Edited by TechFlow

Key Takeaways

• Funding Rates and Market Cap Exhibit Strong Explanatory Power: Lagged multivariate regression analysis shows that perpetual contract funding rates at centralized exchanges and changes in the overall cryptocurrency market capitalization statistically explain a significant portion of the variance in the weekly supply growth of USDC and USDT.

• Carry Spread Reinforces USDC’s On-Chain Asset Positioning: Incorporating the Carry Spread (the difference between on-chain lending rates and traditional risk-free US Treasury yields) into the model, alongside comparing on-chain and off-chain yields, reveals that when the spread widens, the market on average sees net minting of more USDC. This quantitatively confirms USDC’s unique ecosystem role as a core stablecoin deeply integrated with DeFi and on-chain yield generation.

Introduction: The Multidimensional Role of Stablecoins and Macro Mapping

Within public blockchain ecosystems, stablecoins serve as the core token vehicle representing the stable value of user assets. They can be deployed in decentralized lending protocols to earn yield, used for rapid swaps into other crypto assets, or directly paid for various on-chain services integrated into blockchain networks. More importantly, stablecoins profoundly reflect global macroeconomic conditions: in countries and regions facing high inflation or severe local currency depreciation, stablecoins have become the most mainstream tool for the masses to conduct cross-border remittances and access USD liquidity.

Historical Evolution: From the Silicon Valley Bank Crisis to 2026 Growth Stagnation

The development of stablecoins has traversed multiple distinct macro and crypto economic cycles. In March 2023, the failure and bank run of Silicon Valley Bank (SVB), caused by losses on long-term Treasury investments, impacted a portion of cash reserves held by Circle at the institution, leading to a severe depegging event for USDC. This incident triggered sustained capital outflows, causing the total on-chain supply of USDC to plummet from its 2022 peak of $56.6 billion down to a low of $27.9 billion in February 2024.

The turning point arrived in November 2024, when concentrated institutional compliance demands pushed stablecoin supply back onto a track of high growth. USDC's circulating scale successfully reclaimed its previous historical highs by March 2025. However, upon entering 2026, the expansion of stablecoin supply stagnated again, with some competing stablecoins even experiencing continuous contraction. What precisely caused such a massive divergence between the subdued 2026 period and the prosperous 2025? This prompted us to delve deeper into which underlying core variables are truly driving stablecoin supply growth, and how much expansion potential can be expected moving forward.

Econometric Model: Regression of Six Macro and On-Chain Lagged Variables

Not all stablecoins possess the same market cap or distribution breadth. Many smaller tokens' supply fluctuations often rely too heavily on isolated, sporadic factors, making it difficult to extract common driving logic. For instance, certain long-tail stablecoins are highly susceptible to manipulation by project teams through short-term liquidity mining subsidies, a type of growth that is typically unsustainable. Therefore, we first model the two foundational assets, USDC and USDT. Both boast total issuance scales exceeding $25 billion, enjoy deep market penetration across major public blockchains and centralized exchanges, and do not pay native interest or airdrop incentives directly to token holders themselves.

We selected weekly high-frequency time series data spanning full cycles over the past four years, adopting a one-week lag method, and introduced six candidate driver variables to explain the weekly supply changes of USDC and USDT:

• Perpetual Contract Funding Rates (Perp Funding): The annualized average funding rate levels for BTC and ETH perpetual contracts across the top five centralized exchanges by open interest;

• Competing Stablecoin Supply (Competitor Supply): The current total circulating supply of seven other major competitors in the market (BUSD, TUSD, GUSD, PYUSD, EURC, etc.)

• Non-Stablecoin Crypto Market Capitalization (Crypto Market Cap): The weekly percentage change in the total cryptocurrency market capitalization excluding stablecoins (CapMrktCurUSD);

• Two-Year US Treasury Yield Change (2Y Yield Δ): The weekly change difference in the 2-year constant maturity US Treasury yield published by the US FRED database;

• Federal Reserve Net Liquidity Growth (Net Liquidity Growth): The weekly percentage change in actual net liquidity released by the Fed, specifically calculated as the Fed's total assets minus the Treasury General Account (TGA) balance and Overnight Reverse Repurchase Agreement (ON RRP) volume;

• Dollar Nominal Index Returns (Dollar Return): The weekly percentage return of the broad nominal US Dollar Index published by FRED.

Preliminary baseline regression models indicate that these four indicators—perpetual funding rates, competitor supply, total crypto market cap, and the two-year Treasury yield change—are all highly significant at the 5% statistical significance level. Among them, perpetual funding rates exhibit the strongest single explanatory power, independently accounting for 13.3% of the variance in the combined supply changes of USDC and USDT. When the network-wide average annualized funding rates for BTC and ETH rise in one week, the average growth rate of stablecoin supply in the following week significantly exceeds long-term historical averages.

Coin Divergence: USDT Pro-cyclical Expansion vs USDC Counter-cyclical Safe Haven

Although USDC and USDT are highly similar in settlement functionality, they differ significantly in underlying network distribution, regional compliance preferences, and actual user profiles. After building independent regression models for each individual stablecoin, we found that these macro and micro variables impact them along completely different pathways.

First, higher perpetual funding rates exert an extremely strong positive driving effect on the supply expansion of both USDC and USDT for the following week. Controlling for other variables, funding rates are the single largest factor influencing both, independently explaining approximately 10% of USDC's and 4% of USDT's weekly growth variance. High funding rates intuitively reflect robust demand for long leverage in secondary markets. To maintain leveraged long positions and guard against liquidation risks, traders must urgently mint USDC and USDT off-chain to inject as margin collateral into exchanges. Currently, USDT serves as the pricing and margin asset for approximately 42% of trading pairs in the centralized futures market, acting as the primary lifeline for nearly all offshore exchanges; whereas USDC holds roughly 10% of the contract margin share, highly concentrated on decentralized and professional institutional derivatives platforms like Hyperliquid, dYdX, and Deribit, confirming its symbiotic relationship with on-chain DeFi.

Second, the returns of the non-stablecoin crypto market capitalization exhibit the most intriguing "divergent reversal." The model shows that the weekly price changes of the crypto market are significantly negatively correlated with USDC (explaining 4.6% of the variance), yet distinctly positively correlated with USDT (explaining 5.3% of the variance). In other words, when the broader market surges bullishly, USDT typically experiences accelerated net inflows through minting the following week, while USDC's minting pace noticeably slows down; conversely, when the market crashes, USDC faces a surge in net minting demand. This empirical test precisely captures their essential attributes: USDT acts as trading speculative ammunition that expands alongside rising risk appetite (Risk-On), whereas USDC serves as the safe-haven "dry powder" where funds urgently seek refuge and secure on-chain yield generation during deleveraging cycles.

Spread-Driven Dynamics: Quantitative Insights from Introducing the Carry Spread Variable

To further explore the guidance provided by on-chain native yields on capital flows over the past two years, we introduced a key new variable that has garnered significant institutional attention since 2023: the Carry Spread. The Carry Spread is strictly defined as the differential between the average annualized deposit yield (Supply APY) of stablecoins in the Ethereum Aave v3 lending market and the risk-free yield of US 3-month Treasury Bills.

Since complete high-frequency lending data for Aave v3 became fully available only after 2023, we shifted our analysis sample window to cover 2023 to the present. We re-ran the multivariate regression, first comparing the combined USDC and USDT model before and after incorporating the carry spread, followed by independent evaluations for each coin.

Empirical tests show a strong correlation coefficient of up to 0.61 between the carry spread and perpetual funding rates, as both essentially serve as proxy indicators for internal leverage tightness within the crypto system. In both the combined model and the independent USDT model, adding the carry spread yielded no incremental explanatory power; the model merely distributed the overlapping explained variance equally between the two.

However, for USDC, the carry spread provides exceptionally clear and powerful independent explanatory power at the 5% significance level. When combining funding rates with the carry spread, the two jointly explain up to 15.9% of the variance in USDC's weekly supply growth. This finding perfectly aligns with the on-chain micro-ecosystem: USDC is the most fundamental native collateral in the Ethereum DeFi LEGO world. When on-chain lending demand is robust and deposit APYs significantly outperform traditional risk-free Treasury yields, institutional arbitrage capital rapidly wire-funds fiat from traditional banking systems to mint USDC, which is then deposited into lending pools like Aave to lock in this risk-free carry spread. For USDC, the carry spread offers a purer, more direct leading signal for supply expansion than offshore contract funding rates.

Lagged Effect Verification: The Guidance Value of Leading Variables for Supply Decisions

In econometrics, utilizing lagged variables rather than contemporaneous ones not only effectively mitigates endogeneity interference caused by simultaneity but also demonstrates significant advantages in empirical goodness-of-fit.

Data comparisons indicate that with a one-period lag setup, the model explains 13% and 16% of the weekly supply variations for USDC and USDT, respectively, clearly outperforming contemporaneous non-lagged models. More practically valuable is that these core indicators precede the actual inflection points of stablecoin supply in the time series, strongly proving that "market leverage and yields drive stablecoin issuance," rather than a spurious correlation running in the reverse direction.

Nevertheless, regression results still reveal a substantial proportion of "unexplained variance." No single statistical model achieves a total explanatory power exceeding 25%, meaning that over 75% of stablecoin supply fluctuations remain constrained by heterogeneous factors the model failed to capture. These idiosyncratic factors include treasury management operations by stablecoin issuers, fee-waived minting and redemption subsidies within specific ecosystems, the technical integration progress of compliant fiat on/off-ramp channels among providers, and sudden regulatory policy shifts regarding stablecoin licensing across major global jurisdictions.

Conclusion & Reflection: Unexplained Variance Amid Regime Shifts

While statistical models cannot reconstruct every minute detail of stablecoin supply with 100% accuracy, this empirical study definitively confirms that perpetual funding rates, carry spreads, and broader crypto market capitalization trends constitute the most critical quantitative levers driving the medium-to-short-term volatility of USDC and USDT supplies.

The crypto financial system iterates at a breakneck pace, and the greatest challenge for both macro and micro models lies in adapting to fundamental "regime changes" at the base layer. If data from 2022 were fully incorporated, one would have to confront the systemic black swan shocks triggered by the sequential collapses of FTX, Celsius, and Three Arrows Capital (3AC). Models established in recent years, however, more closely align with the genuine norm of traditional institutional capital continuously flowing in via regulated channels. As sovereign wealth funds and Wall Street institutions further adopt blockchain infrastructure, deeper macro-financial factors will inevitably emerge, providing us with a clearer data compass for tracking the evolutionary trajectory of stablecoin supply.

Reference for Key Variables & Data Definitions

• Perp Funding (Perpetual Funding Rates): Based on Talos CM market data, calculating the annualized average rates for BTC and ETH perpetual contracts across the top five platforms by open interest such as Binance, OKX, and Bybit;

• Competitor Supply: Based on Talos CM Network Data Pro, tracking the on-chain circulating supply of mainstream stablecoins excluding the two giants;

• Carry Spread: The average APY of stablecoin lending in Aave v3 core pools minus the FRED 3-month Treasury Bill (DTB3) yield;

• Net Liquidity Growth: The difference between the Federal Reserve's total balance sheet size and the Treasury's TGA deposit accounts plus the Fed's overnight reverse repo volume.

Join the official Coincamps community:

X: https://x.com/coincamps

Telegram: https://t.me/coin_camps