Touchstone financial module improvements

Improvements and fixes which enhance Verisk's business logic modeling are summarized below.

Improvement to step functions

• In the application of limit-at-damage for Building & Contents for use cases, where the start step is less than the deductible amount, the financial module now accurately accounts for market policy language and claims settlement among Japanese insurers for policies with trigger type 1 and type 4.
• For step functions with trigger type 4, we enhanced the accuracy of passing the first step position to every subsequent step and assigning correct probabilities when combining building and content loss; this reflects a more accurate policy gross loss. This applies specifically to cases where content payout is much smaller than ground-up loss. Policy gross loss is expected to decrease for these cases.
• We enhanced the allocation and use of the secondary uncertainty distribution to limit payout for limit-at-damage. For policies with trigger types 1 and 4, we now use the content distribution to limit content payout. For policies with trigger type 2, we now use the building distribution to limit building payout.

Updates to reporting of peril codes

Enhancements to the database table, the user interface, exposure peril coding, and CSV export make peril code reporting for all loss perspectives more consistent and streamlined.

• Event-level results

  We updated peril codes to reflect the requested and effective peril of the analysis instead of the complete and combined peril group. This will not change loss numbers. However, for downstream analytics it will lead to realignment of losses in grouping, marginal impact and other such metrics and operations.

• Summary results (Location Summary, Contract Summary, Annual EP, Annual EP Summary)

  Significant enhancements have improved consistency in peril codes mapping to losses and reporting of results.

  • Losses are reported with consistent peril codes when running multiple models at the same time. They are grouped into main peril buckets (EQ, IF, ST, TC, WF, WS) when saved by peril, regardless of any contributing sub-perils.
  • From version 11.0 to version 12.2: When saved by sub-peril, in database/SQL results and export results, the financial module splits out sub-peril losses, in addition to the main peril bucket for perils and models that report more than one peril or sub-peril.
  • From version 13.0 for Annual EP Summary: Results by sub-peril are also reflected in the Touchstone users view the Summary EP table in the Results Summary pane.

Improved handling of sub-limits and area tags

Area tags link locations to sub-limits. With the introduction of sub-limits for perils, we reworked the management of these pairings to identify losses for each single location term, which results in more accurate gross losses. We introduced two fixes to improve accuracy for complex policies with nested sub-limits and min-max deductibles in hot fix 12.2; these fixes are included with the current release. All of these updates facilitate the full transition to column-wise coding in client exposures.

Enhanced fidelity of loss numbers in granular results’ sets

The fidelity of granular results sets by geography and by exposure attributes is enhanced by consistent matching of top portfolio level numbers. This enhances the credibility of our view-of-risk with sophisticated clients, who perform internal and regulatory model validation.

Improved algorithmic and probabilistic stability

We enhanced the stability and accuracy of algorithmic and probabilistic operations for multiple cases: CAT XOL with line-of-business filters, ML (Max of Deductible Amount or Percent Loss) deductible, annual aggregate multi-peril terms, campus policy structures with disaggregation. These enhancements facilitate the adoption of our new view-of-risk, the advancements which we have made in business logic modeling and the overall adoption of the product.

Improvements to granular back-allocation

Back-allocation allocates losses for any perspective from a portfolio event total or from a portfolio component (such as geography or line-of-business) down to the granular level of location and coverage.

Release13.0 improves back-allocation for a variety of workflows for complex policy structures. Accuracy and coherence enhancements allow for losses from different granularities to accumulate correctly to event and perspective totals and to be consistent between different SQL tables and views of loss.

• For exposures and analysis with policy layers and per-risk reinsurance treaties, we enhanced the consistency of loss summaries and removed discrepancies and imprecisions for the relationships among the following tables:
  • By Treaty vs. TreatyExposureAttribute vs. TreatyExposureAttributeGeo
  • By Layer vs. ContractGeoLayer vs. ContractZoneLayer

• We enhanced exposure and analysis with CAT AGG and XOL treaties discrepancies and imprecisions in pre and post-Cat Net loss back-allocation are removed in the relationships among the following loss tables.

  • By Treaty vs. TreatyExposureAttribute vs. TreatyExposureAttributeGeo
  • By Event vs. LocationSummary vs. ContractSummary
  • By Zone vs. ContractZone vs. ContractZoneLayer
  • By ExposureAttribute vs. ExposureAttributeGeo
• The LOB filter for CAT XOL now works with Location UDF, Contract UDF, and Contract & LOB save results. These results sets can propagate granular information to the CAT XOL filter.

  The equivalency for results in terms of event sums:

  Event total = Ʃ Location UDF = Ʃ Contract UDF = Ʃ Contract = Ʃ LOB

• We enhanced the accuracy of back-allocation of all loss perspectives by coverage for location and sub-limit, layered structure as follows:
  • The financial module provides more consistent synchronization of loss by coverage to the total sum of loss of all coverages.'
  • The financial module performs synchronization and checks accuracy by sub-peril loss and by combined peril loss.

Enhanced accuracy for annual aggregate terms and ML deductible

The following updates improve accuracy:

• The application of annual aggregate terms on location facultative contracts for policies under min-max deductible on layer. As a result, applicable terms for second and third events are now accurately estimated and the fidelity of gross loss is improved.
• The application of annual aggregate terms for treaties has been updated for treaties covering a subset of perils. We now carry over applicable annual aggregate terms between perils whenever the treaty covers a subset of perils. As a result, the fidelity in cross peril erosion of aggregate treaty terms is improved.
• We updated the algorithm for applying ML (Max of Deductible Amount or Percent Loss) type deductibles so that it now produces consistent results with other percentage of loss type deductibles. The percentage value is now defined as in the PL (Percent of Loss) type deductible so that losses can be compared between the two types. As a result, gross loss results are more accurate, regardless of which term types are used.

More consistent layer max losses for varying peril set codes

We improved the consistency of loss results between single peril loss and combined peril loss for layers and sub-limits. Now, when the financial module identifies a single peril loss on any tier of the layer processing, it automatically copies the loss into the combined peril of that risk and then applies and combines the terms. This results in more accurate gross losses for sub-limits that cover a subset of the overarching layer perils.

Correction for gross loss calculation with layer participation

Contracts with limit B, C and E with 0 or 0% participation at the layer level are now producing accurate gross loss. The Layer Occ Participation value correctly defaults to 0 (zero).