Compression and Portfolio Tear-Ups

Compression and portfolio tear-ups—the systematic termination of offsetting OTC derivatives positions—show up in operations as bloated trade counts, inflated gross notional figures, and unnecessary capital consumption that drags on balance-sheet efficiency. In practice, a well-run compression cycle can eliminate 60-80% of gross notional while preserving the exact same net risk profile. The practical point isn't whether to compress (regulators increasingly expect it). It's how to set tolerances, select the right cycle, and avoid the operational pitfalls that turn a capital-efficiency exercise into a reconciliation nightmare.

TL;DR: Compression and tear-ups terminate redundant OTC derivatives trades to reduce gross notional, trade counts, and capital charges—without changing your net risk. Getting it right requires careful tolerance-setting, clean trade populations, and coordination across front office, operations, and risk.

What Compression and Tear-Ups Actually Mean

Compression is the process of replacing a portfolio of trades between counterparties with a smaller set of trades (or fewer trades) that preserve the same net economic exposure. Tear-ups are a specific form: outright termination of trades, settled in cash, with no replacement trade created.

The distinction matters operationally. In a compression cycle, you might terminate four trades and book one replacement. In a tear-up, you terminate trades and exchange a cash payment—nothing new gets booked.

Why this matters: Every live trade in your portfolio generates operational cost—confirmations, payment processing, resets, regulatory reporting, margin calls. A book with 500 trades that nets to the same risk as 150 trades is consuming 3x the operational bandwidth for zero incremental risk benefit.

Core Terminology (What You Need to Know)

Gross notional is the sum of all trade notionals in absolute terms. If you pay fixed on $200M and receive fixed on $150M, your gross notional is $350M. Net notional is the directional exposure after offsetting—in this case, $50M pay-fixed.

Compression ratio measures effectiveness: notional eliminated divided by starting notional. A 70% compression ratio means you removed $7 of every $10 in gross notional. Industry benchmarks for well-constructed cycles typically land between 60% and 80%.

Tolerance is the maximum acceptable change in any risk parameter (DV01, gamma, vega, cash settlement) that a participant will accept. This is the critical control—set it too tight and the algorithm can't find solutions; set it too loose and you're taking unintended risk.

Trade population is the set of trades eligible for a given compression cycle. Getting this right (complete, reconciled, correctly formatted) is where most operational failures originate.

How Compression Works in Practice (The Mechanics)

Bilateral Compression (Two Counterparties)

This is the simplest form. You and a single counterparty identify offsetting trades and terminate them, booking a replacement if needed.

The point is: bilateral compression is something you can do anytime, without a third-party service. It's the low-hanging fruit, and if you're not doing it regularly, you're leaving capital efficiency on the table.

Example: Party A and Party B hold two interest rate swaps:

Gross notional: $180M. Net position: Party A pays fixed on $20M. After compression, you terminate both trades and book one replacement: Party A pays fixed at approximately 4.48% on $20M. The 5-basis-point rate differential on the $80M overlap generates a small cash settlement (roughly $18,000, depending on duration and day count).

Result: Gross notional drops from $180M to $20M. Trade count drops from 2 to 1. Net risk is unchanged.

Multilateral Compression (The Network Effect)

This is where the real power sits. A compression service provider (TriOptima, Quantile, or Capitalab) runs an optimization algorithm across multiple counterparties simultaneously, finding chains of offsetting positions that bilateral compression can't reach.

Why this matters: Party A might owe Party B, who owes Party C, who owes Party A. No bilateral pair can fully net this circular exposure—but a multilateral algorithm can eliminate it entirely.

The typical cycle runs like this:

1. Submission: Participants send their eligible trade populations to the compression service (typically via FpML or CSV feeds)
2. Matching and optimization: The service identifies offsetting chains and runs a constrained optimization—maximizing notional reduction while respecting every participant's tolerances
3. Proposal: Each participant receives a proposal showing which trades would be terminated and what replacement trades (if any) would be created
4. Accept/reject: Participants review the proposal against their tolerances. Most services operate on an all-or-nothing basis for each cycle—you accept the full proposal or reject it
5. Execution: Accepted trades are terminated, replacements booked, and cash settlements processed
6. Reporting: Terminations and new trades are reported to swap data repositories (SDRs) per CFTC/ESMA requirements

Typical cycle metrics (for a large rates compression run): 15-20 participants submit 40,000-60,000 trades with $1.5-2.5 trillion in gross notional. A successful cycle eliminates 25,000-35,000 trades and $1.0-1.8 trillion in notional. These are real numbers from industry compression runs (TriOptima has reported cumulative eliminations exceeding $1 quadrillion in notional since inception).

CCP Compression (Inside the Clearing House)

If your trades are cleared, the CCP itself offers compression services. LCH SwapClear runs weekly compression cycles. CME and Eurex offer similar services.

The advantage here is simplicity. Your clearing relationship doesn't change, the process is automated within the CCP infrastructure, and the margin benefit is immediate—fewer positions means lower initial margin.

What experience teaches: CCP compression is essentially free capital efficiency. If you're a clearing member and not participating in every available cycle, you're overpaying for margin.

Worked Example: Full Compression Cycle (With Numbers)

Initial portfolio between Party A and Party B:

Portfolio metrics before compression:

• Gross notional: $550 million
• Net position: Pay fixed on $50M (pay total $300M, receive total $250M)
• Trade count: 4
• Weighted average pay rate: ~4.52% (on $300M)
• Weighted average receive rate: ~4.50% (on $250M)

Compression proposal: Terminate all four trades. Create one replacement:

The calculation for cash settlement:

The rate on the replacement trade (4.51%) needs to be economically equivalent to the net of the terminated trades. Any difference gets settled in cash.

• DV01 on $50M 5Y swap ≈ $22,500 per basis point
• Rate adjustment: 1 bp difference × $22,500 = $22,500 cash settlement
• Party B pays Party A $22,500 to compensate for the rate differential

Results:

Why this matters: That 91% notional reduction flows directly into SA-CCR exposure calculations, leverage ratio denominators, and G-SIB score inputs. On a $550M portfolio, the capital relief—even at modest risk weights—is meaningful.

Tolerance Settings (The Critical Control)

Tolerances are where compression succeeds or fails. Set them correctly and the algorithm finds maximum compression. Set them wrong and you either get no compression (too tight) or take unintended risk (too loose).

Typical tolerance ranges for interest rate compression:

The point is: tolerances aren't just risk limits—they're optimization constraints. The compression algorithm maximizes notional reduction subject to every participant's tolerances simultaneously. Wider tolerances give the algorithm more room to find solutions. But "wider" doesn't mean "careless"—it means thoughtfully calibrated to your actual risk appetite and hedging precision.

A common mistake: Setting tolerances based on what feels safe rather than what's analytically justified. If your portfolio DV01 is $500,000, a ±$5,000 tolerance represents a 1% change—well within normal hedging precision. But if your DV01 is $20,000, that same ±$5,000 is a 25% shift, which may be unacceptable.

Risks, Limitations, and Common Pitfalls

Risk Profile Changes (The Nuance)

Compression preserves net risk within tolerances, but the distribution of risk across the curve can shift. If you had pay-fixed exposure at 5Y and receive-fixed at 7Y, compression might net these into a single position that doesn't perfectly replicate your original curve exposure.

The test: After every compression cycle, compare your risk ladder (DV01 by tenor bucket) before and after. Any changes should be within your stated tolerances—if they're not, something went wrong in the proposal review.

Operational Risks (Where Things Actually Break)

Incomplete trade populations are the number-one cause of suboptimal compression. If you submit 80% of your eligible trades, you're leaving 20% of potential compression on the table—and possibly creating reconciliation breaks downstream.

Settlement timing matters more than people expect. Trades are terminated and replacements booked simultaneously, but cash settlements may clear on different timescales. If your systems process the termination before the cash settles, your P&L will show a temporary discrepancy (and your risk team will call you).

System reconciliation failures happen when compression results aren't automatically fed back into your booking systems. Manual rebooking of replacement trades is error-prone—one wrong rate or notional and your risk reports are wrong until someone catches it.

Regulatory Considerations

Under EMIR and Dodd-Frank, firms with large OTC portfolios are expected to engage in portfolio compression where possible. ESMA's RTS on risk mitigation techniques for non-centrally-cleared derivatives (Article 14 of the EMIR RTS) explicitly requires firms to have procedures for compression.

Documentation requirements:

• Maintain full audit trail of compressed trades (original terms, termination date, replacement terms)
• Document tolerance settings and rationale
• Report terminations and new trades to SDRs within required timeframes
• Update regulatory capital calculations immediately post-compression

Common Pitfalls (and How to Avoid Them)

Over-tight tolerances result in proposals that compress nothing. Start with wider tolerances on your first few cycles and tighten as you gain experience with the process and understand your actual risk sensitivity.

Missing or mismatched trades between your books and the compression service's records will either exclude trades from the cycle or cause reconciliation failures. Always run a full reconciliation with counterparties before submitting trade populations.

Timing conflicts arise when compression cycles run during active trading hours. New trades booked after submission but before execution create mismatches. Schedule compression runs during market close or weekends when possible.

Cash settlement disputes happen when counterparties disagree on present-value calculations. Pre-agree the calculation methodology (discount curve, day count, valuation date) before the cycle runs—not after.

Capital Impact (Why the CFO Cares)

The calculation for a mid-size derivatives desk:

Annual capital savings: $4.8 million. Against typical compression service fees of $50,000-200,000 per year, the ROI is 24x-96x. This is before counting operational savings from reduced trade count (fewer confirmations, fewer payment processing events, fewer regulatory reports).

Why this matters: Compression isn't a nice-to-have optimization. For any firm with material OTC derivatives exposure, it's a core capital management tool with measurable, auditable P&L impact.

Compression Mitigation Checklist (Tiered)

Essential (high ROI)

These prevent 80% of operational issues:

• Reconcile trade populations with all counterparties before submitting to compression service
• Define tolerance settings per risk parameter with documented rationale
• Pre-agree cash settlement calculation methodology (curves, day count, valuation date)
• Automate system feeds from compression service to booking systems

High-Impact (workflow and process)

For desks running regular compression cycles:

• Schedule compression runs during off-hours to avoid timing conflicts with live trading
• Run pre- and post-compression risk ladder comparisons by tenor bucket
• Coordinate with regulatory reporting team on SDR submissions for terminations and replacements
• Track compression ratios over time to benchmark effectiveness

Post-Compression Verification

After every cycle, confirm these before close of business:

• All terminated trades removed from risk and P&L systems
• Replacement trades booked with correct terms (rate, notional, dates)
• Cash settlements processed and reconciled
• Regulatory capital recalculated and reported
• Margin requirements updated at CCP or bilateral CSA level

Related reading:

• For clearing structures that feed into compression, see Cleared vs. Bilateral Swap Structures
• For valuation adjustments affected by counterparty exposure reduction, see Valuation Adjustments: CVA, DVA, FVA