Evaluating Econometric Forecasts of Economic and Financial Variables

by Clements, Michael P.

• ISBN13: 9781403941565
• ISBN10: 1403941564
• Format: Hardcover
• Copyright: 2005-04-16
• Publisher: Palgrave Macmillan

Summary

In recent years there has been a growing recognition that forecasts of the most likely outcome of the future value of a variable may often be of limited value. Michael P. Clements describes methods of evaluating forecasts which provide a fuller description of the range of possible future outcomes.

Author Biography

Michael P. Clements is an editor of the International Journal of Forecasting.

Table of Contents

List of Tables	x
List of Figures	xi
Author's Preface and Acknowledgements	xii
Introduction	1	(3)
Point Forecasts	4	(42)
Realization-forecast regressions	4	(5)
Testing the rationality of multi-step forecasts	7	(2)
Forecast precision	9	(3)
Rival forecasts, forecast combination and encompassing	12	(9)
Tests of comparative forecast accuracy	12	(3)
Forecast combination (or pooling) and encompassing	15	(6)
Testing model-based forecasts for predictive accuracy	21	(9)
Tests of predictive accuracy	21	(4)
Tests of equal accuracy and encompassing when parameters are estimated	25	(5)
Non-linear models and forecasting	30	(15)
The conditional expectation is the MMSE predictor	30	(2)
Multi-step forecasts and non-linear models	32	(2)
SETAR models and multi-period forecasts	34	(3)
Markov-switching models	37	(2)
Evaluating non-linear model forecasts	39	(6)
Summary	45	(1)
Volatility Forecasts	46	(31)
Introduction	46	(2)
Changing conditional-variances and optimal point forecasts	48	(3)
Time-varying conditional variances and asymmetric loss	51	(3)
Models of conditional variance	54	(14)
ARCH models	54	(4)
Estimation	58	(1)
GARCH models	59	(3)
GARCH model forecasts	62	(1)
IGARCH	63	(1)
Non-linear GARCH	64	(3)
GARCH and forecasts of the conditional mean	67	(1)
Evaluation of volatility forecasts	68	(5)
Recent developments in the evaluation of volatility forecasts	73	(2)
Realized volatility	73	(1)
Intraday range	74	(1)
Utility-based measures and trading rules	75	(1)
Summary	75	(2)
Interval Forecasts	77	(26)
Introduction	77	(1)
Calculating interval forecasts	78	(9)
Bootstrap the forecasts	80	(1)
Allowing estimation uncertainty	81	(1)
Conditional intervals and estimation uncertainty	82	(1)
Bias-correcting the parameter estimates	82	(1)
Monte Carlo evaluation: step-by-step guide	83	(2)
Bootstrapping ARCH processes	85	(2)
Desirable properties of interval forecasts	87	(1)
Tests for conditional efficiency	88	(3)
Unbiasedness	88	(1)
Independence	89	(2)
Regression-based tests of conditional efficiency	91	(1)
Interval forecast construction and ARCH	92	(2)
Empirical illustration	94	(8)
Interval forecasts and intradaily data	94	(1)
Properties of futures returns data	95	(7)
Summary	102	(1)
Density Forecasts	103	(21)
Introduction	103	(1)
Probability distribution forecast evaluation	104	(2)
Joint probability distributions	106	(1)
Calibration	107	(1)
Density and interval forecasts	108	(2)
Empirical illustration (I): the SPF probability distributions	110	(2)
Empirical illustration (II): the MPC inflation forecasts	112	(5)
Point forecast performance	113	(3)
Evaluation of forecast densities	116	(1)
Model-based density evaluation	117	(4)
Model mis-specification	120	(1)
Summary	121	(1)
Appendix: multivariate forecast density probability integral transform tests	121	(3)
Decision-based Evaluation	124	(19)
Introduction	124	(2)
Decision-based evaluation -- some basic results	126	(1)
Quadratic loss functions	127	(2)
Two-state, two-action decision problems	129	(2)
Decision problem for inflation-targeting and interest rate setting	131	(2)
Statistical measures related to economic value	133	(2)
The Bank of England MPC inflation forecasts	135	(2)
Properties of optimal forecasts for general loss functions	137	(4)
General loss functions and the generalized forecast error	140	(1)
Summary	141	(2)
Postscript	143	(3)
Computer Code	146	(10)
Sample Gauss code for the estimation and forecasting of SETAR models	146	(4)
Extensions	147	(3)
Estimation and forecasting GARCH(1, 1) processes	150	(1)
Monte Carlo evaluation of interval lengths and coverages	151	(3)
Extensions	153	(1)
Forecast density evaluation	154	(2)
Notes	156	(4)
References	160	(10)
Index	170