STATISTICS FOR MANAGERS

"An economic forecaster is like a cross-eyed javelin thrower: he doesn't win many accuracy contests, but he keeps the crowd's attention." - Anonymous

1. Types of forecasting methods

a. Qualitative forecasting methods

• Used when historical data are unavailable - subjective

b. Quantitative forecasting methods

• Use historical data

(1) Causal forecasting methods

Determine factors that affect variable to be forecast

(2) Time-series forecasting methods

Forecast future values of a variable based exclusively on past and present observations of the variable

.

2. Component factors of time-series models

a.  Trend

• Overall long-term upward or downward movement in a time series

.

b.  Cyclical effect

• Upward and downward swings , often correlated to the business cycle

.

c.  Seasonal effect

• Variations related to different times of the year

• Could be significant with monthly or quarterly data

.

d.  Irregular or random effect

. 

.

.

.

.

.

.

.

.

.

3. Trend forecasting models

a. Linear trend model

.

.

.

.

.

.

.

.

.

.

b. Quadratic trend model

.

.

.

.

.

.

.

.

.

.

c. Exponential trend model

.

.

.

.

.

.

.

.

.

.

4.  Evaluating forecast models

a.  Residual analysis

.

.

.

.

.

.

.

.

.

.

b.  Squared differences

.

.

.

c.  Mean absolute deviation

.

.

.

d.  Principle of parsimony

• Select the simplest model that gets the job done adequately
  
  - Linear, quadratic, and first-order autoregressive models are the simplest
  
  - Higher order autoregressive and exponential models are more complicated