You, as an actuary, would do well to adopt the mindset of the future and lead the way towards Actuarial work as a Service.
This is the second in our series of articles on Software as a Service (SaaS) principles and how they apply to actuarial work. The first was about Software as a Service providers and the tools they offer actuaries for better productivity. [Read it here.]
Now we’re going to talk about how the “X as a Service” mindset, applied to actuarial work, can create a better output for all involved.
First, a quick reminder of the difference between Software as a Service and Software as a Product.
| Software as a Service | Software as a Product | |
|---|---|---|
| What does it do? | Software stuff | Software stuff |
| What’s included? | Base functionality + modular add-ons | Full suite of functionality for everyone |
| How does the user use it? | Access over an internet connection | Download and install an executable file to a laptop, desktop, or other |
| What is the payment model? | Monthly (or annual) licensing | One-time up-front purchase |
| Who keeps it up-to-date? | End user | Software vendor |
Some of the biggest differences are that X as a Service providers take it upon themselves to maintain all of the underlying parts. They also make the components of X modular, so they can swap out newer, better ways of doing those components and the end-users just get better results for it. And the incentives align with frequent, regular updates and happy customers, which means a continual push for better product delivery.
When X = Actuarial work, then, how might that look? How would an actuary consider providing actuarial work as a service, rather than as a product? Here are a few distinctions.
Timing
Actuarial work as a Product
This traditional view sees actuarial work as a series of major projects happening inconsistently or infrequently. Perhaps this is to “price a new product”, which only comes out every 18 to 24 months as the marketing department schemes up a new rider. Or it’s every year, when Cash Flow Testing “season” comes around. Or when, as recently announced, that life insurance valuation and nonforfeiture interest rates are changing. AGAIN. [To be fair, this is the first change in nearly 8 years. But, really, people, you should have seen this coming for a while now.]
The infrequency and inconsistency of these large updates mean that each new thing becomes a big thing in itself. And in between, processes are forgotten, standards erode, people move to new departments or new jobs, and knowledge is lost.
Which means each time, you’ve probably going to do a lot of building the thing from scratch. That takes a lot of time and, frankly, isn’t the best use of your actuarial skill set.
And you’re likely to take on a lot of new responsibilities in order to get the thing done, on top of your current production requirements. Think you’re going to be giving either one your full attention at that point?
Plus, a rigid structure of timing means that there is no flexibility to adapt to changing situations as they arise. When assumptions are changed only once per year (in Q3, say, before the financial plan), and then new developments impact the situation soon after, those organizations without flexibility to revisit their assumptions will find themselves in trouble. They’ll have 9 months between when they should have changed something, and when they did.
That’s a pretty big risk that may go unnoticed during the annual audits. Often, it’s just been that way for so long that everyone assumes that’s the way it should be.
We think actuaries can do better.
Actuarial work as a Service
In contrast, viewing Actuarial work as a Service (AaaS) would break this down from large projects with infrequent delivery into smaller ones, with very regular delivery.
Instead of repricing a whole product all at the same time, perhaps AaaS providers have a continual churn of pricing happening, for a whole range of potential interest rates and lapse rates. Then, each month the sales department can simply pick and choose from the combinations available. They can take advantage of market opportunities, dial down certain counterparty risk, etc. The point is, those end-users don’t have to wait months for their answers, because they’re always ready.
An AaaS provider would have on-hand the pricing requirements and interventions for those possible insurance valuation rate changes before the decision was made, so that it wasn’t a surprise and it wasn’t a fire drill once it became effective. It could almost be a non-event. Oh, nonforfeiture rates are changing? Here, use this filing. It’s ready to go. Now who saw Monday Night Football?
Scope
When you perform actuarial work as a Product, the scope is often much larger. Since you’re delivering everything all at once, it becomes a very big undertaking each time. Which also means you’ve got to get every single element buttoned up before you deliver any of them.
That delays smaller or less impactful changes from being implemented sooner. It also has the potential to distract from the larger, more critical changes that need a more thorough decision process when there is a competing pressure to evaluate the smaller changes at the same time.
That leads to a distracted thought process, and could lead to a focus on immaterial changes at the expense of understanding the bigger ones.
Or the big issues take all the time, and then the smaller issues don’t get the attention they actually deserve, because there’s a latent model risk that nobody’s looked at for years because “they just never get around to it”.
Splitting the scope into smaller buckets means each piece will be able to stand on its own and have the appropriate amount of review necessary.
Modularity
With Actuarial as a Product, again, you are delivering the whole of the thing all at once. Which may lead to delays if some parts aren’t ready. Or, and this is more likely, to be giving short shrift to certain elements, in order to get them ready to ship on time.
In contrast, AaaS providers will naturally see their offerings as made up of many different, smaller modules, which can be swapped in or out as desired. Those smaller pieces could even be handed off to junior actuaries, without overwhelming them.
Plus you don’t have to delay the mortality assumption update because the expenses update is taking a long time, or vice versa.
For example, if a new macroeconomic situation develops (perhaps, say, a global pandemic?), having modularity means individual components of the lapse rates can be reviewed, analyzed, and updated independently. There is no need to wait for other parts of the model to be updated as well, if those components change more slowly. Since they won’t show the need for a change for quite a while, it’s not necessary to do so all together.
Read more on good model building practices such as modularity, auditability, and inheritability. Click here to get our FREE eBook Non-Technical Principles of Successful Actuarial Model Design.
An AaaS Use Case
An example of this, because it’s a pretty good use case, is Cash Flow Testing (CFT). Here’s a comparison between the typical way that CFT is done under a product mindset and how it might be structured using a service mindset.
| Actuarial as a Product (historical) | Actuarial as a Service (future) | |
|---|---|---|
| Purpose | Asset Adequacy Analysis, to ensure that the insurance company has enough money set aside in reserves to pay reasonable claims. | Asset Adequacy Analysis, to ensure that the insurance company has enough money set aside in reserves to pay reasonable claims. |
| Scope | In-force liabilities and in-force assets. No new liability policies. New asset purchases as available or required to continue to support liabilities. | In-force liabilities and in-force assets. No new liability policies. New asset purchases as available or required to continue to support liabilities. |
| What it does | Project cash flows of insurance policies and assets, determine if the company will have “surplus” at the end of the projection. | Project cash flows of insurance policies and assets, determine if the company will have “surplus” at the end of the projection. |
| Number of times per year | 1 | 4 |
| “Official” Date | 12/31/yyyy | 12/31/yyyy |
| “Model” Date | 9/30/yyyy | 12/31/yyyy |
| True-up from “model” to “official” date | YES – necessary because “it takes so much time to build the models that you can’t get everything done between the official date and the filing date”. | NOT NECESSARY – models are always up-to-date and require no “updating” between quarters |
| Time Commitment throughout the year | 4-6 months, all at once (October through March) | 2-3 months, spread out over increments of 0.5 – 1 month each quarter |
| Processes | OFTEN MANUAL | SIGNIFICANTLY AUTOMATED |
| Sensitivity test dates | 9/30/yyyy | 12/31/yyyy |
| Rollforwards | Incorporate all assumption changes at once, necessitating a large and cumbersome reconciliation between runs | Incorporate assumption changes as they are decided and approved, spread out throughout the year from one quarterly CFT run to the next |
| Useful for | Satisfying regulatory requirements. | Satisfying regulatory requirements. Also, confirming assumptions impact on new business, validating new sales assumptions, incorporating into financial forecasts, more frequent investment allocation recommendations, etc. |
The big reason that many insurance companies have such a headache with Cash Flow Testing is that they do it so infrequently. As a result, there is no incentive to make it much faster. One great reason to adopt a “service” mindset is that it will force you to create more automatic processes which require less manual intervention.
Trust me, if you had to do CFT every quarter, you would absolutely be finding ways to make your processes less intense and more streamlined. The fact that actuaries haven’t yet is less an indication of the difficulty of the work and more about inertia. I.e. “This is the way we’ve always done it.”
It’s just algebra, after all. Every single one of those formulas, from a lapse rate to an income statement roll-forward, is already solved. All that has to happen now is applying those to a series of data sets, and that’s just pipelines. You can do it. You can make CFT great again.
In fact, you can do this for all your actuarial processes: experience studies, pricing, even investment strategy. When you apply a service mindset to your actuarial work, then it becomes much less of a “thing” when asked to do that, and much more of “business as usual”.
The Future is Coming
With GAAP LDTI and IFRS-17 on the horizon, insurance actuaries are waking up to the fact that they will have to create more automation in their processes. These regulatory requirements mean that actuaries will be driven to adopt the actuarial as a service mindset, at least for setting reserves.
That automation mindset is likely to spill over from the actuaries producing results for LDTI and IFRS-17 into other practice areas. They will be setting the bar for regular, smaller updates, rather than large, slow changes.
Frankly, those actuarial areas which are able to make this transition will be, in our opinion, the ones which will produce better results. And with those better results, they will be able to outperform their competition from other professions, such as data scientists. They will produce a better service to their clients (insurance companies). Those companies will also produce better outcomes (more fair rates, faster repricing, better policy administration, etc.), which can lead to greater customer satisfaction, greater market share, and higher returns.
It may seem like a long road to a higher-quality end-product, but it is the better way. The evolution of the software industry from SaaP to SaaS demonstrates the value of this kind of thinking. You, as an actuary, would do well to adopt the mindset of the future and lead the way towards Actuarial work as a Service. We’ll meet you there.