Internal/Portfolio Metrics

Monitors protocol liquidity and outstanding claim levels; and outputs metrics and recommendations to manage the protocol’s portfolio.

Introduction

Concrete’s Internal/Portfolio engine uses the current portfolio of policies and open loans, data about market conditions, historical statistics about C-Scores (produced by collateral models) and L-Scores (produced by Loan&Policy models); and outputs a series of metrics and recommendations:

a. Protocol Reporting

  • Open policies and disbursed loans metrics, with views by credit score decile, by different terms, by origination date and by collateral.

  • Interest currently charged.

  • Capital available

b. Key Metrics

  • Value at Risk.

  • Potential calls for loans.

  • Predicted repayment rates.

c. Theoretical Models Calculations

  • Cramer-Lundberg

Objectives

These outputs will be consumed in 2 main ways:

  • As a means to measure and manage our portfolio metrics; in order to design or fine tune decisioning and pricing strategies. Or more generally to produce data insights and serve as a business intelligence tool.

  • As IR-Scores, which will be supplied, in addition to C-Scores and L-Scores, to the decision and pricing services whenever the protocol needs to approve or price a policy.

Methodology

The Internal/Portfolio Metrics Service will be built by combining:

  • Classic battle tested risk management best practices - “Credit Risk Measurement” book by Saunders & Allen provides good specifications.

  • Novel approaches to risk management such as the dashboards provided by Gauntlet Network for DEFI protocols.

Our framework will be transparent and publicly documented:

  • White paper describing our methodology.

  • Metrics dashboards and reports published quarterly.

This transparency will yield trust; and allow a healthy conversation where all stakeholders and users can drive improvement towards higher levels of safety.

In addition to theoretical models such as Cramer-Lundberg; this service will rely heavily on the simulation engine detailed here.

Integration within the Product Ecosystem

See architecture diagram

Last updated