Financial Market Integration, Exchange Rate Policy, and the Dynamics of Business and. Employment in Korea* Matteo Cacciatore a HEC Montréal and NBER

Transcription

1 Financial Marke Inegraion, Exchange Rae Policy, and he Dynamics of Business and Employmen in Korea Maeo Cacciaore a HEC Monréal and NBER Fabio Ghironi b Universiy of Washingon, CEPR, EABCN, and NBER Yurim Lee c Universiy of Washingon Sepember, 6 Absrac We sudy he consequences of differen degrees of inernaional financial marke inegraion and exchange rae policies in a calibraed, medium-scale model of he Korean economy. The model feaures endogenous producer enry ino domesic and expor markes and search-and-maching fricions in labor markes. This allows us o highligh he consequences of financial inegraion and he exchange rae regime for he dynamics of business creaion and unemploymen. We show ha, under flexible exchange raes, access o inernaional financial markes increases he volailiy of boh business creaion and he number of exporing plans, bu he effecs on employmen volailiy are more modes. Pegging he exchange rae can have unfavorable consequences for he effecs of erms of rade appreciaion, bu more financial inegraion is beneficial under a peg if he economy is subjec o boh produciviy and erms of rade shocks. The combinaion of a floaing exchange rae and inernaionally complee markes would be he bes scenario for Korea among hose we focus on. JEL Codes: E4, E5, F6, F3, F4, F44. Keywords: Business creaion; Exchange rae policy; Inernaional financial inegraion; Moneary policy; Risk sharing; Unemploymen We are graeful for helpful commens o Kosuke Aoki, Michal Fabinger and he paricipans in he NBER-CEPR-TCER (TRIO) Conference on Inernaional Finance in he Global Markes, Tokyo, December 6-7, 5. The views in his paper are hose of he auhors and do no necessarily reflec he views or policies of CEPR and NBER. a Insiue of Applied Economics, HEC Monréal, 3, chemin de la Côe-Saine-Caherine, Monréal (Québec), Canada. maeo.cacciaore@hec.ca. URL: hp:// b Deparmen of Economics, Universiy of Washingon, Savery Hall, Box 35333, Seale, WA 9895, U.S.A. ghiro@uw.edu. URL: hp://faculy.washingon.edu/ghiro. c Deparmen of Economics, Universiy of Washingon, Savery Hall, Box 35333, Seale, WA 9895, U.S.A. yrlee88@uw.edu.

2 . Inroducion Korea ransiioned from limied exchange rae flexibiliy and low inegraion in inernaional financial markes in he 98s o he curren environmen of a floaing won and high capial marke inegraion. In he process, i experienced a severe currency and banking crisis in , which precipiaed he adven of increased exchange rae flexibiliy. Much has been wrien on he crisis ha engulfed Korea and oher Asian counries, and on he ineracion of policy choices wih respec o capial accoun opening and exchange rae policy in exposing counries o such evens. Even absracing from he opic of crises, he effecs of inernaional financial marke inegraion and exchange rae policies are classic subjecs of sudy in inernaional macroeconomics. Imporanly, hese sudies ypically absrac from feaures of economic dynamics ha are becoming increasingly acceped as necessary ingrediens for empirically relevan posiive analysis and for policy conversaion: producer enry ino domesic and expor markes, and labor marke fricions ha resul in unemploymen. 3 The purpose of his paper is o use a medium-scale, dynamic, sochasic, general equilibrium (DSGE) model ha feaures hese ingrediens in addiion o he sandard ingrediens of New Keynesian open economy macroeconomics o shed ligh on he consequences of differen decisions wih respec o inernaional financial marke inegraion and exchange rae policy for Korea. The exercise allows us o highligh he imporance for resuls of channels hihero unexplored in he lieraure and ha sugges ineresing avenues for furher heoreical and empirical exploraion. The model we use is a small open economy version of he benchmark framework for analysis of macro inerdependence and moneary policy wih micro-level dynamics developed by Cacciaore and Ghironi (). In he model, monopolisically compeiive producers decide endogenously on he number of plans (or produc lines) hey operae subjec o sunk coss of new produc creaion. Plans are heerogeneous in heir produciviies and face fixed expor coss as in Meliz (3). Eichengreen (4a, 8) summarizes evens and explanaions and provides references o much lieraure. See Heahcoe and Perri (4) for a survey of lieraure on inernaional financial marke inegraion and risk sharing and Corsei, Dedola, and Leduc () for a survey on moneary policy in open economies. 3 References o much recen lieraure ha inroduces one or boh of hese ingrediens in analyses of macroeconomic flucuaions and/or policy problems can be found in Cacciaore and Ghironi ().

3 Therefore, only he producs of sufficienly producive plans are expored abroad. These micro-level producer dynamics, which have become he benchmark for inernaional rade analysis, are combined wih search-and-maching labor marke fricions as in Diamond (98a,b) and Morensen and Pissarides (994). Hiring workers requires firms o pos vacancies and incur coss of vacancy posing. A sandard maching echnology ranslaes he number of aggregae vacancies and he aggregae unemploymen rae ino new job maches in each period. Wages are deermined by Nash bargaining beween workers and firms. These ingrediens are combined wih he sandard assumpions of wage and price sickiness of a vas New Keynesian lieraure o complee he seup for our exercises. We consider hree scenarios for inernaional financial inegraion and wo possibiliies for moneary policy: financial auarky, inernaionally incomplee markes wih asse rade resriced o nominal bonds, and complee markes, under an exchange rae peg or a floa. In he laer case for exchange rae policy, we assume ha moneary policy ses he ineres rae according o an empirically plausible ineres rae rule as a funcion of inflaion and he oupu gap. We ake he case of incomplee markes under flexible exchange raes as benchmark for comparison of he model s properies o hose of he daa on Korea s business cycle in he period Even if i fails o generae a counercyclical rade balance, he model does well on several dimensions. We hen ackle he following quesion: Suppose Korea operaes under he floaing won regime, bu i does i under financial auarky or complee markes. How would is macroeconomic dynamics differ and wha would be he consequences for welfare? We find ha access o inernaional financial markes increases he volailiy of boh business creaion (producer enry) and he number of exporing plans, bu he effecs on employmen volailiy are more modes. Financial inegraion implies beer consumpion risk sharing, wih lower consumpion volailiy and higher correlaion of consumpion wih he res of he world. As a consequence, welfare coss of business cycles for Korea become significanly lower. Beer access o inernaional finance also implies more volaile business creaion: A given expansionary shock resuls in a larger increase in he number of enrans in he domesic economy. However,

4 since consumpion increases by less (owing o beer smoohing), he combined effec of increasing financial inegraion on he response of employmen o he shock is mued. The nex exercise ha we perform sudies how financial inegraion and he exchange rae regime inerac o shape macroeconomic dynamics in response o shocks and he welfare implicaions of differen scenarios. We find ha an exchange rae peg can have unfavorable consequences especially for he effecs of shocks ha cause erms of rade appreciaion, as pegging consrains he abiliy of policy o counerac exogenous losses of compeiiveness. A he same ime, more financial inegraion is beneficial under a peg if he economy is subjec o boh produciviy and erms of rade shocks. In paricular, given an exogenous compeiiveness loss, he abiliy o borrow inernaionally dampens he increase in unemploymen and he conracion of GDP, and i makes i possible o ake advanage of cheaper impors by increasing consumpion. On welfare grounds, he combinaion of a floaing exchange rae and full insurance in inernaionally complee markes is he bes scenario for Korea among hose we considered. To he bes of our knowledge, our paper makes a novel conribuion o he lieraure by highlighing he role of producer-level dynamics and labor marke fricions in shaping he consequences of differen scenarios for inernaional financial marke inegraion and he exchange rae regime. However, here is much ha he paper does no do: While i suggess mechanisms ha can make an exchange rae peg undesirable, he paper does no include any modeling of he crisis, and i should no be inerpreed as a heory of he crisis. This is so also because he model absracs from financial marke fricions oher han in he menu of inernaionally raded asses, and i is well known ha he banking secor played a crucial role in he crisis. The model depars from he benchmark New Keynesian small open economy framework (as developed, for insance, by Galí and Monacelli, 5) only by inroducing richer micro-level producer dynamics and search-and-maching fricions in labor markes. By doing so, we are absracing from many model feaures of poenial relevance o he specific case of Korea. We are making his choice inenionally: A vas lieraure has applied he off-he-shelf Galí-Monacelli-ype model o a large number of vasly heerogeneous counries. Our ineres here is in performing a similar exercise and 3

5 applying wha we view as a benchmark New Keynesian small open economy model wih producer dynamics and labor marke fricions o he case of Korea. Resricing he exension of he basic framework o he ingrediens we focus on allows us o hone in mos ransparenly on he novel resuls we can learn from he join inroducion of hese model feaures. The developmen of model versions ha incorporae more Korea-specific feaures is an obvious exension of our analysis ha we leave for fuure research. Finally, he paper also does no provide an empirical assessmen of he imporance for Korea of he producer-level dynamics we highligh. Borrowing he erminology of Presco (986), we presen heory ahead of business cycle measuremen, which will require longer ime series of exensive margin daa for rigorous esing han hose currenly available. We view he consrucion of such daa and is use in empirical analysis of he mechanisms we explore as a major ask for fuure research. 4 The res of he paper is organized as follows. Secion presens he model. Secion 3 discusses he properies of he model by presening impulse responses o a produciviy shock and comparing model-generaed business cycle momens o hose of he daa. Secion 4 sudies he consequences of differen degrees of inernaional financial marke inegraion under flexible exchange raes. Secion 5 focuses on he consequences of erms of rade shocks. Secion 6 analyzes he combined effecs of differen exchange rae and financial marke inegraion scenarios. Secion 7 summarizes sensiiviy analysis exercises. Secion 8 concludes. An Appendix presens echnical deails.. The Model The model we employ is an applicaion of he framework developed by Cacciaore and Ghironi (). The difference is ha Korea is a prooype small open economy. As is now sandard pracice in he lieraure, we model he small open economy as a limiing case of a wo-counry dynamic general 4 See Bilbiie, Ghironi, and Meliz () and Cacciaore and Ghironi () for references o lieraure ha suppors he relevance for economic dynamics of he mechanisms we highligh in counries oher han Korea. 4

6 equilibrium model in which one counry (he small open economy, also referred o as Home) is of measure zero relaive o he res of he world (Foreign henceforh). As a consequence, he policy decisions and macroeconomic dynamics of he small open economy have no impac on Foreign. Nex we describe in deail he behavior of households and firms in he small open economy. 5. Household Preferences The small open economy is populaed by a uni mass of aomisic households, where each household is viewed as an exended family wih a coninuum of members along he uni inerval. In equilibrium, some family members are employed, while ohers are unemployed. As is common in he lieraure, we assume ha family members insure each oher perfecly agains variaions in labor income due o changes in employmen saus, so ha here is no ex pos heerogeneiy across individuals in he household (see Andolfao, 996, and Merz, 995). 6 The represenaive household in he Home economy maximizes he expeced ineremporal uiliy funcion E uc lv h, where, is he discoun facor, C is a consumpion baske ha aggregaes domesic and impored goods as described below, l is he number of employed workers, and h denoes hours worked by each employed worker. Period uiliy from consumpion, u (.), and disuiliy of effor, v (.), saisfy he sandard assumpions. C The consumpion baske C aggregaes Home and Foreign secoral consumpion oupus, n, in Dixi-Sigliz (977) form: C C( n) dn, () where is he symmeric elasiciy of subsiuion across goods. The corresponding consumpion-based price index is given by: 5 Cacciaore, Ghironi, and Turnovsky (5) use he model presened here o sudy he consequences of srucural reforms and rade inegraion for New Zealand. 6 Nakajima () and Obiols-Homs () explore he consequences of marke incompleeness in he search-andmaching models of labor marke fricions. We reain he sandard assumpion of marke compleeness o preserve he represenaive household framework of benchmark New Keynesian modeling and focus mos ransparenly on he consequences of exending he framework o include producer dynamics and search-and-maching fricions. 5

7 P P n dn ( ), () where P n is he price index for secor n, expressed in Home currency.. Producion There are wo verically inegraed producion secors. In he upsream secor, perfecly compeiive firms use labor o produce a non-radable inermediae inpu. In he downsream secor, each consumpion-producing secor n is populaed by a represenaive monopolisically compeiive muliproduc firm ha purchases he inermediae inpu and produces differeniaed varieies of is secoral oupu. In equilibrium, some of hese varieies are expored while he ohers are sold only domesically. 7.. Inermediae Goods Producion There is a uni mass of inermediae producers. Each of hem employs a coninuum of workers. Labor markes are characerized by search and maching fricions as in he Diamond-Morensen-Pissarides (DMP) framework. 8 To hire new workers, firms need o pos vacancies, incurring a cos of unis of consumpion per vacancy posed. The probabiliy of finding a worker depends on a consan-reurno-scale maching echnology, which convers aggregae unemployed workers, U, and aggregae vacancies, V, ino aggregae maches, M U V, where and. Each firm mees unemployed workers a a rae q M / V. As in Krause and Lubik (7) and oher sudies, we assume ha newly creaed maches become producive only in he nex period. For an individual firm, he inflow of new hires in period is herefore q, where is he number of vacancies posed by he firm in period. 9 Firms and workers can separae exogenously wih probabiliy (,). Separaion occurs only beween firms and workers who were acive in producion in he previous period. As a resul, he 7 This producion srucure grealy simplifies he inroducion of labor marke fricions and sicky prices in he model. 8 See Diamond (98a, 98b) and Morensen and Pissarides (994). 9 In equilibrium, V. 6

8 law of moion of employmen, l (hose who are working a ime ), in a given firm is given by l l q. The represenaive inermediae firm produces oupu y Zlh, where Z is exogenous I aggregae produciviy. We normalize seady-sae produciviy, Z, o and assume ha Z follows an AR () process in logarihms, log Z Z log Z, where represens iid... draws from a normal disribuion wih zero mean and sandard deviaion. As in Arseneau and Chugh (8), firms face a quadraic cos of adjusing he hourly nominal wage rae, w. For each worker, he real cos of changing he nominal wage beween period and is, /, where is in unis of consumpion, and w / w is he ne w wage inflaion rae. If, here is no cos of wage adjusmen. w, Inermediae goods producers sell heir oupu o final producers a a real price in unis of consumpion. Inermediae producers choose he number of vacancies, employmen, l, o maximize he expeced presen discouned value of heir profi sream:, expressed, and uc, w E Zlh lh w, l, uc, P subjec o he dynamics of employmen, where u C, denoes he marginal uiliy of consumpion in period. Profi in any period consiss of oupu sales less labor coss inclusive of wage adjusmen coss plus vacancy coss. Fuure profis are discouned a he sochasic discoun facor of domesic households, who are assumed o own Home firms. Combining he firs-order condiions for vacancies and employmen yields he following job creaion equaion: w E, Z h h w,, q q P (3) where, uc, / uc, is he one-period-ahead sochasic discoun facor. The job creaion condiion saes ha, a he opimum, he vacancy creaion cos incurred by he firm per curren mach Noe ha he assumpion of a uni mass of inermediae producers ensures ha I y is also he oal oupu of he inermediae secor. 7

9 is equal o he expeced discouned value of he vacancy creaion cos per fuure mach, furher discouned by he probabiliy of curren mach survival, plus he profis from he ime- mach. Profis from he mach ake ino accoun he fuure marginal revenue produc from he mach and is wage cos, including fuure nominal wage adjusmen coss. Wage and Hours The nominal wage is he soluion o an individual Nash bargaining process, and he wage paymen divides he mach surplus beween workers and firms. Due o he presence of nominal rigidiies, we depar from he sandard Nash bargaining convenion by assuming ha bargaining occurs over he nominal wage paymen raher han he real wage paymen. Wih zero coss of nominal wage adjusmen ( ), he real wage ha emerges would be idenical o he one obained from bargaining direcly over he real wage. This is no longer he case in he presence of adjusmen coss. The deails of wage deerminaion are se ou in he Appendix. There we show ha he H J, where w, is he bargaining share equilibrium sharing rule can be wrien as,, w w of firms, H is worker surplus, and J is firm surplus (see he Appendix for he expressions). As in Gerler and Trigari (9), he equilibrium bargaining share is ime-varying due o he presence of wage adjusmen coss. Wihou hese coss, we would have a ime-invarian bargaining share w,, where is he weigh of firm surplus in he Nash bargaining problem. (The seady-sae value of w,, w, differs from if wages are sicky and here is non-zero seady-sae wage inflaion.) The bargained wage saisfies: vh w, E, J w, w,, w, w h w, b w, Zh w, P uc, (4) vh / u b is he worker s ouside opion (he uiliy value of leisure plus an where, C unemploymen benefi b), and is he probabiliy of becoming employed a ime, defined by M / U. Wih flexible wages, he hird erm on he righ-hand side of his equaion reduces o The same assumpion is made by Arseneau and Chugh (8), Gerler, Sala, and Trigari (8), and Thomas (8). 8

10 E, J, or, in equilibrium, / q. In his case, he real wage bill per worker is a linear combinaion deermined by he consan bargaining parameer of he worker s ouside opion and he marginal revenue produc generaed by he worker (ne of wage adjusmen coss) plus he expeced discouned coninuaion value of he mach o he firm (adjused for he probabiliy of worker s employmen). The sronger he bargaining power of firms (he higher ), he smaller he porion of he ne marginal revenue produc and coninuaion value o he firm appropriaed by workers as wage paymens, while he ouside opion becomes more relevan. When wages are sicky, bargaining shares are endogenous, and so is he disribuion of surplus beween workers and firms. Moreover, he curren wage bill reflecs also expeced changes in bargaining shares. As is common pracice in he lieraure, we assume ha hours per worker are deermined by firms and workers in a privaely efficien way, i.e., so as o maximize he join surplus of heir employmen relaion. The join surplus is he sum of he firm s surplus and he worker s surplus, i.e., J H, as defined in (A.) and (A.4). The maximizaion yields a sandard inraemporal opimaliy condiion for hours worked ha equaes he marginal revenue produc of hours per worker o he marginal rae of subsiuion beween consumpion and leisure: vh, / uc, Z, where v h, is he marginal disuiliy of effor... Final Goods Producion A conribuion of Cacciaore and Ghironi () is o show how price sickiness can be inroduced in a racable way in he Ghironi-Meliz (5) model of rade and macroeconomic dynamics, while preserving he aggregaion properies of Meliz s (3) heerogeneous firms model. This is done by inroducing price sickiness a he level of secoral produc bundles for domesic sale and expor ha aggregae individual produc varieies produced by plans wih heerogeneous produciviy. In his sub-secion we describe final goods creaion and producion, he expor decision, and price seing. In each consumpion secor, n, he represenaive, monopolisically compeiive firm n produces he secoral oupu bundle, Y n, sold o consumers in Home and Foreign. Producer n is a muli-produc firm ha produces a se of differeniaed produc varieies, indexed by and defined See, among ohers, Thomas (8) and Trigari (9). 9

11 over a coninuum Ω : (, ) Ω Yn y n d, (5) where is he symmeric elasiciy of subsiuion across produc varieies. 3 Each produc variey y(, n) is creaed and developed by he represenaive final producer n. Since consumpion-producing secors are symmeric in he economy, henceforh we omi he index n y o simplify noaion. The cos of he produc bundle Y, denoed by P, is: y where y ( ), y P p d (6) Ω p is he nominal marginal cos of producing variey. The number of producs creaed and commercialized by each final producer is endogenously deermined. A each poin in ime, only a subse of varieies Ω Ω is acually available o consumers. To creae a new produc, he final producer needs o underake a sunk invesmen, f e,, in unis of inermediae inpu. Produc creaion requires each final producer o creae a new plan ha will produce he new variey. 4 Plans employ differen echnologies indexed by relaive produciviy z. To save noaion, we idenify a variey wih he corresponding plan produciviy z, omiing. Upon produc creaion, he produciviy level of he new plan z is drawn from a common disribuion Gz defined over z. This relaive produciviy level remains fixed hereafer. Each plan min, uses inermediae inpu o produce is differeniaed produc variey, wih real marginal cos: z p (7) y z,. P z A ime, each final Home producer commercializes N d, varieies and creaes N e, new producs ha will be available for sale a ime. New and incumben plans can be hi by a deah shock wih probabiliy, a he end of each period. The law of moion for he sock of producing plans is: 3 Secors (and secor-represenaive firms) are of measure zero relaive o he aggregae size of he economy. Noice ha Y ( n ) can also be inerpreed as a bundle of produc feaures characerizing produc n. 4 Alernaively, we could model produc creaion by assuming ha monopolisically compeiive firms produce produc varieies (or feaures) ha are sold o final producers, in his case inerpreed as reailers. The wo models are equivalen. Deails are available upon reques.

12 N, d N d, Ne,. When serving he Foreign marke, each final producer faces per-uni iceberg rade coss,, and fixed expor coss, f x,. 5 Fixed expor coss are denominaed in unis of he inermediae inpu and are paid for each expored produc. Thus, he oal fixed cos is Fx, Nx, fx,, where N x, denoes he number of produc varieies (or feaures) expored o Foreign. Wihou fixed expor coss, each producer would find i opimal o sell all is produc varieies in Home and Foreign. Fixed expor coss imply ha only varieies produced by plans wih sufficienly high produciviy (above a cu-off level z x,, deermined below) are expored. 6 To proceed furher, we define wo special average produciviy levels (weighed by relaive oupu shares): (i) an average z d for all producing plans, and (ii) an average z x, for all plans ha expor: We assume ha d z z andz min x, G z z min x, zx, zd z dg z, z z dgz. G is a Pareo disribuion wih shape parameer, k. As a resul, x, zx,, where k / k p p p. Thus, he share of exporing plans is given by: k p k p z min Nx, Gzx, Nd, Nd,. z (8) x, follows: The oupu bundles for domesic and expor sale, and associaed uni coss, are defined as 5 Empirical micro-level sudies have documened he relevance of plan-level fixed expor coss see, for insance, Bernard and Jensen (4). Alhough a subsanial porion of fixed expor coss are probably sunk upon marke enry, we follow Ghironi and Meliz (5) and do no model he sunk naure of hese coss explicily. We conjecure ha inroducing hese coss would furher enhance he persisence properies of he model by inducing hyseresis in he exensive margin of expor decisions. Alessandria and Choi (7) develop a model wih heerogeneous firms, sunk expor coss, and Walrasian labor markes (bu no enry in domesic markes). Comin, Loayza, Pasha, and Serven (4) show ha a model wih sunk expor enry coss and endogenous echnology adopion and diffusion makes i possible o replicae lower-frequency flucuaions referred o as medium-erm business cycles (see also Comin and Gerler, 6). Their exended framework hus provides a microfoundaion for resuls in Aguiar and Gopinah (7). 6 Noice ha z x, is he lowes level of plan produciviy such ha he profi from exporing is posiive.

13 d, d, x, x, z min zx, Y y ( z) dg z, Y y ( z) dgz, (9) y y y y d, x, zmin zx, P p ( z) dg z, P p ( z) dgz. () Using equaions (7) and (), he real coss of producing he bundles Y d, and Y x, can hen be expressed as: P P,. P z P z y y d, x, Nd, Nx, d x, () The presen discouned cos facing he final producer in he deerminaion of produc creaion and he expor bundle is hus: E P Y P Y N N f N f y y ds, xs, s s, ds, s xs, s es, s xs, xs, s. s P s Ps The producer chooses N d, and he produciviy cuoff z x, o minimize his expression subjec o (8), (), and x, zx, z. 7 The firs-order condiion wih respec o z x, yields: y Px, p Y x, f x, N x, P k p k. () The above condiion saes ha, a he opimum, marginal revenue from adding a variey wih produciviy z x, o he expor bundle has o be equal o he fixed cos. Thus, varieies produced by plans wih produciviy below z x, are disribued only in he domesic marke. The composiion of he raded bundle is endogenous, and he se of expored producs flucuaes over ime wih changes in he profiabiliy of expor. 7 Equaion (8) implies ha, by choosing z x,, he producer also deermines N x,.

14 The firs-order condiion wih respec o N d, deermines produc creaion: y y Nx, Pd, Yd, Px, Yx, Nx, f, ( e ) E, fe, fx, Nd, P Nd, P N x, N. d, In equilibrium, he cos of producing an addiional variey, fe,, mus equal is expeced benefi (expeced savings on fuure sunk invesmen coss augmened by he marginal revenue from commercializing he variey, ne of fixed expor coss, if i is expored). We are now lef wih he deerminaion of domesic and expor prices. We denoe by P d, he price (in Home currency) of he produc bundle Y d, and le P x, be he price (in Foreign currency) of he expored bundle Y x,. Each final producer faces he following domesic and foreign demand for is produc bundles: Pd, C Px, C Yd, Y, Yx, Y, P P where C Y and C Y are aggregae demands of he consumpion baske in Home and Foreign. Aggregae demand in each counry includes sources oher han household consumpion, bu i akes he same form as he consumpion baske, wih he same elasiciy of subsiuion across secoral bundles. This ensures ha he consumpion price index for he consumpion aggregaor is also he price index for he aggregae demand of he baske. Prices in he final secor are sicky. We follow Roemberg (98) and assume ha final producers mus pay quadraic price adjusmen coss when changing domesic and expor bundle prices, which we assume are se in accordance wih producer currency pricing (PCP): Each final d producer ses P d, and he domesic currency price of he expor bundle, P x,, leing he price in he d foreign marke be Px, Px, / S, where S is he nominal exchange rae (unis of Home currency per uni of Foreign). The nominal coss of adjusing domesic and expor price are, respecively, d d d P Y, and P Y, where deermines he size of he adjusmen d, d, d, d, / x, x, x, x, / coss (domesic and expor prices are flexible if d d d x, Px, / Px,. 3 ), d, Pd, Pd, / and In he absence of fixed expor coss, he producer would se a single price P d, and he law of

15 one price (adjused for he presence of rade coss) would deermine he expor price as P P P S. Wih fixed expor coss, however, he composiion of domesic and expor x, x, d, / bundles is differen, and he marginal coss of producing hese bundles are no equal. Therefore, final producers choose wo differen prices for he Home and Foreign markes even under PCP. We relegae he deails of opimal price seing o he Appendix. We show here ha he (real) price of Home oupu for domesic sales is given by: P P, (3) P y d, d, P d, where: d, Y d, d, d, d, d, E, d,, C Yd, (4) and P P C, /. As expeced, price sickiness inroduces endogenous markup variaions: The cos of adjusing prices gives firms an incenive o change heir markups over ime in order o smooh price changes across periods. When prices are flexible ( ), he markup is consan and equal o /. The (real) price of Home oupu for expor sales is equal o: P P P, (5) y x, x, d x, QP where Q S P P is he consumpion-based real exchange rae (unis of Home consumpion per / unis of Foreign), and: d d d d d d x, Y x, x, x, x, x, E, x,. C Yx, (6) d Absen fixed expor coss zx, zmin and x, d,. Plan heerogeneiy and fixed expor coss, insead, imply ha he law of one price does no hold for he expored bundles. For fuure purposes, define he average real price of a domesic variey, d, Nd, Pd, / P and he average real price of an expored variey, x, Nx, Px, / P 4. Combining equaions (),

16 (3), and (5), we have:,, d, d, x, x, zd Q zx, where d, / d, and / x (7) d, x,. Finally, leing y d, and y x, denoe he average oupu of, respecively, a domesic and expored variey, we have C y N Y, y N Y. (8) C d, d, d, x, x, x,.3 Household Budge Consrain and Ineremporal Decisions In our benchmark scenario, we assume ha inernaional asses markes are incomplee, as he represenaive household can inves only in nominal riskless bonds denominaed in Home and Foreign currency. Home-currency bonds are raded only domesically. Le A and A, denoe, respecively, nominal holdings of Home and Foreign bonds a Home. 8 To ensure a deerminae seady-sae equilibrium and saionary responses o emporary shocks in he model, we follow Turnovsky (985), and, more recenly, Benigno (9), and assume a quadraic cos of adjusing Foreign bond holding, A / /, P. 9 These coss are paid o financial inermediaries whose only funcion is o collec hese ransacion fees and o rebae he revenue o households in lump-sum fashion in equilibrium. The Home household s period budge consrain is: A,, P A S A S P PC i A i A S wl Pb l T T T T, G A I F, where i and i are, respecively, he nominal ineres raes on Home and Foreign bond holdings beween and, known wih cerainy as of. Moreover, G T is a lump-sum ransfer (or 8 Foreign nominal holdings of Foreign bonds are denoed by A,. 9 Given ha idiosyncraic risk is pooled among domesic households, and foreign households only rade foreign currencydenominaed bonds, domesic-currency-denominaed bonds are in zero ne supply. Tha is, in realiy only foreign-currencydenominaed bonds are raded in equilibrium. As a resul, defining he inermediaion coss over he foreign currency bond only is sufficien o pin down he overall seady-sae ne foreign asse posiion. 5

17 ax) from he governmen, inermediaries o which i is paid, and inermediae and final goods producers. le A T is a lump-sum rebae of he cos of adjusing bond holdings from he I T and F T are lump-sum rebaes of profis from Le a / A P denoe real holdings of Home bonds (in unis of Home consumpion) and /,, a A P denoe real holdings of Foreign bonds (in unis of Foreign consumpion). The Euler equaions for bond holdings are:, a i E, C, Q a i E,, Q C, (9) () where C, P P /. We presen below he law of moion for ne foreign asses ha follows from imposing equilibrium condiions in he household s budge consrain. Oher deails on he equilibrium can be found in he Appendix..4 Ne Foreign Asses and he Trade Balance Bonds are in zero ne supply, which implies ha he equilibrium for he domesic bonds, being nonraded, is a in all periods. Home ne foreign asses are deermined by: i Qa Q a QN y N y. (),, x, x, x, x, x, x, C,, he change in ne foreign asses beween and is Defining r i / C, deermined by he curren accoun: In equilibrium, G T Pb( l ) Q a a CA Qr a TB,,,, A, T S P A P I,, 6 T P Z l w P l l V, w, T N y Q N y N f N f x,,,,,,,,,,,, F d,, d d d d x x x x x x e e d, x,.

18 where TB is he rade balance: TB Q N y N y. x, x, x, x, x, x,.5 Moneary Policy and Daa-Consisen Variables Before describing moneary policy in he small open economy, we mus address an issue ha concerns he daa ha are acually available o he cenral bank, i.e. we need o deermine he empiricallyrelevan variables ha should ener he heoreical represenaion of moneary policy (as well as be used for comparison of model properies o he daa in our exercises below). As poined ou by Ghironi and Meliz (5), in he presence of endogenous produc creaion and love for variey in he producion of final consumpion-varieies, variables measured in unis of consumpion do no have a direc counerpar in he daa, i.e., hey are no daa-consisen. As he economy experiences enry of Home and Foreign firms, he welfare-consisen aggregae price index P can flucuae even if produc prices remain consan. In he daa, however, aggregae price indexes do no ake hese variey effecs ino accoun. To resolve his issue, we follow Ghironi and Meliz / P N N P. The average price index (5), and we consruc an average price index d, x, P is closer o he acual CPI daa consruced by saisical agencies han is he welfare-based index P, and, herefore, i is he daa-consisen CPI implied by he model. In urn, given any variable X in unis of consumpion, is daa-consisen counerpar is X, X P / P.. The daa-consisen CPI /. C inflaion rae is P P R We now specify moneary policy for he small open economy. As shown by Kim (3), a sandard ineres rae rule in he spiri of Taylor (993) describes Korean moneary policy quie well. We begin by assuming ha he cenral bank of he small open economy ses he conemporaneous policy ineres rae according o a rule ha can be aken as an operaional descripion of a flexible inflaion argeing policy: There is much empirical evidence ha gains from variey are mosly unmeasured in CPI daa, as documened mos recenly by Broda and Weinsein (). 7

19 i ˆ g R i g i Y i i C, YR,, () ˆ g where Y,, / flex R YR YR, denoes he oupu gap deviaions of real oupu, Y R,, from real oupu flex under flexible prices and wages, Y R, and C, denoes daa-consisen CPI inflaion. R Table summarizes he key equilibrium condiions of he model-small open economy. The able conains 3 equaions ha deermine 3 endogenous variables of ineres: C, d,, l, h, V, Nd,, w / P, z x,, w,, C,, i, a, and, Q. (Oher variables ha appear in he able are deermined as described above.).6 Foreign Aggregaes As summarized in Table, six Foreign variables direcly affec he macroeconomic dynamics in he C small open economy: Y, i,, N, y, x,. Aggregae demand,, he nominal ineres rae, i, and inflaion, C, C, x, x,, are deermined by reaing he res of he world (Foreign) as a closed economy ha feaures he same producion srucure, echnology and fricions ha characerize he small open economy. 3 Here we focus on he deerminaion of he number of Foreign exporers, N x,, he average oupu of Foreign expored varieies, y x, C Y, and heir average relaive price, x,. Since he small open economy is infiniesimally small relaive o he res of he world, hese variables affec C macroeconomic dynamics in he small open economy wihou having any effec on Y, i, and. C, We assume ha Foreign producers solve a profi maximizaion problem ha is equivalen o ha faced by Home producers, including he assumpion ha expor prices are denominaed in producer currency. The number of Foreign exporers is a ime-varying fracion of he number of Foreign producers ha serve heir domesic marke: The sensiiviy analysis of Secion 7 summarizes he implicaions of replacing he policy rule () our represenaion of inflaion argeing under a flexible exchange rae wih a rule ha allows for a response of he policy rae o exchange rae dynamics, capuring feaures of a managed exchange rae regime ha arguably describes imporan feaures of Bank of Korea policymaking since he Asian crisis of We do no repor he deails of he foreign economy. They are discussed in deph by Cacciaore and Ghironi (). 8

20 where x, k p k p z min x, x, d, d, z x, N G z N N, z is deermined by imposing a zero expor-profi condiion ha is he Foreign counerpar o equaion (): In he above expression, and C x, x, Nx, Y f x, k p x, z. f denoe, respecively, iceberg rade coss and fixed expor coss for Foreign firms (boh coss are exogenous). The average oupu of a variey expored by Foreign o Home is: C x, x, x,, y N Y where he average relaive price is given by: x, x, Qx, z x,. In he above expression, denoes he marginal coss of producion of an individual variey in he res of he world; he erm denoes he expor markup: where: x, x, x, x,, x, d x, d d d d d d Y x, x, x, x, x, E, x, x,, Yx, d, x, C, Q Qx, x, Y N y denoes Foreign expor price inflaion, and is a Foreign expor markup shock ha we will use below o inroduce shocks o he erms of rade. 9

21 3. Calibraion and Model Properies 3. Calibraion We inerpre periods as quarers and calibrae he res-of-he-world parameers o mach sandard poswar U.S. macroeconomic daa. We choose he Unied Saes o represen he res-of-he-world economy for our model Korea because Korea sabilized he exchange rae of he won agains he U.S. dollar for much of he pos-breon Woods period, and we will discuss he consequences of differen exchange rae policies vis-à-vis he dollar below. Wih he excepion of he moneary policy coefficiens in he ineres rae rule, and he process of exogenous shocks, we assume ha he parameers ha characerize he small open economy are symmeric o he res of he world. Given ha Korea is an advanced economy, we view his as a plausible assumpion. 4 Table summarizes he calibraion. (In he able and below, variables wihou ime indexes denoe seady-sae levels; parameers denoed wih a sar are specific o he res of he world, i.e., he calibraion of hose parameers is no symmeric across counries.) 3.. Res of he World We se he discoun facor o.99, implying an annual real ineres rae of 4 percen. The period C h uiliy funcion is given by u C / C lh / h. The risk aversion coefficien C is equal o, while he Frisch elasiciy of labor supply / h is se o.4, a mid-poin beween empirical micro and macro esimaes. 5 The elasiciy of subsiuion across produc varieies, is se o 3.8 following Bernard, Eaon, Jensen, and Korum (3), who find ha his value fis U.S. 4 Concerning marke regulaion parameers, OECD indexes for barriers o producer enry and employmen proecion legislaion for Korea and he U.S. are very similar (Online OECD Employmen Daabase, OECD). The same is rue for he unemploymen benefi replacemen rae (Benefis and Wages: Saisics, OECD). 5 The value of his elasiciy has been a source of conroversy in he lieraure. Sudens of he business cycle end o work wih elasiciies ha are higher han microeconomic esimaes, ypically uniy and above. Mos microeconomic sudies, however, esimae his elasiciy o be much smaller, beween. and.6. For a survey of he lieraure, see Card (994). Keane and Rogerson () offer a reconciliaion ha credibly suppors he range of esimaes ypically adoped in macroeconomic simulaions. Our resuls are no affeced significanly if we hold hours consan a he opimally deermined seady-sae level.

22 plan and macro rade daa. Following Ghironi and Meliz (5), we se he elasiciy of subsiuion across Home and Foreign goods,, equal o. Also as in Ghironi and Meliz (5), we se 3.4, and normalize z min o. To ensure seady-sae deerminacy and saionariy of ne foreign asses, we se he bond adjusmen cos parameer o.5 as in Ghironi and Meliz (5). The scale parameer for he cos of adjusing prices,, is equal o 8, as in Bilbiie, Ghironi, and Meliz (8). We choose, he scale parameer of nominal wage adjusmen coss, so ha he model reproduces he volailiy of unemploymen relaive o GDP observed in he daa. This implies ϑ = 6. To calibrae he enry coss, we follow Ebell and Haefke (9) and se f e so ha regulaion coss imply a loss of 5. monhs of per capia oupu. Unemploymen benefis, b, are equal o 54 percen of he seady-sae wage, he average value for he U.S. repored by OECD (4). The seady-sae, flexible-wage bargaining share of workers,, is equal o.4, as esimaed by Flinn (6) for he U.S. The unemploymen elasiciy of he maching funcion,, is also equal o.4, wihin he range of esimaes repored by Perongolo and Pissarides (6) and such ha he sandard Hosios condiion for efficiency in he absence of oher disorions holds in seady sae. The exogenous separaion rae beween firms and workers,, is percen, as repored by Shimer (5). To pin down exogenous plan exi,, we arge he porion of worker separaion due o plan exi equal o 4 percen repored by Haliwanger, Scarpea, and Schweiger (8). Two labor marke parameers are lef for calibraion: he scale parameer for he cos of vacancy posing,, and he maching efficiency parameer,. We se hese parameers o mach he seadysae probabiliy of finding a job and he probabiliy of filling a vacancy. The former is 6 percen, while he laer is 7 percen, in line wih Shimer (5). For he produciviy process, we follow King and Rebelo (999) and se persisence equal o.979 and sandard deviaion of innovaions o.7. In our benchmark scenario, we assume ha here are no shocks o he Foreign expor markup, i.e., we se in all periods. Finally, he parameer values in he policy rule for he Federal Reserve s ineres rae seing are hose esimaed by Clarida, Galí, and Gerler (). The inflaion and GDP gap weighs are.65 and.34, k p

23 respecively, while he smoohing parameer is.7. These are commonly used values for parameers characerizing Federal Reserve ineres rae seing under normal economic condiions since he early 98s. 3.. Small Open Economy As discussed above, parameers are assumed o be symmeric across counries, wih he excepion of he coefficiens appearing in he ineres rae rule (), he persisence of produciviy shocks, and he sandard deviaion of produciviy innovaions. Moreover, hree exogenous variables are specific o he small open economy: he fixed expor cos f x, ; iceberg rade coss relaed o impors, ; and iceberg rade coss relaed o expors,. We assume ha hese coss are consan. Thus, we drop he ime index for simpliciy. Moreover, we assume ha iceberg rade coss relaed o impors T T NT NT (expors) are he sum of ariffs, ( ), and non-ariff barriers, ( ), i.e., T NT ( T NT ). Finally, we le boh ariff and non-ariff componens of rade coss be equal across coss relaed o expors and impors, so ha and rade coss associaed wih expors and impors are fully symmeric. We calibrae rade coss and parameers specific o he small open economy o mach feaures of Korean macroeconomic daa for he period 998-7, which corresponds o he financial inegraion era, prior o he global crisis of 8-9. T NT We se =.7 and calibrae non-ariff barriers so ha oal rade (impors plus expors) over GDP is equal o 66 percen, he average value for Korea over he calibraion period. We choose f x so ha he share of exporing plans is equal o percen, consisen wih he evidence in Aw, Chung, and Robers (). Finally, we se he parameer values in he hisorical rule for Korean ineres rae seing consisen wih he esimaes in Kim (3). The inflaion and GDP gap weighs are.87 and.9, respecively, while he smoohing parameer is.84. We se he persisence of g Y produciviy Z o.999, consisen wih he evidence of a uni roo in Korean produciviy (Kim, Lim, and Park, 9). We calibrae he volailiy of produciviy innovaions o mach he volailiy of Korea s GDP. This requires seing.. i

24 3. Model Properies We now discuss he propagaion of aggregae shocks in he model, and compare business cycle dynamics o he daa. We focus on he scenario of inernaionally incomplee asse markes and endogenous ineres rae seing under flexible exchange raes described in Secion as he benchmark o explain he model s properies and compare hem o daa. Figure shows he impulse responses of our model-korea o a one-percen innovaion in domesic produciviy. Unemploymen ( U ) declines in he periods immediaely following he shock. On impac, he higher expeced reurn of a mach induces domesic inermediae inpu producers o pos more vacancies, which resuls in higher employmen he following period. Firms and workers renegoiae nominal wages because of he higher surpluses generaed by exising maches, and wage inflaion. ( w, ) increases. Wage adjusmen coss make he effecive firm s bargaining power procyclical, i.e., w, rises. 6 Oher hings equal, he increase in w, dampens he response of he renegoiaed equilibrium wage, amplifying he response of job creaion o he shock. Higher produciviy causes enry of domesic produc lines o increase and he expor cuoff, z x, o fall. Accordingly, a larger number of Korean goods are available o domesic and foreign consumers. Korea runs a curren accoun defici in response o he produciviy increase ( CA falls), as i is opimal o borrow from abroad o finance increasing producer enry. Korea s erms of rade (defined as TOT Q ) depreciae he relaive price of Korean expors in erms of Korean x, x, impors falls so ha Korean goods become relaively cheaper. However, he erms of rade depreciaion is mild compared o sandard inernaional business cycle models. Producer enry and he counercyclical response of z x, counerac he effecs of higher produciviy on marginal coss, and domesic expor prices fall by less han in a model ha absracs from enry and heerogeneiy. Table 3 compares he second momens for key macroeconomic aggregaes o hose implied by Korean daa for he period 998:Q-7:Q4. Korea was hi by a combined currency-and-banking crisis in lae 997. Therefore, we ake 998 as he beginning of a floaing exchange rae regime for he won. Excep for he rade balance, all he daa used o compue he momens in Table 3 were 6 Inuiively, increases o ensure opimal sharing of he cos of adjusing wages beween firms and workers. w, 3

25 ransformed by aking logarihms; all series were hen HP-filered wih smoohing parameer,6. In he able, Model I refers o he benchmark case in which he only sochasic shocks are produciviy shocks in he inermediae goods secor. In ha case, he model maches he momens for real GDP, consumpion, and unemploymen fairly well, bu i oversaes he volailiy of invesmen (which, in our model, is given by invesmen in new produc creaion: I fe, Ne,, and IR, PI P.). The model correcly capures he fac ha impors are more volaile han expors, alhough heir absolue volailiy is smaller han in he daa. The model also performs reasonably well a reproducing he observed correlaions of macroeconomic variables wih GDP and successfully generaes cross-counry GDP correlaion ha is higher han consumpion correlaion. This resul is a challenge for sandard inernaional business cycle models. Cosly producer enry and labor marke fricions dampen he resource shifing effec ha ofen leads o counerfacually low GDP correlaion across modelcounries. 7 Model II augmens he produciviy shocks wih an exogenous sochasic componen in erms of rade dynamics (discussed in more deail in Secion 7 below), a resul of which is ha he model maches rade-relaed volailiies more closely, even if i fails o replicae he counercyclicaliy of he rade balance observed in he daa The Macroeconomic Consequences of Financial Inegraion 4. Business Cycles Dynamics We now sudy how inernaional financial marke inegraion affecs business cycle flucuaions and he dynamics of producer enry and employmen in our model-korea. To address his issue, we compare he properies of he model under inernaionally incomplee markes o wo exreme cases: financial auarky or complee markes. Taking he scenario of inegraion in incomplee markes as he mos empirically plausible represenaion of Korea s siuaion in 998-7, he quesion we 7 This resul does no depend on he fac ha he Home economy is of negligible size relaive o Foreign in oher words, on he fac ha Foreign oupu is no affeced by resource shifing oward Home. The resul arises also in Cacciaore and Ghironi s () original version of he model wih symmeric counry size. 8 In conras o Cacciaore and Ghironi s () model wih symmeric counry size, invesmen in produc creaion is no sufficien o generae a counercyclical rade balance for our calibraion of he small open economy framework. The mos sraighforward way o mach his feaure of he daa would be o include physical capial in he producion funcion, and invesmen in capial accumulaion. We refrain from doing ha o resric he deparures from he benchmark New Keynesian small open economy model of Galí and Monacelli (5) o labor marke fricions and he dynamics of marke enry he model feaures we wan o focus on. (By assuming ha oupu is only a funcion of labor and exogenous produciviy, he sandard New Keynesian model canno generae counercyclical rade balances.) 4

26 address in his secion is how would Korea s dynamics in he same period have differed if Korea had operaed is exchange rae floa under financial auarky or in an environmen of full inernaional insurance. Under financial auarky, equaion () is replaced by he condiion ha rade is balanced in every period, TB, ogeher wih he requiremen ha holdings of foreign bonds are zero in every period, a, for any. By conras, under complee markes, equaion () is replaced by he condiion u u, Q,, which represens he opimal risk sharing agreemen beween agens in he / C, C small economy and agens in he res of he world. Risk sharing requires ha he marginal benefi of an exra uni of domesic consumpion obained from Foreign via insurance be equal o he marginal compensaion ha Foreign agens receive for his insurance provision, given by he marginal uiliy of foreign consumpion muliplied by he relaive price of 5 C in erms of C (he inverse of he domesic real exchange rae, Q). When a complee se of sae-coningen securiies is available, he risk-sharing condiion holds in a decenralized equilibrium independenly of oher nominal and real fricions. Figure compares he impulse responses o a one-percen innovaion in Home produciviy under financial auarky (dashed lines) and complee markes (dash-doed lines), as well as he inermediae case of incomplee markes (solid lines). Under financial auarky, households in he small open economy canno borrow from abroad o finance increased producer enry. As a resul, a smaller number of producers ener he domesic marke relaive o he benchmark scenario wih incomplee markes. Meanwhile, consumpion responds by more, as lack of inernaional borrowing reduces he abiliy o smooh is dynamics. Noice ha he persisence of he produciviy process plays an imporan role here: he more persisen he produciviy shock, he sronger he response of consumpion. The reason is ha highly persisen shocks have a larger impac on he household s permanen income. The dynamics of unemploymen reflec wo opposing forces. While smaller producer enry reduces he demand for inermediae inpus (and hus vacancy posing and employmen), higher consumpion of final goods has an opposie effec. As a resul, unemploymen dynamics are no significanly affeced. Since a smaller number of producers eners he expor marke, he marginal y cos of producion for he expor bundle, P / P N / z, falls by less under financial auarky. This leads o erms of rade depreciaing by less. x, x, x, By conras, complee markes resul in srong producion shifing oward he relaively more