Live FX Hedge Surveillance on a Stateless Mark-to-Market API
The Client
The client runs a strip of monthlyUSD/MXN forwards hedging the FX risk on a Latin America private credit facilityof roughly $30M. Hedge coverage is sized to expected monthly income (~2% yield,i.e. ~$600K/month).
The hedge book was managed through Deaglo’s FX advisory,but the client had no internal, real-time view of marks, per-pairmovements, or covenant compliance, every status check meant a manual pull.
The client was among the first to be granted early "research preview"access to the Deaglo R&D API, and has since become an ongoing designpartner, collaborating directly on the endpoint's roadmap, surfacingrequirements, and shaping priorities ahead of broader release.
Key Takeaways
Methodology
Challengue
- No single source of truth. Positions, strikes, tenors, and counterparties lived across confirms and broker statements, not one register.
- No point-in-time history. There was no immutable weekly snapshot to compare movement week over week.
- Covenant exposure. The facility's participation agreement imposed hedging obligations (100% notional cover, reserve mechanics, counterparty standards, and a hedge summary owed with each certificate) that required continuous monitoring.
- Cost of poor visibility, a live example: As the loan amortized from ~$30M to ~$27.5M, required coverage fell to ~$540K/month, leaving the book ~$60K/month over-hedged. The optimal remedy, layering small offsetting forwards matched to each settlement date rather than unwinding (which would crystallize MtM), is only actionable with live marks and raw forward points at the settlement-date level.
Solution
Rather than a heavy weight white-label integration, Deaglo granted whitelisted "research preview" accessto its R&D API, specifically a stateless mark-to-market endpoint.The client could POST a batch of caller-supplied positions and receive livevaluations priced against the Deaglo’s FX Intelligence Engine.
Key properties:
- Stateless. No need to persist positions in the Deaglo’s backend. Send trade parameters (notional, strike, maturity, barrier where relevant); get a live quote on the fly. Supports vanillas, barriers, and forwards, up to 200 positions per request.
- Raw data, not just a number. The endpoint returns executable bid / ask spreads and forwardpoints, not only a final MtM dollar figure, letting the client run its own calculations and reconcile to the platform UI.
- Portfolio-aware. Per-position marks plus portfolio totals across mixed instrument types.
- Scales with usage. A baseline activity allowance is set, expanding as volume and functionality grow.
Implementation
The client wired the feed into a fully internal automation pipeline:
- Single source of truth. An FX Hedge Register plus an immutable weekly snapshots database in the Workspace Platform.
- Live dashboard. A Maestro generated artifact showing the hedge book,surveillance metrics, and the per-pair view, hosted on the client's Internal Platform.
- Scheduled refresh. A Monday 09:00am job that refreshes the register, writes the weekly snapshot, and updates the risk report.
- Compliance layer mapped directly to the facility's hedging covenant.
Work was split by role: Internal IT owned the integration; operations loaded the book; the FX lead validated positions against the live broker book; and the risk analyst produced weekly per-pair movement, MtM, and VaR / Expected Shortfall figures.
Results
- A weekly, automated FX hedge snapshot replacing ad-hoc manual pulls.
- Live per-pair MtM and covenant monitoring tied to actual obligations.
- The ability to identify and structure optimizations proactively, for example, neutralizing theover-hedge with a notional offset instead of realizing a loss on the existing strip.
- A lightweight, low-lift path tovalue a research-preview API rather than a full integration, deferring heavier platform build-out until volume justified it.
