Thursday, March 3, 2016

Jason Smith — More like stock-flow inconsistent

One (but by far not the only) tool Post Keynesians tend to use is a stock-flow consistent (SFC) analysis. My original intent was to show how these could be related to information equilibrium, but instead seem to have found a major flaw. I'd like to show that models like these can sneak in implicit assumptions under the guise of "just accounting".
TL;DR* version: Δ in SFC models has units of 1/time and therefore assumes a time scale.

What follows is from Godley-Lavoie "Monetary Economics" [pdf], specifically their model called SIM (for "simplest").… 
Information Transfer Economics
More like stock-flow inconsistent
Jason Smith

*TL;DR = "Too long, didn't read."



11 comments:

Matt Franko said...

"stock-flow consistency" is not code for "just accounting"

Matt Franko said...

"has units of 1/time and therefore assumes a time scale."

1/time is Frequency and therefore assumes a frequency scale....

Tom this guy may be some kind of unqualified nut-job.....

Matt Franko said...

https://en.m.wikipedia.org/wiki/Frequency

NeilW said...

I have to say that I think Jason is going too far now.

Fundamentally his models and ideas fail the Lucas Critique. The point of economics is to inform policy. Policy works on individuals not aggregates. Therefore you need to know what the behaviour of individuals are in response to policy and how that emerges into the aggregate.

In the mainstream they just sum up the individual 'micro-behaviour' which fails the SMD tests.

Now there are standard time units in economics. They are called 'days'. Every day starts, we do something and it ends. In the banking cycle intraday is a different process from inter-day and it is the same within business. Many of the fundamental processes are day based.

Even the FX system has a cut off point of 10pm GMT where everybody has to settle.

So there is no continuous time. There are intra-day chunks and inter-days chunks with closing boundaries.

Rather than solving silly derivative formula in mathematica, etc. isn't it time for a model that actually looks like what happens in the real world?

Ramanan said...

This guy is poor in Mathematics. Doesn't know how to translate between difference and differential equations:

http://www.concertedaction.com/2016/03/04/stock-flow-inconsistent/

Ignacio said...

Minor caveat, string theory is not physics, is met-physics. Much like macroeconomics though...

Unknown said...

I don't see the blogger understand the necessary concepts or history, either. Mybe it isn't a good idea for science students to open up the Instant Economist! kit.

NeilW said...

"This guy is poor in Mathematics."

The majority of the population is poor in mathematics. I wouldn't say my maths is strong enough to do convert from difference equations to differential equations without making mistakes.

But that is why I never use mathematics directly in my work. Because my work is about constructing models *and discussing them with domain experts to see if they are right*. Therefore I have to have something that makes sense to ordinary people.

Economics uses Cod Mathematics for the same reason medieval priests use Cod Latin - so that ordinary people haven't a clue what they are on about and accept the authority because they are able to use Double Delta symbols in an impressive manner.

It's the tool that says "don't talk to me - fear me".

Matt Franko said...

Well imo we have to get the time domain more involved... Ramanan points out pretty well how the Godley/Lovie stuff has some inherent time domain inclusion which comes with the stock-flow consistency but as it is accounting based it is ex post...

Mainstream will say predictive things which never turn out the way they predict and yet we cant defeat them... its because we dont have the sufficient understanding either which is a part of Smith's point...

Ignacio said...

Neil there are two ways of understanding "mathematics" IMO:

One is from the procedural point of view. Is this the techniques are what we care about, people trained on using the 'tools' have an edge over those who don't or don't use them too often. you care about the procedure, where you get an input and then obtain an output through a procedure. This is the 'applied' part of maths, were you care about the procedure. Understanding this is a matter of repetition and experience, more than true insight, knowledge or understanding of the structures involved.

The second is from the objects and relations point of view, being involved in systems design/review you are probably more aware of this. Maths is the story of collections of objects, and the relations between those objects. This is the 'semantic' way of understanding 'maths', the understanding of the structures and relations between those structure.

True mathematical knowledge involved both things, you have to know the procedures, but also understand the structures underlying those structures. You may not be 'good at math', but you obviously have the intuition/knowledge about structures and relationships, not knowing the procedural part is a matter of lack of experience and practice, or dedication to try understand the procedures. There is no "intrinsically bad at math" in your case (I would bet!).

IMO is overstated people being bad at math, the problem is most people did not have enough exposure to the structural and semantical way of learning maths, but most persons with adequate education should not have problem picking up undergrad maths after a few months of researching and trying to understand it, if their cognitive style lends towards structural and systemic analysis.

Tom Hickey said...

I've posted a link to Ramanan's reply to Jason.