Can policy improve liquidity during a

Transcription

1 Can policy improve liquidiy during a financial crisis? 5 Can policy improve liquidiy during a financial crisis? Ivo Karilaid, Tõnn Talpsepp 1 Absrac This paper empirically examines he developmen and deerminans of he liquidiy posiion in he financial secor during he recen financial crisis in he Balic-Scandinavian region. We look a fiscal and moneary policy implicaions of liquidiy problems arising in he crisis. The resuls are consisen wih heoreical predicions for a small open economy wih he expeced sign of changes and developmens in common economic indicaors. Changes (and he speed of changes) in ineres raes, GDP and money supply have occurred relaively rapidly, meaning ha he rising area of he LM-curve has been shorer han heory would predic. Marke reacions ook place quickly and simulaneously here was no ime for slow resrucuring, so ha liquidiy needs were higher han usual. Keywords: financial crisis, policy implicaions, liquidiy posiion, capial flows JEL classificaion: G1, G21 1. Inroducion and background The economies of he Balic Saes have been amongs he bigges decliners in Europe as well as in he world during he financial crisis ha sared in 27. The crisis has had a disressing effec on hese small and open economies, while liquidiy problems on he world s financial markes and especially in he region have worsened he siuaion even furher. A number of papers draw parallels among previous crises and he curren crisis as well as finding causes for he crisis. Bu he aspec of disappearing liquidiy which causes furher problems for whole counries and wheher policy measures can improve he siuaion has so far gained lile aenion during a ime when we can sill only alk abou an ex-ane view of he curren crisis. The purpose of his sudy is o explain how he capial marke siuaion and he relaive srucure of financial companies asses (money supply) are relaed o changes in main economic indicaors. Thus he paper provides an ex-ane analysis of liquidiy problems during he recen financial crisis and he preceding period. The research is relaed o heoreical 1 Tallinn School of Economics and Business Adminisraion (TSEBA) a Tallinn Universiy of Technology, Akadeemia ee 3, Tallinn, Esonia, ivo.karilaid@seba.u.ee, onn.alpsepp@u.ee

2 6 Balic Journal of Economics 1(2) (21) 5-26 examinaion of he IS-LM model on specific areas 2 of he LM-curve, sequence of changes and lenghs of phases of he curve. All of ha has an impac on raional decisions of capial marke ransacions, resuling in he money marke moving o he nex heoreical equilibrium poin. Theoreical sources of he hypohesis are suppored by he Law of Say and he Law of Verdon. One year afer publishing he heory of Keynes, John Hicks wroe an aricle Mr. Keynes and Classics: Suggesed Inerpreaion 3 in which he presened a simplified conclusion of Keynes s work. Afer elaboraion by Alvin Hansen 4 and oher auhors, his became he model of Hicks (IS-LM model), which in 195 won populariy as Keynesian economic heory. Like any oher model i is also a simplified version of describing processes in he real economy. The reason why his model is sill used nowadays is because i is simple (Brown, 1997). The main advanage may also lie in he fac ha he model includes one of he basic corner-sones of he economy, such as money (in his case money as a sock variable, no income as a flow variable), which can be deermined when referring o he ime of research. Addiionally he erm sock regulaing process 5 is relaed wih he heory of Keynes and he Law of Say, explaining how he marke economy can head lower han full employmen equilibrium despie he fac ha all producion will be bough (Brown, 1997). For developing he sock regulaing process, savings and invesmens should be used. These by naure are also socks and can be aken as sock variables. Under condiions of depression, economy can fall ino he liquidiy rap, which is he Keynes area of he LM-curve when he curve is quie horizonal. The opposie siuaion is he verical area of he LM-curve, which occurs in a booming economy when all money is used for business ransacions. In beween here should be a more realisic area of he LM-curve when he curve is rising, meaning ha money is divided beween bank accouns and business ransacions. Theoreically, shifs from he verical area o he horizonal area should ake some ime bu empirical daa show ha he ime for shifs can be significanly shorer during a crisis. Theoreical models illusrae how imporan i is o manage socks in he economy and how money is a sandard for measuring socks wih is characerisics like ineres raes and volume. Currenly we sill canno say ha he financial and economic crisis ha spread all over he world has ended. Thus, his paper acs as an ex ane sudy of he consequences of financial crisis for he money marke, more precisely liquidiy issues arising and policy implicaions of problems faced in he crisis. 2 The LM-curve is probably seady and rising bu i is efficien o rea hree areas separaely. Under condiions of deep depression, he economy can fall ino a liquidiy rap - his is Keynes s area and he curve is quie horizonal. In he case of fas expansion and high ineres raes, all money will be used for business ransacions and he LM-curve will become quie verical - his is called he classical area. The mos imporan is a rising LM-curve when money is used for business ransacions and money is also on speculaive accouns (Gordon, 1994). 3 John Hicks, Mr. Keynes and he Classics: Suggesed Inerpreaion, Economeria, 5 (April 1997), pp A Guide o Keynes, New York: McGraw-Hill, The sock regulaing process akes place in he following manner. When companies are producing more han hey sell, hen sock will accumulae. When sock increases over he desired level and here is nowhere o realize i hen an enerprise will sop aciviies in ha direcion and will fire employees relaed o ha acion (inpus). In he following period, an enerprise can realize a he expense of previously accumulaed sock more han i produces in ha period. If his process coninues, inpus should be increased for he nex period o increase producion. There is an equilibrium when enrepreneurs realize jus as much as hey produce and he sock remains relaively he same.

3 Can policy improve liquidiy during a financial crisis? 7 We approach he problem from a LM-curve based heoreical background and ake an indirec approach by sudying economic indicaors o see he shif beween he booming area in he LM-curve and he crisis area. We sudy how he level of overnigh ineres raes has flucuaed in he Balic counries (Esonia, Lavia and Lihuania) sudied as well as in he Scandinavian counries (Finland, Sweden and Denmark) sudied. In he LM-curve framework, ineres raes are more easily observable han demand and supply quaniies a differen ineres rae levels. Thus, daa availabiliy problems dicae our indirec approach by looking a differen relevan economic indicaors and drawing mosly qualiaively explained conclusions from he ime series and cross secional differences of economic indicaors wihin he sample used. The sample has been chosen since i enables a simulaneous look a hree small open economies (namely he Balic Saes) o see wheher and how much he policy makers of hose counries can influence he siuaion by providing and improving he overall liquidiy poin of view. A he same ime he Balic Saes are grealy influenced by European Union counries, as daa show (see e.g. Table 1 and Table 2) especially by neighboring Scandinavian counries. In he framework of liquidiy and he money marke, he influence is eviden due o he fac ha mos of he major banks in he Balic Saes are owned by financial insiuions of Scandinavian origin. Alhough our focus is on he money marke, we canno neglec he imporance of oher economic indicaors. From he policy maker s perspecive, we have o look a as complee an economic picure as possible. The money marke as a fundamenal par of he capial marke is influenial in invesmen decisions and gives one of he key indicaions abou he discoun rae. The mos widely used value bases in managemen models in invesmen decisions are NPV (Ne Presen Value), PI (Profiabiliy Index), EP (Economic Profi), EVA TM (Economic Value Added) ec. These models are relaed o he basis of discouned fuure cash flows, which means ha he value depends direcly on he discoun rae. Accordingly here is a fundamenal influence on invesmen decisions and he asses srucure of companies. As o sudying he hisory of financial crises, ypical characerisics are presen in he economy ha precede a crisis and influence how he economy ends o move ou of crisis. Reinhar and Rogoff (29) sudy all he financial and economic crises in he world during he period They find ha financial and economic crises end o be preceded by rapid economic expansion wih increasing consumpion and booming prices in differen real and financial secors. GDP reaches is larges decline one year afer he sar of he crisis. Unemploymen is regarded as a lagging indicaor and sars o decrease during he firs or second year of a crisis, no before. There is no a clear rend concerning inflaion bu depending on he specific broader economic condiions, inflaion can sill be high during he sar of he crisis and sar declining hereafer as economic expansion urns ino recession. Those main characerisics are also presen in he curren sample of counries. Kaminsky and Reinhar (1999) show ha usually inflaion during a domesic deb crisis is clearly higher compared o an exernal deb crisis. As currenly rising levels of domesic deb were he drivers of pre-crisis economic growh in he Balic Saes, a relaively high level of inflaion was more expeced han would have been he case in exernal (governmen) deb problems. Usually a very high co-movemen occurs in he share of counries having high inflaion and defauling on heir sovereign deb (Reinhar and Rogoff 29) bu currenly his

4 8 Balic Journal of Economics 1(2) (21) 5-26 did no concern he Balic Saes sudied. Oherwise boh exernal and inernal deb problems are common causes of crisis. Experience from pas crises (Reinhar and Rogoff 29) shows ha counries experiencing sudden large capial inflows are a high risk of experiencing a deb crisis. This can lead o over borrowing in good imes, leaving counries vulnerable during he ineviable downurns. The same happened in he Scandinavian counries (Mai 28) during he banking crisis of he 199s which should have been a lesson o bankers in he Balic-Scandinavian region bu reoccurred in he Balics during he curren crisis. A high probabiliy exiss of he curren accoun balance being negaive before a crisis as higher consumpion and impors end o overake expors. This is exacly wha happened in he Balic Saes. Eiher by devaluaion of he local currency or resrucuring he economy and producion, exporing goods sars o pick up during a crisis. Thus, weakening he currency or he necessiy for devaluaion (in he case of pegged currencies) are very common arifacs of economic and financial crises. Thus fears of devaluaion of local currencies (worsening he liquidiy siuaion in local markes) in he Balic Saes were jus expecaions based on lessons of pas crises. A number of common policy implicaions help o avoid he wors oucomes during a crisis. Namely, having a complee picure of governmen indebedness is criical. Deb susainabiliy mus be based on plausible scenarios for economic performance and is a mus facor in he possibiliy of sudden sops in capial flows. Inflaionary risks o moneary policy frameworks seem o be linked in imporan ways o levels of domesic deb. Many governmens have a empaion o inflae away domesic deb. Using simulus packages has become widespread during he curren crisis bu such packages have no been widely used (wih some excepions) during pas crises, which makes he success and efficiency of such a simulus harder o predic. The remainder of he paper is organized as follows: Secion 2 gives a brief overview of he daa and mehodology used. Secion 3 empirically analyses he curren crisis in he Balic and Scandinavian counries based on analysis of he dynamics of a large number of economic indicaors and proposes policy measures o sofen he emerging negaive consequences. Concluding remarks are offered in Secion Daa and mehodology The sudy concenraes on he las financial crisis and he preceding period wih daa used from 1994 o 29. The sample consiss of he hree Balic Saes (Esonia, Lavia and Lihuania) and hree Scandinavian counries (Finland, Sweden and Denmark). The sample has been chosen as i enables a view of how he banking secor of hree small open economies is influenced by neighboring more developed counries, which has a clear impac on he inflow of funds o he smaller counries. Daa used in empirical analysis come from wo main sources. These are Daasream and Inernaional Moneary Fund IFS daa collecion. We use he laes available daa. In he case of

5 Can policy improve liquidiy during a financial crisis? 9 seasonally unadjused daa, we use seasonal adjusmen echniques. For beer comparabiliy, we use he average exchange raes in siuaions where comparing he magniude of he series across counries is necessary. Daa on ineres raes come from Daasream as well as mos economic indicaors. Daa abou reserves, foreign rade and invesmens, as well as daa concerning he asses and liabiliies srucure of financial insiuions come from IFS daabank. We look a he dynamics of he asse srucure of he financial secor. By asses we mean financial asses in classical erms. Deposiory insiuions are used because commercial banks are he mos imporan paricipans in money and capial markes in he Balic Saes and he influence of oher players in his field is sill small (Kein, 1999). As he paper analyses he siuaion from he LM curve perspecive, daa availabiliy dicaes he indirec approach used. We look a differen relevan economic indicaors and draw mosly qualiaively explained conclusions from he ime series and cross secional differences of he economic indicaors. We run a number of single regressions o verify he relaionship beween differen economic indicaors boh wihin a single counry and beween he counries viewed. We presen a correlaion marix of he resuls and addiionally run coinegraion ess o verify he long erm relaionships beween he indicaors. The Dickey Fuller es is uilized for esing saionariy of he series and Johansen coinegraion ess are run afer ha. We run coinegraion ess boh wih he original adjused ime series as well as on he firs differences of he ime series. 3. Empirical sudy of he crisis 3.1. Ineres raes When sudying he ineres rae co-movemen of Balic and Scandinavian counries, we can see ha he ineres raes of Denmark, Sweden and Finland coincide grealy wih Euro ineres raes. Alhough Denmark and Sweden have no adoped he Euro, heir cenral banks have lowered heir raes in quie a similar manner o he European Cenral Bank as can be seen from he overnigh ineres rae (see Appendix A) which closely follows he base rae of he counry. The ineres raes of he Balic Saes have behaved slighly differenly during he crisis. During he period 2 o 27 we can see more volailiy, especially in Lavian and Lihuanian raes. A remarkable decrease occurred in overnigh raes afer he end of he crisis. As he currencies of all hree Balic Saes are principally pegged o he Euro, he cenral banks of hese counries do no have he means o direcly influence he rae by money supply. Basically, policy makers canno influence liquidiy on ha level. Sill, he implicaions of he recen model of Brunnermeier and Pedersen (29) sugges ha cenral banks can help o miigae marke liquidiy problems by easing funding or margin requiremens. Even public saemens ha exra funding will be provided during liquidiy dry-ups could help. The dramaic increase (and also decoupling from Euro raes) of he overnigh and also longer erm ineres raes in he Balics clearly illusraes a siuaion where cenral banks are unable

6 1 Balic Journal of Economics 1(2) (21) 5-26 o provide expansive measures, while local markes are more influenced by ouflow of foreign invesmen, which happened afer he beginning of 27. The speed in he change of ineres raes was dramaic. For example i ook 2 days in Sepember 1999 for he Lavian overnigh rae o rise from 3.5% o over 8% and less han 1 days in March 27 o rise from around 2% o 8%. Such quick changes do no allow he economy o adjus o changes in ineres raes and shrinking money supply. Appendix B illusraes he siuaion furher. When Scandinavian counries as well as he Euro area have on average been able o coninue o hold a more or less sable posiive rend in money supply even afer GDP levels sared o deeriorae, he same has no happened in he Balics. Especially Lavia and Lihuania have faced an even more abrup decline in money supply han he dramaic fall in GDP figures. A leas unil he second half of 29, Esonia did a lile beer. A classic policy makers approach in he case of an economic downurn is o sar cuing ineres raes o simulae he economy. Anoher negaive consequence is loss of confidence in he financial sysem during a crisis which, among various negaive effecs, can also induce bank-runs 6. Currenly, cuing ineres raes is exacly wha has happened in he Euro area as well as in Sweden and Denmark, saring from 28. Ineres raes have dropped significanly moving in he same direcion as falling GDP, which we use as a primary proxy for assessing he sae of he economy. The same has no been possible in Esonia, Lavia and Lihuania. Ineres raes in Esonia, Lavia and Lihuania have moved in he opposie direcion due o shrinking money supply, loss of confidence and rising devaluaion expecaions. Thus i should no be oo surprising ha in addiion o a shrinking economy, decreased money supply and increased ineres raes have worsened he siuaion even furher. Alhough ineres raes have no been direcly conrollable by policy makers in he Balics, he high level of ineres raes has parly also been caused by lack of confidence of marke paricipans in he perseverance of currency pegs and he banking sysem. I could be argued ha injecing confidence in marke paricipans as well as preserving a sound economic climae (we mean condiions under which he cenral governmen is able o conrol he level of exernal public deb and is expendiure) could help o smooh he magniude of negaive currency speculaion. On he oher hand, alhough simulus packages have no been oo widely used during pas crises in he world (Reinhar and Rogoff, 29) (wih he no oo successful excepion of Japan), decreased governmen expendiure and lack of exernal financing opions will decrease money supply. Thus, policy makers in pegged currency sysems can have expansive means on he capial markes basically only when borrowing and spending have been conservaive during good imes, so ha exernal financing and assurance of sabiliy remain possible during a crisis when he confidence of financial markes has eroded. In a crisis, ouflow of funds occurs no only from small emerging economies bu also from mos oher counries. We also look a spreads beween shor-erm (overnigh as well as 6-monh) ineres raes and long-erm ineres raes (we use 1-year governmen bonds for ha). Tighening of spreads 6 Allen and Gale (1998) sudy he opimaliy of choice regulaors and cenral banks have o make when dealing wih he risks associaed wih crises o avoid bank-runs.

7 Can policy improve liquidiy during a financial crisis? 11 during a crisis epicener can be regarded as an expeced resul as shor-erm ineres raes rise higher han long erm governmen bond yields in all counries in he sudy. Spreads beween shor-erm and long-erm yields sared o ighen in 24 and urned negaive by he firs half of 27. The adjusmen in spreads back o he levels of 24 was very rapid and ook less han four monhs saring from he end of Sepember 28 in he Scandinavian counries 7. Thus changes in long-erm ineres raes are slighly smooher han in overnigh raes. Bu compared o 1-year raes, shor-erm ineres raes have a clearly sronger impac on curren liquidiy posiions Inflaion and real ineres raes By heir inflaionary levels, Esonia and Lihuania have been in a beer posiion han Lavia. Lavia had no seen lower han 5% inflaion (we look a CPI) since he beginning of 24 unil he over 1% levels of 27 and 28 dropped below 5% in he second quarer of 29 and have been decreasing since. Inflaion also rose above he 1% level in Esonia and Lihuania in 28. Before ha i sayed around he 5% level and was increasing since 27. Pre-crisis years clearly indicae ha he Balics were facing a very high inflaionary environmen, which in urn encouraged invesing (and seemingly also for households o coninue spending) as much as possible and real ineres raes were negaive due o high inflaion. In he ligh of providing liquidiy, he pre-crisis years araced foreign inflows of money. When soaring ineres raes and he inflaionary environmen urned ino deflaionary, his discouraged invesmen as real ineres raes were also expeced o rise in he ligh of diminishing inflaion (by he hird quarer of 29, only Esonia had shown negaive CPI). Real ineres raes also increased because he siuaion of shrinking money supply and devaluaion worries encouraged commercial banks o keep nominal ineres raes high on deposis. Such an effec caused a siuaion where holding money on deposi wih minimum risk was rewarded by high real ineres raes and invesing funds was discouraged by higher risk of invesmen in a shrinking economy. This caused a siuaion where more funds were waiing on he sidelines (read: were lying in deposi accouns) han invesed in an economy facing ouflow of foreign funds a he same ime. Table 1. Correlaion beween changes in counry GDP in EUR from Denmark Esonia Euro Area Finland Lavia Lihuania Esonia.88 Euro Area Finland Lavia Lihuania Sweden Overnigh ineres raes sayed higher han long-erm ineres raes ill he end of 29 in Lavia, which was caused by devaluaion fears.

8 12 Balic Journal of Economics 1(2) (21) Economic growh We look a GDP as he main indicaor of economic well-being. To make he figures comparable, we calculae GDP for all counries in Euro and adjus he ime series for seasonaliy. As he Swedish kronor has been he single mos volaile currency agains he Euro and has weakened nearly 2% agains he Euro since he firs half of 28, his pus Swedish economic performance in a darker ligh han would be measured in local currency. In Table 1 we presen he correlaion beween he GDP (measured in EUR) growh raes of he counries sudied since 2. Viewing longer periods would make he correlaion beween he Balic Saes higher and using local currencies would increase he correlaion beween Sweden and he res of he sample. We would assume significan influence by he Scandinavian counries on he Balic Saes, as much of he financial secor in he Balics is under he conrol or direc influence of banks of Scandinavian origin. Indeed, he correlaion of GDP growh raes is quie significan in all cases. Finland has a nearly.9 correlaion wih all hree Balic Saes, he highes figure wih Esonia. Tha is, considering heir geographical proximiy, an expeced resul. Sweden again has a higher correlaion wih is neighboring counries and a slighly lower correlaion wih Lavia and Lihuania. Alhough we canno show causaliy in hose relaionships, all counries do seem o have a higher correlaion wihin he Balic Sea region han wih he average Euro area. When sudying he saring poins of he economic downurn in he area, we look a he ime series of GDP in boh Euro and local currency. Alhough Sweden showed quie a sable GDP growh in he fourh quarer of 27 in local currency, he weakening of he kronor decreased growh o only.5% when measured in Euro. All oher counries (wih a minor excepion of Denmark) were sill experiencing a srong growh phase (especially Lavia and Lihuania) during ha ime. Already he nex quarer mean a surprising urn ino negaive erriory for Esonia and Denmark, wih Lavia following a quarer laer and all oher counries in he fourh quarer of 28. This effecively urned all hopes of a sof landing for he economy ino an inelucable global crisis ha had already suffered significan sebacks earlier in auumn 28 wih plummeing financial markes and an unseen surge of volailiy afer Lehman Brohers was forced o declare bankrupcy. As Sweden is he larges economy in he sample viewed, one migh expec a greaer influence from here on smaller neighboring counries. We can qualiaively argue ha he influence of Swedish financial secor worries was he mos eviden in Esonia and Lavia. Alhough Swedish GDP growh sayed quie sable unil he fourh quarer of 28 when looking a he figures in kronor, Sweden was - along wih Esonia - he leading counry in slowing and urning negaive when considering he framework of he EU and looking a GDP figures in Euro. We can also see (Appendix A) ha he sock marke (which is considered one of he main leading indicaors) sared o fall firs in Esonia and Sweden. Correlaions wihin counries (see Table 2) beween money supply (eiher M2 or M3) and GDP are srong, ranging beween.82 for Lavia (.84 for Finland) and.93 for Esonia (.91 for Sweden). Especially for he Balic counries, coinegraion es resuls mosly confirm he correlaion findings. The correlaion beween money supply and ineres raes has an expeced

9 Can policy improve liquidiy during a financial crisis? 13 Table 2. Correlaion and coinegraion marix of Euro zone, Esonian, Lavian, Lihuanian, Finnish, Swedish and Danish GDP, money supply (M2 or M3) and average overnigh ineres raes (i) for he period Q o Q3 29.The able presens correlaions in he lower righ par of he marix and resuls of coinegraion ess in he upper righ par of he marix. Coinegraion resuls are presened if eiher series were found o be coinegraed (+) or no (-). Coinegraion es resuls are presened boh for he firs difference of he series (he firs + or -) as well as he original seasonally adjused series (he second + or -). Uni roo EUR M2 EUR i EUR GDP SWE i SWE M2 SWE GDP LIT i LIT M2 LIT GDP DEN i DEN M2 DEN GDP FIN i FIN M2 FIN GDP LAT i LAT M2 LAT GDP EST i EST M2 EST GDP EUR M2 yes EUR i yes EUR GDP yes SWE i yes SWE M2 yes SWE GDP yes LIT i no LIT M2 yes LIT GDP yes ā + + DEN i yes DEN M2 yes DEN GDP yes FIN i yes FIN M2 yes FIN GDP yes LAT i no LAT M2 yes LAT GDP yes EST i yes EST M2 yes EST GDP yes

10 14 Balic Journal of Economics 1(2) (21) 5-26 negaive sign (excep for Lavia - daa problems conneced wih he availabiliy of money supply daa) and falls in he range from -.22 o -.52 bu has very clear coinegraion es resuls confirming he relaionship Reserves, invesmen posiion and foreign rade The curren crisis siuaion is well refleced also in he inernaional reserves which reflec he moneary policy of cenral banks. For example he reserves of Denmark more han doubled in less han six monhs saring from Q3 28 and jumped by 4% in June 29 for Sweden. The changes have no been so drasic in Finland and in he Balics wih pegged currencies and limied abiliy o enforce moneary policy, bu growing reserves are eviden for he whole sample. When comparing inernaional reserves o oal deposiory financial insiuion asses, he changes in reserves are no as abrup bu can sill indicae ha piling up reserves had negaive effecs on liquidiy (and money supply) in he economy. In Q2 29 reserves amouned o approximaely 6%, 4% and 2% of deposiory asses in Denmark, Sweden and Finland respecively, bu 12%, 9% and 15% in Esonia, Lavia and Lihuania which have o hold larger reserves due o pegged currencies. Mainaining he reserve level urned ou o be a challenge for Lavia due o ouflow of foreign funds (Transiion Repor 29). Sudying he invesmen posiion of hese counries shows clearly in all cases invesmens saring o pull ou from foreign counries (see Appendix C). This is similar for boh direc and porfolio invesmens when domesic ineres raes soar and money supply sars o decline along wih a shrinking economy. The Scandinavian counries have a larger amoun of foreign porfolio invesmens, which should be a clear indicaor of more developed financial markes. Figure 1. Boosing inernaional reserves during a crisis. The lef verical axis presens figures in billion SDR for Scandinavian counries and he righ verical axis in billion SDR for he Balic Saes. 45 Inernaional reserves 2 Denmark 3 15 Sweden 15 1 Finland Lihuania Lavia Esonia 1994Q1 1995Q3 1997Q1 1998Q3 2Q1 21Q3 23Q1 24Q3 26Q1 27Q3 29Q1

11 Can policy improve liquidiy during a financial crisis? 15 The Balic Saes end o have less porfolio invesmens and are dominaed by a very large share of direc invesmen. When comparing Esonia, Lavia and Lihuania amongs each oher, we can see ha alhough he Esonian economy is he smalles of hese in absolue numbers, i has araced he highes number of boh direc and porfolio invesmens. In he case of an economic downurn, hose invesmens sar o seek a way ou. In he case of disappearing liquidiy, his is no an easy ask and sars o grealy affec he domesic economy 8. Esonia is he counry mos affeced by ouflow of invesmens as is dependence on invesmen is he greaes (highes share of foreign invesmen per GDP in local economy). We can see a clear correlaion beween falling sock marke prices and decreased value and ouflow of foreign porfolio invesmens, which indicaes ha sock marke prices are grealy influenced by foreign invesors originaing from and wihin he Scandinavian region. In he sample, he correlaion is he sronges for Finland and Esonia. Unforunaely we do no have daa available for Sweden, which probably has he bigges influence on neighboring counries sock markes. For Euro area and Scandinavian counries, he drop in foreign porfolio invesmens has been larger han for direc invesmens. A he same ime hey have also pulled ou (or los value in) heir own porfolio invesmens abroad. As he Balic Saes are grealy influenced by hose counries, plummeing sock indexes and lending by foreign owned banks almos a a sandsill should no be a big surprise in rerospec. Currenly we can only qualiaively argue ha he drop in direc foreign invesmen in he Balic economies can have a longer erm negaive effec. Generally, porfolio invesmens are more mobile, meaning ha hey can move in and ou of he economy faser han direc invesmens. The laes available daa shows some promising signs for he Balics as boh Esonia and Lavia seem o be showing he firs signs of sabilizing he level of foreign direc invesmen in he economy. Due o a high correlaion beween he Balic Saes, Lihuania can be expeced o follow heir lead. The amoun of invesmen abroad is no very high for any of he Balic Saes, 9 which can be one reason why foreign invesmens abroad have no los as much relaive o Scandinavian counries. On he oher hand, Scandinavian counries have been donors for inflow of invesmens ino he Balic Saes and he falling Balic economies have negaively affeced heir invesmen values and decisions. Considering he size and correlaion of he Scandinavian economies viewed, drops in foreign invesmen are probably affeced by pulling ou porfolio invesmens from each oher s economies and he Balic economies are negaively affeced from he ouflow of funds iniiaed by he Scandinavian side. The inflow of funds o all Balic Saes had more han ripled from 24 o he end of 27. As presened in he anaomy of crisis secion of he paper, such a sudden large inflow of funds can lead o borrowing more han necessary, high inflaion and hus can ofen cause rouble laer when he inflow sops or reverses. This is one area where policy makers could heorei- 8 See Masso, Varblane, Vaher (28) and Vaher and Masso (25) for a deailed sudy of spillover effecs of inward and ouward FDI. 9 Esonia has a clearly higher level of boh foreign direc invesmen and foreign porfolio invesmens abroad o GDP compared wih Lavia and Lihuania. The level of ouward FDI o GDP for Esonia is approaching he level of developed Scandinavian counries bu is sill lower and no wih comparable magniude in absolue erms.

12 16 Balic Journal of Economics 1(2) (21) 5-26 cally have imposed regulaive means o discourage domesic borrowing from a cerain level by higher capial requiremens. As inflow of foreign invesmens during good imes helps o boos he economy, moivaion for policy makers o discourage such inflow is exremely low, resuling in higher volailiy in he economy. As expeced, we can see an increasing curren accoun defici in all hree Balic Saes during imes of booming economy (see Appendix D). A he same ime, he capial accoun sayed posiive. Thus he Balic Saes have followed a more classical and expeced road o crisis han he Scandinavian counries, where changes in curren or capial accoun have no been oo noiceable. Boh expor and impor have moved in sync wih GDP and faced significan drops in 28. A clearly larger drop in impors compared o expors in he Balics has been very harsh bu had some sobering effecs on economies ha obained heir previous growh from consumer expendiure. The posiive side of such a drop in impors is ha he rade balance is saring o urn posiive, which is necessary o resar he economy, as pas crises have shown Governmen and household expendiure and invesmen We look a growh in gross fixed capial formaion (GFCF), household and governmen expendiure compared o GDP growh, which show a high correlaion wih each oher 1. We run a number of single linear regressions wih growh raes of GDP as he righ hand side variable and growh raes of expendiure or invesmens as he lef hand side variable o measure he sensiiviy (or slope coefficien bea) of hose indicaors o changes of GDP. GFCF boh increases and decreases wih a magniude larger han GDP (bea around 2 for all counries). As could be expeced, governmen expendiure is more rigid and does no adjus o GDP decline as easily as for GDP growh. Governmen expendiure has a low correlaion wih GDP growh for Scandinavian counries and Esonia bu is slighly higher for Lavia (.37 correlaion coefficien) and Lihuania (.21). Those wo counries also have a higher bea for governmen expendiure reaching near and over.7, which sill means ha over he period viewed ( ) governmen expendiure was no much influenced by he curren period of economic growh. Governmen expendiure is one of he few main economic levers ha can be direcly affeced by policy makers decisions. One of he main problems ha should have been an early warning sign for he Balic Saes is ha saring from 25 and 26, household expendiure sared o increase much more quickly han GDP. For previous periods hroughou he period viewed, such a problem did no exis, as also for Scandinavian counries. For all counries, he correlaion beween household expendiure and GDP has been around.8 for he period viewed bu for he Balics household expendiure has had a bea around and over 1 (i is less han.8 for Scandinavian counries) when compared o GDP growh. High domesic household expendiure growh in he Balics was also one of he main sources of inflaion during he pre crisis period. 1 Daa and figures are available upon reques.

13 Can policy improve liquidiy during a financial crisis? Deposiory insiuions balance shee srucure We look a how deposiory insiuions asses and liabiliies srucure has changed during he period under analysis. Well funcioning financial insiuions is one of he mos criical aspecs for providing liquidiy during a crisis, as disrupions in he banking sysem can have a harsh effec on aggregae economic aciviy (see e.g. Bernanke 1983). As expeced, Figure 2 shows overall growh in he asses of banks o he second half of 28 when he negaive effecs had no reached he balance shees of he banking sysem. The Balics were clearly leading privae secor credi growh among he emerging economies in he world from (Brixiova, Varia, Wörgöer 21) which had a clear posiive effec on he size of commercial banks balance shees. Afer ha poin he Balic Saes and Sweden (which has he mos exposed risks owards he Balics) faced a sligh drop in asses, which is a leas parly explained by loan losses. The rend in bank asse srucure in he Balics (see Appendix F), especially in Lavia and Lihuania, is ha claims on foreigners are decreasing and domesic exposure is rapidly grow- Figure 2. Growh of asses of deposiory financial insiuions. The lef verical axis presens figures for Scandinavian counries and he righ verical axis for he Balic Saes. Esonian daa sar from 24 and have an index saring poin of 1.7 o make i comparable wih he oher Balic Saes. 13 Growh of asses of deposiory financial insiuions Finland Denmark Sweden Lihuania Lavia Esonia Q1 29 Q1 28 Q1 27 Q1 26 Q1 25 Q1 24 Q1 23 Q1 22 Q1 21

14 18 Balic Journal of Economics 1(2) (21) 5-26 ing. In ha sense Denmark and Finland (see Appendix E) have been a quie a sable level bu foreign exposure of Swedish banks has been clearly growing since 24. A leas par of ha foreign exposure growh can be explained by financing Swedish owned banks in he Balic Saes. Claims on he cenral governmen have become less and less imporan in ime. The liabiliies side of bank balance shees clearly reflecs he Balic Saes reliance on inflow of foreign funds. Around 4% of liabiliies in Esonia and Lihuania and 5% in Lavia are o nonresidens. The majoriy of hese are loans from paren companies o local banks. In case of economic rouble, such funds can sar fleeing he counry, making he liquidiy posiion even harder, which has been especially rue for Lavia (see also Transiion Repor, 29). Alhough he liabiliies srucure does no show a very clear decrease in liabiliies o nonresidens during he crisis, we can sill observe a slighly shrinking share of foreign money, on op of which asses and liabiliies sared o decrease in general. This is empirical evidence ha foreign money is flowing ou more quickly han he decrease in asses. A large share of liabiliies o foreigners disors he overall liquidiy picure in he Balics. Leaving foreign liabiliies aside, we can see a sligh piling up of liquid asses on deposis searching refuge from he real economy and securiies invesmens which have been losing value since he beginning of he crisis. Such an effec is he mos eviden for Sweden, Finland and Esonia and also suppors he picure presened abou inernaional reserves. Anoher ineresing aspec is liabiliies o cenral governmen. This does no play an imporan role in Esonia bu has clearly increased in he oher Balic Saes and Denmark. This would be one place where governmen aid packaged o banks would be refleced. I is slighly surprising no o see any noiceable changes here for Sweden in he daa used. 4. Conclusions Empirical daa show ha changes in ineres raes, GDP and money supply occurred relaively rapidly and simulaneously in all counries of he Balic-Scandinavian region bu especially in Esonia and Lavia, so ha i did no give economies much ime o adjus. Tha caused a siuaion where ineres raes in he more fragile Balic economies decoupled from he Euro and Scandinavian area and soared almos o previous crisis heighs. This clearly illusraes a siuaion where cenral banks of open small economies are unable o provide expansive measures and local markes are more influenced by ouflow of foreign invesmens, which sared happening afer he beginning of 27. The correlaion of GDP growh raes is quie significan in all cases, which is posiively conneced wih geographic proximiy. All counries do seem o have a higher correlaion wih he Balic Sea region han wih he average Euro area. Thus, financial secor worries - especially in Sweden, Esonia and Lavia - closely influenced each oher as Sweden and Esonia were he leading counries wih slowing and negaive economic growh. High correlaion of GDP and money supply mean ha along wih soaring ineres raes, liquidiy on he markes shrank significanly. Afer he inflow of invesmens o all hree Balic Saes had more han ripled from 24 o he end of 27, he ouflow of funds iniiaed by he Scandinavian side (seen from he daa abou invesmen posiion and consolidaed balance shees of deposiory

15 Can policy improve liquidiy during a financial crisis? 19 insiuions) and piling up of inernaional reserves worsened he siuaion even furher in he Balics. Experience from previous crises suggess ha having a complee picure of governmen indebedness is criical and inflaing away domesic deb migh no be a good idea. In he Balic Saes, he cenral banks do no have clear moneary policy means o influence he money supply by ineres raes bu even public saemens on providing exra funding when necessary could help o injec confidence in he financial markes during liquidiy crises bu only if necessary buffers exis. Buffers could be achieved by governmen conrolling he level of exernal public deb as well as is expendiure during good imes. Policy makers in pegged currency sysems such as he Balic Saes can posiively affec he liquidiy posiion basically only when borrowing and spending has been conservaive enough during growh in he economy. Tha could improve he chances for exernal financing during a crisis. Operaing in a small and open economy can make policy makers use of levers less effecive due o high dependence on and correlaion wih larger neighboring economies. As he economies of he region sudied are sill under sressed condiions, he curren sudy reflecs only an ex ane view. This could be complemened by a more horough ex-pos sudy ha could ake ino consideraion he whole economic cycle. References Allen, F. and Gale, D. (1998). Opimal financial crises, Journal of Finance, vol. 53, pp Bernanke, B.S. (1983). Nonmoneary Effecs of he Financial Crisis in he Propagaion of he Grea Depression, American Economic Review 73, pp Brixiova, Z., Varia, L., Wörgöer A. (21). Capial Inflows, Household Deb and he Boom-bus Cycle in Esonia, OECD Economics Deparmen Working Papers No. 7. Brown, W. S. (1997). Makroökonoomika Baaseooria, Külim, Tallinn. Brunnermeier, M. K. and Pedersen, L. H. (29). Marke Liquidiy and Funding Liquidiy, Review of Financial Sudies. forhcoming. Daasream. (21). ThomsonReuers, Online Daabase, accessed Jan 15, 21. Fabozzi, F. J. and Modigliani, F. (1992). Capial Markes, Insiuions and Insrumens, Prenice Hall, Englewood Cliffs, NJ. Friedman, M. and Savage, L. J. (1948). The Uiliy Analysis of Choices Involving Risk, The Journal of Poliical Economy, LVI, 4. Gordon, M. J. (1994). Finance. Invesmens and Macro-Economics: The Neoclassical and PosKeynesian Soluion, Aldersho, U.K. Haggard, S., Lee, C. H, and Maxfield, S. (1993). The Poliics of Finance in Developing Counries. Cornell Sudies in Poliical Economy, Cornell Universiy Press, Iacha and London. Hirshleifer, J. (1965). Invesmen Decisions under Uncerainy: Choice Theoreic Approaches, The Quarerly Journal of Economics, LXXIX, 4. Inernaional Financial Saisics (IFS). (29). Inernaional Moneary Fund. IFS CD-ROM, November 29.

16 2 Balic Journal of Economics 1(2) (21) 5-26 Kaminsky, G. Reinhar, C.M. (1999). The Twin Crises: The Causes of Banking and Balance of Paymens Problems, American Economic Review, Vol. 89, pp Kein, A. (1999). Formaion of Securiies Marke in Esonia in 199s, In: Ennuse. Ü. and Wilder, L. (eds.) Harmonisaion wih he Wesern Economics: Esonian Economic Developmens and Relaed Concepual and Mehodological Frameworks. Esonian Insiue of Economics a TTU, Tallinn, pp Krugman, P. (1994). The Age of Diminished Expecaions: U.S. Economic Policy in he 199s, Seconded. Cambridge and London. Mai, N. (28). Lessons from he 199s Scandinavian Banking Crises, JPMorgan Chase Bank, London. Masso, J., Varblane, U. and Vaher, P. (28). The Effec of Ouward Foreign Direc Invesmen on Home-Counry Employmen in a Low-Cos Transiion Economy, Easern European Economics, M.E. Sharpe, Inc., vol. 46(6), pp Reinhar, C.M. and Rogoff, K. (29). This Time is Differen: Eigh Cenuries of Financial Folly, Princeon Universiy Press, New Jersey, USA. Thiessen, U. (1994). Aspecs of Transiion o Marke Economies in Easern Europe, Aldersho, UK. Transiion Repor. (29). EBRD Transiion Repor 29: Transiion in crisis?, European Bank for Reconsrucion and Developmen, November 29. Vaher, P. and Masso, J. (25). Home Versus Hos Counry Effecs of FDI: Searching for New Evidence of Produciviy Spillovers, Bank of Esonia Working Papers 25-13, Bank of Esonia, revised 1 December 25.

17 Can policy improve liquidiy during a financial crisis? 21 Appendix A. Ineres raes, GDP, sock marke and local currency. The boom pane of he chars presens figures for he Balic Saes and he upper pane for he Scandinavian couries for overnigh ineres raes, GDP and sock marke. GDP is presened in quarerly consan prices in million EUR. Lavian currency is presened in 1xLAT per EUR Overnigh ineres rae Euro area Esonia Lavia Lihuania Sweden Denmark Finland Sock marke index 2 Sweden 1 Finland Denmark Q Q1 2 Q1 21 Q1 22 Q1 23 Q1 24 Q1 25 Q1 26 Q1 27 Q1 28 Q Esonia Lavia Lihuania -2 GDP in EUR Local currency per EUR 1 8 Sweden Denmark Finland Sweden Denmark Lavia 1 4 Lihuania Lavia Esonia 4 2 Lihuania Q1 2 Q Q1 22 Q1 21 Q1 27 Q1 26 Q1 25 Q1 24 Q1 23 Q1 29 Q1 28 Q Q Q Q1 2 Q1 22 Q1 24 Q1 26 Q1 28

18 22 Balic Journal of Economics 1(2) (21) 5-26 Appendix B. Changes in ineres raes, GDP and money supply before and during he crisis. The lef verical axis presens changes in money supply and GDP. Money supply and GDP are indexed o heir iniial value a he saring poin of he daa. The righ verical axis presens changes in overnigh ineres raes. Daa cover he period from Q o Q ,5 Finalnd Esonia , ,5 Sweden Lavia , ,5 Denmark Lihuania , ,5 Euro area 6 Q Q Q Q1 21 Q1 23 Q1 25 Q1 27 Q ,5 1 M2 index (1) GDP index (1) Overnigh ineres rae Q1 29 Q1 27 Q1 25 Q1 23 Q1 21 Q Q Q1 1995

19 Can policy improve liquidiy during a financial crisis? 23 Appendix C. Direc and porfolio invesmens in he economy and abroad. Daa are presened in million USD from Q o Q Denmark 2 Esonia Finland 2 Lavia Euro area 2 Lihuania 5 1 Q1 28 Q1 26 Q1 24 Q1 22 Q1 2 Q Q Q1 28 Q1 26 Q1 24 Q1 22 Q1 2 Q Q DIRECT INV ABROAD (ouward FDI) PORTFOLIO INV ABROAD (ouward FPI) DIRECT INV IN ECONOMY (inward FDI) PORTFOLIO INV IN ECONOMY (inward FPI)

20 24 Balic Journal of Economics 1(2) (21) 5-26 Q1 28 Sweden Q1 26 Q1 24 Lihuania GOODS EXPORTS GOODS IMPORTS CURRENT ACCOUNT CAPITAL ACCOUNT Q1 22 Q1 2 Q Q Q Q1 28 Q1 26 Q1 24 Appendix D. Expors, impors, curren accoun and capial accoun. Daaares presened in million USD from Q o Q Esonia Lavia Finland Denmark Q1 22 Q1 2 Q Q Q Q1 28 Q1 26 Q1 24 Q1 22 Q1 2 Q Q Q1 1994

21 Can policy improve liquidiy during a financial crisis? 25 Appendix E. Asse and liabiliy srucure of deposiory financial insiuions in Denmark, Finland and Sweden. The lef pane presens he asse srucure in he following order: claims on oher secors, claims on cenral governmen, claims on nonresidens. The righ pane presens he liabiliy srucure in he following order: shares and oher equiy, oher shares excl. from broad money, liabiliies o cenral governmen, liabiliies o nonresidens, M2 deposis, M1 deposis. 1% 8% Denmark 1% 8% Denmark 6% 6% 4% 4% 2% 2% % % 1% 8% Finland 1% 8% Finland 6% 6% 4% 4% 2% 2% % % 1% 8% Sweden 1% 8% Sweden 6% 6% 4% 4% 2% 2% % % Q3 28 Q1 27 Q3 25 Q1 24 Q3 22 Q1 21 Q3 28 Q1 27 Q3 25 Q1 24 Q3 22 Q1 21 CLAIMS ON OTHER SECTORS CLAIMS ON CENTRAL GOVERNMENT CLAIMS ON NONRESIDENTS SHARES AND OTHER EQUITY SEC.OT.TH.SHARES EXCL.F/BROAD MONEY LIABILITIES TO CENTRAL GOVERNMENT LIABILITIES TO NONRESIDENTS M2 Deposis M1 Deposis

22 26 Balic Journal of Economics 1(2) (21) 5-26 Appendix F. Asse and liabiliy srucure of deposiory financial insiuions in Esonia, Lavia and Lihuania. The lef pane presens he asse srucure in he following order: claims on oher secors, claims on cenral governmen, claims on nonresidens. The righ pane presens he liabiliy srucure in he following order: shares and oher equiy, oher shares excl. from broad money, liabiliies o cenral governmen, liabiliies o nonresidens, M2 deposis, M1 deposis. 1% 8% Esonia 1% 8% Esonia 6% 6% 4% 4% 2% 2% % % 1% 8% 6% 4% Lavia 1% 8% 6% 4% 2% % 2% % Lavia 1% 8% 6% 4% Lihuania 1% 8% 6% 4% 2% % 2% % Lihuania Q3 28 Q1 27 Q3 25 Q1 24 Q3 22 Q1 21 Q3 28 Q1 27 Q3 25 Q1 24 Q3 22 Q1 21 CLAIMS ON OTHER SECTORS CLAIMS ON CENTRAL GOVERNMENT CLAIMS ON NONRESIDENTS SHARES AND OTHER EQUITY SEC.OT.TH.SHARES EXCL.F/BROAD MONEY LIABILITIES TO CENTRAL GOVERNMENT LIABILITIES TO NONRESIDENTS M2 Deposis M1 Deposis

23 LATCOIN: deermining medium o long-run endencies of economic growh in Lavia in real ime 27 LATCOIN: deermining medium o long-run endencies of economic growh in Lavia in real ime Konsanīns Beņkovskis 1,2 Absrac This paper presens a mehod of esimaing he curren sae of Lavia s economy. The evaluaion objec is medium o long-run growh of real GDP, bu no acual GDP iself, which helps o filer ou various one-off effecs and focus on medium and long-run endencies. Our indicaor, called LATCOIN (Lavia s Business Cycle Coincidence Indicaor), could be viewed as a simple adapaion of new EUROCOIN for Lavia wih some changes in mehodology. LAT- COIN is a monhly esimae of he medium o long-run growh of Lavia s real GDP, which is produced on he 9h working day of he nex monh. Using a large panel of macroeconomic variables, a few smooh unobservable facors describing he economy are consruced. Furher, hese facors are used for he esimaion of LATCOIN. Keywords: Lavia s real GDP, band-pass filer, coincidence indicaor, generalised principal componens, real-ime performance JEL classificaion: C22, C5, E32 Inroducion The mos complee and popular indicaor of economic aciviy is real GDP growh. However, wo main drawbacks are usually associaed wih his variable. Firs, informaion on domesic aciviy comes only on a quarerly basis and wih a significan delay. The second drawback is relaed o shor-run flucuaions of real GDP, which creaes a significan problem for analysing, forecasing, and decision making in real ime. Moneary policy makers are usually no ineresed in such flucuaions and are more concerned abou medium-erm and fundamenal endencies in he economy. The firs problem has already been addressed for he case of Lavia in several researches, which analysed shor-erm forecasing possibiliies of Lavia s real GDP. Meļihovs and Rusakova (25) checked he forecasing abiliy of business and consumer survey daa, Ajevskis and Dāvidsons (28) showed ha dynamic facor models provide good forecasing performance in he shor run, and Beņkovskis (28) used bridge equaions wih various conjuncural indicaors o forecas real GDP. All hese papers, however, do no address he second problem of shor-erm flucuaions in real GDP. 1 Moneary Policy Deparmen, Bank of Lavia, Konsanins.Benkovskis@bank.lv 2 The views expressed in his publicaion are hose of he auhor, an employee of he Bank of Lavia Moneary Policy Deparmen. The auhor assumes responsibiliy for any errors or omissions.