In 2022, the abrupt rise in interest rates in Europe caught many life insurers off guard: within a few months, some saw their solvency ratio vary by several dozen points, not because of a change in their portfolio, but as a result of the combined effect of the fall in bond values, the revaluation of liabilities, and more volatile surrender behaviour among policyholders. This episode illustrated in a very concrete way just how much the calculation of solvency has become a dynamic, sensitive, and computationally intensive exercise.
Quantum computing could accelerate key calculations in insurance actuarial science, particularly Monte Carlo simulations of the SCR (Solvency Capital Requirement), within the Solvency II framework, which requires insurers to calibrate their capital and risk management according to an approach structured around three pillars (quantitative requirements, governance/risk management, and transparency/reporting). This article explains what quantum computing can (and cannot) contribute to actuaries.
Key Takeaways
- Why the SCR requires massive simulations (Monte Carlo)
- Where quantum computing can help: acceleration of calculation components (Amplitude Estimation)
- Current limitations: qubits, errors, emerging hardware
- Why it is necessary to prepare now (hybrid classical/quantum approach)
Actuaries Facing Performance and Precision Challenges in Solvency Calculations (H2)
Insurance actuaries are facing growing challenges in the calculation of solvency under
Actuaries facing a challenge of performance and precision in solvency calculation
Insurance actuaries are facing growing challenges in the calculation of solvency under Solvency II, which requires the estimation of the capital needed to withstand an extreme shock (a 0.5% probability of ruin over one year, i.e. a 1-in-200-year event). To achieve this, they must project the insurer’s balance sheet across thousands, or even hundreds of thousands, of economic and insurance scenarios in order to calculate the SCR (Solvency Capital Requirement).
This complexity stems first from the scale of the simulations: a life insurer may run up to 100,000 or 200,000 stochastic trajectories incorporating interest rates, inflation, equity markets, credit spreads, and policyholder behaviour, with projections over 20 to 40 years. A simple 1% rise in interest rates can alter the solvency ratio by 20 to 40 points, because it simultaneously changes the value of bond assets, that of liabilities, and the implicit options embedded in contracts. To this are added interactions that are difficult to model: a market downturn may reduce assets while also triggering mass surrenders, thereby worsening the capital requirement at the worst possible moment.
Actuaries must also incorporate emerging risks, such as climate or cyber risk, which require forward-looking scenarios over several decades despite a lack of robust historical data. Lastly, these calculations are extremely resource-intensive: producing a regulatory ratio may require billions of operations and several hours, or even days, of computation, even as management teams are demanding increasingly frequent projections in order to steer capital in near real time. In practical terms, during reporting or regulatory reporting periods, this results in tight deadlines, a limited ability to recalculate or test alternative scenarios, and a heavy dependence on costly computing platforms. Moreover, this exposes actuarial teams to periods of strain and reduced room for analysis.
In this context, quantum computing appears to be a potential lever: in theory, it could significantly accelerate Monte Carlo simulations. A Monte Carlo estimate is a method consisting in simulating a large number of random scenarios in order to assess the probability and impact of different possible outcomes, to support decision-making in a context of uncertainty, optimise asset allocation under constraints, and improve the modelling of complex correlations. In time, it would open the way to near-instant solvency calculations and a more refined exploration of extreme scenarios, enabling more dynamic and forward-looking capital management.
Is Actuarial Science a Good Candidate for Quantum Computing?
Optimisation Is Already at the Core of Modern Actuarial Practice
Optimisation lies at the heart of both quantum computing and actuarial science. To address the growing complexity of solvency calculation, actuarial science has in fact already begun to explore optimisation approaches. Hybrid and efficient methods have already been implemented in order to reduce the computational burden while preserving the required level of precision. The use of approximations, for example, makes it possible to estimate rapidly the impacts of complex scenarios without recalculating all trajectories. LSMC (Least Squares Monte Carlo) combines stochastic simulations (simulations that incorporate randomness) and regressions to approximate the value of options and guarantees across a large number of scenarios, thereby drastically reducing the number of simulations required. Insurers have also developed internal models to adapt calculations to the specific features of each portfolio and carry out targeted trade-offs between precision and computational capacity, seeking the best compromise between accuracy and computation time.
These approaches nevertheless remain optimisations of how to calculate, without altering the regulatory framework or the scope of the risks to be measured: the SCR must still cover all the risks required under Solvency II. Quantum computing follows the same logic: it does not change what must be calculated, but it could considerably improve the way in which it is done, by making it possible to process an even greater number of scenarios, to model complex correlations more finely, and to produce near-real-time projections, while preserving regulatory rigour.
Why Quantum Computing Is Relevant Without Being Magic?
Quantum computing is not intended to replace actuarial models, but rather to accelerate certain specific computational building blocks.
The potential of quantum computing is particularly evident in tasks that possess certain characteristics:
- Estimation of complex probabilities
- Evaluation of distribution tails (extreme risks)
- Exploration of a very large number of scenarios with exponential complexity
Certain quantum algorithms, such as Amplitude Estimation, therefore hold the potential to transform the way actuaries approach stochastic calculations. In the context of solvency calculation, the assessment of the SCR relies on massive Monte Carlo simulations: in order to estimate the probability that a portfolio will no longer cover its liabilities, the actuary simulates tens or hundreds of thousands of market trajectories and contractual cash flows. Amplitude Estimation makes it possible, in theory, to reduce significantly the number of simulations required to achieve the same level of precision, by exploiting the properties of quantum superposition and interference in quantum computing, as well as other phenomena such as quantum entanglement, which makes it possible to create strong correlations between complex states. In practical terms, this could reduce computation time by orders of magnitude, while maintaining a level of precision equivalent to that of classical methods.
It is important to emphasise that this approach does not replace actuarial models or regulatory rules: it acts solely on the way in which certain critical building blocks are calculated, such as stochastic projections or the valuation of complex guarantees. In other words, the actuary retains responsibility for the choice of models, assumptions, and the interpretation of results, but could benefit from gains in time and capacity that open the way to more refined analyses, the management of more complex portfolios, and a better exploration of extreme scenarios.
At this stage, however, the practical impact remains dependent on the development of quantum hardware: current computers are still limited by the number of qubits and the error rate, but the development trajectory suggests that, in the medium term, these gains could become concrete for actuarial science.
Quantum Computing in Actuarial Science: Why Prepare Now?
Quantum computing could bring tangible gains to actuarial science. It would make it possible to reduce significantly the computation time required for critical simulations, to rerun intermediate scenarios more easily, and to optimise the use of IT resources. Indirectly, these technical gains would translate into less operational stress, more time for business analysis, the upskilling of teams, and the strategic positioning of the insurer as an innovative organisation.
Even though quantum computers have not yet been industrialised, it is relevant to begin preparing for them now. Use cases are becoming clearer, methods are beginning to emerge, and teams can test prototypes through targeted POCs (Proofs of Concept) and anticipate the skills that will be required. Waiting until quantum computing becomes “turnkey” would risk missing the opportunity to learn, to plan, and to establish the right contacts in an ecosystem where talent is scarce.
To remain effective, the approach must be pragmatic and progressive. This means combining classical and quantum computing on targeted sub-problems, experimenting with prototypes or emulators, measuring the potential gains and identifying the limitations, while at the same time preparing teams. As with any disruptive technology, the key is to experiment, measure, and integrate progressively.
Sopra Steria is particularly relevant because it combines strong knowledge of the insurance sector, mastery of the quantum hardware ecosystem, and the technological expertise required to prototype and integrate hybrid solutions, while guiding teams in the exploration of this new frontier.
FAQ :
What Is a Qubit?
Qubit definition: A qubit is the fundamental unit of quantum computing, capable of existing in multiple states simultaneously thanks to the principle of superposition.
What Is Quantum Superposition?
Quantum superposition definition: Quantum superposition is a principle of quantum mechanics whereby a qubit can exist in several states at once, zero and one simultaneously, unlike traditional bits. This enables quantum computing to explore multiple possibilities in parallel.
What Is Quantum Entanglement?
Quantum entanglement definition: Quantum entanglement is a phenomenon in which two qubits become linked such that the state of one instantly depends on the state of the other, even at a distance.