated Industries and Administrative Law
• Browse content in K3 - Other Substantive Areas of Law
• K31 - Labor Law
• K32 - Environmental, Health, and Safety Law
• K34 - Tax Law
• K35 - Personal Bankruptcy Law
• Browse content in K4 - Legal Procedure, the Legal System, and Illegal Behavior
• K42 - Illegal Behavior and the Enforcement of Law
• Browse content in L - Industrial Organization
• Browse content in L1 - Market Structure, Firm Strategy, and Market Performance
• L10 - General
• L11 - Production, Pricing, and Market Structure; Size Distribution of Firms
• L13 - Oligopoly and Other Imperfect Markets
• L14 - Transactional Relationships; Contracts and Reputation; Networks
• L15 - Information and Product Quality; Standardization and Compatibility
• Browse content in L2 - Firm Objectives, Organization, and Behavior
• L21 - Business Objectives of the Firm
• L22 - Firm Organization and Market Structure
• L23 - Organization of Production
• L24 - Contracting Out; Joint Ventures; Technology Licensing
• L25 - Firm Performance: Size, Diversification, and Scope
• L26 - Entrepreneurship
• L29 - Other
• Browse content in L3 - Nonprofit Organizations and Public Enterprise
• L33 - Comparison of Public and Private Enterprises and Nonprofit Institutions; Privatization; Contracting Out
• Browse content in L4 - Antitrust Issues and Policies
• L43 - Legal Monopolies and Regulation or Deregulation
• L44 - Antitrust Policy and Public Enterprises, Nonprofit Institutions, and Professional Organizations
• Browse content in L5 - Regulation and Industrial Policy
• L51 - Economics of Regulation
• Browse content in L6 - Industry Studies: Manufacturing
• L66 - Food; Beverages; Cosmetics; Tobacco; Wine and Spirits
• Browse content in L8 - Industry Studies: Services
• L81 - Retail and Wholesale Trade; e-Commerce
• L85 - Real Estate Services
• L86 - Information and Internet Services; Computer Software
• Browse content in L9 - Industry Studies: Transportation and Utilities
• L92 - Railroads and Other Surface Transportation
• L94 - Electric Utilities
• Browse content in M - Business Administration and Business Economics; Marketing; Accounting; Personnel Economics
• Browse content in M0 - General
• M00 - General
• Browse content in M1 - Business Administration
• M12 - Personnel Management; Executives; Executive Compensation
• M13 - New Firms; Startups
• M14 - Corporate Culture; Social Responsibility
• M16 - International Business Administration
• Browse content in M2 - Business Economics
• M20 - General
• M21 - Business Economics
• Browse content in M3 - Marketing and Advertising
• M30 - General
• M31 - Marketing
• M37 - Advertising
• Browse content in M4 - Accounting and Auditing
• M40 - General
• M41 - Accounting
• M42 - Auditing
• M48 - Government Policy and Regulation
• Browse content in M5 - Personnel Economics
• M51 - Firm Employment Decisions; Promotions
• M52 - Compensation and Compensation Methods and Their Effects
• M54 - Labor Management
• Browse content in N - Economic History
• Browse content in N1 - Macroeconomics and Monetary Economics; Industrial Structure; Growth; Fluctuations
• N10 - General, International, or Comparative
• N12 - U.S.; Canada: 1913-
• Browse content in N2 - Financial Markets and Institutions
• N20 - General, International, or Comparative
• N21 - U.S.; Canada: Pre-1913
• N22 - U.S.; Canada: 1913-
• N23 - Europe: Pre-1913
• N24 - Europe: 1913-
• N25 - Asia including Middle East
• N27 - Africa; Oceania
• Browse content in N3 - Labor and Consumers, Demography, Education, Health, Welfare, Income, Wealth, Religion, and Philanthropy
• N32 - U.S.; Canada: 1913-
• Browse content in N4 - Government, War, Law, International Relations, and Regulation
• N43 - Europe: Pre-1913
• Browse content in N7 - Transport, Trade, Energy, Technology, and Other Services
• N71 - U.S.; Canada: Pre-1913
• Browse content in N8 - Micro-Business History
• N80 - General, International, or Comparative
• N82 - U.S.; Canada: 1913-
• Browse content in O - Economic Development, Innovation, Technological Change, and Growth
• Browse content in O1 - Economic Development
• O11 - Macroeconomic Analyses of Economic Development
• O12 - Microeconomic Analyses of Economic Development
• O13 - Agriculture; Natural Resources; Energy; Environment; Other Primary Products
• O16 - Financial Markets; Saving and Capital Investment; Corporate Finance and Governance
• O17 - Formal and Informal Sectors; Shadow Economy; Institutional Arrangements
• Browse content in O2 - Development Planning and Policy
• O23 - Fiscal and Monetary Policy in Development
• Browse content in O3 - Innovation; Research and Development; Technological Change; Intellectual Property Rights
• O30 - General
• O31 - Innovation and Invention: Processes and Incentives
• O32 - Management of Technological Innovation and R&D
• O33 - Technological Change: Choices and Consequences; Diffusion Processes
• O34 - Intellectual Property and Intellectual Capital
• O35 - Social Innovation
• O38 - Government Policy
• Browse content in O4 - Economic Growth and Aggregate Productivity
• O40 - General
• O43 - Institutions and Growth
• Browse content in O5 - Economywide Country Studies
• O53 - Asia including Middle East
• Browse content in P - Economic Systems
• Browse content in P1 - Capitalist Systems
• P16 - Political Economy
• P18 - Energy: Environment
• Browse content in P2 - Socialist Systems and Transitional Economies
• P26 - Political Economy; Property Rights
• Browse content in P3 - Socialist Institutions and Their Transitions
• P31 - Socialist Enterprises and Their Transitions
• P34 - Financial Economics
• P39 - Other
• Browse content in P4 - Other Economic Systems
• P43 - Public Economics; Financial Economics
• P48 - Political Economy; Legal Institutions; Property Rights; Natural Resources; Energy; Environment; Regional Studies
• Browse content in Q - Agricultural and Natural Resource Economics; Environmental and Ecological Economics
• Browse content in Q0 - General
• Q02 - Commodity Markets
• Browse content in Q3 - Nonrenewable Resources and Conservation
• Q31 - Demand and Supply; Prices
• Q32 - Exhaustible Resources and Economic Development
• Browse content in Q4 - Energy
• Q40 - General
• Q41 - Demand and Supply; Prices
• Q42 - Alternative Energy Sources
• Q43 - Energy and the Macroeconomy
• Browse content in Q5 - Environmental Economics
• Q50 - General
• Q51 - Valuation of Environmental Effects
• Q53 - Air Pollution; Water Pollution; Noise; Hazardous Waste; Solid Waste; Recycling
• Q54 - Climate; Natural Disasters; Global Warming
• Q56 - Environment and Development; Environment and Trade; Sustainability; Environmental Accounts and Accounting; Environmental Equity; Population Growth
• Browse content in R - Urban, Rural, Regional, Real Estate, and Transportation Economics
• Browse content in R0 - General
• R00 - General
• Browse content in R1 - General Regional Economics
• R10 - General
• R11 - Regional Economic Activity: Growth, Development, Environmental Issues, and Changes
• R12 - Size and Spatial Distributions of Regional Economic Activity
• Browse content in R2 - Household Analysis
• R20 - General
• R21 - Housing Demand
• R23 - Regional Migration; Regional Labor Markets; Population; Neighborhood Characteristics
• Browse content in R3 - Real Estate Markets, Spatial Production Analysis, and Firm Location
• R30 - General
• R31 - Housing Supply and Markets
• R32 - Other Spatial Production and Pricing Analysis
• R33 - Nonagricultural and Nonresidential Real Estate Markets
• R38 - Government Policy
• Browse content in R4 - Transportation Economics
• R41 - Transportation: Demand, Supply, and Congestion; Travel Time; Safety and Accidents; Transportation Noise
• Browse content in R5 - Regional Government Analysis
• R51 - Finance in Urban and Rural Economies
• Browse content in Z - Other Special Topics
• Browse content in Z1 - Cultural Economics; Economic Sociology; Economic Anthropology
• Z11 - Economics of the Arts and Literature
• Z13 - Economic Sociology; Economic Anthropology; Social and Economic Stratification
• Advance articles
• Editor's Choice
• Author Guidelines
• Submission Site
• Open Access
• About The Review of Financial Studies
• Editorial Board
• Advertising and Corporate Services
• Journals Career Network
• Self-Archiving Policy
• Dispatch Dates
• Terms and Conditions
• Journals on Oxford Academic
• Books on Oxford Academic

Article Contents

1. data and basic characteristics, 2. cryptocurrency-specific factors, 3. exposures to other assets, 4. additional results, 5. conclusion, acknowledgement, risks and returns of cryptocurrency.

• Article contents
• Figures & tables
• Supplementary Data

Yukun Liu, Aleh Tsyvinski, Risks and Returns of Cryptocurrency, The Review of Financial Studies , Volume 34, Issue 6, June 2021, Pages 2689–2727, https://doi.org/10.1093/rfs/hhaa113

• Permissions Icon Permissions

We establish that cryptocurrency returns are driven and can be predicted by factors that are specific to cryptocurrency markets. Cryptocurrency returns are exposed to cryptocurrency network factors but not cryptocurrency production factors. We construct the network factors to capture the user adoption of cryptocurrencies and the production factors to proxy for the costs of cryptocurrency production. Moreover, there is a strong time-series momentum effect, and proxies for investor attention strongly forecast future cryptocurrency returns.

Cryptocurrency is a recent phenomenon that is receiving significant attention. On the one hand, it is based on a fundamentally new technology, the potential of which is not fully understood. On the other hand, at least in the current form, it fulfills similar functions as other, more traditional assets. Extensive academic attention has focused on developing theoretical models of cryptocurrencies. The theoretical literature on cryptocurrencies has suggested a number of factors that are potentially important in the valuation of cryptocurrencies. The first group of papers builds models stressing the network effect of cryptocurrency adoption (e.g., Pagnotta and Buraschi 2018 ; Biais et al. 2018 ; Cong, Li, and Wang 2019 ) and emphasizes the price dynamics induced by the positive externality of the network effect. The second group of papers focuses on the production side of the coins—the miners’ problem (e.g., Cong, He, and Li 2018 ; Sockin and Xiong 2019 )—and shows that the evolution of cryptocurrency prices is linked to the marginal cost of production. The third group of papers ties the movements of cryptocurrency prices to those of traditional asset classes such as fiat money (e.g., Athey et al. 2016 ; Schilling and Uhlig 2019 ; Jermann 2018 ). There is also a growing literature on the empirical regularities of cryptocurrencies. Borri (2019) shows that individual cryptocurrencies are exposed to cryptomarket tail-risks. Makarov and Schoar (2020) find that cryptocurrency markets exhibit periods of potential arbitrage opportunites across exchanges. Griffin and Shams (2020) study Bitcoin price manipulation. Our paper is the first comprehensive analysis of cryptocurrencies through the lens of empirical asset pricing. Its contribution is twofold. First, it tests the mechanisms and predictions of the existing theoretical models. Second, it establishes a set of basic asset pricing facts for this asset class, which provides a common benchmark that the current and future models of cryptocurrencies should take into consideration.

We start by constructing an index of cryptocurrency (or coin) market returns. This index is the value-weighted returns of all the coins with capitalizations of more than 1 million USD (1,707 coins in total) and covers the period of January 1, 2011, to December 31, 2018. We now describe some basic statistical properties of this index. During the sample period, the averages of the coin market returns at the daily, weekly, and monthly frequencies are 0.46%, 3.44%, and 20.44%, respectively. The daily, weekly, and monthly standard deviations of the coin market returns are 5.46%, 16.50%, and 70.80%, respectively. The coin market returns have positive skewness and kurtosis. We observe that the mean and standard deviation of the coin market returns are an order of magnitude higher than those of the stock returns during the same period. The Sharpe ratios at the daily and weekly levels are about 60% and 90% higher, and the Sharpe ratio at the monthly level is comparable to those of stocks. The returns have positive skewness increasing with the frequencies from daily to monthly. The returns experience high probabilities of extreme losses and gains. For example, an extreme loss of the daily 20% negative return on the coin market happens with a probability of 0.48%, while an extreme gain of the same size occurs with a probability of 0.89%.

We next turn to examine the relationship between the coin market returns and the main cryptocurrency-specific factors that are proposed in the theoretical literature. We formulate and investigate potential drivers and predictors for cryptocurrency returns. Specifically, we construct cryptocurrency network factors, cryptocurrency production factors, cryptocurrency momentum, proxies for average and negative investor attention, and proxies for cryptocurrency valuation ratios. For each of these factors, we aim to provide a number of possible empirical measures, as there are no canonical ways to define them in the cryptocurrency market.

We consider five measures to capture the cryptocurrency network effect. Consistent with the cryptocurrency models based on the network effect, 1 we find that the coin market returns are positively and significantly exposed to cryptocurrency network growth. Furthermore, we show that the evolution of cryptocurrency prices not only reflects current cryptocurrency adoption but also contains information about expected future network growth.

We then study the implications of the cryptocurrency models based on the miners’ production problem. 2 We construct production factors of cryptocurrency to proxy for the cost of mining and test the relationship between these production factors and cryptocurrency prices. To the first approximation, mining a cryptocurrency requires two inputs: electricity and computer power. We separately construct eight proxies for electricity costs and six proxies for computing costs. For electricity, we use time-varying and location-specific measures of the price, consumption, and generation of electricity in the United States and China (including Sichuan province, which hosts the largest mining farm in the world). For proxies of computing costs, we use the prices of Bitmain Antminer, one of the common Bitcoin mining equipments, as our primary measure. We also consider indirect measures—the stock returns of the companies that are major manufacturers of mining chips. Overall, we find that the coin market returns are not significantly exposed to the cryptocurrency production factors.

The existing theoretical models of cryptocurrencies have a number of implications for the predictability of cryptocurrency returns. Some papers argue that the evolution of cryptocurrency prices should follow a martingale, and thus cryptocurrency returns are not predictable (e.g., Schilling and Uhlig 2019 ). Other papers argue that, in dynamic cryptocurrency valuation models, cryptocurrency returns could potentially be predicted by momentum, investor attention, and cryptocurrency valuation ratios (e.g., Cong, Li, and Wang 2019 ; Sockin and Xiong 2019 ). We show that momentum and investor attention strongly predict future cryptocurrency cumulative returns, but cryptocurrency valuation ratios do not.

First, we show that there is a significant time-series momentum phenomenon in the cryptocurrency market. We find that the current coin market returns predict cumulative future coin market returns from one week to eight weeks ahead. For example, a one-standard-deviation increase in the current coin market returns predicts a 3.30% increase in the weekly returns over the next week. Grouping weekly returns by terciles, we find that the top terciles outperform the bottom terciles over the one- to four-week horizons. For example, at the one-week horizon, the average return of the top tercile is 8.01% per week with a t -statistic of 4.30, while the average return of the bottom tercile is only 1.10% per week with a t -statistic of 0.92. The time-series momentum results are valid both in sample and out of sample.

Second, we construct proxies for investor attention with Google searches and show that high investor attention predicts high future returns over the one- to six-week horizons. For example, a one-standard-deviation increase in the investor attention measure yields a 3.0% increase in the 1-week-ahead future coin market returns. At the one-week horizon, the average return of the investor attention tercile is 6.53% per week with a t -statistic of 3.82, while the average return of the bottom tercile is only 0.43% per week with a t -statistic of 0.42. Another proxy for investor attention we construct is Twitter post counts, and we reach similar results with the Twitter measure. Additionally, we construct a proxy for negative investor attention and show that relatively high negative investor attention negatively predicts future cumulative coin market returns.

Research on the equity market (e.g., Hong, Lim, and Stein 2000 ; Hou, Xiong, and Peng 2009 ) shows that there is a strong interaction between momentum and investor attention. Sockin and Xiong (2019) also show that investor attention can generate momentum in the cryptocurrency market, and in their model, the momentum effect disappears controlling for investor attention. We investigate whether there is a similar interaction between momentum and investor attention in the cryptocurrency market. We find that investor attention is high during and after periods of high coin market returns. However, in a bivariate coin market predictability regression with both variables, we show that the two effects do not subsume each other. Finally, we test whether the magnitude of the momentum effect is different during periods of high investor attention and vice versa. In contrast to the equity market, we show that there is limited interaction between cryptocurrency momentum and investor attention.

Moreover, we test whether the cryptocurrency valuation ratios similar to those in the financial markets can predict future coin market returns. In the equity market, the fundamental-to-market ratios are commonly referred to as valuation ratios and are measured as the ratio of the book value of equity to the market value of equity or some other ratio of fundamental value to market value. It is more difficult to define a similar measure of the fundamental value for cryptocurrency. In their dynamic cryptocurrency asset pricing model, Cong, Li, and Wang (2019) show that the cryptocurrency fundamental-to-value ratio, defined as the number of user adoptions over market capitalization, negatively predicts future cryptocurrency returns. Motivated by the theoretical model and studies of other financial markets, we construct six cryptocurrency valuation ratios and test the return predictability of these valuation ratios. Although the coefficient estimates are consistently negative, none of the six cryptocurrency valuation ratios predict future cumulative coin market returns significantly.

Another approach to study what cryptocurrencies represent is to examine the exposures of cryptocurrency returns to other asset classes. In other words, we assess how investors and markets value current and future prospects of cryptocurrencies. The theoretical literature and the community of cryptocurrency have proposed various narratives for what cryptocurrencies represent. Schilling and Uhlig (2019) argue that, in an endowment economy where fiat money and cryptocurrency coexist and compete, the cryptocurrency returns comove with the price evolution of the fiat money. Athey et al. (2016) emphasize the importance of currency exchange rates on cryptocurrency prices. Another popular narrative is that cryptocurrency is “digital gold” and represents a new way to store value. Specifically, we study whether major cryptocurrencies comove with currencies, commodities, stocks, and macroeconomic factors. In contrast to these popular explanations, we find that the exposures of cryptocurrencies to these traditional assets are low. Overall, there is little evidence, in the view of the markets, behind the narrative that there are similarities between cryptocurrencies and these traditional assets.

We note several additional results. First, we acknowledge that we have a short time series and that there is much uncertainty and learning about cryptocurrencies during the sample period. We show that our main results are similar for the first half and the second half of the sample. Second, we discuss the relationship between the cryptocurrency time-series momentum and cross-sectional momentum. Third, we investigate the importance of regulative events in affecting cryptocurrency prices, and show that negative regulative events but not positive regulative events significantly affect cryptocurrency prices. Fourth, we examine the importance of speculative interests in driving cryptocurrency prices. We show that cryptocurrency returns are higher when speculative interests increase, but the coefficient estimates are only marginally significant. Fifth, we construct a direct measure of cryptocurrency investor sentiment and show that the expected coin market return is higher when investor sentiment is high. In the multivariate regressions with the sentiment, investor attention, and momentum measures, all three variables are statistically significant in predicting future cryptocurrency returns. Sixth, we test the role of beauty contests in the cryptocurrency market. Motivated by Biais and Bossaerts (1998) , we use the volume-volatility ratio to capture the degree of disagreement in the cryptocurrency market and show that cryptocurrency return is high when the current volume-volatility ratio is high. Seventh, we conduct a VAR analysis with the coin market returns and the different measures of coin network growth measures. Eighth, we test the effect of production factors with an alternative specification. Lastly, we examine the subsample results based on cryptocurrency characteristics.

Our paper uses standard textbook empirical asset pricing tools and methods, the discussion of which we mostly omit for conciseness. Our findings on momentum are related to a series of papers such as Jegadeesh and Titman (1993) , Moskowitz and Grinblatt (1999) , Moskowitz, Ooi, and Pedersen (2012) , Asness, Moskowitz, and Pedersen (2013) , and Daniel and Moskowitz (2016) . Da, Engelberg, and Gao (2011) use Google searches to proxy for investor attention.

Yermack (2015) is one of the first papers that brought academic attention to the field of cryptocurrency. Several recent articles document individual facts related to cryptocurrency investment. Stoffels (2017) studies cross-sectional cryptocurrency momentum. Hu, Parlour, and Rajan (2018) show that individual cryptocurrency returns correlate with Bitcoin returns. Borri (2019) shows that individual cryptocurrencies are exposed to cryptomarket tail-risks. Makarov and Schoar (2020) and Borri and Shakhnov (2018) find that cryptocurrency markets exhibit periods of potential arbitrage opportunites across exchanges. Griffin and Shams (2020) study Bitcoin price manipulation. Corbet et al. (2019) studies cryptocurrencies as a financial asset. Moreover, a number of recent papers develop models of cryptocurrencies (see, e.g., Weber 2016 ; Huberman, Leshno, and Moallemi 2017 ; Biais et al. 2018 ; Chiu and Koeppl 2017 ; Cong and He 2019 ; Cong, Li, and Wang 2019 ; Cong, He, and Li 2018 ; Sockin and Xiong 2019 ; Saleh 2018 ; Schilling and Uhlig 2019 ; Jermann 2018 ; Abadi and Brunnermeier 2018 ; Routledge and Zetlin-Jones 2018 ).

We collect trading data of all cryptocurrencies available from Coinmarketcap.com. Coinmarketcap.com is a leading source of cryptocurrency price and volume data. It aggregates information from over 200 major exchanges and provides daily data on opening, closing, high, and low prices, as well as volume and market capitalization (in dollars) for most of the cryptocurrencies. 3 For each cryptocurrency on the website, Coinmarketcap.com calculates its price by taking the volume-weighted average of all prices reported at each market. A cryptocurrency needs to meet a list of criteria to be listed, such as being traded on a public exchange with an application programming interface (API) that reports the last traded price and the last 24-hour trading volume, and having a nonzero trading volume on at least one supported exchange so that a price can be determined. Coinmarketcap.com lists both active and defunct cryptocurrencies, thus alleviating concerns about survivorship bias.

We first construct a coin market return as the value-weighted return of all the underlying coins. We use daily close prices to construct daily coin market returns. The weekly and monthly coin market returns are calculated from the daily coin market returns. We require the coins to have information on price, volume, and market capitalization. We further exclude coins with market capitalizations of less than 1,000,000 USD. For earlier years that are not covered by Coinmarketcap.com, we splice the coin market returns with Bitcoin returns from earlier years. The data of the earlier year Bitcoin returns are from CoinDeck and span from January 1, 2011, to April 29, 2013. We start from January 1, 2011, because there was not much liquidity and trading before that date. Altogether, the index of the coin market return covers the period from January 1, 2011, to December 31, 2018.

We use four primary measures to proxy for the network effect of user adoption: the number of wallet users, the number of active addresses, the number of transaction count, and the number of payment count. The data of wallet users are from Blockchain.info. We obtain data on active addresses, transaction count, and payment count from Coinmetrics.io. We use seven primary production factors to proxy for the cost of mining: the average price of electricity in the United States, the net generation of electricity of all sectors in the United States, the total electricity consumption of all sectors in the United States, the average price of electricity in China, and the average price of electricity in Sichuan province. We obtain data on the average price of electricity in the United States, the net generation of electricity of all sectors in the United States, and the total electricity consumption of all sectors in the United States from the U.S. Energy Information Administration. We obtain data on the average price of electricity in China and the average price of electricity in Sichuan province from the National Bureau of Statistics of China and the Price Monitoring Center, NDRC. Our primary computing cost data are the prices of Bitmain Antminer. We extract the Bitmain Antminer data from Keepa.com. The data for Bitmain Antminer start from September 2015.

Google search data series are downloaded from Google. Twitter post counts for the word “Bitcoin” are downloaded from Crimson Hexagon. 4 The spot exchange rates in units of U.S. dollars per foreign currency are from the Federal Reserve Bank of St. Louis. We focus on five major currencies: Australian dollar, Canadian dollar, euro, Singaporean dollar, and U.K. pound. The spot prices of precious metals are from several sources. The gold and silver prices are from the London Bullion Market Association (LBMA). Platinum prices are from the London Platinum and Palladium Market (LPPM).

Aggregate and individual stock returns are from CRSP. Detailed SIC three-digit industry return data series are constructed using individual stock returns. Chinese stock return data are from CSMAR. We build the value-weighted aggregate Chinese stock returns and detailed CIC (China Industry Classification) industry return data series from the individual stocks. The data series of Chinese stock returns last until December 2016. The return series of the 155 anomalies are downloaded from Andrew Chen’s website. 5

We obtain data on the Fama-French three-factor, Carhart four-factor, Fama-French five-factor, and Fama-French six-factor models from Kenneth French’s website. We also collect the return series of Fama-French 30 industries, Europe, Japan, AsiaExJapan, and North America from Kenneth French’s website.

The macroeconomic data series are from the website of the Federal Reserve Bank of St. Louis. Nondurable consumption is defined as the sum of personal consumption expenditures: nondurable goods, and personal consumption expenditures: services.

Stock market prices, dividends, and earnings, as well as the three-month Treasury bill rates, are from Robert Shiller’s website. Using these data series, we construct the stock market price-to-dividend ratio (pd), price-to-earnings ratio (pe), and the relative bill rate (tbill). The relative bill rate is defined as the three-month Treasury bill rate minus its 12 month backward moving average. Credit spread (credit) is defined as the yield spread between BAA corporate bonds and AAA corporate bonds. Term spread (term) is defined as the yield spread between the 10-year Treasury and 3-month Treasury. Data series on the BAA corporate yield, AAA corporate yield, 10-year Treasury yield, and 3-month Treasury yield are from the Federal Reserve Bank of St. Louis’s website.

We now document the main statistical properties of the time series for the coin market returns. Figure 1 shows the return distributions of coin market returns and coin market log returns at daily, weekly, and monthly frequencies. Figure 2 plots the price movements of the coin market compared with those of the three major cryptocurrencies. There are strong comovements across the three major cryptocurrencies. Table 1 compares the properties of the coin market returns with those of Bitcoin returns, Ethereum returns, Ripple returns, and stock market returns.

Coin market return distributions

This figure plots the distributions of daily, weekly, and monthly cryptocurrency returns and log returns.

Cryptocurrency market returns and major coins

This figure plots the cryptocurrency market returns against Bitcoin, Ethereum, and Ripple. The figures show the value of investment over time for one dollar of investment at the starting point of the graphs. The Bitcoin graph starts at April 29, 2013. The Ethereum graph starts at August 8, 2015. The Ripple graph starts at August 5, 2013.

Summary statistics

This table documents the summary statistics of the coin market returns (CMKT). Panel A reports the daily, weekly, and monthly summary statistics of the coin market index and compares them with returns for Bitcoin, Ethereum, Ripple, and the stock market. The mean, standard deviation, t -statistics, Sharpe ratio, skewness, kurtosis, and the percentage of obervations that are positive are reported. Panel B reports the percentage of extreme events based on the daily coin market index returns. The coin market returns, the Bitcoin returns, and the stock market returns are from January 1, 2011, to December 31, 2018. The Ethereum returns are from August 8, 2015, to December 31, 2018. The Ripple returns are from August 5, 2013 to December 31, 2018.

Table 1 shows the statistics of the coin market returns at the daily, weekly, and monthly frequencies compared with those of the stock market returns. Both the average and the standard deviation of the coin market returns are very high. At the daily frequency, the mean return is 0.46% and the standard deviation is 5.46%; at the weekly frequency, the mean return is 3.44% and the standard deviation is 16.50%; at the monthly frequency, the mean return is 20.44% and the standard deviation is 70.80%. Both the means and the standard deviations are an order of magnitude higher than those for the stock market returns. These facts are broadly known.

The Sharpe ratios of the coin market returns are 0.08 at the daily frequency, 0.21 at the weekly frequency, and 0.29 at the monthly frequency. At the daily and weekly frequencies, the Sharpe ratios of the coin market are about 60% and 90% higher than those of the stock market for the comparable time period. At the monthly frequency, the Sharpe ratio is similar to that of the stock market for the comparable time period.

We compare the characteristics of the coin market returns to those of the Bitcoin, Ripple, and Ethereum returns. Note that the Ripple return series starts on August 4, 2013, and the Ethereum return series starts on August 7, 2015. For the Bitcoin returns, the Sharpe ratios are 0.08 at the daily frequency, 0.21 at the weekly frequency, and 0.29 at the monthly frequency. For Ethereum, the Sharpe ratios are 0.08 at the daily frequency, 0.20 at the weekly frequency, and 0.36 at the monthly frequency. The Ethereum returns have a higher mean and standard deviation than the coin market returns. For the Ripple returns, the Sharpe ratios are 0.07 at the daily frequency, 0.13 at the weekly frequency, and 0.24 at the monthly frequency. The Ripple returns have a markedly higher mean and standard deviation compared with those of the coin market returns. The Sharpe ratios of Ripple returns are lower than those of the coin market returns at all three frequencies.

The coin market returns are positively skewed at all frequencies, in contrast to the stock returns, which are negatively skewed. The skewness increases from 0.74 at the daily frequency to 1.74 at the weekly frequency, and to 4.37 at the monthly frequency. The corresponding kurtosis is 15.52 at the daily frequency, 10.22 at the weekly frequency, and 26.54 at the monthly frequency. All three of the major cryptocurrencies have positive skewness and high kurtosis. The coin market returns have high probabilities of exceptional negative and positive daily returns. For example, the probability of a –20% daily return is almost 0.5%, and the probability of a 20% daily return is almost 0.9%.

In the Internet Appendix , we also show the mean, standard deviation, and Sharpe ratios of the returns on different days of the week. In contrast to the stocks, there is no pronounced Monday effect. However, the returns are lower on Saturdays: the average Sunday coin market return is 0.28% with a Sharpe ratio of 0.05, compared with a 0.46% daily average with a Sharpe ratio of 0.08; the average Sunday Bitcoin is 0.29% with a Sharpe ratio of 0.06, compared with a 0.46% daily average with a Sharpe ratio of 0.08; the average Sunday Ethereum is 0.25% with a Sharpe ratio of 0.03, compared with a 0.60% daily average with a Sharpe ratio of 0.08; and the average Sunday Ethereum is –0.15% with a Sharpe ratio of –0.02, compared with a 0.53% daily average with a Sharpe ratio of 0.07. While the coin market and Bitcoin returns are somewhat lower on Sundays, the returns on Saturday are consistently lower.

The theoretical literature has proposed a number of cryptocurrency-specific factors as drivers of cryptocurrency prices and as predictors of cryptocurrency returns. In this section, we develop and investigate the implications of cryptocurrency-specific factors. We first construct cryptocurrency network and production factors. We find that the coin market returns are strongly exposed to the network factors but not the production factors. Then, we test if cryptocurrency returns are predictable by studying whether different cryptocurrency-specific factors can predict future coin market returns. We consider momentum, proxies for investor attention, and proxies for cryptocurrency valuation ratios. All of these variables are specific to the cryptocurrency markets. We find that momentum and proxies for investor attention can account for future coin market returns, and thus strongly reject the notion that cryptocurrency prices are a martingale.

2.1 Network factors

The theoretical literature on cryptocurrency has emphasized the importance of network factors in the valuation of cryptocurrencies (e.g., Cong, Li, and Wang 2019 ; Sockin and Xiong 2019 ; Pagnotta and Buraschi 2018 ; Biais et al. 2018 ). In particular, the network effect of user adoption can potentially play a central role in the valuation of cryptocurrencies. Because users’ adoption of cryptocurrencies generates positive network externality, cryptocurrency prices respond to user adoptions. Hence, variations in user adoptions of the cryptocurrency network could contribute to movements in cryptocurrency prices.

We construct network factors of cryptocurrency and test whether these factors can account for variations in cryptocurrency prices. We use four measures to proxy for the network effect: the number of wallet users, the number of active addresses, the number of transaction count, and the number of payment count. 6 Thus, we measure cryptocurrency network growth using the wallet user growth, active address growth, transaction count growth, and payment count growth. We also construct a composite measure by taking the first principal component of the four primary measures, which we denote as |$PC^{network}$|⁠ . Panel A of Table 2 reports the correlation across the network factors we consider. The four primary measures correlate with each other positively, with correlations ranging from 0.17 to 0.77. The first principal component of the four demand factors strongly correlates with all four of the primary measures. The first principal component has correlations of 0.45, 0.88, 0.88, and 0.90 with the wallet user growth measure, the active address growth measure, the transaction count growth measure, and the payment count growth measure, respectively.

Cryptocurrency return loadings to network factors

This table reports the factor loadings of the coin market returns on the network factors. The network factors include wallet user growth, active address growth, transaction count growth, payment count growth, and the first principal component of the four primary measures. Panel A shows the correlation matrix of the variables. Panel B reports the loadings of the coin market returns on the network factors. The standard t -statistic is reported in parentheses. *, **, and *** denote significance levels at the 10%, 5%, and 1% levels based on the standard t -statistics. The data frequency is monthly.

We regress the coin market returns on each of the four measures of changes in the cryptocurrency network and the composite measure. Panel B of Table 2 presents the results using the network factors. The coin market returns positively correlate with all four of the individual cryptocurrency network factors and the composite measure. The coefficient on the wallet user growth measure is significant at the 10% level, and the three other coefficients are significant at the 1% level. The |$R^2$| s range from 5% for the wallet user growth measure to 30% for the active address growth measure. The |$R^2$| s using the composite measure is 19%. Consistent with the theoretical models, these results suggest that the network factors that measure the network effect of user adoptions are important drivers of cryptocurrency prices.

Moreover, in a dynamic cryptocurrency pricing model with the network effect, cryptocurrency prices not only reflect current cryptocurrency adoption but also contain information about expected future network growth—a key mechanism of Cong, Li, and Wang (2019) . We test this model implication by examining whether current coin market returns contain information about future cryptocurrency network growth. In particular, we predict cumulative future cryptocurrency adoption growth over different horizons using current coin market returns. We investigate cumulative future cryptocurrency adoption growth from one-month to eight-month horizons. We use cumulative wallet user growth, active address growth, transaction count growth, and payment count growth to capture cryptocurrency adoption growth.

Consistent with the prediction that cryptocurrency returns reflect expected future cryptocurrency adoptions, we find that coin market returns positively predict future cryptocurrency adoption growth as shown in Table 3 . Specifically, coin market returns positively and statistically significantly predict cumulative wallet user growth at all the horizons. Coin market returns positively and statistically significantly predict cumulative active address growth and cumulative payment count growth for the first three periods and two periods, respectively, and cease to be significant afterward. The coin market returns positively predict cumulative transaction count growth for the first five periods, but the predictability is not statistically significant. The only exception is the transaction growth measure: there is an insignificant, negative effect on transaction growth over the long horizons. A possible explanation for the negative effect is congestion, as it becomes very expensive to transact in Bitcoin when there is congestion, which deters many of the smaller transactions that would have occurred otherwise (e.g., Easley, O’Hara, and Basu 2019 ).

Predicting future network growth

This table reports the results of predicting cumulative future coin network growth with coin market returns. The network factors include wallet user growth, active address growth, transaction count growth, and payment count growth. Data are monthly. The t -statistics are reported in parentheses and are Newey-West adjusted with |$n-1$| lags. *, **, and *** denote significance levels at the 10%, 5%, and 1% levels. The data frequency is weekly.

2.2 Production factors

Several papers have argued that the costs of mining are essential for the infrastructure and security of cryptocurrencies (e.g., Sockin and Xiong 2019 ; Abadi and Brunnermeier 2018 ; Cong, He, and Li 2018 ). Notably, Sockin and Xiong (2019) show that, in a general equilibrium model with cryptocurrency production, the prices of the cryptocurrency are intimately linked to the marginal cost of mining.

We construct production factors of cryptocurrency to proxy for the cost of mining and test the relationship between these production factors and cryptocurrency prices. To the first approximation, mining a cryptocurrency requires two inputs: electricity and computer power. We separately construct proxies for electricity costs and computing costs. We first discuss our proxies for electricity costs. For electricity, we use seven primary measures. Three of the seven primary measures are U.S.-related: (i) average price of electricity in the United States, (ii) net generation of electricity of all sectors in the United States, and (iii) total electricity consumption of all sectors in the United States. The other four measures are China-related: (i) average price of electricity in China, (ii) electricity generation in China, (iii) average price of electricity in Sichuan province, and (iv) electricity generation in Sichuan province. We include the China proxies, because electricity supply is location specific and because China is considered to have the largest coin-mining operation among all countries. 7 We include Sichuan province proxies because Sichuan province hosts the largest mining farm in the world. Similarly, we also construct a composite measure as the first principal component of these seven primary measures. We denote the composite measure as |$PC^{elec}$|⁠ .

Panel A of Table 4 presents the correlation matrix of the electricity factors. Except for the two electricity price measures in China, the other five primary measures positively and strongly correlate with one another. Electricity prices in China are under strict government control. Unsurprisingly, they have low correlations with other electricity measures. The first principal component of the seven electricity factors strongly and positively correlates with most of the seven primary factors. The correlations are 0.76, 0.93, 0.88, 0.71, and 0.77 with the U.S. electricity price growth measure, the net U.S. generation growth measure, the U.S. electricity consumption growth measure, the China generation growth measure, and the Sichuan generation growth measure, respectively. The correlation between the first principal component and the China electricity price growth measure is –0.15, and the correlation between the first principal component and the Sichuan electricity price growth measure is 0.18. Panel B of Table 4 presents the electricity factor results for the coin market returns. Somewhat surprisingly, the coin market returns are not statistically significantly exposed to any of these production factor proxies. The |$R^2$| s of these regressions are low.

Cryptocurrency return loadings to electricity factors

This table reports the factor loadings of the coin market returns on the production factors that relate to electricity costs. Panel A shows the correlation matrix of the production factors. Panel B reports the factor loadings of the coin market returns on the production factors. Standard t -statistics are reported in parentheses. *, **, and *** denote significance levels at the 10%, 5%, and 1% levels based on the standard t -statistics. The data frequency is monthly.

For proxies of computing costs, we use as our primary measure the prices of Bitmain Antminer, a major piece of Bitcoin mining equipment. We also consider the excess stock returns of the companies that are major manufacturers of either GPU mining chips (Nvidia Corporation and Advanced Micro Devices, Inc.) or ASIC mining chips (Taiwan Semiconductor Manufacturing Company, Limited, and Advanced Semiconductor Engineering, Inc.). 8 We construct a composite measure as the first principal component of these five primary computing factors. We denote the composite measure as |$PC^{comp}$|⁠ .

Panel A of Table 5 presents the correlation matrix of the computing factors. Most of the pairs are positively correlated. The correlation between Antminer price growth and Nvidia return is –0.03, and the correlation between Antminer price growth and AMD return is –0.15. The first principal component is positively correlated with the four return measures and has a low correlation with the Antminer price growth measure. Panel B of Table 5 presents the computing factor results for the coin market returns. The coin market returns have insignificant loadings on the four excess return measures. The coin market returns have some loadings on the Antminer price growth measure, but they are only significant at the 10% level. The coin market returns are not significantly exposed to the first principal component.

Cryptocurrency return loadings to computing factors

This table reports the factor loadings of the coin market returns on the production factors that relate to computing costs. Panel A shows the correlation matrix of the production factors. Panel B reports the factor loadings of the coin market returns on the production factors. Standard t -statistics are reported in parentheses. *, **, and *** denote significance levels at the 10%, 5%, and 1% levels based on the standard t -statistics. The data frequency is monthly.

The model of Sockin and Xiong (2019) primarily concerns utility tokens. Therefore, we conduct our analyses on production factors on Bitcoin, Ethereum, and Ripple, respectively. Because Ethereum and Ripple are utility tokens, while Bitcoin is not, we expect to find that Ethereum and Ripple load significantly on the production factors. We show the results in the Internet Appendix . There is some evidence that Bitcoin returns are exposed to the Bitmain Antminer price growth, but Bitcoin and Ripple returns do not load significantly on these production factors. Overall, there is limited evidence that the computing factors are important drivers of cryptocurrency returns.

Lastly, we test the lead-lag effects between the changes in production factors and cryptocurrency returns to account for possible anticipation effects. We document the results in the Internet Appendix . We show that the one-month-ahead coin market returns are not significantly exposed to most of the production factors. The only exception is the changes in the average price of electricity in the United States, but the significant level is negative and only at the 10% level. However, we find that the current coin market returns positively predict some of the future production factors. In particular, the coin market returns positively and statistically significantly predict future changes in the average price of electricity in the United States, net generation of electricity of all sectors in the United States, total electricity consumption of all sectors in the United States, electricity generation in Sichuan province, and the first principal component of the production factors. Interestingly, we find that the results are stronger for the U.S.-based measures relative to the China-based measures. This is consistent with the fact that electricity prices and generation are heavily regulated in China. These results are consistent with a potential anticipation effect of production costs in the cryptocurrency market.

2.3 Are cryptocurrency returns predictable?

In this section, we test whether the coin market returns are predictable. The existing theoretical models of cryptocurrencies provide various predictions on the predictability of cryptocurrency returns. Schilling and Uhlig (2019) argue that the evolution of cryptocurrency prices should follow a martingale, and thus cryptocurrency returns are not predictable. Other papers predict that, in dynamic cryptocurrency valuation models, cryptocurrency returns could potentially be predicted by momentum, investor attention, and cryptocurrency valuation ratios (e.g., Cong, Li, and Wang 2019 ; Sockin and Xiong 2019 ). Motivated by the existing theoretical development and empirical findings in the financial markets, we test whether the cryptocurrency returns are predictable by momentum, investor attention, and proxies for cryptocurrency valuation ratios.

2.3.1 Cryptocurrency momentum

One of the most studied asset pricing regularities is momentum (e.g., Jegadeesh and Titman 1993 ; Moskowitz and Grinblatt 1999 ). As discussed in Cong, Li, and Wang (2019) , the network effect of user adoption generates a positive externality that is not immediately incorporated into cryptocurrency prices. This channel can potentially lead to a momentum effect in cryptocurrency returns. In their model, Sockin and Xiong (2019) generate momentum in the cryptocurrency market through investor attention—a mechanism similar to De Long et al. (1990) .

In this section, we start by establishing that there is strong evidence of time-series momentum at various time horizons. Panel A in Table 6 documents the time-series momentum results in the regression setting. Specifically, we regress cumulative future coin market returns on current coin market returns from the one-week to eight-week horizons. The current coin market returns positively and statistically significantly predict cumulative future coin market returns at all eight horizons. The results are significant at the 5% level for the one-week to five-week horizons and are significant at the 10% level from the six-week to eight-week horizons. For example, a one-standard-deviation increase in the current coin market return leads to increases in cumulative future coin market returns of 3.30%, 9 8.09%, 13.37%, and 17.66% increases at the one-week, two-week, three-week, and four-week horizons, respectively. Specifically, the one-week-ahead weekly return is that of buying the underlying coin market index at 11:59:59 UTD Sunday and selling the underlying coin market index at 11:59:59 UTD one week later. In the Internet Appendix , we also report the results based on noncumulative returns. The current coin market returns positively and significantly predict one-week- to five-week-ahead returns. The current coin market returns positively but insignificantly predict six-week- and seven-week-ahead returns. The current coin market returns negatively but insignificantly predict eight-week-ahead returns, suggesting some potential long-term reversal effect.

Time-series momentum

This table reports the time-series momentum results. Panel A shows the regression results, and panel B shows the results based on grouping weekly coin market returns into terciles. The first part of panel B reports results for the whole sample. The second part of panel B uses the first two years of data to determine the tercile cutoffs and examine the out-of-sample time-series momentum performance. The t -statistics are reported in parentheses and are Newey-West adjusted with |$n-1$| lags. *, **, and *** denote significance levels at the 10%, 5%, and 1% levels. The data frequency is weekly.

In the first part of panel B in Table 6 , we estimate the magnitude of the time-series momentum by grouping weekly returns into terciles and evaluating their performance going forward. We find that the top terciles outperform the bottom terciles at the one- to four-week horizons, consistent with the time-series regression results presented earlier. For example, at the one-week horizon, the average return of the top tercile is 8.01% per week with a t -statistic of 4.30, while the average return of the bottom tercile is 1.10% per week with a t -statistic of 0.92. The difference between the top and bottom terciles is 6.91% at the one-week horizon. At the two-week horizon, the average of the cumulative coin market returns of the top tercile is 16.22%, and that of the bottom tercile is only 3.59%. The difference between the top and bottom terciles is 12.63%. In the additional results section, we restrict our sample to 2014 onward. Again, we find a strong and significant momentum effect of somewhat smaller magnitude. 10

In the second part of panel B in Table 6 , we use the first two years of data to determine the tercile cutoffs and study the out-of-sample time-series momentum performance. We find a strong and significant momentum effect for the out-of-sample tests. For example, at the one-week horizon, the average return of the top tercile is 6.42%, and that of the bottom tercile is 0.80%. The difference between the top and bottom terciles is 5.62%, which is economically large and slightly smaller than the in-sample result of 6.91%.

Additionally, we test whether the time-series momentum effect is linked to network externalities, as suggested in Cong, Li, and Wang (2019) . In their dynamic cryptocurrency valuation model, the momentum effect is generated by the positive externality of the network effect that is not incorporated into cryptocurrency prices immediately. That is, their model implies that controlling for cryptocurrency adoption growth would subsume the time-series momentum effect. In Table 7 , we show that there is evidence that cryptocurrency adoption growth positively predicts future coin market returns. However, controlling for cryptocurrency adoption growth does not subsume the time-series momentum effect documented presented earlier.

Momentum and network effect

This table reports the results that compare coin market return predictability of momentum and network effect. The table reports the results of predicting cumulative future coin market returns with current coin market returns and each of the network factors. The Newey-West adjusted t -statistics with |$n-1$| lags are reported in parentheses. *, **, and *** denote significance levels at the 10%, 5%, and 1% levels. The data frequency is weekly.

2.3.2 Cryptocurrency investor attention

The theoretical literature of cryptocurrencies has also suggested that investor attention could potentially be linked to future cryptocurrency returns (e.g., Sockin and Xiong 2019 ). In this section, we investigate the role of investor attention in predicting cryptocurrency returns. Specifically, we construct the deviation of Google searches for the word “Bitcoin” in a given week compared with the average of those in the preceding four weeks. We standardize the Google search measure to have a mean of zero and a standard deviation of one. We use Google searches for the word “Bitcoin” to proxy for investor attention of the cryptocurrency market because Bitcoin is by far the largest and most visible cryptocurrency available. In panel A of Table 8 , we report the results of regressing cumulative future coin market returns from one-week to eight-week horizons on the Google search measure. The Google search measure statistically significantly predicts the one-week to six-week ahead cumulative coin market returns at the 5% level. The coefficient estimates of the seven-week and eight-week horizons are positive but are no longer statistically significant. A one-standard-deviation increase in searches leads to increases in weekly returns of about 3% for the one-week ahead cumulative coin market returns and about 5% for the two-week-ahead cumulative coin market returns. 11 In the Internet Appendix , we also report results based on noncumulative returns. The current coin market returns positively and significantly predict one-week- to four-week-ahead returns. The current coin market returns positively but insignificantly predict five-week-ahead returns. The current coin market returns negatively but insignificantly predict six-, seven-, and eight-week-ahead returns.

Google searches

This table reports the time-series Google search results. Panel A shows the regression results, and panel B shows the results based on grouping weekly coin market returns into terciles. The first part of panel B reports results for the whole sample. The second part of panel B uses the first two years of data to determine the tercile cutoffs and examine the out-of-sample time-series performance. The Google search measure is constructed as the Google search data for the word “Bitcoin” minus its average of the previous four weeks, and then normalized to have a mean of zero and a standard deviation of one. The t -statistics are reported in parentheses and are Newey-West adjusted with |$n-1$| lags. *, **, and *** denote significance levels at the 10%, 5%, and 1% levels. The data frequency is weekly.

In the first part of panel B in Table 8 , we investigate the return predictability of the Google search measures by grouping them into terciles and evaluating their performance going forward. Consistent with the regression results, we find that the top tercile outperforms the bottom tercile in terms of cumulative coin market returns at the one- to four-week-ahead horizons. For example, at the one-week horizon, the average return of the top tercile is 6.53% per week with a t -statistic of 3.82, while the average return of the bottom tercile is 0.43% per week with a t -statistic of 0.42. The difference between the top and bottom terciles is 6.09% at the one-week horizon. At the two-week horizon, the average of the cumulative coin market returns of the top tercile is 13.95% with a t -statistic of 4.89, and that of the bottom tercile is only 0.02% with a t -statistic of 0.01. The difference between the top and bottom terciles is 13.93%. In the additional results section, we restrict our sample to 2014 onward and find similar return predictive power of the Google search measures.

In the second part of panel B in Table 8 , we use the first two years of data to determine the tercile cutoffs and study the out-of-sample effect of investor attention, and we find a strong positive investor attention effect as well. For example, at the one-week horizon, the average return of the top tercile is 6.12%, and that of the bottom tercile is 0.70%. The difference between the top and the bottom terciles is 5.42%, which is economically large and slightly smaller than the in-sample estimate of 6.09%.

2.3.3 Negative investor attention

We have shown that unconditionally investor attention positively predicts cryptocurrency returns. However, not all investor attention is positive. For example, in their model, Sockin and Xiong (2019) differentiate positive investor attention and negative investor attention, and show that negative investor attention is followed by cryptocurrency price depreciation in the future.

In this section, we investigate whether negative investor attention predicts cryptocurrency returns. We construct a ratio between Google searches for the phrase “Bitcoin hack” and searches for the word “Bitcoin” to proxy for negative investor attention. We standardize the measure to have a mean of zero and a standard deviation of one. Panel A of Table 9 shows the results of the predictive regressions. The ratio negatively and significantly predicts one- to six-week-ahead cumulative coin market returns. For example, a one-standard-deviation increase in the ratio leads to a 2% decrease of coin market returns in the next week. Panel B of Table 9 reports the in-sample and out-of-sample return predictability of the negative investor attention measures by grouping them into terciles and evaluating their performance going forward. Consistent with the regression results, we find strong negative return predictability results of the negative investor attention measures.

Bitcoin hack

This table reports the time-series Bitcoin hack results. Panel A reports the regression results, and panel B reports the sorting results. The Bitcoin hack measure is constructed as the ratio between Google searches for the phrase “Bitcoin hack” and searches for the word “Bitcoin,” and then normalized to have a mean of zero and a standard deviation of one. Results are based on weekly data. The t -statistics are reported in parentheses and are Newey-West adjusted with |$n-1$| lags. *, **, and *** denote significance levels at the 10%, 5%, and 1% levels. The data frequency is weekly.

Another way to see the results on the investor attention is that our measures of investor attentions proxy for speculative interest and sentiment in cryptocurrencies. Positive investor sentiment is followed by cryptocurrency price appreciation, and negative investor sentiment is followed by depreciation. We further investigate these issues in Section 4 .

2.3.4 Interaction between momentum and attention

We have shown that there are strong effects of time-series momentum and investor attention in the cryptocurrency market. The equity market research (e.g., Hong, Lim, and Stein 2000 ; Hou, Xiong, and Peng 2009 ) shows that there is a strong relationship between momentum and investor attention. It is possible that these two results capture the same underlying phenomenon. For example, Sockin and Xiong (2019) propose a potential channel to generate momentum. In their model, momentum arises because users have incorrect expectations about future prices—a mechanism similar to De Long et al. (1990) . Their model suggests that cryptocurrency momentum and investor attention could potentially arise from the same underlying mechanism. The cryptocurrency momentum and investor attention results could also interact with each other. For example, the cryptocurrency time-series momentum effect may be weaker at times of high investor attention, because there is little information leakage at times of high investor attention.

First, we show that the current investor attention of cryptocurrencies is indeed associated with current and past coin market performance. We regress the current deviation in the Google searches on the contemporaneous and the coin market returns of the previous four weeks. Table 10 documents the results. We find that the deviations in Google searches are positively and significantly associated with contemporaneous and the previous week’s coin market returns. The Google search measures do not significantly correlate with past coin market returns beyond one week. Intuitively, these results suggest that investor attention is elevated after superior cryptocurrency market performance.

Google searches and past returns

This table reports the relationships between the Google search measure and past coin market returns. The Google search measure is constructed as the Google search data for the word “Bitcoin” minus its average of the previous four weeks, and then normalized to have a mean of zero and a standard deviation of one. The standard t -statistic is reported in parentheses, and the bootstrapped t -statistic is reported in brackets. *, **, and *** denote significance levels at the 10%, 5%, and 1% levels based on the standard t -statistics. The data frequency is weekly.

We further test the interaction between the time-series momentum and the investor attention phenomena. The results are reported in Table 11 . In the first test of Table 11 , we regress cumulative future coin market returns on current coin market returns and Google search measures. We find that the coefficients to the current coin market returns are significant for all the horizons, and the coefficients to the Google search measures are significant from the one-week to the five-week horizons. The magnitudes of the coefficients are similar to the standalone estimates. For example, the one-week-ahead coefficients under the univariate regressions are 0.20 and 0.03 for the current coin market returns and the Google search measures, respectively, while they are 0.18 and 0.03 under the bivariate regressions. These results show that the time-series momentum and the investor attention results do not subsume each other.

Interaction between momentum and attention

This table reports the predictive regressions of future cumulative coin market returns on momentum, attention, and the interaction of the two. The indicator variable |$1_{\{Google>0\}}$| equals one if the current Google search measure is above the sample mean and zero otherwise. The indicator variable |$1_{\{R>0\}}$| equals one if the current coin market return is positive and zero otherwise. Results are based on weekly returns. The Newey-West adjusted t -statistics with |$n-1$| lags are reported in parentheses. *, **, and *** denote significance levels at the 10%, 5%, and 1% levels. The data frequency is weekly.

In the second test of Table 11 , we test the performance of the time-series momentum result when investor attention is high. We construct an indicator variable, |$1_{\{Google>0\}}$|⁠ , that equals one if the current Google search measure is above the sample mean and zero otherwise. We regress the cumulative future coin market returns from one-week to eight-week horizons to the current coin market return, the indicator variable, and the interaction term. The interaction term is not significant at any of the eight horizons, suggesting that the magnitude of the time-series momentum effect is similar for high and low investor attention periods. In the third test of Table 11 , we test the performance of the investor attention result when the current coin market return is high. We construct an indicator variable, |$1_{\{R>0\}}$|⁠ , that equals one if the current coin market return is positive and zero otherwise. We regress the cumulative future coin market returns from one-week to eight-week horizons to current Google search measure, the indicator variable, and the interaction term. The interaction term is not significant at any of the eight horizons, suggesting that the magnitude of the investor attention effect is similar for high and low coin market return periods.

Furthermore, we study the cross-section of time-series momentum for high- and low-attention coins. We collect Google attention data for the 10 largest cryptocurrencies from the beginning of 2014 to the end of 2018. The sample period is shorter for this analysis, because before 2014, there are very few cryptocurrencies and the data for alternative coins are hard to get. The list of coins is Bitcoin, Ethereum, Ripple, Litecoin, Tether, Bitcoin-Cash, Tezos, Binance-coin, Monero, and Cardano. At a given point in time, we group the existing coins into two subsamples based on the Google attention data—a group of high-attention coins and a group of low-attention coins. We construct the value-weighted returns of the high-attention group and the low-attention group, separately, and test the time-series momentum strategy effect in each subgroup.

We regress the future cumulative returns on current returns for each of the subsamples and report the results in the Internet Appendix . We find that in this sample, the time-series momentum effect is stronger for the relatively low-attention coins. In particular, the coefficient estimates for both the high-attention and the low-attention subgroups are positive, suggesting that there are time-series momentum effects for both groups. However, the coefficient estimates for the high-attention subgroup is not statistically significant, while the coefficient estimates for the low-attention subgroup is statistically significant up to six weeks out. The magnitudes of the coefficient estimates are also much larger for the low-attention subgroup relative to the high-attention subgroup. The results are consistent with the “underreaction” mechanism of momentum.

2.3.5 Cryptocurrency valuation ratio

Additionally, we test whether the cryptocurrency valuation ratios similar to those in the financial markets can predict future coin market returns. In the equity market, the fundamental-to-market ratios are commonly referred to as valuation ratios and are measured as the ratio of the book value to the market value of equity or some other fundamental value to market value (e.g., dividend-to-price; earnings-to-price). Another value measure used in the literature that has been shown to correlate highly with fundamental-to-market value is the negative of the long-term cumulative past returns (e.g., De Bondt and Thaler 1985 , 1987 ; Fama and French 1996 ; and Moskowitz 2015 ). It is more difficult to define a similar measure of fundamental value for cryptocurrency. However, in their dynamic cryptocurrency asset pricing model, Cong, Li, and Wang (2019) argue that the cryptocurrency fundamental-to-value ratio can be defined as the number of user adoptions over market capitalization, which negatively predicts future cryptocurrency returns.

The market value of cryptocurrency is readily available. However, there is no direct measure of fundamental value for the cryptocurrencies. In its essence, value is a measure of the gap between the market value and the fundamental value of an asset. Because of the lack of a standard “book” value measure of the cryptocurrency market, we use an array of different proxies to capture the idea of fundamental value. We proxy the fundamental-to-market ratio by a number of value measures motivated by the finance literature. The first one is the long-term past performance measure: the negative of the past 100-week cumulative coin market return. The other four measures aim to proxy the cryptocurrency fundamental-to-market value directly: the user-to-market ratio, the address-to-market ratio, the transaction-to-market ratio, and the payment-to-market ratio. The idea of these four measures is to use some measures of the “book” value of the underlying cryptocurrency market and scale by the current market capitalization. The user base of the cryptocurrency market seems to capture the concept of “book” value in the financial markets. This is consistent with the theoretical literature of the cryptocurrency market that emphasizes the notion of the network effect, which can be proxied by the current user base of the cryptocurrencies. On the other hand, the market value of the cryptocurrency provides a market assessment of the current value of the complete cryptocurrency infrastructure. Therefore, the user base-to-market value measure can capture the notion of fundamental-to-market ratio in the financial markets. In panel A of Table 12 , we report the correlations across the different valuation ratios in the cryptocurrency market. The five primary measures are highly correlated with one another, with correlations ranging from 0.73 to 0.91. The first principal component measure for the five fundamental-to-market ratios has correlations of 0.91, 0.91, 0.96, 0.93, and 0.93 with the long-term past returns, the wallet user-to-market ratio, the active address-to-market ratio, the transaction-to-market ratio, and the payment-to-market ratio, respectively.

Cryptocurrency valuation ratio

This table reports the predictive regressions of coin market returns on proxies for cryptocurrency market fundamental-to-market ratio. The proxies for cryptocurrency valuation ratio include the (negative) past 100-week cumulative coin market returns, the user-to-market ratio, the address-to-market ratio, the transaction-to-market ratio, payment-to-market ratio, and the first principal component of the previous five proxies. The ratios are estimated using the cointegration method. Results are based on weekly returns. The Newey-West adjusted t -statistics with |$n-1$| lags are reported in parentheses. *, **, and *** denote significance levels at the 10%, 5%, and 1% levels. The data frequency is weekly.

We regress the coin market returns on the lagged cryptocurrency fundamental-to-market ratios, and the results are reported in panel B of Table 12 . We document the regression results from one-week to eight-week horizons. Although the coefficient estimates are consistently negative, none of the five standalone fundamental-to-market ratios predict future coin market returns significantly over any horizon. The principal component measure also fails to predict future coin market returns over these horizons. Overall, there is a very weak relationship between the future coin market returns and the current cryptocurrency fundamental-to-value ratio.

Both the cryptocurrency literature and the community have debated the nature of cryptocurrencies. For example, Schilling and Uhlig (2019) show that, in an endowment economy with both fiat money and cryptocurrency, the evolution of cryptocurrency prices is linked to that of the fiat money. Athey et al. (2016) emphasize the importance of fiat money risks of cryptocurrencies. The cryptocurrency community has proposed that cryptocurrencies are “digital gold” and serve the purpose of the traditional precious metal commodity. Moreover, Schilling and Uhlig (2019) argue that cryptocurrency returns can have exposure to macroeconomic risks such as monetary policies. In this section, we evaluate these claims by examining the relationship between cryptocurrency returns and traditional asset returns such as currency, commodity, and equity.

3.1 Currency and commodity factor loadings

In an endowment economy where fiat money and cryptocurrency coexist and compete with each other, Schilling and Uhlig (2019) show that the evolution of cryptocurrency prices is correlated with the that of the fiat money prices. Athey et al. (2016) also emphasize the importance of fiat money risks of cryptocurrencies. We test this prediction by investigating the cryptocurrency exposures to traditional currencies. Columns (1) to (6) of Table 13 show the coin market returns’ exposures to the traditional currency returns. For currency returns, we consider five major currencies: Australian dollar, Canadian dollar, euro, Singaporean dollar, and U.K. pound. The exposures of the coin market returns to these major currencies are not statistically significant, and the alpha estimates barely change. We further test cryptocurrency exposures on currency factors as in Lustig, Roussanov, and Verdelhan (2011) instead of individual major currency returns. 12 Columns (7) to (9) of Table 13 report the coin market returns’ exposures to these currency factors. Consistent with the results on individual currency returns, we find that the coin market returns do not have significant exposures to the currency factors. We conclude that there is no consistent evidence of systematic currency exposures in cryptocurrencies.

Currency loadings of coin market returns

This table reports the factor loadings of the coin market returns on returns of different currencies and currency factors. The currencies include Australian dollar, Canadian dollar, euro, Singapore dollar, and U.K. pound. The currency factors are based on Lustig, Roussanov, and Verdelhan (2011) . The returns are in percentage. The results are based on monthly returns. The standard t -statistic is reported in parentheses, and the bootstrapped t -statistic is reproted in brackets. *, **, and *** denote significance levels at the 10%, 5%, and 1% levels based on the standard t -statistics. The data frequency is monthly.

Another popular narrative around cryptocurrencies is that cryptocurrencies serve the same purpose as traditional precious metal commodities. That is, cryptocurrencies are “digital gold.” If the investors of cryptocurrencies hold this belief, we would expect to find that the returns of cryptocurrencies comove with the returns of the traditional precious metal commodities. We test the precious metal commodity exposures of the coin market returns and report the results in Table 14 . For precious metal commodities, we consider gold, platinum, and silver. The exposures of the coin market return to these three major commodities are not statistically significant. Overall, we conclude that there is no consistent evidence of systematic precious metal commodity exposures in cryptocurrencies.

Commodity loadings of coin market returns

This table reports the factor loadings of the coin market returns on returns of different precious metal commodities. The commodities include gold, platinum, and silver. Returns are in percentage. The standard t -statistic is reported in parentheses and the bootstrapped t -statistic is reproted in brackets. *, **, and *** denote significance levels at the 10%, 5%, and 1% levels based on the standard t -statistics. The data frequency is monthly.

3.2 Equity factor loadings

We document the common stock factor exposures of the coin market returns in the Internet Appendix . For the equity risk factors, we choose the Capital Asset Pricing Model (CAPM), Fama-French three-factor, Carhart four-factor, Fama-French five-factor, and Fama-French six-factor models. 13 The alphas for all of the considered models are statistically significant. The average return of the period is 20.44% per month. The CAPM-adjusted alpha decreases to 17.53% per month—a reduction of about 14%. The CAPM beta is large at 3.15 but not statistically significant. The betas are statistically significant at the 10% level only for the five-factor and six-factor models. The corresponding alphas are 15.32% and 15.00% per month for the five-factor model and six-factor model, respectively. The exposures to the other factors are not statistically significant. The exposures to the SMB (small-minus-big) factor are negative but not stable across the specifications: the magnitude of the coefficient decreases when five-factor and six-factor models are considered. The exposures to the HML (high-minus-low) factor are negative and have consistent magnitudes and signs; this suggests that the coin market returns may comove more with growth rather than with value firms. The exposures to the RMW (robust-minus-weak) factor are positive and are estimated slightly more accurately than other statistically not significant factors; this suggests that the coin market returns comove more with high-profit rather than low-profit firms. The point estimates on the MOM (momentum) and CMA (conservative-minus-aggressive) factors are very inaccurate. 14

3.2.1 Exploring the factor zoo

The finance literature has documented more than a hundred factors for predicting the cross-section of stock returns (see e.g., summarizes in Feng, Giglio, and Xiu 2017 and Chen and Velikov 2017 ). To investigate whether any of those factors may be important in pricing cryptocurrencies, we estimate the loadings of the 155 common factors from Andrew Chen’s website. One caveat is that this data set ends at the end of 2016 and thus does not cover the most recent return experiences. We report the results in the Internet Appendix due to the large number of factors. We find that only four out of the 155 factors are significant, but those four factors do not form any discernible patterns.

3.3 Macroeconomic factors

We further examine the macroeconomic factor exposures of the coin market returns. For macroeconomic factors, we consider the nondurable consumption growth, durable consumption growth, industrial production growth, and personal income growth. We document the results in the Internet Appendix . We find that the coin market returns do not significantly load on these macroeconomic factors. We further investigate the three major cryptocurrencies individually. For Bitcoin and Ripple, all of the exposures are not statistically significant. For Ethereum, notably, the durable consumption growth factor has a significant loading.

4.1 Short sample

We have eight years of coin market return data spanning from the beginning of 2011 to the end of 2018. The short sample is a potential barrier to study cryptocurrency that we cannot avoid. Moreover, there is a great deal of uncertainty and learning about cryptocurrencies during the period. As argued by Pástor and Veronesi (2003) , it takes time for investors to fully learn and understand emerging technologies, which can lead to price bubbles.

One approach we take to partially address these concerns is to break the sample into two halves and check whether our results are stable for these subsamples. During the first half of the sample, there are considerably more uncertainty and learning about cryptocurrency as an asset class. We document these results in the Internet Appendix . We find that the directions of all of the results are the same for the first and second halves of the sample. The magnitudes of the results are also comparable between the two subsamples. There is potentially still a lot of uncertainty and learning about cryptocurrencies today, but the assumption we need for the subsample tests is relatively mild: the uncertainty has decreased from the first half of the sample to the second half of the sample. The analysis on the volatility of the coin market returns also supports this assumption. We find that the standard deviation of coin market returns decreased significantly from the first half to the second half of the sample period. The figure in the Internet Appendix shows a significant decrease in the volatility of the coin market returns over time.

4.2 Time-series momentum and cross-sectional momentum

We study the relationship between time-series momentum and cross-sectional momentum. It is difficult to directly compare the time-series momentum and cross-section momentum results. The time-series momentum is a phenomenon on the aggregate coin market returns, while the cross-sectional momentum results are neutral in terms of the aggregate performance of the coin market. We use two different methods to test the relationships between the time-series momentum and the cross-sectional momentum results. In the first method, we use coin market returns to predict the cross-sectional cryptocurrency momentum. This approach gives us a sense about whether the cross-sectional momentum effect is stronger when the time-series momentum is on a positive trajectory. We report the results in the Internet Appendix . The coin market returns do not significantly predict future cumulative cross-sectional momentum returns. This result suggests that the profitable periods of the cryptocurrency time-series momentum and cross-sectional momentum are different.

In the second method, we follow the approach similar to Moskowitz, Ooi, and Pedersen (2012) and construct a portfolio version of the time-series momentum. For our set of instruments, we use one of the following: largest three coins, largest five coins, and largest ten coins. For each instrument and month, we consider whether the excess return over the past three weeks is positive or negative and go long the instrument if positive and short if negative. We hold the position for one week, so there is no overlapping sample. The unadjusted excess returns are positive and significant at the 1% level for all three of the specifications. The economic magnitudes of the excess returns are large, ranging from 3.17% for the top three coins to 4.62% for the top ten coins. Controlling for the coin market returns, the economic magnitudes of the excess returns barely change. Controlling for the cryptocurrency cross-sectional momentum as constructed in Liu, Tsyvinski, and Wu (2019) , the magnitudes of the spreads decrease but remain highly statistically significant. It is not surprising that the magnitudes of the spreads decrease after controlling for the cross-sectional momentum because the excess returns of the strategy contain information about the cross-sectional momentum by construction. However, there is additional information coming from the construction similar to Moskowitz, Ooi, and Pedersen (2012) , which is evidenced by the fact that the magnitudes of the spreads remain positive and statistically significant after controlling for the cross-sectional momentum. Finally, we also test whether the strategies contain information above and beyond the cryptocurrency three-factor model in Liu, Tsyvinski, and Wu (2019) . After controlling for the three-factor model, the magnitudes of the spreads further decrease but still remain positive and significant at the 5% level for the top five and top ten coins and at the 10% level for the top three coins.

4.3 Regulations

A potentially important determinant of cryptocurrency valuation is regulations. To test whether cryptocurrency regulations are important determinants of cryptocurrency valuations, we follow the method of Auer and Claessens (2018) and Shanaev et al. (2019) and determine 120 regulative events. We further categorize these regulative events into positive and negative events based on Auer and Claessens (2018) . We document the list of regulative events and the results in the Internet Appendix . We find that the contemporaneous cryptocurrency returns are lower during the days of regulative events. However, we find that the cryptocurrency returns respond to negative regulative events but not to positive regulative events.

4.4 Speculative interest and sentiment

In this section, we test whether speculation and investor sentiment may be important drivers of cryptocurrency prices. We extract the speculative shares of cryptocurrency usage from Coindesk.com. We test whether the cryptocurrency returns strongly respond to the contemporaneous and expectations of future speculative share growth. We further control for the network growth rates as discussed earlier to examine whether the results of network effects are driven by variations in speculative interests. We document the results on speculative interests in the Internet Appendix . We find some evidence that the cryptocurrency returns positively load on the contemporaneous speculative share growth, but the coefficient estimates are not significant. In the bivariate regressions, we show that the loadings of the cryptocurrency returns to the contemporaneous network growth remain positive and statistically significant. Furthermore, we show that the coin market returns positively forecast future speculative share growth. The coefficients are significant at the three-month and eight-month horizons. These results show that current coin market returns also contain information about expectations of future speculative share growth.

To test the effect of sentiment, we construct a measure that is directly aimed to capture investor sentiment. 15 The measure of cryptocurrency sentiment is defined as the log ratio between the count of positive and the count of negative phrases of cryptocurrencies in Google searches. The positive and negative phrases are described in the corresponding table in the Internet Appendix . Therefore, when the measure is high, investor sentiment is more positive and vice versa. We test whether the sentiment measure predicts future cryptocurrency returns, and compare that to the measures of investor attention and momentum. We find that the sentiment measure positively and significantly predicts future cryptocurrency returns. However, this result is distinct from the investor attention and cryptocurrency momentum results. All three variables are statistically significant in predicting future cryptocurrency returns in the multivariate regressions.

4.5 Beauty contests

One potential theoretical explanation of high volatility in financial markets may be that they are represented by the Keynesian beauty contest model. In this section, we aim to test the role of beauty contests in the cryptocurrency markets. We need a time-varying measure of the degrees of disagreement among the cryptocurrency investors. Ideally, we would like to have the expectations of individual cryptocurrency investors. Practically, this is not feasible due to data limitation at this time. We take a different route and measure the dispersion of investor expectations using the ratio between cryptocurrency volume and return volatility. This choice is motivated by Biais and Bossaerts (1998) , who show theoretically that the volume-volatility ratio summarizes the degree of disagreement among the investors and discriminates between genuine disagreement and mere Bayesian learning with agreeing agents. We test empirically whether the coin market returns respond to the volume-volatility ratio contemporaneously and whether the volume-volatility ratio predicts future coin market returns. We summarize these results in the Internet Appendix . We find that the coin market return is higher when the current volume-volatility ratio is higher. This result is consistent with the idea that investors tend to bid the price up when there is a lot of disagreement in the cryptocurrency market. The flip side is that the volume-volatility ratio does not predict future cumulative coin market returns over any horizon.

4.6 VAR analysis

One concern about the network factor analysis is that the contemporaneous correlations between the network size/activity factors and coin market returns might be mechanical and not truly capture the value of network externalities. To address the concern, we conduct a bivariate VAR analysis with the coin market returns and different coin network growth measures. The results are documented in the Internet Appendix .

To differentiate the network effects from the potential mechanical effects, we examine how changes in the network factors affect valuations in the future—that is, what is the cumulative permanent return associated with a shock to network size/activity factors. The bottom-left graph of each panel shows the associated impulse response function. We find that a shock to the wallet user growth, active address growth, and transaction count growth factors positively predict the coin market returns in the future and that there is not any reversal effect. The effects tend to concentrate on the first couple of weeks. In particular, the wallet user growth and active address growth factors positively and statistically significantly predict coin market returns in the future. The payment count growth measure is the only exception, but the point estimate is not significant. In terms of the cumulative effect, the cumulative return responses to one standard deviation of network factor shock are about 4% based on the changes in wallet user growth, about 2% based on the active address growth, and about 0.45% based on the transaction count growth.

Moreover, there is a bidirectional relationship between the network factors and the cryptocurrency returns. Consistent with the results in Table 3 in the paper, the VAR shows that the coin market returns positively and significantly predict future network growth based on all four specifications. The top-right graph of each panel shows the associated impulse response function. In the bivariate VAR framework, the approach accounts for this bidirectional effect, that returns may affect trading decisions and therefore affect the network size. Additionally, the bivariate VAR again reveals a coin market momentum effect in all four specifications. The top-left graph of each panel shows the associated impulse response function. The VAR approach also helps account for the momentum effect in the cumulative effect of network size/activity factors on the valuation of cryptocurrencies. Overall, the VAR suggests some evidence that a positive shock to the network size/activity factors leads to a permanent increase in the valuations of the coin market in the future. In terms of the timing of the effects, the impulse response functions suggest that it can take a few weeks for the impulse responses to decay to zero.

4.7 Additional production factor test

When the electricity price increases, the return to mining should decrease. However, the reduction in the return to mining would force some miners to exit, which leads to a higher probability of any given miner receiving the reward plus fees and a reduction of the difficulty of the cryptographic puzzles. These two effects would endogenously restore the profitability of the mining. Therefore, the shocks to electricity prices or computing power do not necessarily affect the marginal cost of mining because a change in these costs would affect the profitability of miners, causing an adjustment in the number of miners and therefore an adjustment in the required computing effort to restore profitability (e.g., Easley, O’Hara, and Basu 2019 ).

To address this effect, we include another set of tests. We regress the coin market returns on the number of Bitcoins given as a block reward, controlling for the price of Bitcoins and the fees paid. The rationale of the test is the following: the price of Bitcoin is endogenous, the fees paid could be endogenous as they are driven by network usage, but the number of Bitcoins given as a block reward is exogenous and deterministically changes through time according to the Bitcoin protocol. Therefore, the exogenous variation, controlling for the price of Bitcoin and the fees, could be used to identify the effects of mining costs.

Column (1) of the table documents the baseline specification. The coefficient of interest is in the changes of the number of Bitcoins given as a block reward, or |$\Delta Gen\ Coin$|⁠ . We find that, although the coefficient estimate is positive, it is not statistically significant. Column (2) uses next-month changes of the number of Bitcoins given as a block reward to test any anticipation effect. The point estimate is again positive but not statistically significant. Column (3) includes both the current and next-month changes in the number of Bitcoins given as a block reward as the independent variables, and both coefficient estimates are not significant. Columns (4) to (6) repeat the exercises of columns (1) to (3) but include the first principal component of the production factor, and the results are similar. Columns (7) to (12) repeat the exercises but control for the level of fees instead of changes in fees. We find that the coefficient estimates of |$\Delta Gen\ Coin$| and |$\Delta Gen\ Coin_{+1}$| are not statistically significant, consistent with results in columns (1) to (6).

Easley, O’Hara, and Basu (2019) argue that fee is not the only endogenous variable in the mining process, and that the confirmation time is also endogenously determined as a function of miner competition. Therefore, we also control for the confirmation time in our analysis and examine the results. Columns (13) to (24) report the results controlling for the confirmation time. In particular, columns (13) to (18) control for the changes of confirmation time, and columns (19) to (24) control for the level of confirmation time. We find that the coefficient estimates of |$\Delta Gen\ Coin$| and |$\Delta Gen\ Coin_{+1}$| remain statistically insignificant. Overall, we do not find a significant effect of the exogenous variation in the number of Bitcoins given as a block reward, controlling for the price of Bitcoin and the fees, on the cryptocurrency valuations.

4.8 Subsample by cryptocurrency characteristics

In this section, we consider a number of cryptocurrency characteristics: (i) whether the cryptocurrency is based on Proof-of-Work (PoW) or Proof-of-Stake (PoS), (ii) whether the cryptocurrency is minable, (iii) whether the cryptocurrency is built on an Ethereum blockchain, (iv) whether the cryptocurrency is a stable coin, and (v) whether the cryptocurrency is a smart contract. Based on each characteristic, we form a value-weighted portfolio of all the underlying cryptocurrencies. In the untabulated results, we look at the loadings of returns for each subgroup on the network factors, production factors, currency factors, commodity factors, equity factors, and macroeconomic factors.

First, we examine the loadings on the network factors. We find that the returns for the subgroups generally load positively on the network factors. Based on the first principal component of the four primary measures, we show that the returns of all six subgroups load positively on the network factors. In particular, the returns of PoW, minable coins, Ethereum blockchain coins, and stable coins positively and statistically significantly expose to the first principal component. On the other hand, the returns of PoS and smart contract coins do not statistically significantly expose to the first principal component. Then, we turn to the loadings on the production factors. We find that the returns for most of the subgroups do not significantly load on the production factors. We conclude that the factor exposures of the subgroup returns are largely consistent with the aggregate coin market returns. Lastly, we turn to the loadings of the returns for subgroups on the currency, commodity, equity, and macroeconomic factors. In general, we find that the returns of the subgroups have low exposures to these factor models. We conclude that the factor exposures of the subgroup returns are largely consistent with the aggregate coin market returns.

We find that cryptocurrency returns strongly respond to cryptocurrency network factors, as suggested by the theoretical literature. However, our empirical results do not support the notion that the evolution of cryptocurrency prices is linked to cryptocurrency production factors. At the same time, the returns of cryptocurrency can be predicted by two factors specific to its markets: momentum and investor attention. In contrast to the equity market, we show that the momentum result and the investor attention result are distinct phenomena and that there is only limited interaction between them. Moreover, cryptocurrency returns have low exposures to traditional asset classes such as currencies, commodities, and stocks, and to macroeconomic factors.

Authors have furnished an Internet Appendix, which is available on the Oxford University Press Web site next to the link to the final published paper online.

We thank Andrew Atkeson, Nicola Borri, Eduardo Davila, Stefano Giglio, William Goetzmann, Andrew Karolyi, Ye Li, Stephen Roach, Robert Shiller, Michael Sockin, and Jessica Wachter for their comments. Colton Conley provided outstanding research assistance. Supplementary data can be found on The Review of Financial Studies web site.

1 See, e.g., Cong, Li, and Wang (2019) , Pagnotta and Buraschi (2018) , and Biais et al. (2018) .

2 See, e.g., Cong, He, and Li (2018) and Sockin and Xiong (2019) .

3 Some coins are not tracked by the website because the coins’ exchanges do not provide accessible APIs.

4 We thank William Goetzmann for kindly sharing the Twitter post count data with us.

5 One of the 156 anomalies does not exist during the sample period. The database ends at December 2016.

6 Because Bitcoin is by far the largest and well-known cryptocurrency available, we use Bitcoin network data. The tests in the paper use coin market returns; in the Internet Appendix , we show that our results are qualitatively similar using Bitcoin returns.

7 See Jasper Pickering and Fraser Moore, “How China Become a Haven for People Looking to Cash in on the Bitcoin Gold Rush,” Business Insider, December 12, 2017.

8 See Shanthi Rexaline, “The Companies Behind the Chips That Power Cryptocurrency Minning,” Benzinga, February 2, 2018.

9 The 3.30% weekly return is calculated by multiplying a one-standard-deviation increase of coin market returns (16.50%) and the coefficient estimate (0.20).

10 Stoffels (2017) documents that a cross-sectional momentum strategy based on 15 cryptocurrencies generates abnormal returns during the period between 2016 and 2017.

11 Wang and Vergne (2017) use the level of newspaper mentions of Bitcoin to proxy for the “buzz” of Bitcoin. They document that high “buzz” predicts low Bitcoin returns in the future. Mai et al. (2016) use the level of Twitter post counts to predict Bitcoin returns.

12 We thank Nicola Borri for providing us with the up-to-date currency factors.

13 The Fama-French three-factor model is based on Fama and French (1993) and Fama and French (1996) . The Carhart four-factor model is based on Carhart (1997) . The Fama-French five-factor model is based on Fama and French (2016) . The Fama-French six-factor model is based on Fama and French (2017) .

14 Stoffels (2017) and Gilbert and Loi (2018) examine cryptocurrency loadings on the CAPM and Fama-French three-factor models.

15 Chen et al. (2019) classify cryptocurrency-related positive and negative words of StockTwits and Reddit.

Abadi, J. , and Brunnermeier M. . 2018 . Blockchain economics . Working Paper , Princeton University .

Google Scholar

Google Preview

Asness, C. , Moskowitz T. J. , and Pedersen L. H. . 2013 . Value and momentum everywhere . Journal of Finance 68 : 929 – 985 .

Athey, S. , Parashkevov I. , Sarukkai V. , and Xia J. . 2016 . Bitcoin pricing, adoption, and usage: Theory and evidence . Working Paper , Stanford University .

Auer, R. , and Claessens S. . 2018 . Regulating cryptocurrencies: Assessing market reactions . BIS Quarterly Review 51 – 65 .

Biais, B. , Bisiere C. , Bouvard M. , Casamatta C. , Menkveld A. J. . 2018 . Equilibrium bitcoin pricing . Working Paper , Toulouse School of Economics .

Biais, B. , and Bossaerts P. . 1998 . Asset prices and trading volume in a beauty contest . Review of Economic Studies 65 : 307 – 340 .

Bianchi, D. 2017 . Cryptocurrencies as an asset class: An empirical assessment . Working Paper , Warwick University .

Borri, N. 2019 . Conditional tail-risk in cryptocurrency markets . Journal of Empirical Finance 50 : 1 – 19 .

Borri, N. , and Shakhnov K. . 2018 . Cryptomarket discounts . Working Paper , LUISS University .

Carhart, M. M. 1997 . On persistence in mutual fund performance . Journal of Finance 52 : 57 – 87 .

Chen, A. , and Velikov M. . 2017 . Accounting for the anomaly zoo: A trading cost perspective . Working Paper , Penn State University .

Chen, C. Y. , Despres R. , Guo L. , and Renault T. . 2019 . What makes cryptocurrencies special? Investor sentiment and return predictability during the bubble . Working Paper .

Chiu, J. , and Koeppl T. V. . 2017 . The economics of cryptocurrencies–bitcoin and beyond . Working Paper , Queen’s University .

Cong, L. W. , and He Z. . 2019 . Blockchain disruption and smart contracts . Review of Financial Studies 32 : 1754 – 1797 .

Cong, L. W. , He Z. , and Li J. . 2018 . Decentralized mining in centralized pools . Working Paper , University of Chicago .

Cong, L. W. , Li Y. , and Wang N. . 2019 . Tokenomics: Dynamic adoption and valuation . Working Paper , Columbia University .

Corbet, S. , Lucey B. M. , Urquhart A. , and Yarovaya L. . 2019 . Cryptocurrencies as a financial asset: A systematic analysis . International Review of Financial Analysis 62 : 182 – 199 .

Da, Z. , Engelberg J. , and Gao P. . 2011 . In search of attention . Journal of Finance 66 : 1461 – 1499 .

Daniel, K. , and Moskowitz T. J. . 2016 . Momentum crashes . Journal of Financial Economics 122 : 221 – 247 .

De Long, J. B. , Shleifer A. , Summers L. H. , and Waldmann R. J. . 1990 . Noise trader risk in financial markets . Journal of Political Economy 98 : 703 – 738 .

De Bondt, W. F. , and Thaler R. . 1985 . Does the stock market overreact? Journal of Finance 40 : 793 – 805 .

De Bondt, W. F. , and Thaler R. . 1987 . Further evidence on investor overreaction and stock market seasonality . The Journal of Finance 42 : 557 – 581 .

Easley, D. , O’Hara M. , and Basu S. . 2019 . From mining to markets: The evolution of bitcoin transaction fees . Journal of Financial Economics 134 : 91 – 109 .

Fama, E. F. , and French K. R. . 1993 . Common risk factors in the returns on stocks and bonds . Journal of Financial Economics 33 : 3 – 56 .

Fama, E. F. , and French K. R. . 1996 . Multifactor explanations of asset pricing anomalies . Journal of Finance 51 : 55 – 84 .

Fama, E. F. , and French K. R. . 2016 . Dissecting anomalies with a five-factor model . Review of Financial Studies 29 : 69 – 103 .

Fama, E. F. , and French K. R. . 2017 . Choosing factors . Working Paper , University of Chicago .

Feng, G. , Giglio S. , and Xiu D. . 2017 . Taming the factor zoo . Working Paper , University of Chicago .

Gilbert, S. , and Loi H. . 2018 . Digital currency risk . International Journal of Economics and Finance 10 : 108 .

Griffin, J. M. , and Shams A. . 2020 . Is bitcoin really un-tethered? Journal of Finance 52 : 57 – 87 .

Hong, H. , Lim T. , and Stein J. C. . 2000 . Bad news travels slowly: Size, analyst coverage, and the profitability of momentum strategies . Journal of Finance 55 : 265 – 295 .

Hou, K. , Xiong W. , and Peng L. . 2009 . A tale of two anomalies: The implications of investor attention for price and earnings momentum . Working Paper , Princeton University .

Hu, A. , Parlour C. A. , and Rajan U. . 2018 . Cryptocurrencies: Stylized facts on a new investible instrument . Working Paper , University of California , Berkeley .

Huberman, G. , Leshno J. D. , and Moallemi C. C. . 2017 . Monopoly without a monopolist: An economic analysis of the bitcoin payment system . Working Paper , University of Chicago .

Jegadeesh, N. , and Titman S. . 1993 . Returns to buying winners and selling losers: Implications for stock market efficiency . Journal of Finance 48 : 65 – 91 .

Jermann, U. J. 2018 . Bitcoin and Cagan’s model of hyperinflation . Working Paper , University of Pennsylvania .

Liu, Y. , Tsyvinski A. , and Wu X. . 2019 . Common Risk Factors in Cryptocurrency . Working Paper , Yale University .

Lustig, H. , Roussanov N. , and Verdelhan A. . 2011 . Common risk factors in currency markets . Review of Financial Studies 24 : 3731 – 3777 .

Mai, F. , Bai Q. , Shan Z. , Wang X. S. , and Chiang R. . 2016 . The impacts of social media on Bitcoin performance . Working Paper , University of Cincinnati .

Makarov, I. , and Schoar A. . 2020 . Trading and arbitrage in cryptocurrency markets . Journal of Financial Economics 135 : 293 – 319 .

Moskowitz, T. J. 2015 . Asset pricing and sports betting . Working Paper , Yale University .

Moskowitz, T. J. , and Grinblatt M. . 1999 . Do industries explain momentum? Journal of Finance 54 : 1249 – 1290 .

Moskowitz, T. J. , Ooi Y. H. , and Pedersen L. H. . 2012 . Time series momentum . Journal of Financial Economics 104 : 228 – 250 .

Pagnotta, E. , and Buraschi A. . 2018 . An equilibrium valuation of bitcoin and decentralized network assets . Working Paper , Imperial College Business School .

Pástor, L. and P. Veronesi. 2003 . Stock valuation and learning about profitability . Journal of Finance 58 : 1749 – 1789 .

Routledge, B. , and Zetlin-Jones A. . 2018 . Currency stability using blockchain technology . Working Paper , Carnegie Mellon University .

Saleh, F. 2018 . Blockchain without waste: Proof-of-stake . Working Paper , McGill University .

Schilling, L. , and Uhlig H. . 2019 . Some simple bitcoin economics . Journal of Monetary Economics 106 : 16 – 26 .

Shanaev, S. , Sharma S. , Shuraeva A. , and Ghimire B. . 2019 . Taming the blockchain beast? Regulatory implications for the cryptocurrency market . Working Paper , University of London .

Sockin, M. , and Xiong W. . 2019 . A model of cryptocurrencies . Working Paper , Princeton University .

Stoffels, J. 2017 . Asset pricing of cryptocurrencies and momentum based patterns . Working Paper , Erasmus School of Economics .

Wang, S. , and Vergne J. . 2017 . Buzz factor or innovation potential: What explains cryptocurrencies’ returns? PloS One 12 .

Weber, W. E. 2016 . A Bitcoin standard: Lessons from the gold standard . Working Paper , Bank of Canada .

Yermack, D. 2015 . Is Bitcoin a real currency? An economic appraisal. In Handbook of Digital Currency , ed. Chuen D. L. K. , 31 – 43 . Amsterdam : Elsevier .

Supplementary data

Email alerts

Citing articles via.

• Recommend to your Library

Affiliations

• Online ISSN 1465-7368
• Print ISSN 0893-9454
• Copyright © 2024 Society for Financial Studies
• About Oxford Academic
• Publish journals with us
• University press partners
• What we publish
• New features  
• Open access
• Institutional account management
• Rights and permissions
• Get help with access
• Accessibility
• Advertising
• Media enquiries
• Oxford University Press
• Oxford Languages
• University of Oxford

Oxford University Press is a department of the University of Oxford. It furthers the University's objective of excellence in research, scholarship, and education by publishing worldwide

• Copyright © 2024 Oxford University Press
• Cookie settings
• Cookie policy
• Privacy policy
• Legal notice

This Feature Is Available To Subscribers Only

Sign In or Create an Account

This PDF is available to Subscribers Only

For full access to this pdf, sign in to an existing account, or purchase an annual subscription.

Help | Advanced Search

Computer Science > Cryptography and Security

Title: deciphering the blockchain: a comprehensive analysis of bitcoin's evolution, adoption, and future implications.

Abstract: This research paper provides a comprehensive analysis of Bitcoin, delving into its evolution, adoption, and potential future implications. As the pioneering cryptocurrency, Bitcoin has sparked significant interest and debate in recent years, challenging traditional financial systems and introducing the world to the power of blockchain technology. This paper aims to offer a thorough understanding of Bitcoin's underlying cryptographic principles, network architecture, and consensus mechanisms, primarily focusing on the Proof-of-Work model. We also explore the economic aspects of Bitcoin, examining price fluctuations, market trends, and factors influencing its value. A detailed investigation of the regulatory landscape, including global regulatory approaches, taxation policies, and legal challenges, offers insights into the hurdles and opportunities faced by the cryptocurrency. Furthermore, we discuss the adoption of Bitcoin in various use cases, its impact on traditional finance, and its role in the growing decentralized finance (DeFi) sector. Finally, the paper addresses the future of Bitcoin and cryptocurrencies, identifying emerging trends, technological innovations, and environmental concerns. We evaluate the potential impact of central bank digital currencies (CBDCs) on Bitcoin's future, as well as the broader implications of this technology on global finance. By providing a holistic understanding of Bitcoin's past, present, and potential future, this paper aims to serve as a valuable resource for scholars, policymakers, and enthusiasts alike.

Submission history

Access paper:.

• Other Formats

References & Citations

• Google Scholar
• Semantic Scholar

BibTeX formatted citation

Bibliographic and Citation Tools

Code, data and media associated with this article, recommenders and search tools.

• Institution

arXivLabs: experimental projects with community collaborators

arXivLabs is a framework that allows collaborators to develop and share new arXiv features directly on our website.

Both individuals and organizations that work with arXivLabs have embraced and accepted our values of openness, community, excellence, and user data privacy. arXiv is committed to these values and only works with partners that adhere to them.

Have an idea for a project that will add value for arXiv's community? Learn more about arXivLabs .

Cryptocurrency Special Topic Hub

Decentralized digital assets, currencies, and tokens have garnered attention from academics and practitioners in nearly every scholarly discipline. Cryptocurrencies, and related areas of interest, have wide-ranging business, economic, environmental, legal, political, and regulatory implications. The Cryptocurrency Special Topic Hub presents interdisciplinary scholarship covering all aspects of cryptocurrencies, from the technical platforms, solutions, and tools used, to user identity, authentication, and the encryption empowering it, to the regulatory policies it influences.

It is important to note that many of these papers have not been edited or benefited from the pivotal role of peer-review, which validates and improves the quality of final published journal articles.

SSRN, Elsevier’s world-leading platform devoted to the rapid worldwide dissemination of early-stage research, is committed to advancing societal progress through quality knowledge and education. Research on SSRN is free to download and upload.

Additional resources for cryptocurrency research

• ScienceDirect Topic Page on Cryptocurrencies

Risks and Returns of Cryptocurrency

We establish that the risk-return tradeoff of cryptocurrencies (Bitcoin, Ripple, and Ethereum) is distinct from those of stocks, currencies, and precious metals. Cryptocurrencies have no exposure to most common stock market and macroeconomic factors. They also have no exposure to the returns of currencies and commodities. In contrast, we show that the cryptocurrency returns can be predicted by factors which are specific to cryptocurrency markets. Specifically, we determine that there is a strong time-series momentum effect and that proxies for investor attention strongly forecast cryptocurrency returns. Finally, we create an index of exposures to cryptocurrencies of 354 industries in the US and 137 industries in China.

We thank Andrew Atkeson, Nicola Borri, Eduardo Davila, Stefano Giglio, William Goetzmann, Stephen Roach, and Robert Shiller for their comments. The views expressed herein are those of the authors and do not necessarily reflect the views of the National Bureau of Economic Research.

MARC RIS BibTeΧ

Download Citation Data

Published Versions

Yukun Liu & Aleh Tsyvinski & Itay Goldstein, 2021. " Risks and Returns of Cryptocurrency, " The Review of Financial Studies, vol 34(6), pages 2689-2727.

Mentioned in the News

More from nber.

In addition to working papers , the NBER disseminates affiliates’ latest findings through a range of free periodicals — the NBER Reporter , the NBER Digest , the Bulletin on Retirement and Disability , the Bulletin on Health , and the Bulletin on Entrepreneurship  — as well as online conference reports , video lectures , and interviews .

Blockchain and cryptocurrencies: economic and financial research

• Published: 13 November 2021
• Volume 44 , pages 781–787, ( 2021 )

Cite this article

• Alessandra Cretarola 1 ,
• Gianna Figà-Talamanca 2 &
• Cyril Grunspan 3  

5928 Accesses

8 Citations

3 Altmetric

Explore all metrics

The motivation of proposing and editing the Special Issue “Blockchain and cryptocurrencies” came from the inspirational invited and contributed talks at the 43rd annual A.M.A.S.E.S. conference held in Perugia in September 2019. All the papers have gone through the journal regular refereeing process under the same standards set by the journal, and nine contributions were finally accepted for publication.

Explore related subjects

• Artificial Intelligence

Avoid common mistakes on your manuscript.

1 Literature background

Bitcoin, the pioneer cryptocurrency, was conceived in 2008 by an individual or a group of researchers under the pseudonymous of Nakamoto ( 2008 ) and implemented in 2009 as an electronic payment system via a breakthrough application of the blockchain technology; it is based on an open-source software which generates a peer-to-peer network and does not rely on central banks to regulate the money supply, and it enables essentially anonymous transactions. Two major technological advances set Bitcoin apart in the world of cryptocurrencies:

the extensive and repeated use of proof of work;

the invention of smart contracts.

The first advance (proof of work) establishes a link between computer security and probability theory (specifically, Poisson processes). Prior to Satoshi Nakamoto’s work, there have been attempts at creating decentralized cryptocurrency networks with secure transactions. Nakamoto’s model represents a different paradigm, where double-spending attacks—spending the same amount twice—are unlikely, provided some requirements are met (Grunspan and Pérez-Marco 2018 ). Such repeated attacks prove to be unprofitable, unless one is ready to double-spend enormous amounts on a regular basis—an unreasonable assumption in practice (Grunspan 2021 ). This is achieved using the proof of work concept, which originates in the design of spam filters (Dwork and Naor 1993 ).

The second advance (strangely absent from Bitcoin’s founding paper, which only focuses on describing the protocol itself) turns Bitcoin into a truly programmable currency. It allows for the creation of payment channels between users which, combined, give rise to the Lightning Network, an overlay to the Bitcoin network (Dryja and Poon 2015 ). This new network has incredible performance in terms of scalability, and its transaction throughput (number of transactions processed per second) is far superior to that of traditional centralized networks like Visa or Paypal.

Smart contracts also pave the way to a truly decentralized finance (DeFi). One example is Ethereum, another cryptocurrency created in 2015 in the wake of Bitcoin thanks to Solidity (its flexible programming language). This new finance handles volumes of about 100 billion today, and is growing rapidly ( \(+400\) % in one year). DeFi relies on new offerings that were never implemented in traditional markets. For instance, Automated Market Makers are truly decentralized trading platforms (Angeris and Chitra 2020 ). Defi also offers new products such as flash loans or staking derivatives for cryptocurrencies like Solana or Ethereum 2.0 based on a protocol using PoS (Chitra and Evans 2020 ). The link with traditional finance is made through “stable coins”—crypto-currencies whose price is pegged to the US dollar.

The academic response to this economic and financial breakpoint was initially hesitant and cautious but eventually raised in 2016, and was especially boosted by the exponential price increase in 2017. First economic studies focused on framing Bitcoin among fiat currencies, commodities, or stocks, trying to identify its intrinsic value, see Nian and Chuen ( 2015 ), Yermack ( 2015 ), Cheah and Fry ( 2015 ), Bjerg ( 2016 ) and Böhme et al. ( 2015 ). Then, the advent of thousands of new cryptocurrencies has raised several new research questions: are there specific stylized facts, how are cryptocurrencies related to traditional financial assets, how can they be used to get profitable trading strategies, how are they related to one another, are there common driving factors? Indeed, Bitcoin and other cryptocurrencies are currently seen as an alternative investment class, and, though their high-returns/high-volatility profile has initially motivated speculative investments (Yermack 2015 ; Baur et al. 2018 ), more recent studies claim that investing in these assets does improve the overall portfolio diversification (Platanakis et al. 2018 ). One of the main strands of research in Bitcoin and cryptocurrencies is represented by the econometric modeling of the price dynamics and the identification of possible driving factors. Some contributions, see e.g., Dyhrberg ( 2016 ), describe the relationship between Bitcoin returns and factors in the traditional financial market, such as the price of gold, gold futures, and stock market indexes. Moreover, Kristoufek ( 2015 ) investigates the dependence of Bitcoin price on crypto-related factors, such as the hash-rate, the mining difficulty, the number of transactions and the total number of Bitcoin in circulation. Many researches suggest that Bitcoin returns and volatility are driven by investor attention, sentiment or by specific measures of market attractiveness, see e.g., Ciaian et al. ( 2016 ), Figa-Talamanca and Patacca ( 2019 ), Ahn and Kim ( 2019 ), Eom et al. ( 2019 ) and Figà-Talamanca and Patacca ( 2020 ). Notably, the occurrence of periods of hype for cryptocurrencies, followed by periods of skepticism, has encouraged the analysis of jumps, structural breaks and bubble effects in their price dynamics, see Garcia et al. ( 2014 ), Cheah and Fry ( 2015 ), Fry and Cheah ( 2016 ), Corbet et al. ( 2018b ), Bouri et al. ( 2019 ), Cretarola and Figà-Talamanca ( 2019 ), Scaillet et al. ( 2020 ), Hafner ( 2020 ) and Cretarola and Figà-Talamanca ( 2020 ), among others. The relationship between different cryptocurrencies being very strong, most of the above contributions focus on Bitcoin as a benchmark for the whole asset class. The interdependence between cryptocurrencies is investigated in Ciaian and Rajcaniova ( 2018 ), Yaya et al. ( 2019 ), Chaim and Laurini ( 2019 ), Kumar and Anandarao ( 2019 ), Mensi et al. ( 2019 ), Blau et al. ( 2020 ), Tiwari et al. ( 2020 ) and Figà-Talamanca et al. ( 2021 ), by applying several dependence concepts and alternative methodologies. After the introduction of Bitcoin Futures in the Chicago Board Option Exchange, a few papers have also appeared on the evaluation of Bitcoin derivatives, see Chen et al. ( 2018 ), Corbet et al. ( 2018a ), Cretarola et al. ( 2020 ) and Siu and Elliott ( 2021 ), whereas investments, trading strategies and possible arbitrage opportunities are discussed in Lintilhac and Tourin ( 2017 ), Nakano et al. ( 2018 ), Bistarelli et al. ( 2019 ), Bistarelli et al. ( 2019b ) and Shynkevich ( 2021 ).

2 Special issue contributions

The collected papers span several topics, most of them through econometric analyses: the modeling of returns and volatility dynamics; the identification of profitable investment strategies; the study of the trading volume of cryptocurrencies; the post-sale behavior of tokens issued through initial coin offerings as well as the development of the overall coins and tokens industry.

Majdoub et al. ( 2021 ) attempts to investigate whether Bitcoin can be hedged by selected fiat currencies (EUR, JPY, and GBP). The authors compute optimal hedge ratios between Bitcoin and fiat currencies over the period February 2012-November 2017 based on the VAR-DCC-GARCH model, VAR-ADCC-GARCH model and VAR-component GARCH-DCC model. They evidence that the correlations between Bitcoin and fiat currencies have a time-varying dynamic under different model specifications. Thereafter, they propose suitable dynamic hedging strategies when investing in the Bitcoin market.

The relationship between volatilities of five cryptocurrencies, American indices (SP500, Nasdaq, and VIX), oil, and gold is analyzed in Ghorbel and Jeribi ( 2021 ) in a multivariate BEKK-GARCH model framework. A higher volatility spillover is detected between cryptocurrencies than between cryptocurrencies and other financial assets.

Kyriazis ( 2021 ) investigates the nexus between cryptocurrencies connected to cannabis production and the three highest capitalization digital currencies, Bitcoin, Ethereum and Ripple. The GARCH, EGARCH, TGARCH and GJR-GARCH specifications are employed in order to analyze volatility characteristics. Findings reveal that GARCH and GJR-GARCH specifications are most appropriate to explain the volatility of each cannabis cryptocurrency. This allows to recognize the existence of thresholds in the volatility of cannabis cryptocurrencies when examining their nexus with major digital currencies.

The multivariate relationship between cryptocurrencies is investigated in Figà-Talamanca et al. ( 2021 ) by applying a dynamic factor analysis to the joint behavior of Bitcoin, Ethereum, Litecoin and Monero, as a representative basket of the cryptocurrencies asset class, from 2016 to 2019. The authors show that the basket price is suitably described by a model with two dynamic factors, of which the first is integrated and the second is stationary until the end of August 2019. Based on this evidence, they introduce a trading strategy which proves profitable only when the second factor is stationary.

Angelis et al. ( 2021 ) propose innovative profit-oriented trading strategies on Bitcoins for risk-seeking investors, which are based on buying or selling the so-called Contracts for Difference . Moreover, thanks to the use of a proper machine/deep learning model for predicting aims, i.e., a recurrent neural network with a long short-term memory, the authors are able to forecast possible investment scenarios under a suitable theoretical model specification.

The problem of forecasting the intraday short-term volume and its uncertainty in exchange markets for cryptocurrencies is studied in Antulov-Fantulin et al. ( 2021 ). Precisely, the predictions are built by using transaction and order book data from different markets, and a temporal mixture ensemble model is proposed to identify, at each time step, the set of data which is locally most useful for the forecasting. The authors obtain some empirical findings on the outperformance of their model with respect to time series models both in point and in interval predictions of trading volume. In addition, they also evidence similar results when conditioning to the volume quartile.

In Provenzano and Baggio ( 2021 ), a wide application of complex network analysis to cryptocurrencies markets is provided. Specifically, the authors characterize the dynamics and estimate the synchronicity between the price and volume series of three cryptocurrencies, Bitcoin, Ethereum, and Litecoin. The results show similar complex structures in terms of number and internal composition of communities. In particular, cyclical patterns of similar wavelength and amplitude, and a different degree of synchronization before and after a collapse event over the period considered can be observed in the price and volume dynamics of the three cryptocurrencies, suggesting possible investment opportunities.

The novel funding mechanism of initial coin offerings (ICOs), where digital tokens are issued and sold to investors, is the subject of research of Ante and Meyer ( 2021 ). In particular, they investigate the price effects of 250 cross-listing events of 135 individual tokens and possible abnormal returns when they are immediately traded on secondary markets. They find significant abnormal returns of 6.51% on the listing day and 9.97% over a seven-day window around the event; these abnormal returns are also affected by the exchanges on which the listings occur and by liquidity-related metrics.

Finally, Gandal et al. ( 2021 ) analyze the flourishing industry of cryptocurrency coins and tokens. Even though these terms are commonly used as interchangeable, they are very different in nature and deserve a separate analysis. While cryptocurrency coins have the purpose to replace fiat currency as a store of value, the main use of a token is to fund a blockchain-based project through an ICO. The authors investigate the creation, competition, and destruction of 1082 coins and 725 tokens during the period 2013–2018 and find that 44% of publicly traded coins are abandoned, at least temporarily; 71% of abandoned coins are later resurrected; 18% of coins fail permanently. An unexpected finding is that the bursting of the 2017 Bitcoin bubble has not affected the rise of alternative cryptocurrencies.

Ahn, Y., Kim, D.: Sentiment disagreement and bitcoin price fluctuations: a psycholinguistic approach. Appl. Econ. Lett., pp. 1–5 (2019)

Angeris, G., Chitra, T.: Improved price oracles: constant function market makers. In: Proceedings of the 2nd ACM Conference on Advances in Financial Technologies, Association for Computing Machinery, pp. 80–91 (2020)

Ante, L., Meyer, A.: Cross-listings of blockchain-based tokens issued through initial coin offerings: Do liquidity and specific cryptocurrency exchanges matter? Decis. Econ. Finance, pp. 1–24 (2021)

Antulov-Fantulin, N., Guo, T., Lillo, F.: Temporal mixture ensemble models for probabilistic forecasting of intraday cryptocurrency volume. Decis. Econ. Finance, pp. 1–36 (2021)

Baur, D.G., Hong, K., Lee, A.D.: Bitcoin: Medium of exchange or speculative assets? J. Int. Finan. Markets. Inst. Money 54 , 177–189 (2018)

Article   Google Scholar  

Bistarelli, S., Cretarola, A., Figà-Talamanca, G., Mercanti, I., Patacca, M.: Is arbitrage possible in the bitcoin market? (Work-in-progress paper). In: Coppola M, Carlini E, D’Agostino D, Altmann J, Bañares JÀ (eds) Economics of Grids, Clouds, Systems, and Services. GECON 2018. Lecture Notes in Computer Science, vol. 11113., Springer, Cham, pp. 243–251 (2019a)

Bistarelli, S., Cretarola, A., Figà-Talamanca, G., Patacca, M.: Model-based arbitrage in multi-exchange models for bitcoin price dynamics. Digit. Finance 1 (1), 23–46 (2019b)

Bjerg, O.: How is bitcoin money? Theory, Culture & Society 33 (1), 53–72 (2016)

Blau, B., Griffith, T., Whitby, R., et al.: Comovement in the cryptocurrency market. Economics Bulletin 40 (1), 448–455 (2020)

Google Scholar  

Böhme, R., Christin, N., Edelman, B., Moore, T.: Bitcoin: Economics, technology, and governance. J. Econ. Perspect. 29 (2), 213–238 (2015)

Bouri, E., Gil-Alana, L.A., Gupta, R., Roubaud, D.: Modelling long memory volatility in the bitcoin market: evidence of persistence and structural breaks. Int. J. Finance Econ. 24 , 412–426 (2019)

Chaim, P., Laurini, M.P.: Nonlinear dependence in cryptocurrency markets. N. Am. J. Econ. Finance 48 , 32–47 (2019)

Cheah, E.T., Fry, J.: Speculative bubbles in bitcoin markets? an empirical investigation into the fundamental value of bitcoin. Econ. Lett. 130 , 32–36 (2015)

Chen, C. Y. H., Härdle, W. K., Hou, A. J., Wang, W.: Pricing cryptocurrency options: the case of CRIX and Bitcoin. Tech. rep., IRTG 1792 Discussion Paper (2018)

Chitra, T., Evans, A.: Why stake when you can borrow? (2020). arXiv:2006.11156

Ciaian, P., Rajcaniova, M., Kancs, D.: The economics of bitcoin price formation. Appl. Econ. 48 (19), 1799–1815 (2016)

Ciaian, P., Rajcaniova, M., et al.: Virtual relationships: Short-and long-run evidence from Bitcoin and altcoin markets. J. Int. Finan. Markets. Inst. Money 52 , 173–195 (2018)

Corbet, S., Lucey, B., Peat, M., Vigne, S.: Bitcoin futures-what use are they? Econ. Lett. 172 , 23–27 (2018a)

Corbet, S., Lucey, B., Yarovaya, L.: Datestamping the bitcoin and Ethereum bubbles. Financ. Res. Lett. 26 , 81–88 (2018b)

Cretarola, A., Figà-Talamanca, G.: Detecting bubbles in bitcoin price dynamics via market exuberance. Ann. Oper. Res. 299 , 1–21 (2019)

Cretarola, A., Figà-Talamanca, G.: Bubble regime identification in an attention-based model for bitcoin and Ethereum price dynamics. Econ. Lett. 191 , 108831 (2020)

Cretarola, A., Figà-Talamanca, G., Patacca, M.: Market attention and bitcoin price modeling: theory, estimation and option pricing. Decis. Econ. Finan. 43 (1), 187–228 (2020)

De Angelis, P., De Marchis, R., Marino, M., Martire, A. L., Oliva, I.: Betting on bitcoin: a profitable trading between directional and shielding strategies. Decis. Econ. Finance pp 1–21 (2021)

Dryja, T., Poon, J.: Lightning network (2015). https://lightning.network/

Dwork, C., Naor, M.: Pricing via processing or combatting junk mail. In: Brickell, E.F. (ed.) Advances in Cryptology - CRYPTO’ 92, pp. 139–147. Springer, Berlin (1993)

Dyhrberg, A. H.: Bitcoin, gold and the dollar - a garch volatility analysis. Finance Res. Lett. 16(Supplement C):85 – 92 (2016)

Eom, C., Kaizoji, T., Kang, S.H., Pichl, L.: Bitcoin and investor sentiment: statistical characteristics and predictability. Phys. A 514 , 511–521 (2019)

Figa-Talamanca, G., Patacca, M.: Does market attention affect bitcoin returns and volatility? Decis. Econ. Finance 42 (1), 135–155 (2019)

Figà-Talamanca, G., Patacca, M.: Disentangling the relationship between bitcoin and market attention measures. J. Ind. Bus. Econ. 47 (1), 71–91 (2020)

Figà-Talamanca, G., Focardi, S., Patacca, M.: Regime switches and commonalities of the cryptocurrencies asset class. N. Am. J. Econ. Finance 57 , 101425 (2021)

Figà-Talamanca, G., Focardi, S., Patacca, M.: Common dynamic factors for cryptocurrencies and multiple pair-trading statistical arbitrages. Decis. Econ. Finance, pp. 1–20 (2021)

Fry, J., Cheah, E.T.: Negative bubbles and shocks in cryptocurrency markets. Int. Rev. Financ. Anal. 47 , 343–352 (2016)

Gandal, N., Hamrick, J., Moore, T., Vasek, M.: The rise and fall of cryptocurrency coins and tokens. Decis. Econ. Finance, pp. 1–34 (2021)

Garcia, D., Tessone, C.J., Mavrodiev, P., Perony, N.: The digital traces of bubbles: feedback cycles between socio-economic signals in the bitcoin economy. J. R. Soc. Interface 11 (99), 20140623 (2014)

Ghorbel, A., Jeribi, A.: Investigating the relationship between volatilities of cryptocurrencies and other financial assets. Decis. Econ. Finance, pp. 1–27 (2021)

Grunspan, C., Pérez-Marco, R.: Double spend races. Int. J. Theor. Appl. Finance 21 (08), 1850053 (2018)

Grunspan, C., Pérez-Marco, R.: On profitability of Nakamoto double spend. Probability in the Engineering and Informational Sciences pp. 1–15 (2021)

Hafner, C.M.: Testing for bubbles in cryptocurrencies with time-varying volatility. J. Financ. Economet. 18 (2), 233–249 (2020)

Kristoufek, L.: What are the main drivers of the bitcoin price? Evidence from wavelet coherence analysis. PLoS ONE 10 (4), e0123923 (2015)

Kumar, A.S., Anandarao, S.: Volatility spillover in crypto-currency markets: some evidences from GARCH and wavelet analysis. Phys. A 524 , 448–458 (2019)

Kyriazis, N. A.: the Investigating diversifying or hedging nexus of cannabis cryptocurrencies with major digital currencies. Decis. Econ. Finance, pp. 1–21 (2021)

Lintilhac, P.S., Tourin, A.: Model-based pairs trading in the bitcoin markets. Quant. Finance 17 (5), 703–716 (2017)

Majdoub, J., Sassi, S.B., Bejaoui, A.: Can fiat currencies really hedge bitcoin? evidence from dynamic short-term perspective. Decisions in Economics and Finance, pp. 1–28 (2021)

Mensi, W., Rehman, M.U., Al-Yahyaee, K.H., Al-Jarrah, I.M.W., Kang, S.H.: Time frequency analysis of the commonalities between bitcoin and major cryptocurrencies: portfolio risk management implications. N. Am. J. Econ. Finance 48 , 283–294 (2019)

Nakamoto, S.: Bitcoin: A peer-to-peer electronic cash system. Decent. Bus. Rev., p. 21260 (2008)

Nakano, M., Takahashi, A., Takahashi, S.: Bitcoin technical trading with artificial neural network. Phys. A 510 , 587–609 (2018)

Nian, L.P., Chuen, D.L.K.: Introduction to bitcoin. In: Chuen, D.L.K. (ed.) Handbook of Digital Currency, Elsevier, pp 5–30 (2015)

Platanakis, E., Sutcliffe, C., Urquhart, A.: Optimal vs naïve diversification in cryptocurrencies. Econ. Lett. 171 , 93–96 (2018)

Provenzano, D., Baggio, R.: Complexity traits and synchrony of cryptocurrencies price dynamics. Decis. Econ. Finance, pp. 1–15 (2021)

Scaillet, O., Treccani, A., Trevisan, C.: High-frequency jump analysis of the bitcoin market. J. Financ. Economet. 18 (2), 209–232 (2020)

Shynkevich, A.: Bitcoin arbitrage. Financ. Res. Lett. 40 , 101698 (2021)

Siu, T.K., Elliott, R.J.: Bitcoin option pricing with a SETAR-GARCH model. Eur. J. Finance 27 (6), 564–595 (2021)

Tiwari, A.K., Adewuyi, A.O., Albulescu, C.T., Wohar, M.E.: Empirical evidence of extreme dependence and contagion risk between main cryptocurrencies. N. Am. J. Econ. Finance 51 , 101083 (2020)

Yaya, O. S., Ogbonna, A. E., Olubusoye, O. E.: How persistent and dynamic inter-dependent are pricing of bitcoin to other cryptocurrencies before and after 2017/18 crash? Physica A: Statistical Mechanics and its Applications p 121732 (2019)

Yermack, D.: Is bitcoin a real currency? An economic appraisal. In: Chuen, D.L.K. (ed.) Handbook of Digital Currency, Elsevier, pp. 31–43 (2015)

Download references

Author information

Authors and affiliations.

Department of Mathematics and Computer Science, University of Perugia, via L. Vanvitelli 1, 06123, Perugia, Italy

Alessandra Cretarola

Department of Economics, University of Perugia, via A. Pascoli 20, 06123, Perugia, Italy

Gianna Figà-Talamanca

Léonard de Vinci Pôle Universitaire, Research Center, 92916, Paris, La Défense Cedex, France

Cyril Grunspan

You can also search for this author in PubMed   Google Scholar

Corresponding author

Correspondence to Gianna Figà-Talamanca .

Additional information

Publisher's note.

Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.

Rights and permissions

Reprints and permissions

About this article

Cretarola, A., Figà-Talamanca, G. & Grunspan, C. Blockchain and cryptocurrencies: economic and financial research. Decisions Econ Finan 44 , 781–787 (2021). https://doi.org/10.1007/s10203-021-00366-3

Download citation

Accepted : 25 October 2021

Published : 13 November 2021

Issue Date : December 2021

DOI : https://doi.org/10.1007/s10203-021-00366-3

Share this article

Anyone you share the following link with will be able to read this content:

Sorry, a shareable link is not currently available for this article.

Provided by the Springer Nature SharedIt content-sharing initiative

• Find a journal
• Publish with us
• Track your research

• Open access
• Published: 02 September 2024

Herding and investor sentiment after the cryptocurrency crash: evidence from Twitter and natural language processing

• Michael Cary 1  

Financial Innovation volume  10 , Article number:  142 ( 2024 ) Cite this article

Metrics details

Although the 2022 cryptocurrency market crash prompted despair among investors, the rallying cry, “wagmi” (We’re all gonna make it.) emerged among cryptocurrency enthusiasts in the aftermath. Did cryptocurrency enthusiasts respond to this crash differently compared to traditional investors? Using natural language processing techniques applied to Twitter data, this study employed a difference-in-differences method to determine whether the cryptocurrency market crash had a differential effect on investor sentiment toward cryptocurrency enthusiasts relative to more traditional investors. The results indicate that the crash affected investor sentiment among cryptocurrency enthusiastic investors differently from traditional investors. In particular, cryptocurrency enthusiasts’ tweets became more neutral and, surprisingly, less negative. This result appears to be primarily driven by a deliberate, collectivist effort to promote positivity within the cryptocurrency community (“wagmi”). Considering the more nuanced emotional content of tweets, it appears that cryptocurrency enthusiasts expressed less joy and surprise in the aftermath of the cryptocurrency crash than traditional investors. Moreover, cryptocurrency enthusiasts tweeted more frequently after the cryptocurrency crash, with a relative increase in tweet frequency of approximately one tweet per day. An analysis of the specific textual content of tweets provides evidence of herding behavior among cryptocurrency enthusiasts.

Introduction

Cryptocurrencies have grown rapidly in popularity, especially among non-traditional investors (Mattke et al. 2021 ). Consequently, the motivations underlying the decisions of many cryptocurrency investors are not always purely financial, with investors exhibiting substantial levels of herding behavior with respect to cryptocurrencies (Ooi et al. 2021 ). In fact, the culture developing around cryptocurrency enthusiasts engaging in herding behavior is rich and complex (Dodd 2018 ). The volatility of cryptocurrencies can vary substantially, and smaller cryptocurrencies (e.g., Dogecoin) are especially influenced by the decisions of herding-type investors (Cary 2021 ).

Because cryptocurrency investors are known to hold on to cryptocurrencies for ideological and cultural reasons, even when the return on cryptocurrency investments is negative (Mattke et al. 2021 ), the cryptocurrency crash of 2022 has the potential to drastically affect investor sentiment, especially among those who exhibit loyalty to cryptocurrencies despite negative wealth consequences. Magnifying this concern, Vidal-Tomás et al. ( 2019 ) showed that herding behavior among cryptocurrency investors is particularly strong in down markets.

The May 2022 cryptocurrency crash was one of the largest crashes in the history of cryptocurrency. Sparked by the collapse of the stablecoin Terra, the entire cryptocurrency market crashed (De Blasis et al. 2023 ). Before the crash, Terra was the third-largest cryptocurrency ecosystem after Bitcoin and Ethereum (Liu et al. 2023 ). Terra and its tethered floating-rate cryptocurrency (i.e., Luna) became valueless in only three days, representing the first major run on a cryptocurrency (Liu et al. 2023 ). The spillover effects on other cryptocurrencies have been widespread, with the Terra crash affecting the connectedness of the entire cryptocurrency market (Lee et al. 2023 ). Although an attempt to stabilize the stablecoin was made, the creator was ultimately charged and arrested for securities fraud (Judge 2023 ). While this was the first dissolution of a stablecoin cryptocurrency, it is worth mentioning that scholars have shown that pre-crypto stablecoins have existed, and failed (e.g., the Bank of Amsterdam from the 17th through the early 19th centuries, which tied its currency to silver and gold coins (Frost et al. 2020 )). The cryptocurrency community has much to learn from the history of currency; in many cases, its ideas and attitudes are far from novel.

The consequences of an unregulated cryptocurrency market were not constrained by the cryptocurrency crashes examined in this study. Only months after the cryptocurrency crash of May 2022, the FTX collapsed (i.e., the Futures Exchange, formerly the world’s third largest cryptocurrency exchange and hedge fund). This led to a bank run and, ultimately, multiple internationally renowned figures in the cryptocurrency community being imprisoned for fraud Footnote 1 These failures cost tax payers billions of dollars in the form of bailouts for cryptocurrency investors. These are not the only costs society faces due to cryptocurrency; cryptocurrency is also the medium of exchange for $76 billion of illegal activity, with approximately 46% of Bitcoin transactions representing illegal transactions (Foley et al. 2019 ).

To date, research on this crash has primarily focused on spillovers among different cryptocurrencies or certain commodities. While some research on investor-level sentiment has been published, studies have not explicitly tested the differences in responses between hardcore cryptocurrency enthusiasts and traditional investors who may have held some cryptocurrencies in their portfolios. Thus, given the severity, significance, and recency of this crash in the cryptocurrency market, it is critical to ascertain whether cryptocurrency enthusiasts behave in a fundamentally differently manner than traditional investors, especially in the context of a negative shock. If so, this could potentially lead to greater volatility and is a further reason for regulating the cryptocurrency market. Accordingly, this study seeks to fill the gap in the literature by providing evidence that the May 2022 cryptocurrency crash disproportionately affected herding-type cryptocurrency enthusiasts (relative to traditional investors) as measured by its impact on the sentiment of tweets. This is accomplished by applying natural language processing techniques to harvested Twitter data to quantify the text of tweets, classifying Twitter users into cryptocurrency enthusiasts and a control group, and then feeding this data into a difference-in-differences (DID) model to estimate the potential differential effect of cryptocurrency crashes on herding-type cryptocurrency investors relative to traditional investors. Additionally, this paper analyzes the specific textual content of the tweets in each group to further assess the presence of herding behavior. Such an analysis is important because the presence of herding generates further cause for regulating cryptocurrency markets as herding is known to lead to bubbles (Haykir and Yagli 2022 ).

The results of the DID regressions confirm that the sentiment of tweets from these herding-type cryptocurrency investors became relatively more negative after the cryptocurrency crash, providing evidence of a decline in investor sentiment among herding-type cryptocurrency investors relative to that of traditional investors. The DID estimators estimated in this study are best interpreted as the magnitude of the differential response to the cryptocurrency crash between cryptocurrency enthusiasts and traditional investors. Critically, the significant effect estimated here indicates that these two groups behaved in fundamentally different ways, confirming that they are indeed distinct.

Additionally, the results show that cryptocurrency enthusiasts began to tweet relatively more often after the cryptocurrency crash, suggesting that multiple behavioral changes occurred as a consequence of the crash. This provides further evidence that cryptocurrency enthusiasts and traditional investors are fundamentally different groups, with distinct responses to similar stimuli. Evidentiary, a classification of the specific textual content of tweets in each group, reveals evidence of herding behavior among cryptocurrency enthusiasts but not among traditional investors. Furthermore, a large portion of this herding behavior exhibited by cryptocurrency enthusiasts is centered on related cultural artifacts such as non-fungible tokens (NFTs).

Literature review

Despite the fact that many cryptocurrencies (e.g., Bitcoin) have a history of bubbles (Chaim and Laurini 2019 ), many cryptocurrency enthusiasts routinely invest excessively in them. This seemingly irrational behavior can lead to people tying a large proportion of their financial well-being to cryptocurrency. Because financial distress can lead to deteriorated mental health (Starkey et al. 2013 ) and, by extension, more negative sentiment among investors, the cryptocurrency crash of May 2022 presents a major concern regarding the well-being of herding-type cryptocurrency investors.

The community of investors in cryptocurrencies is diverse, especially among more established cryptocurrencies such as Bitcoin (Dodd 2018 ). However, cryptocurrencies in general, and many smaller, less-established cryptocurrencies in particular, have a core group of ideologues that form the basis of the community (Ooi et al. 2021 ). These ideologically motivated communities are typically very libertarian (Obreja 2022 ), with many members more concerned with belonging to the community and holding cryptocurrency than maximizing the return on their investment (Mattke et al. 2021 ). Understanding the nature of the communities around cryptocurrencies is important because these communities are critical predictors of the growth and popularity of cryptocurrency in terms of both investing and mining (Al Shehhi et al. 2014 ).

The libertarian nature of the cryptocurrency community is particularly relevant given the prevalence of confirmation bias, political and information silos, and the growing number of calls to regulate cryptocurrencies. The strong role of confirmation bias among cryptocurrency investors has been documented (Zhang et al. 2019 ). While regulations may help increase public trust in cryptocurrencies and protect investors and members of this community are among the most likely to benefit from them, (Giudici et al. 2020 ) the cryptocurrency community is largely ideologically opposed to any regulation.

Collectivist behavior exhibits itself in the cryptocurrency community in other ways. Although perhaps unprincipled, herding behavior among cryptocurrency investors is a well-documented phenomenon (Kallinterakis and Wang 2019 ). According to Haykir and Yagli ( 2022 ), herding behavior in cryptocurrency was prominent during the global COVID-19 pandemic. A study of 50 cryptocurrencies also revealed evidence of herding behavior among investors (da Gama Silva et al. 2019 ). Specific events have been found to increase herding behavior among cryptocurrency investors, including the expiration date of Bitcoin futures on the Chicago Mercantile Exchange (Blasco et al. 2022 ). Generally, herding behavior tends to be at its highest when uncertainty is high (Bouri et al. 2019 ). Combining this with the result from Vidal-Tomás et al. ( 2019 ) that herding is strongest when markets are down, we can see that the cryptocurrency crash of 2022 is an important event that can be used to study the behavior of cryptocurrency investors.

Social media is one of the richest sources of data for studying investor behavior. Researchers can study investors’ behavior and motivations by collecting social media data and using natural language processing (NLP) techniques (Zhou 2018 ). The most commonly used NLP technique is sentiment analysis (Liu 2010 ). A prominent example of the use of sentiment analysis in the finance literature comes from Gao et al. ( 2022 ), who used the announcement of the Chinese “dual carbon” target to study the impacts of investor sentiment on the volatility of (green) stock returns.

Several studies generally consider the role of investor sentiment in stocks (Baker and Wurgler 2006 , 2007 ; Baker et al. 2012 ; Da et al. 2015 ). In addition, Seok et al. ( 2019 ) and Xu and Zhou ( 2018 ) examined the role of investor sentiment in Korean and Chinese stocks, respectively. However, the application of sentiment analysis to financing does not end with the stock market. Using data on bettor sentiment, Avery and Chevalier ( 1999 ) showed that bettor sentiment affects the point spread in football games.

Other important papers related to this one include Kou et al. ( 2023 ) which demonstrates how fuzzy methods of studying facial expressions are relevant to making sustainable financial decisions. Kou et al. ( 2024 ) used fuzzy methods to impute expert financial decisions that are “based on the golden ratio”. Another paper by Kou et al. ( 2021 ) used fuzzy methods to study fintech investments in the context of European banks. Leaving the realm of fuzzy methodologies, Kou et al. ( 2021 ) developed a model to predict bankruptcies among small and medium sized businesses. Footnote 2

Turning to the effects of investor sentiment on cryptocurrencies, the literature remains plentiful. Cryptocurrencies do not always respond to new information in the same manner as traditional investments Rognone et al. ( 2020 ). This is particularly important because the sentiment analysis of both news (Lamon et al. 2017 ) and social media (Philippas et al. 2019 ) has been linked to changes in cryptocurrency prices. Mai et al. ( 2018 ) built on these results by showing that not only did social media sentiment affect cryptocurrency markets but also that such effects were driven by the sentiment of low-frequency posters, not high-frequency posters. Furthermore, relevant sentiment data from social media have been shown to affect the volatility of cryptocurrency markets (Ahn and Kim 2021 ) and liquidity (Yue et al. 2021 ) and can predict bubbles in cryptocurrency markets (Phillips and Gorse 2017 ). Several studies have considered the effects of the sentiment of (or pertaining to) influential figures on cryptocurrency prices, most notably Ante ( 2023 ) and Cary ( 2021 ).

In the case of Cary ( 2021 ), there was a severe negative impact on the price of Dogecoin attributable to the action of the crypto-tastemaker, who affixed their celebrity to the cryptocurrency. This raises an important, more general concern: given that Anastasiou et al. ( 2021 ) showed that sentiment correlates with the risk of a crash in the cryptocurrency market, what happens when there is a major drop in the market?

The May 2022 crash was not the first to occur in cryptocurrency markets. For instance, crashes occurred during 2017–2018 (Cross et al. 2021 ) and 2013–2014 (Bouri et al. 2017 ).

Herding behavior among investors is common in cryptocurrency crashes (Li et al. 2023 ). Examples of observed herding in cryptocurrency markets include a study by Vidal-Tomás et al. ( 2019 ), who presented evidence of herding in the lead up to the 2017–2018 cryptocurrency crash. Similarly, Shu et al. ( 2021 ) found proof that herding caused a bubble in Bitcoin in 2021. Bouri et al. ( 2019 ) studied herding over a longer period of time, finding it to be a persistent feature of cryptocurrency markets that ebbed and flowed over time. Raimundo et al. ( 2022 ) found that herding behavior was particularly prominent in cryptocurrency markets during periods of market stress. A typical approach to measuring sentiment throughout the literature is to find a source of relevant text, typically from social media, perform sentiment analysis on the text, and relate the results from the sentiment analysis to the price of a cryptocurrency (Abraham et al. 2018 ).

Although there are abundant studies on herding, not all such papers can be cited reasonably in this literature review. For example, although they did not study cryptocurrency specifically, Yousaf et al. ( 2018 ) found that Ramadan does not lead to herding behavior in the Pakistani stock market. Similarly, Yousaf and Ali ( 2020b ) studied spillovers between Bitcoin, Ethereum, and Litecoin before and during the COVID-19 pandemic by comparing spillovers during the periods October 3, 2018 to December 31, 2019, and January 1, 2020 to April 1, 2020 using a vector autoregressive asymmetric generalized autoregressive conditional heteroskedasticity (VAR-AGARCH) framework. Building on this, Yousaf and Ali ( 2020a ) also studied spillovers between Bitcoin, Ethereum, and Litecoin before and during the COVID-19 pandemic by comparing spillovers during the periods January 1, 2019 to December 31, 2019 and January 1, 2020 to April 22, 2020, using a VAR-DCC-GARCH framework. Building on this line of research, Yousaf and Ali ( 2021 ) examined spillovers between Bitcoin, Ethereum, and Litecoin, as well as the S &P 500, before and during the COVID-19 pandemic by comparing spillovers from May 21, 2019 to December 31, 2019 and from January 1, 2020 to May 21, 2020 using a BEKK-AGARCH framework. All three studies on cryptocurrency spillovers and COVID-19 have consistent results and, as of April 1, 2020, report a combined market cap of 76% for these three cryptocurrencies. The author of this paper highly recommends that the reader look at these three papers. Footnote 3

While much literature exists on how herding and sentiment affect prices, the literature on the opposite direction is sparse and considerable progress remains to be made regarding the effects of returns on sentiment. This study builds on the existing literature by providing empirical evidence that returns on financial investments affect investor sentiment, but, in the case of cryptocurrencies, in a non-homogeneous manner across different types of investors. Furthermore, this difference in behavior is tantamount to herding.

Theoretical motivation

Before presenting the data and methodology, it is important to explain why we should expect to see behavioral differences between cryptocurrency enthusiasts and traditional investors. While other assumptions are possible, the model presented here specifically assumes expected utility theory (Mas-Colell et al. 1995 ). This framework is ideal because it allows for a straightforward economic interpretation of behavioral expectations derived from the model, which helps build a simple intuition for why we should expect to see the empirical results that we do indeed observe, considering the following setup with two types of agents: type \(\theta _{E}\) represents cryptocurrency enthusiasts and type \(\theta _{I}\) represents traditional investors. All agents are utility maximizers that maximize the following intertemporal utility function in Eq.  1 , where agents derive utility from their wealth \((W_{t})\) and consumption \((C_{t})\) , which is proportional to wealth (i.e., \(C_{t} = \rho W_{t}\) , where \(0<\rho <1\) to ensure that some consumption and savings occur), and from investments in Bitcoin \((B_{t})\) . The parameters \(\alpha _{B}>0\) , \(\alpha _{W}>0\) , and \(\alpha _{C}>0\) describe the relative contributions of held Bitcoin, wealth, and consumption, respectively, to an individual’s total utility. The key assumption of the model is that the function \(f(\theta )\) is equal to \(\alpha _{B}\) for the cryptocurrency enthusiast type \(\theta = \theta _{E}\) ; however, for the traditional investor type \(\theta = \theta _{I}\) , the value of \(f(\theta )\) is zero. This finding implies that traditional investors gain utility from cryptocurrencies as their wealth increases. Thus, the parameter \(\alpha _{B}\) describes the additional value (cultural or otherwise) ascribed to Bitcoin held by cryptocurrency enthusiast type agents. Importantly, this study assumes that \(\alpha _{B}\le \alpha _{W}\) , such that \(\theta _{E}\) -type agents still aspire to grow their wealth.

Next, we describe the evolution of wealth by using Eq.  2 . Here, \(I_{t} = W_{t} - B_{t}\) represents traditional (non-cryptocurrency) investments. The parameters \(r_{I}\) and \(r_{B}\) denote the rates of return on traditional investments and Bitcoin, respectively. Importantly, this study assumes that the rate of return \(r_{I}\) is equal to the highest rate of return available to investors; if Bitcoin has the highest rate of return available to a traditional investor, \(r_{I} = r_{B}\) . However, because this study focuses on investor behavior during the aftermath of the cryptocurrency crash of 2022, we assume that \(r_{i} > r_{B}\) for the remainder of this analysis.

By simply rearranging the terms and using the definitions of \(I_{t}\) and \(C_{t}\) , we obtain the following description of the change in investor wealth over time (Eq.  3 ):

Given this simple model, the natural questions are as follows: (1) How much wealth will be invested in Bitcoin by \(\theta _{E}\) -type investors? (2) How does the evolution of wealth differ between the two investor types?

To answer the first question, we set equal partial derivatives of utility with respect to wealth and Bitcoin (Eq.  4 ).

Rearranging the terms and using the definition of \(f(\theta )\) gives us the ratio of wealth invested in Bitcoin over time for each investor type:

As can be seen from Eq.  5 , the optimal proportion of Bitcoin holdings is equal to the relative utility a \(\theta _{E}\) -type investor obtains from holding Bitcoin compared to their utility from wealth in general.

Turning our attention to the second question, the goal is to describe the difference in \(\frac{\dot{W}}{W}\) between the two investor types. Equation  6 states the evolution of wealth for \(\theta _{I}\) -type investors, and 7 states the evolution of wealth for \(\theta _{E}\) -type investors. The " Appendix " provides the derivation of these results.

Taking the difference between these two equations, we find that traditional investors’ wealth grows at a faster rate than in (Eq.  8 ; derivation in the " Appendix ").

Thus, using a simple model, we show that cryptocurrency enthusiasts will experience a lower growth rate for wealth as a consequence of the utility they gain from holding Bitcoin. Given that changes in wealth can be reasonably expected to affect the sentiment embedded in relevant tweets, this derivation provides a formal justification for why we should expect to see changes in the sentiment of tweets among cryptocurrency enthusiasts in the aftermath of the cryptocurrency crash of 2022.

It is important to acknowledge that an expected utility framework is not the only way to motivate the empirical analysis in this study. The prospect theory is another means of framing this study. However, there is extensive value in establishing and deriving this expected utility model. Specifically, this study shows how non-financial factors, such as belonging to a community, can affect the utility-maximizing behavior of cryptocurrency enthusiasts. Essentially, while the cryptocurrency enthusiast’s position of holding crypto assets during a crash is not what a traditional investor would consider rational, it is rational from the perspective of a cryptocurrency enthusiast. This is important for policymakers when designing regulations for cryptocurrency markets.

Data and identification strategy

The data used in this study were obtained from Twitter. To identify a differential effect linking the cryptocurrency crash to changes in the sentiment of cryptocurrency enthusiasts relative to traditional investors, we ultimately need to quantify the relevant aspects of tweets using sentiment analysis. These relevant aspects of tweets are referred to as affective states in the sentiment analysis literature (Xie et al. 2021 ) as a “positive,” “negative,” “neutral,” and an aggregate or “compound” score. Once all tweets are assigned scores for each affective state, all tweets by a given user during each of the two time periods (before and after the cryptocurrency market crash) the scores are combined and averaged to give a mean score for each user and affective state pair. This dataset also contains the frequency of tweets made by each user before and after the cryptocurrency crash. Because the state of the cryptocurrency market itself is likely to affect investor sentiment, the price of Bitcoin is also included. Table 1 presents the summary statistics, and the process for generating these data is described below.

The data were generated using a four-step process. The first step was to curate a list of Twitter users for the potential treatment and control groups. This was done by searching for and collecting tweets containing at least one of a series of finance related terms including cryptocurrency specific terms such as “wagmi.” (An acronym for “We are all gonna make it,” “wagmi” is a rallying cry among herding-type cryptocurrency enthusiasts). This approach was chosen over other sample selection methods (e.g., the seed-based method proposed by Yang et al. ( 2015 )) because it allows for a straightforward classification of users. While many different clustering methods could have been used to (bi)partition the users, such methods rely on the full sample of users having already been determined and, therefore, are not appropriate for use here (Li et al. 2021 ). Footnote 4 Twitter was selected as a data source for several reasons. First, when the data for the study were collected, the Twitter API was freely accessible to researchers. Second, Twitter users tend to post frequently, with short yet expressive posts, which is an ideal combination for this study. Third, a body of literature exists on extracting a representative sample of users from Twitter for a given research purpose (Vicente 2023 ; Mislove et al. 2011 ).

Once the tweets were collected, the second step was to partition the users into the treated and control groups for the DID regression. The treated group; that is, herding-type cryptocurrency enthusiasts, was defined via the existence of herding-type cryptocurrency enthusiast-specific keywords in tweets. If a herding-type cryptocurrency enthusiast-specific keyword was present in any of a user’s tweets, the user was classified as “treated.” The remaining users were classified as “controls” and may be thought of as more traditional investors (i.e., investors who did not exhibit outward evidence of herding-type behavior in the cryptocurrency market prior to the cryptocurrency crash of May 2022). It is important to note that these users may still invest in cryptocurrencies; however, such investment decisions are no different from any other investment decision.

With a large number of tweets (614,116) collected and users classified into treated and control groups, the third step was to define the time periods for “pre” and “post” event (pre = January 1–April 30; post = June 1–August 31), representing before and after the initial cryptocurrency crash in May 2022. All of the tweets were classified accordingly, and tweets between these two periods were omitted from the study. This choice of period allows us to look at persistent changes in investor sentiment, as opposed to the more transient changes that may have been observed during the omitted intermediate period.

Finally, the various affective states considered in the study were quantified as “positive,” “neutral,” “negative,” and an aggregate measure. This was achieved by performing sentiment analysis of the tweets using the Sentiment Intensity Analyzer tool from the Natural Language Toolkit (NLTK) module in Python. Individual tweets were tokenized, lemmatized, and stemmed, and all usernames were removed from the tweets. The output of the Sentiment Intensity Analyzer tool consists of four scores that quantify the presence of specific affective states in tweets: “positive,” “neutral,” “negative,” and “compound” (an aggregate measure). The scores assigned to each affective state come from the unit interval [0, 1], with the exception of neutral sentiment, which is mapped to the interval \([-1,1]\) . For more details on performing sentiment analysis, please see Feldman ( 2013 ). For replicability, all code pertaining to this study is available from https://github.com/cat-astrophic/cryptocrash .

Relating these affective states to investor sentiment, if the cryptocurrency crash had negatively impacted investor sentiment, one consequence we might expect to observe would be relatively less positive sentiment expressed in tweets, meaning a greater prevalence of negative and neutral tweets in the post-crash period. This is because typically positive herding-type cryptocurrency enthusiasts may have either a higher sentiment baseline or employ a more positively measured diction relative to other users, including other herding-type cryptocurrency enthusiasts. Tweets by these users may become more “neutral,” meaning that although they no longer express explicitly positive sentiment on Twitter, they do not necessarily express explicitly negative sentiment. A practical example of this would be unimpassioned appeals within the herding-type investor community to hold a course that does not explicitly express dismay at the current state of the cryptocurrency market.

In addition to these four broad affective states, more nuanced emotion-specific variables were created in the same manner as the data for the four broad affective states, namely “anger,” “disgust,” “fear,” “sadness,” “negativity,” “surprise,” ‘trust,” “joy,” and “positivity.” As was the case for the broader affective states discussed above, these emotions were calculated using the NLTK module in Python and followed the data preprocessing methods outlined above.

Finally, we acquired data on the number of tweets that each user tweeted during each period. These data are included because significant results indicate that cryptocurrency enthusiasts changed not only their sentiment but also their behavior regarding Twitter usage.

Identification strategy

Given the nature of the research question and the data, two sets of ID models were used to determine whether cryptocurrency enthusiasts behaved fundamentally differently from traditional investors. The standard interpretation of the DID estimator is the average treatment effect of the treated units (ATT). However, in the context of this study, where the treated units are cryptocurrency enthusiasts and the control units are traditional investors, this tells us whether there is a differential response to the cryptocurrency crash between the two groups. If so, these two groups behave fundamentally differently from one another and thus represent two distinct types of investors.

In the first model in each set, we estimated the effect of cryptocurrency crashes on the mean tweet sentiment across all Twitter users in the sample. Here, the model is specified as shown in Eq.  9 , where \(Y^{s}_{x,i,t}\) is the mean score for affective state s for tweet x by user i in period t , \(Treated_{x}\) is a dichotomous variable indicating whether tweet x was written by a herding-type cryptocurrency enthusiast, \(Post_{t}\) is a dichotomous variable indicating whether the day t of tweet x was before or after the cryptocurrency crash, \(\delta _{i}\) is a user-specific fixed effect, \(\epsilon _{i,t}\) is the error term, and \(\beta\) is the parameter of interest. Separate regressions were run for each affective state, and this approach was used for both the sentiment and emotion analyses.

To provide additional support for these regressions, we estimate the regression shown in Eq.  10 , where we examine the user-level average values for each affective state in each of the two time periods. \(\beta\) remains the parameter of interest.

To test for another potential change in behavior among herding-type cryptocurrency enthusiasts, another DID model was specified, as shown in Eq.  11 . Here, \(Tweets_{i,t}\) becomes the dependent variable; otherwise, we follow the specification from Eq.  10 and are again interested in estimating the parameter \(\beta\) . Changes in the frequency at which a tweet can be an additional indicator of sentiment have not been previously considered in the literature.

Finally, these models cannot be presented without a discussion of endogeneity. When running a DID model, a key assumption for causality is that there can be no self-selection in the treatment. In this study, individual investors can choose whether they are willing to participate in a broader cryptocurrency culture (type \(\theta _{E}\) investors in the theoretical model). However, we do not estimate the causal effect of a policy but rather exploit an exogenous market shock to directly observe the differential responses of two distinct groups of investors to this shock. Therefore, the only potential endogeneity concern in this paper lies in potential correlations between the use of the term “wagmi” in tweets and the sentiment of the words surrounding “wagmi” in the tweet. However, as will be shown in the following section, if anything, this study underestimates the magnitude of the differential response to the May 2022 cryptocurrency crash between cryptocurrency traders and traditional investors.

This section presents and discusses the regression results and textual evidence suggestive of herding behavior. First, we focus on the results of the tweet- and user-level regressions for broad affective states (i.e., compound, positive, negative, and neutral). The results are presented in Tables 2 and 3 , respectively. Next, we take a more nuanced look at these affective states using the results from the tweet- and user-level regressions for the presence of specific emotions in the tweets. The results are presented in Tables 4 and 5 , respectively. Third, we address the results of the regressions on the frequency at which users tweet (see Table 6 ). Finally, we analyze the specific textual content of the tweets and provide evidence of herding among herding-type investors but not among traditional investors.

Broad affective states

Beginning with the regressions for the four broad affective states (Tables 2 and 3 ), cryptocurrency enthusiasts saw a decrease and increase in negative sentiments and neutral sentiments in their tweets, respectively. The increase in neutral sentiment should not be surprising as it can be explained by a more subdued discourse among cryptocurrency enthusiasts that contains both increases in negative sentiment attributable to the cryptocurrency crash and the deliberately positive, collectivist, and perhaps even dogmatic, “wagmi” mantra. Conversely, the decrease in negative sentiment might be surprising given the negative nature of the cryptocurrency crash and its impact on cryptocurrency enthusiasts. Given that the cryptocurrency enthusiast community made a deliberate, collective effort to stay positive (“wagmi”), a decrease in negative sentiment makes sense. Since “wagmi” is a deliberate positive rallying cry, its use appears to have offset a decline in positive sentiment, leading to statistically insignificant results for both positive sentiment and the compound score. This is particularly important because the decrease in the price of Bitcoin (assuming it is correlated with investor sentiment), may have been partially offset by the collective effort to hold Bitcoin, despite the financial implications of the herding-type cryptocurrency enthusiasts, thus validating the model presented in Sect. " Theoretical Motivation ".

These results suggest that cryptocurrency enthusiasts behave in a fundamentally different manner compared to traditional investors. From an economic policy standpoint, this is particularly important because it suggests the potential for herding. (The next section of this paper explores herding further.) A market flooded with participants who engage in herding behavior is more likely to have bubbles and eventually runs. As the cryptocurrency market is still largely unregulated, there is a need for regulations to prevent runs. While this would seem a dramatic claim in other contexts, a run in the cryptocurrency market occurred between the beginning of this study and its publication. Footnote 5

In the user-level regressions (Table 3 ), we can see that cryptocurrency enthusiasts are overall more positive, less negative, and less neutral and have higher compound scores than traditional investors. The statistical insignificance of the treated indicator in the tweet-level regressions suggests that user-level fixed effects account for the differences between the two user types. We also find that the change in the price of the Bitcoin variable was statistically significant and negative for neutral sentiment. This suggests that increased emotionality was present among finance-oriented Twitter users when Bitcoin prices went up.

These results confirm and build on the findings of previous studies (e.g., (Baker and Wurgler 2007 )) that link investor sentiment to market conditions. In particular, these results expand the literature by showing that investor sentiment responds to market conditions and that there is profound heterogeneity in responses to changes in market conditions across different types of investors.

Next, we focus on the emotion-specific regression results (Tables 4 and 5 ), which indicate that cryptocurrency enthusiasts experienced significant changes in their expression of two specific emotions: surprise and joy. Specifically, cryptocurrency enthusiasts expressed less surprise and joy in their tweets than more traditional investors. These results, especially the decrease in the expression of joy, suggest that there is indeed some fundamental difference between how cryptocurrency entrants and traditional investors invest in, utilize, and experience cryptocurrencies such as Bitcoin. The decrease in the expression of surprise is interesting because it suggests that cryptocurrency enthusiasts may have begun to anticipate the persistence of cryptocurrency crashes. This implies that cryptocurrency enthusiasts possibly held Bitcoin despite poor expectations of its future performance. These results strongly support the model provided in Sect. " Theoretical Motivation " in which cryptocurrency enthusiasts derive extra utility simply from holding cryptocurrency assets.

Similar to the regressions for the four broad affective states, the user-level regressions suggest stark differences in how the two groups communicate. Cryptocurrency opportunists appear to express less anger, disgust, fear, surprise, trust, joy, and positivity and tend to express more sadness and negativity. Finally, changes in the price of Bitcoin lead to a decrease in disgust and fear, which, in turn, results in an increase in trust. These results confirm the existing literature on the psychology of cryptocurrency enthusiasts.

As in the previous subsection, these results confirm and build on the literature that links investor sentiment and market conditions. In particular, we find that (1) the cryptocurrency crash caused a decrease in the expression of surprise and joy among herding-type cryptocurrency enthusiasts and (2) herding-type cryptocurrency enthusiasts have a distinct emotional profile compared to traditional investors. Cryptocurrency enthusiasts are prone to express themselves in sadder and more negative ways, with less trust, joy, anger, disgust, fear, and surprise than traditional investors. This suggests that a certain type of person (i.e., a certain set of personality traits) self-selects into a herding-type cryptocurrency group.

Tweet frequency

The final set of regressions examines the actual tweet behavior of users by studying the frequency of their tweets. Here, we see that cryptocurrency enthusiasts increased the frequency with which they tweet by over 100 tweets during the post-cryptocurrency crash period compared to the control group, which translates to an increase of more than one tweet per day relative to that of traditional investors. As shown in Table 6 , these results are highly consistent across the specifications, demonstrating their robustness to the sentiments contained in the tweets. Moreover, they suggest that behavioral changes in cryptocurrency enthusiasts may be numerous and correlated as we found changes in both sentiment/emotionality and tweet frequency attributed to the same event. This builds on the existing literature by providing the first evidence that market conditions differentially affect investors’ use of social media when discussing investment-related topics.

Textual evidence of herding

In this section, we present evidence suggesting the presence of herding among cryptocurrency enthusiasts by analyzing the specific textual content of tweets. To this end, we apply a manually augmented hierarchical clustering method to the most frequent terms found in the tweets using the following process.

First, after following the same data preprocessing steps outlined in the methodology for performing sentiment analysis; that is, tokenization, lemmatization, and stemming, a bag of (unique) words was created for each group. Once the unique words were identified, the frequency with which they appeared in the tweets was computed, and words appearing in at least 1/1000 tweets were identified. In total, 19 words are associated with traditional investors, whereas 57 are associated with cryptocurrency enthusiasts.

To provide evidence of herding, these frequent terms were classified using a hierarchical clustering method from SciPy in Python (scipy.cluster.hierarchy). This algorithm clusters terms based on their co-occurrence in tweets. The results (classes) of this algorithm were then manually updated to the final classes listed in Table 7 .

These classifications support the notion of herding for two primary reasons. First, the disjoint nature of terms between the two groups of investors suggests that cryptocurrency enthusiasts represent their own “clique” within the online investing community. Second, across the classes for the terms commonly used by cryptocurrency enthusiasts, clear themes emerge as the dominating discourse. Class 1, a class of terms related to cryptocurrencies, is not surprising and does not necessarily imply the existence of herding behavior. However, Classes 2 and 3 suggest otherwise. Regarding Class 2, the fact that the NFT bubble was not observed as a common topic of discourse among traditional investors but was important enough to constitute its own class among cryptocurrency enthusiasts is qualitative evidence suggesting that cryptocurrency enthusiasts engage in herding behavior, at least regarding NFTs. Class 3 (i.e., the (“wagmi” class) suggests that this behavior extends to cryptocurrencies as well since it is, by definition, representative of the discourse related to holding cryptocurrency despite the nature of the market at that time. This is direct evidence of herding behavior among cryptocurrency enthusiasts but not traditional investors in the cryptocurrency market in the aftermath of the cryptocurrency crash in May 2022.

Finally, other important trends became apparent during the analysis. First, cryptocurrency enthusiasts use more current Internet vocabulary than traditional investors do. Examples include the use of emojis; no emojis were among the most frequent terms used by traditional investors, while five emojis appeared among the most common terms used by cryptocurrency enthusiasts. While this certainly reflects a significant cultural difference between the two groups, it could also reflect meaningful demographic differences. These differences and the elevated risk-seeking behavior observed among cryptocurrency enthusiasts fits the social identity model of risk-taking (Cruwys et al. 2021 ).

The second theme that emerged is the gendered nature of online investment communities. “He,” “bro,” “guy,” “ser,” “fam,” and “they,” were all among the most commonly used words used by the two groups in this study, yet no female-gendered words (e.g., “she”) appeared among the most common words. This suggests that online investment communities are largely male-dominated.

The last trend that emerged was within the community of cryptocurrency enthusiasts, concerns the use of the term “ser.” This term is used as a synonym for “sir,” but it also has a racist second meaning; it is used to mock Indian and East Asian cryptocurrency enthusiasts for their relatively more frequent use of “sir” in the online discourse (Limbu 2022 ). To emphasize the toxic nature of the online cryptocurrency enthusiast community, the source cited on the previous line, an article published on the popular cryptocurrency news site Blockchain.News, classified this behavior as “trolling,” minimizing the iniquity.

In the aftermath of the cryptocurrency crash of 2022, investor sentiment among cryptocurrency generators has changed relative to that of traditional investors, specifically, an increase in neutral sentiment and, surprisingly, a decrease in negative sentiment. This is particularly significant as the deliberate, collectivist approach to publicly displaying positivity and holding Bitcoin (“wagmi”) could have mitigated the magnitude of the crash to a small extent. These findings are also important as they provide further support that cryptocurrency enthusiasts will hold on to a cryptocurrency even when they could earn better returns by investing elsewhere. These results validate the model described in Sect. " Theoretical Motivation " of this paper. In summary, cryptocurrency enthusiasts and traditional investors exhibit visibly distinct behavioral patterns.

In addition to changes in investor sentiment, two other changes were observed in the behavior of cryptocurrency enthusiasts. First, there were changes in the specific emotional content of their tweets, specifically a decrease in surprise and joy. Second, the frequency at which these cryptocurrency enthusiasts tweet increased in the aftermath of the cryptocurrency crash of 2022, suggesting that public displays of loyalty to Bitcoin and/or the Bitcoin community are an important cultural practice that manifests itself in herding behavior. This reinforces the notion that herding and other collectivist behaviors are central to cryptocurrency community membership.

From a policy perspective, cryptocurrency markets must be regulated. The prevalence of herding behavior among cryptocurrency enthusiasts is not only present but also a core cultural component in this community. This could lead to the formation of bubbles and subsequent runs. As stated in the body of this paper, runs are not an abstract and unlikely concern but an observed consequence of this behavior. Given the gradually increasing role of cryptocurrencies in traditional portfolios, a failure to regulate the cryptocurrency market could lead to spillovers to other markets and negatively impact all investors. Only months after the cryptocurrency crash studied in this paper, the FTX Exchange collapsed, a bank run occurred, multiple leading global figures in the cryptocurrency market (as of May 2022) are now in prison, illicit activities have been financed, and taxpayers have lost billions of dollars bailing out cryptocurrency investors.

Another implication of this study is that we can identify potential herding-type cryptocurrency investors via social media. As researchers continue to study herding and other disconcerting phenomena in markets, this can be useful for various reasons, including targeting individuals for surveys or online experiments on social media. Additionally, the ability to identify herding investors on social media could allow targeted nudges designed to prevent herding in markets and increase market efficiency.

This study has two limitations. First, the herding results are largely, although not exclusively, qualitative. Causal analysis of herding behavior would be an excellent extension of this study. Second, the main results from the DID models may actually understate the true effects of the cryptocurrency crash on cryptocurrency enthusiasts as they deliberately emphasized positivity (“wagmi”), which could have impacted the sentiment scores assigned to tweets and users in this data set. An econometric consequence is a potential downward bias in the point estimates for negativity and a potential upward bias in the point estimates for positivity. If these biases are present, this further confirms the conclusions drawn in this study, and further analyses of this (and other related) phenomenon would be valuable extensions of this research. One possible way to expand the scope of this analysis is to collect data from a broader set of source materials.

So far, the importance of the results has been framed in the context of fundamental differences between traditional investors and cryptocurrency enthusiasts regarding their perceptions of, experiences with, and motivations for investing in cryptocurrencies and the consequential regulatory implications. While this consequence is incredibly important, there is another potential consequence of these results. While the literature is not conclusive regarding the sentiment of social media posts and mental health, if such a relationship exists, these results suggest that the mental health of cryptocurrency enthusiasts can be linked to the state of cryptocurrency markets. This is concerning for both financial and humanitarian reasons.

Availability of data and materials

All data and code is available at https://github.com/cat-astrophic/cryptocrash .

Sam Bankman-Fried and Do Kwon are the two most notable, but many other fraud schemes have been caught across the world.

I’d like to thank a particularly dedicated reviewer for suggesting these four papers.

I’d like to thank another particularly dedicated reviewer for suggesting these four papers and bringing my attention to all that they imply.

This paper was also suggested by one of the afore-thanked reviewers.

This is a reference to the FTX collapse and ensuing run.

Abraham J, Higdon D, Nelson J, Ibarra J (2018) Cryptocurrency price prediction using tweet volumes and sentiment analysis. SMU Data Sci Rev 1(3):1

Google Scholar  

Ahn Y, Kim D (2021) Emotional trading in the cryptocurrency market. Financ Res Lett 42:101912

Article   Google Scholar  

Al Shehhi A, Oudah M, Aung Z (2014) Investigating factors behind choosing a cryptocurrency. In 2014 IEEE international conference on industrial engineering and engineering management , pp 1443–1447. IEEE

Anastasiou D, Ballis A, Drakos K (2021) Cryptocurrencies’ price crash risk and crisis sentiment. Financ Res Lett 42:101928

Ante L (2023) How elon musk’s twitter activity moves cryptocurrency markets. Technol Forecast Soc Chang 186:122112

Avery C, Chevalier J (1999) Identifying investor sentiment from price paths: the case of football betting. J Bus 72(4):493–521

Baker M, Wurgler J (2006) Investor sentiment and the cross-section of stock returns. J Financ 61(4):1645–1680

Baker M, Wurgler J (2007) Investor sentiment in the stock market. J Econ Perspect 21(2):129–152

Baker M, Wurgler J, Yuan Y (2012) Global, local, and contagious investor sentiment. J Financ Econ 104(2):272–287

Blasco N, Corredor P, Satrústegui N (2022) The witching week of herding on bitcoin exchanges. Financ Innov 8(1):1–18

Bouri E, Gupta R, Roubaud D (2019) Herding behaviour in cryptocurrencies. Financ Res Lett 29:216–221

Bouri E, Jalkh N, Molnár P, Roubaud D (2017) Bitcoin for energy commodities before and after the december 2013 crash: diversifier, hedge or safe haven? Appl Econ 49(50):5063–5073

Cary M (2021) Down with the #dogefather: Evidence of a cryptocurrency responding in real time to a crypto-tastemaker. J Theor Appl Electron Commer Res 16(6):2230–2240

Chaim P, Laurini MP (2019) Is bitcoin a bubble? Phys A 517:222–232

Cross JL, Hou C, Trinh K (2021) Returns, volatility and the cryptocurrency bubble of 2017–18. Econ Model 104:105643

Cruwys T, Greenaway KH, Ferris LJ, Rathbone JA, Saeri AK, Williams E, Parker SL, Chang MX, Croft N, Bingley W et al (2021) When trust goes wrong: a social identity model of risk taking. J Pers Soc Psychol 120(1):57

Da Z, Engelberg J, Gao P (2015) The sum of all fears investor sentiment and asset prices. Rev Financ Stud 28(1):1–32

da Gama Silva PVJ, Klotzle MC, Pinto ACF, Gomes LL (2019) Herding behavior and contagion in the cryptocurrency market. J Behav Exp Financ 22:41–50

De Blasis R, Galati L, Webb A, Webb RI (2023) Intelligent design: stablecoins (in) stability and collateral during market turbulence. Financ Innov 9(1):85

Dodd N (2018) The social life of bitcoin. Theory, Culture Soc 35(3):35–56

Feldman R (2013) Techniques and applications for sentiment analysis. Commun ACM 56(4):82–89

Foley S, Karlsen JR, Putniņš TJ (2019) Sex, drugs, and bitcoin: how much illegal activity is financed through cryptocurrencies? Rev Financ Stud 32(5):1798–1853

Frost J, Shin HS, Wierts P (2020) An early stablecoin? the bank of amsterdam and the governance of money

Gao Y, Zhao C, Sun B, Zhao W (2022) 0effects of investor sentiment on stock volatility: new evidences from multi-source data in china’s green stock markets. Financial Innovation 8(1):1–30

Giudici G, Milne A, Vinogradov D (2020) Cryptocurrencies: market analysis and perspectives. J Ind Bus Econ 47(1):1–18

Haykir O, Yagli I (2022) Speculative bubbles and herding in cryptocurrencies. Financ Innov 8(1):1–33

Judge B, Eichengreen B, Zysman J (2023) The mirage of decentralized finance. Available at SSRN 4459315

Kallinterakis V, Wang Y (2019) Do investors herd in cryptocurrencies-and why? Res Int Bus Financ 50:240–245

Kou G, Dinçer H, Yüksel S, Alotaibi FS (2024) Imputed expert decision recommendation system for qfd-based omnichannel strategy selection for financial services. Int J Inf Technol Decis Mak 23(01):141–170

Kou G, Olgu Akdeniz Ö, Dinçer H, Yüksel S (2021) Fintech investments in european banks: a hybrid it2 fuzzy multidimensional decision-making approach. Financ Innov 7(1):39

Kou G, Pamucar D, Dinçer H, Yüksel S (2023) From risks to rewards: a comprehensive guide to sustainable investment decisions in renewable energy using a hybrid facial expression-based fuzzy decision-making approach. Appl Soft Comput 142:110365

Kou G, Xu Y, Peng Y, Shen F, Chen Y, Chang K, Kou S (2021) Bankruptcy prediction for smes using transactional data and two-stage multiobjective feature selection. Decis Support Syst 140:113429

Lamon C, Nielsen E, Redondo E (2017) Cryptocurrency price prediction using news and social media sentiment. SMU Data Sci Rev 1(3):1–22

Lee S, Lee J, Lee Y (2023) Dissecting the terra-luna crash: evidence from the spillover effect and information flow. Financ Res Lett 53:103590

Li T, Chen H, Liu W, Yu G, Yu Y (2023) Understanding the role of social media sentiment in identifying irrational herding behavior in the stock market. Int Rev Econ Financ 87:163–179

Li T, Kou G, Peng Y, Philip SY (2021) An integrated cluster detection, optimization, and interpretation approach for financial data. IEEE Trans Cybern 52(12):13848–13861

Limbu A (2022) Slangs used in crypto space

Liu B et al (2010) Sentiment analysis and subjectivity. Handbook of natural language processing. Chapman and Hall/CRC, Boca Raton, pp 627–666

Liu J, Makarov I, Schoar A (2023) Anatomy of a run: The terra luna crash. Technical report, National Bureau of Economic Research

Mai F, Shan Z, Bai Q, Wang X, Chiang RH (2018) How does social media impact bitcoin value? A test of the silent majority hypothesis. J Manag Inf Syst 35(1):19–52

Mas-Colell A, Whinston MD, Green JR et al (1995) Microeconomic theory, vol 1. Oxford University Press, New York

Mattke J, Maier C, Reis L, Weitzel T (2021) Bitcoin investment: a mixed methods study of investment motivations. Eur J Inf Syst 30(3):261–285

Mislove A, Lehmann S, Ahn Y-Y, Onnela J-P, Rosenquist J (2011) Understanding the demographics of twitter users. In: Proceedings of the international AAAI conference on web and social media. vol 5, pp 554–557

Obreja DM (2022) The social side of cryptocurrency: exploring the investors’ ideological realities from romanian facebook groups. New Med Soc 26(5):14614448221092028

Ooi SK, Ooi CA, Yeap JA, Goh TH (2021) ‘Embracing bitcoin: users’ perceived security and trust. Qual Quant 55(4):1219–1237

Philippas D, Rjiba H, Guesmi K, Goutte S (2019) Media attention and bitcoin prices. Financ Res Lett 30:37–43

Phillips RC, Gorse D (2017) Predicting cryptocurrency price bubbles using social media data and epidemic modelling. In: 2017 IEEE symposium series on computational intelligence (SSCI), pp 1–7. IEEE

Raimundo G Jr, Palazzi RB, Tavares RdS, Klotzle MC (2022) Market stress and herding: a new approach to the cryptocurrency market. J Behav Financ 23(1):43–57

Rognone L, Hyde S, Zhang SS (2020) News sentiment in the cryptocurrency market: an empirical comparison with forex. Int Rev Financ Anal 69:101462

Seok SI, Cho H, Ryu D (2019) Firm-specific investor sentiment and daily stock returns. N Am J Econ Financ 50:100857

Shu M, Song R, Zhu W (2021) The 2021 bitcoin bubbles and crashes-detection and classification. Stats 4(4):950–970

Starkey AJ, Keane CR, Terry MA, Marx JH, Ricci EM (2013) Financial distress and depressive symptoms among African American women: identifying financial priorities and needs and why it matters for mental health. J Urban Health 90(1):83–100

Vicente P (2023) Sampling twitter users for social science research: evidence from a systematic review of the literature. Qual Quant 57:1–41

Vidal-Tomás D, Ibáñez AM, Farinós JE (2019) Herding in the cryptocurrency market: CSSD and CSAD approaches. Financ Res Lett 30:181–186

Xie H, Lin W, Lin S, Wang J, Yu L-C (2021) A multi-dimensional relation model for dimensional sentiment analysis. Inf Sci 579:832–844

Xu H-C, Zhou W-X (2018) A weekly sentiment index and the cross-section of stock returns. Financ Res Lett 27:135–139

Yang SY, Mo SYK, Liu A (2015) Twitter financial community sentiment and its predictive relationship to stock market movement. Quant Financ 15(10):1637–1656

Yousaf I, Ali S (2020) The covid-19 outbreak and high frequency information transmission between major cryptocurrencies: evidence from the var-dcc-garch approach. Borsa Istanbul Rev 20:S1–S10

Yousaf I, Ali S (2020) Discovering interlinkages between major cryptocurrencies using high-frequency data: new evidence from covid-19 pandemic. Financ Innov 6(1):1–18

Yousaf I, Ali S (2021) Linkages between stock and cryptocurrency markets during the covid-19 outbreak: an intraday analysis. Singap Econ Rev. https://doi.org/10.1142/S0217590821470019

Yousaf I, Ali S, Shah SZA (2018) Herding behavior in Ramadan and financial crises: the case of the Pakistani stock market. Financ Innov 4(1):1–14

Yue W, Zhang S, Zhang Q (2021) Asymmetric news effects on cryptocurrency liquidity: an event study perspective. Financ Res Lett 41:101799

Zhang S, Zhou X, Pan H, Jia J (2019) Cryptocurrency, confirmatory bias and news readability-evidence from the largest Chinese cryptocurrency exchange. Account Financ 58(5):1445–1468

Zhou G (2018) Measuring investor sentiment. Annu Rev Financ Econ 10:239–259

Download references

Acknowledgements

I would like to thank the reviewers for the information they shared throughout the review process.

No external funding was used in this research.

Author information

Authors and affiliations.

Department of Agricultural and Applied Economics, Virginia Tech, 250 Drillfield Drive, Blacksburg, VA, 24061, USA

Michael Cary

You can also search for this author in PubMed   Google Scholar

Contributions

All work is the responsibility of Michael Cary.

Corresponding author

Correspondence to Michael Cary .

Ethics declarations

Ethics approval and consent to participate.

Not applicable.

Consent for publication

I consent to publication in Financial Innovation.

Competing interests

No competing interests are declared

Additional information

Publisher's note.

Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.

Derivation of Eq.  6

Derivation of eq.  7, derivation of eq.  8.

First, we derive the difference between \(\frac{\dot{W}_{I}}{W_{I}}\) and \(\frac{\dot{W}_{E}}{W_{E}}\) as

All that remains to be shown is that the difference is positive. First, we consider the numerator. As \(\rho > 0\) , the term \(1+\rho\) is positive. Similarly, because \(\alpha _{B}>0\) and \(\alpha _{W}>0\) , ratio \(\frac{\alpha _{B}}{\alpha _{W}}\) is positive. Finally, assuming that \(r_{I} > r_{B}\) , the difference \(r_{I}-r_{B}\) is positive. This means that the numerator is the product of three positive terms and is therefore positive.

Moving on to the denominator and following an argument similar to that showing that the numerator is positive, we must show that \(1+r_{I} > \left( \frac{\alpha _{B}}{\alpha _{W}}\right) (r_{I}-r_{B})\) . Since \(0<(r_{I}-r_{B})<r_{I}\) , and since \(0\le \left( \frac{\alpha _{B}}{\alpha _{W}}\right) \le 1\) it follows that \(1+r_{I}>r_{I}\ge \left( \frac{\alpha _{B}}{\alpha _{W}}\right) (r_{I}-r_{B})\) and thus the claim that \(\frac{\dot{W}_{I}}{W_{I}} - \frac{\dot{W}_{E}}{W_{E}} > 0\) is verified.

Rights and permissions

Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made. The images or other third party material in this article are included in the article’s Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article’s Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http://creativecommons.org/licenses/by/4.0/ .

Reprints and permissions

About this article

Cite this article.

Cary, M. Herding and investor sentiment after the cryptocurrency crash: evidence from Twitter and natural language processing. Financ Innov 10 , 142 (2024). https://doi.org/10.1186/s40854-024-00663-x

Download citation

Received : 20 June 2023

Accepted : 20 August 2024

Published : 02 September 2024

DOI : https://doi.org/10.1186/s40854-024-00663-x

Share this article

Anyone you share the following link with will be able to read this content:

Sorry, a shareable link is not currently available for this article.

Provided by the Springer Nature SharedIt content-sharing initiative

• Cryptocurrency
• Investor sentiment
• Natural language processing
• Sentiment analysis

JEL classification