Robust monetary policy in the micro-founded two-country model of the. currency union. Kuznetsova Olga, Higher School of Economics, Moscow, 2009

Size: px
Start display at page:

Download "Robust monetary policy in the micro-founded two-country model of the. currency union. Kuznetsova Olga, Higher School of Economics, Moscow, 2009"

Transcription

1 Robus moneary polcy n he mcro-founded wo-counry model of he currency unon Kuznesova Olga, gher School of Economcs, Moscow, 29 For wo-counry model of currency unon wh scy prces he problem of model uncerany s analyzed. By mehods of robus conrol we derve robus moneary polcy ha wors reasonably well even n he wors-case of model perurbaons. We fnd some volaon of Branard prncple and show ha cenral ban s opmal reacon o he economc shocs becomes more aggressve wh an ncrease n s fear of msspecfcaon. Inroducon...2 Chaper. Leraure...5. Types of uncerany Mehods of analyss Choce crera General resuls Robus analyss for euro area...9 Chaper 2. Reference model of moneary unon Consumer problem Fscal polcy Toal demand Frms Deermnsc seady sae Equlbrum wh scy prces Welfare creron Calbraon Cenral ban opmzaon problem Sae space form of he model...24

2 Chaper 3. Robus conrol specfcaon Model uncerany specfcaon Consran robus conrol problem Mulpler or penaly robus conrol problem Relaons beween wo robus programs Robus conrol program soluon Defnon of possble model se θ Some compuaonal resuls...4 Graph. Terms of rade shoc n he wors-case, approxmang model and raonal expecaons whou robusness...43 Graph 2. Welfare benefs from robus polcy applcaon...44 Concluson...45 Appendx A. Dervaon of he reference model...47 ) Soluon of consumer problem ) Governmens ) Frms program...5 4) Deermnsc equlbrum...5 5) Lnearzaon of scy-prces equlbrum ) Dervaon of mcro-founded loss funcon...55 Appendx B. Compuaonal resuls...65 References...7 Inroducon A usual way of elaborang opmal moneary polcy s o desgn n he conex of a parcular model of he economy. owever nobody nows he rue and exremely complex srucure of he economy. In aemp o capure a leas some of he mos mporan economc regulares by means of racable model a ceran degree of smplfcaon s needed. The resulng analycal framewor appears o be more or less sylzed. Consequenly, nobody can be absoluely confden abou he predcng power of any parcular model employed for he moneary polcy 2

3 analyss. Thus, he problem of dealng wh hs model uncerany or uncerany abou he rue srucure of he economy arses. There are wo common approaches o hs problem. The frs one mples polcymaer s learnng abou he rue srucure of he economy. Bu hs process requres a lo of me and possbly he second approach s more approprae o acle uncerany qucly. Ths second mehod s a searchng for he robus moneary polcy ha wors reasonably well across a ceran se of models or model specfcaons. Accordng o he ype of he model se and specfc ype of uncerany he analyss can be made n dfferen ways. The man queson whch s usually n he cenre of aenon n he robus leraure concerns he comparson of robus polces and smple opmal ones, desgned for he parcular model. The semnal resul called Branard conservasm assumes ha robus polcy s less aggressve n he reacon o he economc shocs han a polcy, consruced for a sngle model whou ang no accoun model uncerany. There are many examples of such analyss for he USA and for he euro area and some auhors agree wh hs general fndng whle he ohers do no. All he models used n such analyss n he applcaon o he euro area represen he areawde models whch operae wh aggregaed daa. Bu seems o be no he bes approach, as a euro area represens a se of neracng heerogeneous counres. So n our analyss we adop for hs problem wo-counry mcro-founded model of currency area. We derve a mcro-founded welfare creron on whch a polcy evaluaon s based. I s shown ha erms of rade beween hese regons maer for he populaon welfare and so usng a one-counry model for analyss of such consrucons as moneary unon s unjusfed. 3

4 Then we calbrae hs model for he European Unon ang no accoun he dfferences beween regons. Afer ha we consruc a robus o model uncerany moneary polcy under commmen. Ths consrucon s based on he robus conrol mehodology. We llusrae how he robus conrol echnques naed by ansen, Sargen (2) can be appled for mul-counry models wh raonal expecaons. We show, conrary o Bhan (22), Zacovc, Weland, Rusem (25), ha he man characerscs of robus polcy and economc oucome depends crucally on he parcular parameer a wllngness of he cenral ban o consruc a robus polcy or a fear of model msspecfcaon. To choose a proper exen of model msspecfcaon we adop an error deecon probably approach frsly naed by ansen, Sargen. The res of he paper s organzed as follows: Chaper oulnes exsng leraure on he robus quesons. The wo-counry model s presened n Chaper 2 where he cenral ban s mcro-founded loss funcon s also derved. In Chaper 3 we apply robus conrol echnques for hs model and derve he characerscs of he robus polcy. The las secon concludes and oulnes he possble drecons for he fuure research. 4

5 Chaper. Leraure. Types of uncerany The noon of uncerany can be referred o some emprc varables or o he model srucure as a whole. To clearly undersand he applcaon area of robus heory accurae dsncon of dfferen sources of uncerany s needed. Kngh (92) conrased rs wh uncerany. Rs refers o he processes whch have nown dsrbuons, whle uncerany arses from unnown facors and processes or from he processes whch hardly can be descrbed by sasc erms. Speang abou emprc varables nghan noon of rs corresponds o he aleaory or sascal uncerany, whle he erm uncerany refers o he epsemc or sysemc uncerany. Uncerany abou he model srucure or model uncerany can arse boh from aleaory (possble esmaon errors, par example) and epsemc sources (unnown facors or neracons of economc varables). In boh cases robus echnques can be mplemened, bu he praccal mehod dffers wh he parcular ype of model uncerany. All possble cases of model uncerany can be roughly dvded no wo broad classes. The frs class can be named more or less paramerc uncerany. In hs case he overall srucure of he model s aen as rue bu he values of specfc parameers are unceran. Ths very class can be dvded no 3 ypes accordng o he exen o whch model uncerany s couned as a paramerc (Telow, Muehlen 24): Bayesan uncerany, he mos paramerc case. Under hs ype of model uncerany parameers of he model are assumed o have nown dsrbuons. 5

6 Srucured Knghan uncerany. In hs case uncerany s assumed o be locaed n some parcular parameers of he model. The rue values of hese parameers are supposed o le beween he nown mnmum and maxmum possble levels. Unsrucured Knghan uncerany, he less paramerc varan. Under such uncerany neher s locaon nor s naure are specfed. Ths case can be analyzed as a game played by he cenral baner agans a malevolen naure (Onas, Soc 2). The second class aes no accoun much more severe exen of uncerany han he frs one so called specfcaon uncerany. No only he precse parameers values bu even core srucural model characerscs are reaed as unnown. In he macroeconomc models such core characerscs can represen a characer of expecaons (forward- or bacward-loong), adherence o mcrofoundaons or, par example, lag srucure. In all cases hs class of uncerany adms he rue economy o be consderably dfferen from he gven model, and no only by he value of some parameers..2 Mehods of analyss Le possble nds of model uncerany, all analyzng robus echnques can also be dvded n wo classes. The frs one refers o he fs class of uncerany he paramerc one and mples ha he polcymaer or economs has only one reference model. The real srucure of he economy may dffer from hs model and he polcymaer beleves ha he rue economy les n he specfed neghborhood of a baselne model (Branard, 967). Ths neghborhood ncludes 6

7 all possble devaons from he reference framewor and one can nerpre hs approach as analyss of a se of smlar bu no dencal models (Gannon, 22). The second approach o he analyss mples a much more serous exen of uncerany specfcaon uncerany. In hs case such mporan characerscs as a nd of expecaons (bacward- or forward-loong or even a hybrd case) or lag srucure of varables are aen as unceran. Thus he rs of specfcaon error s so hgh ha he use of a sngle reference model s no approprae. Moreover, moneary polcy rules based only on a sngle model may cause very bad performance f he rue economy dffers from he baselne model by characer of expecaons mplemened (Levn, Wllams 23). Tha s why he followers of hs approach propose he analyss of a se of dsnc or compeng models wh very dfferen core characerscs..3 Choce crera Despe of general approach chosen we need o have crera accordng o whch one can evaluae he robusness of moneary polcy rule (a rule whch wors reasonably well n all models from he model se (eher smlar or compeng models)). The fs possble way s model or Bayesan averagng (Broc, Durlauf, Wes (24)) when he parameers of fnal rule mnmze a weghed sum of polcymaer s losses under dsnc models from he usng se. The weghs used are poseror probables of each model o be he rue one. There are several problems concernng hs mehod. The frs one s he dffculy wh compung hese probables. The second concerns he fac ha large dffcul models f daa beer han he smaller ones only due o he ncluson of large se of varables, 7

8 no because of her beer qualy. Bu n hs case he weghs of hese large models wll be unreasonably hgh n comparson wh smpler consrucons. Boh problems concerned here are no a queson f equal weghs are chosen as n Levn, Wllams (23). Ths mehod dffers from he Bayesan averagng because model weghs do no correspond o poseror probables. Mnmax creron mples mnmzaon of polcymaer s losses n he wors varan from he se of possble oucomes. Ths choce allows expecng a reasonably well performance of hs robus rule n all possble cases. Some auhors ry o combne advanages of dfferen choce varans and some hybrd crera arse. Kueser, Weland (28) ulzes an ambguy averson prncple when model weghs are aached le n he Bayesan mehod bu o he wors possble oucomes exra weghs are gven. Performance maxmzaon creron s very smlar o he mn-max one. In hs case he problem s represened as a zero-sum game beween a polcymaer and a malevolen naure, whch s amed o maxmze he losses (Gannon, 22). Under sably maxmzaon creron (for example n Onas and Soc (22)) polcy s chosen o maxmze a se of models, whch are sable under such a rule. Fnally, faul olerance approach naed by Levn, Wllams (23) mples frsly consrucng he opmal polcy rule for each model from he usng se. Then a researcher assesses a model s faul olerance by analyzng devaons of one of he polcy parameers from s opmal value, holdng he oher parameers fxed. If he loss funcon s relavely nsensve o changes n all parameers of he polcy rule, he model s consdered as a faul oleran one. Then zones of muual olerance are compued. For hese zones a compromse polcy can be 8

9 consruced. And hs polcy s consdered as a robus moneary polcy under model uncerany..4 General resuls The core characerscs of obaned robus rules crucally depend on he choce crera appled. Many auhors ry o compare hese dfferen rules n order o recognze some regulary n her performance. The frs resul was so called Branard s prncple accordng o whch a polcy under uncerany mus be less aggressve han a ceran one. Ths prncple s confrmed by some auhors (Bhan (22), Zacovc, Weland, Rusem (25), ec) and rejeced by ohers (Onas, Soc (22), Leemo and Södersröm (28)), who have found a more aggressve han a ceran polcy robus rules. Generally speang model averagng echnques lead o comparavely moderae polcy reacon whle mn-max approach mples more aggressve moneary rules (Zacovc, Weland, Rusem, 25). Bu even hs very cauous saemen s no always he ruh. Leemo and Södersröm (25) sae ha he exen of polcy reacon depends on he shoc ype, he source of uncerany and even on he wllngness of polcymaer o mplemen robus polcy (hs robus preferences)..5 Robus analyss for euro area The exen of model uncerany for euro area and consequenly he pernence of robus echnques are very hgh. There are several reasons for. Frs of all n comparson wh he USA models, consrucng models of euro area s relavely new and unelaboraed doman. Secondly he rue neracons beween European auhores are really complex. Whle moneary polcy s decded unquely by he 9

10 European Cenral Ban he fscal measures are underaen n each counry ndependenly whch n urn provoe addonal quesons when consrucng moneary polcy. Moneary auhory canno monor enrely he fscal moves and measures, so ha s an addonal source of uncerany. Thrdly he queson of daa aggregaon arses. There s also a more normave problem of necessary cenral ban s reacons on he counry specfc shocs or wha s he rue welfare creron for hese models. As he frs ECB s presden Wllem Dusenberg once sad, hs s no a rval as gven he large unceranes ha we are facng due o he esablshmen of a mul-counry moneary unon. No only can we expec some of he hsorcal relaonshps o change due o hs shf n regme, bu also, n many cases, here s a lac of comparable and cross-counry daa seres ha can be used o esmae such relaonshps. So, many auhors deal wh he problem of model uncerany relave o he euro area. Alavlla, Cccarell (28) shows on he models of euro area ha model uncerany really represens a danger for he moneary polcy and model averagng can be a useful ool for he polcy conduc. The whole doman of analyss concerns moneary polcy under nflaon uncerany. Unguelon, Coenen, Smes (23), Jaasela (25), Adald, Coenen, McAdam, Svero (25) and Coenen (27) found ha under hgh degree of nflaon he opmal moneary polcy rules are more aggressve and moreover, such rules are more robus. Some of hese arcles are based on he sole baselne model whle ohers examne dfferen ses of compeng framewors. The man general recommendaon for he moneary auhory s o assume hgher exen of nflaon perssence when consrucng moneary polcy. So, opmal robus polcy should mply relavely hgh degree of nera.

11 The smlar resul s obaned by Bhan, Sahuc (22) who examne parameer uncerany and fnd ha for he model of euro area Branard prncple holds. Smlar resul s obaned by Zcovc, Weland, Rusem (25) n spe of mn-max creron appled. Kueser, Weland (28) search for he robus polcy usng a hybrd approach whch combnes Bayesan and mn-max feaures. The weghs aached o he compeve euro models are compued accordng o he Bayesan prncple bu he exra weghs are added o he wors possble varans. All he papers concernng robus polcy characerscs n he euro zone deal wh area-wde models. Bu Bengno (24) showed ha erms of rade beween dfferen counres also maer and hus mus be n he opmal polcy consrucon. In our paper we llusrae a robus polcy consrucon for he wo-counry model of moneary unon.

12 Chaper 2. Reference model of moneary unon In hs paper we assume a unque cenral ban ha decdes for he moneary polcy n a wo-counry currency unon. Ths ban has n s possesson a sngle mcro-founded model wh scy prces ha s aen as reference bu here are some doubs concernng s qualy, hus he moneary auhory acles wh a model uncerany problem. So, he cenral ban needs o consruc no smply an opmal polcy for hs model bu a robus polcy for he possble msspecfcaon. Possble model dsorons are nroduced as choces of some malevolen agen whch ams a he cenral ban s losses maxmzaon. So he man as s a resoluon of mn-max problem by means of robus conrol echnques naed by ansen, Sargen (2). In hs paper we apply he wo-counry opmzng model wh scy prces descrbed n Bengno (24). The dfference of hs paper from he base model of Bengno s he choce of α - parameer of prce nera. Whle Bengno doesn specfy s value for he fuure calculaons, n our analyss we need o have some specfc numbers. So on a base of European daa n dvde counres of EMU n wo regons accordng o her prce rgdy and compue he parameer α for each of regons. Then hese values are used o derve he robus moneary polcy for he unon. The whole dervaon of he fnal equaons s presened n he Appendx A and here we nroduce only he man feaures of he model. The currency unon consss of wo counres or regons ( and F). The populaon of hs unon represens a un-connuum where agens from [, n) nerval belong o he -counry and he res [n, ] are F-counry s nhabans. 2

13 Each counry has an ndependen local governmen, whch deermnes fscal polcy n a correspondng regon. Moneary measures are defned by a sngle cenral ban. Each agen s smulaneously producer of a sngle dfferenaed good and consumer of all good ypes manufacured n he unon. So here s an nerregonal rade whle mgraon of labor force s absen. Number of goods, produced n he regon, equals o n, so hs parameer represens also an economc sze of hs regon or he share of unon GDP produced n he regon. 2. Consumer problem j j s j M s j max U = E β U ( Cs ) + L ; ε s V ( ys ; z s ) () s= P s.. E, { q B } B M + + j j j P ( + R ) P ( + ) M + B p j y j Q B + + C + j j, j j, j ( τ ) P P R P (2) where j [,] s agen s ndex, - counry s ndex ( or F ). β - neremporal dscoun rae. Uly funcon of agen j a perod s posvely depends on hs consumpon j C s, hs soc of real money balances M P j s s (hs can be nerpreed as some lqudy preferences of agens) and negavely on he j ' s supply of a dfferenaed good ( j y s ), as represens labor dsuly. j y s s a funcon of labor hours and a erm V ( ) ε s s a counry-specfc lqudy preference shoc, whle z s represens a producvy shoc n counry. 3

14 E X + sands for he expeced n he perod value of varable X n he perod + ; E s an operaor of raonal expecaons. Every agen consumes home and foregn good bundles, whch are subsues. So ndex j C s a composon of consumpon ndces for home and foregn goods: С j = j n j ( C ) ( CF ) n n ( n) n n (3) Whn each bundle dsnc producs are subsues wh an elascy of subsuon σ. Indces of j agen s consumpon of home and foregn goods are represened by he followng relaons: σ n σ σ σ j σ j С = c ( h) dh n (4) σ σ σ σ j σ j СF = c ( f ) df n (5) n j where c h s a quany of a good h [, n) j he j agen and F regon. Correspondng prce ndces are: from he counry consumed by c f s he same for he good f [ n,] produced n he Regon s prce ndex: n P P P P T = We also nroduce erms of rade n he followng way: P F n F. 4

15 Prce ndces of and F bundles sold on he regon s n σ P p ( h) dh n mare: σ σ PF p ( f ) df n n σ (6) where p ( h) s a prce of a good h sold n he regon, he same s assumed for he good f. Under assumpon of zero ransacon coss every good j mus be F sold a equal prces n boh regons, or p ( j) = p ( j) j. Consumer s budge consran (2) ncludes: B he real value a me of he agen j s porfolo of conngen j, secures ssued n regon and denomnaed n uns of he consumpon-based prce ndex wh one-perod maury q he vecor of he secury prces j B he agen j s holdng of he nomnal one-perod non-conngen bond denomnaed n he unon currency R he nomnal neres rae, moneary auhory s nsrumen τ a regonal proporonal ax on nomnal ncome Q nomnal lump-sum ransfers from he fscal auhory of regon j, o he agen j 2.2 Fscal polcy Fscal polces are deermned by he local governmens n each counry separaely. Each governmen collecs axesτ, deermnes ransfers 5 Q and j,

16 purchases only producs produced n s own counry. We don deal wh he problem of fscal polcy deermnaon, so we don solve any programs for he ransfers or axes. We only ae hese values as gven under assumpon ha he neremporal budge consran s held: E τ Y Q G = s= ( + R ) s = So, he whole sums of governmen expendures are gven by G and F G respecvely. I should be noed ha we assume a cenral ban whch aes he fscal measures as gven and auonomous and so here s no space for he fscal-moneary neracons and even for he opmal consrucon of fscal polcy. Bu one can ncorporae an addonal assumpons concernng he fscal polcy and even he relaons beween fscal and moneary auhores neracons o expand an analyss o he opmal consrucon of fscal measures or ang no accoun game aspecs of polcy deermnaon. For example, he smlar analyss can be made for he model of Beesma, Jensen (25), where polcy neracons can be aen no accoun. 2.3 Toal demand The oal demand for each good s a sum of all prvae agens demands and he demand of he correspondng governmen: σ d p h n j = + P y h T C dj G σ d p f n j F = + P F y f T C dj G (7) 6

17 Or, namng a unon-wde ndex of consumpon C W = j C dj d p h n W y ( h) = T C + G P σ σ d p f n W F y ( f ) = T C + G P F (8) Or on he regon level: Y Y F n W T C + G = P n W F T C + G = F P (9) 2.4 Frms Producers n he model are monopolss on her producs mares. They se prces accordng o he Calvo (983). Each seller faces a probably ( α ) of adjusng hs prce. Ths ype of nomnal rgdy mples ha dynamcs of prce level n every counry can de represened by he followng ln: J ( α ) P = P + p% () σ J σ J, α J, σ So, he erm α can be nerpreed as an exen of nomnal rgdy n he correspondng regon. Wh a help of hs parameer heerogeney beween regons can be nroduced no analyss (le n Brssms, Soda, 28). 7

18 2.5 Deermnsc seady sae Unforunaely, hs model canno be solved n he closed-form. Consequenly, he analyss s made n he erms of devaons from he deermnsc equlbrum. Deermnsc seady sae whou any shocs and zero nflaon rae s aen as a base for he analyss. In he followng analyss a noaon X corresponds o he value of varable X n he deermnsc seady sae. So we assume ha C = cons ε = G = z = π = and T = cons Y = cons We smplfy he fuure analyss resrcng our aenon only o he case of equal ax raes τ F = τ and T = and F Y = Y = C 2.6 Equlbrum wh scy prces In he followng analyss only equlbra whch are close o he deermnsc case descrbed earler are dscussed. So he soluon can be made n erms of small flucuaons from he deermnsc case and an approprae log-lnearzaon of he model s usually conduced. In he log-lnearzed equlbrum condons by X % we denoe he devaon of logarhmc of varable X from he seady sae when prces are flexble, whle X ˆ s he same devaon under scy prces. In oher words X% Xˆ X ( fl. prces) ln X X ( s. prces) ln X Dfferences beween hese wo devaons are named by small leers: 8

19 x = Xˆ X % In some cases s useful o deermne a unon-wde varable X W, whch represens a weghed average of specfc counres values: = + ( ) X nx n X W F, or a relave varable, whch s smply he dfference beween some values for dfferen counres: X X X R F. So, he law of moon of he economy s presened by he followng equaons: ( π ) E Cˆ = Cˆ + ρ Rˆ E () W W W + + ( ) ˆ ˆ ˆ W Y = n T + C + g (2) ˆ F ˆ ˆ W F Y = nt + C + g (3) ˆ ˆ F T = T + π π (4) ( n) ( Tˆ T% ) y E π W = T + C + β π + ( ˆ % ) F F F W F π = nt T T + C y + β Eπ + (5) (6) where C T UCCC ρ ( α β )( α ) ρ + η UC = α + ση VyyC and η + η Vy = C ρ + η VY εε Y V C yy The frs hree equaons deermne he relaons beween consumpon, governmen spendng, oupu gap, expeced fuure nflaon and he value of nomnal neres rae. To smplfy fuure calculaons we assume ha a fscal polcy s auonomous from any moneary decsons and s reaed as nown wh cerany, so g + = E g +. In hs case we can rearrange equaons (-3) no (7): 9

20 E y = y + Rˆ E π ρ W W W + + (7) Ths equaon s a usual form of IS-curve for he hole currency area, whch deermnes an oupu gap, whch depends posvely on s fuure expeced value, expeced fuure nflaon and negavely on he nomnal neres rae. Equaons (5-6) descrbe supply sde of he unon economy and sand for he New Keynesan Phllps curves. From hese equaons nflaon raes n he unon regons are deermned by he unon-wde oupu gap, expecaons of fuure nflaon and he unon erms of rade. Usually nsde erms of rade are omed from such ype of analyss, based on he unon-wde models, so he opmal polcy s consruced for he aggregae levels of nflaon and oupu. Bu from he equaons (5-6) s clear ha ang no accoun rade flows beween regons can be mporan for polcy consrucng. For hs very purpose we offer a more-hanone-counry model for analyzng robusness characerscs. Equaon (4) goes explcly from he defnon of erms of rade and represens a dynamc of hs varable deermned by s pas value and he curren nflaon raes n boh counres. So he cenral ban s as has o se a nomnal rae mnmzng s welfare funcon subjec o he equaons (4-7). Thus, n he model here are 4 forwardloong varables ( T π,, π y F W ) and one polcy conrol varable ( R ). 2.7 Welfare creron We assume ha cenral ban s benevolen and res o maxmze socal welfare W + gven by W = E β w - an expeced weghed sum of all fuure values of average uly n he unon w U ( C ) V y ( j), z dj. 2

21 The second-order approxmaon of he welfare funcon s based on he Beesma (25) and gves he followng form of welfare creron (see Appendx for deals): + W = E β L, where one perod loss s gven by ( ) ˆ γ ( π ) γ ( π ).. ( ε ) W 2 F 3 L = Λ y + n n Γ T T% + + F + p + o (8) Where..p. sands for he erms ndependen from polcy and he las par of hs relaon 3 ε ncludes all parameers of more-han-second order of approxmaon / σ and Λ = n / + n / C F C ( + η ) / σ F ( ρ η ) Γ = ( n / + n / ) + C n / C γ = n / + n / F C C F ( n) / C + ( ) γ = n / n / F F C C C (9) (2) (2) (22) We can see ha n hs model he loss funcon L has a usual quadrac form. Bu conrary o he semnal represenaon where he loss value s defned by he oupu gap and nflaon raes, erms of rade are also presen n hs funcon. I can be explaned by he reference model srucure, as he Phllps curves srucure ncludes rade beween regons and aes no accoun no only a currency area as a whole bu also relaonshps whn a unon. 2.8 Calbraon In our calbraon we parly follow Bengno (24). Thus we chose a value of elascy of producng dfferenaed goods η equal o.67. Parameer of 2

22 neremporal subsuon β equals o.99. The degree of monopolsc compeon σ s aen o be equal o Rs-averson coeffcen ρ s assumed o be /6. Moreover, s assumed ha he shoc T % follows he auo-regressve process of he nd T% =.95T% + ε, where he erm ε sands for he whe-nose process wh varance.86. The man dffculy concerns he choce of prce nera parameerα. In hs secon we don follow Begnno (24) who allows hese parameers o vary across a wde range of possble values. In conras, our choce of hese values s based on he esmaons of by Vermeulen a al (27). Table Frequency of prce changes and counry weghs n Euro GDP (%) (2) (3) Belgum France Germany Ialy Porugal.23.5 Span Euro area.22 Noe: (2): Frequency of prce changes ( α ) Source: Vermeulen a al (27) (3): Counry wegh n Euro GDP (%) 22

23 We ae a frequency of prce changes as a proxy for he probably o change a prce ( α ). We dvde counres n wo groups accordng o he followng scheme: f a frequency of prce changes s lower or equal o.22 (average frequency for he unon), he counry belongs o he regon. If hs frequency s hgher han.22, he counry s a par of F regon. Thus for he counres wh avalable daa regon consss of Germany, Span and Ialy, whle regon F consss of France, Belgum and Porugal. Accordng o he Table regon produces around 7% of unon oupu, so we calbrae he regon sze o he.7. Accordng o he correspondng weghs we assume ha an average frequency of prce change n he regon equals o.7, whle he same rao for he regon F comes o.23. F These values correspond o he model parameers α =.83 and α =.77 Now we compue he correspondng weghs n he welfare funcon: Λ =.58 Γ =.24 γ F =.82 γ = γ =.8 So, we can see ha he wegh aached o he scer regon ( ) n he loss funcon s hgher han a wegh of a more flexble one whch corresponds o he man fndngs of papers sudyng quesons of nomnal nera n he models of euro-area (see, for example, Adald, Coenen, McAdam, Svero (25)). 2.9 Cenral ban opmzaon problem Accordng o he made assumpons and he calbraon from he prevous secon we can compue values of he man parameers of he model. So, he cenral ban s program can be rewren as: 23

24 max E + β L ( π ) ( π ) 2 W ˆ F = + % L y T T (23) W W W E y+ = y + 6 R Eπ + π =.3( Tˆ T% ) +.5y +.99E π s.. π =.4( Tˆ T% ) +.y +.99E π ˆ ˆ F T T% = T T% + π π + e e =.95e + ε W + F W F + Ths specfcaon represens he problem of search for he opmal moneary polcy on he base of one reference model. From here s seen ha he man concern of he cenral ban s nflaon n he regon wh he hghes prce scness. In he followng secons we dscuss a polcy of cenral ban whch aes hs model as reference one bu fears ha realy can be raher dfferen from hs base framewor. In he nex secon we sae he program of robus polcy consrucon and hen we fnd parameers of he moneary polcy under model uncerany. 2. Sae space form of he model For he convenence of furher analyss he program (23) can be rewren n he usual sae space form. For hs we frsly expand he low of moon of erms of rade o he perod + and ae he raonal expecaon: Tˆ T% = Tˆ T% + π π +.95e F E Tˆ T% = Tˆ T% + E π E π +.95e F (24) From hs, usng (4), we oban: 24

25 ( ˆ % ) ˆ % F π ( ˆ % + + ) ( ˆ W π % ) ( ˆ % F W ) π π E T T = T T +. +.*.4 T T.*.y W. +.*.3 T T +.*.5y +.95e =.44 T T y +.95e (25) From oher equaons of (23) we oban expressons for he all expecaons of forward-loong varables: ( ˆ % ) ( ˆ % ) ( ˆ % ) E y =.2 T T +.4y 4.24π.82π + 6R W W F + E π =.3 T T.5y +.π W + E π =.4 T T.98y +.π F W F + (26) So he sae space form of our model s he followng: e =.95e + ε + + ( ˆ % + + ) ( ˆ %) ( ˆ + %) ( ˆ + % ) ( ˆ + %) W F E T T =.44 T T.55 y.π +.π +.95e W W F E y =.2 T T +.4y 4.24π.82π + 6R E π =.3 T T.5 y +.π W F W F E π =.4 T T.98y +.π (27) Or n he bref form: e e + A BR C E z = z ε + (28) Where T% ˆ T W y z = π F π - a vecor of forward-loong varables, E z + s an expeced n he perod fuure value of vecor z. A s marx of sze 5 5 of correspondng coeffcens from (27). B = and C =, showng ha only a predeermned varable e s allowed o be affeced be shocs. 25

26 Mnmzng loss funcon can be also rewren n he followng form: E β ( x' Qx ) (29) = where x e = z and Qs marx of sze 5 5 : n( n) Γ Q = Λ γ γ (3) So, he problem (53) can be rewren as: mn E β ( x' Qx ) R = e e + s.. A BR C E z = z ε + (3) 26

27 Chaper 3. Robus conrol specfcaon 3. Model uncerany specfcaon In modelng he robus program we mplemen a ansen-sargen s approach whch s also called robus conrol. We assume ha he cenral ban has (57) as a reference model descrbng he economy. Bu a he same me hs moneary auhory fears ha hs reference consrucon doesn correspond o he real sae of naure properly here s a rs of msspecfcaon. In oher words some perurbaons of modeled economy from he real one are allowed. The possble sources of hese perurbaons are some unnown varables or processes. To accoun hs possble msspecfcaon moneary auhory analyses only a class of alernave models, whch canno be dsngushed from he reference one wh help of sascal mehods. In oher words a se of possble perurbaons s lmed - ncludes only such perurbaons, whch wh some fxed probably wll no be dscovered. The reason o mpose hs resrcon on he possble msspecfcaon s que clear for grea perurbaons, when he real economy dffers consderably from he reference one, here s no any reason o ae any decson on he base of hs concree model and adapaon of he model o realy s needed. So he as for he cenral ban s o consruc no an opmal for hs model polcy bu a polcy, whch performs reasonably well even f here s any perurbaon. For hs purpose a mn-max creron s appled a robus polcy s such one ha produces he smalles losses n he case of he wors model perurbaon. For hs wors-case creron we can suppose ha perurbaons from he reference model ae he form of some addonal shocs 27 υ + s whch are added o he

28 sandard ε + s n he model (57) and are nduced by so called malevolen naure or evl agen, whch res o maxmze losses of he cenral ban. So he robus program can be represened by smulaneous wo-agens game, where he evl agens chooses a perurbaon from he reference model υ + s, he cenral ban decdes he values of s nsrumen neres rae. The se of possble perurbaons s modeled by he resrcon on he evl agen s measuresυ + s. In general here are hree possbles for he cenral ban o nerac wh prvae agens who form expecaons n he economy: ) Commmen o he opmal polcy 2) Commmen o some smple polcy rule (for example, semnal Taylor rule) 3) Dscreon ere we assume commmen case. In hs suaon several assumpons concernng he prvae percepons of he possble msspecfcaon can be made: ) Populaon has he same fear of msspecfcaon as he cenral ban 2) Populaon s more cauous han moneary auhory 3) Populaon s more careless In our analyss we consder he frs smples case, when populaon and cenral ban have he same reference model and he same deas concernng possble msspecfcaons. In he res wo cases he populaon s reference model and s se of model devaons should be specfed. Moreover, we need decde f he evl agen comms or no. Followng ansen, Sargen we analyze a case when boh opmzng agen comm. The man reason for hs assumpon s he fac ha our malevolen agen s only a meaphor used o consruc and solve he mn-max problem. The cenral ban under uncerany 28

29 derves s polcy as f hs evl naure exsed. Tha s why we suppose ha evl agen arrves and solves s program only when he cenral ban opmzes Consran robus conrol problem As we ve already dscussed n he prevous secon, here s some resrcon on he evl agen s measures. We assume he followng neremporal consran of he malevolen naure: + ' + = E β υ υ η (32) where υ s a vecor of dsurbances naed by he malevolen naure n he economy. And η s a possble oal exen of model msspecfcaon. In oher words, (32) represens he allowed se of perurbaons dscussed earler. If we magne all possble perurbed models as a cloud around he reference one and η s s radus. Ths represenaon of possble model se corresponds o he consran robus conrol problem of he cenral ban n he followng form: mn max E β ( x' Qx ) R = e e = s.. A BR C E z + z E υ + ' + = β υ υ η ( ε υ ) + + (33) Sze of possble perurbaons, η s defned by he cenral ban s fear of msspecfcaon: hgher fear allows hgher possbles of he malevolen agen and bgger devaons of he real economy from he reference one and s refleced by hgher η n (32). gher fear of msspecfcaon a he same me sands for hgher preferences for he robusness. 29

30 For example, f we assume ha he cenral ban has no any fear of msspecfcaon or has no preferences for robusness, η =, and we oban he program equvalen o (3). Noce ha we analyze polcy decsons for he momen and evl acons for he momen +. Ths means ha he value of hs sraegc shoc s nown a he me. (Alernave assumpon supposes ha he evl agens acons are no nown a he prevous momen or are hdden by he sochasc errors of hs perod). For a gven law of moon of economy he oal welfare depends only on he polcymaer s and malevolen agen s seps and he economc shocs. So we rewre problem (33) n he followng form: mn max β L ( R, υ) R = υ + = β D ( υ) η (34) where R s a sequence of cenral ban s decsons and υ s a sequence of sraegc shocs naed by malevolen agen: D υ = υ ' υ Mulpler or penaly robus conrol problem Anoher possble represenaon of robus conrol program s so called mulpler or penaly problem. The program (33) aes he followng form: β x Qx θυ+ υ+ = mn max E ( ' ' ) R υ e e = s.. A BR C E z + z ( ε υ ) + + (35) where θ represens a se of possble devaons of reference model from he real economy. When hs erm θ s low hs se s very large, when θ s hgh, here 3

31 are only very lmed dsorons. The case of θ corresponds o he usual opmal conrol program when no accoun of possble msspecfcaons s aen. Rewrng n he same form as (34), we oban: β L R υ θ β D + υ (36) = = mn max (, ) R υ 3.4. Relaons beween wo robus programs In hs secon we derve relaons beween wo robus programs presened earler. Frs of all we rename he overall losses of he consran robus program as K ( η ) = mn max β L ( R, υ) R υ Η ( η ) = Η ( η ) = υ : β D + ( υ) η = (37) The resulng level of opmzed funcon for he penaly problem s named: (38) = = = + V θ mn max β L ( R, υ) θ β D ( υ) R υ Now we reformulae consran robus problem (34) n he Lagrangan form: mn max mn (, ) R υ θ β L R υ θ β D + υ η (39), = = where θ s a Lagrange mulpler for he malevolen agen s consran. Changng he order of opmzaon (we can do f we consder only cases when equlbrum exss) we oban mn mn max β L ( R, υ) θ β D ( υ) η θ R υ + = = = = mnv + θη θ ( θ ) (4) From (4) s seen ha for a gven θ he las erm of hs expresson (θη ) doesn nfluence he choce of R andυ. So he followng saemen s rue: 3

32 Saemen. For he gven η * R and * υ are soluons of consran robus program. Then exss a * θ (equal o he opmal Lagrange mulpler from he consran robus program) such ha he correspondng penaly robus program has he same soluon and K mnv η * = θ + θη * (4) θ Now we consder ha for some follows ha for any η and θ K V V K θ * η θ * η * θ he penaly program s resolved. From (4) η θ + θη, so for any η (42) From here we conclude ha V max K (43) θ * η θ * η η Or, usng he defnon of K ( η ), V * θ maxmn max β L (, ) * R υ θ η η R υ Η ( η ) = (44) We maxmze he rgh sde of he expresson f we pu β D + ( υ) nsead of η, as he consran n (64) holds. = * max mn max β L ( R, υ) θ β D + ( υ) η R υ Η ( η ) = = * β L R υ θ η η R υ Η ( η ) = max mn max (, ) = (45) We rewre he penaly robus program 32

33 * * R υ = = V θ = mn max β L ( R, υ) θ β D ( υ) = * max mn β L ( R, υ) θ β D + ( υ) υ R = = = = + * max max mn β L ( R, υ) θ β D + ( υ) η υ Η( η ) R = = = (46) * Tang no accoun (45) and (46) we rewre he defnon of V ( θ ) : V η υ Η ( η ) * * max max mn L ( R, ) R = θ = β υ θ η = ( η ) = max K η * θ η (47) For gven η he las par of expresson, * θ η doesn nfluence he choce of R and υ, so he soluon of penaly robus corresponds o he soluon of he consran robus program. Saemen 2. For some * θ * R and * υ are soluons of penaly robus program. Then * * he consran robus program wh η = β D + υ has he same soluon. = So, we ve showed ha wo robus programs are ghly relaed. The sze of possble uncerany s gven by η n he consran robus program and by θ n he penaly program. These wo measures are nerconneced n he way presened by Saemens and 2. From (4) and (47) we can see ha hgher θ means lower correspondng value of η and vce versa. So when we spea abou hgher uncerany we mean hgher η and lowerθ Robus conrol program soluon 33

34 In hs and he nex secons we acle wh he penaly robus program rememberng ha all hese resuls can be easly represened as he soluon of correspondng consran problem. Ths problem s solved numercally and here we presen some echncal deals of soluon proposed n Gordan, Soderlnd (24). Gordan, Soderlnd (24) assume he followng robus polcy program: { u} { υ} ' β + + θυ+ υ+ = mn max E ( x ' Qx u ' Ru 2 xuu ' ) x x ( ε υ ) + s.. A Bu C E x = x 2 C C = Οn 2 n, where x s a n vecor of predeermned or bacward-loong varables, x2 s n 2 vecor of forward-loong varables, vecor ' ' neres equal o (, 2 ) x x. ' x s a vecor of all varables of u s a vecor of cenral ban nsrumens of sze. ε + s an d n vecor of shocs and υ + s an n vecor of sraegc shocs, naed by evl agen. Q and R are symmerc marces. Our program (35) s a prvae case of such of Gordan, Soderlnd (24) wh zero-marx and he sole polcy nsrumen neres rae, so =. n =, as we have only one predeermned varable e and here are 4 forward-loong varables n he model, so our vecor of forward-loong varables s of sze n 2 = 4. Because he core srucure of our model and such proposed n Gordan, Soderlnd (24) are dencal we now pass o he soluon algorhm of our specfc program. Frs of all we defne he error of expecaons for he perod + : = z E z (48) z ξ Then we noe he oal error of he model as: 34

35 ξ + ε = ξ + z + 5 (49) So he program (36) can be rewren as follows: β x Qx θυ + υ+ = mn max E ( ' ' ) (5) s.. x = Ax + BR + ξ + Cυ The Lagrangan of hs represenaon s: L ( x' Qx θυ ' υ ) = E β = + 2 ρ ' + ( Ax + BR + ξ+ + Cυ + x + ) (5) where ρ + s a vecor of sze 5 of correspondng langrangan mulplers. The frs-order condons wh respec o ρ +, x, R and υ + are: I5 Ο5 Ο5 Ο5 5 x + A5 5 B5 C5 Ο5 5 x ξ + Ο5 5 Ο5 Ο5 β A' 5 5 R + βq5 Ο5 Ο5 I 5 R = Ο5 + (52) Ο 5 Ο Ο B' 5 υ + 2 Ο 5 Ο Ο Ο 5 υ + Ο Ο 5 Ο Ο C ' 5 E ρ+ Ο 5 Ο θi Ο 5 ρ Ο where Οn sands for zero-marx of n sze. Or f we rename he marces of correspondng coeffcens n (52) we oban x + x ξ+ R R 5 G + Ο = D + υ + 2 υ + Ο Eρ+ ρ Ο (53) e Then we ae condonal expecaon of (53) and nong ha ρ s a shadow prce correspondng o he predeermned varable shadow prces for he z we oban: e and ρ s a vecor ( 4 ) z of 35

36 e + e z z ρ ρ + z z z + e R GE = D, or posng = z R + R ρ and λ = we oban: υ+ e υ + 2 υ + ρ e e ρ + ρ GE = D λ + λ + (54) Vecor e = z ρ has nal condons: he frs erm s he predeermned varable wh some nal condon e and he second represens shadow prces for he forward-loong varables and hus can be chosen freely a he nal perod o be z equaled o zero, so ha ρ =. Then we use decomposon proposed n he Gordan, Soderlnd (24). Square marces G and D can be represened by he followng decomposon: G = VSZ D = VTZ, when Z s a ranspose of conjugae of Z. Then, V V = Z Z = I and S and T are upper rangular. ence, (54) can be rewren as: + VSZ E VTZ λ = + λ (55) Premulply (55) by V + SZ E TZ λ = + λ (56) As S and T are upper rangular, (56) can be rewren so ha a he frs places sable soluons go. Sably of he soluon s checed by he correspondng egenvalue s modulus should be less han. If we denoe now 36

37 θ = Z δ λ, (57) where θ corresponds o he sable roos and δ - o he unsable ones, we oban: SE θ θ = T δ, or, as S and T are upper rangular, + δ + S S θ T T θ E =, (58) θθ θδ + θθ θδ S δδ δ + T δδ δ Ο Ο where J j comes for he correspondng par of J marx. For every sable soluon δ =, so Sθθ Eθ = + Tθθ θ or Eθ + = Sθθ Tθθθ (59) z z Then, usng e + Ee + = Ceε + and from Gordan (24) ρ E ρ =, we have + + e+ e + Ceε + E z z + E+ ρ = = + ρ + Ο (6) θ Z θ Z δ θ From (42) Z λ = δ = Z λθ Z λδ δ (6) And as δ =, Z θ = θ λ Z δ (62) ence, E Z E C ε Ο e = θ ( θ+ θ + ) = (63) In assumpon ha Z θ s nverble and usng (44), we oban: θ C ε e + + = Sθθ Tθθθ + Z θ Ο 4 (64) Then e + C eε + + = Z z θθ + Z θ Sθθ Tθθ θ ρ = = + + Ο 4 (65) + e + e + Or Z z θ Sθθ Tθθ Z θ M z z ρ = ρ + = + + Ο 4 ρ Ο 4 37 e e C ε e C ε (66)

38 z R e e And λ = = Zλθ Z θ = N z z υ + ρ ρ e ρ (67) From he las equaon (67) one can derve an opmal robus polcy, whch s consruced as some reacon on he values of predeermned varables and shadow prces of forward-loong varables: e R = ( Zλθ Z θ ) R z ρ (68) Equlbrum dynamcs of he model depends crucally on he acons aen by he evl agen n realy. And here we dsngush wo varans: wors-case model and approxmang one. The wors case aes place when he evl agen s acon realzes n realy. So such a model s descrbed by (66-67). The oppose suaon, when he cenral ban consrucs he robus polcy bu he evl agen doesn ac refers as an approxmang model. We oban from approxmang model by pung all sraegc shocs n (67) equal o zero. Surely, he resulng dynamcs of he economy depends crucally on he value of msspecfcaon fears, θ. The nex secon s devoed o he defnon of reasonable value for hs parameer Defnon of possble model se θ Followng Denns (29), we choose he value of hs parameer on he bass of error deecon probably mehod. The man dea under hs approach s ha model from he possble model se canno be easly dsngushed usng he avalable daa. In oher words, he cenral ban canno decde f real daa are 38

39 generaed wh he acons of malevolen agen or n he realy here s no any an-auhory measures. The frs suaon, when he malevolen naure aes all possble resources o nfluence he cenral ban s losses, s named wors case model ( W ). The second case, when he cenral ban nsure agans hs agen by robus polcy bu he evl naure doesn aes any measures s reaed as approxmang model ( A ). So, accordng o he error deecon mehod, he cenral ban should no be able o dsngush beween hs o models, ( W ) and ( A ), usng all avalable nformaon. In he oppose case, when he can ruly decde f here s any model deecon, a polcymaer has no any need of robus polcy, he smply adops he model o he realy. Probably of error π θ : ( LA LW W ) ( LW LA A) π θ = Pr > / 2 + Pr > / 2 (69) Where L s a lelhood funcon. The frs par of rgh-hand expresson sands for he probably o rea he model as an approxmang case whle n realy malevolen naure nerrups he daa generang process and he second par s probably o ae he model as a wors-case one whle here are no any naure acons. Ths probably s compued wh a help of smulaons and depends crucally on he value ofθ. We need o choose allowed error probably and fnd correspondng sze of msspecfcaon. 39

40 For example, 5% error probably corresponds o zero robusness (he case of sandard opmal polcy n he model wh raonal expecaons). Lower π ( θ ) means hgher msspecfcaon fear and hgher robus preferences (or lower θ ) Some compuaonal resuls We ve analyzed several varans of polcy robusness. Correspondng resuls of compuaons are summarzed n Appendx B, where noons M and N sand for he wors-case, M a and Na descrbe law of moon of approxmang model. Fv represens a malevolen agen reacon on he predeermned varables. Wha we are neresed n are he coeffcens of he robus polces, whch are e = represened by R ( Zλθ Z θ ) R z ρ The moneary polcy reacons under dfferen preferences for robusness are summarzed n Table2. Table 2. Parameers of robus moneary polcy e R = ( Zλθ Z θ ) R z ρ Error deecon probably θ r r2 r3 r4 r5 2% % % %

41 The mos mporan s he frs coeffcen r, whch represens reacon on he erms of rade. As hs parameer s subjec o he shoc nfluence, hs coeffcen also reflecs polcy reacon o he shocs. If we rac he module of he correspondng coeffcen as a degree of polcy aggressveness, we can see ha aggressveness of polcy reacon on he predeermned varable e rses wh ncrease n he preferences for robusness. Ths relaon can be explaned by he fac ha wh hgher fear of msspecfcaon he cenral ban supposes ha here s hgher possbly ha he shoc n he economy s no smply..d process bu s naed by he malevolen agen, and he reacs more aggressvely. ere we see some volaon of Branard prncple for he moneary unon and resul smlar o he sandard mn-max sraegy of more aggressve reacon o any shocs. Anoher neresng pon s he sgn of polcy reacon on he shoc: f here s a posve shoc, he cenral ban rases he neres rae. There s a clear nuon under hs fac: accordng o (23) hs shoc ncreases nflaon n he regon and decreases nflaon n he second par of he unon. Bu accordng o he coeffcens n he polcy loss funcon, he hng of s mos concern s nflaon n he regon wh hgher nflaon perssence, as s wegh n losses s much more mporan han oher varables. So n response o such shoc he reasonable response of cenral ban s an ncrease n neres rae wha decreases oal oupu n he unon and hus sablzes nflaon n he regon of he hghes neres. 4

42 Now we smulae conduc of economy under a shoc of unon s erms of rade, Impulse-response funcons are summarzed n he Graph, where he followng noaons are used: ε TT Terms of rade V Value of sraegc evl shoc Y Oupu gap L Value of loss funcon ph Inflaon n he regon WC Wors-case model pf Inflaon n he F regon A Approxmang model R Moneary polcy nsrumen RE Raonal expecaons whou robusness 42

43 Graph. Terms of rade shoc n he wors-case, approxmang model and raonal expecaons whou robusness TTWC YWC phwc pfwc RWC VWC LWC x 2 TTA x 22 YWA x 2 pha x 2 pfa x 2 RA VA x 42 LWA 5 5 TTRE YRE phre pfre RRE x -5 VRE 5 5 LRE

44 We can see from hs pcure ha malevolen naure reacs on hs shoc by edng some sraegc shoc (wors-case model). Knowng hs, he cenral ban ncreases neres rae. The resul of hs measure s a moderaon of he s nflaon rse. Resul of polcy-naure game n he wors case s a rse of oupu gap and nflaon n he regon rse, whle regon F s nflaon decreases. The las graph n he frs column represens losses assocaed wh he wors-case model. We can see, ha malevolen acons cause suffcen welfare losses. When here s no malevolen acon, he losses of naon are praccally nl (see column 2), even f robus polcy s appled. So, n hs case he robus polcy can properly couneraac any exernal shoc. If we consder RE case, when cenral ban aes he possbly of msspecfcaon as zero, he losses are hgher han n he case of a smple robus polcy. We compare hese wo polces defnng he dfference beween oal welfare under robus polcy and under non-robus one. Resul s presened n he followng graph: Graph 2. Welfare benefs from robus polcy applcaon

45 We can see ha robus polcy enals some welfare benefs n comparson wh he alernave. So here s an evden bass for robus polcy consrucon. Concluson For he mcro-founded wo-counry model of currency unon we have consruced a robus polcy under commmen. We ve found ha he cenral ban reacs on he shocs more aggressvely when hgher exen of possble msspecfcaon s admed, hus Branard prncple s volaed. We ve shown echncal deals of robus conrol opmzaon n he adapaon for wo-counry models. There are many possble applcaons of hs mehod for he consrucon of moneary polcy n he currency area. Frs of all we ve consdered only he case of erms of rade shoc. The analyss can be easly exended o oher economc shocs. For example, echnologcal changes can be aen no accoun. Secondly, we ve consruced a full commmen polcy. Bu he conrol of such polcy s raher dffcul, so he problem of polcy nconssency can arse. Conrol of smple moneary rules (for example, of a Taylor ype) s easer and so such rules can conrbue o he populaon confdence o he cenral ban measures. So he fs possble exenson s consrucon of robus moneary polcy rule. Thrdly, n our model here s no sance for moneary and fscal polcy neracons. The nfluence of cenral ban preferences for robusness on he economy under assumpon of sraegc neracons beween moneary and fscal 45

46 auhores can be analyzed, for example, on he base of wo-counry model of Beesma, Jensen (25). Then, our analyss can be easly exended o he case of ree or more counres. For example, represenaon of EMU as a unon of Germany, France and Ialy, he larges European economes, mgh provde he research wh some mporan exensons. Fnally, we ve dscussed only a case of wo-counry world. Neverheless, neracons wh he res of he world va nernaonal rade and capal flows can be analyzed. 46

47 Appendx A. Dervaon of he reference model ) Soluon of consumer problem Soluon of he consumer problem consss of 3 sages: ) Solvng a problem () subjec o (2) one deermnes an opmal value of consumpon ndex j C, an opmal real balances soc M P,an opmal oupu suppled. Ths problem can be easly rewren: j s j M + s j max E β U ( C+ s ) + L, ε+ s V ( y+ s, z+ s ) + s= P + s + E M + B p j y j Q j j, j, j + s + s + s + s j + s B + s + + ( τ ) C + P s P s P s λ s j j s=, j B + s M + s E + s { q + sb } P + s ( + R + s ) P + s From where we oban he relevan frs-order condons: M R LM, ε = U C ( C ) P P + R (A.) P U C ( C ) = ( + R ) β E U C ( C + ) P + j ( + ) λ EU C = E + P C s s + s (A.2) (A.3) (A.) ells us ha n he opmal consumer choce margnal rae of subsuon beween consumpon and real money balances holdngs mus concde wh he prce of real money balances n erms of consumpon ndex. (A.2) represens a usual Euler equaon for he forward-loong model. Accordng o (A.2) he margnal uly of consumpon n he perod mus equals o he expeced 47

Graduate Macroeconomics 2 Problem set 5. - Solutions

Graduate Macroeconomics 2 Problem set 5. - Solutions Graduae Macroeconomcs 2 Problem se. - Soluons Queson 1 To answer hs queson we need he frms frs order condons and he equaon ha deermnes he number of frms n equlbrum. The frms frs order condons are: F K

More information

UNIVERSITAT AUTÒNOMA DE BARCELONA MARCH 2017 EXAMINATION

UNIVERSITAT AUTÒNOMA DE BARCELONA MARCH 2017 EXAMINATION INTERNATIONAL TRADE T. J. KEHOE UNIVERSITAT AUTÒNOMA DE BARCELONA MARCH 27 EXAMINATION Please answer wo of he hree quesons. You can consul class noes, workng papers, and arcles whle you are workng on he

More information

V.Abramov - FURTHER ANALYSIS OF CONFIDENCE INTERVALS FOR LARGE CLIENT/SERVER COMPUTER NETWORKS

V.Abramov - FURTHER ANALYSIS OF CONFIDENCE INTERVALS FOR LARGE CLIENT/SERVER COMPUTER NETWORKS R&RATA # Vol.) 8, March FURTHER AALYSIS OF COFIDECE ITERVALS FOR LARGE CLIET/SERVER COMPUTER ETWORKS Vyacheslav Abramov School of Mahemacal Scences, Monash Unversy, Buldng 8, Level 4, Clayon Campus, Wellngon

More information

Solution in semi infinite diffusion couples (error function analysis)

Solution in semi infinite diffusion couples (error function analysis) Soluon n sem nfne dffuson couples (error funcon analyss) Le us consder now he sem nfne dffuson couple of wo blocks wh concenraon of and I means ha, n a A- bnary sysem, s bondng beween wo blocks made of

More information

John Geweke a and Gianni Amisano b a Departments of Economics and Statistics, University of Iowa, USA b European Central Bank, Frankfurt, Germany

John Geweke a and Gianni Amisano b a Departments of Economics and Statistics, University of Iowa, USA b European Central Bank, Frankfurt, Germany Herarchcal Markov Normal Mxure models wh Applcaons o Fnancal Asse Reurns Appendx: Proofs of Theorems and Condonal Poseror Dsrbuons John Geweke a and Gann Amsano b a Deparmens of Economcs and Sascs, Unversy

More information

Dynamic Team Decision Theory. EECS 558 Project Shrutivandana Sharma and David Shuman December 10, 2005

Dynamic Team Decision Theory. EECS 558 Project Shrutivandana Sharma and David Shuman December 10, 2005 Dynamc Team Decson Theory EECS 558 Proec Shruvandana Sharma and Davd Shuman December 0, 005 Oulne Inroducon o Team Decson Theory Decomposon of he Dynamc Team Decson Problem Equvalence of Sac and Dynamc

More information

Econ107 Applied Econometrics Topic 5: Specification: Choosing Independent Variables (Studenmund, Chapter 6)

Econ107 Applied Econometrics Topic 5: Specification: Choosing Independent Variables (Studenmund, Chapter 6) Econ7 Appled Economercs Topc 5: Specfcaon: Choosng Independen Varables (Sudenmund, Chaper 6 Specfcaon errors ha we wll deal wh: wrong ndependen varable; wrong funconal form. Ths lecure deals wh wrong ndependen

More information

Robustness Experiments with Two Variance Components

Robustness Experiments with Two Variance Components Naonal Insue of Sandards and Technology (NIST) Informaon Technology Laboraory (ITL) Sascal Engneerng Dvson (SED) Robusness Expermens wh Two Varance Componens by Ana Ivelsse Avlés avles@ns.gov Conference

More information

THE PREDICTION OF COMPETITIVE ENVIRONMENT IN BUSINESS

THE PREDICTION OF COMPETITIVE ENVIRONMENT IN BUSINESS THE PREICTION OF COMPETITIVE ENVIRONMENT IN BUSINESS INTROUCTION The wo dmensonal paral dfferenal equaons of second order can be used for he smulaon of compeve envronmen n busness The arcle presens he

More information

Comb Filters. Comb Filters

Comb Filters. Comb Filters The smple flers dscussed so far are characered eher by a sngle passband and/or a sngle sopband There are applcaons where flers wh mulple passbands and sopbands are requred Thecomb fler s an example of

More information

Linear Response Theory: The connection between QFT and experiments

Linear Response Theory: The connection between QFT and experiments Phys540.nb 39 3 Lnear Response Theory: The connecon beween QFT and expermens 3.1. Basc conceps and deas Q: ow do we measure he conducvy of a meal? A: we frs nroduce a weak elecrc feld E, and hen measure

More information

Political Economy of Institutions and Development: Problem Set 2 Due Date: Thursday, March 15, 2019.

Political Economy of Institutions and Development: Problem Set 2 Due Date: Thursday, March 15, 2019. Polcal Economy of Insuons and Developmen: 14.773 Problem Se 2 Due Dae: Thursday, March 15, 2019. Please answer Quesons 1, 2 and 3. Queson 1 Consder an nfne-horzon dynamc game beween wo groups, an ele and

More information

( ) () we define the interaction representation by the unitary transformation () = ()

( ) () we define the interaction representation by the unitary transformation () = () Hgher Order Perurbaon Theory Mchael Fowler 3/7/6 The neracon Represenaon Recall ha n he frs par of hs course sequence, we dscussed he chrödnger and Hesenberg represenaons of quanum mechancs here n he chrödnger

More information

Mechanics Physics 151

Mechanics Physics 151 Mechancs Physcs 5 Lecure 9 Hamlonan Equaons of Moon (Chaper 8) Wha We Dd Las Tme Consruced Hamlonan formalsm H ( q, p, ) = q p L( q, q, ) H p = q H q = p H = L Equvalen o Lagrangan formalsm Smpler, bu

More information

Mechanics Physics 151

Mechanics Physics 151 Mechancs Physcs 5 Lecure 9 Hamlonan Equaons of Moon (Chaper 8) Wha We Dd Las Tme Consruced Hamlonan formalsm Hqp (,,) = qp Lqq (,,) H p = q H q = p H L = Equvalen o Lagrangan formalsm Smpler, bu wce as

More information

2 Aggregate demand in partial equilibrium static framework

2 Aggregate demand in partial equilibrium static framework Unversy of Mnnesoa 8107 Macroeconomc Theory, Sprng 2009, Mn 1 Fabrzo Perr Lecure 1. Aggregaon 1 Inroducon Probably so far n he macro sequence you have deal drecly wh represenave consumers and represenave

More information

Technical Appendix to The Equivalence of Wage and Price Staggering in Monetary Business Cycle Models

Technical Appendix to The Equivalence of Wage and Price Staggering in Monetary Business Cycle Models Techncal Appendx o The Equvalence of Wage and Prce Saggerng n Moneary Busness Cycle Models Rochelle M. Edge Dvson of Research and Sascs Federal Reserve Board Sepember 24, 2 Absrac Ths appendx deals he

More information

Let s treat the problem of the response of a system to an applied external force. Again,

Let s treat the problem of the response of a system to an applied external force. Again, Page 33 QUANTUM LNEAR RESPONSE FUNCTON Le s rea he problem of he response of a sysem o an appled exernal force. Agan, H() H f () A H + V () Exernal agen acng on nernal varable Hamlonan for equlbrum sysem

More information

Advanced Macroeconomics II: Exchange economy

Advanced Macroeconomics II: Exchange economy Advanced Macroeconomcs II: Exchange economy Krzyszof Makarsk 1 Smple deermnsc dynamc model. 1.1 Inroducon Inroducon Smple deermnsc dynamc model. Defnons of equlbrum: Arrow-Debreu Sequenal Recursve Equvalence

More information

An introduction to Support Vector Machine

An introduction to Support Vector Machine An nroducon o Suppor Vecor Machne 報告者 : 黃立德 References: Smon Haykn, "Neural Neworks: a comprehensve foundaon, second edon, 999, Chaper 2,6 Nello Chrsann, John Shawe-Tayer, An Inroducon o Suppor Vecor Machnes,

More information

On One Analytic Method of. Constructing Program Controls

On One Analytic Method of. Constructing Program Controls Appled Mahemacal Scences, Vol. 9, 05, no. 8, 409-407 HIKARI Ld, www.m-hkar.com hp://dx.do.org/0.988/ams.05.54349 On One Analyc Mehod of Consrucng Program Conrols A. N. Kvko, S. V. Chsyakov and Yu. E. Balyna

More information

Variants of Pegasos. December 11, 2009

Variants of Pegasos. December 11, 2009 Inroducon Varans of Pegasos SooWoong Ryu bshboy@sanford.edu December, 009 Youngsoo Cho yc344@sanford.edu Developng a new SVM algorhm s ongong research opc. Among many exng SVM algorhms, we wll focus on

More information

Technical Appendix for Central Bank Communication and Expectations Stabilization

Technical Appendix for Central Bank Communication and Expectations Stabilization Techncal Appendx for Cenral Bank Communcaon and Expecaons Sablzaon Sefano Eusep Federal Reserve Bank of New York Bruce Preson Columba Unversy and NBER Augus 0, 008 Absrac Ths echncal appendx provdes some

More information

( t) Outline of program: BGC1: Survival and event history analysis Oslo, March-May Recapitulation. The additive regression model

( t) Outline of program: BGC1: Survival and event history analysis Oslo, March-May Recapitulation. The additive regression model BGC1: Survval and even hsory analyss Oslo, March-May 212 Monday May 7h and Tuesday May 8h The addve regresson model Ørnulf Borgan Deparmen of Mahemacs Unversy of Oslo Oulne of program: Recapulaon Counng

More information

In the complete model, these slopes are ANALYSIS OF VARIANCE FOR THE COMPLETE TWO-WAY MODEL. (! i+1 -! i ) + [(!") i+1,q - [(!

In the complete model, these slopes are ANALYSIS OF VARIANCE FOR THE COMPLETE TWO-WAY MODEL. (! i+1 -! i ) + [(!) i+1,q - [(! ANALYSIS OF VARIANCE FOR THE COMPLETE TWO-WAY MODEL The frs hng o es n wo-way ANOVA: Is here neracon? "No neracon" means: The man effecs model would f. Ths n urn means: In he neracon plo (wh A on he horzonal

More information

Notes on the stability of dynamic systems and the use of Eigen Values.

Notes on the stability of dynamic systems and the use of Eigen Values. Noes on he sabl of dnamc ssems and he use of Egen Values. Source: Macro II course noes, Dr. Davd Bessler s Tme Seres course noes, zarads (999) Ineremporal Macroeconomcs chaper 4 & Techncal ppend, and Hamlon

More information

Department of Economics University of Toronto

Department of Economics University of Toronto Deparmen of Economcs Unversy of Torono ECO408F M.A. Economercs Lecure Noes on Heeroskedascy Heeroskedascy o Ths lecure nvolves lookng a modfcaons we need o make o deal wh he regresson model when some of

More information

2 Aggregate demand in partial equilibrium static framework

2 Aggregate demand in partial equilibrium static framework Unversy of Mnnesoa 8107 Macroeconomc Theory, Sprng 2012, Mn 1 Fabrzo Perr Lecure 1. Aggregaon 1 Inroducon Probably so far n he macro sequence you have deal drecly wh represenave consumers and represenave

More information

Ordinary Differential Equations in Neuroscience with Matlab examples. Aim 1- Gain understanding of how to set up and solve ODE s

Ordinary Differential Equations in Neuroscience with Matlab examples. Aim 1- Gain understanding of how to set up and solve ODE s Ordnary Dfferenal Equaons n Neuroscence wh Malab eamples. Am - Gan undersandng of how o se up and solve ODE s Am Undersand how o se up an solve a smple eample of he Hebb rule n D Our goal a end of class

More information

Mechanics Physics 151

Mechanics Physics 151 Mechancs Physcs 5 Lecure 0 Canoncal Transformaons (Chaper 9) Wha We Dd Las Tme Hamlon s Prncple n he Hamlonan formalsm Dervaon was smple δi δ Addonal end-pon consrans pq H( q, p, ) d 0 δ q ( ) δq ( ) δ

More information

Online Supplement for Dynamic Multi-Technology. Production-Inventory Problem with Emissions Trading

Online Supplement for Dynamic Multi-Technology. Production-Inventory Problem with Emissions Trading Onlne Supplemen for Dynamc Mul-Technology Producon-Invenory Problem wh Emssons Tradng by We Zhang Zhongsheng Hua Yu Xa and Baofeng Huo Proof of Lemma For any ( qr ) Θ s easy o verfy ha he lnear programmng

More information

Lecture Notes 4. Univariate Forecasting and the Time Series Properties of Dynamic Economic Models

Lecture Notes 4. Univariate Forecasting and the Time Series Properties of Dynamic Economic Models Tme Seres Seven N. Durlauf Unversy of Wsconsn Lecure Noes 4. Unvarae Forecasng and he Tme Seres Properes of Dynamc Economc Models Ths se of noes presens does hree hngs. Frs, formulas are developed o descrbe

More information

CS434a/541a: Pattern Recognition Prof. Olga Veksler. Lecture 4

CS434a/541a: Pattern Recognition Prof. Olga Veksler. Lecture 4 CS434a/54a: Paern Recognon Prof. Olga Veksler Lecure 4 Oulne Normal Random Varable Properes Dscrmnan funcons Why Normal Random Varables? Analycally racable Works well when observaon comes form a corruped

More information

Volatility Interpolation

Volatility Interpolation Volaly Inerpolaon Prelmnary Verson March 00 Jesper Andreasen and Bran Huge Danse Mares, Copenhagen wan.daddy@danseban.com brno@danseban.com Elecronc copy avalable a: hp://ssrn.com/absrac=69497 Inro Local

More information

Lecture 1 The New Keynesian Model of Monetary Policy. Lecturer Campbell Leith University of Glasgow

Lecture 1 The New Keynesian Model of Monetary Policy. Lecturer Campbell Leith University of Glasgow Lecure The New Keynesan Model of Moneary Polcy Lecurer Campbell Leh Unversy of Glasgow The New Keynesan model of moneary polcy s becomng ncreasngly sandard n he analyss of moneary polcy. Ths parcular reamen

More information

New M-Estimator Objective Function. in Simultaneous Equations Model. (A Comparative Study)

New M-Estimator Objective Function. in Simultaneous Equations Model. (A Comparative Study) Inernaonal Mahemacal Forum, Vol. 8, 3, no., 7 - HIKARI Ld, www.m-hkar.com hp://dx.do.org/.988/mf.3.3488 New M-Esmaor Objecve Funcon n Smulaneous Equaons Model (A Comparave Sudy) Ahmed H. Youssef Professor

More information

Chapter 6: AC Circuits

Chapter 6: AC Circuits Chaper 6: AC Crcus Chaper 6: Oulne Phasors and he AC Seady Sae AC Crcus A sable, lnear crcu operang n he seady sae wh snusodal excaon (.e., snusodal seady sae. Complee response forced response naural response.

More information

Fall 2009 Social Sciences 7418 University of Wisconsin-Madison. Problem Set 2 Answers (4) (6) di = D (10)

Fall 2009 Social Sciences 7418 University of Wisconsin-Madison. Problem Set 2 Answers (4) (6) di = D (10) Publc Affars 974 Menze D. Chnn Fall 2009 Socal Scences 7418 Unversy of Wsconsn-Madson Problem Se 2 Answers Due n lecure on Thursday, November 12. " Box n" your answers o he algebrac quesons. 1. Consder

More information

CS286.2 Lecture 14: Quantum de Finetti Theorems II

CS286.2 Lecture 14: Quantum de Finetti Theorems II CS286.2 Lecure 14: Quanum de Fne Theorems II Scrbe: Mara Okounkova 1 Saemen of he heorem Recall he las saemen of he quanum de Fne heorem from he prevous lecure. Theorem 1 Quanum de Fne). Le ρ Dens C 2

More information

(,,, ) (,,, ). In addition, there are three other consumers, -2, -1, and 0. Consumer -2 has the utility function

(,,, ) (,,, ). In addition, there are three other consumers, -2, -1, and 0. Consumer -2 has the utility function MACROECONOMIC THEORY T J KEHOE ECON 87 SPRING 5 PROBLEM SET # Conder an overlappng generaon economy le ha n queon 5 on problem e n whch conumer lve for perod The uly funcon of he conumer born n perod,

More information

Appendix H: Rarefaction and extrapolation of Hill numbers for incidence data

Appendix H: Rarefaction and extrapolation of Hill numbers for incidence data Anne Chao Ncholas J Goell C seh lzabeh L ander K Ma Rober K Colwell and Aaron M llson 03 Rarefacon and erapolaon wh ll numbers: a framewor for samplng and esmaon n speces dversy sudes cology Monographs

More information

Math 128b Project. Jude Yuen

Math 128b Project. Jude Yuen Mah 8b Proec Jude Yuen . Inroducon Le { Z } be a sequence of observed ndependen vecor varables. If he elemens of Z have a on normal dsrbuon hen { Z } has a mean vecor Z and a varancecovarance marx z. Geomercally

More information

FTCS Solution to the Heat Equation

FTCS Solution to the Heat Equation FTCS Soluon o he Hea Equaon ME 448/548 Noes Gerald Reckenwald Porland Sae Unversy Deparmen of Mechancal Engneerng gerry@pdxedu ME 448/548: FTCS Soluon o he Hea Equaon Overvew Use he forward fne d erence

More information

Time-interval analysis of β decay. V. Horvat and J. C. Hardy

Time-interval analysis of β decay. V. Horvat and J. C. Hardy Tme-nerval analyss of β decay V. Horva and J. C. Hardy Work on he even analyss of β decay [1] connued and resuled n he developmen of a novel mehod of bea-decay me-nerval analyss ha produces hghly accurae

More information

FI 3103 Quantum Physics

FI 3103 Quantum Physics /9/4 FI 33 Quanum Physcs Aleander A. Iskandar Physcs of Magnesm and Phooncs Research Grou Insu Teknolog Bandung Basc Conces n Quanum Physcs Probably and Eecaon Value Hesenberg Uncerany Prncle Wave Funcon

More information

Online Appendix for. Strategic safety stocks in supply chains with evolving forecasts

Online Appendix for. Strategic safety stocks in supply chains with evolving forecasts Onlne Appendx for Sraegc safey socs n supply chans wh evolvng forecass Tor Schoenmeyr Sephen C. Graves Opsolar, Inc. 332 Hunwood Avenue Hayward, CA 94544 A. P. Sloan School of Managemen Massachuses Insue

More information

Outline. Probabilistic Model Learning. Probabilistic Model Learning. Probabilistic Model for Time-series Data: Hidden Markov Model

Outline. Probabilistic Model Learning. Probabilistic Model Learning. Probabilistic Model for Time-series Data: Hidden Markov Model Probablsc Model for Tme-seres Daa: Hdden Markov Model Hrosh Mamsuka Bonformacs Cener Kyoo Unversy Oulne Three Problems for probablsc models n machne learnng. Compung lkelhood 2. Learnng 3. Parsng (predcon

More information

Lecture 6: Learning for Control (Generalised Linear Regression)

Lecture 6: Learning for Control (Generalised Linear Regression) Lecure 6: Learnng for Conrol (Generalsed Lnear Regresson) Conens: Lnear Mehods for Regresson Leas Squares, Gauss Markov heorem Recursve Leas Squares Lecure 6: RLSC - Prof. Sehu Vjayakumar Lnear Regresson

More information

. The geometric multiplicity is dim[ker( λi. number of linearly independent eigenvectors associated with this eigenvalue.

. The geometric multiplicity is dim[ker( λi. number of linearly independent eigenvectors associated with this eigenvalue. Lnear Algebra Lecure # Noes We connue wh he dscusson of egenvalues, egenvecors, and dagonalzably of marces We wan o know, n parcular wha condons wll assure ha a marx can be dagonalzed and wha he obsrucons

More information

TSS = SST + SSE An orthogonal partition of the total SS

TSS = SST + SSE An orthogonal partition of the total SS ANOVA: Topc 4. Orhogonal conrass [ST&D p. 183] H 0 : µ 1 = µ =... = µ H 1 : The mean of a leas one reamen group s dfferen To es hs hypohess, a basc ANOVA allocaes he varaon among reamen means (SST) equally

More information

Robust and Accurate Cancer Classification with Gene Expression Profiling

Robust and Accurate Cancer Classification with Gene Expression Profiling Robus and Accurae Cancer Classfcaon wh Gene Expresson Proflng (Compuaonal ysems Bology, 2005) Auhor: Hafeng L, Keshu Zhang, ao Jang Oulne Background LDA (lnear dscrmnan analyss) and small sample sze problem

More information

P R = P 0. The system is shown on the next figure:

P R = P 0. The system is shown on the next figure: TPG460 Reservor Smulaon 08 page of INTRODUCTION TO RESERVOIR SIMULATION Analycal and numercal soluons of smple one-dmensonal, one-phase flow equaons As an nroducon o reservor smulaon, we wll revew he smples

More information

EEL 6266 Power System Operation and Control. Chapter 5 Unit Commitment

EEL 6266 Power System Operation and Control. Chapter 5 Unit Commitment EEL 6266 Power Sysem Operaon and Conrol Chaper 5 Un Commmen Dynamc programmng chef advanage over enumeraon schemes s he reducon n he dmensonaly of he problem n a src prory order scheme, here are only N

More information

Reactive Methods to Solve the Berth AllocationProblem with Stochastic Arrival and Handling Times

Reactive Methods to Solve the Berth AllocationProblem with Stochastic Arrival and Handling Times Reacve Mehods o Solve he Berh AllocaonProblem wh Sochasc Arrval and Handlng Tmes Nsh Umang* Mchel Berlare* * TRANSP-OR, Ecole Polyechnque Fédérale de Lausanne Frs Workshop on Large Scale Opmzaon November

More information

Cubic Bezier Homotopy Function for Solving Exponential Equations

Cubic Bezier Homotopy Function for Solving Exponential Equations Penerb Journal of Advanced Research n Compung and Applcaons ISSN (onlne: 46-97 Vol. 4, No.. Pages -8, 6 omoopy Funcon for Solvng Eponenal Equaons S. S. Raml *,,. Mohamad Nor,a, N. S. Saharzan,b and M.

More information

Midterm Exam. Thursday, April hour, 15 minutes

Midterm Exam. Thursday, April hour, 15 minutes Economcs of Grow, ECO560 San Francsco Sae Unvers Mcael Bar Sprng 04 Mderm Exam Tursda, prl 0 our, 5 mnues ame: Insrucons. Ts s closed boo, closed noes exam.. o calculaors of an nd are allowed. 3. Sow all

More information

There are a total of two problems, each with multiple subparts.

There are a total of two problems, each with multiple subparts. eparmen of Economcs Boson College Economcs 0 (Secon 05) acroeconomc Theory Problem Se Suggesed Soluons Professor Sanjay Chugh Fall 04 ue: ecember 9, 04 (no laer han :30pm) Insrucons: Clearly-wren (yped

More information

Knowing What Others Know: Coordination Motives in Information Acquisition Additional Notes

Knowing What Others Know: Coordination Motives in Information Acquisition Additional Notes Knowng Wha Ohers Know: Coordnaon Moves n nformaon Acquson Addonal Noes Chrsan Hellwg Unversy of Calforna, Los Angeles Deparmen of Economcs Laura Veldkamp New York Unversy Sern School of Busness March 1,

More information

Lecture 18: The Laplace Transform (See Sections and 14.7 in Boas)

Lecture 18: The Laplace Transform (See Sections and 14.7 in Boas) Lecure 8: The Lalace Transform (See Secons 88- and 47 n Boas) Recall ha our bg-cure goal s he analyss of he dfferenal equaon, ax bx cx F, where we emloy varous exansons for he drvng funcon F deendng on

More information

[ ] 2. [ ]3 + (Δx i + Δx i 1 ) / 2. Δx i-1 Δx i Δx i+1. TPG4160 Reservoir Simulation 2018 Lecture note 3. page 1 of 5

[ ] 2. [ ]3 + (Δx i + Δx i 1 ) / 2. Δx i-1 Δx i Δx i+1. TPG4160 Reservoir Simulation 2018 Lecture note 3. page 1 of 5 TPG460 Reservor Smulaon 08 page of 5 DISCRETIZATIO OF THE FOW EQUATIOS As we already have seen, fne dfference appromaons of he paral dervaves appearng n he flow equaons may be obaned from Taylor seres

More information

Should Exact Index Numbers have Standard Errors? Theory and Application to Asian Growth

Should Exact Index Numbers have Standard Errors? Theory and Application to Asian Growth Should Exac Index umbers have Sandard Errors? Theory and Applcaon o Asan Growh Rober C. Feensra Marshall B. Rensdorf ovember 003 Proof of Proposon APPEDIX () Frs, we wll derve he convenonal Sao-Vara prce

More information

Problem 1 / 25 Problem 2 / 15 Problem 3 / 15 Problem 4 / 20 Problem 5 / 25 TOTAL / 100

Problem 1 / 25 Problem 2 / 15 Problem 3 / 15 Problem 4 / 20 Problem 5 / 25 TOTAL / 100 Deparmen of Appled Economcs Johns Hopkns Unversy Economcs 60 Macroeconomc Theory and Polcy Fnal Exam Suggesed Soluons Professor Sanjay Chugh Fall 009 NAME: The Exam has a oal of fve (5) problems and pages

More information

GENERATING CERTAIN QUINTIC IRREDUCIBLE POLYNOMIALS OVER FINITE FIELDS. Youngwoo Ahn and Kitae Kim

GENERATING CERTAIN QUINTIC IRREDUCIBLE POLYNOMIALS OVER FINITE FIELDS. Youngwoo Ahn and Kitae Kim Korean J. Mah. 19 (2011), No. 3, pp. 263 272 GENERATING CERTAIN QUINTIC IRREDUCIBLE POLYNOMIALS OVER FINITE FIELDS Youngwoo Ahn and Kae Km Absrac. In he paper [1], an explc correspondence beween ceran

More information

Optimal environmental charges under imperfect compliance

Optimal environmental charges under imperfect compliance ISSN 1 746-7233, England, UK World Journal of Modellng and Smulaon Vol. 4 (28) No. 2, pp. 131-139 Opmal envronmenal charges under mperfec complance Dajn Lu 1, Ya Wang 2 Tazhou Insue of Scence and Technology,

More information

. The geometric multiplicity is dim[ker( λi. A )], i.e. the number of linearly independent eigenvectors associated with this eigenvalue.

. The geometric multiplicity is dim[ker( λi. A )], i.e. the number of linearly independent eigenvectors associated with this eigenvalue. Mah E-b Lecure #0 Noes We connue wh he dscusson of egenvalues, egenvecors, and dagonalzably of marces We wan o know, n parcular wha condons wll assure ha a marx can be dagonalzed and wha he obsrucons are

More information

Exchange Rate Policy: The euro area, the US, and Asia

Exchange Rate Policy: The euro area, the US, and Asia Exchange Rae Polcy: The euro area, he, and Leonor Counho Unversy of Cyprus, Economcs Deparmen June, 2009 Absrac Ths paper uses a hree-counry model of he, he euro area and, o analyze alernave polcy responses

More information

Data Collection Definitions of Variables - Conceptualize vs Operationalize Sample Selection Criteria Source of Data Consistency of Data

Data Collection Definitions of Variables - Conceptualize vs Operationalize Sample Selection Criteria Source of Data Consistency of Data Apply Sascs and Economercs n Fnancal Research Obj. of Sudy & Hypoheses Tesng From framework objecves of sudy are needed o clarfy, hen, n research mehodology he hypoheses esng are saed, ncludng esng mehods.

More information

Bundling with Customer Self-Selection: A Simple Approach to Bundling Low Marginal Cost Goods On-Line Appendix

Bundling with Customer Self-Selection: A Simple Approach to Bundling Low Marginal Cost Goods On-Line Appendix Bundlng wh Cusomer Self-Selecon: A Smple Approach o Bundlng Low Margnal Cos Goods On-Lne Appendx Lorn M. H Unversy of Pennsylvana, Wharon School 57 Jon M. Hunsman Hall Phladelpha, PA 94 lh@wharon.upenn.edu

More information

Lecture 11 SVM cont

Lecture 11 SVM cont Lecure SVM con. 0 008 Wha we have done so far We have esalshed ha we wan o fnd a lnear decson oundary whose margn s he larges We know how o measure he margn of a lnear decson oundary Tha s: he mnmum geomerc

More information

Lecture VI Regression

Lecture VI Regression Lecure VI Regresson (Lnear Mehods for Regresson) Conens: Lnear Mehods for Regresson Leas Squares, Gauss Markov heorem Recursve Leas Squares Lecure VI: MLSC - Dr. Sehu Vjayakumar Lnear Regresson Model M

More information

Machine Learning Linear Regression

Machine Learning Linear Regression Machne Learnng Lnear Regresson Lesson 3 Lnear Regresson Bascs of Regresson Leas Squares esmaon Polynomal Regresson Bass funcons Regresson model Regularzed Regresson Sascal Regresson Mamum Lkelhood (ML)

More information

Fall 2010 Graduate Course on Dynamic Learning

Fall 2010 Graduate Course on Dynamic Learning Fall 200 Graduae Course on Dynamc Learnng Chaper 4: Parcle Flers Sepember 27, 200 Byoung-Tak Zhang School of Compuer Scence and Engneerng & Cognve Scence and Bran Scence Programs Seoul aonal Unversy hp://b.snu.ac.kr/~bzhang/

More information

INTERNATIONAL MONETARY POLICY ANALYSIS WITH DURABLE GOODS

INTERNATIONAL MONETARY POLICY ANALYSIS WITH DURABLE GOODS INTERNATIONAL MONETARY POLICY ANALYSIS WIT DURABLE GOODS A Dsseraon by KANG KOO LEE Submed o he Offce of Graduae Sudes of Texas A&M Unversy n paral fulfllmen of he reuremens for he degree of DOCTOR O PILOSOPY

More information

( ) [ ] MAP Decision Rule

( ) [ ] MAP Decision Rule Announcemens Bayes Decson Theory wh Normal Dsrbuons HW0 due oday HW o be assgned soon Proec descrpon posed Bomercs CSE 90 Lecure 4 CSE90, Sprng 04 CSE90, Sprng 04 Key Probables 4 ω class label X feaure

More information

Dual Approximate Dynamic Programming for Large Scale Hydro Valleys

Dual Approximate Dynamic Programming for Large Scale Hydro Valleys Dual Approxmae Dynamc Programmng for Large Scale Hydro Valleys Perre Carpener and Jean-Phlppe Chanceler 1 ENSTA ParsTech and ENPC ParsTech CMM Workshop, January 2016 1 Jon work wh J.-C. Alas, suppored

More information

Density Matrix Description of NMR BCMB/CHEM 8190

Density Matrix Description of NMR BCMB/CHEM 8190 Densy Marx Descrpon of NMR BCMBCHEM 89 Operaors n Marx Noaon Alernae approach o second order specra: ask abou x magnezaon nsead of energes and ranson probables. If we say wh one bass se, properes vary

More information

[Link to MIT-Lab 6P.1 goes here.] After completing the lab, fill in the following blanks: Numerical. Simulation s Calculations

[Link to MIT-Lab 6P.1 goes here.] After completing the lab, fill in the following blanks: Numerical. Simulation s Calculations Chaper 6: Ordnary Leas Squares Esmaon Procedure he Properes Chaper 6 Oulne Cln s Assgnmen: Assess he Effec of Sudyng on Quz Scores Revew o Regresson Model o Ordnary Leas Squares () Esmaon Procedure o he

More information

J i-1 i. J i i+1. Numerical integration of the diffusion equation (I) Finite difference method. Spatial Discretization. Internal nodes.

J i-1 i. J i i+1. Numerical integration of the diffusion equation (I) Finite difference method. Spatial Discretization. Internal nodes. umercal negraon of he dffuson equaon (I) Fne dfference mehod. Spaal screaon. Inernal nodes. R L V For hermal conducon le s dscree he spaal doman no small fne spans, =,,: Balance of parcles for an nernal

More information

THEORETICAL AUTOCORRELATIONS. ) if often denoted by γ. Note that

THEORETICAL AUTOCORRELATIONS. ) if often denoted by γ. Note that THEORETICAL AUTOCORRELATIONS Cov( y, y ) E( y E( y))( y E( y)) ρ = = Var( y) E( y E( y)) =,, L ρ = and Cov( y, y ) s ofen denoed by whle Var( y ) f ofen denoed by γ. Noe ha γ = γ and ρ = ρ and because

More information

Density Matrix Description of NMR BCMB/CHEM 8190

Density Matrix Description of NMR BCMB/CHEM 8190 Densy Marx Descrpon of NMR BCMBCHEM 89 Operaors n Marx Noaon If we say wh one bass se, properes vary only because of changes n he coeffcens weghng each bass se funcon x = h< Ix > - hs s how we calculae

More information

January Examinations 2012

January Examinations 2012 Page of 5 EC79 January Examnaons No. of Pages: 5 No. of Quesons: 8 Subjec ECONOMICS (POSTGRADUATE) Tle of Paper EC79 QUANTITATIVE METHODS FOR BUSINESS AND FINANCE Tme Allowed Two Hours ( hours) Insrucons

More information

2. SPATIALLY LAGGED DEPENDENT VARIABLES

2. SPATIALLY LAGGED DEPENDENT VARIABLES 2. SPATIALLY LAGGED DEPENDENT VARIABLES In hs chaper, we descrbe a sascal model ha ncorporaes spaal dependence explcly by addng a spaally lagged dependen varable y on he rgh-hand sde of he regresson equaon.

More information

CHAPTER 10: LINEAR DISCRIMINATION

CHAPTER 10: LINEAR DISCRIMINATION CHAPER : LINEAR DISCRIMINAION Dscrmnan-based Classfcaon 3 In classfcaon h K classes (C,C,, C k ) We defned dscrmnan funcon g j (), j=,,,k hen gven an es eample, e chose (predced) s class label as C f g

More information

HEAT CONDUCTION PROBLEM IN A TWO-LAYERED HOLLOW CYLINDER BY USING THE GREEN S FUNCTION METHOD

HEAT CONDUCTION PROBLEM IN A TWO-LAYERED HOLLOW CYLINDER BY USING THE GREEN S FUNCTION METHOD Journal of Appled Mahemacs and Compuaonal Mechancs 3, (), 45-5 HEAT CONDUCTION PROBLEM IN A TWO-LAYERED HOLLOW CYLINDER BY USING THE GREEN S FUNCTION METHOD Sansław Kukla, Urszula Sedlecka Insue of Mahemacs,

More information

Welfare Maximizing Operational Monetary and Tax Policy Rules

Welfare Maximizing Operational Monetary and Tax Policy Rules Forhcomng n Macroeconomc Dynamcs (also publshed n 24 as CEPR DP 4782) Welfare Maxmzng Operaonal Moneary and Tax Polcy Rules Rober Kollmann (*) Deparmen of Economcs, Unversy of Pars XII 6, Av. du Général

More information

Motion in Two Dimensions

Motion in Two Dimensions Phys 1 Chaper 4 Moon n Two Dmensons adzyubenko@csub.edu hp://www.csub.edu/~adzyubenko 005, 014 A. Dzyubenko 004 Brooks/Cole 1 Dsplacemen as a Vecor The poson of an objec s descrbed by s poson ecor, r The

More information

F-Tests and Analysis of Variance (ANOVA) in the Simple Linear Regression Model. 1. Introduction

F-Tests and Analysis of Variance (ANOVA) in the Simple Linear Regression Model. 1. Introduction ECOOMICS 35* -- OTE 9 ECO 35* -- OTE 9 F-Tess and Analyss of Varance (AOVA n he Smple Lnear Regresson Model Inroducon The smple lnear regresson model s gven by he followng populaon regresson equaon, or

More information

The Dynamic Programming Models for Inventory Control System with Time-varying Demand

The Dynamic Programming Models for Inventory Control System with Time-varying Demand The Dynamc Programmng Models for Invenory Conrol Sysem wh Tme-varyng Demand Truong Hong Trnh (Correspondng auhor) The Unversy of Danang, Unversy of Economcs, Venam Tel: 84-236-352-5459 E-mal: rnh.h@due.edu.vn

More information

RELATIONSHIP BETWEEN VOLATILITY AND TRADING VOLUME: THE CASE OF HSI STOCK RETURNS DATA

RELATIONSHIP BETWEEN VOLATILITY AND TRADING VOLUME: THE CASE OF HSI STOCK RETURNS DATA RELATIONSHIP BETWEEN VOLATILITY AND TRADING VOLUME: THE CASE OF HSI STOCK RETURNS DATA Mchaela Chocholaá Unversy of Economcs Braslava, Slovaka Inroducon (1) one of he characersc feaures of sock reurns

More information

CH.3. COMPATIBILITY EQUATIONS. Continuum Mechanics Course (MMC) - ETSECCPB - UPC

CH.3. COMPATIBILITY EQUATIONS. Continuum Mechanics Course (MMC) - ETSECCPB - UPC CH.3. COMPATIBILITY EQUATIONS Connuum Mechancs Course (MMC) - ETSECCPB - UPC Overvew Compably Condons Compably Equaons of a Poenal Vecor Feld Compably Condons for Infnesmal Srans Inegraon of he Infnesmal

More information

The Nonlinear Phillips Curve and Inflation Forecast Targeting - Symmetric Versus Asymmetric Monetary Policy Rules Schaling, E.

The Nonlinear Phillips Curve and Inflation Forecast Targeting - Symmetric Versus Asymmetric Monetary Policy Rules Schaling, E. Tlburg Unversy The Nonlnear Phllps Curve and Inflaon Forecas Targeng - Symmerc Versus Asymmerc Moneary Polcy Rules Schalng, E. Publcaon dae: 998 Lnk o publcaon Caon for publshed verson (APA: Schalng, E.

More information

Pendulum Dynamics. = Ft tangential direction (2) radial direction (1)

Pendulum Dynamics. = Ft tangential direction (2) radial direction (1) Pendulum Dynams Consder a smple pendulum wh a massless arm of lengh L and a pon mass, m, a he end of he arm. Assumng ha he fron n he sysem s proporonal o he negave of he angenal veloy, Newon s seond law

More information

Advanced time-series analysis (University of Lund, Economic History Department)

Advanced time-series analysis (University of Lund, Economic History Department) Advanced me-seres analss (Unvers of Lund, Economc Hsor Dearmen) 3 Jan-3 Februar and 6-3 March Lecure 4 Economerc echnues for saonar seres : Unvarae sochasc models wh Box- Jenns mehodolog, smle forecasng

More information

Forecasting customer behaviour in a multi-service financial organisation: a profitability perspective

Forecasting customer behaviour in a multi-service financial organisation: a profitability perspective Forecasng cusomer behavour n a mul-servce fnancal organsaon: a profably perspecve A. Audzeyeva, Unversy of Leeds & Naonal Ausrala Group Europe, UK B. Summers, Unversy of Leeds, UK K.R. Schenk-Hoppé, Unversy

More information

Relative controllability of nonlinear systems with delays in control

Relative controllability of nonlinear systems with delays in control Relave conrollably o nonlnear sysems wh delays n conrol Jerzy Klamka Insue o Conrol Engneerng, Slesan Techncal Unversy, 44- Glwce, Poland. phone/ax : 48 32 37227, {jklamka}@a.polsl.glwce.pl Keywor: Conrollably.

More information

Expectations and Exchange Rate Policy

Expectations and Exchange Rate Policy Expecaons and Exchange Rae Polcy Mchael B. Devereux UBC Charles Engel Wsconsn Aprl 24, 2006 Absrac Boh emprcal evdence and heorecal dscusson have long emphaszed he mpac of `news on exchange raes. In mos

More information

Clustering (Bishop ch 9)

Clustering (Bishop ch 9) Cluserng (Bshop ch 9) Reference: Daa Mnng by Margare Dunham (a slde source) 1 Cluserng Cluserng s unsupervsed learnng, here are no class labels Wan o fnd groups of smlar nsances Ofen use a dsance measure

More information

Disclosure Quality, Diversification and the Cost of Capital. Greg Clinch University of Melbourne June 2013

Disclosure Quality, Diversification and the Cost of Capital. Greg Clinch University of Melbourne June 2013 Dsclosure Qualy, Dversfcaon and he Cos of Capal Greg Clnch Unversy of Melbourne clnchg@unmelb.edu.au June 03 I hank Cynha Ca, Kevn L, and Sorabh Tomar for helpful commens and suggesons on an earler (ncomplee)

More information

Introduction ( Week 1-2) Course introduction A brief introduction to molecular biology A brief introduction to sequence comparison Part I: Algorithms

Introduction ( Week 1-2) Course introduction A brief introduction to molecular biology A brief introduction to sequence comparison Part I: Algorithms Course organzaon Inroducon Wee -2) Course nroducon A bref nroducon o molecular bology A bref nroducon o sequence comparson Par I: Algorhms for Sequence Analyss Wee 3-8) Chaper -3, Models and heores» Probably

More information

Advanced Machine Learning & Perception

Advanced Machine Learning & Perception Advanced Machne Learnng & Percepon Insrucor: Tony Jebara SVM Feaure & Kernel Selecon SVM Eensons Feaure Selecon (Flerng and Wrappng) SVM Feaure Selecon SVM Kernel Selecon SVM Eensons Classfcaon Feaure/Kernel

More information