Comparing Predictive Accuracy

By Francis X. Diebold, Roberto S. Mariano. 1995

We propose and evaluate explicit tests of the null hypothesis of no difference in the accuracy of two competing forecasts. In contrast to previously developed tests, a wide variety of accuracy measures can be used (in particular, the loss of function need not be quadratic and need not even be symmetric), and forecast errors can be non-Gaussian, nonzero mean, serially correlated, and contemporaneously correlated. Asymptotic and exact finite-sample tests are proposed, evaluated, and illustrated.

How researchers see this paper

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

Citations in the corpus

Deep Learning for Predicting Asset Returns (2018, methodology): “The most popular method for testing signiﬁcant difference in MSPE is the Diebold and Mariano (2002) and West (1996) test.”
History and trends in solar irradiance and PV power forecasting: A preliminary assessment and review using text mining (2018, methodology): “For instance, twenty years after the original paper (Diebold and Mariano, 1995) was proposed, the lead D.”
Modeling and Forecasting Medium-Term Electricity Consumption Using Component Estimation Technique (2020, methodology): “To verify the superiority of the results listed in Table 2, we performed Diebold and Mariano (DM) test for each pair of models [70].”
Fitting Vast Dimensional Time-Varying Covariance Models (2020, background): “The related tests of Diebold and Mariano (1995) and West (1996) focus solely on comparing forecast models and are thus well-suited for our problem.”
Panel parametric, semiparametric, and nonparametric construction of counterfactuals (2018, unknown): “Unfortunately, neither the conventional hypothesis testing approach, nor the predictive approach (e.g., Diebold and Mariano (1995), White (2000)) appears feasible to assess the appropriateness of which method is more likely to generate more accurate counterfactuals in a given situation because…”
Analyzing Crude Oil Prices under the Impact of COVID-19 by Using LSTARGARCHLSTM (2020, methodology): “The Wilcoxon signed-rank (WS) and Diebold–Mariano (DM) tests were applied to confirm the equivalence of forecast accuracy [86].”
Harnessing jump component for crude oil volatility forecasting in the presence of extreme shocks (2019, methodology): “First, we used the Diebold and Mariano (DM) (Diebold and Mariano, 1995) test(6) to evaluate the pairwise forecasting differences between each of our extensions and the benchmark model.”
The GDP-Temperature relationship: Implications for climate change damages (2021, unknown): “Moreover, we invoke the substantial literature on data-driven, predictive model comparison (e.g., White 1996; Diebold and Mariano 1995; Hansen 2005; Hansen et al. 2011) to identify the set of models that are statistically superior to alternatives.”
Macroeconomic forecast accuracy in a data‐rich environment (2019, methodology): “(9)The tables with MSPE relative to the ARD (autoregressive direct) model used as a benchmark are available in the supplementary material, along with the Diebold and Mariano (1995) (DM) test.”