Procrustean Art of Backtracking: “Okun’s Law”
For the sake of convenience, we copy the
following from somewhere else:
[The
policy “guidance” called Oku’s law] is sometimes described as g= 3%- 2Δu, where “g” is the percentage increase in GDP [(ΔY/ Y) over two consecutive
periods, and Δu is the percentage
decrease in human power (Δu≒ - ΔL/L)
between two time points apart by a year. The linkage can be justified only when
the unemployment ratios at the two points effectively represent the labor input
in the respective period. This assumption is counterfactual if in a “cyclical”
economy.
Moreover, this law’s prediction that the
percentage gap in GDP is twice (2) as big as the gap in the labor input is the polar
opposite of the Solow model’s spirit for the long run: the law of diminishing
marginal product of labor puts the number small than the unity (1). Incidentally,
the elasticity coefficient is hardly
larger than 0.8 for the short or long run if the Cobb-Douglas function is any
guide.[1]
¿Y
qué? Well, for the sake of verifying a
hypothesis you might cook the data with handy assumptions and metric-free mathematical
equations. Who knew “virtual” equaled “real”!
[1]
Assume
that the human power (L), or all
employees, work for the same hours per
annum and that the labor force (F)
grows at n% PA. With the unemployment
ratio u= (F- L)/F, or L= F∙(1– u),
the annual growth in the labor input (ΔL/L)= (F/L)∙(n–
Δu– n∙u- n∙Δu)≒ (F/L)∙(n –Δu)=
[1/(1- u)]∙(n– Δu). Now, apply this to the
Cobb-Douglas function, Y= A∙Kα∙L1-α, and we get: g= ΔY/Y = [ΔA/A+
α∙(ΔK/K)]+ (1- α)∙(ΔL/L)= [ΔA/A+ α∙(ΔK/K) + n/(1-
u)]– [(1- α)/(1- u)]∙Δu, The elasticity (1- α)/(1- u) is probably smaller than 0.8 where α is often thought to be
0.3 and u rarely larger than 0.1,
where, with n≒
0
over a year or less, ΔL/L= - [1/(1-u) ]∙Δu≒
-Δu.
Comments
Post a Comment