Thread by @RomanchukBrian

First part of a primer on Minsky’s “Financial Instability Hypothesis.” bondeconomics.com/2019/04/primer…

The FIH has a lot of nasty implications for people who think formal mathematical models are essential for macro.
This comes out more in the second part of my primer, which I am finishing off (plan to publish Sunday).

Why care about the FIH? My argument is that if someone formalises it properly, we can demonstrate formally that formal macro models will fail for predictable reasons.

Since everyone uses physics analogies, the argument is that formal macro models are like perpetual motion machines.

(Control theory has directly applicable analogies, but nobody else has heard of them.)

This is why I mock everyone who asks “where’s the formal model?” or tells me to “build a better model.” Replace “model” with “perpetual motion machine” and go talk to a mech eng prof who teaches thermodynamics and see how well that theory works out.

(At lunch, can add some background.)
I split the FIH into two parts: changes to borrower behaviour, and changes to lender behaviour. There’s a symbiotic relationship between them, but simpler to split. Part I deals with borrowers.

If we look at Minsky’s text, we see he writes about “units”: Ponzi units, etc.
He really was talking about a “unit- (agent-)based model.”

Unfortunately for him, there was no way to deal with agent-based models back when he developed the FIH. (Have fun with the punch cards, pal!) So he was stuck with hand-waving about aggregate model behaviour.

What happens in a unit-based model under conditions of macro stability? The Ponzi units eat the lunch of the conservative units. More profitable, more leverage = faster expansion, etc.

To top it off, the conservative units aren’t dumb. They see the Ponzi units making out like bandits, and so they flip behaviour.

The end result: the Ponzi units are in ur aggregates, eating ur behavioural parameters.

Good luck with your “reduced form” formal model!

Part II of my discussion of Minsky’s Financial Instability Hypothesis is published.
bondeconomics.com/2019/04/primer…

This part focuses on the evolution of the financial system. Has a symbiotic relationship with the better known hedge-speculative-Ponzi unit (agent) scheme.

We can view macro models that model all aggregate flows as being boring video games.
The FIH argues that the financial sector re-writes the rules while the game is underway.

This is toxic for formal models. To apply a model to the real world, you need to fit to data. If the rules change, your parameters change.

If you are approaching macro from a literary perspective, this isn’t too bad: we know capitalism is dynamic. However, the people that insist that only math models are formal, they need to ask how statements about sets can be adapted to “rules changing.”

The reason why this matters is the financial system is evolving so that Ponzi units can get their financing. They will try to accomplish this outside the realms of regulated finance,

So the net result is that the system inevitably builds up toxic concentrations of people who can’t evaluate credit risk lending to entities with no capacity to pay loans back, and this is happening out of sight.

The thing that keeps this dynamic under control is the residual common sense in the system. Ten years after the Financial Crisis, wholesale lenders are still cautious. (Equity investors may be loons, but they don’t matter for the economy.)