Orbis Securities
Back to Research
01 May 2026 · LIQUIDITY OPS · QUANT

Tier 1 Aggregation: Why LP Selection Outperforms LP Count

Orbis Research DeskSYDNEY
Tier 1 Aggregation

The conventional metric for prime broker liquidity quality — "how many LPs do you connect to?" — has become a noise variable. Across a sample of six institutional FX desks reviewed in Q1 2026, the relationship between LP count and effective execution quality was statistically weak. What did matter was stream curation: which LPs, in what weighting, with what fallback logic.

Part I — What LP Count Doesn't Tell You

1. Stream Overlap

Best for understanding

Why a desk with 18 LPs may be functionally identical to one with 9.

Mechanism

  • Multiple Tier 1 LPs source from overlapping prime-of-prime pools, yielding correlated quotes.
  • Internal aggregator collapse logic often hides this from the trading desk.
  • "Effective" LP count after dedup typically runs 40–60% of nominal.

Takeaway: Diligence questions should target unique pricing sources, not connection count.

2. Selection Bias Under Volatility

What we observed

  • During the March 2026 yen volatility window, 4 of 12 sampled LPs widened spreads by 3–8x within 200ms.
  • Aggregators that statically retained those LPs in the top-of-book stream materially degraded fill quality.
  • Curated streams that dynamically de-weighted under volatility outperformed by ~22 bps on the affected pairs.

Takeaway: Volatility is when LP curation logic earns its keep. Static rosters underperform.

Part II — What Actually Drives Execution Quality

3. Stream Curation Methodology

  • Active per-pair LP weighting based on rolling fill quality.
  • Time-of-day curation (different LP mix per session).
  • Adversarial flow detection and per-LP rejection asymmetry monitoring.

Takeaway: Counterparty diligence should ask about curation methodology, not roster size.

Putting It All Together

The next generation of LP diligence questions is less "how many" and more "how do you decide". Counterparties that can answer the second question with operational specifics — per-pair weighting, volatility de-weighting, mark-out feedback loops — are demonstrably better-aligned with execution outcomes than those who lead with their LP roster.

Sources

  • Orbis Securities aggregation logs, Jan–Apr 2026 (internal sample, n=6 institutional desks).
  • BIS Triennial FX Survey 2025 — APAC venue concentration data.
  • Industry TCA reports, 2024–25 (Bestex Research, ITG, public summaries).