Why Event Collateral Is Different

Overcollateralized lending works because most collateral is forgiving: it moves in steps, it can be sold, and a bad position still leaves something to recover. Event positions are none of those things. Four properties do the damage.

Four properties that break lending

• It jumps: information does not leak in gradually. A position can fall from one level to a far lower one between two blocks, the instant a ruling, a count, or a print lands. No liquidation that reacts after the move can catch it.
• Liquidity leaves with the news: the event that moves the price is the same event that clears the book. So depth is thinnest precisely when a lender needs to sell, not when it is convenient.
• The payoff is binary: the claim resolves to a fixed value. Collateral that looked well-covered an hour earlier can settle to nothing, and there is no partial recovery to fall back on.
• The borrower may know more than you: someone who wants leverage against a specific outcome often has a sharper read on that outcome than the lender pricing the loan. Adverse selection is baked into the demand.

What that forces the design to do

These are not footnotes to work around. They decide the architecture.

Mark collateral at what it would fetch, not what it last printed. CUSP values a position at its conservative mark, the most pessimistic of several prices, then haircuts it further. Value the system cannot defend on an actual sale does not back credit.

Pull borrowing capacity before resolution, not at it. Capacity decays as the event approaches, and the origination gate shuts off new credit entirely near it. Nothing gets freshly financed through the window where it can gap to zero.

Liquidate through specialists, not bots. A generic keeper sells into a book that is no longer there and misprices outcomes it does not understand. CUSP clears positions through auctions among bonded specialists who price the collateral correctly because it is their job to.