Cloud computing and industry adoption trends

New accounting and regulatory changes, including GAAP long duration targeted improvements (LDTI), IFRS 17 and principle-based reserving (PBR), require increased computational demands for actuarial valuation and projections, albeit within the same close calendar time constraints. This has led many insurance carriers to re-evaluate how to strategically position their financial modelling infrastructure for the long term, including computation data and automation capabilities. A key decision point in formation of this transformation strategy is whether to establish hosted modelling infrastructure through a vendor off-premises (“cloud-managed”) or establish distributed processing capabilities directly “on-premises”. Companies seek fast, inexpensive, and robust technology solutions for their modelling operations.

This article will explore considerations for cloud-managed and on-premises infrastructure for actuarial modelling applications and explores the latest industry trends from survey data of 20 different US-based insurance companies.

Some key questions for on-premises vs. cloud-managed infrastructure for actuarial modelling applications include:

• What are the core components of modelling infrastructure and how can they be procured?
• How does cloud-managed compare to on-premises for people, process, and technology?
• What are industry trends for adoption and usage of cloud-managed infrastructure?

Modelling infrastructure component overview

Large actuarial modelling operations were traditionally hosted on-premises. The arrival of cloud-managed operations created a compelling reason to evaluate the cost-benefit of where to host modelling technology infrastructure. The industry has seen a shift of migrating modelling infrastructure to the cloud from what used to largely be on-premises solutions to a variety of new platforms. Adding to the choices, many actuarial software vendors now offer cloud-managed solutions.

There are several core components to modelling infrastructure as shown in Figure 1.

Options for procuring modelling infrastructure

For on-premises solutions, an organisation procures its own infrastructure (as shown above), locating it on the premises of the organisation’s data centre (or similar) rather than acquiring through a service provider. The modelling infrastructure and the applications are under the ownership of the company and not rented.

In contrast, companies can explore a variety of cloud-managed solutions. Cloud-managed is a method of enabling on-demand network access to a shared pool of resources, including networks, servers, storage, applications and services.

Comparisons

The insurance sector has seen a shift from on-premises to cloud-managed. However, cloud-managed operations may not always be cheaper than on-premises, and flexibility of hosting arrangements can vary, hence it’s crucial to understand options, benefits, costs and trade-offs.

Industry trends

To help understand current adoption trends and costs associated with cloud computing supporting actuarial modelling use cases, EY recently conducted a survey of 20 different insurance companies, which included a diverse mix of type (stock and mutual), size and actuarial software usage. The goal was to understand the industry practice of hosting (e.g., on-premises, cloud-managed), utilisation of distributed processing capacity, computing cost and expectations about future cloud usage.

Some of the key findings from the survey include:

On-premises and cloud-managed adoption

• Among the participants, 50% utilise cloud-based solutions exclusively, with an even distribution between those using cloud providers and those opting for the modelling platform vendor.
• Meanwhile, 5% of respondents rely solely on on-premises solutions, and 45% use a combination of on-premises and cloud-managed solutions.

On-premises and cloud-managed usage

• The industry continues to trend toward increased use of cloud-managed services for hosted technology solutions to support its actuarial modelling process, both as a complement to dedicated on-premises grids servicing peak demand periods (e.g., quarter-end), as well as its own stand-alone solution (more than 50% of the participants indicated that cloud-managed solutions are used to facilitate all their model runs).

Usage by use-case

• It is worth nothing that all survey participants indicated that on-premises grids were established for general-purpose use rather than dedicated to a particular use case (e.g., hedging) or purpose.
• When it comes to concentrated usage of cloud computing, the primary use cases are valuation, cash flow testing and capital modelling.

Computing costs

• As expected, larger companies tended to have larger overall spend. The survey results indicated that larger companies spent an average of $8.1 million (with a range from $750k to $10m+) for the year 2022, while small- to medium-sized companies averaged $1.6 million (none of them exceeding $5m).

Conclusion

Insurance companies are now addressing the challenges presented by recent accounting and regulatory changes through a strategic reassessment of their modelling infrastructure, contemplating on-premises and cloud-managed solutions. While cloud solutions provide scalability, their cost-efficiency and suitability should be assessed with care.

By understanding these opportunities, actuaries and technologists can partner to transform to a future state that chooses optimal solution sets, balances cost and usability, and brings transparency to a process that must be managed like a well-constructed factory to remain competitive going forward.

We would love to hear your views on the perspectives we have shared. Please share your thoughts, insights, and experiences in the comments section.

For further insights, please click here.

The views reflected in this article are the views of the author and do not necessarily reflect the views of the global EY organisation or its member firms.