adaptive smoothing, 60
adjustments, 5
based on error evaluations, 51–52
holdout-sample evaluations and, 55–56
arbitrage pricing theory (APT) model, 182–183
Aristotle, 243
associative analyses, 6, 86, 158, 173
autocorrelation, 100–101, 103–104, 107–108, 110–111, 127
autoregressive (AR) models, 244
standard error of the regression, 162–163
autoregressive integrated moving average (ARIMA) models, 158, 244
testing for stationarity, 164–166
average consumer expectations for business conditions, 225
average forecast value, 59
average seasonal index (AS), 145–146
average weekly hours in manufacturing sector, 221–222
best linear unbiased estimators (BLUE), 87
bond markets. See also stock markets
expected yield to maturity (YTM) of a bond, 178–180
interest rate forecasting (IRF), 180
Box–Jenkins procedure, 164
business forecast models. See diffusion models on sales and demand; financial forecasting; operational forecasting
capital asset pricing model (CAPM), 181–182
cautions on forecast results
confusion between precision and accuracy, 247
global financial conditions, 248
mistaking correlation for causality, 246
overestimation or underestimation of forecast values, 247
theory-less forecasting, 247
center moving average (CMA), 145
central tendency, measures of, 8–11. See also probability
charting tools of Excel, 13–15
composite index, 6
composite seasonal index, 145–146
constant price, 35
consumer price index (CPI), 36–37
convergence (or catching up), 231–236
required growth rate to keep a constant gap, 235–236
required growth rate to reach a target, 234–235
covariance, 9
credit index, 224
logarithm of zero and, 250
missing observations and, 249
current prices, 35
cyclical-random index, 147–148
changing units, 250
cleaning of data, 249
preliminary steps of, 251
data manipulations in Excel, 16–17
decomposition
calculating trend-random values (TRt), 146–147
concept, 144
constructing a composite seasonal index, 145–146
constructing seasonal-random indices for individual periods, 144–145
cyclical component in, 147–148
obtaining forecast values, 147
steps in, 144
three-component, 144
Delphi method, 6
de-randomizing process, 154
descriptive statistics, 18–20, 251
de-seasoning process, 154
determinants of land use, 211
deviation in percentage (DPE) measures, 46–47
diagnostic checking, 164
diffusion indices, 225–228. See also turning point, forecasting
diffusion models on sales and demand, 188–194. See also financial forecasting; operational forecasting
Lawrence–Lawton model, 191–193
discrete variable, 7
dispersion of a distribution, 9
dividend discount model (DDM), 183–184
double exponential smoothing (DE), 74–81, 244. See also Exponential smoothing (ES)
forecasts against actual demand, 80–81
double moving averages (DMs), 65–74
equation for one-period forecasts using, 67
early adopters, 190
early majority, 190
cross-sectional regressions, 116
linear regression analysis, 86–87
multiple linear regressions, 115–116
economic indicators, 6
economic models. See input–output demands model; production–allocation model
error measurements
evaluations, 5
error. See Error measurements
Excel applications
autocorrelation, 107–108, 110–111
autoregressive (AR) models, 168–169
autoregressive integrated moving average (ARIMA) models, 166–167
capital asset pricing model (CAPM), 187–188
composite seasonal index, obtaining, 151
converting nominal values to real values, 39
cross-sectional data for regression analysis, 91–93, 120–122
cyclical component values, obtaining, 152
descriptive statistics, performing, 18–20
diffusion models on sales and demand, 193–194
double exponential smoothing (DE), 72–74
expected yield to maturity (YTM) of a bond, 185–186
exponential smoothing (ES), 34–35
FB, 49
forecast values, obtaining, 151
higher-order exponential smoothing (HOE), 156–158
Holt–Winters Exponential Smoothing, 155–156
interest rate forecasting (IRF), 186–187
MAE, 49
mathematical operations in, 11–12
MSE, 49
multiple linear regressions, 120–123, 131–134
operational forecasting, 176–177
panel-data forecasting, 137–140
production–allocation model, 205–210
seasonal-random indexes for individual quarters, constructing, 150
time-series data for regression analysis, 93–94, 122–123
trend-random values, calculating, 151
triple exponential smoothing, 155–158
weighted MA, performing, 30–31
expected yield to maturity (YTM) of a bond, 178–180
exponential smoothing (ES), 23, 244. See also Double exponential smoothing (DE)
calculating an average values, 32
next-period forecast and current-period forecast, 33
obtaining forecast values, 34–35
financial forecasting, 177–188, 196–197. See also diffusion models on sales and demand; operational forecasting
Fisher, M.E., 243
fixed effect estimators 135-136
focus forecasting, 60
forecast bias (FB) measures, 47
forecasting
methods, 4, 253–256. See also cautions on forecast results
forecast models
with a quadratic term, 154
forecast of forecast technique, 91, 174
forecast results, cautions on
confusion between precision and accuracy, 247
global financial conditions, 248
mistaking correlation for causality, 246
overestimation or underestimation of forecast values, 247
theory-less forecasting, 247
future information for forecasting, 4
future period, 4
Gauss–Markov theorem, 87
generalized least squares (GLS), 102
Global Indicator Program, 248
good forecaster, 5
goodness of fit, 130
trip distribution forecasts, 212–215
gross domestic product (GDP), 35
gross domestic product (GDP) deflator (GDPD), 36–38
heteroskedasticity, 98–100, 102–103, 106–110, 127, 129–130
higher-order exponential smoothing (HOE), 154–155
Holt–Winters exponential smoothing (HWE), 153–154
individual judgment, 5
innovators, 190
input–output demands model, 199, 203–205. See also production–allocation model
Institute for Supply Management (ISM) index, 222–223
interest rate forecasting (IRF), 180
interest rate spread of 10-year Treasury Less Federal Funds Target, 224–225
interval forecasts
Excel application, 83–84, 94–96
for t-distribution for a sample of N observations, 89
interval prediction, 90
laggards, 190
Lagrange Multiplier (LM) test, 99, 101
late majority, 190
Lawrence–Lawton model, 191–193
least squares dummy variable (LSDV) estimation, 135, 139
left-skewed (negatively skewed) distribution, 10
left-skewed sample, 10
linear optimization, 200
linear regression analysis, 85, 244–245. See also multiple linear regressions
assumptions, 87
evaluations and adjustments, 96–111
of personal consumption (CONS), 86
for time-series data, 87
logarithm of zero, 250
log-log model, 101
longitudinal (panel) data, 12
manufacturers’ new orders
of consumer goods and materials, 222
for nondefense capital goods, 223
mean absolute deviation (MAD), 44
mean absolute error (MAE), 44–45
approximations, 82
mean absolute percentage error (MAPE), 45–46, 188, 243
mean squared error (MSE), 44, 59
modified Hausman test for endogeneity, 132–133
moving averages (MA), 23
concept, 26
moving average of order one MA(1), 163
pros and cons, 243
multicollinearity problem, 116
multiperiod forecasts, 91
multiple linear regressions. See also linear regression analysis
for cross-sectional data, 116
evaluations and adjustments, 123–134
Excel applications, 120–123, 131–134
interval estimates, 118
multiple regression model specifications, 130–131
of personal consumption (CONS), 115–116
point estimates, 117
time-series data, 117, 122–123
negative bias, 47
Newton’s law of gravitational attraction, 210–211
nonregression analysis, 6
normal distribution, 8
one-period interval forecasts, 91
operational forecasting. See also diffusion models on sales and demand; financial forecasting
running forecast technique, 174–176
ordinary least squares (OLS) estimation, 87, 88, 137
pooled, 134
panel-data forecasting
detecting different characteristics, 137
fixed effect estimators 135-136
panel of experts method, 5
point estimates, 88
pooled OLS estimation, 134
positive bias, 47
preliminary steps of data analysis, 251
probability, 7–8. See also central tendency, measures of
probability distribution function (pdf), 7–8
problem of endogenous regressors, 128
production–allocation model, 199–203. See also input–output demands model
linear function, 201
production limits, 202
p-value, 104
rate of change, estimation. See diffusion indices
rates of change, models on
convergence (or catching up), 231–234
real gross domestic product (RGDP), 220, 228
regression analysis, 6
right-skewed (positively skewed) distribution, 10
right-skewed sample, 10
right-tail test, 97
running forecast technique, 174–176
seasonal-random indices for individual periods, 144–145
serial correlation, 100
simple linear regressions. See linear regression analysis
simple moving averages (MA), 26–28
skewness, 10
smoothing constant, 32
standard deviation (SD), 10
of forecast (SDF) measures, 43–44
standard error of the forecast, 82, 83, 90, 105–106, 162–163
standard error (SE), 10
Standard & Poor’s 500 (S&P 500), 224
τ-statistic (tau-statistic), 164
stock markets. See also Bond markets
arbitrage pricing theory (APT) model, 182–183
capital asset pricing model (CAPM), 181–182
dividend discount model (DDM), 183–184
t-distribution, 89
three-component decomposition, 144
three-parameter exponential smoothing, 154
Thurman, W.N., 243
missing observations, 250
time series data, 12
trend-random values (TRt), 146–147
trip distribution forecasts, 212–215
triple exponential smoothing
higher-order exponential smoothing (HOE), 154–155
Holt-Winters exponential smoothing (HWE), 153–154
turning point, forecasting. See also diffusion indices
average consumer expectations for business conditions, 225
average weekly hours in manufacturing sector, 221–222
building permits and new houses, 223–224
credit index, 224
Institute for Supply Management (ISM) index, 222–223
interest rate spread of 10-year Treasury Less Federal Funds Target, 224–225
manufacturers’ new orders
of consumer goods and materials, 222
for nondefense capital goods, 223
measures of unemployment, 222
Standard & Poor’s 500 (S&P 500), 224
two stages least squares (2SLS) estimators, 129
two-tail test, 98
unemployment, measures of, 222
unit-root test, 164
variance measures, 9
weighted forecast value, 59
weighted moving averages (WA), 28, 243
Z-critical value of a normal distribution, 57–58, 82
zero covariance, 9
3.15.226.120