LTVs, Liquidity, and Lessons from the Frontlines of Crypto

In the BRANESTORM podcast’s most technically rich episode to date, host Craig Birchall, Head of Product at Membrane Labs, sat down with two of the digital asset industry’s sharpest minds in institutional risk: Asaf Iram, Co-Founder and Chief Risk Officer at Bitpulse, and Mihai Popa, Head of Risk Analytics at CorPrime. Together, they unpacked the mechanics behind lending risk frameworks, collateral modeling, and navigating liquidity across CeFi and DeFi.

At the core of risk underwriting in crypto is setting the right Loan-to-Value (LTV) ratios, which strike a delicate balance between how much one can borrow and the collateralization threshold at which liquidation is triggered.
Mihai laid out a traditional, data-driven approach:

• Start with historical returns for an asset like BTC.
• Use the 99th percentile drawdown as your stress-loss input.
• Set LTV accordingly. For example, if Bitcoin’s worst-case drop is 25%, the minimum safe collateralization becomes 133%.

Asaf, in contrast, explained Bitpulse’s simulation-based risk engine, where clients define margin levels and loan parameters. The system then runs Monte Carlo simulations to produce probabilities of margin calls, underwater scenarios, and liquidations. This helps institutions calibrate LTVs around their risk appetite, rather than relying solely on backward-looking data.

Both guests agreed that credit risk in crypto is uniquely challenging to quantify. Mihai emphasized that, even with audited financials, understanding counterparty risk remains murky due to a lack of transparency regarding indirect exposures. His method? A combination of:

• Ratio analysis (e.g., debt-to-assets, liquidity ratios)
• Scorecards based on financial and operational factors
• Qualitative overlays on organizational quality and governance

Asaf shared a personal approach from his Anchorage days – create a quality score for financials, discounting unreliable documentation, and never assume solvency without verification. In his view, every institution must ask, “How much money can we afford to lose and still survive?”

Liquidity risk isn’t just about volatility, it’s about where you can unwind a position when things go south.

CeFi Context (per Mihai):

• Use order book data from centralized exchanges.
• Model slippage explicitly.
• Incorporate these estimates into margin models.

DeFi Complexity (per Asaf):

• Real liquidity ≠ Total Value Locked (TVL). TVL is often misleading without knowing how it’s composed across pools and layers.
• Bit Pulse builds direct RPC adapters to different DeFi protocols (e.g., Uniswap, Curve) to calculate usable liquidity.
• Redemption risk is real: protocols may claim “instant liquidity,” but if only 5 WBTC is actually available, that promise is meaningless.

Both emphasized that real-time liquidity monitoring is essential. Pools can lose hundreds of millions in TVL in minutes.

The episode ended with a candid reflection on the cultural challenges of risk management in crypto. Asaf noted the tension between traders and risk teams, where one is wired for upside, the other for caution. Mihai added that while risk is often undervalued during bull markets, it’s indispensable in a crisis.

Ultimately, the takeaway was clear: risk isn’t about saying no, it’s about giving teams the tools and clarity to make better decisions.

Whether you’re an institutional lender navigating CeFi or a protocol operator deep in DeFi, this episode offers a masterclass in how to approach risk rigorously and realistically.

Want to explore risk analysis for Crypto lending? Learn more about Membrane Risk Analysis.

Watch the full episode on YouTube