Procrustean Art of Backtracking: “Phillips Curve”
For the sake of convenience, we copy the
following from somewhere else:
The
Phillips curve relates the inflation rate (π)
to the unemployment “rate” (u).
Basically, there is dimension aberration: the inflation rate has the time
dimension (T-1, or per peiod) while the unemployment rate, a ratio as
a matter of fact, does not (T0, or at a moment). In order to link
the two, we need a “coefficient” with the time dimension, but finding one may
not be easy. To tell the truth, a rate (T-1) can if ever be bridged
to a ratio (T0) with a third variable (T1) rather than a constant coefficient.
The same holds true for any type of level as
well. For example, a price level at t1 (Pt1) may be connected to another at t2 (Pt2) with the inflation rate
per certain period (π) times a
multiple or a fraction (n) of the
period; that is, Pt2= Pt1∙(1+
n∙π), where n
is a variable in the time dimension (T1).
Not to mention, high unemployment (u) is no more shrinking human power (Δu) than a high price (P) is a rising price (ΔP). In the Phillips curve,
macroeconomists apparently commit the error of dimension aberration. This
mistake is probably caused by the normalization trap due to the granted assumption:
the accounting period to be the unity (1). More specifically, π= (ΔP/ P)/ Δt= ΔP/ P, where Δt= 1: then, the time dimension (T-1) conveniently disappears
and might be legitimately neglected (T0).
Back to basics, both the price level and the
unemployment ratio are conceptually a number available as of a moment.
Therefore, if anything, the change rate in the price level (π) must be compared to the change rate in
the unemployment ratio (Δu/ u).
With that said, could we still find a curve between the two rates as such? If
any, usability of the curve would not be worth the research cost of the
discovery.
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”!
Comments
Post a Comment