Olivier Blan­chard, Equi­lib­rium, Com­plex­ity, And The Future Of Macro­eco­nom­ics

Flattr this!

I have observed and appre­ci­ated Olivier Blanchard’s intel­lec­tual jour­ney over the last decade. It began in August 2008, with what must be regarded as one of the worst-timed papers in the his­tory of eco­nom­ics. In a sur­vey of macro­eco­nom­ics enti­tled “The State of Macro”, he con­cluded, one year after the finan­cial cri­sis began, that “The state of Macro is good” (Blan­chard, 2008). How­ever, Blan­chard did not remain locked into that posi­tion, and he had the rare intel­lec­tual courage to say so in pub­lic and in aca­d­e­mic papers. His most recent post, before the one I am respond­ing to today (“Fur­ther Thoughts on DSGE Mod­els: What we agree on and what we do not”), stated that, far from the state of macro being good:

There are many rea­sons to dis­like cur­rent DSGE mod­els. First: They are based on unap­peal­ing assump­tions. Not just sim­pli­fy­ing assump­tions, as any model must, but assump­tions pro­foundly at odds with what we know about con­sumers and firms. (Blan­chard, “Do DSGE Mod­els Have a Future?”, August 2016)

I have com­mented on sev­eral of Olivier’s papers on the progress between “The State of Macro” and “Do DSGE Mod­els Have a Future?”, and he ref­er­ences one of my com­ments (“The need for plu­ral­ism in eco­nom­ics”) in this most recent piece (as well as oth­ers by  Narayana Kocher­lakotaSimon Wren-LewisPaul RomerAnton KorinekPaul Krug­manNoah SmithRoger Farmer, and Brad Delong). So he includes me in his sum­mary of the dis­cus­sion, and this neces­si­tates a reply because—while again I appre­ci­ate his engagement—I dis­agree with his sum­mary from its very first point.

Click here to read the rest of this post.

About Steve Keen

I am Professor of Economics and Head of Economics, History and Politics at Kingston University London, and a long time critic of conventional economic thought. As well as attacking mainstream thought in Debunking Economics, I am also developing an alternative dynamic approach to economic modelling. The key issue I am tackling here is the prospect for a debt-deflation on the back of the enormous private debts accumulated globally, and our very low rate of inflation.
Bookmark the permalink.
  • twowith­inthree­thati­sone

    What if the actual solu­tion was an inte­gra­tion of the
    truths of both equi­lib­rium and dis­e­qui­lib­rium the­o­ries, a true inte­gra­tion of micro­foun­da­tions and macro-eco­nom­ics, an inte­gra­tion of Lorenz’s three fac­tors, a new eco­nomic phi­los­o­phy whose poli­cies are based on the con­cept of grace as in mon­e­tary gift­ing whose anatomy is described as an inte­grated dual­ity within an inte­gra­tive trin­ity-unity and for­mu­lated thusly:

    [ ( A x B ) –> G/C ] where ( A x B) is an inte­grated dual­ity and 

    every­thing within [ ] is an inte­gra­tive and dynamic trin­ity-unity process in the tem­po­ral uni­verse

    It’s a sim­pli­fi­ca­tion like Lorenz’s three inte­grated fac­tors, but if also applied to addi­tional fac­tors like the ones you men­tion or other cur­rently unper­ceived real­i­ties like the cost infla­tion­ary nature of mod­ern economies due to the cost account­ing con­ven­tion that all costs go into prices being enforced despite the fact that the rate of flow of total costs increas­ingly exceeds the rate of flow of total indi­vid­ual incomes.

  • twowith­inthree­thati­sone

    .…it would fit within the for­mula and result in the inte­gra­tive trin­ity-unity.

  • twowith­inthree­thati­sone

    What I refer to as The Cos­mic Code can describe basic forces like elec­tro-mag­net­ism

    [ ( + charge x — charge) –> electricity/electro-magnetism ]

    The Hegelian dialec­tic

    [ (the­sis x antithe­sis) –> syn­the­sis ]

    the human state of Grace

    [ (Space x Time ) –> Self Aware­ness ]

    the res­o­lu­tion of any prob­lem

    [ ( effort x counter effort ) –> res­o­lu­tion via inte­gra­tion ]
    [ ( idea x counter idea ) –> res­o­lu­tion via inte­gra­tion ]

    (An actual inte­gra­tion is the com­bi­na­tion of only truth(s), work­a­bil­i­ties, applic­a­bil­i­ties and/or rel­e­vant exis­tences)

    the sig­na­ture of good sci­ence

    [ (the dual­is­tic sci­en­tific process) –> open­ness to new inte­grat­able fac­tors ]

    the sig­na­ture of sci­en­tific break­throughs

    [ (the dual­is­tic sci­en­tific process) –> an aspect of human con­scious­ness like visu­al­iza­tion, the abil­ity to expe­ri­ence the moment as new or sim­ply the exis­ten­tial fact of human con­scious­ness itself ]

  • james­g11

    Satire only works when it engages appro­pri­ately with the text satirised — ie the satire com­pre­hends its tar­get. These ‘com­ments’ are just hope­less sar­casms …

  • twowith­inthree­thati­sone

    The satire I am aim­ing at is the folly of the unwill­ing­ness to inte­grate. With­out inte­gra­tion both spir­i­tu­al­ity and sci­ence are equally frag­mented and reac­tionary pur­suits no mat­ter how seri­ous their efforts. Con­scious­ness exists and so does the sci­en­tific method. With­out sci­ence spir­i­tu­al­ity is the movie Monty Python and the Holy Grail. With­out spir­i­tu­al­ity sci­ence is John Cleese and the Depart­ment of Silly Walks. That’s all I’m say­ing.

  • stanis­laus2

    Steve, I’m a fan of your Debunk­ing Eco­nom­ics. I’m a retired quan­ti­ta­tive psy­chol­o­gist (author of Mulaik, S. A. (2009) Lin­ear causal mod­el­ing with struc­tural equa­tions. Boca Raton, FL: CRC Press Tay­lor and Fran­cis Group). I became inter­ested in MMT
    and found your book as a result. Any­way, have you ever run across the book Odum Howard T. (1971) Envi­ron­ment, Power and Soci­ety. New York: Wiley-Inter­science? Odum was an ecol­o­gist at the uni­ver­sity of Geor­gia in the 1970’s and 1980’s. He stud­ied energy flows in envi­ron­ments (includ­ing human envi­ron­ments) and devel­oped a sys­tems approach for rep­re­sent­ing them. For him money just flowed in the oppo­site direc­tion from energy flows. But what is sig­nif­i­cant is that he was way ahead of his time in regard­ing eco­nom­ics as just a spe­cial branch of his energy flows mod­el­ing. So, in way he has sought to ground ecol­ogy and eco­nom­ics in the the­ory of phys­i­cal energy flows in sys­tems. I think you would like that very much. 

    Another point. I recently attempted to for­mu­late a gen­eral equa­tion of mon­e­tary flows (I’m sure it’s not new, but it has a fea­ture I don’t see in some MMT treat­ments: it con­tains the bank­ing sec­tor as the orig­i­na­tor of dol­lars in cir­cu­la­tion, not the Fed. Any­way it’s a dif­fer­en­tial equa­tion (I am told) adapted from a fun­da­men­tal equa­tion of hydrol­ogy:

    ?C = [X, G, I, L] — [ M, T, S, P]

    where ?C is the change in quan­tity C of money in cir­cu­la­tion (buy­ing, sell­ing goods and mate­ri­als) as func­tion of the dif­fer­ence between inflows and out­flows of money (dol­lars) into cir­cu­la­tion, with
    Inflows:
    X — exports,
    G — gov­ern­ment spend­ing (includ­ing deficit spend­ing)
    I — invest­ing
    L — bank loans

    Out­flows:
    M — imports pur­chases
    T — tax rev­enues
    S — sav­ings (non­spent money)
    P — pay­back of bank loans

    There exists a quan­tity of money in cir­cu­la­tion C’ where the econ­omy is at full pro­duc­tion and employ­ment at sta­ble prices, below which you have defla­tion and room for growth in money sup­ply and above which you have infla­tion.

    To me this basic equa­tion is some­thing every Pres­i­dent and every Sec­re­tary of the Trea­sury and every Con­gress­man needs to know and under­stand,. It shows that there are many ways to reach C’ with dif­fer­ent com­po­si­tions of the basic inflows and out­flows. It shows the folly of fis­cal bal­ances of bud­gets because C’ is what we seek, not bal­ances of gov­ern­ment spend­ing against taxes. It also shows that a fiat money sys­tem (con­trary to the Austrian’s views) need not be inher­ently infla­tion­ary because any­one who knows this equa­tion will know not to intro­duce new money into the econ­omy that causes C to exceed C’. 

    I also think that C is equiv­a­lent to the GDP, which lacks explic­itly the rep­re­sen­ta­tion of bank activ­ity L — P loans minus pay­backs. I put it into my equa­tion above after read­ing your book and learn­ing that the last reces­sion was due to bank debt not fed­eral spend­ing or non-spend­ing.

    In fact, I have also reached the con­clu­sion that quan­ti­ta­tive eas­ing could not pos­si­bly be infla­tion­ary because it rep­re­sents buy­ing up the debt of the gov­ern­ment to banks in pur­chases of secu­ri­ties. I’m not sure why the Fed (and the Aus­tri­ans) think that buy­ing secu­ri­ties will increase the banks’ dol­lars to lend. In fact, and I think some of the folks at the Fed have come around to this view, the Fed plays very lit­tle role in money cre­ation for deficit spend­ing, other than to relieve the banks of car­ry­ing this debt of the gov­ern­ment on their books. Well, my point is that when some­one buys a secu­rity he pays off the loan to the pre­vi­ous holder of the security–in this case the gov­ern­ment for the loan for deficit spend­ing. Those dol­lars were cre­ated out of thin air by the banks, and had to be exchanged into reserve bal­ance dol­lars at a cor­re­spon­dent bank in order to be able to pur­chase the secu­ri­ties of the gov­ern­ment. (That puts the pur­chase on the Fed’s books), The cor­re­spon­dent bank gets the bank-cre­ated dol­lars and swaps them for reserve bal­ance dol­lars that it assem­bles together on loan from var­i­ous banks in the bank­ing sys­tem. OK. What hap­pens when the Fed buys the secu­rity from the bank to the reserve bal­ance dol­lars the bank gets in return for giv­ing up to the Fed the secu­rity? The bank still has the account it cre­ated when it cre­ated the dol­lars out of thin air to buy the secu­rity. It now has to close that account and extin­guish the dol­lars it cre­ated for the loan. It’s a loan! What hap­pens to loans on the books when the loan is paid off by some­one? The loan is extin­guished along with the dol­lars used to pay off the loan. The banks don’t have extra dol­lars with which to make new loans. But they can cre­ate them out of thin air as new loans to bank cus­tomers who demand them. Fiat money sys­tems don’t need to do frac­tional reserve bank­ing because they don’t have to use money in bank deposits that are backed by a com­mod­ity (gold) to make a legit­i­mate loan. 

    The Fed will likely hold the secu­ri­ties it pur­chases until infla­tion arises and then will swap the mature secu­ri­ties for new secu­ri­ties with the Trea­sury. No cash exchanged, just secu­ri­ties. The Fed will then sell these secu­ri­ties to pri­vate and for­eign investors to drain their dol­lars out of cir­cu­la­tion for the time being. They may seek to have the investors roll over the secu­ri­ties with fur­ther swaps with the Trea­sury until infla­tion abates. The Trea­sury will extin­guish the mature secu­ri­ties it gets in the swap with the Fed. The only new money that enters cir­cu­la­tion is the orig­i­nal pur­chase money for the secu­ri­ties sold to fund deficit spend­ing. That could be infla­tion­ary, depend­ing
    on the cir­cum­stances.

    In the mean­time there is no national debt prob­lem. The Trea­sury will swap secu­ri­ties for­ever with banks that buy them for deficit spend­ing by the gov­ern­ment. That means the debt is not really a debt because no one expects the ‘debt’ to be repaid. The banks get a free source of inter­est money and will be loath to give up the secu­ri­ties as long as it gets the inter­est at each swap. The Trea­sury can also bor­row money from banks to cover the inter­est pay­ments. (That’s why the cap on Con­gres­sional bor­row­ing can cause the gov­ern­ment to default. I think it is also uncon­sti­tu­tional for con­gress to have a law that puts an a pri­ori cap on bor­row­ing. The Con­sti­tu­tion of the United States gives an absolute, uncon­di­tional power to bor­row and Con­gress can­not over­ride this by pass­ing a law; it needs a con­sti­tu­tional amend­ment.)

    See also my paper, Mulaik, S. A. (2001) The curve fit­ting prob­lem: An obje­crtivist view. Phi­los­o­phy of Sci­ence, 68, 218–241. It shows why esti­mat­ing more and more para­me­ters loses power to test a hypoth­e­sis in mod­els with many esti­mated para­me­ters.

  • stanis­laus2

    Also the guy com­ment­ing above me is on the right track. I’m a neo-Kant­ian and I see the aim of thought is to be able to syn­the­size as much as pos­si­ble, which is why sys­tems are bet­ter in sci­ence than sim­ple the­o­ries.

  • Hi Stan­ley, and thanks for your com­ment. You’re spot on in every regard, and your equa­tion is very sim­i­lar to how I’ve rede­fined Friedman’s old money equa­tion to show that the expen­di­ture side is not merely MV but MV + dD/dt (since change in debt is credit and it cre­ates spend­ing power 1:1).

    Mak­ing C equiv­a­lent to GDP though is slightly wrong since that leaves out the turnover of exist­ing money.

    Can you send me that paper BTW? My email is debunk­ing AT gmail.com.

  • twowith­inthree­thati­sone

    stanis­laus2,

    The Hegelian dialec­tic

    [ (the­sis x antithe­sis) –> syn­the­sis ]

    also describes the dual­is­tic process of the sci­en­tific method which requires the inte­gra­tive rule of incor­po­rat­ing only truth(s), only exis­tent rel­e­van­cies, and only work­a­bil­i­ties result­ing in a third state of more uni­fied knowl­edge-real­ity.

    The code for progress and greater con­scious­ness of real­ity is an inte­grated dual­ity within an inte­gra­tive process and ethic of trin­ity-unity.

    With­out the will­ing­ness to inte­grate, dis­cus­sions will almost inevitably degrade into obses­sive ego­is­tic con­tention and thus no progress, no new knowl­edge and no enlight­en­ment. So it is with our cur­rent pol­i­tics, eco­nomic the­ory etc.

    Now one may say, “every­body knows this”, how­ever, the truth is every­body doesn’t know it CONSCIOUSLY or we’d have a lot more agree­ment, coop­er­a­tion in inte­grat­ing and con­se­quent progress and increase in knowl­edge in all of its forms. Such is the value of hav­ing a for­mula that con­sciously enables and encour­ages the entire process.

  • twowith­inthree­thati­sone

    Every truth is an expres­sion of a nat­ural philo­soph­i­cal aspect of the con­cept of Grace, and The Cos­mic Code can help the indi­vid­ual become more aware of that fact

    Eco­nomic and mon­e­tary exam­ple:

    [ (Key­ne­sian addi­tional money into the sys­tem x Direct Gift­ing to the indi­vid­ual) –> Wisdomics/Gracenomics/Social Credit ]

    In other words inte­grate the truth that addi­tional money does need to go into the econ­omy to help sta­bi­lize it and the more resolv­ing and valid truth of direct to the indi­vid­ual mon­e­tary poli­cies of a uni­ver­sal div­i­dend and a retail dis­count leads to the more effec­tive and com­plete eco­nomic solu­tion of Wisdomics/Gracenomics/Social Credit

  • twowith­inthree­thati­sone

    This also works in the sci­en­tific aspect of The Cos­mic Code, that is, the rejection/non-inclusion of unwork­a­bil­i­ties, inap­plic­a­bil­i­ties and non-resolv­ing pal­lia­tives, for instance:

    [ ( Social­is­tic re-dis­trib­u­tive attempt to solve an inher­ent sys­temic scarcity of indi­vid­ual income x direct poli­cies of a div­i­dend and retail dis­count) –> Wisdomics/Gracenomics/Social Credit ]

    In other words the rejec­tion of unwork­able and non-resolv­ing re-dis­tri­b­u­tion which is at best a pal­lia­tive not a solu­tion and hence is con­trary to the laws of sci­ence and inte­gra­tion leaves the cost­less solu­tion of poli­cies of a div­i­dend and retail dis­count as the more effec­tive mon­e­tary and eco­nomic solu­tions and these are aligned/lead to Wisdomics/Gracenomics/Social Credit.

  • Bhaskara II

    RE: Pro­fes­sor Keen’s Tweet

    Steve Keen ?@ProfSteveKeen 8h8 hours ago

    I’ve just signed to oppose @Number10gov’s plan to force firms to list ‘for­eign work­ers’. Join me:”

    I think a bet­ter sys­tem is that coun­tries should just award cit­i­zen­ship out­right on obtain­ing work or entry for work. Then every worker is a fel­low coun­try­woman or coun­try­man. This is bet­ter than a employer con­trolled work visa for the new and indige­nous res­i­dents.

    Why? It makes a sit­u­a­tion of more equal foot­ing. In the US the (H1B) work visa means the per­son is stuck with the employer for years and has min­i­mal mobil­ity com­pared to per­ma­nent res­i­dency or cit­i­zen­ship. When immi­grants are totally free agents wages might be less depressed. As, employ­ers might have to com­pete for their ser­vices, rather than monop­o­lize their ser­vices.

    Also, it leaves peo­ple free­dom to start busi­nesses, or make other con­tri­bu­tions rather than just to be an employee at one com­pany.

  • Bhaskara II

    Impor­tant Cor­rec­tion about H1B visa

    My state­ments about the H1B visa are prob­a­bly wrong and or dated. I checked what I said above and below are some links that mov­ing jobs is pos­si­ble. Links are below but bet­ter advice or doc­u­men­ta­tion is prob­a­bly avail­able.

    There is men­tion of a porta­bil­ity act of year 2000 allow­ing trans­fer from one employer to an other, by get­ting approved paper work. There have also been other leg­isla­tive revi­sions.

    Please do not con­sider any com­ments as legal advice. I am some what igno­rant.

    Links below are sec­ondary sources and should not be relied upon as being absolutely cor­rect.

    http://www.nolo.com/legal-encyclopedia/part-time-work-multiple-employers-h-1b-workers.html
    http://hackinglawpractice.com/faq-items/can-i-work-2-jobs-on-a-single-h1b-visa/
    https://en.wikipedia.org/wiki/H-1B_visa

  • ken

    Two things not really related to the post:
    1. I just can’t believe how much at the moment econ­o­mists are ignor­ing how the econ­omy is behav­ing much like it did dur­ing the great mod­er­a­tion. Expand­ing debt brings sta­bil­ity at a cost of increased risk.
    2. This post from The Onion http://www.theonion.com/article/new-study-finds-most-earths-landmass-will-be-phoen-54107 sums up the prob­lems of assump­tions of expo­nen­tial growth.

  • twowith­inthree­thati­sone

    Would not the poli­cies of a new eco­nomic phi­los­o­phy of grace as in gift­ing, namely a suf­fi­cient free gift of a div­i­dend to every­one 18 and older and a gen­eral dis­count to prices at retail sale that was greater than any com­puted rate of infla­tion and that was rebated back to the mer­chants who gave it, resolve both the require­ment of con­tin­ual bor­row­ing and the con­tra­dic­tory par­a­digm of debt and loan only that Finance now uses to dom­i­nate every other busi­ness model and manip­u­late nations and regions?

  • twowith­inthree­thati­sone

    Are they not more thor­ough and eco­nom­i­cally insight­ful solu­tions to the sys­temic scarcity of indi­vid­ual income as well as the answer to the creep­ing ero­sion of prof­its and pur­chas­ing power that infla­tion has had on mod­ern economies, more so than the incom­plete and so non-resolv­ing poli­cies of a basic income guar­an­tee, a min­i­mum wage increase or even QE directly to the indi­vid­ual ALONE ?

  • twowith­inthree­thati­sone

    Vir­tu­ally every prob­lem in exis­tence remains unre­solved because of

    1) lack of aware­ness of work­able third alter­na­tives to its two ortho­dox “solu­tions” and/or

    2) a rigid refusal to even con­sider them.

  • twowith­inthree­thati­sone

    Finan­cial reforms are well and good, but if the idea of mon­e­tary grace as in gift­ing ever took root it would be the end of the rule of finance.

  • Bhaskara II

    His­tory:

    Has any one seen the East Ger­man hun­dert Marks note?

    http://www.banknotes.com/ddr31.htm

    Maybe, they should have put the British Library on the back, since Marx stud­ied there!

    The Read­ing Room was used by a large num­ber of famous fig­ures, includ­ing notably Sun Yat-sen, Karl Marx, Oscar Wilde, Friedrich Hayek, Bram Stoker, Mahatma Gandhi, Rud­yard Kipling, George Orwell, George Bernard Shaw, Mark Twain, Vladimir Lenin (using the name Jacob Richter), Vir­ginia Woolf, Arthur Rim­baud, Moham­mad Ali Jin­nah, H. G. Wells[4] and Sir Arthur Conan Doyle.”

    https://en.wikipedia.org/wiki/British_Museum_Reading_Room

  • Bhaskara II

    His­tory:

    I came across the hun­dret marx note on this a historian’s data.

    http://www.history.ucsb.edu/faculty/marcuse/projects/currency.htm

  • You sound like Steve Hum­mel, and you behave the same way: mul­ti­ple posts on a topic before any­one replies to you.

    I let Hum­mel 1.0 degrade the dis­cus­sion on this list for far too long. I don’t check this blog often enough these days, but I’ll hap­pily block you if next time I check in I see more than 1 com­ment from you which is not in reply to some­one else.

  • Actu­ally, I don’t care whether you are Hum­mel or you aren’t: you’re an obses­sive that over­whelms dis­cus­sion here and you are about to be blocked.

  • Bhaskara II

    Pro­fes­sor Keen,

    Yes, the words and phrases are often the same as Hum­mel 1.0.

    Sur­pris­ingly, I was able to link the two together!

    In one of the pre­vi­ous twowith­ininthree­blabla com­ments, gave quoted text attrib­uted to Hum­mel.

    He, she, or it seems to be com­ment­ing as a sock pup­pet. I’m sure there is a bet­ter older clas­sic term for it.

    I found a site that appeared to archive com­ments twointhree­blabla com­ments archived also.*

    *I left a link to in in a pre­vi­ous site in a com­ment a while ago.
    http://www.debtdeflation.com/blogs/2016/08/19/incorporating-energy-into-production-functions/comment-page-2/#comment-44619

  • Bhaskara II

    Pro­fes­sor Keen,

    Please block.

    Here is twowith­ingth­ree…. quot­ing Hum­mel on August 16, 2016

    http://www.debtdeflation.com/blogs/2016/08/19/incorporating-energy-into-production-functions/comment-page-1/#comment-44564

  • New­town­ian

    Thanks for this arti­cle Steve. You sup­plied me with another intrigu­ing piece into the puz­zle for me. Not being a clas­si­cal econ­o­mist (I only dab­ble in risk and eco­log­i­cal eco­nom­ics) I hadnt appre­ci­ated any details of DSGE mod­els par­tic­u­larly the use of Bayesian sta­tis­tics.

    Bayesian stats in the form of Bayes nets are rapidly becom­ing pop­u­lar among ecol­o­gists and envi­ron­men­tal peo­ple, includ­ing me, in recent years because they offer a very good way to con­cur­rently con­cep­tu­al­ize and quan­tify com­plex sys­tems. My spe­cific inter­est is in Risk Man­age­ment and they appear to pro­vide a means for cap­tur­ing all the dif­fer­ent types.

    As good Bayesians will do, many papers also warn about the seduc­tive nature of mod­els. Mod­els can look very good but do they rep­re­sent real­ity? There are lots of tricks now avail­able and even the good data we have often turns out to have low explana­tory power. A good one is MARCOT, B. G. 2012. Met­rics for eval­u­at­ing per­for­mance and uncer­tainty of Bayesian net­work mod­els. Eco­log­i­cal Mod­el­ling, 230 50 — 62. (which makes me won­der how seri­ously econ­o­mists eval­u­ate their models.…and even then intel­li­gent use is needed.….if in 2007 they resam­pled data between 1990 and 2005 of course they would have found their mod­els reli­able I’m sure). 

    One acute lim­i­ta­tion I’ve been strug­gling over is how hard it is to do time­series mod­el­ling because very quickly the num­ber of vari­ables in expanded series exceeds the quan­tity of data and the prob­lem of Par­si­mony rears its head. Have you ever heard the story of Gauss and mod­el­ling an ele­phant. “How many para­me­ters does it take to fit and ele­phant?”. The answer is about 4 and 1 more and it will wig­gle its head.

    The par­tial BN solu­tion is “Dynamic Bayesian Nets”. But these look to be a way of under­tak­ing what I under­stand you would call Kalman fil­ter­ing and are cen­tral to DSGE mod­els which is really scary. A quick look at the BN ver­sion indi­cates the maths is ele­gant but the pre­ci­sion is utterly depen­dent on how accu­rate and pre­cise your prior data is and as you say lit­tle devi­a­tions get rapidly mag­ni­fied. Poor recog­ni­tion of the uncer­tainty in Bayesian input data is one of the major prob­lems that fre­quen­tist sta­tis­ti­cians have with Bayesians i.e. how good is the input data.

    Reflect­ing this it looks like what the DSGE mod­els may be doing is pre­defin­ing the sys­tems they are mod­el­ling as equi­lib­rium with the sit­u­a­tion at any time con­strained by their ini­tial pre­sump­tions and are not free to go off in new direc­tions when the actual sys­tem being mod­elled changes.

    As I was once told.…..all mod­els are wrong but some are use­ful (imply­ing there is also a lot of junk out there espe­cially when mod­els are mis­taken for real­ity.)