logo image

Insurance Solutions

Formerly Moody’s RMS

At the recent Moody’s RMS Exceedance® conference in New York, I had the pleasure of launching unified modeling for the Moody’s RMS Intelligent Risk Platform™ (IRP). The architecture supporting unified modeling includes the new Open Modeling Engine (OME), offering the prospect of over 700 risk models being available on the IRP.

If you are a Moody’s RMS client using the Risk Modeler™ application on the IRP, you already have access to over 400 models. So what could you do with more than 700 available models that you cannot already do with 400+ models?

Here are just a few things you could do using the new unified modeling capabilities:

  • You could combine and blend Moody’s RMS, custom, and third-party risk models more seamlessly, to create your own unique view of risk. Using more risk models also means you can more easily carry out sensitivity tests for loss frequency and severity.
  • Take advantage of Moody’s RMS geocoding, hazard lookup, and exposure data to enrich your exposure input, not just with Moody’s RMS models, but also with custom and third-party models.
  • Utilize Moody’s RMS financial modeling, from loss grouping operations to post-analyses treaty editing (PATE) to help ensure a more uniform application of financial terms.
  • By using Risk Modeler for all these models, you could also optimize risk transfer using the support for rich data formats within Risk Modeler, and circulate model results from any of the 700+ models in the Results Data Module (RDM), Year Loss Table (YLT), and Period Loss Table (PLT) formatted outputs.
  • You could also greatly simplify exposure data conversion as the system would allow any of the 700+ models to be executed using Catastrophe Exposure Data Exchange (CEDE™), Open Exposure Data (OED), or Exposure Data Module (EDM) formatted exposures without dealing with any data conversion complexities.

In this post, I’d like to go ‘under the hood’ and show you how the Open Modeling Engine works in the platform. I should note that the OME is still under development, and the capabilities I am describing here are being previewed by several of our customers and partners, and their feedback may result in changes to the final version of the product.

As part of our unified modeling architecture announcement, in addition to the Open Modeling Engine, I also outlined the new Native Modeling Engine. I’ll go deep into the Native Modeling Engine in a future blog post.

Under the Hood with the Open Modeling Engine

The Open Modeling Engine (OME) is a new model execution engine similar to the Aggregate Level Model (ALM), Detailed Level Model (DLM), and high-definition (HD) modeling engines already used in Moody’s RMS models.

However, it differs from the existing engines as the OME is being built to connect to external modeling services. The first application of the OME is to connect the Moody's RMS Intelligent Risk Platform with the Nasdaq Risk Modeling for Catastrophes service (NRMC).

Moody's RMS and Nasdaq
Figure 1: Overview of the Open Modeling Engine Architecture

Let’s look at how the OME operates Oasis Loss Modelling Framework (LMF)-based models through its connection to the Nasdaq Risk Modeling for Catastrophes service. To operate the NRMC environment, the first step will be to enter your Nasdaq service credentials into the Intelligent Risk Platform.

Then, you will be able to operate the Nasdaq environment from the Risk Modeler application on the platform, to execute models hosted on the Nasdaq service, side-by-side with Moody’s RMS models.

As shown in Figure 1 above, by selecting various Moody’s RMS or third-party models to execute using the Risk Modeler user interface (UI) or API, the OME will manage the necessary conversions to deliver the required input for the Nasdaq engine – an Oasis Framework OED, and then the OME will collect the Nasdaq engine’s modeled loss output – an Open Results Data (ORD) file, into the Risk Data Lake.

Using this approach, the model execution is performed by the Nasdaq Risk Modeling for Catastrophes service. The benefit for our clients is that the Open Modeling Engine then performs comprehensive data conversions to accommodate the execution of the risk models to and from the Nasdaq service.

Combined with the rich exposure import capabilities within Risk Modeler, clients do not have to deal with multiple versions of EDM, OED, or CEDE formatted data files, or learn multiple APIs and user interfaces to provide access to all these models. Clients can simply import any of these popular exposure formats and execute any destination model.

Let’s go deeper with a step-by-step overview of how a cat modeler would work with both Moody's RMS and third-party models. If you were at our Moody’s RMS Exceedance 2023 conference recently, you would have seen a preview demonstration of how a catastrophe modeler using Risk Modeler will be able to perform import, model execution, and export operations:

Import:

1. The client would first import the EDM, MRI, OED, and CEDE-formatted exposures into the Risk Modeler application.

2. The exposure files are then converted to the Risk Data Open Standard (RDOS), the unified data format used by the IRP.

3. The files are then saved into our platform’s storage.

Note: The RDOS is the default, systemwide data format for the Intelligent Risk Platform. This ensures that all applications, microservices, and model engines utilizing exposure, policy, account, treaty, hazard, and other risk data are all built to operate on this data format, regardless of the format they arrived in.

Figure 2: Importing Exposures
Figure 2: Importing Exposures

Model Execution:

As an example, the client then uses the imported exposures to trigger a run of both Moody’s RMS U.S. Flood HD model and the third-party U.S. flood model.

1. The Moody’s RMS model runs on Moody’s RMS HD modeling engine operating natively on the IRP, without requiring any data conversion. However, the third-party model goes over to the Open Modeling Engine (OME).

2. For the third-party model, the OME converts the RDOS-formatted exposures into the OED format required by the third-party model and then sends job execution instructions to the Nasdaq Risk Modeling for Catastrophes service.

3. The OME then receives the ORD-based results outputted from the Nasdaq Risk Modeling service.

4. Finally, the OME converts and saves the losses into the IRP’s storage.

Figure 3: Model Execution with Open Modeling Engine
Figure 3: Model Execution with Open Modeling Engine

The output from both the HD and Open Modeling Engines is also in our RDOS format. The loss format in RDOS is also a superset that ensures we can manage one comprehensive data format regardless of which engine may have produced the losses.

Export and RDM Output:

The client then triggers a grouping operation to compile a single output from the multiple views of risk, to generate an RDM.

1. Read losses from the Moody’s RMS and the third-party flood models and group them.

2. Export the combined losses into the RDM formatted output.

With recent updates to RDM, the RDM formatted losses now can carry both ELT- and YLT-based outputs. This means clients can choose from multiple options in the RDM output that can help simplify communications with brokers and reinsurers who may be using different styles of loss formats.

Figure 3: Grouping and Exporting an RDM
Figure 4: Grouping and Exporting an RDM

It is important to note that data transfer between the Intelligent Risk Platform and the Nasdaq service is secured using encrypted communications over REST APIs.  

Another consideration is that data transfer between services can be expensive, especially for large results. To ensure that the communication can be well optimized, when configuring the Nasdaq service connection we will require customers to ensure both the IRP instance and Nasdaq Service are set up within the same Amazon Web Services (AWS) region.

Open Exposure Data (OED)

Even though the Nasdaq Risk Modeling for Catastrophes service is performing the model execution, to get the most accurate output possible from third-party models hosted on the Nasdaq service, the Open Modeling Engine is needed to produce the best possible OED.

Therefore, we are working with model vendors to ensure that OEDs are optimized, allowing for a better understanding of the nuances required for the input.

This is a complex area given the multiple versions of the OED we may need to produce to support various versions of the Oasis Platform and Nasdaq Risk Modeling Service currently in use, to ensure that we can still maintain fidelity with the original imported data that may have arrived in CEDE, OED, EDM or MRI formats, each with their own version support.

One important aspect to mention here is that, as the capabilities in Risk Modeler mature, I expect Moody's RMS UnderwriteIQ™ and TreatyIQ applications will also utilize the Open Modeling Engine to access these models.

This means that in the future, clients may be able to execute both account and treaty program modeling with these models within dedicated experiences from other Moody’s RMS IQ applications.

If you attended the Exceedance 2023 conference, you would have heard from a few of our clients and partners who are trying out the capabilities through the Risk Modeler application, and we are adding more customers and partners to that list.

If you are interested in previewing the capabilities yourself, or for additional information, please reach out to your Moody’s RMS representative or simply email us at info@rms.com

Share:
You May Also Like
link
Team meeting
May 02, 2023
How are Reinsurers Utilizing the Intelligent Risk Platform? …
Read More
link
Team meeting
March 13, 2023
How are Insurers Utilizing the Intelligent Risk Platform? …
Read More
Related Products
link
risk-intelligence
Intelligent Risk Platform

The Moody's RMS Intelligent Risk Platform™ is an…

Learn More
Cihan Biyikoglu
Cihan Biyikoglu
Managing Director - Head of Product for Moody's RMS

Cihan is the Managing Director - Head of Moody's RMS Product, responsible for product management across the full suite of Moody's RMS models and risk management tools. He has extensive experience in leading product management for innovative machine learning and big data analytics solutions at Fortune 500 companies over the last 20 years.

As a former Vice President of Product at Databricks and Redis Labs, Cihan developed the product strategy and road map for open-source technologies such as Apache Spark and Redis and respective enterprise offerings in the public and private cloud platforms.

Cihan also worked on products at Microsoft, Couchbase, and Twitter, where he focused on on-premises and cloud offerings in the data and analytics space. At Microsoft, Cihan focused on the incubation of the Azure Cloud Platform in its early days and the SQL Server product line, both of which have grown into multi-billion-dollar businesses for Microsoft.

Cihan holds several patents in the data management and analytics space, and he has a master’s degree in database systems and a bachelor’s degree in computer engineering.

cta image

Need Help Managing Your Portfolio?

close button
Overlay Image
Video Title

Thank You

You’ll be contacted by an Moody's RMS specialist shortly.