Credibility Weighted

The Actuarial Intelligence Newsletter

January 2026

Credibility Weighted scans academic preprints, regulatory publications, industry news, and professional body updates daily, then surfaces what's relevant to actuaries.

We're named after the principle: blend prior knowledge with new evidence, weight by reliability, update beliefs. That's what actuaries do. That's what this newsletter does.

1/1 Renewals: Capacity Returns

The January 1st, 2026 reinsurance renewals tell a consistent story: capacity is available and pricing is softening.

What the Brokers Report

Guy Carpenter notes price reductions across Marine, Energy, and Technical lines. Capacity continues flowing into specialty, diversifying from property into an already competitive market.

Gallagher Re highlights cyber as particularly capacity-rich—levels far exceeding those available a few years ago.

Reading the Market

The hard market that followed 2017-2022 catastrophe losses is easing:

Property cat: Still the tightest segment, but capacity is returning
Specialty lines: Competitive across most classes
Cyber: Moved from capacity-constrained to capacity-abundant in roughly three years
Casualty: Capacity available, though social inflation concerns persist

For cedants, this means better terms and more negotiating leverage. For reinsurers, return pressure intensifies.

ILS and Cat Bonds

Gallagher Securities observes that the "innovation premium" on cyber cat bonds has largely disappeared. Early cyber cat bonds commanded extra spread for unfamiliarity; investors are now comfortable with the asset class.

Lane Financial projects cat bond market returns around 6% for 2026—reasonable in absolute terms, but reflecting spread compression from capital inflows.

Practical Implications

Buyers: Consider locking multi-year terms while conditions favour cedants. Explore alternative structures—sidecars, collateralised quota shares—while capacity is abundant.

Sellers: Differentiation becomes more important when capacity is plentiful. Speed, service quality, and genuine specialty expertise matter more than price-cutting.

Themes From This Month

Systemic Risk Thinking Expands

The Bank of England's Financial Policy Committee December record focuses increasingly on system-wide dynamics—liquidity channels, interconnectedness, feedback loops—rather than firm-level solvency alone.

Their stablecoin holding limits paper applies similar thinking: how does stress in a systemically important stablecoin propagate through counterparty exposures?

This macro-prudential approach will likely extend to insurance supervision, particularly for large life offices with significant asset-liability interdependencies.

Alternative Capital Scale

Current figures: $124 billion in alternative reinsurance capital, $19.6 billion in sidecars ($17.9B property, $1.7B casualty).

The practical implication: ILS investors evaluate risk differently from traditional reinsurers—portfolio diversification matters as much as loss distributions. This affects program structures and negotiation dynamics.

AI Implementation Phase

Industry commentary for 2026 focuses on operationalising AI rather than exploring it—governance frameworks, performance monitoring, documentation requirements.

The CAS Artificial Intelligence Working Group continues developing guidance. Model risk management expectations are rising across the profession.

Research Papers

Fairness Constraints in Insurance Pricing

arxiv.org/abs/2512.24747v1

As machine learning enters production pricing, questions about discriminatory outcomes become unavoidable. This paper addresses the technical challenge directly: how do you embed fairness constraints in a pricing model while still managing profitability and risk selection?

The approach uses multi-objective optimization—treating fairness metrics as explicit objectives alongside traditional actuarial goals. Practically useful because it moves the conversation from "is our model fair?" to "what fairness-accuracy trade-off are we willing to accept?"

Machine Learning for Catastrophe Bond Pricing

arxiv.org/abs/2512.22660v1

Applies ML methods to predict catastrophe bond coupons. The paper examines which features drive coupon levels: sponsor characteristics, structural features, peril type, market conditions. Useful for anyone involved in ILS structuring or trying to understand secondary market pricing dynamics.

Bank Systemic Risk Under Geopolitical Shocks

arxiv.org/abs/2512.20515v1

Emerging markets focus, but the methodology has broader application. The core problem: how do you stress test for correlated shocks that don't appear in your historical loss data?

Parametric Insurance Economics

arxiv.org/abs/2512.21973v1

Examines when parametric structures make more economic sense than traditional indemnity coverage. Useful framing for conversations about parametric products—moves beyond "parametric is innovative" to "here's when parametric actually creates value."