Market scenario generator of the year: Conning

Aakash Dhage/Unsplash

Key points Summary

• Conning’s GEMS^® Economic Scenario Generator meets demand for transparent modelling in a volatile market environment.
• The platform models the global economy, producing realistic outcomes across various asset classes and indicators.
• GEMS captures interconnectedness between markets, allowing users to test business models under different scenarios.
• The transition to a cloud-based software-as-a-service model enhances scalability and reduces costs for clients.

These summaries are generated by an LLM, but edited by humans.

In an environment where market shocks are more frequent and more complex, financial institutions are demanding transparent scenario models that capture real market behaviour and offer greater transparency. Conning’s GEMS^® Economic Scenario Generator has met that challenge.

The simulation platform models the global economy over multiple time horizons and across different currencies. It produces realistic distributions of possible future market outcomes, covering a broad range of asset classes and economic indicators. This allows users to test business models, capital strategies and investment decisions under expected conditions and more extreme scenarios. The platform combines this breadth of modelling – spanning sovereign and corporate bonds, inflation measures, equity markets and derivatives – with a consistent framework that links the underlying economic factors.

Daniel Finn, Conning

This coherence stems from a single, integrated process that parameterises the global economy as a single interconnected system. By capturing dependencies within and across markets, GEMS represents how stress in one economy or asset class can ripple through others. This integrated design supports a wide range of uses, from asset allocation and capital modelling to climate risk, valuation and hedging.

The approach proved its value during the market turbulence of 2025, when volatility spiked across asset classes. “In recent years, we’ve seen shocks spreading more quickly across markets and asset classes,” says Daniel Finn, managing director at Conning. “That interconnectedness means models need to reflect not just isolated risks, but how stresses are transmitted through the global system.”

GEMS is built on robust models and a disciplined approach to calibration. Unlike systems that focus narrowly on recent volatility, it uses a through-the-cycle method based on long histories of market data combined with expert judgement from Conning’s quantitative finance team. The model recognises the potential of extreme but plausible events, allowing institutions to explore tail risks before they occur.

Conning’s models already captured such scenarios within their simulations, so no emergency recalibration was required. The firm later offered an optional ‘global jump’ parameterisation, which amplifies correlations between asset classes to explore the impact of more severe crises on capital requirements and strategic allocation decisions. This allows users to adjust the model to reflect different risk appetites or regulatory settings while keeping relationships between markets internally coherent.

“As markets evolve, we see a real divergence in how firms want to use stochastic models,” Finn adds. “Some are stress-testing resilience under severe downturns, while others are exploring long-term climate or demographic scenarios. Our role is to give them a framework that is both technically robust and adaptable to those different perspectives.”

Conning is unusual among software providers in that it also manages institutional assets as both modeller and investor. That dual perspective informs its approach to calibration and scenario design. This helps ensure the models are mathematically rigorous while remaining grounded in how markets behave under stress and recovery.

Every model and parameter within GEMS is documented and accessible through the user interface, supported by quarterly validation reports that allow clients to review and approve each parameter update. This transparency supports internal governance processes and builds confidence in model outputs. “This is a key consideration for our clients,” says Finn. “We’re seeing far greater scrutiny of scenario frameworks from regulators and boards.”

In the past year, GEMS has added the ability to model credit default swaps (CDS), aligning CDS pricing with the platform’s corporate bond model to deliver consistent valuations across instruments. This allows users to assess the risk profile of hedged portfolios and evaluate different hedging strategies with greater precision.

GEMS’ calibration toolkit has also been refined so that users can more easily re-parameterise models to align with internal capital assumptions, long-term economic views or regulatory targets such as one-in-200 shocks. Enhanced fitting algorithms for interest rate and credit curves improve the representation of illiquid markets, while new options smooth initial data inputs to eliminate noise. Together, these tools allow the institutions to explore the models’ behaviour at a more granular level.

All the winners

Markets Technology Awards 2026: March of the riskbots 

Will a new wave of AI-powered assistants live up to expectations?

GEMS also supports risk-neutral modelling, used for International Financial Reporting Standard (IFRS) 17 and long-duration targeted improvement requirements. Conning simplified the calibration interface and enhanced the least-squares Monte Carlo method to help new users adopt the framework more easily and meet the stringent martingale tests required for pricing consistency.

Finally, the firm’s move toward a full software-as-a-service model is reshaping how clients deploy and scale their scenario generation. By migrating from desktop installations to a cloud-based architecture written in modern programming languages such as Julia, GEMS now offers faster run times, on-demand computing power and reduced infrastructure costs.

Looking ahead, Finn sees greater demand for transparency and flexibility rather than ever-greater model complexity. “Clients increasingly want to integrate scenario generation into wider decision-making processes – whether that’s strategic planning, capital management or sustainability analysis,” he says. “The opportunity is to make these tools more accessible, connected and responsive to fast-changing economic realities.”

Sponsored content