9781118041581

Forecasting 151

future economic developments. (The stock market is one of the best-known

leading indicators of the course of the economy.)

Time-Series Models

Time-series modelsseek to predict outcomes simply by extrapolating past

behavior into the future. Time-series patterns can be broken down into the fol-

lowing four categories.

1. Trends


2. Business cycles


3. Seasonal variations


4. Random fluctuations

A trendis a steady movement in an economic variable over time. For exam-

ple, the total production of goods and services in the United States (and most

other countries) has moved steadily upward over the years. Conversely, the

number of farmers in the United States has steadily declined.

On top of such trends are periodic business cycles. Economies experience

periods of expansion marked by rapid growth in gross domestic product

(GDP), investment, and employment. Then economic growth may slow and

even fall. A sustained fall in (real) GDP and employment is called a recession. For

the United States’ economy, recessions have become less frequent and less

severe since 1945. Nonetheless, the business cycle—with periods of growth fol-

lowed by recessions, followed in turn by expansions—remains an economic

(and political) fact of life.

Seasonal variationsare shorter demand cycles that depend on the time of

year. Seasonal factors affect tourism and air travel, tax preparation services,

clothing, and other products and services.

Finally, one should not ignore the role of random fluctuations. In any short

period of time, an economic variable may show irregular movements due to

essentially random (or unpredictable) factors. For instance, a car dealership

may see 50 more customers walk into its showroom one week than the previous

week and, therefore, may sell eight more automobiles. Management is grateful

for the extra sales even though it can identify absolutely no difference in eco-

nomic circumstances between the two weeks. Random fluctuations and unex-

pected occurrences are inherent in almost all time series. No model, no matter

how sophisticated, can perfectly explain the data.

Figure 4.3 illustrates how a time series (a company’s sales, let’s say) can be

decomposed into its component parts. Part (a) depicts a smooth upward trend.

Part (b) adds the effect of business cycles to the trend. Part (c) shows the regular

seasonal fluctuations in sales over the course of the year added to the trend and

c04EstimatingandForecastingDemand.qxd 9/5/11 5:49 PM Page 151