Spine: Coinbase's on-chain loan book on Base

Independent analysis, not affiliated with Coinbase. USDC lent against cbBTC, ETH and six alts on Morpho Blue. Snapshot loading

What the words mean
LLTV
Liquidation loan-to-value: the debt-to-collateral ratio at which a position can be liquidated. 86% on cbBTC and WETH, 77% on cbETH, 62.5% on the alts.
Liquidation bonus
The discount a liquidator gets on the collateral it seizes, 4.38% at 86% LLTV. A lower LLTV comes with a larger bonus.
Bad-debt LTV
1 / (1 + bonus), about 95.8% at 86% LLTV. Above it the seized collateral no longer covers the debt plus the bonus, and lenders take the difference.
Health factor
LLTV divided by the position's LTV. Below 1.0 the position can be liquidated.
Capacity
How much collateral liquidators can sell within the bonus in one step: on-chain pools plus Coinbase and Kraken bids, after the stress multiplier.
Scenarios A / AB / ABC
A: on-chain bots only. AB: plus liquidators selling on Coinbase and Kraken, limited by the capital they bring each day. ABC: plus Coinbase itself redeeming cbBTC 1:1, so depth is unlimited.
Book
Every open position in a market (or in all nine), marked at today's prices.
Stress multiplier
The fraction of today's measured depth assumed to survive a crash; the Stress test note prints the values in use.

Now

How big the book is, how far it sits from trouble, and what the model says about the worst path.

The book

-

borrowed, all nine markets

Distance to capacity, cbBTC

-

    Red at 10% or under, amber at 20% or under. Alts are 3% of the book.

    Worst modeled path

    loading

    Book

    Who owes what, at what loan-to-value, and how close each position sits to its liquidation line.

    Market

    LTV distribution

    Borrow USD per 2% LTV bin; the tooltip carries the number of positions. Dashed lines: the LLTV where liquidation starts and the LTV past which seized collateral no longer covers the debt (bad debt).

    Top 25 borrowers

    Liquidations

    How much debt a price drop would put up for liquidation, what liquidators can absorb, and what they have cleared before.

    Liquidatable borrow vs price drop

    Borrow that crosses LLTV at each instantaneous price drop. The second series is the part of it already past the bad-debt LTV, where the seized collateral no longer covers the debt plus the bonus. Dashed lines: what Coinbase bids and DEX pools would absorb today within the liquidation bonus. The shaded region starts at the stressed total of the two (the capacity); the dot marks the drop at which the book first outruns it.

    Liquidation history

    Daily repaid USD. The three crashes the book has lived through are named.

    Trend

    Each snapshot since the tracker started: book LTV, distance to capacity and borrow liquidatable at a 20% drop, for the selected market.

    Top 10 liquidators (all time, by repaid USD)
    Liquidator leaderboard (stress windows)

    Top-10 liquidators of each stress window, merged across cbBTC and WETH. Gross bonus is seized minus repaid, before gas. Latency: median seconds from the oracle crossing to the liquidation, where the address has at least 5 measured events.

    Stress test loading

    Today's book replayed through seven historical crashes to see how much lenders would lose.

    The two paths that cost money, at today's terms

    Today's book at its current LLTV, liquidators selling on-chain and on exchanges (scenario AB): what lenders lose by the end of the path, what sat underwater at the lowest print, and the loss if the low had held for a day.

    Who carries a loss

    The vaults funding the selected market, largest first, and their cut of the worst modeled path (March 2020, scenario AB, the row above). Morpho socializes bad debt pro rata to suppliers, so each vault's cut is its share of market supply; the last column sets the held-a-day loss against the vault's total assets. Other suppliers are wallets supplying the market directly.

    Bad debt by LLTV and crash path

    A: on-chain bots only. AB: plus exchange-hedged liquidators (what we observe). ABC: plus a Coinbase backstop (hypothetical).
    Loss: realized by the end of the path. Exposure: underwater at the lowest print.

    Liquidator capital, cbBTC on March 2020 (AB)

    How much liquidator capital would have to exist for the March 2020 path to leave nothing exposed. Bad debt as the daily cap on exchange-hedged liquidators rises from the largest day the book has seen (1x).

    Warning time

    How long a position that ended up liquidated spent between 80% and 86% LTV before crossing, median, at today's terms (scenario AB). A borrower warning needs hours; the March 2020 bar is the one to read.