Broker-dealers predominantly send uncompressed (atomic/gross) execution-level trades to Financial Market Infrastructures (FMIs)—such as exchanges, Central Counterparties (CCPs), Central Securities Depositories (CSDs), and Trade Repositories—for initial post-trade processing. Compression (or netting) is generally performed by the market infrastructure itself or through specialized multilateral compression utilities after trade capture and matching, though specific pre-submission aggregation practices exist depending on the asset class and workflow.
1. Compressed vs. Uncompressed: Industry Practices
Whether trades travel gross or aggregated depends on the stage of the post-trade lifecycle and the type of market infrastructure:
- Exchange-Traded & CCP-Cleared Flows (Uncompressed Ingestion → CCP Netting):
- At Ingestion: Execution venues feed individual, uncompressed fills (“street-side” executions) directly into the CCP (e.g., DTCC/NSCC in the US, LCH, Eurex Clearing, or NCL/ICCL in India). Broker-dealers also report individual, granular fills to Trade Repositories (TRs) and regulators (e.g., FINRA CAT, MiFIR) to preserve an exact audit trail.
- During Clearing: Once matched and novated, the CCP applies Multilateral Netting—offsetting buys and sells in the same CUSIP/ISIN for the same settlement date—to compress thousands of gross executions into a single net delivery or payment obligation per clearing member.
- Institutional Block & Allocation Workflows (Bunched/Average-Price Compression):
- When an institutional broker-dealer executes a large order via hundreds of partial fills on an exchange, sending every tiny fill to the buy-side client or custodian creates operational drag.
- Practice: The broker-dealer aggregates those uncompressed street fills into a single bunched block trade using a Volume-Weighted Average Price (VWAP) or average price, and then splits it into client-level allocations sent to matching utilities (e.g., DTCC CTM) and custodians/CSDs for settlement.
- OTC Derivatives & Portfolio Compression (Post-Novation / Multilateral Compression):
- Bilateral OTC trades (like interest rate swaps or CDS) are submitted uncompressed to clearinghouses or trade repositories.
- Periodically, dealers run Multilateral Portfolio Compression (via services like TriOptima or LCH SwapClear), tearing up offsetting economically redundant contracts across multiple dealers to reduce gross notional exposure and capital requirements.
- Bilateral Netting / Pair-Offs (Pre-Settlement Compression):
- In fixed income or repo markets where a CCP isn’t used, two broker-dealers with offsetting trades for the same security and settlement date may bilaterally agree to “pair off” the trades—canceling the physical security movement and settling only the net cash difference.
| Lifecycle Stage / Channel | State of Trades Sent | Primary Mechanism / Reason |
| Execution Venue $\rightarrow$ CCP | Uncompressed (Atomic fills) | Real-time novation, risk management, and exact trade matching. |
| Broker $\rightarrow$ Trade Repository / Regulator | Uncompressed (Atomic fills) | Regulatory surveillance, auditability, and market abuse detection. |
| Broker $\rightarrow$ Matching Utility / CSD | Aggregated / Allocated (Average price) | Grouping partial fills into institutional allocations before settlement. |
| CCP / CSD $\rightarrow$ Settlement Bank | Compressed (Netted obligations) | Minimizing liquidity needs and physical security transfers. |
2. What Should Be the Unit of Measurement for Trades?
Because post-trade infrastructure spans ingestion, risk, clearing, and settlement, relying on a single metric distorts capacity and cost analysis. You should measure trades across four distinct units, depending on the layer you are sizing or billing:
- Trade / Execution Count (Atomic Fills) — For Ingestion, Matching & Latency
- What it measures: Every individual fill generated by a matching engine.
- Why use it: Algorithmic trading has shrunk average order sizes while exploding fill counts. Ingestion gateways, trade capture engines, and regulatory reporting pipelines scale based on message/fill count, regardless of dollar size.
- Net Settlement Instructions (Deliveries/Receipts) — For CSD Settlement & Custody Costs
- What it measures: The final post-netting obligations sent to a CSD (e.g., DTC, Euroclear, NSDL/CDSL).
- Why use it: CSDs and custodians typically charge per settlement instruction, and settlement fails happen at the instruction level. Comparing Atomic Fills to Net Settlement Instructions gives you your Netting Efficiency Ratio (often 95%+ compression in liquid equities).
- Notional Value / Principal Amount — For Risk, Margin & Capital
- What it measures: Total monetary value of the underlying assets ($\text{Quantity} \times \text{Price}$ for equities; face/par value for fixed income and derivatives).
- Why use it: CCP initial/variation margin models, intraday liquidity buffers, and regulatory capital charges (like Basel III leverage ratios) are driven by notional exposure—not how many messages generated it.
- Quantity (Shares / Par Value / Contracts) — For Asset-Class Specific Normalization
- Equities & ETFs: Number of shares/units.
- Fixed Income / Bonds: Par (face) value (e.g., $1,000 denominations), since clean/dirty prices fluctuate with yields and accrued interest.
- Listed Derivatives: Number of contracts (standardized lot multipliers).
- FX: Base currency amount.
The mechanics of compression vs. gross ingestion and unit sizing cover the structural flow, yet several operational, architectural, and regulatory realities dictate how post-trade pipelines actually behave in production—especially as global markets compress toward T+1 (and upcoming 2027 European T+1 deadlines) and expand central clearing.
1. The “Netting Set” Boundary (Why Trades Often Fail to Compress)
In theory, a CCP nets all buys and sells in the same ISIN/CUSIP. In practice, netting only occurs within a strict “netting set” (or netting bucket). If any attribute in the tuple below differs, the CCP or CSD cannot compress those trades together, forcing separate settlement instructions:
- Clearing Account Type: House (proprietary) trades cannot be netted against Client (agency) segregated accounts.
- Settlement Location / CSD: A dual-listed stock settled in Euroclear cannot net against a leg settled in DTCC without a cross-border realignment instruction.
- Currency & Settlement Date: Even a 1-day mismatch (e.g., T+0 cash trade vs. T+1 regular-way trade) breaks the netting bucket.
- Capacity & Tax/Regulatory Flags: Short sales, specific tax-lot designations, or restricted securities frequently require gross reporting or segregated settlement buckets.
2. Unique Identifiers & Lineage Across Compression
When 5,000 atomic fills are compressed into a single net obligation or institutional allocation, regulators and auditors still demand end-to-end traceability back to the original execution:
- UTI (Unique Transaction Identifier) & UPI (Unique Product Identifier): Mandated globally (under EMIR Refit, CFTC, SEC rules) to link parent executions to child allocations, novated CCP legs, and compressed swap tear-ups.
- The Lineage Trap: If your post-trade architecture drops the mapping table between the Execution IDs and the Netted Settlement ID, you cannot answer regulatory inquiries (like SEC Rule 13h-1 Large Trader or FINRA CAT) or process corporate action claims on failed/unsettled portions of a netted block.
3. Settlement Fails, Partialing, and “Shaping”
Compression creates a single point of failure: if a broker-dealer is due to deliver a net block of 100,000 shares to a CSD and is short by just 100 shares (e.g., due to a delayed stock loan), the entire 100,000-share instruction risks failing.
- Shaping: To prevent massive settlement fails and steep regulatory penalties (such as CSDR cash penalties in Europe), brokers and CSDs intentionally de-compress large net obligations into smaller standardized chunks (e.g., maximum clips of €10M or 50,000 shares).
- Auto-Partialing: Modern CSDs and CCPs (like DTCC’s Continuous Net Settlement) automatically split a netted delivery so the 99,900 available shares settle on time, leaving only the 100-share deficit rolling forward into the next day’s net obligation.
4. Accelerated Settlement (T+1) & Intraday Liquidity Compression
With North America and India on T+1 (and the EU/UK transitioning in October 2027), the operational window between trade execution, matching, and netting cutoffs has shrunk from ~24 hours to just a few hours on Trade Date (T):
- Missed Cutoffs = Gross Settlement: If an institutional allocation or prime brokerage give-up is affirmed after the CCP’s evening netting cutoff (often ~22:00–23:59 on
T), it misses the multilateral netting batch. It must then settle gross (uncompressed) trade-for-trade, spiking intraday liquidity demands and CSD fees. - Intraday Netting Cycles: Rather than a single end-of-day batch, FMIs are increasingly shifting to multiple intraday netting and settlement runs to recycle cash and collateral faster.
5. Messaging Standards & Payload Bloat (ISO 20022 vs. FIX)
- Front-to-Middle (FIX): Uncompressed fills travel via lightweight FIX (
ExecutionReport/TradeCaptureReport) messages optimized for microsecond throughput. - Middle-to-FMI (ISO 20022): Clearinghouses, CSDs, and payment rails are standardizing on verbose ISO 20022 XML schemas (
sese,semt,pacs). Sending uncompressed atomic fills in ISO 20022 format creates massive payload bloat—making pre-CSD aggregation/allocation essential not just for settlement fees, but for network and parser throughput.
6. Expanding Regulatory Clearing Mandates (e.g., U.S. Treasuries)
Historically, massive segments of fixed income and repo markets settled bilaterally without a CCP. Under the SEC U.S. Treasury Clearing Mandate (phasing in through 2026–2027), cash Treasury and repo trades executed by broker-dealers must be centrally cleared. For firms building or scaling post-trade infrastructure, this dramatically increases the volume of gross fixed-income trades routed to FICC (Fixed Income Clearing Corporation) for multilateral netting, shifting balance-sheet capacity calculations from bilateral pair-offs to CCP margin and sponsored-repo models.