Three ways capital modelling is evolving

Published in: Risk, Risk management, Risk Models, Capital management, Capital Models, Regulation, Solvency II, RBC Worldwide, Software - IT, People

Companies: Conning

Ahead of the Insurance ERM Americas conference in New York on 16 September, InsuranceERM asked Daniel Finn, managing director and head of North America risk solutions at Conning, to share his thoughts on how capital and economic modelling is evolving, and how new technology might influence developments. 

Extending the platform

“In the past, a lot of the focus has been on proving capital modelling is a worthwhile process. Specifically, there was a lot of time and effort spend on showing these models can help companies make better decisions.

"Boards take it as a given they're going to do this type of modelling"

“Now, with a wide range of regulatory requirements like the own risk and solvency assessment (Orsa), and Solvency II, most companies’ boards take it as a given they're going to do this type of modelling.

"That has allowed modellers to focus on extending the platform to create a more realistic simulation of the company's future results.”

Three areas of focus

Daniel Finn, Conning“There are three areas where we're seeing a lot of focus from our clients:

  • The first is incorporating operational risk with a particular focus on cyber risk. It seems like once a month another company is in the news announcing a major breach. So, senior management is keenly aware of the need to incorporate this type of risk, and especially the associated cleanup costs, into any risk model.
  • Second, companies are starting to build out their models to be able to address the fungibility of capital. Especially with large, multi-national organisations, there is a distinct risk the organisation may have enough capital, but it may not be able to be deployed to a specific troubled subsidiary.
  • Third, with the increasing use of multi-year models, clients are starting to put more emphasis on the impact of management decisions.
    Historically, these models were built with the idea that the company's business plan will just roll out over the next three to five years. One of the key roles of senior management is to adjust the plan to counter problems that occur along the way. Without this, it can be difficult to get internal buy-in about the projected results.
    At the same time, though, modellers need to be careful when building in adjustments to ensure that they are reasonable both in terms of timeliness and magnitude.” 

Driving the change

“Most of these changes are being driven by the increased buy-in from senior managers. Being forced to file and then defend these types of projections with regulators and rating agencies gives them a huge incentive to make sure the results are reasonable.

"There seems to be a lot more feedback driving the process"

“Now that we seem to be reaching critical mass, there also seems to be a lot more feedback driving the process.

"Specifically, when regulators see a key model improvement at one company, they're becoming much more proactive in pressing other companies to extend their models as well.”

Technology’s influence 

“Technology is influencing this type of modelling in two key ways. The first is the dramatic increase in power, especially with respect to the cloud.

“In the past, some of these model extensions were simply impractical because of how complex models had become: it was not unusual for large company models to take over 24 hours to run.

“But, once these models are redesigned to take advantage of the massive computing power available through the cloud, run times can drop to a few hours. At that point, it is feasible to start adding in these additional components to improve the reasonability of the projections.

"Technology providers had to develop better ways to trace and understand their projection"

“Another area that has grown out of this increasing complexity is a heightened need to be able to trace unexpected results. In simpler models, it is easier to figure out what's causing bad outcomes: was there a huge catastrophe? Did the equity market tank?

“But, with the increasing interactions between components, that's no longer quite so simple. As a result, the technology providers had to develop better ways to trace and understand their projections at a much finer level. Without those tracing capabilities, these models can quickly become ‘black boxes’ that companies don't feel comfortable relying on.”

Future developments

“The biggest changes we're going to see is the integration of these models into product design similar to what we saw with catastrophe models.

"We're going to see is the integration of these models into product design"

“When catastrophe models were first deployed, companies used them solely to determine how much risk they had on their books.

“But, as the models gained acceptance, companies started realising they weren't just for monitoring: they could actually help companies design a better book of business.

“As a result, we've seen the rise of large percentage deductibles in both coastal and earthquake exposed areas. We've seen companies augment their regular reinsurance with capital market solutions. In short, we've seen companies use these models to help change the basic nature of property catastrophe insurance.

“And, while I don't know what the changes will look like going forward as companies leverage their investment in capital modelling, I'm excited to help companies figure it out.”