Home > Main > MMT Modeling

MMT Modeling

September 26, 2011

We crazy MMTers are always saying “You need to follow the accounting!” to professional economists.  They say they do.  But I don’t think they do.

In a very deep way, any economy is a totally closed system.  All of the books and numbers add up somehow. Every balance sheet can be reconciled with every other balance sheet in the economy, and every statement of cash flows, and every income statement – they can all be reconciled if you could just take the time to do it.

There aren’t any gaps.  It’s all filled in.

This premise drives MMT.  All of this money stuff adds up – so it can be cross referenced against other parts of the economy.  Something interesting to note is that it also adds up in the future too, and added up in the past.

The cool thing about this type of thinking is that you can create models of the economy using the “it all adds up” idea as an end point at any given moment, and then give a bit of an impluse to parts of the economy, and move forwards in time.

This kind of thinking is not really that new.  You can go and download simple predator/prey models, and find out what happens in closed economies over time.   There are lots of things like traffic modeling that has taken huge steps forward over the last decade.

All of this is due to thinking about systems as having inputs that impact the model, and there is an interaction between the parts that “adds up”.

In an MMT world, a Steve Keen world, a Wynne Godley world, this type of modeling is very possible.   We have computers that can keep track of lots of numbers, and are easily able to apply rules to how the numbers change, given inputs to the economy.

These models are pretty hard to make in some ways, but very easy to understand.   You know how the parts fit together and what they must add up to be.

The new MMT Blogger Fred Decker is creating models like this, and they seem to be pretty cool.  I don’t have the software to be able to use these models yet, or the time until probably next year.

But I am not surprised that these models are coming and being built.  And the great, great thing about these models is that many, many of the numbers that go into the models are observable and measurable in something like real time.

So we can see not just the inputs to what goes into the model, these models can also be calibrated against real world data in something like complete manner.


Categories: Main Tags:
%d bloggers like this: