Affiliate member and CEO of Reitigh Software Limited, Brian Walsh, describes the dangers and opportunities as actuaries get to grips with new tools.
Data is at the heart of all actuarial tasks. Traditionally, actuaries have relied heavily on tools like Excel or Access to handle data manipulation requirements. However, the advent of data science and the increasing accessibility of programming languages such as R and Python are reshaping how actuaries approach their work. This evolution heralds the rise of the ‘citizen developer’ within the actuarial community, signalling a shift towards more traditional programming practices in the search for automated and scalable solutions. But as with any evolution, change brings both opportunities and challenges.
For actuaries venturing into the realm of programming, the benefits are obvious. The ability to automate tasks, analyse vast datasets, and create scalable solutions is a significant leap from the constraints of Excel and Access.
However, software engineering is a discipline in its own right, governed by principles that ensure code is not only functional, but also efficient, maintainable, and scalable.
An algorithm that works flawlessly with a small dataset or a few lines of code often fails when scaled up. This reality requires that actuaries stepping into the role of citizen developers must equip themselves with core software engineering principles to ensure their solutions are robust.
Excel has been the actuary's Swiss army knife for decades, automating tasks and handling data with a certain degree of efficiency. Yet the limitations of Excel, particularly its lack of auditability, as well as row cap and performance issues, are well known.
These constraints have led to what is often referred to as ‘end user computing’ issues (EUCs). Excel's capabilities are stretched beyond their limits, resulting in operational risks, inefficiencies, and potential errors.
The transition to programming languages like R and Python opens a new realm of possibilities.
These tools do not suffer from the same limitations as Excel, offering the ability to handle larger datasets and perform more complex analyses with greater speed and efficiency. However, this power also comes with increased responsibility. If you can shoot yourself in the foot with Excel, you can blow your leg off with Python.
As actuaries embrace coding, the distinction between citizen developers and professional developers becomes increasingly relevant. This distinction raises questions about code maintenance, execution, and responsibility.
In particular, a lot of thinking is required around where the responsibility boundaries exist between actuarial and IT departments. When actuaries develop solutions, should these be considered mere specifications to be handed over to developers in the IT department for ‘professionalisation’. Or should actuaries retain ownership, being responsible for both the maintenance and execution of their code?
This dilemma extends to the execution environment – whether code should run on individual laptops and desktops or be deployed on servers. If actuaries are responsible for their solutions, does this extend to managing the server infrastructure as well?
These questions underscore the need for clear guidelines and policies to navigate the responsibilities and roles of citizen developers within the actuarial profession.
R has been included in the core actuarial syllabus for all trainee actuaries. But the syllabus revolves around how to use R to solve statistical challenges, rather than core software engineering principles.
If actuaries dive headfirst into writing millions of lines of code without the appropriate training, the current challenges around EUCs will pale in comparison to the challenges the industry will face in years to come.
That said, the move towards programming and data science within the actuarial profession is not merely a matter of following trends. It's a response to a real need for faster, more efficient solutions that traditional IT departments, often bogged down by processes and a lack of specific domain knowledge, cannot always provide.
There is a real issue that needs to be solved here. By empowering actuaries to develop their own solutions, the profession is becoming more agile, more engaged, and better positioned to tackle the challenges of the future.
As actuaries become more proficient in programming and data science, they are not only enhancing their capabilities, but also facing new responsibilities and challenges.
This evolution requires a delicate balance between embracing the opportunities presented by these new tools and recognising the importance of software engineering principles, collaboration with IT professionals, and the ongoing need for education and training.
As we navigate this transition, the actuarial profession stands on the cusp of a new era, one that promises greater innovation, efficiency, and impact.
