Show Summary Details

Page of

PRINTED FROM OXFORD HANDBOOKS ONLINE (www.oxfordhandbooks.com). © Oxford University Press, 2018. All Rights Reserved. Under the terms of the licence agreement, an individual user may print out a PDF of a single chapter of a title in Oxford Handbooks Online for personal use (for details see Privacy Policy and Legal Notice).

date: 19 November 2019

(p. 691) Index

(p. 691) Index

absolute deviations of MAPE (ADMAPE), 643, 644
“absorbing barrier,” 328
accuracy, forecast. See forecast accuracy
ACD (autoregressive conditional duration) model, 518–19
ACFs (autocorrelation functions), 533–35, 534f
ACR model, 336
adaptive expectations model, 474–78, 492
ADL (autoregressive distributed lag)-MIDAS regressions, 231–35
ADMAPE (absolute deviations of MAPE), 643, 644
administrative data sets, for health care cost models, 627
aggregated disaggregate forecast error taxonomy, 291–92, 292t
aggregated-disaggregate robust forecast error taxonomy, 294–95, 295t
aggregate forecast error taxonomy, 292–94, 293t
aggregating disaggregated forecasts, 291–92
AIC (Akaike's information criterion), 16, 23, 24t, 25, 26, 27t, 29t
AIM algorithm, 96
Akaike's information criterion. See AIC
algebra, of consensus forecasts, 464–67
Almon weighting scheme, 228, 229, 237, 240
alternative hypothesis, 391. See also null hypothesis
American Statistical Association–National Bureau of Economic Research (ASA-NBER), 166, 481
analogy forecasting, 76
analogy usage, 173t, 179–80
analytical point forecasts, 72–73
anchoring and adjustment heuristic, 168, 170, 173t
ANNs. See neural networks
anticipated changes, 478–81. See also multidimensional panel data of survey forecasts
approximate factor models, 39, 41
approximating logistic transition function, 330–31
approximators. See universal approximators
AR (univariate), 361, 365t, 366t
ARCH (autoregressive conditional heteroskedasticity) models. See also GARCH models
forecast failure and, 280
HARCH, 541
MIDAS regressions and, 239
volatility forecasting and, 239, 507–9
AR_FAC (multivariate factor-based models), 361. See also forecast combinations illustration: survey and time series forecasts
ARFIMA (autoregressive fractionally integrated moving average) models, 526
ARMA structure and, 526–27
electricity spot prices and, 616, 617
GARCH models v., 540–41
nonlinearity and, 543
volatility forecasting and, 526–27, 540–41
white-noise process and, 514
ARIMA (autoregressive integrated moving average) models
random walk hypothesis and, 505
reduced-form volatility forecasting and, 527
sampling the future algorithm and, 102
structural time series models v., 131
ARI model, linear, 77, 78f
ARMA (autoregressive moving average) models
ACD model and, 519
ADL-MIDAS regressions v., 232
(p. 692) ARFIMA framework and, 526–27
Bass model and, 677
conditional predictability testing and, 453
cycle component as, 136
data revisions model and, 253
expert forecasts and, 686
multivariate version of, 151
smooth trend and, 139–41
UC-RV model and, 541
vector fractionally integrated, 541
Arrow, Kenneth, 561
Artificial neural networks. See neural networks
ASA-NBER (American Statistical Association–National Bureau of Economic Research), 166, 481
assessing forecast performance. See forecast performance assessment
asymptotic approximation
with growing in-sample size, 443
with growing out-of-sample size, 443–44, 453–54
autocorrelation functions. See ACFs
automatic model selection algorithms, 72, 332, 348
automatic statistical procedure, nowcasting v., 195
Autometrics approach, 72, 82, 82t, 83t, 319, 319n1, 326, 327, 332, 333, 335, 336, 347, 348
autoregressive conditional duration ACD model, 518–19
autoregressive conditional heteroskedasticity. See ARCH models
autoregressive distributed lag MIDAS. See ADL-MIDAS regressions
autoregressive fractionally integrated moving average. See ARFIMA models
autoregressive integrated moving average. See ARIMA models
autoregressive neural network models, 79
autoregressive representation, of unobserved components time series models, 148
availability heuristic, 168, 169–70, 173t
Bachelier, Louis, 504–5
backcasts, 207, 600–601, 602
Finance Act of 1984, 321
Bank of England's Survey of External Forecasters, 467
Bass model, 674–82. See also sales forecasting
forecasting moment of peak sales with, 680–81
generalized, 682
important extensions of, 681–82
LCD television sets illustration, 678–80, 679f, 680t
sales forecasts generated from, 676–78
Bayesian estimation, of dynamic factors, 44–46
Bayesian forecasting. See sampling the future method
Bayesian information criterion. See BIC
Bayesian reduced-rank VECMs, 18–19
Bayesian VARs. See BVARs
Bayesian VECMs, 25
BCEI. See Blue Chip Economic Indicators
bellwethers, 657, 658
benchmark revisions, 3, 248, 249, 250, 251, 253, 259, 262, 264, 307. See also data revisions
Beta weight function, 228, 240, 241, 539
Beveridge-Nelson (BN) decomposition, 11, 12, 14
BFGS (Broyden-Fletcher-Goldfarb-Shannon) method, 143
biases (in judgment), 164, 167–72. See also judgment
anchoring and adjustment heuristic, 168, 170, 173t
availability heuristic, 168, 169–70, 173t
confirmation, 165, 171, 173t
contrarianism and, 167
correcting for, 173t, 176–77
egocentric attribution, 171
false consensus effect, 171
group, 173t, 181
herding behavior and, 167, 172, 181
heuristics, 168–70
hindsight, 171, 173t
illusion of control, 171
information use, 170, 172
motivational, 167–68, 173t
optimism, 171–72, 173t
overconfidence, 170–71, 173t
psychological, 168, 170
“publicity hypothesis” and, 167
representativeness heuristic, 168–69
“reputational cheap talk” hypothesis and, 167
selective recall of information and, 170, 172
solutions to, 172–82, 173t
strategic, 167–68, 173t
systematic patterns in random events, 169, 173t
“wishful expectations hypothesis” and, 168
BIC (Bayesian information criterion), 16, 23, 24t, 25, 26, 27t, 29t
bilinear models, 67, 73
binary health care cost models, 626
“black arts,” 174
Black–Scholes option pricing model, 507, 509
Blue Chip Economic Indicators (BCEI), 473, 480n3, 487, 488f, 489, 490, 492
BN (Beveridge-Nelson) decomposition, 11, 12, 14
Board of Governors of the Federal Reserve, 195
Bonferroni's one-step procedure, 392, 393, 400, 404
bootstrapping method, psychological, 173t, 177–78
Box-Cox models, 631–32, 639, 642t, 646, 647
Box–Jenkins time series models, 130, 684
(p. 693) breakpoint Chow test rejection frequencies, 341, 341f
breaks. See structural breaks and location shifts
bridge equations, 194, 203
Brier score, 444, 563, 568, 569, 569f, 570
Brownian motion, 501, 504–5
Broyden-Fletcher-Goldfarb-Shannon (BFGS) method, 143
business cycle component. See cycle component
business cycle theory, real, 10, 17, 122
BVARs (Bayesian VARs). See also NAWM benchmarks
advances in specification/estimation of, 9
DSGEs with, 104
VECMs and, 19
Canonical correlation analysis (CCA), 10, 11, 13n1, 14, 18, 19, 23, 32–34
Capital asset pricing model(CAPM), 520
CAPM. See capital asset pricing model
Captain Cook/Maori example, 321
carbon emissions markets, 609. See also electricity spot markets
energy commodities bundle and, 609
EUAs and, 609
UK spot carbon prices (time series), 610f
case studies, weather/climate forecasts, 570–73
CCA. See canonical correlation analysis
Center for Economic Studies ifo Institute World Economic Survey, 467
CFNAI (Chicago Fed's National Activity Index), 430
chance, election forecasting and, 669
chaos theory, 61, 75, 84, 517
chartism (technical analysis), 500, 501
Chicago Fed index, 204
Chicago Fed's National Activity Index (CFNAI), 430
choice questions, 574
civil war example, 322
cleaning data, 532
climate forecasts. See weather and climate forecasts
coal market. See also electricity spot markets
carbon emissions markets and, 609
cointegration analyses and, 608
as economic sentiment driver, 608, 620
UK spot coal prices (time series), 610f
volatility of, 620
cobreaking
common nonlinear features and, 17
defined, 17
in VECMs, 17
Cochrane–Orcutt estimator, 42
cognitive feedback, 175
coincident indicators of economic activity, 195, 204, 213. See also nowcasting
cointegrating vectors, 10, 13, 15, 17, 19, 19n2, 20, 31, 286, 521
cointegration
CCA and, 10
coal market and, 608
common trends and, 263
DFMs and, 53–54
forecasting impacted by, 16–17
foreign exchange rates and, 521
fractional, 521
interest rates term structure and, 521
oil market and, 608
real business cycle theory and, 10
VECMs and, 10, 17
combination forecasting. See forecast combinations
combining survey forecasts. See survey forecasts
common nonlinear features, 17
common trends and cycles
cointegration and, 263
common cycles hypothesis, 14
multivariate time series models with, 130, 150–52
complex dynamic systems, 61, 75–76
conditional predictive ability tests, 4, 441–55
asymptotic approximation (with growing out-of-sample size), 443–44, 453–54
decision rules for forecast selection, 452–53
econometric methodology, 444–52
extensions, to different testing problems, 450–52
“fixed forecasting scheme” and, 442
fluctuation test, 446f, 447–48, 449t, 454
illustration of, 441–42
one-time reversal test, 446f, 447, 448–50, 454
open research questions, 453–54
recursive forecasting scheme and, 442, 443, 444
time-varying predictive ability testing, 446–50
unconditional predictive ability tests v., 416, 438, 441, 442–43, 444
when to use, 443–44
confirmation bias, 165, 171, 173t
conjoint analysis method, 574, 676
Consensus Economics, Inc., 492
consensus forecasts, 457–58, 464–67. See also survey forecasts
accuracy of, 464–65
algebra of, 464–67
anchoring and adjustment bias and, 170
arguments against, 469
multidimensional panel data of survey forecasts and, 469, 474, 484, 485
“wisdom of crowds,” 465
contrarianism, 167
convex loss functions
Jensen's inequality for, 458, 465, 485, 565
mean forecasts and, 465–66
Copas test, 628–29, 629n2, 643, 647
cost-effectiveness analysis, health care costs models and, 626
(p. 694) cost-loss decision-making model, 560, 567–70, 571. See also weather and climate forecasts
dynamics in, 570
notation and expense matrix for, 568t
quality/value curves for, 569, 569f
real-world situations and, 570
cost regression health care costs models, linear, 627–29
count data health care costs models, 626
cross-country panels, long-horizon growth forecasting and, 599–600, 602
cross-equation restrictions
NAWM, 93, 117, 119
VARs, 1–2, 10, 11, 17
cross-forecaster dispersion, in point predictions, 462–63
cross-sectional averaging estimators, 39–40
cross-sectional shocks, 478–80, 479f, 488
cumulative shocks, 478–80, 479f, 487, 488
cumulative sum (CUSUM) test, 447
“curse of knowledge,” 172
CUSUM (cumulative sum) test, 447
cycle component, 129–30. See also smooth cycle component
ARMA process, 136
common trends and cycles, 130, 150–52
multiple cycles, 137
random walk plus smooth cycle model, 140–41
smooth time-varying trigonometric, 137
specification, 136–37
time-varying trigonometric, 136–37
data/data sets. See also high-frequency data; information; MIDAS regressions; mixed-frequency data; multidimensional panel data of survey forecasts; panel-data sets; real-time data
administrative, for health care cost models, 627
cleaning, 532
DFMs and, 35
forecasting combinations illustration, 361, 362t363t
Google Trends, 319, 320, 326, 348
large real-time databases (online list), 252
Litterman, 25–26, 27t, 28, 29t
Microeconometrics Using Stata, 644
nowcasting model, 205–6, 206t, 217t219t
data-generating processes. See DGPs
data issues. See also missing data; mixed-frequency data; nowcasting; real-time data vintages
models and, 1
overview, 3
data measurement errors, 278. See also forecast origin
data mining, 391, 429
data revisions. See also data revisions example; real-time data
benchmark revisions, 3, 248, 249, 250, 251, 253, 259, 262, 264, 307
DFMs and, 261–62
G7 countries (study), 254
impact on forecasts, 3, 247, 255–59, 263–64
location shifts and, 307–8
model of, 253–54
news revisions v. noise revisions issue, 260–61
optimal forecasting when data revisions exist, 259–64
real-time data analysis bibliography (online), 254
real-time data structure, 248–49, 248t
real-time data v., 257–58
repeated observation forecasting, 255–56, 256f
RMAR and, 253
sizes of, 252–53
state-space models and, 262, 307
unconditional predictive ability tests and, 428–29
unresolved questions, 264
data revisions example (residential investment)
growth rate (1976 Q3), 249–50, 249f
growth rates with revisions from initial to latest, 250–51, 250f
half-decade growth rates, 251, 251t
Real-Time Data Set for Macroeconomists, 248t, 251t, 252
real-time data structure, 248–49, 248t
repeated observation forecasts (1976 Q4), 255–56, 256f
data snooping, 391–92
data vintages. See also real-time data
location shifts and, 307–8
real-time, 247–67
vintage dates, 248, 248t, 250, 251t, 255, 256, 257
DCG (Diagnostic Cost Group) system, 626
DDD (double-differenced device), 283–84
disaggregation by time and, 301
disaggregation over variables and, 294–96
LDGP nonconstancy and, 284–87
VEqCM v., 283–87
decision analytic valuation studies, 560, 567–73. See also weather and climate forecasts
case studies, 570–73
limitations of, 579
prototypical decision-making models, 567–70
decomposition, of judgment, 173t, 180–81
Delphi method, 173t, 181–82
demography. See also long-horizon economic growth forecasting
approaches/methods for demographic-economic connections, 587, 592–602
demographic inertia, 586, 590–92, 602
economic growth and, 578–90, 602–3
(p. 695) fertility intentions and, 461–62
GDP process and, 586
stable populations and, 589–90
density forecasts
NAWM and, 113–17, 116f, 117f
Prequential approach and, 113
det(MSFE), 22t, 24t, 26, 27t, 29t
det (MSFEh), 20, 21
deterministic breaks, 277, 278, 285, 287
DFMs. See dynamic factor models
DGPs (data-generating processes). See also LDGPs
forecasting during breaks and, 342t
location shifts and, 17, 280–81
reduced-rank VECMs and, 19–25
structural breaks and, 272
unconditional predictive ability Monte Carlo experiments and, 430–36
Diagnostic Cost Group (DCG) system, 626
diagnostic tests. See health care costs models; specific tests
differenced VEqCM, 280, 287–88, 347
direct forecasts, iterated v., 48–49, 226
direct point forecasts, 75, 81–83
disaggregation by time, 273, 296–301, 309
break prediction and, 326
comparison of forecasts (disaggregated data v. time-aggregated data), 297–301, 309
DDD taxonomy, 301
disaggregation over variables, 273, 289–96, 309. See also forecasting in presence of unanticipated location shifts
aggregated disaggregate forecast error taxonomy, 291–92, 292t
aggregated-disaggregate robust forecast error taxonomy, 294–95, 295t
aggregate forecast error taxonomy, 292–94, 293t
aggregating disaggregated forecasts, 291–92
break prediction and, 326
DDD and, 294–96
forecasting aggregate directly by its past, 292–94
with unanticipated breaks, 291
discrete conditional density estimator, 626, 640–41, 645n16
discrete shocks, 478–81, 479f
discrete-time GARCH models, 543
distributed lag (DL) models
MIDAS regression models and, 227–28
VARS models and, 10
volatility forecasting and, 537–40
DL-MIDAS models, 230–31, 232, 235
DL models. See distributed lag models
dollar/sterling exchange rate changes, 502, 504f, 506, 506t, 509–13, 510f
double differenced device. See DDD
DSGEs (dynamic stochastic generale quilibrium models). See also NAWM
BVARs with, 104
DFMs v., 52, 55, 90
increase in scale of, 123
misspecification and, 124
reduced-form models v., 90
VARs v., 2, 89, 90, 104
DSGEs, log-linearized, 89–124. See also NAWM estimating predictive distribution of, 101–3, 123
state-space form for, 90
DSGE-VAR models, 104, 104n2, 124
durable products models, 674–82, 687. See also Bass model; sales forecasting
DVARs (VARs in differences), 17, 26
dynamic factor models (DFMs), 2, 35–59. See also dynamic factors; model instability illustration-DFMs combination; nowcasting model; static factors
ADL-MIDAS models and, 233–35
appeal of, 54
cointegration and, 53–54
data revisions and, 261–62
data sets and, 35
DSGE estimation with, 52, 55, 90
early research, 35
error correction in, 53–54
extensions of, 52–54
forecast combinations and, 306–7
forecast combinations experiment with, 357–58, 382–86
hierarchical, 54
motivation for using, 36
nonlinear, 45–46
nonlinear time series models v., 2
nowcasting and, 55, 195, 198–200, 202, 203, 205, 209, 212, 213
outlook for, 54–55
pseudo out-of-sample forecasting exercises and, 49
recent literature on, 36
with structural breaks, 52–53
SVARs and, 55
with time-varying parameters, 52–53
unobserved components time series model (multivariate version) and, 130, 152–53
dynamic factors
Bayesian estimation of, 44–46
determining number of, 46–48, 58
innovations, estimation of, 42
as instrumental variables, 50
MLE of, 37, 38–39, 43, 45
in second-stage regressions, 48–50
state-space model with, 43–44
time-domain estimation of, 37–44
uses of, 48–52
dynamic principal components, 42
(p. 696) dynamic stochastic general equilibrium models. See DSGEs
ECM (exponential conditional mean) health care costs models, 632–33, 634, 636, 639, 645, 646, 648t, 649t
econometric framework
conditional predictive ability testing, 444–52
nowcasting model(euro-area GDP), 198–202, 213
econometric models, robust forecasting models v., 279–88
economic growth. See long-horizon economic growth forecasting
economic value of imperfect information, 560, 564–67
economic value of weather and climate forecasts. See weather and climate forecasts
ECX (European Climate Exchange), 609
EEE models. See extended estimating equations health care costs model
efficiency condition, Nordhaus's, 474–75
efficient capital markets theory, 505
EGARCH (exponential GAR CH) model, 508, 509, 510, 511, 512, 515, 541
egocentric attribution, 171
Einstein, Albert, 505
election forecasting, 5, 655–71
approaches to, 658–59
history of, 657–58
political betting markets, 658, 659
political forecasting and, 655–56, 668
polls, 657, 658–59
probability and, 657
election forecasting statistical models
assessment, 663–66
generic model with examples, 660–62
presidential elections, 666–68, 668t
reasons for selecting, 668
success of, 668–69
UK example, 662–63
electricity daily spot prices (Nord Pool exchange market), 130–31, 153–59
bivariate analysis, 156
forecasting results, 156–59
time series decomposition of electricity prices, 153–55, 155f
time series decomposition of electricity prices and consumption, 156, 157f
univariate analysis, 153–55
electricity spot markets
marginal cost supply function (for power-generating facilities in Germany), 612, 612f, 614
oligopolistic nature of, 614, 617, 618–19
power market and, 620–21
price formation in, 611–15
UK demand evolution for load periods, 612–13, 613f
electricity spot prices
ARFIMA models and, 616, 617
modeling with multiple explanatory drivers, 617–20
power prices, 609–10, 610f
power spot modeling, 610
stochastic models of, 615–17
time series class of models and, 611
UK spot power prices (time series), 610f
UK spot prices for load periods, 614f
volatility of, 609–10, 613
EM (expectation maximization) algorithm, 39, 143, 153, 201–2, 203–4, 221–23
energy commodities (forecasting/modeling), 5, 607–21. See also electricity spot markets
as bundle, 607–9
carbon emissions market, 609, 610f
challenge of, 620
linkages of, 607–9
power market and, 620–21
supply shocks, 607, 610, 613, 616, 620
equal weights (EW), 364, 365t, 366t
equilibrium correction models, 273, 520, 521. See also VEqCMs
error correction, in DFMs, 53–54
estimated factors. See dynamic factors
estimation uncertainty, parameter, 277, 278, 285
EU emissions allowances (EUAs), 609
euro-area GDP. See nowcasting model
euro-area model. See NAWM
Eurocoin index, 204
European Climate Exchange (ECX), 609
EW(equal weights), 364, 365t, 366t
exact static factor model, 41
expanding window estimates, 18, 26, 27, 28, 29t, 364, 384, 385. See also rolling window estimates
expectation maximization algorithm. See EM algorithm
expert sales forecasts, 673, 674, 684–87
exponential Almon weighting scheme, 228, 229, 237, 240
exponential conditional mean health care costs models. See ECM health care costs models
exponential GAR CH. See EGARCH model
extended estimating equations (EEE) health care costs model, 632, 639, 642, 642t, 643, 645, 646, 647, 648t, 649, 649t
factor-augmented vector auto regression models (FAVARs), 2, 37, 42, 45, 48, 50–51
factor error-correction model, 54
factor models. See dynamic factor models
FADL-MIDAS model, 235, 236
fallowing/planting case study, 572
false consensus effect, 171
false discovery, 392
false discovery rate. See FDR
family-wise error rate. See FWER
(p. 697) FAVARs. See factor-augmented vector autoregression models
FDR (false discovery rate)
controlling for, 393, 404–7
controlling for, with given λ, 406–7
defined, 393
estimating, 405–6
optimal choice for λ, 407
type 1 error, 392, 393, 406, 410, 411
Federal Open Market Committee, 484
Federal Reserve Economic Data (FRED), 25, 25n4
Federal Reserve System
Board of Governors, 195
CFNAI, 430
Chicago Fed, 204, 430
Federal Open Market Committee, 484
Green books, 195, 478
Livingston Survey, 167, 467, 473, 481, 492
Philadelphia Fed, 204, 252, 264, 361, 417n2, 467, 468
Real-Time Data Set for Macroeconomists, 248t, 251t, 252, 361
SPF, 356, 467
St. Louis Fed, 25, 417n2
feedback, 173t, 174–76
cognitive feedback, 175
outcome feedback, 174–75
performance feedback, 175
task properties feedback, 175–76
fertility intentions, interpreting, 461–62
FIEGARCH (fractionally integrated EGARCH) model, 511f, 512f, 515
FIGARCH (fractionally integrated GARCH) model, 510f, 515
final forecasts, based on judgment, 166–67
financial crisis of 2007-2010, 316, 318, 336
financial time series forecasting, 4, 499–524. See also volatility forecasting
difficulty of, 521
dollar/sterling exchange rate changes, 502, 504f, 506, 506t, 509–13, 510f
long-horizon, 517–19
martingale assumption and, 505, 506, 513, 519
random walk hypothesis and, 501–6
S&P 500 stock market index returns, 499, 502, 502f, 506, 506t, 509–13, 511f
Technical analysis and, 500
Treasury bill yield changes, 499, 502, 503f, 506, 506t, 509–13, 512f
usefulness of published papers on, 500, 521
volatility forecasting and, 506–13
finite mixture health care costs models, 626, 639–40, 644, 647, 649
finite-sample predictive ability tests, 423–26
Clark/McCracken study, 425–26
Giacomini/White study, 423–24
null hypothesis and, 417
firms class, NAWM, 92, 95, 99
first arc sine law of probability, 502
first-order Markov chain, 65, 356n1, 570
fiscal authority, NAWM, 92
fixed dummy seasonal component, 133–34
“fixed forecasting scheme,” 442
fixed trigonometric seasonal component, 134–35
Fletcher. See BFGS method
fluctuation test, 446f, 447–48, 449t, 454
forecast accuracy
consensus forecasts, 464–65
NAWM, 107
relative, performance assessment and, 277
unconditional predictive ability tests and, 415–16, 418, 419t
forecast combinations, 3, 355–88. See also model instability
assigning weights in, 356
data revisions/real-time data, 263
design of model universe in, 356
DFMs and, 306–7
empirical studies, 305–6
GVARs and, 305–6
individual forecast models v., 355, 386
misspecification biases and, 386
model instability's impact on, 357
as pooling, 301
portfolio theory v., 304–5
risk and, 386
SPF and, 356, 356n1, 364, 366t, 369t
statistical/judgmental, 173t, 178–79
structural breaks and, 273, 301–7
VARs and, 305
forecast combinations experiment: in presence of breaks, 357–58, 382–86
empirical results, 384–86, 385t
MSFE values, 384–86, 385t
forecast combinations illustration: survey and time series forecasts, 3, 358–61
all models combination, 373, 374t
data, 361, 362t363t
design of model universe, 358–61
empirical analysis, 358, 361–74
equal-weighted combination scheme, 359–61
individual time series models, in universe of models, 361
individual time series models, performance, 364, 365t366t, 367
linear/nonlinear models combinations, 361, 370, 371t372t, 373
model-based/survey forecasts combinations, 361, 367, 368t369t, 370
pseudo out-of-sample forecasts and, 364
restricted universes of models, 361
survey forecasts, in universe of models, 361
survey forecasts, performance, 365t366t, 367
variables in, 361, 362t363t
forecast densities, nonlinear models and, 84
forecast errors, 277–79. See also GFESM measure; MSFE matrix
(p. 698) forecast error taxonomies, 277–79
aggregate, 292–94, 293t
aggregated disaggregate, 291–92, 292t
aggregated-disaggregate robust, 294–95, 295t
data measurement errors, 278
deterministic breaks, 277, 278, 285, 287
forecast origin mismeasurement, 278, 285, 292t, 293t, 295, 295t, 307
LDGP innovation error, 279, 280, 285, 287, 292t, 293t, 295t, 308
location shifts, 278
misspecification of in-sample deterministic factors, 278
misspecification of stochastic components, 278
paramete restimation uncertainty, 277, 278, 285
stochastic breaks, 277, 278, 279, 285, 357
forecast evaluation, 3–4. See also conditional predictive ability tests; forecast performance assessment; multidimensional panel data of survey forecasts; multiple forecast model evaluation; survey forecasts; unconditional predictive ability tests
forecast failure
ARCH models and, 280
DDD and, 283–84
defined, 163, 279
location shifts and, 17, 166, 273, 278, 279, 280, 282, 283
nowcasting and, 163
reasons for, 17, 166
forecasting (macroeconomic forecasting). See also election forecasting; financial time series forecasting; nonlinear time series forecasting; sales forecasting; volatility forecasting
cointegration's impact on, 16–17
judgment in, 2, 163–89
models v., 287
structural breaks' impact on, 346–47, 353
weather and climate forecasting v., 561
forecasting breaks, 3, 315–53. See also forecasting during breaks; forecasting models; information
forecasting regime shift, 336–39
necessary conditions for, 316–18, 336
predictability and, 317
strategy outline, 319–20
unpredictability and, 321–24, 348
forecasting during breaks, 336. See also forecasting breaks
forecasting during new exponential break, 345–46
forecasting during new location shift, 339–41
forecasting models, 341–42, 342t
ME and, 342f, 343f, 344, 345f
Monte Carlo analysis, 341–45
MSFE and, 339, 346, 346f
RMSFEs and, 340f, 341, 342f, 343f, 344, 345, 345f
forecasting in presence of unanticipated location shifts (with misspecified models), 271–314. See also forecast error taxonomies
data revisions and structural breaks, 307–8
disaggregation over variables, 273, 289–96, 309
econometric models v. robust forecasting models, 279–88
forecast combination and breaks, 273, 301–7
forecast error taxonomies, 277–79
forecast performance assessment, 273, 274–77
time disaggregation, 273, 296–301, 309
forecasting models, for forecasting breaks. See also forecast combinations; forecasting breaks
approximating logistic transition function, 330–31
automatic model selection, 72, 332, 348
conventional, 328–29
forecast combinations v., 355, 386
for forecasting during breaks, 341–42, 342t
formulation of, 327–31
IIS and, 333–34
modeling nonlinearity, 334–35
nonlinear functions, 329
reduction to theory-based form, 335–36
selecting, 320, 332–36
testing for nonlinearity, 331
threshold models, 329–30, 337
forecast origin
improved data at, 327
mismeasurement, 278, 285, 292t, 293t, 295, 295t, 307
forecast performance assessment. See also forecast combinations; forecasting in presence of unanticipated location shifts
general purpose loss functions for, 274–75
GFESM measure and, 276–77
model instability and, 3, 355–56, 357
MSFE matrix and, 275–77
NAWM, 107
relative accuracy in, 277
RMSFEs and, 277
VECMs, 17, 23, 24t, 26, 27t, 28, 29t
forecast period location shifts, 278, 280
forecast uncertainty
density forecasts and, 113
multidimensional panel data of survey forecasts and, 481–85
nowcast model and, 209–12, 210f, 211t
repeated observation forecasting and, 255–56, 256f
foreign exchange market
ACD model and, 518–19
cointegration and, 521
dollar/sterling exchange rate changes, 502, 504f, 506, 506t, 509–13, 510f
fractional cointegration, 521
(p. 699) fractionally integrated EGARCH (FIEGARCH) model, 511f, 512f, 515
fractionally integrated GARCH (FIEGARCH) model, 510f, 515
fractional marginal likelihood criterion, 16
“framing” effects, 172
FRED (Federal Reserve Economic Data), 25, 25n4
frost-fruit case study, 572
fruit-frost case study, 572
FWER (family-wise error rate). See also RC procedure; SPA test
Bonferroni's one-step procedure, 392, 393, 400, 404
controlling for, 392–93, 396–400, 404
defined, 392
GARCH models and, 393
Holm's step-down method, 392, 393, 399, 400, 404
G7 countries data revisions (study), 254
Galton, F., 465
GARCH (generalized autoregressive conditional heteroskedastic) models
ACD model v., 518–19
ARFIMA models v., 540–41
controlling for FWER, 393
discrete-time, 543
EGARCH, 508, 509, 510, 511, 512, 515, 541
FIEGARCH, 511f, 512f, 515
FIGARCH, 510f, 515
GARCH(1,1), 393, 411, 507–8, 544, 545
GARCH-in-mean, 519
GARCHX, 544–46, 547
HYBRID GARCH structure, 549
IGARCH, 508–9, 509, 510, 515
parallel GARCH structure, 547
Realized GARCH model, 547–49
volatility forecasting and, 507–13, 543–49
gas market. See also electricity spot markets
carbon emissions markets and, 609
oil pricing and, 608
pricing, 608
UK spot gas prices (time series), 610f
VARs and, 608
Gaussian MLEs. See maximum likelihood estimations
Gaussian state-space models. See state-space models
GDP. See long-horizon economic growth forecasting; nowcasting model
generalized autoregressive conditional heteroskedastic models. See GARCH models
generalized Bass model, 682
generalized gamma model (GGM), 634, 635, 642, 643
generalized linear models. See GLMs
generalized method of moments (GMM), 50, 485, 486, 492, 633
generalized principal components estimators, 41–42, 44
general polynomial approximation [Ch11, Section 6.3]
general purpose loss functions, 274–75. See also forecast performance assessment
general unrestricted model(GUM), 335–36
generational accounting, 585
GFESM (generalized forecast error second moment) measure, 20, 21, 22t, 23, 276
forecast performance assessment and, 276–77
GFESM1/h, 24t, 26, 27t, 29t
GFESMh, 20, 21
GGM (generalized gamma model), 634, 635, 642, 643
GLMs (generalized linear models for health care costs), 5, 627, 636–39, 642, 642t, 645, 646–47, 648, 648t, 649, 649t
basis approach, 636–39
EEE approach, 632, 639, 642, 642t, 643, 645, 646, 647, 648t, 649, 649t
global panel model, 601f
global vector autoregressive (GVAR) models, 305–6
GMM (generalized method of moments), 50, 485, 486, 492, 633
Goldfarb. See BFGS method
goodness of fit measures, health care costs case study, 647, 649t
Google Trends data, 319, 320, 326, 348
Granger's representation theorem, 10, 13, 17, 520
“Great Moderation,” 367
Greenbooks, 195, 478
group bias, 173t, 181
GUM (generalunr estricted model), 335–36
GVAR (global vector autoregressive) models, 305–6
HAC (heteroskedasticity and autocorrelation consistent) estimator, 421, 422, 426, 445, 448, 451
Hannan-Quinn criterion. See HQ criterion
Hansen's SPA test, 398–99, 404, 411, 541. See also SPA class of tests
HARCH (heterogeneous ARCH) model, 541
HAR (heterogeneous autoregressive) models, 528, 541–42
Harmonized Index of Consumer Prices. See HICP
HAR-RV (heterogeneous autoregressive model of realized variance), 239–40, 541–42
hazard models, 627, 633, 634–36, 640, 641
Health and Retirement Study (HRS), 464
health care costs models, 5, 625–54
administrative data sets for, 627
binary models, 626
Box-Cox models, 631–32, 639, 642t, 646, 647
challenges for, 625
comparing model performance, 627, 641–44
(p. 700) Copas test, 628–29, 629n2, 643, 647
cost effectiveness analysis and, 626
count data models, 626
discrete conditional density estimator, 626, 640–41, 645n16
ECM models, 632–33, 634, 636, 639, 645, 646, 648t, 649t
EEE approach, 632, 639, 642, 642t, 643, 645, 646, 647, 648t, 649, 649t
finite mixture models, 626, 639–40, 644, 647, 649
GLMs, 5, 627, 636–39, 642, 642t, 645, 646–47, 648, 648t, 649, 649t
hazard models, 627, 633, 634–36, 640, 641
Hosmer-Lemeshow test, 628, 647
linear regression models, 626, 627–32
log transformations, 629–31
MEPS and, 627, 642–43, 644, 650
negative binomial model, 633
nonlinear regression models, 626, 627, 628–29, 632–36
Park test, 638, 639, 646
Poisson regression, 633–34, 643, 646, 649
Pregibon's link test, 628, 638, 646, 647, 648t
RESET test, 628, 629
risk adjustment and, 626
semiparametric transformation models, 632
square root transformations, 631
two-part models, 625–26, 629, 640
usage areas for, 626
health care costs models: empirical application, 644–49
data set for, 644
goodness of fit measures, 647, 649t
MEPS and, 627, 644
specification tests, 647, 647t
HEAVY model, 547
herding behavior, 167, 172, 181
Herfindahl–Hirschman index (HHI), 619
heterogeneous ARCH (HARCH) model, 541
heterogeneous autoregressive model of realized variance. See HAR-RV
heterogeneous autoregressive models. See HAR models
heterogeneous survey forecasts. See survey forecasts
heteroskedasticity and autocorrelation consistent (HAC) estimator, 421, 422, 426, 445, 448, 451
heuristics, 168–70
HHI (Herfindahl–Hirschman index), 619
HICP (Harmonized Index of Consumer Prices), 204, 206t, 207, 208f, 209, 210f, 211t, 212, 218t
“hidden layer,” 70. See also neuralnetworks
hidden Markov regression models. See Markov switching regression models
hierarchical DFMs, 54
high-frequency data. See also disaggregation by time; disaggregation over variables; volatility forecasting
break detection and, 326
dynamic properties of volatility and, 526–27
Google Trends data, 319, 320, 326, 348
SV modele stimation and, 528
volatility forecast evaluation and, 527
volatility forecasting impacted by, 525–26
high-frequency predictors
DL-MIDAS and, 230–31
nowcasting and, 203
high-frequency price changes
ACD model and, 518–19
GARCH-in-mean model and, 519
hindsight bias, 171, 173t
Holm's step-down method, 392, 393, 399, 400, 404
Hooker, R., 471
Hosmer-Lemeshow test, 628, 647
households class, NAWM, 92, 93, 94, 119
house price changes, 316, 316f
HQ (Hannan-Quinn) criterion, 16, 23, 24t, 25, 26, 27t, 29t
HRS (Health and Retirement Study), 464
hurricane forecasts study, 575–79, 576f, 578t. See also weather and climate forecasts
HYBRID GARCH structure, 549
hybrid principal components/state-space methods, 37, 42–44
IEEE Monthly Business Survey, 467
IGARCH (integrated GARCH) models, 508–9, 510, 515
IIS (impulse-indicator saturation), 278, 319n1, 320, 326, 327, 331, 332, 333–34, 335
“I knew it all along effect,” 171
illusion of control, 171
implicit expectations framework, 475–77, 492
impulse-indicator saturation. See IIS
IMSE, 365t, 366t
Indian Ocean tsunami, 316–17
inequality, Jensen's, 458, 465, 485, 565
information (for predicting breaks), 317. See also information sets
conditions for forecasting breaks and, 316–18
disaggregated data over time and variables, 326
formal description of, 323–24
improved data at forecast origin, 327
leading indicators, 325–26
new sources for, 319
prediction markets, 326–27
separation into sets, 320
survey data, 326
unpredictability and, 321–24, 348
information criteria. See also AIC; BIC; HQ criterion
estimation of static factors with, 46–47
VARs and, 15–16
(p. 701) information sets, 320, 321, 348
information use, bias and, 170, 172
innovation error, LDGP, 279, 280, 285, 287, 292t, 293t, 295t, 308
insanity filter, 79, 82, 83, 364n6
instrumental variables, dynamic factors as, 50
INTC (Intel), ACFs for, 534, 534f
integrated GARCH. See IGARCH models
interaction model, long-horizon growth forecasting and, 600, 601
interest
term structure of interest rates, 10, 521
variable of interest forecasting, 130
intergenerational transfer system, 592, 593, 594, 599
interpreting and combining survey forecasts. See survey forecasts
irregular component (regression component), 129
in electricity spot prices illustration, 154, 155f, 157f
regression coefficients, 143–44
specification, 138
irrelevant models, 393, 398, 402, 409, 411. See also multiple forecast model evaluation
ISIC, 364, 365t, 366t
iterated forecasts, direct v., 48–49, 226
iterated logarithm, law of, 398, 404, 410
jagged edge (ragged edge) problem, 194, 194n1, 196, 198, 202, 204, 205
Jensen's inequality for convex functions, 458, 465, 485, 565
JT information set, 319, 320, 321, 324–25
judgment (in economic forecasting), 2, 163–89. See also biases
analogy usage and, 173t, 179–80
“black arts” and, 174
combination (statistical/judgmental) forecasting, 173t, 178–79
dangers of, 164, 167–72
decomposition of, 173t, 180–81
Delphi method and, 173t, 181–82
feedback and, 173t, 174–76
final forecasts based on, 166–67
in formulation of models, 164–65
news revisions v. noise revisions issue and, 261
nowcasting and, 194, 195
prediction markets and, 182
psychological bootstrapping method and, 173t, 177–78
in revising components of models, 166
sales forecasting and, 5, 673
“twilight world” and, 174
Kalman filter. See also maximum likelihood estimations
MIDAS regressions v., 236–38
missing data and, 43, 44, 146, 202, 225–26, 235
nowcasting and, 198, 201, 202, 204n8, 209, 223, 225, 235–38
recursive lemma and, 141–42
state-space models and, 130, 141–42
KG polynomials. See Kolmogorov-Gabor polynomials
K hypotheses testing, 392, 404, 404t
Kolmogorov-Gabor (KG) polynomials, 69–70, 73, 81
KT information set, 319, 320, 321, 324–25
large real-time databases, 252, 264, 361
large-scale forecast comparisons, 78–81
law of the iterated logarithm, 398, 404, 410
LCD television sets, Bass model, 678–80, 679f, 680t
LDGPs (local DGPs). See also DGPs defined, 272
innovation error, 279, 280, 285, 287, 292t, 293t, 295t, 308
nonconstancy,VEqCM/DDD and, 284–87
leading indicators, 325–26. See also information
leads, MIDAS with, 235–36
learning functions, nonlinear, 329, 337
Lehman Brothers, 318, 329
lemma, Kalman filter and, 141–42
likelihood evaluation. See maximum likelihood estimations
Lindley's statistical paradox, 113
linear ARI model, 77, 78f
linear Gaussian state-space models. See state-space models
linear-quadratic models, 122
linear regression health care costs models, 626, 627–32. See also transformed health care costs models
cost regression models, 627–29
tests for, 628–29
transformed costs models, 629–32
linear time series models. See also forecast combinations illustration: survey and time series forecasts; nonlinear time series models
nonlinear v., 61, 76–77
state-space form and, 138
linear time series/nonlinear time series forecasts comparison, 61, 76–77
ARI v. LSTAR model, 77, 78f
forecasting with same model for each forecast horizon, 80–81
forecasting with separate model for each forecast horizon, 78–80
large-scale forecast comparisons, 78–81
link test, Pregibon's, 628, 638, 646, 647, 648t
Litterman data set, 25–26, 27t, 28, 29t
Livingston Survey, 167, 467, 473, 481, 492
(p. 702) LMSV (long-memory stochastic volatility) model, 540
local DGPs. See LDGPs
local linear trend component, 132
location shifts. See also forecasting during breaks; forecasting in presence of unanticipated location shifts; structural breaks
data revisions and, 307–8
data vintages and, 307–8
DGPs and, 280–81
examples of, 316, 316f
financial crisis of 2007–2010, 316, 318, 336
forecast error taxonomy and, 278
forecast failure and, 17, 166, 273, 278, 279, 280, 282, 283
house price changes, 316, 316f
improving robustness to, 283
subprime loans, 316, 316f
UK mortgage lending, 316, 316f
unmodeled, 291, 304, 315
VEqCMs impacted by, 282–83
world liquidity changes, 316, 316f
logistic STAR. See LSTAR models
logistic STR. See LSTR models
logistic transition function, 64, 330–31
log-linearized DSGEs. See DSGEs, log-linearized
log transformations, cost regression health care models, 629–31
long-horizon economic growth forecasting (demographically based), 5, 585–605. See also demography
advantages of, 586
backcasts, 600–601, 602
benefits of, 602
cross-country panels, 599–600, 602
demographic-economic connections, approaches/methods for, 587, 592–602
economic growth theory and empirics, 587–90
GDP development and, 585–602
global panel model, 601f
interaction model forecast, 600, 601
intergenerational transfer system, 592, 593, 594, 599
Sweden economic growth studies, 597–98, 597f, 600–601, 601f
time series approaches, 592–99, 602
uncertainty measures of, 603
long-horizon financial time series forecasting, 517–19
long-memory models, 513–15
HAR models and, 542
UC-RV model and, 541
long-memory stochastic volatility (LMSV) model, 540
loss functions, 274–75. See also forecast performance assessment
LSTAR (logistic STAR) models. See also forecast combinations illustration: survey and time series forecasts
ARI model v., 77, 78f
forecasting regime shifts and, 337–38, 339
neuralnetw orks and, 515–17
LSTR (logistic STR) models, 64, 335–36, 337, 360, 366, 515, 619, 620
“lucky” model, 392
macroeconomic forecasting. See forecasting
Malthus, T., 589, 590
Maori/Captain Cook example, 321
MAPE (mean absolute percentage errors), 157, 158, 643, 647, 649, 649t
marginal cost supply function, for power-generating facilities in Germany, 612, 612f, 614
market microstructure effects, 531, 538, 613
Markov chain, first-order, 65, 356n1, 570
Markov chain Monte Carlo (MCMC) methods, 44, 45, 52, 90, 509, 521
Markov switching (MS) regression models, 45, 61, 65–66, 383
ARFIMA and, 543
GNP and, 77
MS-VAR models, 66, 320, 337
as nonlinear time series models, 65–66
Markov switching (MS)-VAR models, 66, 320, 337
martingale assumption, 505, 506, 513, 519, 530, 531
Matlab
Toolbox for MIDAS regressions, 241
YADA program, 96
maximum likelihood estimations (MLEs)
DSGE estimation, 52
dynamic factors estimation, 37, 38–39, 43, 45
nowcasting model, 201–2
state-space models, 142–43
MCS (model confidence set) approach, 393, 407–10
ME (mean error), 342f, 343f, 344, 345f. See also forecasting during breaks
mean absolute percentage errors. See MAPE
mean error. See ME
mean forecasts, convex loss functions and, 465–66
mean prediction error. See MPE
mean squared error. See MSE
median forecasts, unimodal loss functions and, 466
Medical Expenditure Panel Study (MEPS), 627, 642–43, 644, 650
Meese-Rogoff puzzle, 446
MEM (multiplicative error model), 547
MEPS (Medical Expenditure Panel Study), 627, 642–43, 644, 650
Michigan Monthly Survey, 464
Microeconometrics Using Stata data set, 644
MIDAS (mixed-data sampling) regressions, 227–35, 241
ADL-MIDAS, 231–35
(p. 703) ARCH and, 239
direct/iterated approaches v., 226
DL-MIDAS, 230–31, 232, 235
FADL-MIDAS, 235, 236
HAR-RV and, 239–40
HYBRID GARCH structure and, 549
Kalman filter v., 236–38
with leads, 235–36
Matlab Toolbox for, 241
MIDAS-NIC, 240–41
MIDAS-RV, 240
nowcasting and, 203, 235–38
volatility forecasting and, 227, 238–41
minimum mean squared linear estimator (MMSLE), 141, 147
Minnesota prior, 103–4, 104n2, 106
missing data. See also EM algorithm; Kalman filter
EM algorithm and, 203–4, 221
Kalman filter and, 43, 44, 146, 202, 225–26, 235
misspecification
for DSGEs, 124
forecast combinations and, 386
of in-sample deterministic factors, 278
for NAWM, 124
of stochastic components, 278
misspecified models. See also forecast error taxonomies; forecasting in presence of unanticipated location shifts
forecast error taxonomies and, 277
structural breaks and, 302–5
mixed-data sampling. See MIDAS regressions
mixed-frequency data, 225–45. See also MIDAS regressions
computer technology innovations and, 225
forecasting with, 225–27
MLEs. See maximum likelihood estimations
MMSLE (minimum mean squared linear estimator), 141, 147
model-based volatility forecasting, 529, 543–50
realized measures and, 528
reduced-form volatility forecasting v., 527–28
model confidence set (MCS) approach, 393, 407–10
model instability, 3, 357–58, 375–82. See also structural breaks
forecast combinations impacted by, 357
forecast performance and, 3, 355–56, 357
model instability illustration-DFMs combination, 357–58, 375–82, 386–87
equal-weighted forecast v. best model forecast, 381–82, 381f, 386–87
MSFE values, 380–81, 380f
models. See also DSGEs; dynamic factor models; forecasting models; health care costs models; judgment; nonlinear time series models; unobserved components time series models; VARs; specific models
advantages of, 163–64
data issues and, 1
forecasting v., 287
judgment's role in, 164–66
limitations, 163–64
“lucky,” 392
structural breaks' impact on, 315, 346–47
susceptibility of, to unanticipated structural breaks, 272
model selection [Ch11, section 6] also see Autometrics
model universes. See also forecast combinations
design of, 356, 358–61
restricted, 361
modified Hosmer-Lemeshow test, 628, 647
monetary authority, NAWM, 92–93, 94, 95, 99
Monte Carlo analysis. See also forecast combinations experiment: in presence of breaks; Markov chain Monte Carlo methods
forecasting during breaks, 341–45
health care cost models comparison, 643
reduced-rank VECMs, 19–25, 30
unconditional predictive ability experiments, 429–38
Monthly Business Survey, IEEE, 467
monthly factor model, 199–200. See also nowcasting model
mortgage lending, UK, 316, 316f
motivational bias, 167–68, 173t
MPE (mean prediction error), 643, 644, 647, 648, 649
MS. See Markov switching regression models
MSE (mean squared error)
correcting for biases and, 176–77
log determinant statistics of scaled MSE matrices (NAWM), 110–12, 111f, 112f
MMSLE and, 141
MSFE (mean squared forecast error) matrix
det(MSFE), 22t, 24t, 26, 27t, 29t
det (MSFEh), 20, 21
forecast combinations experiment and, 384–86, 385t
forecasting during breaks and, 339, 346, 346f
forecast performance assessment and, 275–77
limitations of, 275–77
model instability illustration and, 380–81, 380f
trace(MSFE), 22t, 24t, 26, 27t, 28, 29t
trace(MSFEh), 20, 21
MS (Markov switching)-VAR models, 66, 320, 337
multidimensional panel data of survey forecasts, 4, 473–95
adaptive expectations model, 474–78, 492
anticipated changes measures, 478–81
BCEI and, 473, 480n3, 487, 488f, 489, 490, 492
consensus forecasts and, 469, 474, 484, 485
cross-sectional shocks, 478–80, 479f, 488
cumulative shocks, 478–80, 479f, 487, 488
discrete shocks, 478–81, 479f
(p. 704) forecast uncertainty measurement, 481–85
implicit expectations framework, 475–77, 492
Livingston Survey and, 473, 481, 492
rational expectations hypothesis, 473, 474–78, 485, 492–93
rationality tests, 485–92
SPF and, 473, 480n3, 482, 484, 485, 487, 487f, 488–90, 492
volatility measures, 478–81
multinominal health care costs models, 626
multiple cointegrating vectors, 521
multiple cycles component, 137
multiple forecast model evaluation, 4, 391–413
Bonferroni's one-step procedure, 392, 393, 400, 404
controlling for FDR, 393, 404–7
controlling for FWER, 392–93, 396–400, 404
empiricale vidence, 410–11
Hansen's SPA test, 398–99, 404, 411, 541
Holm's step-down method, 392, 393, 399, 400, 404
irrelevant models, 393, 398, 402, 409, 411
MCS approach, 393, 407–10
RC procedure, 392–93, 398, 399–400
SPA class of tests, 393, 400–404
stepM approach, 399–400
type 1 error and, 392, 393, 406, 410, 411
multiplicative error model (MEM), 547
multivariate factor-based models (AR_FAC), 361. See also forecast combinations illustration: survey and time series forecasts
multivariate regression models, 33, 34, 519
multivariate time series models, 148–49. See also unobserved components time series models
challenge of, 148–49
with common trends and cycles, 130, 150–52
dynamic factor analysis and, 130, 152–53
estimation methodology, 152
forecasting methodology, 152
multivariate trend model, 149–50
state-space representation, 152
multivariate trend model, 149–50
Muth's unbiasedness condition, 474–75
National Institute of Statistics and Economic Studies (INSEE) Monthly Business Survey, 467
National Longitudinal Survey of Youth, 464
NAWM (New Area-Wide Model), 91–101
bird's eye view of, 92–93
cross-equation restrictions, 93, 117, 119
density forecasts, 113–17, 116f, 117f
empirical implementation of, 96–101
estimating predictive distribution of DSGE, 101–3, 123
firms class, 92, 95, 99
fiscal authority, 92
forecast accuracy evaluation, 107
households class, 92, 93, 94, 119
key equations in, 93–96
log predictive scores (forecasting changes of variables), 116–17, 116f, 117f
misspecification matters for, 124
monetary authority, 92–93, 94, 95, 99
performance/structure relationship, 117–22
point forecasts, 108–12, 108f, 111f, 112f
quarterly nominal wage growth and nominal interest rate forecast paths, 119, 120f121f
RMSEs, 107, 108, 108f, 109, 111, 112, 118, 122
steady-state values of variables, 118–19, 118t
structural parameters, posterior mode estimates of, 100f, 101
structural shocks, 99, 101, 102, 103, 104n1
SVAR and, 93, 97, 99, 109
variables, 96–99, 98f, 108–9, 108f
NAWM benchmarks (reduced-form models), 103–7, 123
density forecasts, 113–17, 116f, 117f
forecast accuracy evaluation, 107
large BVAR, mixed prior, 91, 104, 106, 107
large BVAR, white-noise prior, 91, 106, 107
log predictive scores (forecasting changes of variables), 116–17, 116f, 117f
mean, 91, 107
point forecasts, 108–12, 108f, 111f, 112f
quarterly nominal wage growth and nominal interest rate forecast paths, 119, 120f121f
small BVAR, 91, 105
small VAR, 91, 106–7
nearest neighbor forecast method, 76
negative binomial model, 633
neural networks. See also forecast combinations illustration: survey and time series forecasts
autoregressive, 79
forecasting with, 330
regime switching models and, 515–17
RMSFE, 72
‘single hidden-layer feedforward’ model, 70
as universal approximators, 70–72
New Area-Wide Model. See NAWM
news
news revisions v. noise revisions, 260–61
nowcasting model and, 194, 197–98, 203, 207–9, 213
news impact curve (NIC), 240, 241
NIC. See news impact curve
NLS (nonlinear least squares), 229, 516, 633, 646, 675
noise revisions, news revisions v., 260–61
nonconstancies
parameter, 271, 282, 289
VEqCM/DDD and, 284–87
nonlinear DFMs, 45–46
nonlinear functions, 329
(p. 705) nonlinearity. See also forecasting models; volatility forecasting
in reduced-form volatility forecasting, 543
testing for, 331
nonlinear learning functions, 329, 337
nonlinear least squares (NLS), 229, 516, 633, 646, 675
nonlinear/non-Gaussian state-space models, 45–46
nonlinear regression health care costs models, 626, 627, 632–36
ECM models, 632–33, 634, 636, 639, 645, 646, 648t, 649t
hazard models, 627, 633, 634–36, 640, 641
Poisson regression, 633–34, 643, 646, 649
tests for, 628–29
nonlinear threshold autoregressive model, 77
nonlinear time series forecasting, 72–75
analytical point forecasts, 72–73
direct point forecasts, 75, 81–83
recursive point forecasts, 73–74, 81–83
nonlinear time series/linear time series forecasts comparison, 61, 76–77
ARI v. LSTAR model, 77, 78f
forecasting with same model for each forecast horizon, 80–81
forecasting with separate model for each forecast horizon, 78–80
large-scale forecast comparisons, 78–81
nonlinear time series models, 61–87. See also forecast combinations illustration: survey and time series forecasts; Markov switching regression models; neural networks; STAR models; STR models; switching regression models; TAR models
bilinear models, 67, 73
DFMs v., 2
forecast densities and, 84
linear v., 61, 76–77
random coefficient models, 67–68
universal approximators and, 61, 68–72, 73, 81
VARs v., 2
nonmarket valuation, of weather and climate forecasts, 560, 573–79
nonparametric averaging methods, 37, 39–42
nonscientific election forecasting, 657, 658. See also election forecasting
Nordhaus's efficiency condition, 474–75
Nord Poolel ectricity market. See electricity daily spot prices
nowcasting, 193–223. See also nowcasting model
alternative modeling strategies, 194, 203
automatic statistical procedure v., 195
bridge equations strategy and, 194, 203
coincident indicators of economic activity v., 195, 204, 213
DFMs and, 55, 195, 198–200, 202, 203, 205, 209, 212, 213
estimation approach, 199, 203–4
forecast failure, 163
future research on, 212–13
high-frequency predictors and, 203
importance, in forecasting literature, 195
jagged edge problem, 194, 194n1, 196, 198, 202, 204, 205
judgmental process and, 194, 195
Kalman filter and, 198, 201, 202, 204n8, 209, 223, 225, 235–38
MIDAS and, 203, 235–38
new developments, 212–13
related literature and, 202–4
single equation approaches, 203
structural breaks and, 327
surveys and, 194
VARs and, 212
nowcasting model (euro-area GDP), 213
benchmark model, 205
data set, 205–6, 206t, 217t219t
econometric framework, 198–202, 213
EM algorithm, 201–2, 221–23
empirical results, 204–12
estimation approach, 199, 203–4
forecast uncertainty in, 209–12, 210f, 211t
forecast updates, 207–9
HICP, 204, 206t, 207, 208f, 209, 210f, 211t, 212, 218t
key feature, 202–3
limitations, 212–13
maximum likelihood estimation, 201–2
methodology, 194–95, 198–99
monthly factor model, 199–200
news and, 194, 197–98, 203, 207–9, 213
quarterly variables model, 200–201
reasons for GDP emphasis, 194
RMSFE in, 209, 210, 212
state-space representation, 201, 220–21
null hypothesis. See also FDR; finite-sample predictive ability tests; population-level predictive ability tests
alternative hypothesis and, 391
data snooping and, 391–92
defined, 411
false discovery and, 392
forms of, 417
objective probability of personal outcomes, point predictions and, 460–61
observable factor models, 2, 11, 14, 18, 30
OECD (Organization for Economic Cooperation and Development), 318, 588, 607
ogive, 329, 332, 348
oil market. See also electricity spot markets
carbon emissions markets and, 609
cointegration analyses and, 608
as economic sentiment driver, 607, 620
energy commodities bundle and, 607–8
models/conceptual frameworks for, 607–8
pricing, VARs and, 608
(p. 706) UK spot oil prices (time series), 610f
volatility of, 620
OLG (overlapping generations) model, 589
oligopolistic nature, of electricity spot markets, 614, 617, 618–19
one-step procedure, Bonferroni's, 392, 393, 400, 404
one-time reversal test, 446f, 447, 448–50, 454
OPEC (Organization of Petroleum Exporting Countries), 607
optimism bias, 171–72, 173t
Organization for Economic Cooperation and Development (OECD), 318, 588, 607
Organization of Petroleum Exporting Countries (OPEC), 607
Ornstein–Uhlenbeck process, 526
outcome feedback, 174–75
overconfidence bias, 170–71, 173t
overlapping generations (OLG) model, 589
overparameterization problem, 9–10
ox weighing, 465
panel-data models, 337
panel-data sets, 467, 473. See also multidimensional panel data of survey forecasts; SPF; survey forecasts
Bank of England's Survey of External Forecasters, 467
Blue Chip Economic Indicators, 473, 480n3, 487, 488f, 489, 490, 492
Center for Economic Studies ifo Institute World Economic Survey, 467
Livingston Survey, 167, 467, 473, 481, 492
National Institute of Statistics and Economic Studies (INSEE) Monthly Business Survey, 467
panels of forecasters, temporal variation of forecasts by, 458, 467–69
paradox, Lindley's, 113
parallel GARCH structure, 547
parameter estimation uncertainty, 277, 278, 285
parameter nonconstancies, 271, 282, 289
Park test, 638, 639, 646
partial canonical correlation analysis (PCCA), 10, 13, 14, 32–34
PCCA. See partial canonical correlation analysis
performance assessment. See forecast performance assessment
performance feedback, 175
Philadelphia Fed, 204, 252, 264, 361, 417n2, 467, 468
PIC. See posterior information criterion
planting/fallowing case study, 572
point predictions. See also survey forecasts
of binary outcomes, 459–62
cross-forecaster dispersion, interpreting, 462–63
of fertility intentions, 461–62
objective probability of personal outcomes and, 460–61
of real-valued events, 462–63
subjective probabilities and, 459–60
Poisson distributions, 159, 633, 638, 645, 646
Poisson process, 518, 615
Poisson QML, 195, 202, 633, 638, 639, 639n7
Poisson regression, 633–34, 643, 646, 649
policy
differenced VEqCM and, 287
DSGEs and, 2
DSGE-VARs and, 124
forecasting v., 287
health care costs models and, 626
models v., 287
quantitative models and, 164
stochastic breaks and, 278
structural breaks' impact on, 315, 346–47
political betting markets, 658, 659
political forecasting, 655–56, 668. See also election forecasting
polls, election forecasting, 657, 658–59
pooling. See forecast combinations
population-level predictive ability tests, 419–23, 438–39
populations, stable, 589–90
portfolio theory, forecast combinations v., 304–5
posterior information criterion (PIC), 16
power markets. See electricity spot markets
predictability (of breaks), 317. See also information; unpredictability
prediction markets, 182
Delphi method v., 182
forecasting breaks and, 326–27
predictive ability testing. See conditional predictive ability tests; unconditional predictive ability tests
Pregibon's link test, 628, 638, 646, 647, 648t
prequential approach, 113
presidential election forecasting models, 666–68, 668t
principal components estimators, 40–41
dynamic, 42
generalized principal components estimators v., 41–42, 44
hybrid principal components/state-space methods, 37, 42–44
probability forecasting. See also multidimensional panel data of survey forecasts; survey forecasts; weather and climate forecasts
election forecasting and, 657
learning through feedback and, 174
prediction markets and, 327
in survey forecasts, 463–64
weather/climate forecasting and, 562–64
prototypical decision-making models, 567–70. See also weather and climate forecasts
(p. 707) pseudo out-of-sample forecasting exercises
defined, 416
DFMs and, 49
forecast combinations illustration and, 364
unconditional predictive ability tests and, 416, 417–19, 438
vintage data comparisons and, 307
psychological biases, 168, 170
psychological bootstrapping method, 173t, 177–78
“publicity hypothesis,” 167
puzzle
forecast combination puzzle, 356
Meese-Rogoff puzzle, 446
p-value, 406
QML (quasi-maximum likelihood), 195, 202, 633, 638, 639, 639n7
quantitative models. See models
quarterly variables model, 200–201. See also nowcasting model
quasi-maximum likelihood (QML), 195, 202, 633, 638, 639, 639n7
QuickNet algorithm, 72, 82, 82t, 83, 83t
ragged edge. See jagged edge problem
random coefficient models, 67–68
random walk benchmark. See NAWM benchmarks
random walk hypothesis, 501–6
random walk plus smooth cycle model, 140–41
random walk seasonal component, 135–36
rational expectations hypothesis, 122, 460, 461, 462, 473, 474–78, 485, 492–93
rationality tests, 485–92. See also multidimensional panel data of survey forecasts
“raw” estimates, 19, 19n2
RC (reality check) procedure. See also SPA test
defined, 392–93
SPA test v., 398
stepM approach v., 399–400
real business cycle theory, 10, 17, 122
reality check. See RC procedure
Realized GARCH model, 547–49
realized measures
benefits of, 526
defined, 525
model-based volatility forecasting and, 528
reduced-form volatility forecasting and, 527–28
of volatility, 531–32
in volatility forecasting, 527–28, 535–37
“realized volatility,” 507
real-time data. See also data revisions
data revisions v., 257–59
large real-time databases, 252, 264, 361
real-time data analysis bibliography (online), 254
real-time datasets (online), 252
structure, 248–49, 248t
unconditional predictive ability tests and, 428–29
Real-Time Data Set for Macroeconomists, 248t, 251t, 252
real-time data vintages, 247–67. See also data revisions
recursive forecasting scheme, 437, 442, 443, 444
recursive lemma, Kalman filter and, 141–42
recursive point forecasts, 73–74, 81–83
recursive regression strategies, 520
reduced-form models, 90
reduced-form volatility forecasting, 527–29, 537–43, 549–50
distributed lag models, 537–40
HAR models, 528, 541–42
model-based volatility forecasting v., 527–28
nonlinearity in, 543
realized measures and, 527–28
separating continuous and jump components, 542–43
reduced-rank error-correction models, of Litterman data, 25–26, 27t, 28, 29t
reduced-rank structures, 14, 15
reduced-rank VECMs
Monte Carlo analysis, 19–25, 30
shrinkage and, 18–19
regime shifts, 336–39, 348. See also forecasting breaks
regime switching models, 330, 515–17
forecasting regime switch with, 337
two-regime, 64, 330
regression coefficients, 143–44
regression component. See irregular component
Reifschneider and Tulip measure, 484, 485
relative accuracy, in forecast performance assessment, 277
relative mean absolute revision (RMAR), 253
repeated observation forecasting, 255–56, 256f
representation theorem, Granger's, 10, 13, 17, 520
representativeness heuristic, 168–69
“reputational cheap talk” hypothesis, 167
RESET test, 628, 629
residential in vestment. See data revisions example
restricted vector error-correction models (RVECMs), 24t, 25, 26, 27t, 28, 29t
revealed preference methods, 573. See also stated preference methods
revisions. See data revisions
risk, forecast combinations and, 386
risk adjustment, health care costs models and, 626
RMAR (relative mean absolute revision), 253
RMSEs. See root mean squared errors
RMSFEs. See root mean squared forecast errors
robust forecasting models, econometric models v., 279–88
(p. 708) rolling window estimates, 18, 26, 28, 29t, 305, 394n2, 395, 424, 447. See also expanding window estimates
root mean squared errors (RMSEs), NAWM and, 107, 108, 108f, 109, 111, 112, 118, 122
root mean squared forecast errors (RMSFEs)
forecast during breaks and, 340f, 341, 342f, 343f, 344, 345, 345f
forecast performance assessment and, 277
neural networks and, 72
nowcasting model and, 209, 210, 212
roulette wheel, 501
RVECMs (restricted vector error-correction models), 24t, 25, 26, 27t, 28, 29t
sales forecasting, 5, 673–89
Bass model, 674–82
durable products, 674–82, 687
expert forecasts, 673, 674, 684–87
judgment and, 5, 673
nearest neighbor method and, 76
SKU-level sales, 674, 683–87
sampling the future method, 90. See also NAWM
algorithm, 90–91, 102–3
alternative forecasting models, 103–7
estimating predictive distribution of log-linearized DSGE model, 101–3
scientific election forecasting. See election forecasting
scree plots, 46, 47
seasonal component, 129
in electricity spot prices illustration, 154, 155f, 157f
fixed dummy, 133–34
fixed trigonometric, 134–35
random walk, 135–36
specifications, 133–36
time-varying dummy, 134
time-varying trigonometric, 135
second-stage regressions, dynamic factors and, 48–50
SEE (Survey of Economic Expectations), 464
SEE (standard error of estimates), 655, 661, 664, 665, 666
selective recall of information, 170, 172
self-exciting threshold autoregressive models. See TAR models
semimartingales, 530, 531
semiparametric approaches. See discrete conditional density estimator; finite mixture health care costs models
semiparametric transformation models, 632
sequential testing bias problem, 392
Shannon. See BFGS method
shocks. See also energy commodities; multidimensional panel data of survey forecasts
cross-sectional, 478–80, 488
cumulative shocks, 478–80, 487, 488
discrete, 478–81
FAVARs and structural shocks, 50, 51
NAWM structural shocks, 99, 101, 102, 103, 104n1
supply shocks, 607, 610, 613, 616, 620
shrinkage
Minnesota prior and, 103–4, 104n2, 106
reduced-rank VECMs and, 18–19
single-equation approaches, 203, 263
ʼsingle hidden-layer feedforward’ model, 70
“single source of error” models, 12
SKU-level sales, 674, 683–87. See also sales forecasting
expert forecasts, 684–87
models, 683–84
SM-AR (switching-mean autoregressive) model, 71
smooth cycle component
in electricity spot prices illustration, 154, 155f, 157f
plus random walk, 140–41
time-varying trigonometric, 137
smooth transition autoregressive models. See STAR models
smooth transition regression models. See STR models
smooth trend component, 132–33
in electricity spot prices illustration, 153, 154, 155f, 157f
plus ARMA model, 139–40
S&P 500 stock market index returns, 499, 502, 502f, 506, 506t, 509–13, 511f
SPA class of tests, 393, 400–404
SPA (superior predictive ability) test, 398–99, 404, 411, 541
special interest areas, 5. See also election forecasting; energy commodities; health care costs models; long-horizon economic growth forecasting; sales forecasting; weather and climate forecasts
specification tests, health care costs case study, 647, 647t
specific-to-generalt echnique, 71, 72. See also Autometrics approach; QuickNet algorithm
SPF (Survey of Professional Forecasters)
forecast combinations and, 356, 356n1, 364, 366t, 369t
multidimensional panel data of survey forecasts and, 473, 480n3, 482, 484, 485, 487, 487f, 488–90, 492
survey forecasts and, 463, 467, 468, 469
spurious regression problem, 53, 593, 594, 596
square root transformations, cost regression health care models, 631
SR models. See switching regression models
St. Louis Fed, 25, 417n2
stable populations, 589–90
STAMP software package, 154, 156, 158
(p. 709) standard error of estimates (SEE), 655, 661, 664, 665, 666
STAR (smooth transition autoregressive) models, 64. See also LSTAR models
neural networks v., 516–17
STR v., 64, 65
vector, 65
stated choice methods, 574, 576
stated preference methods, 560, 573–74, 579
stated value (SV) studies, 574, 575
state-space models, 138
data revisions and, 262, 307
diagnostic checking, 144–46
with dynamic factors, 43–44
hybrid principal components/state-space methods, 37, 42–44
Kalman filter and, 130, 141–42
likelihood evaluation, 142
linear time series models and, 138
log-linearized DSGE, 90
MLE of, 142–43
multivariate version (unobserved components time series models), 152
nonlinear/non-Gaussian, 45–46
nowcasting model, 201, 220–21
parameter estimation, 142–44
with static factors, 43
time-domain maximum likelihood (via Kalman filter) and, 38–39
unobserved components time series models, 130, 139–41
static factors, 38. See also dynamic factors
determining number of, 46–47
estimation of dynamic factor innovations from, 42
exact static factor model, 41
state-space model with, 43
statistical models. See election forecasting statistical models; sales forecasting
step-down methods
Holm's step-down method, 392, 393, 399, 400, 404
stepM approach, 399–400
stepwise multiple testing approach (stepM), 399–400
sterling/dollar exchange rate changes, 502, 504f, 506, 506t, 509–13, 510f
stochastic breaks, 277, 278, 279, 285, 357
stochastic models, of electricity spot prices, 615–17
stochastic volatility (SV) models, 508–9
estimation, high-frequency data and, 528
IGARCH(1,1) models v., 508–9
stop-break model, 336
strategic bias, 167–68, 173t
straw polls, 657
STR (smooth transition regression) models, 64–65, 320, 329
logistic, 64, 335–36, 337, 360, 366, 515, 619, 620
STAR v., 64, 65
switching regression models v., 64
strong form reduced-rank structure, 15
structural breaks (unanticipated structural breaks). See also forecasting breaks; forecasting during breaks; forecasting in presence of unanticipated location shifts; location shifts; model instability
as “absorbing barrier,” 328
causes of, 272
DFMs with, 52–53
DGPs and, 272
disaggregation by variable and, 291
forecast combinations and, 273, 301–7
forecast combinations experiment and, 357–58, 382–86
forecasting impacted by, 315, 346–47
misspecified models and, 302–5
modeling impacted by, 315, 346–47
models' susceptibility to, 272
nowcasting and, 327
parameter nonconstancies and, 271, 282, 289
policy impacted by, 315, 346–47
poor forecast performance and, 3, 271
stochastic breaks, 277, 278, 279, 285, 357
testing, 447
unmodeled, 346–47, 418, 429
structural shocks
FAVARs and, 50, 51
in NAWM, 99, 101, 102, 103, 104n1
structural time series models, 131
structural VARs. See SVARs
subjective probabilities, point predictions and, 459–60
subprime loans, 316, 316f
superior predictive ability test. See SPA test
supply shocks, 607, 610, 613, 616, 620
Survey, Livingston, 167, 467, 473, 481, 492
survey data, 326. See also information
survey forecasts (interpreting and combining), 4, 457–72. See also forecast combinations illustration: survey and time series forecasts; multidimensional panel data of survey forecasts
assessing temporal variation of forecasts (by panels of forecasters), 458, 467–69
consensus forecasts, 457–58, 464–67
heterogeneity of, 4, 458
interpreting, 4, 457–69
mean forecasts and convex loss functions, 465–66
nowcasting and, 194
point predictions of binary outcomes, 459–62
point predictions of real-valued events, 462–63
point predictions of uncertain events, 457, 458–64
probabilistic forecasting, 463–64
SPFs and, 463, 467, 468, 469
Survey of Economic Expectations (SEE), 464
(p. 710) Survey of External Forecasters, Bank of England's, 467
Survey of Professional Forecasters. See SPF
SV. See stated value studies; stochastic volatility models
SVARs (structural VARs)
DFMs and, 55
FAVARs v., 50–51
NAWM and, 93, 97, 99, 109
Sweden economic growth studies, 597–98, 597f, 600–601, 601f. See also long-horizon economic growth forecasting
switching-mean autoregressive (SM-AR) model, 71
switching regression (SR) models, 329. See also regime switching models
as nonlinear time series models, 2, 62–63
STR models v., 64
systematic patterns in random events (bias), 169, 173t
TAQ database, 532, 533
TAR (threshold autoregressive) models, 63, 84, 329–30
task properties feedback, 175–76
taxonomy, for forecast errors. See forecast error taxonomies
Taylor expansion 330, 330n2
technical analysis (chartism), 500, 501
temporal variation of forecasts, by panels of forecasters, 458, 467–69
term structure of interest rates, 10, 521
testing for nonlinearity, 331. See also forecasting models
tests. See conditional predictive ability tests; health care costs models; unconditional predictive ability tests; specific tests
three-dimensional panel data of survey forecasts. See multidimensional panel data of survey forecasts
threshold autoregressive (TAR) models, 63, 84, 329–30
threshold models, 329–30, 337
time disaggregation. See disaggregation by time
time-domain estimation, of dynamic factors, 37–44
time series approaches, long-horizon growth forecasting, 592–99, 602
time-varying dummy seasonal component, 134
time-varying factor loading (TVFL) models, 28, 29t
time-varying parameters, DFMs with, 52–53
time-varying predictive ability testing, 446–50. See also conditional predictive ability tests
time-varying trigonometric cycle component, 136–37
time-varying trigonometric seasonal component, 135
trace(MSFE), 22t, 24t, 26, 27t, 28, 29t
trace(MSFEh), 20, 21
Tracy–Widom law, 47
transformed health care costs models, 629–32
Box-Cox models, 631–32, 639, 642t, 646, 647
log transformations, 629–31
semiparametric transformation models, 632
square root transformations, 631
Treasury bill yield changes, 499, 502, 503f, 506, 506t, 509–13, 512f
trend component, 129. See also smooth trend component
common trends and cycles, 130, 150–52
local linear, 132
multivariate, 149–50
specifications, 131–33
with stationary drift, 132
trigonometric components
fixed trigonometric seasonal component, 134–35
smooth time-varying trigonometric cycle component, 137
time-varying trigonometric cycle component, 136–37
time-varying trigonometric seasonal component, 135
trimming, 79, 178, 179, 356, 643, 644. See also forecast combinations; insanity filter
tsunamis, 316–17
Tulip and Reifschneider measure, 484, 485
TVFL (time-varying factor loading) models, 28, 29t
“twilight world,” 174
2004 Indian Ocean tsunami, 316–17
2007-2010 financial crisis, 316, 318, 336
two-part health care costs models, 625–26, 629, 640
two-regime switching regression model, 64, 330
type 1 error, 392, 393, 406, 410, 411. See also FDR
UC-RV (unobserved components-realized volatility) model, 541
UK demand evolution for load periods, 612–13, 613f
UK election forecasting models, 662–63
UK mortgage lending, 316, 316f
UK spot prices
evolution, for load periods (electricity prices), 614f
time series, 609, 610f, 614
unanticipated structural breaks. See structural breaks
unbiasedness condition, Muth's, 474–75
uncertainty. See also forecast uncertainty; parameter estimation uncertainty
uncertainty information and weather/climate forecasts, 560, 561–64, 579–80
uncertainty measures of demographically based long-horizon growth forecasting, 603
(p. 711) unconditional predictive ability tests, 4, 415–40
conditional predictive ability tests v., 416, 438, 441, 442–43, 444
data revisions and, 428–29
existing extensions of research, 427–28
forecast accuracy measures, 415–16, 418, 419t
Monte Carlo experiments, 429–38
other dimensions, 427–29
population-level predictive ability, 419–23, 438–39
pseudo-out-of-sample tests, 416, 417–19, 438
real-time data and, 428–29
topics for future research, 429
unimodal loss functions, median forecasts and, 466
univariate model (AR), 361, 365t, 366t. See also forecast combinations illustration: survey and time series forecasts
universal approximators, 61, 68–72. See also neural networks
illustration of, 68
Kolmogorov-Gabor polynomials, 69–70, 73, 81
universes of models. See also forecast combinations
design of, 356, 358–61
restricted, 361
unmodeled breaks. See location shifts; structural breaks
unobserved components-realized volatility (UC-RV) model, 541
unobserved components time series models, 129–62. See also cycle component; irregular component; seasonal component; trend component
autoregressive representation of, 148
diagnostic checking, 130, 144–46
electricity daily spot prices (illustration), 130–31, 153–59
extending, to non-Gaussian class, 159
forecasting, 146–48
interest variable forecasting and, 130
multivariate extensions of, 130, 148–53
observation weights of forecast function, 147–48
in state-space form, 130, 139–41
unpredictability
formal description, 322–23
information and, 321–24, 348
predictability of breaks, 317
unrestricted VARs. See UVARs
UVARs (unrestricted VARs), 17, 23, 24t, 26, 27t, 28, 29t
valuation of weather and climate forecasts. See weather and climate forecasts
VARFIMA (vector fractionally integrated ARMA) model, 541
variable of interest forecasting, 130
variables disaggregation. See disaggregation over variables
VARs (vector autoregressive models), 1–2, 9–34. See also BVARs; NAWM benchmarks; SVARs
cross-equation restrictions, 1–2, 10, 11, 17
DSGEs v., 2, 89, 90, 104
DSGE-VARs, 104, 104n2, 124
DVARs, 17, 26
FAVARs, 2, 37, 42, 45, 48, 50–51
forecast combinations and, 305
gas pricing and, 608
GVARs, 305–6
information criteria and, 15–16
Minnesota prior and, 103–4
MS-VARs, 66, 320, 337
nonlinear time series models v., 2
nowcasting and, 212
observable factor models and, 2, 11, 14, 18, 30
oil price dynamics and, 608
overparameterization problem in, 9–10
sequential testing and, 15–16
UVARs, 17, 23, 24t, 26, 27t, 28, 29t
VARs in differences (DVARs), 17, 26
VEC (vector error-correction), 10
VECMs (vector error-correction models)
Bayesian, 25
Bayesian reduced-rank, 18–19
BVARs and, 19
cobreaking in, 17
cointegration and, 10, 17
forecast performance, 17, 23, 24t, 26, 27t, 28, 29t
reduced-rank, 18–25, 30
RVECMs, 24t, 25, 26, 27t, 28, 29t
vector autoregressive models. See VARs
vector equilibrium correction models. See VEqCMs
vector error-correction. See VEC
vector error-correction models. See VECMs
vector fractionally integrated ARMA (VARFIMA) model, 541
vector STAR models, 65
vector switching regression model, 63
vector threshold autoregressive model, 63
VEqCMs (vector equilibrium correction models), 280
DDD v., 283–87
differenced VEqCM, 280, 287–88, 347
in-sample forecast model, 281–82
LDGP nonconstancy and, 284–87
location shift's impact on, 282–83
verification, weather and climate forecasts, 562–63
vintage dates, 248, 248t, 250, 251t, 255, 256, 257. See also data vintages
volatility
assessing persistence of, 533–35
dynamic properties, high-frequency data and, 526–27
(p. 712) of electricity spot markets, 609–10
measuring, 531
multidimensional panel data of forecasts and, 478–81
of oil market, 620
realized measures of, 531–32
“realized volatility,” 507
statistical properties of, 523–35
volatility forecasting
ARCH models, 239, 507–9
Black–Scholes option pricing model and, 507, 509
financial time series forecasting and, 506–13
GARCH models and, 507–13
long memory models and, 513–15
methods for, 507
MIDAS and, 227, 238–41
“realized volatility” and, 507
survey of, 507
SV models and, 508–9
volatility forecasting, using high-frequency data, 4, 525–56
ARFIMA models, 526–27, 540–41
data cleaning, 532
distributed lag models, 537–40
future research, 550
GARCH models and, 543–49
model-based volatility forecasting, 527–29, 543–50
notational framework, 529–30
realized measures in, 527–28, 535–37
reduced-form volatility forecasting, 527–29, 537–43, 549–50
TAQ database, 532, 533
volcanic eruptions, 318
vox populi, 465
Wald test statistic, 53, 445
weak form reduced-rank structure, 14
weather and climate forecasts (economic value assessment), 5, 559–83
background on methods, 560–64
case studies, 570–73
climate v. weather forecasts, 561
cost-loss decision-making model, 560, 567–70, 568t, 569f, 571
decision analytic valuation studies, 560, 567–73, 579–80
economic forecasting v., 561
economic value of imperfect information, 560, 564–67
“good,” 562
hurricane forecasts study, 575–79, 576f, 578t
nonmarket valuation of, 560, 573–79
probability, 562–64
prototypical decision-making models, 567–70
reliability and, 562–63
revealed preference methods, 573
stated choice methods, 574, 576
stated preference methods, 560, 573–74, 579
stated value studies, 574, 575
uncertainty information and, 560, 561–64, 579–80
verification, 562–63
weather v. climate forecasts, 561
weights
assigning, in forecast combinations, 356
Beta weight function, 228, 240, 241, 539
equal-weighted combination scheme (forecast combinations illustration), 359–61
equal-weighted forecast v. best model forecast (model instability illustration), 381–82, 381f, 386–87
EW, 364, 365t, 366t
exponential Almon weighting scheme, 228, 229, 237, 240
observation weights of forecast function (unobserved components time series models), 147–48
ox weighing and, 465
Wharton Research Data Services (WRDS) system, 534
white chaos, 517
white noise process, 503, 514
“wisdom of crowds,” 465. See also consensus forecasts
“wishful expectations hypothesis,” 168
Wold forecast error, 197
Wold representation, 12
World Economic Survey, Center for Economic Studies ifo Institute, 467
world liquidity changes, 316, 316f
WRDS (Wharton Research Data Services) system, 534
YADA program, 96