Cryptocurrency Market as an Asymmetric Duopoly: Complementarity and Competition Between Bitcoin and Ethereum README.txt 1. DATASET OVERVIEW Title: Replication Data for: Cryptocurrency Market as an Asymmetric Duopoly: Complementarity and Competition Between Bitcoin and Ethereum Short description: This dataset provides the replication materials for the empirical analysis published in the article "Cryptocurrency Market as an Asymmetric Duopoly: Complementarity and Competition Between Bitcoin and Ethereum". It includes the analytical R script, derived tabular outputs, and figures that reproduce all tables and graphs presented in the article. The analysis covers daily closing prices and trading volumes for Bitcoin (BTC-USD) and Ethereum (ETH-USD) sourced from Yahoo Finance for the period from 1 January 2017 to 20 March 2026. The empirical framework tests the implications of Bertrand price competition and Cournot quantity competition models, and evaluates three hypotheses: Bitcoin price leadership (H1), Cournot quantity equilibrium (H2), and partial asset complementarity (H3). Keywords: Bitcoin; Ethereum; cryptocurrency market; asymmetric duopoly; Bertrand model; Cournot model; price leadership; market shares; complementarity; capital rotation; digital assets. Language: Ukrainian; English Unit of analysis and coverage: - Unit of analysis: cryptocurrency asset (Bitcoin, Ethereum). - Temporal coverage: 1 January 2017 – 20 March 2026 (daily data). - Analytical frequency: daily (raw) and monthly averages (regression inputs). - Cut-off date: 2026-03-20. Version: v1.0.0 Release date (ISO 8601): 2026-03-20 Update schedule: Ad hoc 2. CREATORS AND CONTACT Data creators: 1) Telnova, Hanna State University of Trade and Economics (Kyiv, Ukraine) Email: g.telnova@knute.edu.ua ORCID: 0000-0002-5724-7229 Funding: N/A 3. IDENTIFIERS AND REPOSITORY INFORMATION Internal identifier: CRYPTO-DUOPOLY-BTC-ETH-v1 Repository record (DSpace): [URL after deposit] Persistent identifier (Handle): [Handle after deposit] DOI: [DOI after deposit] Article DOI: [DOI of published article] 4. FILE INVENTORY AND FORMATS Files included: - crypto_duopoly_analysis.R (main analytical script) - tables/Table1_yearly_shares.csv (Table 1 of the article: annual market shares) - tables/Table2_Bertrand_models.csv (Table 2 of the article: Bertrand model estimates) - tables/Table3_Cournot_model.csv (Table 3 of the article: Cournot model estimates) - tables/TableA1_correlations.csv (Appendix: correlation coefficients) - tables/DataS1_monthly_regression_data.csv (monthly regression inputs, 100 observations) - tables/DataS2_daily_data.csv (full daily dataset used in the analysis) - tables/SessionInfo.txt (R session information for reproducibility) - figures/Figure1_price_dynamics.png (Figure 1 of the article: synchronous price dynamics) - figures/Figure2_elasticity_ETH_BTC.png (Figure 2: ETH elasticity with respect to BTC) - figures/Figure3_elasticity_BTC_ETH.png (Figure 3: BTC elasticity with respect to ETH) - figures/FigureS1_normalized_price_log.png (supplementary: normalised price index, log scale) - figures/FigureS2_volume_shares.png (supplementary: volume-based market shares) - figures/FigureS3_Bertrand_panel.png (supplementary: Bertrand panel figure) - README.txt (this documentation) File formats: R; CSV; PNG; TXT Required software: R (version >= 4.2.0) with the following packages: quantmod, tidyverse, lubridate, fixest, ggplot2, patchwork, scales. Any spreadsheet tool supporting CSV (e.g., Microsoft Excel, LibreOffice) for tabular outputs. Any image viewer for PNG figures. Any text viewer for README and SessionInfo. 5. METHOD AND SOURCES Method: The empirical analysis proceeds in four analytical blocks: (1) Correlation analysis of logarithmic price levels and log-returns of BTC and ETH to distinguish short-term co-movement from long-term trend co-integration. (2) Annual market share analysis based on both price-based and volume-based metrics to assess the structure and stability of the duopoly. (3) Bertrand model testing via regression of logarithmic first differences: a levels model (ln_eth ~ ln_btc) and two cross-elasticity models (d_ln_eth ~ d_ln_btc; d_ln_btc ~ d_ln_eth). Asymmetry in estimated elasticities serves as the key test for price leadership. (4) Cournot model testing via regression of changes in ETH volume-based market share on the lagged BTC market share (d_share_eth ~ lag_share_btc). The sign and significance of the slope coefficient are interpreted in terms of competitive displacement versus complementary demand. All regressions are estimated using OLS via feols() from the fixest package (R). Price and volume data are downloaded directly from Yahoo Finance using the quantmod package with a fixed end date of 20 March 2026. Processing: Data are downloaded programmatically. No manual editing or imputation was applied. Missing observations are removed using drop_na(). Monthly averages are computed from daily data using floor_date() and group_by()/summarise(). Log transformations and first differences are computed within the script. No external data sources other than Yahoo Finance are used. Sources: Yahoo Finance. BTC-USD and ETH-USD daily closing prices and trading volumes. Retrieved via quantmod::getSymbols(). Period: 2017-01-01 to 2026-03-20. 6. ACCESS, LICENCE, AND REUSE Rights / licence: CC BY 4.0 Access information: Publicly available via the DSpace repository record; files can be downloaded directly from the record page. FAIR alignment note: This documentation is provided together with the dataset to support identification, understanding, and reuse, in line with the FAIR Guiding Principles and Ukrainian RDM guidance (MESU Methodological Recommendations on Research Data Management, 2024). The script is fully self-contained and downloads all source data automatically, enabling complete reproduction of all reported results. 7. CONTENT DESCRIPTION (VARIABLES) 7.1 DataS2_daily_data.csv — daily dataset - date: calendar date (YYYY-MM-DD) - btc_price: Bitcoin daily closing price (USD) - eth_price: Ethereum daily closing price (USD) - btc_volume: Bitcoin daily trading volume (USD) - eth_volume: Ethereum daily trading volume (USD) - year: calendar year (integer) - total_price: sum of btc_price and eth_price - share_btc_price: BTC price share within the two-asset system - share_eth_price: ETH price share within the two-asset system - total_volume: sum of btc_volume and eth_volume - share_btc_vol: BTC volume share within the two-asset system - share_eth_vol: ETH volume share within the two-asset system 7.2 DataS1_monthly_regression_data.csv — monthly regression inputs - month: first day of the month (YYYY-MM-DD) - btc_price: monthly average BTC closing price (USD) - eth_price: monthly average ETH closing price (USD) - btc_volume: monthly average BTC trading volume (USD) - eth_volume: monthly average ETH trading volume (USD) - ln_btc: natural logarithm of btc_price - ln_eth: natural logarithm of eth_price - d_ln_btc: first difference of ln_btc (monthly log-return of BTC) - d_ln_eth: first difference of ln_eth (monthly log-return of ETH) - total_vol: sum of btc_volume and eth_volume - share_btc: BTC volume share within the two-asset system - share_eth: ETH volume share within the two-asset system - lag_share_btc: one-period lag of share_btc - d_share_eth: first difference of share_eth 7.3 Table1_yearly_shares.csv — annual market shares (Table 1 of the article) - year: calendar year - btc_price_share: annual average BTC price share - eth_price_share: annual average ETH price share - btc_volume_share: annual average BTC volume share - eth_volume_share: annual average ETH volume share - diff_btc: difference between btc_price_share and btc_volume_share 7.4 Table2_Bertrand_models.csv — Bertrand regression results (Table 2) - Model: model label - Constant: estimated intercept - Coefficient: estimated slope coefficient - SE: standard error of the slope - R2: coefficient of determination - R2_adj: adjusted R-squared - N: number of observations 7.5 Table3_Cournot_model.csv — Cournot regression results (Table 3) - Variable: regressor name - Estimate: OLS point estimate - SE: standard error - t_value: t-statistic - p_value: p-value (scientific notation) - N: number of observations - RMSE: root mean squared error - R2_adj: adjusted R-squared - Interpretation: economic interpretation of the result 7.6 TableA1_correlations.csv — correlation coefficients - Показник: correlation measure label - Значення: correlation coefficient value Codes, abbreviations and conventions: - BTC: Bitcoin - ETH: Ethereum - USD: United States Dollar - log: natural logarithm (base e) - d_: prefix denoting first difference - ln_: prefix denoting natural logarithm - lag_: prefix denoting one-period lag - H1: hypothesis of Bitcoin price leadership - H2: hypothesis of Cournot quantity equilibrium - H3: hypothesis of partial asset complementarity - S1, S2: supplementary data files - FigureS1–S3: supplementary figures not included as main article figures 8. NOTES ON REPRODUCIBILITY To fully reproduce all tables and figures, run crypto_duopoly_analysis.R in R with an active internet connection. The script will automatically download the required data from Yahoo Finance and create the tables/ and figures/ output directories. All random seeds are not applicable (no stochastic elements). Results may differ marginally if Yahoo Finance revises historical data after 2026-03-20. END OF README