💰 quantitative finance

Tokens All the Way Down: A Money View of Decentralized Finance

This paper applies the Money View framework to reveal that Decentralized Finance operates as a layered credit hierarchy mirroring traditional banking, where each additional layer of tokenized claims increases systemic leverage and yield premiums while introducing compounding dependency risks that are particularly pronounced during crises.

Original authors: Wenbin Wu

Published 2026-03-03

📖 6 min read🧠 Deep dive

CC BY 4.0

Original authors: Wenbin Wu

Original paper licensed under CC BY 4.0 (http://creativecommons.org/licenses/by/4.0/). ✨ This is an AI-generated explanation of the paper below. It is not written or endorsed by the authors. For technical accuracy, refer to the original paper. Read full disclaimer

The Big Idea: A Tower of Financial Legos

Imagine the world of Decentralized Finance (DeFi) not as a collection of separate apps, but as a giant, wobbly tower built out of financial Legos.

In traditional banking, you have cash (the base), then a bank deposit (a claim on that cash), and maybe a loan against that deposit. It's a simple hierarchy.

In DeFi, this gets crazy. You take a digital coin (like ETH), lock it in a "staking" app to get a receipt token. Then, you take that receipt token and lock it in a "lending" app to get a new receipt token. Then you take that new token and lock it in a "yield farming" app.

You end up with a token that is a claim on a claim on a claim. The paper calls this a "Token Hierarchy." The authors argue that instead of trying to count the "real" money and ignoring the receipts (which is what most people do), we should study the tower itself. The tower is the system.

The Core Metaphor: The "Money View"

The authors use a framework called the "Money View." Think of it like a family tree, but for money.

Tier 0 (The Parents): The original assets. Native coins like Ethereum (ETH) or Bitcoin (BTC). These are the "real" money.
Tier 1 (The Children): Tokens that are directly backed by the parents. For example, "stETH" (a token you get when you stake ETH) or "WETH" (a wrapped version of ETH).
Tier 2 (The Grandchildren): Tokens created by locking up the children. For example, "aWETH" (a token you get when you lend your WETH to a lending app).
Tier 3+ (The Great-Grandchildren): Tokens built on top of the grandchildren.

The paper finds that by late 2025, for every $1 of original "parent" money (Tier 0), there was $4.70 of "child" and "grandchild" money floating around. This is called the Layering Multiplier. It's like a money multiplier in a bank, but it happens automatically through code.

The Great Puzzle: Why Do Deeper Tokens Pay Less?

Here is where the paper gets interesting. You might think: "If I'm holding a 'Great-Grandchild' token (Tier 3), it's riskier because it depends on so many apps working correctly. So, it should pay me a huge interest rate to compensate for that risk, right?"

Surprisingly, the raw data says NO.
The data showed that deeper tokens often reported lower interest rates than the simple ones. This seemed backwards.

The Solution: The "Hidden Yield" Illusion
The authors realized the data was lying to us, but not because of fraud—because of bad accounting.

Imagine you own a rental house (Tier 1). You rent it out and get $1,000/month.
Now, imagine you put that house into a complex investment fund (Tier 2) which then lends the house out again.

The Tier 1 report says: "I am earning $1,000."
The Tier 2 report says: "I am earning $0." (Because the Tier 2 app only sees the new loan it made, not the original rent the house was already earning).

The deeper you go in the tower, the more "upstream" income gets hidden from the report. The Tier 3 token is actually earning the rent from the house plus the interest from the loan, but the app only shows you the interest from the loan.

The Fix: When the authors added back all the "hidden" income from the layers below, the puzzle was solved.

The Composition Effect: Once you fix the math, deeper tiers do pay more. Why? Because the deeper you go, the more likely you are to be in a "Lending" app (which pays high rates) rather than a "Staking" app (which pays lower rates).
The Structural Discount: However, even after fixing the math, there is still a penalty for being deep in the tower. For every step you go down (every "hop" away from the original money), the yield drops by about 2.9 percentage points.
- Why? Because it's harder to find borrowers for these complex, nested tokens. People prefer to borrow simple, liquid assets. Less demand for borrowing = lower interest rates.

The Crisis Test: When the Tower Shakes

The paper also looked at what happens during scary times (crises like the collapse of Terra, the FTX bankruptcy, or the SVB bank run).

They found that during normal times, the difference in interest rates between the top and bottom of the tower is small. But during a crisis, the gap explodes.

The Analogy: Imagine a game of "Musical Chairs" where the chairs are the different layers of tokens. When the music stops (a crisis), everyone rushes to the safest chair (Tier 0, the original money).
The people holding the "Great-Grandchild" tokens (Tier 3) suddenly realize their token depends on three other apps that might fail. They demand a massive "safety fee" (higher yield) to hold the risk, or they sell their tokens to move up the tower.
This proves that the tower is real and that the risk is cumulative. If the bottom layer breaks, the whole tower above it wobbles.

Why Should You Care?

For Investors: Don't just look at the interest rate an app shows you. You need to know how "deep" your token is. A high rate might just mean you are in a deep, risky layer where the app is hiding the fact that the underlying asset is struggling.
For Regulators: This "Layering Multiplier" (the $4.70 for every $1) is a new way to measure how much leverage (debt) the whole crypto system has. If that number gets too high, the system is fragile and ready to collapse.
The Big Picture: DeFi isn't just a bunch of cool apps; it's a credit system that looks a lot like traditional banking, just built with code. It has a hierarchy, it creates money through layers, and it carries the same risks of "double counting" and cascading failures.

In a nutshell: The paper tells us that DeFi is a tower of financial promises. The higher you climb, the more complex and risky it gets. The "interest rates" we see are often misleading because they hide the income from the layers below. When the system gets scared, the whole tower shakes, and the people at the top pay the price.

1. Problem Statement

Decentralized Finance (DeFi) protocols create a complex web of financial claims by stacking derivative tokens atop one another (e.g., depositing ETH $\to$ receiving stETH $\to$ depositing stETH $\to$ receiving astETH). While existing literature acknowledges this "recursive nesting," it largely treats the resulting structure as a measurement problem (i.e., "double counting" that needs to be filtered out to find "true" Total Value Locked).

The paper argues that this view is flawed. Instead of noise, this layering is the credit structure itself. The current lack of a mapped hierarchy obscures systemic risk, prevents accurate leverage measurement, and fails to explain how yields are determined across the ecosystem. The core problem is the absence of a framework to map the "who owes what to whom" relationships in DeFi and how this hierarchy impacts risk and pricing.

2. Methodology

The author applies the "Money View" framework (rooted in Chartalist theory and Perry Mehrling's work) to the entire DeFi ecosystem, treating protocols as financial intermediaries similar to traditional banks.

A. Conceptual Framework: The Token Hierarchy

Intermediation: Protocols accept input tokens (assets) and issue output tokens (liabilities/claims).
Tier Assignment: Tokens are organized into a hierarchy based on their distance from "Base Money" (Tier 0).
- Tier 0: Native assets (ETH, BTC) and fiat-backed stablecoins (USDC).
- Tier 1: Direct derivatives (Wrapped tokens, Liquid Staking Tokens like stETH).
- Tier 2+: Claims on claims (Lending receipts like aUSDC, restaking derivatives).
Layering Multiplier: A macro-prudential metric defined as $\text{Total Mapped TVL} / \text{Tier 0 TVL}$ , measuring the system-wide leverage (claims per dollar of base assets).

B. Data and Graph Construction

Data Source: DefiLlama data covering ~7,000 protocols and 18,000 liquidity pools across 200+ blockchains (Feb 2022 – Dec 2025).
Token Graph ( $G_{deriv}$ ): A directed graph where nodes are tokens and edges represent derivation (transformation).
- Filtering: Excludes trading protocols (DEXs) and bridges; focuses on credit creation (staking, lending, wrapping).
- Algorithm: Uses a Breadth-First Search (BFS) weighted by TVL to assign tiers. Tier 0 tokens are identified as those with zero in-degree in the derivation graph (subject to TVL and connectivity thresholds).
Panel Data: 2.0 million pool-month observations.

C. Empirical Strategy

The paper employs panel regressions to test the relationship between hierarchical position and yields (APY), controlling for TVL, stablecoin status, time, protocol type, and blockchain.

Key Variables:
- Tier: Discrete level (0–3).
- Graph Distance: Integer hops from the nearest Tier 0 token.
- Embedded Yield Correction: A novel adjustment that adds upstream yields (e.g., staking rewards embedded in a token) to the reported APY to correct for systematic underreporting by yield aggregators.

3. Key Contributions

Theoretical Extension: Extends the Money View from stablecoins to the full DeFi ecosystem, redefining "double counting" not as an error but as the money multiplier of the crypto economy.
Methodological Innovation: Develops an automated, balance-sheet-based algorithm to discover token hierarchies from raw pool data, creating a "Token Graph" that explicitly maps dependency chains.
Yield Decomposition: Disentangles two opposing forces in DeFi yields:
- A Compositional Premium (deeper tiers have more lending protocols, which pay higher rates).
- A Structural Distance Discount (deeper tokens attract less borrowing demand, lowering rates).
Macro-Prudential Metric: Introduces the Layering Multiplier as a real-time tracker of system-wide leverage.

4. Key Results

A. Structure and Leverage

Hierarchy Stability: Despite market volatility, the relative distribution of TVL across tiers remains stable (Tier 0 and 1 hold ~75–85% of value).
Layering Multiplier: By late 2025, the multiplier reached 4.7x (meaning $1 of base assets supported $4.70 of total claims). This recovered from a low of 2.1x during the 2024 deleveraging.
Drivers: Lending and staking account for over 80% of the layering. Lending dominates the deepest tiers (Tier 2 $\to$ 3), mirroring the "stake then lend" pattern of traditional finance.

B. The Yield Puzzle and Correction

Initial regressions showed a negative relationship between tier depth and yield (deeper tiers appeared to earn less), which contradicted risk-premium theories.

The Artifact: This was largely a measurement error. Yield aggregators (like DefiLlama) report only the outermost protocol's rate, omitting upstream yields (e.g., the staking yield embedded in a token used as collateral).
The Correction: After applying the Embedded Yield Correction, the tier coefficient vanished, while the Graph Distance coefficient remained robustly negative.
Decomposed Findings:
- Compositional Effect: Controlling for structural distance, each tier step up adds +2.0 percentage points (pp) to yield (driven by the shift from staking to lending protocols).
- Structural Effect: Controlling for protocol type, each additional hop from base assets reduces yield by -2.9 pp. This reflects lower borrowing demand for nested, less liquid tokens.

C. Crisis Dynamics

Countercyclical Premium: The yield gradient widens during crises. During the SVB crisis, the tier premium rose from ~0.6 pp (normal) to ~1.6 pp.
Implication: This confirms that upstream dependency risk is repriced during stress. The SVB crisis specifically demonstrated how a shock to a Tier 0 asset (USDC) cascaded through the entire hierarchy, breaking par at multiple levels.

5. Significance and Implications

Risk Assessment: The Token Graph makes dependency chains explicit. A Tier 3 token (e.g., aweETH) depends on the solvency of Aave, ether.fi, Lido, and Ethereum simultaneously. This quantifies "cascade risk."
Regulatory Tool: The Layering Multiplier serves as a macro-prudential indicator for regulators, signaling leverage cycles similar to capital ratios in traditional banking.
Investor Insight: Investors must distinguish between reported APY and total return. A Tier 3 token's reported yield understates total return (due to embedded upstream yields) but overstates safety (due to structural distance discounts and crisis sensitivity).
Systemic Anchoring: The study reveals that despite DeFi's "decentralized" aspirations, the hierarchy remains anchored in traditional finance (e.g., USDC's link to Silicon Valley Bank), proving that crypto credit systems are not immune to traditional banking shocks.

In summary, the paper reframes DeFi not as a collection of isolated protocols, but as a layered credit hierarchy where risk and yield are structurally determined by a token's position in the derivation chain. It provides the first quantitative framework to measure and monitor this systemic leverage.