VaR and cVar Calculations
Budget: $30 – $250 USD
1. User will define a set of financial assets codes and the calculation period.
2. Download the time series of closing prices from Yahoo and Google.
3. Check for consistency and appoint discrepancies and outliers.
4. Calculates VaR and cVaR (expected shortfall) according to past prices, using the following methods for the selected set of financial assets (1 to N):
a. Parametric;
b. Historical;
c. Filtered Historical Simulation http://filteredhistoricalsimulation.com/;
d. Monte Carlo simulation, assuming Gaussian, Generalised Pareto, Levy, Truncated Levy Flights and empirical (verified) distributions.
5. cVaR must be calculated using Extreme Value Theory as well.
6. Simulates forward/future prices using Filtered Historical Simulation with and without GARCH, and Monte Carlo simulation, assuming Gaussian, Levy and empirical (verified) distributions.
7. A selected number (1 to N) of price simulations (past, forward/future) must be performed considering a number of X daily steps (one day to be divided in X pieces) and for Y days ahead.
8. Covar matrix and other relevant statistics must be informed and visualized.
9. Graphs, charts, distributions and statistics must be presented.
10. Code must be delivered in Python for Linux, well commented and documented.
11. Sources (articles, books, sites) must be indicated.
2. Download the time series of closing prices from Yahoo and Google.
3. Check for consistency and appoint discrepancies and outliers.
4. Calculates VaR and cVaR (expected shortfall) according to past prices, using the following methods for the selected set of financial assets (1 to N):
a. Parametric;
b. Historical;
c. Filtered Historical Simulation http://filteredhistoricalsimulation.com/;
d. Monte Carlo simulation, assuming Gaussian, Generalised Pareto, Levy, Truncated Levy Flights and empirical (verified) distributions.
5. cVaR must be calculated using Extreme Value Theory as well.
6. Simulates forward/future prices using Filtered Historical Simulation with and without GARCH, and Monte Carlo simulation, assuming Gaussian, Levy and empirical (verified) distributions.
7. A selected number (1 to N) of price simulations (past, forward/future) must be performed considering a number of X daily steps (one day to be divided in X pieces) and for Y days ahead.
8. Covar matrix and other relevant statistics must be informed and visualized.
9. Graphs, charts, distributions and statistics must be presented.
10. Code must be delivered in Python for Linux, well commented and documented.
11. Sources (articles, books, sites) must be indicated.