Thread by @MacRoweNick

A thread on why using Fiscal Policy (not Monetary Policy) to control Y (and so control inflation) is a Bad Idea.
This is a followup to my previous thread (RTd below).
It's also going to be a bit harder.
Sorry.
Praying I don't mess it up. 1/n

https://twitter.com/MacRoweNick/status/1149301819098705920

To Keep It Simple (for me) I'm only going to consider a Closed Economy.
Y=C+I+G
Remember (from previous thread):
We want 2 things:
1. We want the right *Total* Y (fullish employment, keeping inflation controlled)
2. We also want the right *Mix* of Y (between C & I & G)

If we had the True Model of the economy, this would be easy.
Any mathematician could solve for the endogenous {C,I,G} given the policy instruments {G,T,R}.
Or solve it in reverse, for the settings of the policy instruments needed to get the right Total and right Mix.

If we asked the mathematician whether we should use Monetary Policy (R) or Fiscal Policy (G &/or T) to control Total Y(=C+I+G) we would get a funny look, and a lecture on why simultaneous equations systems aren't like that.
Everything (usually) depends on everything else.

But we don't know the True Model of the Economy.
And even what we *think* we know is too much for any One Mind.
(Hayek "Use of Knowledge" etc.)
So we:
1. Decentralise decisions, and;
2. "Gradient Climb"
to try to do the best we can.

I cannot solve the True Model of my Car to calculate where to set the gas pedal to get the right (100km/h) speed, given variable hills & headwinds.
But I can "Gradient Climb" to get it roughly right:
to increase speed press down on the gas;
to decrease speed ease up.

So the right question to ask is:
How should we assign targets to instruments (decentralise responsibility for "solving" this simultaneous equations problem) so that "Gradient Climbing" can work reasonably well?

(It pains me to admit this:)

The "Macro" part (getting Total Y right) of this assignment question is easy for everyone (yes, there's expectations, and long & variable lags, etc., but...)

It's the "Micro" part (getting the Mix of Y right) that's harder for Central Bankers.

"Macro" part:
If Total Y is too high (so inflation might get too high), and Total Y is your responsibility, you can:
Cut G (if you control G)
Raise T (if you control T)
Raise R (if you control R)

It's easy to Gradient Climb, whoever is assigned Total Y.

But getting the Mix right (dividing a given Y into components C, I, and G) is easier for the Fiscal Authority (with instruments G and T) than for the Central Bank (with instrument R).

For example, it would be daft to assign Total Y to instrument G, and the Mix G/Y to instrument R.
If it thought G/Y was too low, the Central Bank would need to raise R, which would reduce Y, which would cause the Fiscal authority to respond by raising G to get Y back up again.

It's much simpler for both to swap the assignments, so the Mix G/Y is assigned to G itself.
So let's suppose we do this.
Which way round should we assign T and R to Total Y and Mix C/I?

Suppose we assigned the Mix C/I to R, and Total Y to T.
If the Central Bank thought C/I was too high, what should it do?

The answer isn't obvious (to me).........

I *think* the CB should cut R, because I *think* that would cause a bigger percentage increase in I than in C, and induce the Fiscal authority to raise T (to offset the rise in Y), which I *think* would cause a bigger percentage decrease in C than in I (depending on *which* Tax).

But it would be simpler if we assigned targets to instruments the other way around.

Bottom Line:
Macroeconomists spend too much time arguing whether Monetary or Fiscal Policy has an *Absolute* Advantage in getting the Macro question (Total Y) right.
What matters instead is whether Monetary or Fiscal Policy has a *Comparative* Advantage in Macro vs Micro.