Powell Logistigs

I. Double-log

lnY=β0 + β1lnx1+ ε

Ceteris paribus, if x increases by 1%, then, on average, Y increases by β1%.

Ceteris paribus, if education increases by 1%, then, on average, wage will increase by 0.71%.

II. Semilog:

lnY=β0 + β1x1+ ε

Formula for β1:

100[exp(β1)-1]

1. TERR

Coefficient on TERR = 100[exp(-0.21)-1]=100[0.81-1]=-0.19*100=-19.7%

Ceteris paribus, if the house is terraced, then, on average, house price will be lower by 19.7% compared to that of a detached house.

Ceteris paribus, if education increases by 1year, on average, wage will increase by 6.59%.

Ceteris paribus, if age increases by one year, on average, house price will decrease by 0.3%.

Ceteris paribus, if age increases by 1%, then price will decrease, on average, by 0.06%.

Ceteris paribus, if education increases by one year, then, wage will increase, on average, by 5.2%.

Ceteris paribus, if age increases by one year, then, on average, wage will increase by 4.45%.

III. Semilog:

Y=β0 + β1lnx1+ ε

Ceteris paribus, if x increases by 1%, then, on average, Y increases by β1 units (if price, then in Euros).

IV. Polynomial

Y=β0 + β1x1+β2x12+ ε

Formula for turning point:

Wage equation where Experience has a nonlinear impact on wages

X*=I β1/ (2β2)I=I0.298/(2*0.0061)I=24.4

* Maybe few people in the sample have 24 years of experience, so the part of the curve beyond 24 years of Exp can be ignored

* Maybe the impact of experience on wages is biased, due to omitted variables