Log-Level Model

ln⁡y\ln y regressed on xx: a one-unit rise in xx changes yy by about 100β^1100\hat\beta_1 percent.

When to use

Use it when the outcome moves in percentages per unit of xx; its marginal effect is β^1y\hat\beta_1 y and its rate of change is β^1\hat\beta_1. Compare with the Log-Log Model, where the slope is an elasticity.

See Lecture 3 and Lecture 13.

Appears in