flattr this!

Minsky: Latest News

Russell Standish has been working on Minsky thanks to the extra funds raised by the donate button below.

Keep Russell Standish on the Minsky Project

Other Amount:

Your Email Address (and comment if you wish to add one) :

He has produced a beta which is a substantial improvement over the release version currently on the Sourceforge home page. It adds:

  • an equation viewer--so you can design a model and instantly see the equations behind it from within Minsky (as well as export them to LaTeX);
  • A differential operator, so you can calculate rates of change (and in typical Russell style, this is no ordinary numerical approximation but a full symbolic differential system, so the approximation problems that apply with (to my knowledge) all other system dynamics program differential operators don't apply to Minsky);
  • A large num­ber of bug fixes.

With the caveat that beta ver­sions almost always have bugs, this ver­sion is very close to a new release can­di­date and I would rec­om­mend using it rather than the sta­ble release:

Min­sky Beta Page

Keep Rus­sell Stan­dish on the Min­sky Project

Other Amount:

Your Email Address (and com­ment if you wish to add one) :

Keep Rus­sell Stan­dish cod­ing Minsky

Min­sky has been devel­oped thanks to two grants: an ini­tial US$128,000 grant from INET, and crowd­fund­ing of $78,000 from Kick­starter. With other amounts from indi­vid­ual donor, that has funded about 2500 hours of pro­gram­ming by the bril­liant pro­gram­mer behind Min­sky, Pro­fes­sor Rus­sell Stan­dish. I will be apply­ing for other grants now that I am based in the UK, but in the mean­time more fund­ing is needed to keep the devel­op­ment rolling and to reserve Russell’s skills for Min­sky. So please donate via the Pay­pal link above: the funds go directly to Rus­sell, and the money raised so far–about $8,000–has helped pro­duce the lat­est beta, which

Source­Forge Project of the Month for Jan­u­ary 2014

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..

31 Responses to Minsky

  1. Steve Keen says:

    That fea­ture has been superceded. To move an ele­ment now, sim­ply click on it and drag. To wire one to another, click on the input or out­put ports. This bit is kludgy right now, but as you move the mouse over an object you will see cir­cles appear on them. Click­ing in a cir­cle drags an arrow which you can then attach to another object.

  2. Pingback: digi_owl comments on “What’s wrong with finance” | Exploding Ads

  3. Steve, would you be will­ing to endorse a much sim­pler ver­sion of Min­sky that could serve as a pro­gram sup­port­ing a “com­mu­nity cur­rency. What you have accom­plished so far would occupy a social space where peo­ple con­cerned about the effects of stupid/neo-classical and even lower forms of fis­cal and eco­nomic assump­tions. Part of the point, is to have avail­able a rel­a­tively well informed fiscal/economic model, tat would would replace the non­sense of the LETS and TIME BANK mod­els, which are just debt based, and tend to be pro­moted as a “com­mu­nity” cen­tered alter­na­tive to the neo-classical ide­ol­ogy though absent mea­sure­able fis­cal or eco­nomic lit­er­acy value.

    In prac­tice the dis­tri­b­u­tion and uncrit­i­cal adop­tion of all sorts of blather, besides sim­ply occu­py­ing and divert­ing energy and occu­py­ing com­mu­nal inter­est in effect divert­ing eco­nomic lit­er­acy. The mod­i­fied “Min­sky” would serve as a pre-collegiate eco­nomic lit­er­acy process, to teach by doing toward higher lev­els of fiscal/economic lit­er­acy. Now that you are at Kingston U. you are pos­si­bly aware of the “Tran­si­tion Town” move­ment in Eng­land and on the con­ti­nent to which this would imme­di­ately appeal.

    I have been pro­mot­ing and pre­sent­ing related to pub­lic fis­cal and eco­nomic lit­er­acy for a num­ber of years, and this sort of learn­ing by doing would be a great resource. The UMKC con­tin­gent gen­er­ally has ignored the value of pub­lic edu­ca­tion other than through an aca­d­e­mic con­text, which is at least lim­it­ing if not also very shortsighted.

    Although the Bucka­roos sys­tem has been used for pos­si­bly two decades now, and the Deni­son Unv., and the DVD/Denison Vol­un­teer Dol­lars pro­gram (thanks to Dr. Fad­hel Kaboub{Phd from UMKC) both are oper­at­ing based upon a spread sheet util­ity. Also there is a small rev­enue stream that would be avail­able through host­ing such a utility.

    The level of communal/community dis­course needs to be raised beyond the ivied ram­parts as well. The present Sen­a­tor Sanders pres­i­den­tial cam­paign is a mea­sure of peo­ple want­ing social and eco­nomic change, though Sanders is so far address­ing the pos­si­ble against the adversely self serv­ing ide­ol­ogy of neo-liberalism. Stephanie Kel­ton sev­ing in Wash­ing­ton as the advis­ing econ­o­mist to the minor­ity por­tion of the US Sen­ate Bud­get Cmte is a pos­i­tive, but not one that will engage as a sub­stan­tive alter­na­tive except in Wash­ing­ton DC as the the­o­ret­i­cal affir­ma­tion of pro-community fis­cal poli­cies and pro­pos­als and some fis­cal and eco­nomic lit­er­acy at that level. ideasinc@ee.net Tadit

  4. arijan says:

    As is, pro­gram barely works and is as such com­pletely unus­able. I could not keep it up run­ning for more than 5 min­utes. After an hour of try­ing I finally unin­stalled it even though I would des­per­ately like to get my hands on a decent tool to model finan­cial flows.

    All I can say is you could have spent €100k more wisely, but that’s my opinion.

    Func­tion­al­ity wise I would find it use­ful if the pro­gram would come with some basic eco­nomic tem­plates which would be explained in more detail. Zip file con­tains no exam­ples, and the ones on github have no doc­u­men­ta­tion. If this tool is to be of any use for begin­ners, exten­sive help file is prob­a­bly more needed than an elab­o­rate list of mod­ules for data processing.

    Which brings me to the sec­ond point. I’m sorry it’s slightly off topic, but the above is such a per­fect exam­ple it’s impos­si­ble for me to resist.

    Namely, it is my con­tention that you’ve mas­sively over­paid a project that ulti­mately does not work. Now if you have had paid for it out of your own pocket you would prob­a­bly be upset as you could not mar­ket and or sell it prop­erly. As it was paid for by the grant your first instinct is to apply for another grant. It’s not a lot of money involved, but it’s an illustration.

    You see where I’m get­ting at? It this was a busi­ness, it would fail.

    That seems to be the case with most gov­ern­ment financed projects. Just a recent exam­ple from the coun­try I come from: gov­ern­ment over­paid a coal fired power plant 3 times over only to find that now it is finally fin­ished it will start to loose money every day it oper­ates. (about €4M/month) Sure it was a huge stim­u­lus for local econ­omy, but it came at a huge cost and loans acquired.

    This type of behav­ior is what I believe got my coun­try into trou­ble and imho got Greece and a host of other coun­tries into trou­ble as well. So your con­tention that more gov­ern­ment spend­ing should be in order in times of cri­sis rather that aus­ter­ity is prob­a­bly going to make the sit­u­a­tion worse as this spend­ing is never ratio­nal spend­ing. To this extent Peter Schiff, as much as you seem to dis­like him, is right.

    As long as the sums are small, such as €100k in the exam­ple above, this is non issue. How­ever €1.5B was spent on a power plant I wrote about. And this is not the only exam­ple, hence the prob­lem fac­ing these coun­tries: falling GDP, huge exter­nal deficit, tax hikes and polit­i­cal instability.

    More loans to boost spend­ing will not solve the under­ly­ing prob­lem, nor will debt jubilee, nor will pen­nies from the sky (or rather Euro­copter if only we had a Fed equiv­a­lent here). What needs to change is a social con­tract between peo­ple to form new kind of state, “new deal” so to speak, but that’s just to plain dif­fi­cult as peo­ple are unwill­ing to change and are unwill­ing to face real­ity and are unwill­ing to learn. And btw “new deal” from US was not the exam­ple of what’s needed, but the name is too good not to be reused.

    Any­way this is what I wanted to com­ment on, and I would appre­ci­ate a reply if you would be so kind to indulge me with one.


  5. Steve Keen says:

    Part of what you say here is jus­ti­fied: I’m the first to admit that Min­sky is incom­plete and has bugs, and your com­ment about need­ing sam­ple files has moti­vated a new ticket on Source­forge that I hope will be com­pleted in the next cou­ple of weeks: we’ll embed a large num­ber of sam­ple files in the pro­gram to guide new users.

    But your inter­pre­ta­tion of why it has prob­lems is com­pletely false. For starters, not a cent of gov­ern­ment money has gone into Min­sky. It has been entirely funded (a) by a too-small grant from INET (www.ineteconomics.org) and sec­ondly by pub­lic dona­tions. I asked for $250K from INET to fund a ground-up devel­op­ment of Min­sky in C++ or the like; they have me $128K but spec­i­fied the C++ imple­men­ta­tion (as opposed to the Math­e­mat­ica add-on that I said I could get done for $125K) and required it to be Open Source. The lat­ter was of course fine by me, but it means it’s been released far ahead of when a com­mer­cial prod­uct would be released–and hence many of the prob­lems for first time users.

    Gov­ern­ment fund­ing used to exist for “blue sky projects” like Min­sky, but the relent­less cut­backs to gov­ern­ment spend­ing over the years have vir­tu­ally elim­i­nated that source of fund­ing, and robbed gov­ern­ments of the exper­tise needed to do such things properly–which may be an issue in the exam­ple you give of your own econ­omy. Five decades ago, gov­ern­ment depart­ments had engi­neers, sci­en­tists, etc., in them who could eval­u­ate projects prop­erly. Now at best they’ve got accoun­tants, and worst econ­o­mists, who couldn’t read an engi­neer­ing spec­i­fi­ca­tion if it mugged them in a dark alleyway.

    Con­se­quently the kind of intel­li­gent blue sky stuff that gov­ern­ments used to fund–like NASA in its heyday–don’t get fund­ing any more, or are done incom­pe­tently as you’ve iden­ti­fied with your own coun­try. Then projects like Min­sky fall between the stools of what gov­ern­ments would once have funded, and what com­mer­cial ven­tures might fund if they could see a poten­tial profit.

    The real weak­nesses in Min­sky come from a lack of fund­ing com­pared to what a com­mer­cial enter­prise (startup or oth­er­wise) would have put into it. You wouldn’t attempt a com­mer­cial prod­uct with­out at least 3 developers–one lead, two sup­port, and alter­nat­ing devel­op­ment and bug-catching roles between them. We’ve just had one, with me being a part-time design leader since I am so busy with numer­ous other tasks.

    Would Min­sky have been devel­oped by a com­mer­cial enter­prise? Unlikely, since there are numer­ous func­tional com­peti­tors in the prod­uct space already (Simulink, Vis­sim, Ven­sim, etc.). As it hap­pens Min­sky improves on all of them in a num­ber of ways, some of which I detailed here (https://sourceforge.net/blog/january-2014-potm/) when it won SourceForge’s Prod­uct of the Month award in Jan­u­ary 2014. But these improve­ments alone wouldn’t have been enough to per­suade a profit-oriented enter­prise to give devel­op­ing it a crack.

    What it does that no other pro­gram does is enable mon­e­tary flows to be mod­elled eas­ily. We’ve only just devel­oped the start of that func­tion­al­ity, but I demon­strate it in numer­ous videos on YouTube (https://www.youtube.com/channel/UCM1ubsbE-tG9ru61mc3zX8A). Now once it’s up and run­ning, I am get­ting approaches from groups that want to com­mer­cialise this. But not before.

    On usabil­ity, yes there are many issues–which if I had more fund­ing I would have had solved ages ago. But since I haven’t, I’ve con­tin­ued to work with the pro­gram with my own obvi­ously very com­plete knowl­edge of what it can do and how to work around its prob­lems. Given that knowl­edge, it does things eas­ily that no other pro­gram enables me to do–and I have every com­mer­cial pro­gram you can think of for such jobs. But what Min­sky doesn’t have is the help sys­tems and sup­port files to let a neo­phyte user get up to speed with it.

    I also don’t know the extent to which your dif­fi­cul­ties arose from Minsky’s own weak­nesses, or your own level of knowl­edge of the sys­tem dynam­ics fam­ily of soft­ware to which it belongs. A lot of the “why is Min­sky so hard to use?” feed­back that I get, when I track it down, comes from people’s lack of knowl­edge of the very estab­lished flow­chart equa­tion devel­op­ment par­a­digm that has dom­i­nated engi­neer­ing for 3 decades now. But eco­nom­ics stu­dents often don’t even know that this par­a­digm exists. So I face a dou­ble chal­lenge in teach­ing them how to use Minsky.

  6. Bill K. says:

    Per­haps it would be a good idea to plunk Min­sky into some Uni­ver­sity as a course. I am a lit­tle biased but the Uni­ver­sity
    of Water­loo, On. would be a good choice. You could lend your per­sona to the eco­nom­ics dept. as Mr. Hawk­ing has to the physics
    dept. I can’t guar­an­tee you will have a build­ing named in your
    hon­our, or some fund­ing from out Black­berry cre­ators. How­ever,
    you may get a very capa­ble group that can put in some brute force.

Leave a Reply