Testing Density Forecasts, With Applications to Risk Management

By Jeremy Berkowitz. 2001

881 citations recorded by the source corpus.

Subfields: Demography, Economics and Econometrics, Finance

View Paper: https://www.semanticscholar.org/paper/329f503547b1a7bbf55b69b12d716dbab268af3a

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

In Demography

In Economics and Econometrics

In Finance

Citations in the corpus

International Asset Allocation Under Regime Switching, Skew and Kurtosis Preferences (2005, background): “Under the regime-switching process (7)-(8) and m−moment preferences (6), the expected utility is given by”
Combining forecast densities from VARs with uncertain instabilities (2010, methodology): “The goodness-of-fit tests employed include the Likelihood Ratio (LR) test proposed by Berkowitz (2001), for which results are presented at h > 1 using a two degrees-of-freedom variant (without a test for autocorrelation; see Clements (2004)). For h = 1, we use a three degrees-of-freedom variant with”
Markov Switching Models in Empirical Finance (2011, background): “Berkowitz (2001) and Vlaar and Palm (1993) show that inaccuracies in the specified density will be preserved by the transformed z-scores.”
Nowcasting GDP in Real Time: A Density Combination Approach (2014, positive): “We use a test of uniformity of the pits proposed by Berkowitz (2001). The Berkowitz test works with the inverse normal cumulative density function transformation of the pits. Then we can test for normality instead of uniformity. For 1-step ahead forecasts, the null hypothesis is that the transformed”
A Component GARCH Model with Time Varying Weights (2009, methodology): “By proposition 1 in Berkowitz (2001), the transformed series xt+h = Φ−1 (∫ yt+h −∞ ft(r)dr ) = Φ−1(Ft(yt+h)) where Φ(.) denotes the standard normal cdf, is distributed as a sequence of iid N(0,1) random variables.”
Monte Carlo methods for estimating, smoothing, and filtering one- and two-factor stochastic volatility models (2006, methodology): “This approach to model diagnostics is also discussed by Berkowitz (2001) and Bontemps and Meddahi (2005) among others, but not applied in the context of models with latent state variables.”
A Multivariate Generalized Orthogonal Factor GARCH Model (2007, methodology): “The limiting distribution of the estimator θ̂ can be obtained using theorem 4 of Comte and Lieberman (2003) and wellknown results of constrained (quasi-)ML estimation based on the use of Lagrange multipliers.”
Asset Allocation Under Multivariate Regime Switching (2005, neutral): “(8) Our approach is consistent with the notion that investors never observe the true state.”
Adaptive models and heavy tails with an application to inflation forecasting (2017, methodology): “For the one step-ahead we evaluate the calibration of the densities looking at the results of Berkowitz’s (2001) LR test and the nonparametric test of Rossi and Sekhposyan (2014).(26) For horizons beyond the one step-ahead we report the test of Knüppel (2015), which is robust to the presence of seri”
Probabilistic forecasts, calibration and sharpness (2007, background): “Smith (1985), Fruhwirth-Schnatter (1996) and Berkowitz (2001) proposed an assessment of independence based on the transformed PIT values Φ−1.pt/, which are Gaussian under the assumption of ideal forecasts.”
Flexible modeling of conditional distributions using smooth mixtures of asymmetric student t densities (2010, background): “Berkowitz (2001). It is clear from Figure 3 that even the SMR with largest LPDS produces much to large residuals during the most volatile period, and so does the GARCH(1,1) and t-GARCH(1, 1).”
VOLATILITY AND CORRELATION FORECASTING (2006, methodology): “Methods for density forecast evaluation are developed in Berkowitz (2001), Diebold, Gunther and Tay (1998), Giacomini (2002) and Hong (2000), as well as the chapter by Corradi and Swanson in this Handbook.”
The Hidden Dangers of Historical Simulation (2006, neutral): “VaR distributional assumptions are often judged in practice by whether the VaR measures provide the correct conditional and unconditional coverage for risk (Christoffersen, 1998; Diebold et al., 1998; Berkowitz, 2001).”
A Generalized Extreme Value Approach to Financial Risk Measurement (2007, background): “Moreover, as indicated by Kupiec (1995), Christoffersen (1998), and Berkowitz (2001), the unconditional coverage tests have low power against alternative hypotheses if the sample size is small.”
Backtesting for Risk-Based Regulatory Capital (2004, neutral): “A ﬁnal drawback is that by transforming the information of the distribution into one characteristic (exceeding of value-at-risk or not) we lose relevant information of the return distribution (see also Berkowitz (2001)).”