Tables
2.1Confidence intervals of for one-day 99% VaR under different methods.
2.2One day 99% VaR confidence intervals on Hypothetical P&L as a percentage of the VaR estimate.
4.3Count and percent of subportfolios that pass the independence test.
4.4Count and percent of subportfolios that pass the conditional coverage test.
4.5Count and percent of subportfolios that pass the Ljung–Box test.
4.6Logit and LPM regressions of exceedances on lagged VaR or lagged P&L.
4.7Count and percent of subportfolios that pass the Kolmogorov–Smirnov test.
4.8Count and percent of subportfolios that pass the Anderson–Darling test.
4.9Count and percent of subportfolios that pass the Cramér–von Mises test.
4.10Linear regression of transformed PITs on lagged transformed PIT and lagged P&L.
6.1Predicting backtesting exceptions from lagged exceptions.
6.2Predicting backtesting exceptions from lagged exceptions by asset class.
6.3Predicting backtesting exceptions from lagged market factors.
6.4Predicting backtesting exceptions from lagged market factors by asset class in 2020.
6.5Associating backtesting exceptions with contemporaneous market movements.
6.6Associating backtesting exceptions with Contemporaneous market movements by asset class in 2020.
6.7Linear regression coefficients compared to logit marginal effects.
6.8Linear regression coefficients compared to logit marginal effects.
13.3Descriptive statistics for ratio of alternative estimates of the 99.9th quantile to lognormal estimates.
13.5Descriptive statistics for ratio of alternative estimates to lognormal estimates.
13.11Descriptive statistics on AMA capital and its benchmark.
15.1Information criteria and goodness-of-fit tests results for all benchmark copula models.
15.2Diversification benefits in percentage terms for each model at different quantiles.
15.3Backtesting results. Realized and expected violation counts for each model at different quantiles.
