Available via license: CC BY 4.0
Content may be subject to copyright.
MAE
original
RAE depending on damping factor (in %)
Model 0% 10 20 30 40 50 60 70 80 90 100
Abramowitz (16) 1.69 1.00 1.05 1.16 1.34 1.53 1.74 1.94 2.15 2.36 2.57
Campbell (17) 1.97 1.03 1.06 1.09 1.12 1.27 1.45 1.64 1.82 2.01 2.20
Cuzán (15) 2.05 1.00 1.06 1.17 1.34 1.54 1.73 1.93 2.12 2.32 2.52
Bellucci (8) 2.21 0.97 0.95 0.93 0.92 0.90 0.88 0.86 0.86 0.87 0.89
Lewis-Beck & Tien, US (18) 2.25 0.98 0.98 1.05 1.14 1.26 1.42 1.57 1.73 1.89 2.05
Lewis-Beck. Nadeau & Bélanger (14) 2.33 1.03 1.08 1.12 1.18 1.24 1.30 1.36 1.46 1.58 1.71
Erikson & Wlezien (20) 2.47 0.97 0.99 1.03 1.10 1.23 1.36 1.49 1.62 1.75 1.88
Jackman (5) 2.54 0.98 0.97 0.96 0.94 0.93 0.92 0.91 0.91 0.91 0.91
Holbrook (19) 2.59 0.96 0.91 0.95 1.07 1.19 1.31 1.43 1.55 1.67 1.79
Lockerbie (21) 2.73 1.00 1.01 1.05 1.08 1.15 1.25 1.36 1.46 1.57 1.67
Fair (3) 2.80 0.97 0.96 0.99 1.08 1.21 1.34 1.46 1.59 1.72 1.85
Magalhães & Aguiar-Conraria (11) 3.17 0.97 1.00 1.11 1.24 1.44 1.64 1.84 2.04 2.25 2.47
Cameron & Crosby (4) 4.22 0.97 0.94 0.91 0.90 0.88 0.87 0.86 0.86 0.85 0.86
Dassonneville, Lewis-Beck & Mongrain (10) 4.34 1.02 1.04 1.06 1.10 1.14 1.22 1.30 1.38 1.47 1.57
Bélanger & Godbout (6) 4.59 0.96 0.93 0.90 0.89 0.89 0.92 0.95 1.02 1.12 1.24
Lewis-Beck & Tien, JP (9) 4.79 0.96 0.96 0.98 1.00 1.02 1.06 1.09 1.13 1.25 1.38
Nadeau & Blais (7) 7.01 0.96 0.92 0.88 0.85 0.82 0.78 0.75 0.73 0.70 0.67
Magalhães, Aguiar-Conraria, Lewis-Beck (12) 9.86 0.95 0.91 0.86 0.82 0.79 0.76 0.73 0.70 0.67 0.67
Toros (13) 12.44 1.02 1.08 1.18 1.27 1.36 1.46 1.55 1.68 1.82 1.96
Mean 4.00 0.98 0.99 1.02 1.07 1.15 1.23 1.32 1.41 1.52 1.62
Notes:
Models are ordered from most to least accurate by mean absolute error (MAE) of the original individual regression models’ out-of-sample
error (determined by N-1 cross validation) across all available observations.
The MAEs are the percentage point error that one would achieve without any (i.e., 0%) damping.