Short note: statistical approach to demand analysis

The statistical approach uses past data of demand and its determinants to estimate the demand function mathematically and to forecast future demand. It is objective and based on numbers rather than opinion.

Main techniques

1. Trend (time-series) analysis. Demand is plotted against time and a trend line is fitted, e.g. by the least squares method: Y = a + bX, where b = (nΣXY − ΣXΣY) / (nΣX² − (ΣX)²) and a = (ΣY − bΣX) / n. Other forms are moving average, exponential smoothing and exponential growth curves. The line is extended to the future years.
2. Regression (causal) analysis. Demand is related to explanatory variables such as price, income and price of substitutes: Q = a + b₁P + b₂Y + ... The sign and size of the coefficients show how demand responds. Used for elasticity estimates.
3. Correlation analysis. The coefficient r (−1 to +1) measures how strongly demand and a factor move together, and R² how well the equation fits.
4. Econometric models. A set of simultaneous equations for the market.

Merits and limits

• Merits: scientific, can be tested for accuracy, uses large amounts of data.
• Limits: needs reliable past data, assumes past patterns continue, and sudden changes in tastes or policy are not captured.