MSquare and the Laws of Model Monitoring

What is Model Monitoring?

Model Monitoring is a branch of governance and regulations theory defined for financial services. It is proven that unregulated and ungoverned business often leads to inefficiency.  More precisely, it studies the effects of changes in the internal and environmental systems at the macroscopic scale by analysing the collective motion of models using analytics. Though there are many important model monitoring principles that govern the behaviour of models, the most critical principles are defined in three basic laws of Model Monitoring,

The Law of Large Models

The Law of Inertia (Model Performance)

The Law of Governance

The math itself should make it abundantly clear that that the need for Model Governance across the model lifecycle is driven by both Business Risk and Regulatory Risk. Let look at a platform M2 which can manage all the laws of model monitor to promise a better regulation process.

What is M2?

At BRIDGEi2i, we believe that there are 4 essential elements that must converge to make Model Governance work:

  • Process, roles and responsibilities: A good model governance solution needs to set up clear roles and responsibilities to meet for internal stakeholder and regulatory requirements
  • Data & Metrics: Standardized data schema, pre-programmed metrics and custom dashboards are a few ways to make the data and metrics robust
  • Resource Optimization: Businesses are increasingly moving towards automated workflow systems that speed up resolution time across businesses, and
  • irrespective of size, there is a need to develop a model inventory, and drive clarity in Documentation

M2 is an enterprise server platform built with pride at BRIDGEi2i’s research labs in Bangalore that has baked in all three laws of Model Monitoring and simplifies model management across the model lifecycle from initiation to regular tracking and pro-active alerts on model performance, and enhancement audits.

Model monitor m2-diagram 1

Let’s better understand Model Monitoring laws and also how these manifest themselves within the M2 platform.

The Law of Large Models

States that all the statistical models build for a business should converge towards the unique business strategy.  Financial services use large number of models for analyzing business strategies, informing business decisions, identifying and measuring risks, valuing exposures, instruments or positions, conducting stress testing and many more. The fundamental requirement is building of single repository for all the models which are auditable and transparent. For the ideal, Model variables, Model type and Detailed documentations are characteristic described by the law.

model monitor m2-diagram2

M2 platform is one stop repository for all the models used in the business with detailed documentation. The data can be setup and new models can be added with approvals.  Models can be selected for further investigation and notifications can be sent to corresponding business and models owners.

The Law of Inertia

States that, when viewed in an inertial reference frame, a model either remains at rest or continues to perform at constant accuracy, unless acted upon by an extra data point. Model performance changes over time based on the change in environment and corresponding data. Risk models are those whose performance reduces exponentially with time. A stable model is defined as the following set , Stable _Model ={m | (m)  (when t1 ==>t2  then  p(t2) >= p(t1)) }

where p(t) is accuracy of the model at time t.

Identification of the stable and risk models in terms of accuracy should be tracked periodically to verify the law of inertia.

model monitor m2-diagram3

As per the law any model can go off in terms of performance with an additional data used inside it, M2 provides reports on historical and current model performance at overall level as well as for a specific model selected. M2 can also identify key drivers of model performance and shows alerts based on defined rules.

The Law of Governance

The strength of the governance on every model is directly proportional to health of the portfolio.

model monitor m2-diagram4

Key characteristics of governance are Building models for tracking, managing models and Withdraw models,

model monitor m2-diagram5

The law is supported through the corollary that Model risk can lead to financial loss, poor business and strategic decision-making, or damage to a banking organization’s reputation. Challenging and changing risk regulations requires financial services to implement a good governance practice across all individual models and aggregate  to identify, remediate, monitor, exploit and manage model risks.

M2 provides strong governance supports in terms of all the three laws which gives a structure to risk management functions and hence increases the stability across different customer life cycles. A case study with an indepth analysis of M2 can be downloaded at . Feel free to comment below or reach out to us via an email at

This blog is authored by Rajani Rai, Analytics Consultant at BRIDGEi2i

About BRIDGEi2i: BRIDGEi2i provides Business Analytics Solutions to enterprises globally, enabling them to achieve accelerated business impact harnessing the power of data. Our analytics services and technology solutions enable business managers to consume more meaningful information from big data, generate actionable insights from complex business problems and make data driven decisions across pan-enterprise processes to create sustainable business impact. To know more visit

Connect with us:
facebook  BRIDGEi2i on twitter  BRIDGEi2i on LinkedIn  BRIDGEi2i on Google+  BRIDGEi2i on YouTube

The views and opinions expressed in this article are those of the author and do not necessarily reflect the official position or viewpoint of BRIDGEi2i.