Deadline extended for Request for Proposals: Creating a Simulation Engine That Simulates Reserve Development for Individual Claims
1. Casualty Actuarial Society (CAS)
The CAS was organized in 1914 as a professional society for the promotion of actuarial and statistical science as applied to insurance other than life insurance, such as automobile, liability other than automobile, workers compensation, fire, homeowners, commercial multiple peril, and others. Such promotion is accomplished by communication with those affected by insurance, presentation and discussion of papers, attendance at seminars and workshops, collection of a library, research and other means. The membership of the CAS includes over 7000 actuaries employed by insurance companies, industry advisory organizations, national brokers, accounting firms, educational institutions, state insurance departments, the federal government and independent consultants.
2. Committee on Committee on Dynamic Risk Modeling
The Committee on Dynamic Risk Modeling (the DRM Committee) is charged with facilitating research and providing direction, guidance, and support to the profession, regulators, and others regarding dynamic modeling of property & casualty risks. The DRM Committee will sponsor the research to be conducted under this Request for Proposal (‘RFP’) on behalf of the CAS.
3. CAS Interest in the Subject
Insurance entities are required to record a reserve due to their liability for the cost of settling all claims that have been incurred regardless of whether they have been reported or not. This “unpaid loss reserve” is usually represented as three parts: (1) total 'case reserves' for all open claims, (2) expected development on open claims, and (3) total cost of incurred but not reported (IBNR) claims.
One of the most important functions that property & casualty actuaries perform is to quantify the reserves that an insurance entity needs. An actuarial reserve analysis involves the assessment of the stochastic loss process, and an estimate of the amount of reserves to cover the unknown future claim loss payments.
Actuaries employ multiple methods to analyze loss reserves. A common type are triangulation methods, which consists of aggregating historical claim loss payments and arranging them by accident year and by development year. Triangulation methods have been used to obtain a point estimate of the loss reserve amount, as well as to measure the uncertainty around that point estimate. However, by compacting loss data into a triangle, a large amount of information is lost, and thus some details of the underlying loss emergence process cannot be considered.
In recent years, actuaries have put forward methods that do not compact data into a loss triangle and that, instead, use data at an individual-claim level to develop models that describe the probability distribution of key features of the underlying behavior of those claims.
4. Research Problem Description
Background: The Loss Simulation Model (LSM) is an existing software tool that is designed to carry out stochastic simulation modeling at an individual-claim level. It is a tool that is publicly available, owned by the CAS, and maintained by the DRM Committee. In its current form, the LSM is able to model individual claims stemming from new accident years only; it is not designed to take an existing inventory of individual claims and simulate the development of those claims, and hence it is not suitable for reserve analysis at an individual-claim level.
The DRM wishes to develop a tool that can be used as a simulation engine for stochastic reserve development at an individual-claim level that would be used to aggregate the probabilistic models of individual claims. We envision this to be a collection of modules, separate from a Graphical User Interface (GUI), which users can call from other platforms, as outlined below.
5. Project Requirements
The DRM Committee requests proposals from qualified researchers to develop a simulation engine that can be used to estimate both expected total loss reserves as well as its potential volatility. The DRM Committee has identified paper, “Triangle-Free Development: a non-traditional framework for estimating reserves and reserve uncertainty” by Pietro Parodi1, which should be the basis for the methodological framework and the input parameters that will guide how the simulation engine should carry out the simulation modeling. The Committee will be available to provide further technical recommendations to the tool development.
The remainder of this section sets out requirements regarding the deliverables of the simulation engine that concerns this RFP.
- It should be possible to provide inputs through the Application Programmer Interface (‘API’, see below) or as CSV files.
- Inputs are identified in the paper referenced above and include:
- Existing inventory of individual claims, open and closed.
- Past earned exposures
- Parametric assumptions regarding lag distributions for reporting, payments, and case reserve adjustments.
- It should allow to input multiple claim classes, each with its own set of parametric assumptions.
- The proposal should include a means for the storage and maintenance of multiple sets of inputs, as would be needed to run multiple claim classes.
It should be noted that the setting of the parameters that would be an input to this simulation engine is outside of the scope of this RPF. Although such parameterization is critical to the tool’s functioning, the DRM Committee plans to attend to it separately.
B) Execution and Performance
- It should be possible to execute the simulation from the command line to generate output files, based on input files.
- It should be possible to execute the simulation engine through the API.
The committee has set a performance benchmark of being able to generate 250,000,000 simulations in 10 minutes. The committee will give preference to proposals that can demonstrate they can attain this benchmark.
C) Outputs and Reports
A reporting functionality must be included as part of the simulation engine, to allow the user to confirm that the results are as expected. The committee will give preference to proposals that include a solution for a fast high-volume output data store.
The tool should be able to:
- Execute the reporting functionality from the command line.
- Store the simulation in output files to facilitate retrieval and reporting.
- Calculate and output the closed-form expected value of each claim.
The committee will work with the vendor to design specific reports, but for the purpose of this RFP, we expect a proposal for a reporting engine with the ability of aggregating across claim classes, accident years and development ages in order to calculate key simulation statistics.
The simulation engine should be developed with an Application Programmer Interface (‘API’), in order for other applications or user interfaces (UI) to be able to call the engine’s functions or procedures.
It is expected that the engine will be connected to the LSM through its .NET API.
- The simulation engine should be developed using an open source language(s).
- The committee will give preference to proposals that use R's statistical libraries for the distribution functions required for the actual simulation and perhaps calculations of statistics.
- The technology chosen should be common and readily understood by a large group of developers.
The committee will prefer an architecture that does not depend on open source components that are not expected to survive the immediate future.
- The vendor shall provide documentation explaining how to use the simulation engine.
- The vendor shall provide documentation on how to install and build the environment necessary to run this module.
The development of this project will be performed under contract and funded by the CAS. It is expected that the researcher will work with representatives of the Committee on the specific design of the project.
6. Proposal Requirements
Proposals should include:
- Clear outline of the work that will be performed. The outline should include a validation methodology that provides comfort to the Committee that the simulation engine functions according to its specification.
- Timeframe in which it will be performed (including key dates). The more specific, the better.
This proposal should be reviewed in conjunction with the attached Research Agreement which defines the terms and conditions under which the work is to be performed.
The proposal should be accompanied by the resumes of the researcher(s), indicating how their background, education, and experience bear on their qualifications to undertake the research.
Respondents should demonstrate their interest in and familiarity with the application of software development and/or dynamic risk modeling by including a resume (if a firm, of the principal consultant(s) performing or directing the work) showing relevant work/research experience and professional accomplishments (e.g., papers published).
The contract will be awarded by the CAS to the respondent who, in the judgment of the DRM Committee and entirely on the basis of his or her written proposal, is best able to perform the work as specified herein. If the DRM Committee determines that no proposal meets the requirements of the RFP, then no contract will be awarded.
Receipt of proposals will be acknowledged by providing each respondent with a list of all respondents. Respondents not awarded the contract will be so informed shortly thereafter.
Interested researchers should submit their proposals and any questions in writing to:
Casualty Actuarial Society
4350 N. Fairfax Dr. Suite 250
Arlington, VA 22203
Phone: (703) 276-3100
Fax: (703) 276-3108
The proposals will be reviewed by members of a subcommittee of DRM Committee. The current membership of the DRM Committee includes:
Fernando Alvarado, Chairperson
Derek P. Cedar
Sara J. Hemmingson
Daniel W. Lupton
Joseph O. Marker
7. Schedule (details negotiable)
20 Jun 2017
RFP circulates to CAS sources including ARIA listserve for academics.
23 Jul 2017
Deadline for questions (must be written) from researchers regarding the RFP.
02 Aug 2017
All written questions together with their answers will be distributed to all proposers.
16 Aug 2017
Proposal deadline - end of business day
26 Aug 2017
Proposal selection by DRM Committee
26 Sept 2017
Progress report due from research team
04 Dec 2017
First progress report due
04 Feb 2018
Second progress report due
04 Mar 2018
Final tool delivered and tested, and approved by DRMC
A total of $35,000 is available to be awarded to one or more researchers. Payment of award(s) will be contingent upon delivery of a simulation engine that demonstrates adequate functional testing and is fit for the stated purpose.
9. Presentation, Ownership and Publication of Report
If asked, the researcher(s) agree to be available to present the report at a CAS meeting or seminar. If travel is required, reasonable expenses will be paid in addition to the compensation provided in Section 8.
As a condition of selection, the CAS requires that all right, title, and interest, including copyright and patent, in and to the report be owned by the CAS. The selected researcher must sign a formal Agreement (attached) that assigns all such rights to the CAS. Of course, in any publication of the report, the researcher will receive appropriate credit. The CAS may publish the report in any CAS publication, including electronic versions such as on its Web site or on compact discs.
1 Paper available from the Institute and Faculty of Actuaries website: