Paper: Ethereum Tokenomics Insights for Web3 Entrepreneurs, a Quantitative Study
Authors: Zishan Ashraf Mohammad; Nicolas Harkiolakis, PhD
Date: January 26
Estimated Reading Time: 24 minutes
This study examines how Ethereum’s token design parameters are associated with ETH price dynamics over a 52 month period from August 2021 to September 2025. Using a quantitative, non-experimental correlational design, the authors analyze bi-weekly data covering inflation through token unlocks, burned tokens, total value locked, gas fees, and governance concentration measured by the Gini coefficient. Spearman correlations and log-linear multiple regression are employed to test relationships between these variables and ETH price. Results show strong positive associations between ETH price and total value locked, as well as significant relationships with gas fees and token burns. Token unlocks exhibit a significant negative effect in the regression model, indicating supply-driven dilution effects. The Gini coefficient does not display statistical significance in relation to price. The findings suggest that Ethereum’s price behavior is closely linked to demand-utility metrics and supply adjustments embedded in protocol design.
Core insights
- Demand-utility dominance: Total value locked and gas fees demonstrate statistically significant positive associations with ETH price. The regression results indicate that liquidity retention and transaction activity are primary explanatory variables in price variation over the observed period.
- Supply contraction effects: Burned tokens show a positive and significant relationship with ETH price. This supports the view that the EIP-1559 fee burn mechanism links network activity to monetary contraction.
- Supply expansion pressure: Token unlocks exhibit a negative regression coefficient despite positive monotonic correlation. This indicates that while unlock events may coincide with broader market cycles, increased circulating supply exerts downward pressure when controlling for other variables.
- Governance concentration neutrality: The Gini coefficient of voting power does not show statistical significance with ETH price. Within the study window, wealth concentration appears unrelated to short-term or medium-term valuation changes.
- High explanatory power of network variables: The reduced log-linear model excluding Gini achieves adjusted R² of approximately 0.98. This suggests that variation in ETH price is largely associated with network activity metrics and supply mechanics within the specified timeframe.
The study frames Ethereum’s tokenomics around two interacting dimensions: demand-utility intensity and supply mechanics. Demand is operationalized through total value locked and gas fees, both reflecting on-chain economic activity. Supply is represented by token unlocks and burned tokens, capturing expansion and contraction forces within the monetary base. By modeling ln(price_t) as a function of ln(fees_t), ln(TVL_t), ln(unlocks_t), and ln(burned_t), the analysis interprets coefficients as elasticities, allowing proportional changes to be compared directly.
The strongest elasticity is associated with TVL (β ≈ 0.80). This implies that a 1 percent increase in TVL corresponds to an approximate 0.8 percent increase in ETH price, holding other factors constant. This raises a structural question: does TVL function primarily as a proxy for speculative capital cycles, or does it represent persistent staking commitment that constrains circulating supply? The paper treats TVL as a demand-utility anchor, but overlapping dynamics with staking and liquidity incentives may amplify cyclical volatility.
Gas fees (β ≈ 0.20) and burned tokens (β ≈ 0.15) both exhibit positive coefficients, reinforcing the interaction between usage and deflation. Because fees and burns are highly correlated, multicollinearity is present among network activity indicators. The regression nonetheless retains statistical significance, suggesting that network activity intensity is systematically associated with valuation. However, when fees and burns move almost in tandem, how stable are individual elasticity estimates across alternative specifications?
Token unlocks show a negative elasticity (β ≈ −0.22) in the log-linear regression, indicating that proportional increases in circulating supply reduce price when other factors are controlled. This result aligns with standard monetary intuition where circulating_supply = total_supply − locked_tokens, and increases in available tokens can dilute price. Yet the positive Spearman correlation between unlocks and price suggests that supply expansions often coincide with bullish phases. This divergence underscores the importance of multivariate modeling when interpreting token supply effects.
The absence of significance for the Gini coefficient suggests that governance concentration does not materially influence price within the observed period. One interpretation is that transparent on-chain ownership mitigates concentration risk perceptions. Another possibility is that price formation in liquid crypto markets is more sensitive to liquidity and activity variables than to governance dispersion metrics. Would alternative governance measures, such as effective voting participation rather than token balance concentration, yield different associations?
The adjusted R² of approximately 0.98 indicates that the selected variables explain most of the variance in ETH price over the sample. While the study does not claim causality, the magnitude of explanatory power suggests strong alignment between Ethereum’s token design parameters and market valuation. This reinforces the interpretation of Ethereum as a system where programmable burns, staking-based locking, and transaction-driven fees jointly structure both supply contraction and demand expansion dynamics across market cycles.
