← Projects
Portfolio Theory · Asset Allocation · Risk Analytics

Portfolio Management & Risk Analytics

A portfolio-construction tool that moves from allocation choices to risk decomposition, modern portfolio theory, and scenario stress work. The emphasis is on tradeoffs: expected return versus volatility, diversification benefit versus concentration, and how different macro regimes hit the same allocation differently.

Analytical Question What kind of portfolio am I actually building once allocation, covariance, and stress outcomes are considered together?
Outputs Risk score, MPT metrics, diversification signals, concentration read-through, and scenario stress returns.
Use Case Useful for translating abstract asset-allocation theory into a concrete risk-return conversation.
Interactive Model

Construct, Diagnose, Stress-Test

Begin with a preset or custom allocation, confirm the weights, then move through three layers: basic risk composition, modern portfolio theory diagnostics, and macro stress testing. The result should tell you what kind of portfolio you own, not just what percentage sits in each bucket.

What This Models

A compact portfolio-construction framework using expected returns, volatilities, correlations, and regime shocks to approximate how allocations behave.

How To Use It

Select a preset or build your own allocation, validate that the weights sum to 100%, then compare the risk score, MPT profile, and stress outcomes before changing the mix.

How To Read The Output

The key signal is whether the same allocation that looks attractive on expected return still holds up under concentration, correlation, and macro stress.

Methodology

  • Risk scoring uses weighted asset-class risk contribution as a quick first-pass diagnostic.
  • MPT metrics estimate expected return, volatility, Sharpe, beta, drawdown proxy, and diversification ratio.
  • Stress tests apply stylized macro shocks to expose regime sensitivity that a mean-variance view can hide.

Limits

  • Expected return and covariance assumptions are simplified and static.
  • No optimization frontier or security-level position data is modeled here.
  • The goal is decision support and intuition, not a production allocator or portfolio-management system.