Hello Aspose team,
We are using Aspose.word .net (Net 2) version 15.7 version of dll to generate word document as response object using vb.net 2010.
When we paste attached description from Source.docx following XML generated with full content. But when we generate word document with aspose.word, last 2 paragraphs are missing from item 3 and sample output attached AsposeWord-WithMissinglast2Paragraphs.docx.
Attached part of the function adding description (FunctionFillIssueDetails.docx) to word template.
This is a major concerns with our application users and we need to resolve this issue ASAP. Please help to resolve the issue.
Thank you.
Akshay Patel
IAD identified the following three related issues associated with management’s aggregation and calculation of the CCAR buffer:
· Fragmented ownership of the CCAR capital buffer calcuation and methodology
· Incomplete documentation or no formal documentation of assumptions based on management judgment
· Inadequate segregation of duties in the process to aggregate and estimate the CCAR buffer.
Each of these issues is described in more detail below.
-
Fragmented ownership of the CCAR capital buffer calculation and methodology
It is not clear who owned the CCAR capital buffer calculation and methodology from end to end in the 2014 November CCAR calculation. With the recent staff changes, it is not clear who will own the CCAR capital buffer calculation and methodology going forward. -
Incomplete documentation or no formal documentation of assumptions based on management judgment
The documentation associated with the end to end calculation of the CCAR capital buffer did not fully describe the process and / or provide the rationale behind management’s decisions. For example, the documentation did not answer the following questions:
a) For Consumer (L&R) and Commercial (UGC) component liability models, the model risk charge was calculated based on a 95 percentile parameter risk while the documentation misled readers by implying that the model risk charges were calculated using “macro shock”. Documentation should be clarified.
b) The Consumer (L&R) inherent modeling limitations capital covers only Variable and Fixed Annuities. The instruction did not provide rationale for not considering similar impact on other product lines such as Interest Sensitive Life.
c) Inherent Model Risk Aggregation: The total stand-alone component model risk charge is $9.3B and the aggregated (diversified) total is $3.5B - $4.5B depending on correlation and confidence level assumptions. Rationale on the use and selection of correlation assumptions, and confidence level assumptions was not evidenced.
d) Model Validation and Adequacy Risk Charge: The asset model risk for non-validated models was calculated directly from granular model and asset information at the position level. The liability model risk for non-validated models was derived from the above asset model risk using a formula. The calculation of the liability portion of the model adequacy charge relies heavily on management’s assumptions and judgments, which are not clearly documented in the methodology.
e) Process Limitations Risk Charge: The capital for process limitations was based on several adjustments made in Consolidated Analysis and Reporting Tool (CART) due to process and controls and these adjustments were combined judgmentally to arrive at the $1B estimate for this buffer. The management judgments were not clearly documented.
f) The selection of the overall capital buffer that covers (1) inherent modeling limitations, (2) model validation and adequacy, and (3) process limitations was based on management’s judgments that were not clearly explained or documented.
g) The Natural Catastrophe model risk charge was based on a two year projection instead of 9 quarters. Additionally, documented rationale was not maintained to support the two-year projection. Also, it is not clear if management took the position of this being immaterial to the overall results or if the model could not project for 9 quarters. Adding one additional quarter would result in $79M of additional model risk charge for Natural Catastrophe; however, after diversification and rounding, the impact is significantly reduced.
h) Although the aggregate inherent model risk charge and overall capital buffer documentation was presented to ERM senior management for review, there are no meeting minutes documenting the output of such discussions and conclusions reached.
i) Consumer (Life and Retirement, L&R) proposed a stand-alone inherent model risk charge of $689M. However, during the aggregation process of stand-alone inherent model risk charge for various business units, Corporate ERM used an earlier version of the model risk charge submitted by L&R Business, causing the total stand-alone inherent model risk charge to be overstated by $283M.
j) The independent challenge over the investments inherent model risk charge is not sufficiently documented.
Management’s estimation of the buffer is informed by numerous pieces of analysis (for example, the aggregation of individual model risk estimates to determine an overall model risk estimate). These pieces of analysis depend on assumptions that are based exclusively on management judgment / business intuition. These assumptions and management judgment are not formally documented.
3) Inadequate segregation of duties in the process to aggregate and estimate the CCAR buffer
The roles and responsibilities over the CCAR inherent model risk aggregation are not properly segregated. The Model Risk Analytics group (reporting to the head of Model Risk Management Team) aggregates the stand-alone inherent model risk charges from the operational units, runs the Aggregation model and assesses the results. While the aggregation model meets AIG’s criteria for a “model”, it has not been registered in Model Administration and Reporting System (MARS) and therefore not validated. However, under AIG’s current model risk management policy, the model would need to be validated by the Independent Validation Group (reporting to the head of Model Risk Management Team) which could lead to a potential conflict of interest.
As a result, the capital buffer results may not include all relevant components.