Name: ESRR Model- Limitations of existing Excel Model and Challenges in application of Statistical Modeling
Uploaded: 2024-11-04
Duration: 6 min 26 s
Description: ESRR Model- Limitations of existing Excel Model and Challenges in application of Statistical Modeling.

ESRR Model- Limitations of existing Excel Model and Challenges in application of Statistical Modeling. 

Scene 1

ESRR Model- 
Limitations of existing Excel Model and Challenges in application of Statistical Modeling

By- Disha Joshi Date – November 4th, 2023

AGENDA. Introduction Overview of Current Excel Model. Limitations of Current Excel Model-oversimplification. Limitations – Data Dependency & it’s impacts Statistical Modelling Challenges Case starting from scratch Leveraging existing model Expected outcomes & Benefits. 

Scene 2

Introduction  Overview of Current Excel Model.	 Limitations of Current Excel Model-oversimplification. Limitations – Data Dependency & it’s impacts Statistical Modelling Challenges Case starting from scratch  Leveraging existing model Expected outcomes & Benefits

Introduction. Our current Excel-based risk rating model, inspired by ESRS standards, faces significant limitations that hinder its effectiveness and reliability. This presentation will examine the key shortcomings of the existing model, explain why applying statistical modeling to it is not feasible, and propose an action plan to develop a more robust and accurate risk assessment framework. By understanding these challenges and embracing a data-driven approach, we can create a model that truly captures the complexity of environmental, social, and governance risks in our operations.. 

Scene 3

Our current Excel-based risk rating model, inspired by ESRS standards, faces significant limitations that hinder its effectiveness and reliability. This presentation will examine the key shortcomings of the existing model, explain why applying statistical modeling to it is not feasible, and propose an action plan to develop a more robust and accurate risk assessment framework. By understanding these challenges and embracing a data-driven approach, we can create a model that truly captures the complexity of environmental, social, and governance risks in our operations.

Overview of Current Excel Model. Residual Risk score The final residual risk rating is determined by subtracting the sum of averaged control factors from the overall inherent risk score, providing a simplified view of the organization's risk exposure.. 

Scene 4

Residual Risk score The final residual risk rating is determined by subtracting the sum of averaged control factors from the overall inherent risk score, providing a simplified view of the organization's risk exposure.

Control Score Assessment  Control factors are applied to certain risks, with scores based on a control ratio multiplied by the average of the affected factors. Controls are categorized as policies, targets, actions, or metrics based on which control scoring range is defined.

Inherent Risk Calculation The model calculates overall inherent risk by summing natural, climate, and human rights risks. The inherent risks represents the susceptibility to loss or harm stemming from the factors before any mitigation measures introduced. Each factor is an impact, or a dependency based on the sector on the industry, weighted based on a Biodiversity Risk Rating (BRR) derived from WWF data and revenue percentages by geography.

The current excel model functions as a rudimentary calculator for risk ratings, operating primarily on linear equations without complex algorithms. The fundamental equation of the model is , Overall Inherent Risk Score – Overall Control Score = Residual Risk Score

Scene 5

Limitations: Oversimplification of Complex Relationships. 

Scene 6

Limitations: Oversimplification of Complex Relationships

Linear Calculations The current model relies on simple linear calculations, which fail to capture the intricate relationships between various risk factors. Environmental, social, and governance risks often have complex, non-linear interactions that can amplify or mitigate each other in unexpected ways.

Averaged Weights  The use of averaged weights for risk factors oversimplifies their relative importance. This approach assumes that all factors within a category have equal significance, which is rarely the case in real-world scenarios. Some factors may have disproportionate impacts that are not reflected in the current model.

Lack of Interdependencies The model does not account for interdependencies between different risk categories or control factors. In practice, changes in one area can have cascading effects on others, creating complex feedback loops that are not captured by the current linear approach.

Limitations: Data Quality and Availability Issues. 

Scene 7

Limitations: Data Quality and Availability Issues

Missing VLOOKUP Tables  The model includes cell references to VLOOKUP tables, but these tables are not actually present. This creates a significant gap in the model's functionality and raises questions about the accuracy of any calculations that rely on these missing data sources.

Lack of Documentation There is no comprehensive documentation explaining the model's methodology, assumptions, or data sources. This absence of context makes it difficult to understand the rationale behind certain calculations or to verify the model's accuracy.

Manual Input Dependency Most of the inputs in the current model require manual entry, increasing the risk of human error and inconsistency. This reliance on manual data entry also limits the model's scalability and makes it challenging to update regularly with new information.

Absence of Qualitative Data The model lacks integration of qualitative data that could provide crucial context and nuance to the risk assessment process. Without this information, the model may miss important factors that are not easily quantifiable but still have significant impacts on overall risk.

Scene 8

Challenges of transitioning from Basic to Advanced Methodologies. 

Scene 9

Challenges of transitioning from Basic to Advanced Methodologies

Why Statistical Modelling Cannot be Applied to Existing Model

The current model's reliance on manual inputs and limited data sources means there are not enough data points to perform meaningful statistical analysis. Statistical modeling requires large datasets to identify patterns and relationships accurately Without a comprehensive historical dataset, it's impossible to validate statistical models or test their predictive power. This absence of time-series data hampers our ability to understand trends and develop reliable forecasts. The current model structure may introduce multicollinearity, especially for certain regions where risk factors are closely related. This can lead to unreliable estimates and difficulty in isolating the effects of individual variables. Many statistical modeling techniques assume linear relationships between variables. However, the complex nature of ESG risks often involves non-linear interactions that are not captured by the current model structure

Developing an effective risk model necessitates clear guidelines and principles that shape its construction. Key considerations encompass adaptability and user engagements, which are vital for model acceptance. A careful assessment of statistical methods, coupled with defined data inputs, will enhance the model’s robustness.. 

Scene 10

Developing an effective risk model necessitates clear guidelines and principles that shape its construction. Key considerations encompass adaptability and user engagements, which are vital for model acceptance. A careful assessment of statistical methods, coupled with defined data inputs, will enhance the model’s robustness.

The limitations of the current model are so significant that attempting to build upon it would be more challenging than starting anew. A fresh start allows us to incorporate modern risk assessment methodologies and data science techniques from the ground up.

By starting from scratch, we can design a data collection and management system that supports advanced analytics. This includes identifying key data sources, establishing automated data pipelines, and implementing robust data quality controls.

A new model can be built with scalability and adaptability in mind, allowing for easy incorporation of new risk factors, changing regulations, and emerging best practices in ESG risk assessment.

Starting anew enables us to leverage cutting-edge statistical and machine learning techniques that can capture complex, non-linear relationships between risk factors and provide more accurate risk predictions.

Leveraging Existing Model for Scope Definition. Identify Key Risk Areas. 

Scene 11

Leveraging Existing Model for Scope Definition

Use the current model's structure to map out the primary risk categories (natural, climate, human rights) and their components. This provides a starting point for a comprehensive risk taxonomy in the new model.

Catalogue the existing control factors (policies, targets, actions, metrics) to ensure all relevant risk mitigation strategies are considered in the new model's development.

Review the current model's inputs to identify critical data sources, including external references like the WWF biodiversity risk filter, which can inform data collection efforts for the new model.

Analyze the existing model's users and contributors to create a comprehensive stakeholder map, ensuring all relevant perspectives are included in the new model's design process.

Proposed Action Plan. Conduct a thorough analysis of current risk assessment needs, stakeholder views, and industry best practices. Develop a detailed project plan and logic and assemble a cross-functional literature relayed to the ESRS Standards. Phase 1: Discovery and Planning (Months 2-3). 

Scene 12

Conduct a thorough analysis of current risk assessment needs, stakeholder views, and industry best practices. Develop a detailed project plan and logic and assemble a cross-functional literature relayed to the ESRS Standards.  Phase 1: Discovery and Planning (Months 2-3)

Phase 2: Data Strategy and Architecture (Months 3-4)

Design a comprehensive data architecture, identifying key data sources and establishing data governance protocols. Develop data collection methodologies and build automated data pipelines to ensure consistent, high-quality inputs for the new model.

Phase 3: Model Development and Testing (Months 5-8)

Create a prototype of the new risk assessment model using advanced statistical and machine learning techniques. Conduct rigorous testing and validation, including sensitivity analyses and back-testing against historical data where available.

Phase 4: Implementation and Training (Months 8-9)

Roll out the new model in phases, starting with key business units. Provide comprehensive training to users and stakeholders. Develop detailed documentation and establish a feedback loop for continuous improvement.

Expected Outcomes & Benefits. Improved Accuracy Enhanced risk predictions through advanced modeling techniques and comprehensive data integration.. 

Scene 13

Improved Accuracy Enhanced risk predictions through advanced modeling techniques and comprehensive data integration.

Regulatory Compliance Improved alignment with evolving ESG reporting standards and regulatory requirements.

Increased Efficiency Automated data collection and processing, reducing manual effort and potential for errors.

Scalability Flexible framework capable of adapting to new risk factors and changing business landscapes.

Better Decision Making More granular and timely risk insights to support strategic planning and risk mitigation efforts.

Competitive Advantage Enhanced ability to identify and address risks proactively, potentially leading to better market positioning.

ESRR Model- Limitations of existing Excel Model and Challenges in application of Statistical Modeling

Scene 1 (0s)

Scene 2 (9s)

Scene 3 (22s)

Scene 4 (46s)

Scene 5 (1m 33s)

Scene 6 (1m 39s)

Scene 7 (2m 16s)

Scene 8 (3m 1s)

Scene 9 (3m 7s)

Scene 10 (3m 47s)

Scene 11 (4m 36s)

Scene 12 (5m 13s)

Scene 13 (5m 56s)