Repository Public
1{"job":{"id":"da55283b-b66e-4466-b74b-df7c275bd77b","tool":"var","params":{"symbol":"SPY","portfolio_value":100000,"confidence_levels":[0.9,0.95,0.99],"time_horizon":1,"num_simulations":10000,"data_period":504},"status":"done","startedAt":1795041547976,"finishedAt":1795041549084,"error":null,"code":"import numpy as np\nfrom scipy import stats\nfrom fmpClient import get_historical_prices\n\n# Fetch historical data for SPY\nsymbol = 'SPY'\nprint(f\"Fetching historical prices for {symbol}...\")\nhist_data = get_historical_prices(symbol)\nprices = np.array([h['close'] for h in reversed(hist_data['historical'])])\n\n# Calculate daily returns\nreturns = np.diff(prices) / prices[:-1]\n\n# Portfolio parameters\nportfolio_value = 100000\nconfidence_levels = [0.9,0.95,0.99]\ntime_horizon = 1 # days\n\n# Scale returns for time horizon\nif time_horizon > 1:\n returns_scaled = returns * np.sqrt(time_horizon)\nelse:\n returns_scaled = returns\n\n# Calculate VaR using different methods\nfor conf_level in confidence_levels:\n # Historical VaR\n hist_var = -np.percentile(returns_scaled, (1 - conf_level) * 100)\n hist_var_dollar = hist_var * portfolio_value\n\n # Parametric VaR (assumes normal distribution)\n mean = np.mean(returns_scaled)\n std = np.std(returns_scaled)\n z_score = stats.norm.ppf(1 - conf_level)\n param_var = -(mean + z_score * std)\n param_var_dollar = param_var * portfolio_value\n\n # Monte Carlo VaR\n simulated = np.random.normal(mean, std, 10000)\n mc_var = -np.percentile(simulated, (1 - conf_level) * 100)\n mc_var_dollar = mc_var * portfolio_value\n\n # Conditional VaR (CVaR)\n var_threshold = -hist_var\n tail_losses = returns_scaled[returns_scaled <= var_threshold]\n cvar = -np.mean(tail_losses) if len(tail_losses) > 0 else hist_var\n cvar_dollar = cvar * portfolio_value\n\n print(f\"\\n{int(conf_level*100)}% Confidence Level:\")\n print(f\" Historical VaR: {hist_var*100:.2f}% (${hist_var_dollar:,.2f})\")\n print(f\" Parametric VaR: {param_var*100:.2f}% (${param_var_dollar:,.2f})\")\n print(f\" Monte Carlo VaR: {mc_var*100:.2f}% (${mc_var_dollar:,.2f})\")\n print(f\" CVaR (Expected Shortfall): {cvar*100:.2f}% (${cvar_dollar:,.2f})\")","result":{"success":true,"symbol":"SPY","parameters":{"portfolio_value":100000,"time_horizon":1,"time_horizon_description":"1-day VaR","num_simulations":10000,"data_points":503,"data_period_days":504},"var_by_confidence":{"90%":{"confidence_level":0.9,"historical":{"var_percentage":1.096,"var_dollar":1096.04},"parametric":{"var_percentage":1.3178,"var_dollar":1317.79},"monte_carlo":{"var_percentage":1.3402,"var_dollar":1340.25},"conditional_var":{"cvar_percentage":1.8668,"cvar_dollar":1866.8}},"95%":{"confidence_level":0.95,"historical":{"var_percentage":1.5641,"var_dollar":1564.14},"parametric":{"var_percentage":1.7047,"var_dollar":1704.65},"monte_carlo":{"var_percentage":1.7302,"var_dollar":1730.18},"conditional_var":{"cvar_percentage":2.4227,"cvar_dollar":2422.74}},"99%":{"confidence_level":0.99,"historical":{"var_percentage":2.702,"var_dollar":2702},"parametric":{"var_percentage":2.4304,"var_dollar":2430.35},"monte_carlo":{"var_percentage":2.4858,"var_dollar":2485.78},"conditional_var":{"cvar_percentage":4.2094,"cvar_dollar":4209.43}}},"distribution_statistics":{"mean_daily_return":0.0469,"std_daily_return":1.0649,"annual_return":11.82,"annual_volatility":16.9,"skewness":1.0045,"kurtosis":21.3238,"is_normally_distributed":false,"jarque_bera_pvalue":0},"extreme_values":{"worst_daily_loss_pct":5.8543,"worst_daily_loss_dollar":5854.29,"best_daily_gain_pct":10.5019,"best_daily_gain_dollar":10501.93},"interpretation":{"distribution_type":"non-normal","tail_risk":"high","recommended_method":"historical","notes":["Historical VaR uses actual distribution of returns","Parametric VaR assumes normal distribution","Monte Carlo VaR uses simulated returns","CVaR shows average loss beyond VaR threshold","Distribution is non-normal (Jarque-Bera test)"]}}}}