Basis Risk

Basis risk is the risk that a derivative payout doesn't perfectly match the client's actual loss. It's inherent in all parametric products — and managing it is central to product design.

Five Sources of Basis Risk

Source	Description	Impact
Asset mismatch	The derivative covers weather; the actual exposure is to revenue or cost	Derivative pays out based on temperature, but the client's loss is driven by a more complex relationship
Time mismatch	Contract measurement period doesn't match exposure period	A monthly rainfall contract may not capture a critical 3-day event within the month
Location mismatch	Weather station doesn't represent actual project site conditions	Reference station is at the airport; the project site is 20km inland with different microclimate
Liquidity constraints	Can't perfectly hedge the desired notional amount	Available capital limits the maximum payout, leaving residual exposure
Measurement errors	Data quality issues in the reference index	Station equipment malfunction, delayed reporting, or data gaps

Basis Risk by Variable

Variable	Typical Basis Risk	Reason
Temperature	Low	Spatially uniform over mesoscale distances; airport stations representative of nearby areas
Rainfall	High	Highly localised — convective rainfall can vary significantly over a few kilometres
Wind	High	Terrain effects, turbine wake, surface roughness create hyper-local variation
Irradiance	Moderate	Cloud cover is more spatially coherent than rainfall but still variable

This ordering supports Cliff Horizon's product sequencing: temperature-based products first (lowest basis risk, most tractable), then rainfall and irradiance, then wind.

How Cliff Horizon Mitigates Basis Risk

Asset Mismatch → Scenario Simulator

The Scenario Simulator explicitly models the weather → operational impact → financial exposure chain. Rather than selling a temperature derivative and hoping it correlates with delay costs, the engine maps the causal pathway and prices accordingly.

Time Mismatch → Flexible Contract Windows

Contracts are structured with measurement windows that match the client's exposure period — daily, weekly, monthly, or seasonal. Rolling windows (e.g., "any 7-day period within the construction phase") capture event clustering that fixed-period contracts miss.

Location Mismatch → SatSure Hyper-Local Data

This is where Layer 1 provides the greatest value. SatSure satellite data provides observations at the actual project site, not just at the nearest weather station. For agricultural contracts, farm-level soil moisture data directly verifies water availability — a much better reference than a rain gauge 30km away.

NWP ensemble spatial interpolation also helps: by using multiple grid points surrounding the project site, the engine produces a location-specific probability that accounts for spatial variability.

Liquidity Constraints → Ensuro Pool

Ensuro's USDC liquidity pool provides the counterparty capital that makes parametric products possible in markets without existing weather derivative liquidity. The pool scales dynamically — if demand grows, LP allocations can increase.

Measurement Errors → Independent Oracle Strategy

Settlement uses independent, redundant data sources: NWS Climatological Reports, SatSure satellite observations, and Chainlink oracle infrastructure. Redundancy protects against single-source data quality issues.

Critically, the engine's own output is never used as the settlement reference. The engine prices the risk; independent oracles determine whether the trigger was hit. This separation avoids benchmark manipulation risk (which is a criminal offence under SFA Part 12 Division 2 in Singapore).

Measuring Basis Risk

Basis risk is measured by conditional probability β — the probability that the derivative does NOT pay out given that the insured event actually occurs.

β = P(derivative does not trigger | client experiences loss)

A perfect hedge has β = 0. Temperature derivatives typically achieve β < 0.1 (low basis risk). Rainfall derivatives may have β = 0.2–0.4 depending on the spatial distance between reference station and project site.

The engine reports estimated β for each derivative structure on the Derivatives tab, alongside the pricing waterfall — so clients can make informed decisions about the basis risk they're accepting.