Portfolio Optimization for Index Tracking with Constraints on Downside Risk and Carbon Footprint

Battle Count: 81

The paper is relevant to quantitative trading in the context of index tracking, portfolio construction, and risk management. It provides a framework for constructing decarbonized portfolios that track benchmarks while minimizing downside risk. However, it focuses more on passive index construction rather than active trading strategies. The VaR/ES optimization framework and factor model decomposition are directly applicable to quantitative portfolio management. The out-of-sample performance analysis during climate events provides actionable insights for risk-aware portfolio managers.

Implementation requires: (1) multi-factor model estimation (Fama-French regressions), (2) constrained nonlinear optimization using Trust-Region Constrained Algorithm, (3) carbon data integration and ranking, (4) rolling window in-sample/out-of-sample framework, (5) handling of missing data and data alignment across multiple sources. The mathematical formulation is clear but the optimization with multiple constraints and the factor model calibration add complexity. Python with scipy.optimize or similar libraries could implement TRCA.

Data sources are identified (Yahoo Finance via yfinance, Bloomberg Terminal at IIM-B, Kenneth French Data Library, IIM-A Data Library). Python modules (yfinance, pandas-datareader) are mentioned. However, the Bloomberg Terminal data is not freely accessible, and the specific optimization code (TRCA implementation) is not provided. The methodology is well-described mathematically, but exact parameter settings and code are not shared.