Minsky

flattr this!

Minsky: SourceForge Project of the Month for January 2014

MinskyLorenzScreenshot
We're delighted that SourceForge has seen fit to make Minsky it's Project of the Month. Minsky is both a general purpose system dynamics program, and the first such program specifically designed to support the building of monetary models of the economy.

Our design objective has been to make it as intuitive as possible to use, while supporting various advanced features--such as maintaining consistency between multiple monetary flow tables--in as seamless a way as possible. We hope that users will find it useful, and we would be delighted to get coding assistance from developers, user assistance in bug testing and building the help system, etc.

Dr. Russell Standish and I have been working on Minsky now for almost two years now--first using a $125K from INET's Spring 2011 grant round and then another $80K from a successful Kickstarter campaign: Russell as builder (coding in C++ and Tcl/Tk) and me as architect (playing with each release, spotting bugs and suggesting features). It's been a part-time endeavor: Russell, as a contract programmer, has to keep more than one iron in the fire, while I have a fair few balls in the air myself. Russell has put in about 2000 hours of coding over that time.

That amount of coding is chicken feed compared to the millions of hours that have gone into a standard commercial product like Microsoft Word or Excel, so Minsky is still rough at the edges, and lacks a number of features that you would expect in a commercial package. But there is enough core functionality and stability in place for a first major release--the "Mun" version which can now be downloaded from SourceForge (major iterations of Minsky are named after major thinkers in economics--we started with Aristotle).

Minsky Endogenous Money Simulation

There is still only a rudimentary and incomplete Help file (did I say I had too many balls in the air?), so to make up for it I've recorded the following set of 15 videos that cover most of the features of the program. They're all up on YouTube (where the number is the series changes from 8 to 11 to 15...), but here they are for easy reference, with some notes about what each video covers. All the Minsky (mky) files used in the presentations are in this Zip file.

So if you're into economic modeling and you'd like to develop dynamic monetary models, Minsky now has enough features to be worth using--and it won't face the limitations that apply to more conventional system dynamics programs, like Stella, Vensim, Vissim, etc., that make it inherently challenging to model the financial system. With Minsky and its Godley Tables, monetary modeling is a cinch (and it will get easier as we add more capabilities to the program). So please, download the program, join our beta-testing group, and start building monetary macroeconomics.

If you're a developer with time to devote to this project, immediate needs include:

  • Sprucing up plots;
  • Copy and paste support files;
  • Multiple files open at once;
  • Building a LaTeX viewer into Minsky;

Medium term needs include:

  • A direct LaTeX editor with seamless transfer between the equation pane and the canvas pane;
  • Adding Social Accounting Matrix support for equation consistency checking;

The long term objectives include:

  • Transform a "scalar" economic model into a "vector" with monetary and physical input-output dynamics with one click;
  • Multiple economy support with financial and physical flows between economies, currency conversion, etc.;
  • Data importing, manipulation and nonlinear parameter fitting

Minsky 1.0 Demo #1 Basics And Godley Table

Installing Minsky from a ZIP file, the canvas, wiring to build equations, installing a graph, graphing a basic function like sin(t), outputting the equations to LaTeX (which you can import into MathType in Word), and a basic "endogenous money" Godley Table.

Minsky Files: sine.mky (using Minsky to plot a sine wave); Godley.mky (a basic purely monetary model, with no simulation--just the double-entry bookkeeping table).

Minsky 1.0 Demo #2 Godley Simulation

Using a Godley Table to define a numerical simulation: right-click to copy flows and variables ("stocks"), defining monetary equations, and running a simulation.

Minsky file: GodleySim01.mky (the same model as in the previous video with definitions of flows added so that it can be numerically simulated).

Minsky 1.0 Demo #3 Godley-Goodwin Integration

Using time lags, entering Greek characters and subscripts, building a dynamic model using the flowchart part of the program, using integration blocks in dynamic models, making x-y graphs as well as graphs of variables against time.

Minsky file: GodleySim02.mky (the same model as in the previous video, but using time constants rather than arbitrary numbers for system parameters); GodleyGoodwin01 and GodleyGoodwinSim01 (adding a Goodwin model of production using the flowchart aspects of the program, and adding XY charts, slowing down the simulation speed)

Minsky 1.0 Demo #4 Godley-Goodwin Integration

Starting to integrate a monetary and a physical model, using sliders to vary parameters during a simulation.

Minsky files: GodleyGoodwinSimInt01 (beginning to integrate the two models: making the wage a monetary phenomenon)

Minsky 1.0 Demo #5 Godley Goodwin Integration #2

Doing it right: starting by modifying the real cyclical model to include monetary variables.

Minsky files: GodleyGoodwinSimInt01C (adding Price to the Goodwin model before attempting the integration).

Minsky 1.0 Demo #6 Godley-Goodwin Integration

Continuing the process.

Minsky files: GodleyGoodwinSimInt01D (completing the integration)

Minsky 1.0 Demo #7 Godley-Goodwin Integration #4

Continuing the process.

Minsky files include GodleyGoodwinSimInt03, where I add a linear investment function to the linear wage change function that is the core of the Goodwin model.

Minsky 1.0 Demo #8 Godley-Goodwin Integration #5

Getting cycles in the monetary economy as well as the physical. Introducing an investment function as well as a wage change function. Showing the debt-deflation model.

Minsky files include DebtDeflation02, which is the monetary version of my 1995 debt-crisis model in "Finance and Eco­nomic Breakdown

Min­sky 1.0 Demo #09 Mul­ti­ple God­ley Tables

Com­bin­ing MMT and MCT with two God­ley Tables, one for the Cen­tral Bank, the other for Pri­vate Banks. Minsky’s built-in logic to main­tain con­sis­tency between assets and lia­bil­i­ties across mul­ti­ple God­ley Tables. A poten­tially sur­pris­ing result about the rela­tion­ship between the change in pri­vate debt and the change in pub­lic debt.

Min­sky files: MinskyOnePointZeroDemo09MultipleGodleyTables (com­bines MMT and MCT and shows that the change in pri­vate debt and the change in pub­lic debt are not of equal and oppo­site mag­ni­tudes in a closed economy).

Min­sky 1.0 Demo #10 Zooming

Group­ing enti­ties, zoom­ing in to see a group, edit­ing a group, using the Lorenz Strange Attrac­tor as first exam­ple, and my model of debt-deflation as the second.

Min­sky 1.0 Demo #11 Bugs

Bugs and how to cope with them.

Min­sky 1.0 Demo #12 Lorenz model

Using the flow­chart side of the pro­gram to build the famous Lorenz model. Chang­ing the Runge-Kutta para­me­ters to get a smoother sim­u­la­tion. Set­ting ranges on a graph.

Min­sky 1.0 Demo13 # Lorenz model with Sliders

Using slid­ers to vary para­me­ters in a model while it runs.

Min­sky files: Vari­a­tions on Lorenz, espe­cially LorenzWithSliders02.

Min­sky 1.0 Demo #14 Competitors

How does Min­sky stack up as a sys­tem dynam­ics pro­gram com­pared to the estab­lished play­ers Simulink, Ven­sim, and Vis­sim? It doesn’t have the power of these estab­lished pro­grams of course–matching the fea­ture set of even Ven­sim would require tens thou­sands more hours of pro­gram­ming time, let alone that of Vis­sim or the mar­ket leader Simulink. But on a user-friendliness scale, I think it’s easy to see that Min­sky is much eas­ier to use than Ven­sim, sig­nif­i­cantly eas­ier than Simulink, and within reach of Vissim–though it lacks some basic edit­ing features.

Min­sky 1.0 Demo #15 Why God­ley Tables?

This is the clincher. If you want to model mon­e­tary dynam­ics, there is no other pro­gram that offers any­thing like the God­ley Table for mod­el­ing the double-entry flows that char­ac­ter­ize the finan­cial sys­tem. A flow­chart is an inher­ently weak tool for build­ing a mon­e­tary model..

19 Responses to Minsky

  1. Laszlo Orosz says:

    Hello guys, kind of new here. Can any­body help me, how to down­load Min­sky?
    I’ve played around with QED and the first Min­sky ver­sion from source­forge, but I’m unable to get the recent ver­sion.
    Thanks

  2. Charlotte van Dixhoorn says:

    Hello all,

    Very inter­est­ing mod­els. How­ever, does any­one also get the “Error: No input for port 147″ when run­ning the exam­ple model:

    4MonetaryMinskyModelLessUnstableStart

    Cheers

  3. Steve Keen says:

    There’s a new ver­sion Charlotte:

    http://www.debtdeflation.com/blogs/wp-content/uploads/2012/08/Minsky.1.0.D99-distrib.tar.gz

    I’m also about to post a set of videos on how to use it, and a call for beta testers. Lovely to have your help on that already!

  4. Graham Wideman says:

    @Steve and Char­lotte: I just down­loaded both the ver­sion in the “Min­sky for Win­dows link, and the one linked to in your Aug 9 post. They are both named Min­sky 1.0 D099, and both pro­duce the error men­tioned by Char­lotte.
    Look­ing for­ward to user doc­u­men­ta­tion, and also docs for the mky xml file for­mat. I will be happy to be involved in beta testing.

  5. Russell Standish says:

    @Graham & Char­lotte: Its vital that these bug reports are entered inot Min­sky tick­ets so that they get a chance to be looked at:

    I did a quick search, but can’t see whether your par­tic­u­lar issue has been reported or fixed, though.

    As for beta test­ing, I should have a new build for beta test­ing next week, but you need to sub­scribe to the beta testers mail­ing list to get noti­fi­ca­tion of new builds for testing.

    Start­ing next month, I will be work­ing on for­mal­is­ing the mky for­mat as an XML DTD. This is impor­tant, as the exist­ing for­mat is imped­ing the program’s evo­lu­tion. The descrip­tions will be checked into the source code repos­i­tory, and posted to the Min­sky wiki.

    As for user doc­u­men­ta­tion, the idea is for the soft­ware to be quite intu­itive. Feed­back on fail­ings in this regard will be appre­ci­ated! More impor­tant doc­u­men­ta­tion is more a ped­a­gog­i­cal one, show­ing how to cre­ate eco­nomic and other mod­els in Min­sky. Steve’s videos are a good start.

  6. Graham Hodgson says:

    I’m puz­zled by the spec­i­fi­ca­tion of stock vari­ables, which are gen­er­ated when col­umn head­ings of the God­ley Table are named. Stock vari­ables don’t have input ports. How are stock lev­els replenished?

  7. Steve Keen says:

    By mak­ing row entries Gra­ham. Add an entry for loans from the bank to a bor­rower for exam­ple, or check some of the sam­ple files.

  8. Russell Standish says:

    A more recent ver­sion of Min­sky is avail­able from Source­Forge. Cur­rently, the released ver­sion is 1.D9, also known as “Aristotle”.

  9. Kai Wakerman Powell says:

    Dear Steve,

    Any word on what’s hap­pened to this? I mean… take my money already.

  10. Kai Wakerman Powell says:

    Addi­tion­ally, will you put out an email when the cam­paign opens?

  11. Steve Keen says:

    Thanks Kai,

    At the moment I intend launch­ing the cam­paign on Mon­day next week–January 28. I had planned to start ear­lier but–to put it mildly–I have had some dis­trac­tions recently. I have some work to do on the cam­paign, which I hope to do in the morn­ings next week before attend­ing the Aus­tralian Open dur­ing the after­noons and evenings this week. I should get them done in time for a Mon­day launch.

    And I will def­i­nitely email, blog and twit­ter about it when it goes live!

  12. Steve Roth says:

    Hey Steve you might want to update this post.

  13. Pingback: Asymptosis » Reveal Your Preferences! Show Your Support for Accounting-Based Economic Modeling

  14. Bill Walters says:

    Hello, unfor­tu­nately I missed the Kick­starter cam­paign. Please advise how to con­tribute after the fact if possible.

  15. Pingback: Steve Keen: Is QE Quantitatively Irrelevant? « naked capitalism

  16. Pingback: How QE works and what this means for asset prices and credit | Credit Writedowns

  17. Pingback: Steven Keen – “I will be wrong on house prices”

  18. Pingback: Der Geld-Illusions-Schock der Bank of England | Lars Schall

  19. Steve, this is a minor spur pro­posal and/or an invi­ta­tion toward a col­lab­o­ra­tion. I believe that it would a valu­able edu­ca­tional tool to have a “com­mu­nity cur­rency” pro­gram based upon the MMT/Endogenuous Econ model woulld be a wor­thy project. In essence it would take the UMKC Bucka­roos model and then con­vert it into a mod­el­ing by which peo­ple can expe­ri­ence first have the effects and dynam­ics of a eco­nom­i­cally lit­er­ate and func­tional eco­nomic model. An alter­na­tive is to involve the back­ers in the crowd source financ­ing of Min­sky to this much smaller scale bit of mod­el­ing. My email is: ideasinc@ee.net thanks, Tadit

Leave a Reply