Post-Trade Compression: The Subscription Netting Never Cancels

Imagine five overlapping subscriptions to the same streaming service: a legacy plan from three years ago nobody remembered to cancel, a family plan added when a sibling moved in for a few months, and a premium tier tested once and left running. The billing app is smart enough to net all five into a single combined monthly charge, so paying it feels simple enough. But the company's own books still show five separate, fully live contracts, because netting the bill and canceling the subscriptions are two completely different acts. One changes what gets paid. The other changes how many contracts actually exist.

Banks run exactly that problem with derivatives, at a scale where the difference isn't a confusing credit card statement — it's a regulatory capital charge.

Netting settles a bill. Compression cancels the contract.

Netting collapses a web of obligations down to the smallest number that actually needs to move at settlement — but it never touches the underlying trades themselves. Every original contract stays alive, still legally binding, still sitting on the books, still counted. Portfolio compression (also called trade compression) is a separate operation entirely: identifying trades that are economically redundant — positions whose cash flows already cancel each other out — and tearing them up completely, replacing a large population of offsetting contracts with a much smaller one that produces exactly the same net market risk. Netting changes what has to be paid. Compression changes how many contracts exist in the first place.

That distinction sounds academic until you ask why anyone would bother eliminating contracts that were already netted down to zero net risk. The answer is that modern bank capital rules don't only look at net risk. The Basel III leverage ratio was built specifically as a blunt backstop to risk-based capital requirements, and it counts a bank's derivatives exposure on a gross basis, with only limited recognition for netting. A bank can run a portfolio that nets to zero market risk and still be forced to hold capital against it, purely because the leverage ratio counts the size of the book it's still carrying, not the number that survives after everyone's promises cancel out. That's a different kind of risk than the one netting was ever built to solve — and it's why compression exists as its own discipline instead of a footnote inside netting.

Scale the subscriptions picture from one person's account to a market and the mechanism stretches without changing shape. Real compression is almost always multilateral: Bank A's trade with Bank B, Bank B's trade with Bank C, and Bank C's trade back with Bank A can all be torn up together and replaced with a smaller set of new contracts that leave every one of the three banks holding exactly the net position they started with. Nobody's individual subscription list even has to look redundant on its own — the redundancy only shows up once someone lines up everyone's contracts against everyone else's at the same time, which is exactly why compression cycles run as one coordinated exercise across dozens of participants at once rather than something any single desk can do by looking at its own book.

Example 1: the global market that shrank once gross size started costing money

Trade compression predates the leverage ratio — TriOptima launched its triReduce service back in 2003, originally to cut the plain operational and counterparty risk of banks carrying thousands of offsetting contracts on their books. The practice accelerated sharply once Basel III's leverage ratio came into force and started rewarding exactly the kind of gross-notional reduction that earlier, purely risk-weighted capital rules had never rewarded. The Bank for International Settlements recorded a 14% drop in interest rate derivatives notional outstanding in the six months to June 2015 alone — from $505.4 trillion down to $434.7 trillion — and named compression, not any actual change in market activity, as the primary driver. ISDA's own research went further: adjusted for the separate effect of central clearing, the interest rate derivatives market would be 162% larger today without any compression activity at all. By 2018, triReduce alone had compressed $250 trillion of gross notional at LCH SwapClear in that single year, on top of more than $750 trillion eliminated cumulatively since the service began. None of that changed a single bank's actual net exposure to interest rate movements. It changed how many contracts were sitting on their books when a regulator measured the leverage ratio.

Some regulators didn't just create a capital incentive to compress — they mandated the process outright. Under the EU's EMIR framework, a counterparty holding 500 or more non-centrally-cleared OTC derivative contracts with the same counterparty has to analyze the possibility of a portfolio compression exercise at least twice a year, and actually perform one where the analysis says it's possible, with a formal explanation owed to regulators if it isn't. In the US, CFTC Rule 17 CFR 23.503 requires every swap dealer and major swap participant to maintain written policies for bilateral offset, bilateral compression, and multilateral compression with other swap dealers and major swap participants. Two different regulatory philosophies — reward it through capital math, or require it outright — arriving at the same conclusion: a portfolio that nets to zero risk still isn't finished if it's still carrying contracts nobody needs.

Example 2: India runs the same exercise through its own clearinghouse

India's Clearing Corporation of India Ltd (CCIL) runs its own recurring compression cycles for the interbank interest rate swap market benchmarked to MIBOR, rather than routing the exercise through a private vendor the way most global banks use TriOptima. On September 11, 2025, CCIL completed the 39th cycle of this exercise with 32 participating banks and primary dealers. Of the 30,855 trades among those participants found eligible for compression, 27,622 were torn up entirely — an 89.4% compression rate — cutting market-wide notional outstanding by ₹11,06,820.38 crore, roughly $130 billion, in a single exercise. Like every compression cycle anywhere in the world, the point wasn't to change what any of those 32 institutions owed each other. It was to stop carrying 27,622 contracts that were doing nothing except sitting on a ledger, getting reconciled, getting reported, and counting toward capital calculations that had already priced in their real net effect twice over.

Notice what's different about how India gets to that same outcome. EMIR and the CFTC compel compression through a compliance deadline sitting on top of a market that runs the exercise itself, bank by bank. CCIL's cycles run the other way around: the clearinghouse itself operates as the market's shared infrastructure for compression, built into the same institution that already clears and nets these swaps, rather than a rule forcing each bank to go find a vendor and justify why it hasn't. The gross notional disappears either way; a market-infrastructure design got India there instead of a compliance mandate.

Reality check: compression can't do what netting does in a crisis

It's worth being precise about why compression isn't just netting with extra steps. Close-out netting activates automatically the moment a counterparty defaults — the contract itself is what tells both sides how offsetting positions collapse into one enforceable number the instant something goes wrong. Compression is the opposite kind of event: it's voluntary, scheduled, and it depends on every participant agreeing in advance, using consistent and mutually accepted trade valuations, that a given set of contracts really is economically redundant enough to destroy outright. Nobody can compress their way out of a Tuesday-morning default the way netting activates on its own. That's precisely why every serious compression cycle — CCIL's included — runs as a coordinated multilateral exercise with a neutral party doing the matching, rather than something any single bank can trigger unilaterally on its own book. Netting is a rule contracts already contain, waiting to be invoked. Compression is a project someone has to run.

Why this matters for a Business Analyst

Back to the subscriptions

A requirement that says "reduce our derivatives exposure" is unfinished until it specifies which of two entirely different problems it's actually describing. If the goal is reducing net market risk, that's usually already solved — through netting, margining, or both — before anyone opens a new project. If the goal is reducing gross notional or trade count for a leverage-ratio or operational reason, that's compression, and it touches a completely different set of systems: trade repository updates reflecting terminated contracts, legal tear-up confirmations from every counterparty involved, and notional reporting to regulators that has to reconcile against a smaller population of trades than existed the day before. Scoping a "compression project" as if it were a netting or reconciliation project — assuming the trades stay put and only a number gets recalculated — builds the wrong system for what actually has to happen: contracts getting permanently destroyed, not just added up differently.

Lighthouse Insight

A combined subscription bill only tells you what gets paid this month. It says nothing about how many separate contracts are still quietly running behind it, still counted, still on file, still someone's job to track. CCIL tore up 27,622 of them in a single September morning in 2025. Globally, compression has eliminated more than $750 trillion of derivatives contracts that a net settlement number had already made irrelevant — proof that knowing what you owe and actually canceling what you don't need are two different jobs, and a market only finishes the second one on purpose.

Reference anchors

• BIS Quarterly Review: Outstanding OTC Derivatives Positions Dwindle as Compression Gains Further Traction (December 2015)
• ISDA: The Impact of Compression on the Interest Rate Derivatives Market (July 2015)
• Hedgeweek: TriOptima Compression Exceeds USD750 Trillion in Eliminated Notional Principal
• Benzinga: TriOptima Sets New triReduce Record — $250 Trillion Gross Notional Value Compressed at LCH SwapClear During 2018
• CCIL: Derivatives Quarterly, July–September 2025
• ESMA: Clearing Obligation and Risk Mitigation Techniques Under EMIR
• eCFR: 17 CFR § 23.503 — Portfolio Compression