AfterLife 2/4 – Future-proof professionalism

Posted by

(Reading time: about 5 min)

I had written about what next after actuarial qualification in the previous blog: AfterLife 1/4 – I have qualified; what do I do now? Feels empty

The most important aspect of professionalism, to be ready for the future, needed a separate note. In this world of rapidly transforming capabilities, a young actuary must be keenly aware of how professional expertise needs to evolve.

So the critical advice for professional excellence is:

4. One that is future-ready 

Real learning starts after you qualify. Without continuous sharpening, any foundation withers away. Actuarial expertise is data-based. We are witnessing tectonic changes in generation, accessibility, and capability to process data. Actuaries need to understand the implications of exabytes of data on actuarial science and how risk-sharing frameworks are transformed. More importantly, the insurance sector will be disrupted like banks, hotels, and transport sectors have witnessed. 

Three main disruptive innovations are cloud, blockchain, and quantum computing. 

What does cloud mean to actuaries? Cloud enables an API (Application Programming Interface) driven economy where very different platforms speak to each other, and all the data is on the web. The traditional expertise of dealing with structured data to derive risk parameters will be challenged with vast volumes of unstructured data. Group-aggregate analytics will gain importance rather than individual events (for example, target group average of specific sickness). It might be apt for insurance, as social media influences people big time. Up to 40% of an individual’s general health can be influenced by lifestyle and behavioral factors alone [decadoo]. As a large volume of group statistics becomes available, they become more reliable and statistically credible.

Traditionally, Excel-based data processing has been by columns of data, with massive data on the cloud that wouldn’t work. You don’t want to copy the entire database stored in a distributed structure onto your hard drive. One needs to start thinking about a bunch of related rows that need to be processed; welcome PySpark! Knowing how to query the data requires getting comfortable with PySpark’s SQL Aggregate window functions and WindowSpec. More importantly, it is crucial to shift logical steps to solve problems in terms of many related rows, rather than one column linked to the other columns of the entire database.

In an API-driven economy, the pricing, reserving, and risk systems may not be seen as separate workflows. As a result, reusable actuarial models can proliferate. These systems will need to be redesigned into many parts that take in structured data input to provide universal outputs to be used downstream. This means one will need to know JSON and other data output formats that can be parsed through R, python, or PySpark. .

What about blockchain? Blockchain enables peer-to-peer value exchange without the need for costly market makers and legal backing. These are Regulators and Governments for now. As a result, very different constructs for risk management become possible, which are simply not possible with the current ecosystem. 

For example, insurance capital is the sum of the capital covering the risk associated with the expected value of the insurance pay-outs plus the capital covering the operational risk. It is not feasible to separate these two unrelated parts of the capital in the current insurance structure. Blockchain enables tokenizing these two capital parts separately, creating the possibility of a totally different group of capital providers or investors. 

Think of this as the two sides of a coin; the first is the better returns to policyholders from separating risk cover and investment needs (like in popular unbundled unit-linked plans). The other side of the coin is shareholders; they get better returns by unbundling risk capital and operating capital. 

And then, there is the possibility of project-based professional services or mutuals enabled by blockchain. Why not white-labeled underwriting, actuarial, or accounting services like white-labeled ATMs? The technology makes it feasible to develop verifiable/immutable history and work documentation. The relevance of institutional knowledge will reduce, which is there because of the non-transparent or bundled nature of these capital constructs in the first place. 

On top of these structures, smart-contracts can provide significantly better enforcement constructs with minimum trust costs. If this sounds far-fetched, read about Etherisc here. Etherisc completed the token generating event and rolled out flight delay and crop insurance on Ethereum. 

Insurance is fundamentally a P2P construct, more so than basic banking. Insurance thrives on the ability to share between large numbers; hence expect this to be a very different space enabled by the blockchain when you retire. 

If you want to know more about how the Blockchain economy can disrupt the current economic framework and how it can lead to new social structures, please read here

Even more disrupting will be quantum computing. While the chip density is approaching the limits of the deterministic physical world, quantum computing is set to transcend these physical limits. Traditionally separate areas, physics, mathematics, and coding, have converged into quantum computing. This new frontier will need a new breed of talents navigating effortlessly in this multidimensional space. For actuaries, the stochastic within stochastic scenarios that we can only dream of will be within reach. We have just become aware of the capabilities; it is like a thick fog that hides the mountains of possibilities that we yet don’t know exist.  

New actuaries should start getting comfortable with these tech innovations to be future-ready. 

  1. Revisit some of the linear algebra, differential equations, and Bayesian statistics. I had forgotten all my linear algebra. While going through quantum computing learning (a good starting place is Quantum computing for the very curious ) had to be refreshed. This led to the discovery of the MIT Open Learning Library. You will find some excellent MIT courses on linear algebra and quantum mechanics. I really loved most the visualized algebra learning at 3blue1brown.  
  2. You will need good knowledge of the current actuarial modeling ground to adapt to new ways. Particularly the theory behind multistate decrements and GLMs will be helpful. Getting hands-on mode on the R-codes on the following prescribed books will help; for those brave, I suggest replicating in python and JavaScript. [Predictive Modeling Applications in Actuarial ScienceGeneralized Linear Models for Insurance DataModeling Mortality with Actuarial applications & Regression Modeling with Actuarial and Financial Applications]. 
  3. MOOC platforms like Udemy or Coursera are very convenient. Start with python or R courses; if you can code in both, even better. Coding/architectural literacy will help to converse with hordes of data analysts and developers. 
  4. MOOCS will help you get the dummies version of knowledge, which should be enough. You can begin with knowing a bit about HTML, CSS, JavaScript, JASON, relational databases, and distributed database storage. 
  5. Highly recommend Andrew N G’s introductory courses for machine learning and AI.   
  6. Keep watching Humble book bundles on an excellent bundle of books, video courses on python, machine learning, or blockchain; you also get to do charity alongside. 
  7. Subscribe to Medium, and you will get faster updates and innovative ideas. 
  8. For the blockchain highly recommend the books by Andreas Antonopoulos.
  9. Twitter space is exploding with good quality advice; if you scroll through likes in my Twitter profile, you can find links to many quality books available online for coding, algebra, and machine learning [example].  
  10. Many well-known education institutes are partnering with new-gen edTech to offer learning courses. I am currently going through the Digital Health and Imaging course by TalentSprint & the Indian Institute of Science. I find the course content, teaching, and sheer knowledge and talent depth in my learning cohort quite impressive! I always regretted that I didn’t complete my ME Integrated from there; finally, I will get some consolation.

This seems like a lot, but gaining awareness in these areas isn’t that hard, and you have got a lifetime of learning anyway!

It helps to be tuned into InsurTech space. The Digital Insurer (TDI) is a good repository of InsurTech start-up information and articles. In addition, the Big 4s and tech giants regularly publish recent trends. I liked the white papers from BCE start-ups, detailed and fascinating articulations. They better be; venture capitalists invested a record $25.2 billion in blockchain companies in 2021.

Overwhelming space of information

Beginning to learn new tech can be pretty exhausting. One quickly experiences course fatigues, accumulating tons of e-books, flagging many medium articles, bookmarking hundreds of pages, or subscribing to many newsletters. All happen. But you will quickly get over them and find your pace. 

The beauty is that this vast amount of ever-increasing material also makes it possible to curate a way for you to gain expertise, but discovering that suitable method can be overwhelming. Small chunks at a time will surely help, rather than tangentially chasing every link the way I did. Sites like Medium have good articles on approaches that helped others to get grips with a fast-changing landscape and early insights into the fast-evolving landscape. 

Most importantly, keep linking your learning to core fundamental actuarial concepts. Our specialty lies in bringing the holistic risk sense to a data-rich world.