Implement Risk Metrics: Value At Risk

[AlexCatarino]

Expected Behavior

Portfolio Statistics include Value at Risk (VaR)

Actual Behavior

Not supported

Checklist

• [x] I have completely filled out this template
• [x] I have confirmed that this issue exists on the current master branch
• [x] I have confirmed that this is not a duplicate issue by searching issues
• [x] I have provided detailed steps to reproduce the issue

[wtindall1]

Hi @AlexCatarino - please could I pick up this issue?

[AlexCatarino]

Hi @wtindall1 , thank you! I suggest adding your thoughts about this problem before working on it. VaR is a statistical metric, an "indicator" (see guidance on GitHub issues for Indicators), so we must create a unit test that compares the calculation with an external source.

[wtindall1]

When you say VaR should be an "indicator", do you mean that the scope of the issue is an implementation of IIndicator? Rather than a property of PortfolioStatistics (as I had assumed from the Expected Behaviour)

[AlexCatarino]

If VaR can be an IIndicator and a property of PortfolioStatistics, the better, right? :-) For example, see Sharpe and Sortino ratios. If not, yes, it should be property of PortfolioStatistics. The point I was trying to make is that it should be validated like an indicator (external data file generated with third-party software, you need to tell us how it was created).

[wtindall1]

Makes sense - thank you. I will have a look into this and comment back here

[wtindall1]

@AlexCatarino - proposed script for generating the test data, based on: https://blog.quantinsti.com/calculating-value-at-risk-in-excel-python/ spy_var.csv

#https://blog.quantinsti.com/calculating-value-at-risk-in-excel-python/

import pandas as pd
from scipy.stats import norm

spy = pd.read_csv("https://github.com/QuantConnect/Lean/raw/master/Data/equity/usa/daily/spy.zip",
                       index_col=0, names=["open", "high", "low", "close", "volume"])

spy_close = spy[["close"]]
spy["returns"] = spy_close.close.pct_change()

window_length = 252
min_periods = 2
spy["mean"] = spy["returns"].rolling(window_length, min_periods).mean()
spy["stdev"] = spy["returns"].rolling(window_length, min_periods).std()

spy["VaR_99"] = norm.ppf(0.01, spy["mean"], spy["stdev"]) 
spy["VaR_95"] = norm.ppf(0.05, spy["mean"], spy["stdev"]) 
spy["VaR_90"] = norm.ppf(0.1, spy["mean"], spy["stdev"]) 

columns_to_export = ["open", "high", "low", "close", "volume", "returns", "VaR_99", "VaR_95", "VaR_90"]

spy[columns_to_export].to_csv("spy_valueatrisk.csv", index_label="date")