Single Period Mean-Variance Optimization (MVO, Markowitz) with scope 1+2 carbon intensity constraints
Python script for running a single period mean variance optimization (Markowitz, 1952) with a weighted portfolio scope 1+2 carbon intensity target on top of the "usual" long-only constraints and having the portfolio weights sum up to 1. Carbon intensity, or carbon emissions per dollar of revenue, adjusts for company size and is generally accepted to be a more accurate measurement of the efficiency of output rather than a portfolio's absolute carbon footprint.
NB - Markowitz's Modern Portfolio Theory assumes (amongst others):
- frictionless markets
- market liquidity is infinite
- investors are risk averse
- returns are normally distributed
Also, kindly note that:
- this script is highly time period sensitive - whilst it also considers the mean daily return to be a good estimator of future returns
- the Quandl module isn't actively supported anymore for returns post April 11, 2018
This script requires the following packages / modules in order to function properly:
