Hanke, Wichern and Reitsch,

//\\

//\\

Business Forecasting,

\I7l \I7l O)('HAPTER

7th ed, 2001

\\ /l\\ //\\ //\\

7/\I7l \I/' \7'

Introduction
to
Forecosting
Tni, Uoot is concerned with methods used to predict the uncertain future in an effort
to help the managersof a businessmake better decisions.Such efforts involve the study
of historical data and the manipulation of thesedata to searchfor patterns that can be
effectively extended into the future.
In this text,we regularly remind readersthat soundjudgment must be used along with
numerical resultsif good forecastingis to result.The casesat the end of text chaptersemphasizethis point, and there are discussionsof this matter in this chapter and in the concluding chapter of the text.The example in this chapter emphasizesthis point as well.

Historyof Forecosling
Many of the forecastingtechniquesused today and discussedin this book were developed in the nineteenthcentury;regressionanalysisproceduresare an example.By contrast, some of the topicsin this book were developedand have receivedattention only
recently.The Box-Jenkins procedures and neural networks fall into this category.
With the development of more sophisticatedforecastingtechniqueg along with the
advent of computers,especiallythe proliferation of the small,personal computer and associatedsoftware,forecastinghas receivedmore and more attention. Every managernow
has the ability to utilize very sophisticateddata analysistechniques for forecasting purposes,and an understandingof these techniquesis now essentialfor businessmanagers
For this same reason,consumersof forecasts(managers)must be alert to the improper
use of forecastingtechniquesbecauseinaccurate forecastscan lead to poor decisions
New techniquesfor forecastingcontinue to be developedas managementconcern
with the forecasting processcontinues to grow.A particular focus of this attention is on
the errors that are an inherent part of any forecastingprocedure.Predictionsas to future outcomes rarely are preciselyon the mark; the forecastercan only endeavor to
make the inevitable errors as small as possible.

Need for Forecosting

In view of the inherent inaccuraciesin the process,why is forecastingnecessary?The answer is that all organizationsoperate in an atmosphere of uncertainty and that, in spite
of this fact, decisionsmust be made that affect the future of the organization.Educated
guessesabout the future are more valuable to organization managersthan are uneducated guesses.
This book discussesvarious ways of making forecaststhat rely on logical
methodsof manipulatingthe data that have been generatedby historicalevents.

cHAPTER 1 Introduction Io Forecasting

This is not to say that intuitive forecastingis bad. On the contrary,the "gut" feelings of personswho manage organizationsoften provide the only forecastsavailable.
This text discussesforecastingtechniquesthat can be used to supplementthe common
senseand management ability of decision makers;judgmental elements in the forecastingprocessare discussedin Chapter 10.It is our view that decisionmakers are better off understanding both quantitative and qualitative forecasting techniques and
using them wisely than they are if forced to plan for the future without the benefit of
valuable supplemental information.
The role of judgmental forecasting has changed in recent years.Before the advent
of modern forecastingtechniquesand the power of the computer,the manager'sjudgment was the only forecasting tool available. There is now evidence according to
Makridakis (1986) that forecastsusing judgment only are not as accurateas those involving the judicious application of quantitative techniques:
Humans possessunique knowledge and inside information not available to
quantitative methods.Surprisingly,however, empirical studies and laboratory
experiments have shown that their forecastsare not more accurate than those
of quantitative methods.Humans tend to be optimistic and underestimate future uncertainty. In addition, the cost of forecasting with judgmental methods
is often considerably higher than when quantitative methods are used.l
It is our belief that the most effective forecaster is able to formulate a skillful mix
of quantitative forecasting techniques and good judgment and to avoid the extremes
of total reliance on either. At the one extreme we find the executive who, through ignorance and fear of quantitative techniques and computers, relies solely on intuition
and feel. At the other extreme is the forecasterskilled in the latest sophisticateddata
manipulation techniques who is unable or unwilling to relate the forecasting processto
the needs of the organization and its decision makers.We view the quantitative forecasting techniques discussedin most of this book to be only the starting point in the effective forecasting of outcomes important to the organization:Analysis, judgment,
common sense,and businessexperiencemust be brought to bear at the point at which
these important techniques have generated their results.
Since the world in which organizationsoperate has always been changing,forecasts
However, recent years have brought about increasedrehave always been necessary.
liance on methods that involve sophisticateddata manipulation techniques.New technology and new disciplineshave sprung up overnight;governmentactivity at all levels
has intensified;competitionin many areashas become more keen;international trade
has steppedup in almost all industries;socialhelp and serviceagencieshave been created and have grown; and the Internet has become an important source of data and
decision-making information. These factors have combined to create an organizational
climate that is more complex,more fast-paced,and more competitive than ever before.
Organizations that cannot react quickly to changing conditions and cannot foresee the
future with any degree of accuracyare doomed to extinction'
Computers,along with the quantitative techniquesthey make possible,have become more than a conveniencefor modern organizations;theyhave become essential.
The complexitiesjust discussedgenerate tremendous amounts of data and an overwhelming need to extract useful information. The modern tools of forecasting,along
with the capabilitiesof the computer,have becomeindispensablefor organizationsoperating in the modern world.
p.17.
M-mauti, (1986),

CHAPTER I

Introduction to Forecasting 3

Who needs forecasts?Almost every organizalion,large and small, private and public, uses forecastingeither explicitly or implicitly, becausealmost every organization
must plan to meet the conditions of the future for which it has imperfect knowledge.
In addition, the need for forecastscuts acrossall functional lines as well as all types of
organizations.Forecastsare needed in finance,marketing, personnel,and production
areas,in government and profit-seekingorganizations,in small social clubs,and in national political parties.Considerthe following questionsthat suggestthe need for some
forecastingprocedures:
. If we increaseour advertisingbudget by 70"h, how will salesbe affected?
o What revenue might the state government expect over the next two-year period?
r How many units might we sell in an effort to recover our fixed investment in production equipment?
r What factors can we identify that will help explain the variability in monthly unit
sales?
o What is a year-by-yearprediction for the total loan balanceof our bank over the
next 10 years?
. Will there be a recession?If so,when will it begin, how severewill it be, and when
will it end?

Typesof Forecosts
When organizationmanagersare faced with the need to make decisionsin an atmosphere of uncertainty, what types of forecasts are available to them? Forecasting proceduresmight first be classifiedas long term or short term. Long-term forecastsare
necessaryto set the general course of an organization for the long run; thus, they become the particular focus of top management.Short-term forecastsare used to design
immediate strategiesand are used by middle managementand first-line management
to meet the needs of the immediate future.
Forecastsmight also be classifiedin terms of their position on a micro-macro continuum, that is, on the extent to which they involve small details versus large summary
values.For example, a plant manager might be interested in forecasting the number of
workers neededfor the next severalmonths (a micro forecast),whereasthe federal government is forecasting the total number of people employed in the entire country (a
macro forecast).Again, different levels of management in an organization tend to focus
on different levels of the micro-macro continuum.Top management would be interested in forecasting the sales of the entire company, for example, whereas individual
salespersonswould be much more interested in forecasting their own salesvolumes.
Forecastingprocedurescan also be classifiedaccordingto whether they tend to be
more quantitative or qualitative.At one extreme,a purely qualitative technique is one
requiring no overt manipulation of data.Only the "judgment" of the forecasteris used.
Even here,of course,the forecaster's"judgment" is actually a result of the mental manipulation of past historical data.At the other extreme,purely quantitative techniques
need no input of judgment; they are mechanicalproceduresthat produce quantitative
results.Some quantitativeproceduresrequire a much more sophisticatedmanipulation
of data than do others,of course.This book emphasizesthe quantitative forecasting
techniquesbecausea broader understandingof thesevery useful proceduresis needed
in the effective managementof modern organizations.However, we emphasizeagain
that judgment and common sensemust be used along with mechanicaland data manipulative procedures.Only in this way can intelligent forecastingtake place.

CHAPTER I

Introduction to Forecasting

Mocroeconomic
Forecosting
We usually think of forecasting in terms of predicting important variables for an individual company or perhaps for one component of a company.Monthly company sales,
unit salesfor one of a company'sstores,and absenthours per employee per month in
a factory are examples.
By contrast, there is growing interest in forecasting important variables for the entire economy of a country.Much work has been done in evaluatingmethods for doing
this kind of overall economic forecasting,called macroeconomic forecasting.Examples
of interest to the federal government of the United States are unemployment rate,
grossdomestic product, and the prime interest rate.Economic policy is based,in part,
on projections of important economic indicatorssuch as the three just mentioned.For
this reason,there is great interest in improving forecastingmethods that focus on overall measuresof a country's economic performance.
One of the chief difficulties in developing accurate forecasts of overall economic
activity is an unexpected and significant shift in a key economic factor. Among such
factors are significant changesin oil prices, inflation surges,and broad policy changes
by a country's government that affect the global economy.
The possibility of such significant shifts in the economic scene has raised a key
question in macroeconomic forecasting: Should the forecasts generated by the forecasting model be modified using the forecaster'sjudgment? Current work on forecasting methodology often involves this question.
Much work, both theoretical and practical, continues on the subject of macroeconomic forecasting.An issueof the International Journal of Forecastingis devoted to this
subject(Vol.6, No.3, October 1990).Consideringthe importanceof accurateeconomic
forecasting to economic policy formulation in this country and others, continuing attention to this kind of forecasting can be expected in the future.

Choosingo ForecosfingMethod
The preceding discussionsuggestsseveral factors to be considered in choosing a forecasting method. The level of detail must be considered.Are forecastsof specific details
needed (a micro forecast)? Or is the future status of some overall or summary factor
needed (a macro forecast)? Is the forecast needed for some point in the near future (a
short-term forecast) or for a point in the distant future (a long-term forecast)?And to
what extent are qualitative fiudgment) and quantitative (data manipulative) methods
appropriate?
The overriding consideration in choosing a forecasting method is that the results
must facilitate the decision-making processof the organizatioh's managers.The essential requirement,then, is not that the forecastingmethod involve a complicatedmathematical processor that it be the latest sophisticatedmethod. Rather, the method
chosen should produce a forecast that is accurate,timely, and understood by management so that the forecast can help produce better decisions.Also, the use of the forecasting procedure must produce a benefit that is in excessof the cost associatedwith
its use.

Forecosting
Steps
All formal forecasting procedures involve extending the experiences of the past into
the uncertain future. Thus, they involve the assumption that the conditions that generated past data are indistinguishable from the conditions of the future except for those

CHAPTER 1. Introduction to Forecasting 5

variables explicitly recognizedby the forecastingmodel. If one is forecastingjob performance ratings of employees using only the company entrance examination score as
a predictor,for example,it is being assumedthat each person'sjob performancerating
is related only to the entranceexaminationscore.To the extent that this assumptionof
indistinguishablepast and future is not totally met, inaccurateforecastsresult unless
they are modified by the judgment of the forecaster.
The recognition that forecastingtechniquesoperate on the data generatedby historical events leads to the identification of the following five steps in the forecasting
process:
1.
2.
3.
4.
5.

Data collection
Data reduction or condensation
Model building and evaluation
Model extrapolation (the actual forecast)
Forecastevaluation

Step 1, data collection,suggeststhe importanceof getting the proper data and making sure they are correct.This step is often the most challenging part of the entire forecasting process and the most difficult to monitor since subsequent steps can be
performed on data whether relevant to the problem at hand or not. Collection and
quality control problems usually abound wheneverit becomesnecessaryto obtain pertinent data in an organization.
Step 2, data reduction or condensation,is often necessarysince it is possible to
have too much data in the forecasting processas well as too little. Some data may not
be relevant to the problem and may reduce forecasting accuracy.Other data may be
appropriate but only in certain historical periods.For example, in forecasting the sales
of small cars one may wish to use only car salesdata since the oil embargo of the 1970s
rather than data over the past 50 years.
Step 3, model building and evaluation,involves fitting the collected data into a forecasting model that is appropriate in terms of minimizing the forecasting error. The simpler the model, the better it is in terms of gaining acceptanceof the forecasting process
by managerswho must make the firm's decisions.Often a balance must be struck between a sophisticatedforecastingapproach that offers slightly more accuracyand a simple approach that is easily understood and gains the support of-and, hence,is actively
used by-the company's decision makers.Obviously,judgment is involved in this selection process.Since this book discussesnumerous forecasting models and their applicability, it is our hope that the reader's ability to exercisegood judgment in the choice and
use of appropriate forecastingmodels will increaseafter studying'this material.
Step 4, model extrapolation, consistsof the actual model forecasts that are generated once the appropriate data have been collected and possibly reduced and an appropriate forecasting model has been chosen.Often forecasting for recent periods in
which the actual historical values are known is used to check the accuracy of the
process.The forecasting errors are then observed and summarized in some way. This
procedure is discussedin Step 5.
Step 5, forecast evaluation,involves comparing forecast values with actual historical
values.In this process,a few of the most recent data values are often held back from the
data set being analyzed.After the forecasting model is completed, forecasts are made
for these periods and compared with the known historical values.Some forecastingprocedures sum the absolute values of the errors and may report this sum, or divide it by
the number of forecastattempts to produce the averageforecast error. Other procedures
produce the sum of squared errors,which is then compared with similar figures from alternative forecastingmethods.Some procedures also track and report the magnitude of

CHAPTER 1. Introduction to Forecasting

the error terms over the forecastingperiod. Examination of error patterns often leads
the analyst to a modification of the forecastingprocedure.Specific methods of measuring forecastingerrors are discussednear the end of Chapter 3.

Process
Monogingthe Forecosting
The discussionin this chapter servesto underline our belief that management ability
and common sensemust be involved in the forecasting process.The forecaster should
be thought of as an advisor to the manager rather than as the monitor of an automatic
decision-making device.Unfortunately, the latter is sometimes the case in practice, especially with the aura of the computer. Again, quantitative techniques in the forecasting process must be seen as what they really are, namely, tools to be used by the
manager in arriving at better decisions.According to Makridakis (1986),
The usefulnessand utility of forecasting can be improved if management
adopts a more realistic attitude. Forecasting should not be viewed as a substitute for prophecy but rather as the best way of identifying and extrapolating
establishedpatterns or relationships in order to forecast.If such an attitude
is accepted,forecasting errors must be considered inevitable and the circumstancesthat cause them investigated.2
Given that, several key questions should always be raised if management of the
forecasting processis to be properly conducted.
.
.
o
o
e
o
.
o
.

Why is a forecast needed?

Who will use the forecast, and what are their specific requirements?
What level of detail or aggregationis required, and what is the proper time horizon?
What data are available, and will the data be sufficient to generate the needed
forecast?
What will the forecast cost?
How accurate can we expect the forecast to be?
Will the forecast be made in time to help the decision-making process?
Does the forecaster clearly understand how the forecast will be used in the
organization?
Is a feedback processavailable to evaluate the forecast after it is made and to adjust the forecasting processaccordingly?

Pockoges
ComputerForecosling
The developmentthat has had the greatestimpact on forecastingin the past several
yearsis that of computersoftwarepackagesspecificallydesignedto deal directly with
variousforecastingmethods.Two typesof computerpackagesare of interestto foretime seriesanalysis,and
casters:(1) statisticalpackagesthat includeregressionanalysis,
and (2) forecastingpackagesthat are
other techniquesusedfrequentlyby forecasters;
specificallydesignedfor forecastingapplications.
Hundredsof statisticaland forecastingpackageshave been developedfor both
mainframesand microcomputers(or personalcomputers,frequently referred to as
techniquesare
PCs).Managerswith PCson their desksand knowledgeof forecasting
managers
are
taking advanModern
their
forecasts.
no longer dependenton staff for

46F+::

CHAPTER 1 Introduction to Forecasting 7

tage of the easeand availability of sophisticatedforecastingmethods afforded by personal computers.
Becauseof the rapid turnover in available forecastingsoftware,it would not be
wise in a text such as this to detail and discussthe dedicatedforecastingpackagesavailable as this text is written. Softwareauthorsare constantlystrivingto improve the number of techniquesavailableon their programs and their easeof use and interpretation.
Forecastersmust constantlybe alert to new software developmentsand, without constantly jumping from one package to another, realize that software improvements are
inevitable. They should be in a position to take advantage of such improvements.
In spite of this caveat,it remainsessentialto use computer software packagesfor
many forecastingapplicationsbecauseof the extensivecalculationsinvolved. For this
reasonwe have chosentwo typical computer programs to illustrate many of the techniques in this text: Minitab 13 and Excel 2000.You will find screensfor theseprograms
throughout the text. Keep in mind that any programs you use will have the samebasic
calculations programmed into them, even though the input and output screenswill not
be identical.
A listing of forecasting packageswas published in The Forum, The Joint Newsletter of the International Association of BusinessForecastingand the International Institute of Forecasters.Vol. 9. No. 3. Fall 1996.

Forecosling
Exomple
Discussionsin this chapter emphasizethat forecastingrequires a great deal of judgment along with a mathematicalmanipulation of collected data.The following example shows the kind of thinking that often precedes a forecasting effort in a real firm.
Notice that the data valuesthat will produce useful forecasts,even if they exist,are not
apparent at the beginning of the processand may not have been identified. In other
words,the initial efforts may turn out to be uselessto management.The computer results of the forecasting effort using the identified variables are not shown here as they
involve topicsthat are describedthroughout this text. However,look for the techniques
describedin later chaptersto be applied to thesedata. Example 1.1 will be resolvedin
Chapter 11. For the moment, we hope that this example underlines the scope of the
forecasting effort that real managersface.
E x o m p l el . l
AlomegaFoodStoresis a retail food providerwith2"/storesin a midwesternstate.The company
engages
in variouskindsof advertisingand,until recently,hasneverstudiedthe effectits advertisingdollarshaveon sales,althoughsomedatahad beencollectedfor threeyears.
The executivesat Alomega decidedto begin trackingtheir advertisingefforts alongwith
the salesvolumesfor eachmonth.Their hope wasthat after severalmonthsthe collecteddata
could be examinedto possiblyrevealany relationshipsthat would help in future advertising
expenditures.
Theaccountingdepartmentbeganextendingits historicalrecordsby recordingthe salesvolume for eachmonth alongwith the advertisingdollarsfor both newspaperads and TV spots.
They alsorecordedboth thesevalueslaggedone and two months.This wasdone becausesome
peopleon the executivecommitteethoughtthat salesmight lag advertisingexpendituresrather
than respondin the month the adsappeared.
It wasalsobelievedthat salesexperienced
a seasonal
effect.For thisreasona dummyor categoricalvariablewasusedto indicateeachmonth.Managementalsowonderedaboutany trend
in salesvolume.
Finally,it wasbelievedthat Alomega'sadvertisingdollarsmight havean effecton its major
competitors'advertisingbudgetsthe followingmonth.For eachfollowingmonth it wasdecided
that competitors'
advertising
couldbe classified
aslittle (1),a moderateamount(2),or a great
amount(3).

cHAPTER 1 Intro duction to Fo recasting

After a few months of data collection and analysis of past records,the accounting department completed a data array for 48 months using the following variables.
.
.
.
r
e
.
.
.
.
.
.

Salesdollars
Newspaper advertising dollars
TV advertising dollars
Month code where January = 1, February = 2, through December = 12
A seriesof 11 dummy variables to indicate month
Newspaper advertising lagged one month
Newspaper advertising lagged two months
TV advertising lagged one month
TV advertising lagged two months
Month number from 1 to 48
Code 1,2, or 3 to indicate competitors' advertising efforts the following month

Alomega management, especially Julie Ruth, the company president, now wants to learn
anything it can from the data it has collected.In addition to learning about the effects of advertising on salesvolumes and competitors' advertising,Julie wonders about any trend and seasonal
effeJts on sales.However, the company's production manager,JacksonTilson, does not share her
enthusiasm.At the end of the forecasting planning meeting, he makes the following statement:
"I've been trying to keep my mouth shut during this meeting, but this is really too much. I think
we're wasting alot of people's time with all this data collection and fooling around with computers. All you have to do is talk with our people on the floor and with the grocery store manug"rs to understand what's going on. I've seen this happen around here before, and here we go
alain. Some of you people need to turn off your computers, get out of your fancy offices, and
talk with a few real peoPle."

Summory
The purpose of a forecast is to reduce the range of uncertainty within which manage-

two primary rules to which the

ment judgmentsmust be made.This purposesuggests
process
adhere:
must
forecasting
1. The forecastmustbe technicallycorrectand produceforecastsaccurateenoughto
meet the firm's needs.
2. T\e forecastingprocedureand its resultsmust be effectivelypresentedto managementso that the forecastsare utilized in the decision-makingprocessto the
basis.
mustalsobe justifiedon a cost-benefit
firm's advantage;results
The latter considerationis often misunderstoodand can be frustratingto profesYet if forecastsare to be usedto the firm's benefit,thosewho have
sionalforecasters.
authoritymustutilizethem.Thisraisesthe questionof whatmight
the decision-making
and sometimesmajor expenditures
Substantial
be calledthe "politics"of forecasting.
view of the course
and resourceallocationswithin the firm often reston management's
of future events.As the movementof resourcesand power within an organizationis
it is not surprisingto
often basedon the perceiveddirectionof the future (forecasts),
process.This
forecasting
find a certain amount of political intrigue surroundingthe
considerationunderlinesthe importanceof the secondprimaryrule:The forecastsgeneratedwithin the firm must be understoodand appreciatedby its decisionmakersso
that they are usedby the managementof the firm.
variousforecastingmodelsand procedures.
The remainderof this book discusses
appears,includingan introductionto correconcepts
statistical
basic
First,a review of
One chapteris then devotedto methodsof collecting
analysis.
lation and regression
data and exploringdata setsfor underlyingpatterns.Many specificforecastingmethods are detailedin the chaptersthat follow,and the final two text chaptersare devoted
processasit relatesto the decisionmakersof the firm.
to the forecasting