Reserve Requirements and Economic Growth : the Case of South Korea

Transcription

1 Reserve Requiremens and Economic Growh : he Case of Souh Korea EUN YOUNG OH 1 DURHAM UNIVERISTY Absrac This paper examines he relaionship beween reserve requiremens, and long-erm economic growh. I presen a model which shows ha he governmen migh repress he banking secor as his is he easy way of channelling resources o producive secors. In he overlapping generaions model, I employ he governmen s capial inpu in he firm producion funcion. The implicaions of he model are confirmed in ha, an increase in reserve requiremens has wo differen reverse effecs on growh; he negaive effec on he financial secor and a growh enhancing effec from effecive public spending on he real secors. More precisely, his paper uses coinegraion echniques and Vecor Auoregressive (VAR) analysis in order o derive he relaionship beween reserve requiremens and economic growh. Key Words: Reserve Raio, Governmen Spending, Growh JEL Classificaion: E40, E51, O42, 1 PhD Candidae & Teaching Fellow a Deparmen of Economics, e.y.oh@durham.ac.uk

2 Reserve Requiremens and Economic Growh 1. Inroducion Since a high reserve requiremen in a moneary economic model was severely criicized by Chari e al. (1995) and Roubini and Sala-i-Marin (1995), a reserve requiremen being a highly conroversial issue. They demonsraed how he reserve raio leads o an increase in he cos of financial inermediaion, hence deerioraing he financial sysem's funnelling channels o he efficien producive secors and households. On he oher hand, Basu (2001) argues ha heoreically, reserve requiremens can help o accelerae economic growh hrough public spending under he assumpion ha seigniorage revenue by imposing reserve requiremens is spen producively. Many developing counries wih a lower ax base use seigniorage as an easy source of revenue. In addiion, a number of papers demonsrae how his influences he growh induced by governmen spending. 2 The case of Souh Korea is of a paricular ineres, since i has shown asonishing GDP growh as well as exhibiing he differen policies of is financial marke by allowing is governmen o inerfere significanly wih he conrol of he bank reserve raio, and he credi allocaion. For insance, he Bank of Korea se he reserve raio in he banking secor as high as 11.5 percen a he end of 1980s. I can be seen as a case ha imposing a higher reserve requiremen and seigniorage may be an imporan channel of financing public spending which could impac growh in developing counries. In addiion he public invesmen raio in Souh Korea was relaively higher han ha of oher developed counries. 3 Wihin hese conexs, his paper addresses he following quesion: In he process of he upurn and downurn of he Korean economy during he las 40 years, have he sauory reserve requiremens served imporan roles? How do reserve requiremens, a form of financial repression policies, impac he economic growh? Wha are he channels hrough which he reserve requiremens effec or hinder oupu growh? I presen a heoreical model which shows ha he governmen uses he financial secor as way of financing public spending which could posiively impac growh. In his paper he overlapping generaions model is employed by considering an economy where he governmen revenue is colleced by imposing he ax such as reserve requiremens on he banking secor. This ax is hen spen by he governmen on economic infrasrucures which migh enhance he produciviy of he indusrial secor. The implicaions of he model indicae ha an increase in reserve requiremens has wo differen reverse effecs on growh. One is he negaive effec on he banking secor and he oher is a growh enhancing effec from effecive public invesmen. Wih his background in mind, his paper esimaes a heoreical model for Souh Korean economy by applying a coinegraed Vecor Auoregressive (VAR) model which helps o es he effecs ha reserve requiremens may have on economic growh in boh he shor and long run. 2 Aghevili(1977) 3 Public invesmen o oal governmen expendiure in 1990h (Naional Bureau of Saisic,1997): Souh Korea(42.1%), Japan (9.3%), Germany (15.2%), Unied Saes (8.0%), France (11.75), Thailand (47.3%) 1

3 Reserve Requiremens and Economic Growh The res of his paper is organized as follows. Secion 2 describes he heoreical model. Descripion of daa and uni roo ess are in secion 3. Shor-and Long-run effecs based on a VAR model are invesigaed in secion 3. Finally Secion 5 offers he concluding remark. 2. Theoreical Framework I employ he overlapping generaions model which sae he economy is characerised by an infinie sequence of wo-period-lived-overlapping-generaions. Time is discree and is indexed by 1,2,3,.... A each dae, here are wo coexising generaions, young and old wih hem being in he firs and second period of heir lives respecively. N people are born a each ime poin 1. A 1, here exis N people in he economy, called he iniial old, who live for only one period. Hereafer N is normalised o 1. Each agen is endowed wih one uni of working ime when young and is reired when old. ln c y ln c, (0,1) (1) o s.. cy d w, d c 1 i d, o ( c, c ) 0, y o Given ( w ) and ( i d ), he allocaion cy, c o solves he represenaive household s problem. d w (1 ) (2) There are compeiive banks in his model which ake deposis from he households and make loans o he firms. This inermediaion service is assumed o be cosless. A governmen ses he reserve requiremen which he bank should keep a a fracion of nominal deposi Aside from his sauory requiremen; here is no oher reason o hold fia money in his model. I follow Diaz e al.(1992) s argumen ha banks solve he following saic problem. l d max (1 i ) L M (1 i ) D (3) LD, s.. D L D M D (4) 2

4 Reserve Requiremens and Economic Growh Where M is he cash reserves held by he bank a ; L is he nominal loan; i l is he ineres rae on loans a, and is he sauory reserve raio. The banks hold he reserve as he form of fia money and in his paper. I focus on he siuaion when he consrain (4) is binding. I is assumed ha financial inermediaries behave compeiively and free enry drives profis o zero. Here a zero profi condiion holds when banks are compeiive. i l d i (5) 1 Reserve requiremens hus end o induce a wedge beween he ineres rae on savings and lending raes for he financial inermediaries. Haslag (1998a) demonsraes he Cenral Bank in his economy is recommied o a sequence of money supplies, M and reserve requiremen, aking he sequence of household s deposis D, ineres raes, i l. i d as given. Formally, his means ha he Cenral Bank ses he ime pah for he money supply in such a way ha a each dae he money supply equals money demand. M d P (6) The equilibrium condiion for he governmen budge consrain is: G M P (7) All loans are convered o capial wihou any cos. G k 1 (8) The model forms a producion funcion, in which he marginal produc of capial is consan. Final oupu is produced by wo inpus: Capial ( k ), and Governmen inpu ( g ). I assume a Cobb-Douglas aggregae producion funcion, which is wrien as follows: y Ak n g a 1 a ( ) (9) Where 0a 1, A is he a fixed echnology parameer. k, n and g respecively denoing capial, labour, and governmen expendiure inpus a ime. I follow Barro (1990) in assuming ha g is a non-rival and non-excludable inpu in he producion process. 3

5 Reserve Requiremens and Economic Growh The represenaive firm a any poin of ime maximises he discouned sream of profi flows in erms of he capial accumulaions and loan consrains. Formally, he problem of he firm can be oulined as follows. i l Max p y (1 i ) pl pw n (10) k1, n i0 k 1 (1 k ) k i (11) k p i k L (12) Where is he discoun facor, and k 1is he rae of capial depreciaion for simpliciy. The firm solves he above problem o idenify he demand for labour and invesmen in period, or he capial accumulaion o be carried over o period 1. The firm s problem using recursive formulaion can be wrien as follows: V ( k ) max p y (1 i ) p k (1 k p w n V ( k ) (13) l nk, 1 1 The firs condiion for he above dynamic programming problem is V ( k ) / ( k ) (1 i ) p V ( k ) 0 l 1 1 (1 i ) p V ( k ) l 1 1 g V ( k ) / ( n ) A(1 ) k w k (14) 1 g V ( k) p A k Opimisaion for he firm s problem reduces o 4

6 Reserve Requiremens and Economic Growh 1 i l 1 1 g 1 A k 1 1 (15) This equaion provides he condiion which leads o he opimal invesmen decisions of he firm. w Indicaes ha he firm hires labour up o he poin where he marginal produc of labour equaes o he real wage. The firm esimaes he cos of increasing invesmen in wih he fuure sream of profis creaed from he exra capial invesed in. Equilibrium Given he iniial k 0 and g 0, an equilibrium for our economy is a sequence of; prices,, d l P w i, i 0 variables,, g 0, allocaions c, i k, n, socks of financial asse 0 d 0 such ha, and policy - The sock of financial asses solve he bank s profi maximisaion problem (3), given prices and policy variables - The real allocaions solve he firm s profi maximisaion problem (10) - The loanable funds marke equilibrium condiion: pik L is saisfied for all 0 - n 1for all 0 - The governmen budge is balanced, meaning (6) holds. 1 i, 1 i, d and P mus be posiive a all daes. d l - Thus, in Equilibrium, 1a 1 (1 )(1 ) 1 1 K A a k (16) I focus on he balanced growh pah. This is o observe ha consumpion, privae capial, oupu, and public capial will all grow a he same rae. The growh rae of privae capial is reduced o ln k ln k (17) 1 5

7 Reserve Requiremens and Economic Growh ln ln A ln(1 a) (1 a)ln( ) aln(1 ) (1 a)ln (18) 1 The growh equaion (18) demonsraes how reserve augmened seigniorage impacs growh. There are wo aspecs of (18) ha need o be highlighed. Firs, he opimal growh rae is relaed o he money growh rae, Reserve raio, he differences, and he subjecive discoun facor ( ). As long as hese hree are consan, privae capial will always grow a he same consan rae in ha period. Second, he growh rae is a Laffer-Curve ype in. in oher words, he effecs in reserve raio are non-linear because hey have wo differen effecs-i.e., promoing he growh as well as disoring i. 3 Descripion of Daa and Uni Roo Tess 3.1 Descripion of Daa This secion examines he reserve requiremens and oupus for he Souh Korean economy over he period Annual daa for hese 40 years have been used o esimae he heoreical model in he previous secion. These are gahered from various sources including he Bank of Korea and Inernaional Daabase. Based on he model described in he previous chaper, he following ime series daa has been uilised in he es as summarised in Table 1 below. (Table 1. Descripion of Variables and Daa Sources) Variable Daa Daa source Symbol Oupu Reserve Requiremen Raio Producive Governmen Inpu Seigniorage Domesic Credi by Bank Real Per Capia GDP Reserve Requiremens o Bank Deposis Public Fixed Capial Formaion on Economic Aciviies Reserve Money Loans by Banks IMF Inernaional Financial Saisics The Bank of Korea Economic Saisics Year Book : Various Issues The Bank of Korea : Economics Saisics Sysem World Bank World Developmen Indicaors The Bank of Korea : Economics Saisics Sysem LGDP DRR LGSE LSEG LDBL 6

8 Reserve Requiremens and Economic Growh Mos of daa se excep he reserve requiremen raio is provided in nominal erms. This curren US dollar daa se has been convered o real value by he GDP deflaor for All dollar figures are convered by using he 2005 official exchange raes. LDBL is defined as domesic credi provided by he banking secor. I represens he availabiliy of bank loans o households and firms. In addiion, hey are measured in naural logarihms so ha hey can be inerpreed as growh erms when aking he firs differences. We se DRR as a raio beween sauory reserve requiremens and bank asses. 5 Wih regard o he governmen expendiures, I employed he series on gross fixed capial formaion by governmens o measure public invesmen. Specifically governmen capial expendiures on economic aciviies are employed. 6 I use he gross fixed capial formaion insead of capial sock for economics aciviies using he sandard perpeual invenory mehod. Alhough i is widely applied in he economic lieraure, here are many drawbacks Uni Roo Tess As Nelson and Plosser (1982) saed, mos macroeconomic daa ofen show sochasic rends, which can be lead o a spurious regression. Hence before running any regression, he saionariy es is necessary o disinguish beween saionary series and non-saionary series. Since all variables conain a srong auocorrelaion, as well as ime rend, we have chosen o adop he Augmened Dickey Fuller Tes (ADF). The Table 2 provides srong evidence ha all of daa series are non-saionary. The null hypohesis for esing a uni roo es is H : (1) 0 y I, and he alernaive hypoheses under each esing approach is H : (0) 1 y I. I (1) Indicaes ha y is inegraed of order one which has a uni roo and I(0) shows ha y is saionary. Using es saisics and a p-value, he exisence of uni roo may be examined; for he variables a a level less negaive han 5% criical value, bu for he firs differenced daa more negaive han he 5% criical value. Hence he level es confirms ha he exisence of a uni roo in LGDP, DRR, DGSE, LSEG, and LDLB and firs difference es resuls suppor ha here is no uni roo in LGDP, DRR, LGSE, LSEG, and LDLB. In oher words, he repored saics sugges ha hey are inegraed of order 1, I (1). 4 Nominal per capia GDP a GD GD Commercial banks and specialised banks are considered. 6 Agriculure, Foresry, Fishing and huning, Mining, manufacuring and Consrucion, Elecriciy, Gas, Seam and Waer, Roads, Inland and cosal waerways, Oher ransporaion and communicaion and oher economic services 7 Fein-sein (1968) and Surm and De Haan (1995) 7

9 Reserve Requiremens and Economic Growh (Table 2. Uni roo es: Augmened Dickey Fuller Tes) Level Firs Difference Variables Lag -saisics Prob Lag -saisics Prob LGDP * DRR * LGSE * LSEG * LDLB * Noe: * indicaes significan a he 5% level. Lag Lengh is based on SIC, MAXLAG=9. 4 Empirical Resuls There are hree differen VAR specificaions followed by a-priori heoreical model. In order o esimae he effec of he sauory reserve requiremens on economic growh, we observe wo secors: he governmen secor and he bank secor. The Firs VAR (VAR1): Reserve Requiremen Raio, Seigniorage, and Governmen Expendiures on Economic Aciviies (DRR, LSEG, LGSE) The Second VAR (VAR2): Governmen Expendiures on Economic Aciviies and Gross Domesic Oupu (LGSE, LGDP) The Third VAR (VAR3) : Reserve Requiremen Raio, Loans by banks and Gross Domesic Oupu (DRR, LDLB, LGDP) 4.1 The Co-inegraed Sysem Table 3 presens Johansen s maximum likelihood mehod es for coinegraion relaionship. The order of lags is seleced by he Akaike informaion crierion (AIC) and he Schwarz informaion crierion (SC) as six for VAR(1), one for VAR(2) and VAR(3) 8

10 Reserve Requiremens and Economic Growh (Table 3: Coinegraion Tes: Johansen s Tes) (a) VAR1: DRR, LSEG, LGSE Trace es Max-eigenvalue es H Trace Max-Eigen H 1 H Saisics 1 Saisics Criical Value 0.05 Criical Value r=0 r> * r= * r=1 r> r= Noe: * indicaes significan a he 5% level. r is he number coinegraing vecors. (b) VAR2: LGSE, LGDP Trace es Max-eigenvalue es H Trace Max-Eigen H 1 H Saisics 1 Saisics Criical Value 0.05 Criical Value r=0 r> * r= * r=1 r> r= Noe: * indicaes significan a he 5% level. r is he number coinegraing vecors. (c) VAR3: DRR, LDLB, LGDP Trace es Max-eigenvalue es H Trace Max-Eigen H 1 H Saisics 1 Saisics Criical Value 0.05 Criical Value r=0 r> * r= * r=1 r> r= Noe: * indicaes significan a he 5% level. r is he number co-inegraing vecors. I is clear from he ables ha he boh race saisics and maximal eigenvalue saisics unanimously indicae one coinegraing relaion a 5% significan in hree differen VAR sysems. More precisely, in each VAR model, has reduced rank of 1 so ha one coinegraion vecor is presen in β. In sum, herefore, he coinegraion es suggess ha here exis, for Souh Korea one long-run relaionship in a VAR1, VAR2 and VAR3. 9

11 Reserve Requiremens and Economic Growh The resuling normalised parameers esimae of our model is obained as follows: (Table 4. Normalized coinegraing coefficiens) (a) VAR1: DRR, LSEG, LGSE ˆLGSE ˆLSEG ˆDRR Coefficien ( ) * ( ) Noe: he number in parenheses denoes he sandard error, * indicaes significan a he 5% level (b) VAR2: LGSE, LGDP ˆLGDP ˆLGSE Coefficien ( ) * Noe: he number in parenheses denoes he sandard error, * indicaes significan a he 5% level (c) VAR3: DRR, LDLB, LGDP ˆLGDP ˆDRR ˆLDBL Coefficien ( ) * ( ) * Noe: he number in parenheses denoes he sandard error, * indicaes significan a he 5% level Normalized co-inegraing coefficiens are esimaed o inerpre he long run elasiciy of dependen and is independen variables. In VAR1, in a long run, a one percen increase in seigniorage(lseg) leads o an increase in governmen expendiures(lgse) by 0.39 percen. 8 Since he coefficien is saisically significan, he effec of governmen inpu (LGSE) on oupu (LGDP) is quie subsanial in magniude. In he long erm, a one percen increase in governmen spending resuls in an increase of oupu by 0.89 percen. In VAR3, he oupu elasiciies wih respec o reserve raio and domesic credi by banks are 0.68 percen and 1.2 percen respecively. 8 LGSE LSEG LSEG ( ) LSEG LGSE 10

12 Reserve Requiremens and Economic Growh 4.2 Shor- Run Dynamics The relaionships beween he variables in hree differen models in he shor run are esimaed using a Vecor Auoregressive (VAR) model. This analysis can reveal in greaer deail if here is a causal relaionship beween variables and can also accuraely indicae he direcion of causaliy. Toda and Yamamoo(1995) demonsrae ha even when variables of he model are non-saionary, sandard asympoic heory sill applies in a VAR sysem condiion, ha he order of inegraion of he daa generaing process is less han he seleced lag lengh of he model. Wih regard o heir finding, employing non-saionary ime series in a VAR can be used when n lags have been chosen, where is he seleced leg lengh of he model and n is he order of inegraion of he variables. As Sims (1992) menions, a differen selecion in ordering leads o a differen recursive srucures of VAR. He argues ha he earlier lised variables in he VAR impacs he laer lised variables, whils he same effec does no apply vice-versa. Hence he soluion o his problem is ordering exogenous variables firs and endogenous variables hereafer. 9 Turning o he impulse-response analysis repored in Figure 1, he middle of figure 1a examines how he governmen expendiures respond o shocks in seigniorage and reserve requiremen raio. I shows ha wha migh happen o he level of governmen expendiures if here were a one-ime on sandard-deviaion shock o he seigniorage or reserve requiremen raio. The doed lines represen 95 percen confidence limis for he impulse-response funcions. I oulines ha he 1 % increase of seigniorage leads o a 0.6% increase in governmen expendiures by he second year. The level of governmen expendiure would coninue o decrease unil i effecively reaches zero. Furhermore, he response of governmen expendiure shock o innovaion of reserve requiremen raio shows an increase by he second year and decrease aferwards. The response of reserve raio o seigniorage increases 0.08 percen by he hird year. Figure 1b oulines ha he 1 % increase in he governmen spending recorded during he 10-year horizon deermines a 0.012% increase in he GDP, implying ha i is realised in he long run. In Figure 1c, a shock o reserve raio decreases domesic credi delivered by banks by around 0.03 percen. In addiion, he reacion of oupu o he domesic credi by banks exhibis an insan increase and decrease by 0. Surprisingly he response of oupu o reserve raio presens a similar paern o he Laffar- Curve in a heoreical model. I goes up o 0.01% and decrease by 0.2 percen. 9 VAR1- Reserve Requiremen Raio, Seigniorage and Governmen Expendiures, VAR2- Governmen Expendiure and Oupu, and VAR3- Reserve Requiremen raio, Domesic Credi by Banks and Oupu. 11

13 Reserve Requiremens and Economic Growh (Figure 1: Impulse Response Funcions) (a) VAR1: DRR, LSEG, LGSE 12

14 Reserve Requiremens and Economic Growh (b) VAR2: LGSE, LGDP (c) VAR3: DRR, LDLB, LGDP 13

15 Reserve Requiremens and Economic Growh 5 Conclusion In his paper, I develop a model of reserve requiremens, and growh ha allows us o examine he effecs of financial and moneary policy on long-erm economic growh. I can also provide some evidence as o why some emerging economies, such as Souh Korea, have shown he remarkable economic growh under a higher reserve requiremen policy in he early sage of economic growh. The following is he main resuls of he paper. The overlapping generaions model is developed here and I assume ha he governmen spends is revenue from reserve requiremens ino he real secors. I discover ha he reserve requiremens have wo differen effecs on economic growh. One is he posiive effec ha a governmen uses his source as an inpu for he producive secors. Hence i helps o allocae capial owards more profiable invesmens, increasing he average reurn o capial. On he oher hand, he negaive effecs are from he financial secor, as a higher reserve requiremen reduces aggregae invesmen for given levels of savings so ha i leads o adverse effecs on growh. In his paper, here is he presence of a nonlinear effec in he relaionship beween hese wo financial repression policies and economic growh, suggesing ha an increase in he reserve requiremens may have a posiive impac on growh up o a hreshold level. Moreover his heoreical finding is jusified wih he case of Souh Korea by a VAR analysis. This paper confirms ha public invesmen, which migh depend on sources of governmen revenue and he domesic credi by disribued banks, induces oupu. I find a posiive relaionship beween governmen spending, seigniorage and reserve requiremen raio in a shor run while here is a negaive relaionship beween he reserve requiremen raio and domesic credi by banks. The more imporan inerpreaion, noneheless, is ha Souh Korea migh have shown asonishing GDP growh by allowing is governmen o inerfere significanly wih he conrol of he bank reserve raio and inves on producive economic aciviies in he iniial sage of economic developmen. However i is imporan o noice a limiaion of his es since he reserve requiremens is considered in a domesic framework. Therefore, in fuure research i would be worh invesigaing if he effecs of he globalised financial markes, such as capial inflows, have an impac on he reserve requiremens. 14

16 Reserve Requiremens and Economic Growh Bibliography Aghevli, B. B. (1977). Inflaionary Finance and Growh. Journal of Poliical Economy, Vol.85, pp Basu, P.(2001). Reserve Raio, Seigniorage and Growh, Journal of Macroeconomics, Vol.23, pp Barro, R.J. (1990). Governmen Spending in a Simple Model of Endogenous Growh. Journal of Poliical Economy, Vol.98, pp Chari, V.V., Jones, L.E. and Manuelli, R.E.(1995). The Growh Effecs of Moneary Policy. Federal Reserve Bank of Minneapolis Quarerly Review, Vol.19, pp Diaz-Gimenez, J.,Presco,E.C.,Fizgerald,T. and Alvarez,F. (1992). Banking in Compuable General Equilibrium Economies. Journal of Economic Dynamics and Conrol, Vol.16, pp Feinsein, C. H. (1968).Domesic Capial Formaion in he Unied Kingdom, Cambridge: Cambridge Universiy Press. Haslag, J.H.(1998a). Moneary Policy, Banking and Growh. Economic Inquiry, Vol.36, pp Nelson, C.R. and Plosser,C.I.(1982). Trends and Random Walks in Macroeconmic Time Series: Some Evidence and Implicaions. Journal of Moneary Economics, Vol.10, pp Roubini, N. and Sala-i-Marin, X.(1995). A Growh Model of Inflaion, Tax Evasion and Financial Repression. Journal of Moneary Economics, Vol.35, pp Sims, C.A. (1980). Macroeconomics and Realiy. Economerica, Vol.48, pp Surm, J.E. and Hann, D.J. (1995). Is Public Expendiure Really Producive? New Evidence from he USA and he Neherlands. Economic Modelling, Vol.12, pp Toda, H.Y., and Yamammoo, T. (1995). Saisical Inerence in Vecor Auoregressions wih possibly inegraed processes. Journal of Economerics, Vol.66, pp