Explanation

When modeling quarterly seasonality with an intercept term, we need to use 3 dummy variables.

Reasoning:

• There are 4 quarters in a year
• If we include an intercept term (β₀), it represents the baseline level for one quarter
• To avoid perfect multicollinearity (the dummy variable trap), we omit one quarter
• The omitted quarter becomes the reference category
• Therefore: 4 quarters - 1 = 3 dummy variables

For example:

• D₁ = 1 for Q1, 0 otherwise
• D₂ = 1 for Q2, 0 otherwise
• D₃ = 1 for Q3, 0 otherwise
• Q4 becomes the reference category (when all dummies = 0)