irwin/mcgraw-hill andrew f. siegel, 1997 and 2000 14-1 l chapter 14 l time series: understanding...
DESCRIPTION
Irwin/McGraw-Hill © Andrew F. Siegel, 1997 and Cross-Sectional and Time-Series Cross-Sectional Data Expect next observation to be about S away from Time-Series Data Next will probably not be about S away from (not a random sample) S S S STRANSCRIPT
Irwin/McGraw-Hill © Andrew F. Siegel, 1997 and 2000
14-1
l Chapter 14 l
Time Series: Understanding
Changes over Time
Irwin/McGraw-Hill © Andrew F. Siegel, 1997 and 2000
14-2
Time Series AnalysisGoals
Understand the past Forecast the future
Different from Cross-Sectional Data Time-series data are not independent of each other
Not a random sample Does not satisfy the random-sample assumption for confidence intervals (in
Chapter 9) or hypothesis testing (in Chapter 10) New methods are needed to take account of the interdependence
Irwin/McGraw-Hill © Andrew F. Siegel, 1997 and 2000
14-3
Cross-Sectional and Time-SeriesCross-Sectional Data
Expect nextobservation to beabout S awayfrom
Time-Series Data Next will probably
not be about S away from(not a random sample)
X
X
XSS
S XS
Irwin/McGraw-Hill © Andrew F. Siegel, 1997 and 2000
14-4
ForecastingUse a model
A system of equations that can produce data that “look like” your time series data
Estimate the model Your forecast will be the expected (mean) value of the future
behavior of the model The forecast limits are the confidence limits for your forecast (if your
model can produce them) Computed from the appropriate standard error If model is correct, the future observation has a 95% chance of being
within the forecast limits
Irwin/McGraw-Hill © Andrew F. Siegel, 1997 and 2000
14-5
Trend-Seasonal and Box-JenkinsTrend-Seasonal Analysis
Direct and intuitive, with four components: (1) Long-term Trend, (2) repeating Seasonal, (3) medium-term wandering
Cyclic, and (4) random Irregular Forecast comes from extending the Trend and Seasonal
Box-Jenkins ARIMA Process Flexible, but complex, probability models for how current value of
the series depends upon Past values, past randomness, and new randomness
A better way to describe the Cyclic component Forecast is
Expectation of random future behavior, given past data
Irwin/McGraw-Hill © Andrew F. Siegel, 1997 and 2000
14-6
Example: Computer RevenuesSteady growth
not perfectly smooth Nonlinear (curved)
Suggests constant growth rateLogarithm of revenues
Log plot looks linearif constant growth rate
Can use regression tomodel relationship Points appear randomly
distributed about the line,so serial correlation is not a problem
Fig 14.1.4, 6
0
10
20
30
40
50
1986 1988 1990 1992 1994 1996
Rev
enue
s (bi
llion
s)
2
3
4
1986 1988 1990 1992 1994 1996
Log
of R
even
ues
Irwin/McGraw-Hill © Andrew F. Siegel, 1997 and 2000
14-7
Example: Retail SalesU.S. Retail Sales (Monthly)
Growth Repeating seasonal variation
High in December Low in January, February
Seasonally-Adjusted Sales Growth Seasonal pattern removed
Shows how sales went up(or down) relative to whatyou expect for time of year
Fig 14.1.7, 8
150
200
250
1995 1996 1997 1998
Sale
s ($b
illio
ns)
150
200
250
1995 1996 1997 1998
Sale
s ($b
illio
ns)
Irwin/McGraw-Hill © Andrew F. Siegel, 1997 and 2000
14-8
Example: Interest RatesU.S. Treasury Bills, Yearly
Generally rising
Substantial variation
Cyclic pattern Rising and falling Increasing magnitude Not perfectly repeating Not expected to continue rising indefinitely!
Fig 14.1.9
0%
5%
10%
15%
1960 1970 1980 1990 2000In
tere
st ra
te
Irwin/McGraw-Hill © Andrew F. Siegel, 1997 and 2000
14-9
Trend-Seasonal AnalysisDecompose a Time Series into Four Components
Data = Trend Seasonal Cyclic Irregular Trend
Long-term behavior (often straight line or exponential growth) Seasonal
Repeating effects of time-of-year Cyclic
Gradual ups and downs, not repeating each year, not purely random Irregular
Short-term, random, nonsystematic noise
Irwin/McGraw-Hill © Andrew F. Siegel, 1997 and 2000
14-10
Ratio-to-Moving-Average MethodMoving Average Represents Trend and Cyclic
Eliminates Seasonal and Irregular by averaging a yearDivide Data by Moving Average
Represents Seasonal and Irregular Group by season, then average, to obtain Seasonal
Seasonal Adjustment: Divide Data by SeasonalRegress Seasonally-Adjusted Series vs. Time
Represents TrendForecast by Seasonalizing the Trend
Multiply (future predicted Trend) by (Seasonal index)
Irwin/McGraw-Hill © Andrew F. Siegel, 1997 and 2000
14-11
Example: Ford Motor CompanyTime-series Plot
Quarterly data with strong Seasonal pattern Revenues typically highest in second quarter Does not repeat perfectly (due to Cyclic and Irregular)
Fig 14.2.1
$0
$10
$20
$30
1990 1995 2000
Sale
s ($
billi
ons)
Irwin/McGraw-Hill © Andrew F. Siegel, 1997 and 2000
14-12
Example: Moving AverageAverages one year of data
2 quarters before to 2 quarters after each data value Smooths the data, eliminating Seasonal and Irregular Shows you Trend and Cyclic
Original data
Moving average
Fig 14.2.4
$0
$10
$20
$30
1990 1995 2000
Sale
s ($
billi
ons)
Irwin/McGraw-Hill © Andrew F. Siegel, 1997 and 2000
14-13
Example: Seasonal IndexFig 14.2.6
Average “Ratio-to-Moving-Average” by Quarter Seasonal index for each quarter, repeating each year
Shows how much larger (or smaller) this quarter is compared to a typical period throughout the year
0.8
0.9
1.0
1.1
1990 1995 2000
Seas
onal
inde
x
Irwin/McGraw-Hill © Andrew F. Siegel, 1997 and 2000
14-14
Example: Seasonal AdjustmentDivide Data by Seasonal Index
To get Seasonally Adjusted Value Eliminates the expected seasonal component Shows changes that are not due to expected seasonal effects
Original data
Seasonallyadjusted series
Fig 14.2.7
$0
$10
$20
$30
1990 1995 2000
Sale
s ($
billi
ons)
Irwin/McGraw-Hill © Andrew F. Siegel, 1997 and 2000
14-15
Example: Trend LineRegress Seasonally-Adjusted Data vs. time
The resulting line can be extended into the future This gives a Seasonally-Adjusted Forecast
Fig 14.2.8
$0
$10
$20
$30
$40
1990 1995 2000
Sale
s ($b
illio
ns)
Seasonallyadjustedforecast
Seasonallyadjusted series
Trend line
Irwin/McGraw-Hill © Andrew F. Siegel, 1997 and 2000
14-16
Example: ForecastSeasonalize the Trend
Multiply Trend by Seasonal Index Can be extended into the future
Use future predicted Trend with quarterly Seasonal index
Fig 14.2.9
$0
$10
$20
$30
$40
1990 1995 2000
Sale
s ($b
illio
ns)
Original data
Seasonalized trendForecast
Irwin/McGraw-Hill © Andrew F. Siegel, 1997 and 2000
14-17
Box-Jenkins ARIMA ProcessesA Collection of Linear Statistical Models
Can describe many different kinds of time-series Including medium-term “cyclic” behavior
Compared to trend-seasonal analysis, Box-Jenkins Has a more solid statistical foundation Is more flexible Is somewhat less intuitive
Outline of the steps involved Choose a type of model and estimate it using your data Forecast using average future random behavior of this model Find standard error (variability in this future behavior) Find forecast limits, to include 95% of future behavior
Irwin/McGraw-Hill © Andrew F. Siegel, 1997 and 2000
14-18
Random Noise ProcessA Random Sample, with No Memory
Data = (Mean value) + (Random Noise) Yt = + t
The long-term mean of Y is
Mean
Irwin/McGraw-Hill © Andrew F. Siegel, 1997 and 2000
14-19
Autoregressive (AR) ProcessRemembers the Past, Adds Random Noise
Data = + (Previous value) + (Random Noise) Yt = + Yt–1 + t
The long-term mean value of Y is –
Mean
Irwin/McGraw-Hill © Andrew F. Siegel, 1997 and 2000
14-20
Moving-Average (MA) ProcessRemembers Previous Noise, Adds New Noise
Data = + (Random Noise) – (Previous Noise) Yt = + t – t–1
The long-term mean value of Y is
Mean
Irwin/McGraw-Hill © Andrew F. Siegel, 1997 and 2000
14-21
ARMA ProcessAutoregressive Moving Average Process
Remembers the Past, Previous Noise, Adds New NoiseData = + (Previous value) + (Noise) – (Previous Noise) Yt = + Yt–1 + t – t–1
The long-term mean value of Y is –
Mean
Irwin/McGraw-Hill © Andrew F. Siegel, 1997 and 2000
14-22
Example: Unemployment
Estimated ARMA Process for this Time SeriesYt = + Yt–1 + t + t–1
where random noise has standard deviation 0.907
0%
5%
10%
1960 1970 1980 1990 2000
Une
mpl
oym
ent r
ate
Irwin/McGraw-Hill © Andrew F. Siegel, 1997 and 2000
14-23
Example (continued)Random Simulations from Estimated Process
Look similar to actual unemployment rate history Because of estimation using actual data Looking at “what might have happened instead”
0%
5%
10%
1960 1970 1980 1990 2000
Une
mpl
oym
ent r
ate
Irwin/McGraw-Hill © Andrew F. Siegel, 1997 and 2000
14-24
Example (continued)Forecast and 95% Forecast Limits (10 years ahead)
Using the average of random future possibilities And their lower and upper 95% limits
0%
5%
10%
1960 1970 1980 1990 2000 2010
Une
mpl
oym
ent r
ate
Forecast
Irwin/McGraw-Hill © Andrew F. Siegel, 1997 and 2000
14-25
Example (continued)Three Simulations of the Future
With forecast and 95% Forecast Limits To see how forecast represents future possibilities
0%
5%
10%
1960 1970 1980 1990 2000 2010
Une
mpl
oym
ent r
ate
Irwin/McGraw-Hill © Andrew F. Siegel, 1997 and 2000
14-26
Pure Integrated (I) ProcessA Random Walk from the Previous Value
Data = + (Previous value) + (Random Noise) Yt = + Yt–1 + t
Over time, Y is not expected to stay close to any long-term mean value
Irwin/McGraw-Hill © Andrew F. Siegel, 1997 and 2000
14-27
ARIMA ProcessAutoregressive Integrated Moving AverageRemembers its Changes
The differences, Yt – Yt–1, follow an ARMA process Over time, Y is not expected to stay close to any long-term mean
value