Regression-Scenario Modeling of Long-Term Capitalization Dynamics of the Global Cryptocurrency Market: Methodological Limitations of Extrapolative Approaches (Підприємництво і торгівля)

Summary

• A log-linear regression of crypto market capitalization for 2011–2025 yields a coefficient of determination of 0.866 and a slope significant at the 1 percent level.p.2
• The implied average annual capitalization growth is about 109% in the full-sample model versus about 61% in the shortened-sample model.p.5
• The upper bound of the 95% confidence interval at a two-year horizon exceeds any economically realistic estimate by several orders of magnitude.p.7
• Three alternative market scenarios up to 2028 were built from forecasts of ARK Invest, Bloomberg Intelligence, Pantera Capital and Standard Chartered.p.6

Key insights

1. The crypto market grew rapidly but extremely unevenly during 2011–2025, which complicates the use of traditional trend models.p.3
2. A model built on a long series with an explosive low-base growth phase overstates the slope relative to the model for the more mature period.p.5
3. The almost twofold gap in implied growth rates supports the hypothesis that market growth slows as the market matures.p.5
4. Leading researchers usually limit forecasting to one or two periods ahead and report confidence intervals with point estimates.p.6
5. Regression equations remain useful for confirming a significant long-term trend but must be accompanied by disclosed confidence intervals and scenario analysis.p.7
6. Future work should add nonlinear logistic growth specifications and factors such as a regulatory index and institutional capital flow indicators.p.8