Clearing & SettlementAssumes the mechanics

Between agreeing a trade and owning the thing sits an industry nobody thinks about until it fails. It decides who you are actually exposed to, and for how long.

Three separate things, in order

• Execution — the price and size are agreed. This is the part everyone watches, and it is over in microseconds.
• Clearing — the obligations are established, netted and collateralised. Who owes what, secured how, until delivery.
• Settlement — the asset and the money actually change hands. Only now do you own anything.

Between the second and the third you hold a claim on a counterparty, not an asset. That gap is the entire subject of this page, and shortening it has been the industry's main project for forty years.

The central counterparty, and what novation really does

• Novation: after a cleared trade, your original contract is torn up and replaced by two — you against the clearing house, the clearing house against your counterparty. You no longer care who you traded with. You care a great deal about the clearing house.
• What that buys: counterparty risk concentrated in one heavily capitalised, heavily margined entity instead of scattered across a web nobody can see whole. Archegos is the counter-example — five prime brokers, each blind to the other four.
• What it costs: initial and variation margin on everything, posted daily, in eligible collateral. Cleared markets are safer and consume far more liquidity, which is a trade the system has made deliberately.
• What it concentrates: a CCP is a single point of failure by design. It is engineered not to fail, which is a different statement from cannot.
• Not everything clears. Bespoke derivatives remain bilateral, secured under a collateral annex rather than by a clearing house — the counterparty is a named institution again, and the credit work returns with it.

Netting: the reason the plumbing can carry the volume

Gross obligations in a large market run to many multiples of the cash available to settle them. Netting is what closes that gap:

• Payment netting — offsetting same-day, same-currency flows with one counterparty. Housekeeping.
• Bilateral close-out netting — on a default, every contract with that counterparty collapses to a single net claim. This is what makes a collateral agreement work at all; without it, an administrator could enforce the contracts in their favour and disclaim the rest.
• Multilateral netting — only a central counterparty can do this: A owes B, B owes C, C owes A, and almost nothing needs to move. It is where the large compression numbers come from.

The settlement window, and why it keeps getting shorter

While a trade is agreed but unsettled, the price can move. If the counterparty then fails, you must replace the trade at the new price, and the difference is your loss. That exposure scales with the square root of the window, which is why cutting T+2 to T+1 removes about 29% of it and cutting T+1 to same-day removes the rest:

Delivery versus payment, and the risk it removes

• DvP means the asset moves only if the cash moves, simultaneously and irrevocably. Without it, one side delivers and then hopes — a risk expensive enough to have its own name in FX, where the two legs settle in different time zones and the spot market built an entire settlement utility to close the gap.
• Fails happen anyway: the securities were lent out, an instruction was wrong, a chain of deliveries stalled on one missing link. Most fails are operational rather than credit events, and modern regimes charge a daily penalty precisely so that they stay boring.
• A buy-in is the last resort: the failed-to party goes to the market, buys the security, and bills the difference. Rare, and unpleasant in an illiquid name — liquidity and settlement failure are the same problem seen from two ends.

The chain you actually own through

• Almost nobody holds securities directly. The chain runs broker → custodian → central securities depository, and your name usually appears only at the first link.
• Segregation is the thing that matters. Client assets held separately from the firm's own are recoverable if the firm fails; assets pooled with the firm's are a claim in an insolvency. That distinction, not the brand on the app, is the protection — see investor protection.
• Securities lending sits inside the chain. Your shares may be on loan against collateral, which is how some fee-free services are paid for. It is disclosed, it is legal, and it is a counterparty exposure you were not thinking about.
• Omnibus accounts complicate recovery. Pooled client holdings are faster and cheaper to run and slower to untangle when something goes wrong — a trade-off made on your behalf, in a document you agreed to.

The default waterfall

When a clearing member fails, the losses are absorbed in a fixed, published order. Knowing it is the only way to know what a cleared position is really exposed to:

• First: the defaulter's own margin, then the defaulter's contribution to the default fund. The polluter pays, as far as they can.
• Then: a slice of the clearing house's own capital — deliberately placed before other members' money, so that the CCP has something at stake in setting margin properly. This is the layer everyone argues about.
• Then: the surviving members' default-fund contributions — mutualised loss. At this point being a member of a clearing house means paying for someone else's failure.
• Last: assessment powers, position tear-ups, variation-margin haircutting. Recovery tools that exist so the alternative — an unmanaged CCP failure — never has to be tested.