The Core Idea

Rava as an On-Chain Clearinghouse

Financial markets do not derive stability or capital efficiency from trade execution alone. Execution venues facilitate price discovery but do not bound downside risk or guarantee settlement under stress. These functions are performed by clearinghouses, which operate as neutral post-trade utilities that aggregate exposure, compute risk, and enforce margin continuously.

In traditional finance, clearinghouses enable leverage and liquidity by transforming gross exposures into net exposures and constraining those exposures using statistically grounded risk measures. In decentralized finance, this clearing layer is largely absent. Rava provides it.

What Clearinghouses Actually Do

A clearinghouse interposes itself between counterparties to guarantee settlement and manage systemic risk. It does not speculate, provide directional liquidity, or price assets for profit. Its objective is to bound plausible loss over a defined margin period of risk and to ensure that sufficient collateral is posted to absorb that loss.

Clearinghouses control risk through three tightly coupled functions:

Exposure Aggregation:Collecting position information and computing net exposures. Netting reduces the collateral required to support a given level of activity.

Loss Distribution Estimation:Estimating the distribution of potential losses over the margin period. This incorporates volatility, correlation, liquidity, and stress scenarios. The output is Value at Risk.

Haircut Enforcement:Translating VaR into collateral requirements. Each asset receives a haircut reflecting its risk contribution. Haircuts adjust continuously as conditions change.

These functions operate as a continuous loop. Position changes flow into aggregation. Aggregated exposures flow into loss estimation. Loss estimates flow into haircut adjustments. Haircut adjustments constrain positions. The loop runs continuously, not at discrete intervals.

Haircuts and Discounts

From VaR, RAVA derives two related but distinct values that serve different purposes.

Haircut: Always Applied (The Normal Case)

The haircut is always applied as part of clearing. It determines how much margin must be posted and how much exposure can be safely supported during the settlement period.

Collateral Value = Market Value × (1 − Haircut)

Even for a simple T+2 token, the haircut is applied because a two-day settlement window creates price, delivery, and operational risk. The haircut ensures that if anything goes wrong, there's enough margin posted to absorb losses without breaking settlement.

The haircut is decomposed into two components:

Static Component (S):Structural risk that cannot change over short time horizons. For some assets this includes legal complexity, redemption mechanics, or manager quality. For liquid assets, S may be minimal.

Dynamic Component (D):Market risk that updates continuously. Volatility, credit spreads, liquidity conditions, correlation shifts. This component changes as markets move.

Haircut = S + D(t)

This decomposition provides transparency. Market participants can see which components are structural versus market-driven. The dynamic component can be hedged using standard instruments.

Discount: Only If Settlement Fails (The Failure Case)

The discount is NOT applied upfront. It only becomes relevant if settlement fails and the clearing system must replace, finance, or exit the position before delivery.

Execution Price = Market Value × (1 − Discount)

The discount incorporates the haircut plus additional execution costs: market impact, time-to-exit risk, adverse selection, and slippage.

Discount = Haircut + Execution Costs

In normal conditions, no discount is realized. But the haircut exists to ensure that IF the discount becomes relevant, losses are absorbed without breaking settlement.

Haircut is insurance you always pay. Discount is the cost you hope to never realize.

Discounts are always greater than or equal to haircuts. The gap widens as asset liquidity decreases.

VaR as the Foundation

Value at Risk is defined as a quantile of the loss distribution over a fixed horizon. VaR at confidence level α is the smallest loss threshold such that losses exceed this threshold with probability at most (1 − α).

Clearinghouses typically use horizons of one to five days and confidence levels of 99% or higher. The horizon reflects the margin period of risk: the time required to detect a problem, close out positions, and settle obligations.

VaR is not a pricing model. It is a loss bound under adverse but plausible conditions.

Haircuts are the deterministic transformation of VaR into collateral constraints. As inputs change, VaR changes continuously, and haircuts move smoothly along a curve. This produces gradual tightening of leverage rather than discrete liquidation events.

Asset-Specific Methodology

Different assets require different approaches to VaR estimation, producing different haircuts and discounts:

Liquid assets:Historical simulation with short lookback windows. Volatility is directly observable. Haircuts respond quickly to market changes. Haircuts: 1-5%. Discounts: 2-7%.

Credit instruments:Factor models incorporating spread, duration, and liquidity risk. Haircuts: 5-15%. Discounts: 8-20%.

Structured products:Scenario-based analysis with stress correlation assumptions. Limited secondary market requires wider buffers. Haircuts: 15-30%. Discounts: 25-45%.

Private assets:Explicit modeling of valuation lag, exit time, and forced sale costs. VaR includes uncertainty about the VaR estimate itself. Haircuts: 30-50%. Discounts: 40-60%+.

The gap between haircut and discount widens as liquidity decreases. For liquid assets, execution costs are minimal. For illiquid assets, forced sale conditions impose significant additional costs.

Governance defines methodology for each asset class. Individual parameters are derived from methodology using current market data.

Clearing Example: Monthly-Redeemable Private Credit

This example illustrates how RAVA handles gated, illiquid assets where the primary risk is time-to-cash uncertainty rather than price volatility.

The Asset

A private credit RWA token with the following properties:

• Reported NAV: $100 per token
• Redemption frequency: Monthly
• Redemption subject to gating
• Secondary liquidity: Limited

The Trade

A trader sells $10 million of this token with expected economic settlement at T+2.

From a clearing perspective, the risk is not price volatility alone but the possibility that cash cannot be realized for up to one month.

VaR Computation

RAVA computes risk using a liquidation-horizon-adjusted VaR model that accounts for:

• NAV variability over one month
• Correlation to market drawdowns
• Execution uncertainty under forced exit
• Operational delay risk

Resulting loss estimate over the liquidation horizon: 35%

Margin Calculation

Applied haircut: 35%

Margin required from the seller:

• $10 million notional
• 35% haircut
• $3.5 million USDC posted as margin

If redemption occurs as scheduled, the trade settles normally and the $3.5 million USDC is returned minus clearing fees. If redemption is delayed or gated, the margin bridges the delay so the buyer settles on time.

Comparison to Liquid Assets

The difference is driven by time-to-cash uncertainty, not asset quality.

A high-quality private credit fund with excellent credit metrics still requires a larger haircut than a liquid equity token because the risk being priced is settlement timing, not creditworthiness.

The Key Insight

Private credit RWAs with gated liquidity are not incompatible with on-chain markets. They require clearing systems that explicitly price settlement timing risk.

By separating asset design from settlement guarantees, RAVA allows these instruments to become tradable and marginable without changing their redemption terms or forcing artificial liquidity.

The Default Waterfall

Traditional clearinghouses absorb losses through a structured waterfall: variation margin, initial margin, defaulting member collateral, mutualized default funds, member assessments, and finally clearinghouse capital. This layered structure means clearing does not require instant liquidation to remain solvent.

DeFi protocols rely almost exclusively on liquidation. When a position becomes undercollateralized, liquidators must find buyers and execute immediately. If liquidation fails, the protocol absorbs bad debt directly. No default fund, no assessment mechanism, no buffer beyond collateral.

Rava introduces clearing mechanics that bound loss before liquidation becomes necessary. Continuous haircut adjustment means positions are margined proactively rather than reactively.

Why This Matters

Without a clearing layer, every protocol builds its own internal risk logic. This creates fragmentation, inconsistent leverage, and systemic risk.

With a shared clearing layer, all protocols reference the same valuation foundation. This enables composability, consistent credit standards, and safe leverage across the ecosystem.

Regulatory Alignment

Clearinghouses, VaR models, and haircut schedules are well-established regulatory constructs. Regulators understand them. Compliance frameworks exist for them. Risk officers can audit them.

An on-chain clearinghouse aligns with existing regulatory intuition far more closely than static collateral systems. The methodology is familiar. The math is standard. The enforcement is continuous and transparent.

Asset Evolution Under Clearing Pressure

Assets that seek tighter haircuts will adapt. They will improve pricing frequency, shorten redemption cycles, enhance disclosure, and develop secondary liquidity.

Clearing creates economic pressure toward better asset design. Assets that remain opaque or hard to exit face wider haircuts. Assets that improve data continuity and exit mechanics earn tighter haircuts and better financing terms.

The Settlement Layer

RAVA is not a lending protocol. It is risk infrastructure that lending protocols integrate. It does not lend, trade, or take directional risk.

Just as price oracles provide spot prices for liquid assets, RAVA provides settlement values for all tokenized assets. Protocols consume these values to compute margin requirements and liquidation triggers.

This is the first time clearinghouse mechanics have been expressed in open, programmable form for tokenized assets.