Skip to main content

Table 6 Diebold and Mariano (1995) test with respect to benchmark models

From: Forecasting inflation in the euro area: countries matter!

Horizon   HICPex HICP
  BVAR-Base BVAR-Base
One quarter RW 3.05 1.11
UC-SV \(-\) 1.46 \(-\) 0.94
Two quarters RW 2.68 1.51
UC-SV \(-\)1.97 0.15
One year RW 2.45 1.30
UC-SV 0.88 0.98
Two years RW 0.19 0.66
UC-SV 3.56 2.78
  1. RW is the random walk model; BVAR-Base is our large multi-country model; UC-SV is the unobserved component stochastic volatility model of Stock and Watson (2007). A negative value of the t statistic indicates that the model on the table’s row is more accurate than the model on the table’s column. The forecasts are computed over the period 2006Q1-2017Q2. Numbers marked in bold indicate that the test is significant at 10% level