Dislocations in the FX swap and money markets in Hong Kong SAR during the global credit crisis of

Transcription

1 Dislocaions in he FX swap and money markes in Hong Kong SAR during he global credi crisis of 7 8 Laurence Fung and Ip-wing Yu I. Inroducion The subprime crisis emerged in he Unied Saes in mid-7 and spilled over o oher economies. From mid-7 o mid-8, he spillovers were relaively modes, bu he siuaion began o change in mid-8. Following he bankrupcy of Lehman Brohers in mid- Sepember 8, developmens ook a dramaic urn. One channel for hose spillovers was severe disrupions in inernaional money markes. Uncerainy abou losses incurred by banks increased banks liquidiy needs as well as heir relucance o lend o each oher in money markes. Reflecing hese, and possibly oher, facors, inerbank shor-erm ineres raes surged subsanially afer he failure of Lehman Brohers and persised a high levels, promping cenral banks around he world o adop unprecedened policy measures o supply funds o banks. The Hong Kong dollar (HKD) money marke was also hi hard by he liquidiy squeeze. The urbulence in he global inerbank markes also spilled over o he foreign exchange (FX) swap markes, including in Hong Kong SAR. In response o he sress in he local financial markes, he Hong Kong Moneary Auhoriy (HKMA) and he Hong Kong SAR Governmen announced a series of measures o help conain he global sresses from spilling over o he domesic banking sysem. In his paper, we examine he anomalies and disress experienced in he local financial markes and aemp o provide answers o he following quesions:. To wha exen was he disress ha originaed from US dollar (USD) money markes ransmied o he HKD money marke?. How effecive were he policy acions underaken by he HKMA and he Hong Kong SAR Governmen in miigaing he sress wihin he inerbank marke in Hong Kong SAR? 3. To wha exen did he policy acions help o improve he efficiency of he money and FX swap markes in Hong Kong SAR? II. Transmission of erm funding shock and volailiy linkage In his sudy, we look a he Hibor-OIS spread as he indicaor of inerbank sress. This inerbank sress indicaor is derived by subracing he HKD overnigh index swap (OIS) rae of corresponding mauriy from he Hong Kong inerbank offered rae (Hibor) of he same Research Deparmen, Hong Kong Moneary Auhoriy. Noe ha in his paper he words money marke and inerbank marke are inerchangeable. BIS Papers No 54 8

2 mauriy, 3 and is relevan o cenral banks in heir assessmen of inerbank marke disress and applicable policy acions. 4 Char shows he hree-monh Hibor less he hree-monh OIS rae (he Hibor-OIS spread) agains ha of he USD (he Libor-OIS spread) from January 7 o Sepember 9. I can be seen ha, before he onse of he crisis on 9 Augus 7, he Hibor-OIS spread averaged around 8.5 basis poins (bp), compared o 8.4 bp of is USD counerpar. 5 As he financial crisis unfolded, he HKD money marke responded o he developmens in inernaional financial markes wih a sharp increase in he Hibor-OIS spread. Wih an increasing number of inernaional banks reporing subsanial asse wriedowns due o subprime-relaed invesmens and he collapse of Lehman Brohers in Sepember 8, boh HKD and USD inerbank sress series rose sharply from lae Sepember o mid-ocober 8. Global effors o suppor financial sabiliy and confidence in he banking sysem finally eased he sress in he money markes in November December 8. Char Three-monh Hibor-OIS and Libor-OIS spreads Basis poins 4 Basis poins Aug 7 Onse of crisis Three-monh HIBOR - OIS spread Three-monh LIBOR - OIS spread 5 Sep 8 Lehman failure 4 Mar 8 Bear Searns acquired Jan-7 Jul-7 Jan-8 Jul-8 Jan-9 Jul-9 Source: Bloomberg In principle, he inerbank rae of a given enor reflecs he curren and expeced fuure overnigh ineres rae and premia associaed wih liquidiy and credi risks. The OIS rae, meanwhile, is closely relaed o he average overnigh ineres rae expeced o prevail over he erm of he swap. The spread beween he inerbank rae and he OIS rae can be used as an indicaor o gauge he level of sress in he inerbank marke. A widened inerbank rae-ois spread reflecs he increase in liquidiy and/or counerpary defaul risks. Throughou he paper, he inerbank sress indicaor for he money markes under sudy will be referred o as he inerbank sress of he respecive money marke. We follow Taylor and Williams (8) and choose 9 Augus 7 as he onse of he urmoil when BNP Paribas froze redempions for hree of is invesmen funds. 8 BIS Papers No 54

3 One possible ransmission mechanism of he USD shock o he HKD money marke during he crisis may have been he increased use of he HKD money marke o secure USD funding. As noed in Imakubo e al (8), financial insiuions were facing a shorage of USD funding when he USD Libor surged o an unprecedened level. US banks, which encounered increased financing difficulies and had o preserve funds on hand, were relucan o lend heir USD o peers. To secure USD funds, non-us financial insiuions increased heir borrowings in currencies oher han USD and acively convered hem ino USD hrough FX swaps. Such a sraegy could have led o he ighening of condiions in he HKD money marke. 6 The dynamic inerrelaionship beween he USD and HKD inerbank sress series is examined hrough he impulse response funcion of a bivariae vecor error correcion model (VECM). 7 Char shows he responses o changes in HKD inerbank sress o a onesandard-deviaion innovaion in he changes in he USD series as well as in is own series for a period up o 3 days. 8 Char Responses o changes in HKD inerbank sress o one-sandard-deviaion of USD and domesic shocks Changes in bps 8 6 USD shock Domesic shock Source: HKMA esimaes Genberg e al (9) analyse he link beween he urbulence in money markes and he FX swap markes. For deails, see Fung and Yu (9a). An innovaion refers o a shock o he random disurbance erm on a series in he VECM sysem. Through he dynamic srucure of he VECM sysem, a shock o one of he series is also ransmied o he oher series in he sysem. BIS Papers No 54 83

4 Char shows ha he responses o he USD shock are gradual and persisen. The impac from he USD shock dies ou o zero afer 5 days. In relaion o he impac of domesic shocks on is own money marke, while overwhelming on he firs day, he impac rapidly dies ou o zero afer nine days. The resuls from he impulse response funcion show ha he changes in HKD erm funding sress are subjec o pronounced and persisen influences from he USD shock. Hence, i appears ha changes in inerbank sress in he USD money marke fed hrough o he HKD inerbank marke. This resul is no surprising, as Hong Kong SAR is a small and open economy wih a Linked Exchange Rae Sysem (LERS), linking he Hong Kong dollar o he US dollar, and an inernaional financial cenre wih some of he major inernaional financial insiuions paricipaing in he local inerbank marke. In fac, as he funding sress firs unfolded in he USD money marke, exering heavy pressure on many major inernaional financial insiuions, heir abiliy o obain erm funding in he local inerbank marke could also have been affeced. 9 The above finding provides an assessmen of he ransmission of money marke ensions from USD o HKD. If he variaions in hese inerbank sress series are highly synchronised, he linkages may have implicaions for financial sabiliy. To furher examine wheher hese inerbank sresses are synchronised, we model he possible swich in variance regimes by using a univariae regime-swiching ARCH (SWARCH) model wih wo volailiy regimes o idenify periods of unusually high volailiy. The graphs in Char 3 feaure he smoohed probabiliies of he high-volailiy sae in each money marke s inerbank sress along wih he changes in inerbank rae-ois spreads. Char 3 Smoohed probabiliies of being in a high-volailiy sae and changes in inerbank sress The Unied Saes Hong Kong SAR Probabiliy of being in a high-volailiy sae Changes in LIBOR-OIS spread (in bps) Jan-7 Jul-7 Jan-8 Jul-8 Jan-9 Jul-9 Changes in LIBOR-OIS spread (rhs) Probabiliy ha he changes in LIBOR-OIS spread are in a high-volailiy sae (lhs) - Probabiliy of being in a high-volailiy sae Change in HIBOR-OIS spread (in bps). Jan-7 Jul-7 Jan-8 Jul-8 Jan-9 Jul-9 Changes in HIBOR-OIS spread (rhs) Probabiliy ha he changes in HIBOR-OIS spread are in a high-volailiy sae (lhs) Source: HKMA esimaes. 9 This may be due o he reappraisal of counerpary risks and he need o preserve liquidiy by local banks. The smoohed probabiliy provides informaion abou he likelihood ha he changes in spread are in a paricular volailiy sae a ime based on he full sample of observaions. 84 BIS Papers No 54

5 By comparing he paern of he smoohed probabiliies across he USD and HKD money markes, i is possible o examine wheher he high-volailiy sae occurred concurrenly during he financial urmoil from he second half of 7. As shown in Char 3, he daily changes in inerbank sress in he wo money markes experience high volailiy simulaneously on several occasions, such as in Augus 7 and mid-sepember 8, suggesing ha a volailiy linkage or co-movemen exiss beween he money markes. The inerbank sress in he wo money markes had shifed o a low-volailiy sae by he end of January 9. The univariae SWARCH model is exended o a bivariae one in order o examine he issue of volailiy linkages, especially in he high-volailiy sae. Char 4 graphs he smoohed probabiliies when he inerbank sresses of he USD and HKD money markes are boh in a high-volailiy sae. If we focus on he siuaion in Sepember 8, he money marke pairs responded significanly o he collapse of Lehman Brohers and shifed o a high-volailiy sae for over a monh. Therefore, a very significan adverse shock (ie a subsanial increase in he variaions of inerbank sress) in he USD money marke migh have a desabilising impac on he HKD money marke. Char 4 Smoohed probabiliies of boh markes in a high-volailiy sae Probabiliy. Probabiliy Jan-7 Jul-7 Jan-8 Jul-8 Jan-9 Jul-9. Source: HKMA esimaes. Resuls from he bivariae SWARCH esimaion show ha he expeced duraion for he USD and HKD money markes o joinly be in a high-volailiy sae is 5.5 days. Tha means, on average, ha he wo money markes are expeced o say in he high-volailiy sae for 5.5 days before hey shif ino oher saes of volailiy. Hence, he expeced duraion provides useful informaion on he exen of he volailiy linkage beween money markes during a crisis period. Overall, he above analysis shows ha he ransmission of USD money marke sress o he HKD money marke is rapid and persisen. The impac of a USD shock on he changes in BIS Papers No 54 85

6 HKD inerbank sress can las for 5 days before he impac dies ou. There is also an indicaion of volailiy co-movemen beween he inerbank sress of hese wo money markes and he expeced duraion of heir join high-volailiy sae is 5.5 days. In he nex secion, we conduc an even sudy and examine he effeciveness of policy acions aken by he HKMA and he Hong Kong SAR Governmen in miigaing he sress in he HKD inerbank marke. III. Effeciveness of policy acions on HKD inerbank marke sress and he dislocaions in he FX swap marke. Dislocaions in he FX swap marke In Secion II, we examined how disress originaing in USD money markes was ransmied o he HKD inerbank marke. When uncollaeralised USD money markes malfuncioned and he inerbank ineres raes sho up during he urmoil, many non-us financial insiuions relied heavily on FX swap markes o raise US dollars using local currencies. Heighened concerns over liquidiy and counerpary risks raioned hem ou of he USD money markes, and hey all bid for USD in he FX swap marke, creaing a one-sided marke. This one-sided marke induced an FX swap-marke premium, ie a deviaion from he covered ineres pariy (CIP) condiion, as many non-us financial insiuions found hemselves facing similar USD funding shorages. This unusual pricing behaviour refleced dislocaions in he FX swap markes and a similar siuaion was also observed in Hong Kong SAR. The CIP condiion is almos always observed. However, here are imes and siuaions in which he condiion breaks down. One possibiliy is ha, in imes of financial urmoil, he risks as perceived by marke paricipans migh change, rendering he assumpions of he CIP condiion inapplicable. Indeed, Baba and Packer (8) and Genberg e al (9) find ha, in he recen global financial crisis, he urbulence in money markes spilled over o FX swap markes amid a reappraisal of counerpary risks. During he financial crisis, he urbulence in he global inerbank markes spilled over o he local financial markes in which deviaions from he CIP condiion were observed in erms of he difference beween he FX swap-implied USD rae and he USD Libor as shown in he following equaion: CIP deviaion HKD USD r r F,x, x S FX swapimplied USD rae, x USD in erbank rae where S is he HKD per USD spo rae a ime., x is he HKD per USD forward rae HKD conraced a ime for exchange a ime + x. r, x and r USD, x are he corresponding uncollaeralised HKD and USD ineres raes a ime wih a enor of x, proxied by he HKD inerbank rae (Hibor) and he USD inerbank rae (Libor) respecively wih he same enor. In Char 5, he red line measures how much he FX swap-implied hree-monh USD funding rae deviaes from he corresponding USD Libor he risk premium demanded by dollar F () For a deailed discussion, see Fung and Yu (9b). See Taylor (989) for such occasions during he floaion of serling in 97 and he incepion of he European Moneary Sysem in BIS Papers No 54

7 lenders in he swap marke or he deparure from CIP in he local financial markes. As can be seen, before he summer of 7, i oscillaed around basis poins (bp), bu subsequenly sared o follow an upward rend. Around he beginning of Sepember 8, i flucuaed wildly. Char 5 FX swap-implied USD rae, USD Libor and CIP deviaion Ineres rae (% p.a.) Aug 7 Onse of crisis 4 Mar 8 Bear Searns acquired 5 Sep 8 Lehman failure Three-monh FX swap-implied USD rae CIP deviaion (in bps) 8 Deviaion from CIP (rhs) 4-4 Three-monh USD LIBOR -8 Jan-7 Apr-7 Jul-7 Oc-7 Jan-8 Apr-8 Jul-8 Oc-8 Jan-9 Apr-9 - Sources: Bloomberg and HKMA esimaes. In addiion o he dislocaions in he local FX swap marke refleced by he CIP deviaions, he HKD inerbank marke, as referred o above, was also under sress during he crisis. Again, we use he hree-monh Hibor-OIS spread as a measure of sress in he local inerbank marke. Char shows ha he hree-monh Hibor-OIS spread increased when he crisis emerged in Augus 7 and surged o more han bp afer he failure of Lehman Brohers in mid-sepember 8.. Key policy acions and analysis framework In response o he sress in he HKD inerbank marke following he collapse of Lehman Brohers in mid-sepember 8, he HKMA and he Hong Kong SAR Governmen announced a series of measures o help conain he liquidiy and solvency risks in he domesic banking sysem. Table oulines hese policy iniiaives from Sepember 8 o March See HKMA (8) for more deails on he measures ha were implemened in response o he crisis. BIS Papers No 54 87

8 Table Policy measures announced by he HKMA and he Hong Kong SAR Governmen from Sepember 8 o March 9 Announcemen dae Measure 8 9 () 3 Sepember Five emporary liquidiy measures () 8 Ocober Modificaion of he Base Rae formula (3) 4 Ocober Two precauionary measures o suppor confidence in he Hong Kong SAR banking sysem (4) Ocober Addiional supply of hree-monh Exchange Fund Bills (5) 6 November Two refinemens o he emporary liquidiy measures (6) 4 November Addiional supply of hree-monh Exchange Fund Bills (7) 6 March HKMA o coninue he provision of liquidiy assisance o banks Source: HKMA (8) and HKMA press releases. 3. Model specificaion To capure he effecs of policy acions or announcemens as deailed in Table, a policyacion dummy variable is consruced where i is equal o one on days wih policy acions or announcemens and zero on oher days. Given ha he policy acions may influence boh he level and volailiy of he CIP deviaion and hree-monh Hibor-OIS spread, we model he effec of policy acions or announcemens under a sandard exponenial GARCH (EGARCH) model proposed by Nelson (99). The EGARCH model has been widely used in analysing he effecs of policy evens on financial markes, as he model capures he asymmeric effec in he volailiy of financial ime series. The dislocaions in he HKD-USD FX swap marke, measured by he changes in he CIP deviaions, are assumed o be associaed wih he relaive risk of he banking sysems of Hong Kong SAR and he Unied Saes, along wih he policy-acion dummy. 4 A similar framework is applied o analyse he relaionship beween he inerbank sress measured by he Hibor-OIS spread and he effeciveness of he policy acions Effeciveness of he policy acions Resuls from he EGARCH esimaion show ha he policy acions had no maerial impac on he hree-monh Hibor-OIS spread (he indicaor of inerbank marke sress), indicaing ha hey did no ease he disress in he longer end of he HKD inerbank marke following he collapse of Lehman Brohers. 6 This finding is, however, consisen wih he recen sudy by he IMF (9) which shows ha he liquidiy suppor measures iniiaed by he cenral banks The relaive risk of he banking sysem can be considered as he funding liquidiy risk (see Hui e al (9)). Deails of he model specificaion are se ou in Annex I. Shorly afer he measures were implemened, he overnigh Hibor gradually eased. 88 BIS Papers No 54

9 in Japan, Sweden, Swizerland, he Unied Kingdom, he Unied Saes and he European Cenral Bank had an insignifican impac on he Libor-OIS spreads afer he Lehman Brohers failure. The IMF sudy suggess ha he finding does no necessarily mean ha he policy acions o provide liquidiy o he banking sysem were ineffecive, bu ha hose acions may have been anicipaed by marke paricipans. Therefore, heir effecs on he Libor-OIS spreads are no noiceable in he empirical ess. The same marke reacions migh also have happened in Hong Kong SAR, making he effecs of he HKMA and Governmen policy acions on he Hibor-OIS spread no visible in our empirical analysis. In relaion o he dislocaions in he FX swap marke, i is found ha he policy acions have a negaive impac (wih saisical significance) on he CIP deviaions, suggesing ha he policy acions effecively reduced he CIP deviaions during he crisis. This implies ha arbirage opporuniies diminished as he financial markes reurned o normal, o a cerain exen. In oher words, following he failure of Lehman Brohers, he policy acions aken by he HKMA and he Governmen helped o miigae he dislocaions and herefore improved he efficiency in he money and FX swap markes o faciliae arbirage ransacions. In paricular, he five emporary measures provided addiional longer-erm funding o banks agains a wider range of collaeral a a poenially lower ineres cos. Banks were reassured abou he availabiliy of funds and more willing o lend in he inerbank marke. Furhermore, policy acions aimed a conaining he solvency risk in he banking sysem relaxed he counerpary risk consrain on he markes and hus removed he financial dislocaions. 7 IV. Conclusions This paper invesigaes he ransmission of erm funding sress from USD o HKD and examines he effeciveness of he policy acions aken by he HKMA and he Hong Kong SAR Governmen o miigae he anomalies and sress in he FX swap and inerbank markes in Hong Kong SAR during he financial crisis of 8. On he quesion of ransmission of erm funding sress, he resuls show a srong inerdependence beween he variaions in USD and HKD inerbank sress. Following a shock in he USD money marke, he HKD inerbank sress increased immediaely and he impac died ou in around 5 days. This suggess ha for a marke which is as open as Hong Kong SAR and wih such a high degree of paricipaion of foreign insiuions, he US influence can be very profound. Similarly, he analyses find evidence of co-movemens in inerbank sress volailiy in USD and HKD during he crisis. The expeced duraion when he wo money markes are boh in a high-volailiy sae can be as long as 5.5 days. The shor-lived impac of a shock in he USD money marke on he HKD money marke can be aribued o he coordinaed effors by cenral banks and policymakers worldwide o conain he credi crisis. On he effeciveness of he policy acions in miigaing he dislocaions in he FX swap marke and he sress in he HKD inerbank marke, he resuls sugges ha he policy acions had no visible impac on he Hibor-OIS spread, indicaing ha hey did no ease he disress in he HKD inerbank marke. However, his may be due o he fac ha he policy acions migh have been anicipaed by marke paricipans and, herefore, heir effecs migh no have been visible in he HKD inerbank marke. On he oher hand, he policy acions underaken effecively reduced he dislocaions in he FX swap marke afer he failure of Lehman Brohers. The reducion in he CIP deviaions showed ha he policy acions improved he efficiency of he money and FX swap markes where liquidiy resumed o eliminae arbirage opporuniies. 7 The EGARCH esimaion resuls are repored in Annex II. BIS Papers No 54 89

10 Appendix I: EGARCH model specificaions The dislocaions in he FX swap marke beween USD and HKD, measured by he changes in he CIP deviaion (in basis poins (bp)), are assumed o be associaed wih he relaive risk of he banking sysems of Hong Kong SAR and he Unied Saes, along wih he policyacion dummy. The policy-acion dummy is also pu ino he condiional variance equaion o sudy wheher i has any effecs on he degree of volailiy. The condiional mean equaion of he EGARCH model is wrien as: n CIP a bi CIP i crrb d PA CIP,, CIP, i ~ N(, ), (A) where is he firs difference operaor. CIP is he CIP deviaion (in bp) a ime for he respecive crisis period. RRB is he variable for he relaive risk of he banking sysems of Hong Kong SAR and he Unied Saes a ime, which is measured by he difference (in bp) beween he Hibor-OIS spread and he USD Libor-OIS spread during he global credi crisis of 8. PA is he dummy variable for policy acion announcemens a ime. Lags of he dependen variable are included in equaion (A) o conrol for he serial correlaion, if necessary. If he relaive risk of he banking sysems is a deerminan of he premium or discoun as refleced in he swap-implied USD rae, he esimaed coefficien c in equaion (A) should be posiive and saisically significan. If policy acions are helpful in reducing he CIP deviaions, he esimaed coefficien d should be negaive (and i should also be saisically significan if he policy acions have a maerial impac on he CIP deviaions). In he EGARCH (p, q, r) model, he condiional variance equaion is given as: ln( PA (A) q p r CIP, i CIP, k CIP, ) j ln( CIP, j ) i k j i CIP, i k CIP, k The coefficien measures he poenial impac of he policy acions on he degree of volailiy. If he policy acions have he desirable effec of reducing he volailiy of he CIP deviaion, he esimaed coefficien is expeced o have a negaive sign. The coefficien measures he persisen effec in he dynamics of he condiional variance, while he coefficien k capures he asymmeric effec of news. A similar framework is applied o he empirical analysis of he relaionship beween he inerbank marke sress and he effeciveness of policy acions. Fung and Yu (9a) find ha, during he credi crisis of 8, he disress in he USD inerbank marke had a maerial impac on he HKD inerbank marke. Thus, in he condiional mean equaion of he EGARCH model, he inerbank sress indicaor for HKD is assumed o have a linear relaionship wih he sress measure of he USD inerbank marke. The condiional mean equaion is specified as: HK n HK HK US biisi cis dis epa IS, IS, i IS a, CIP, ~ N(, ), (A3) HK US where is he firs difference operaor. IS and IS are he inerbank sress indicaors for HKD and USD (in bp) respecively a ime, which are he Hibor-OIS spread and he USD HK Libor-OIS spread. The lagged erm of he HKD inerbank sress indicaor ( IS ) is included as a conrol variable in case he changes in he spread depend on is level. PA is he dummy variable for policy acion announcemens a ime. Lags of he dependen variable IS, 9 BIS Papers No 54

11 are included in equaion (A3) o conrol for he serial correlaion, if necessary. If he policy acions ease he disress in he HKD inerbank marke, hen i is expeced ha he esimaed coefficien e should be negaive (and should also be saisically differen from zero if he policy acions have a maerial impac on he inerbank sress indicaor). Similarly, he condiional variance equaion in he EGARCH model is given as: ln( PA (A4) q p r IS, i IS, k IS, ) j ln( IS, j ) i k j i IS, i k IS, k The coefficien measures he poenial impac of policy acions on he degree of volailiy. Again, if policy acions have any effec in reducing he volailiy of he Hibor-OIS spread, hen he esimaed coefficien is expeced o have a negaive sign. BIS Papers No 54 9

12 Appendix II: EGARCH model esimaion resuls Table A Esimaion resuls of he CIP deviaion, Hibor-OIS spread and he effeciveness of policy acions for he global credi crisis of 8 Sample period: 6 Sep 8 o 3 Mar 9 Dependen variable: changes in CIP deviaion ( CIP ) Dependen variable: changes in Hibor-OIS spread ( IS ) Esimaed coefficien in he condiional mean equaion n CIP a bicip i crrb d PA CIP, i a. (.84) b.8 (.5) c.* (5.) PA (d).** (.86) n HK HK HK US IS a biis i cis dis epa IS, i a 5.4 (.94) b c.8* (.4) d.3* (.) PA (e) 3.67 (.33) Esimaed coefficien in he condiional variance equaion ln( q p r CIP, i CIP, k CIP, ) j ln( CIP, j ) i k j i CIP, i k CIP, k PA ln( q p r IS, i IS, k IS, ) j ln( IS, j ) i k j i IS, i k IS, k PA.4. (.84) (.).98* (65.3).99* (68.46).7 (.86).9* (.7). (.34). (.5) PA ( ).7 (.) PA (.33 ) (.36) Noes: Figures in parenheses are z-saisics based on Bollerslev-Wooldrige robus sandard errors. * indicaes significance a he 5% confidence level. ** indicaes significance a he % confidence level. Source: HKMA esimaes. 9 BIS Papers No 54

13 References Baba, N and F Packer (8): Inerpreing derivaives from covered ineres pariy during he financial marke urmoil of 7 8, BIS Working Papers, no 67. Fung, L and I-W Yu (9a): A sudy on he ransmission of money marke ensions in EMEAP economies during he credi crisis of 7 9, Hong Kong Moneary Auhoriy Working Paper, 9/9. (9b): Dislocaion in FX swap and money markes in Hong Kong and policy acions during he financial crisis of 8, Hong Kong Moneary Auhoriy Working Paper, 7/9. Genberg, H, C-H Hui, A Wong and T-K Chung (9): The link beween FX swaps and currency srengh during he credi crisis of 7 8, Hong Kong Moneary Auhoriy Research Noe, /9. Imakubo, K, T Kimura and T Nagano (8): Cross-currency ransmission of money marke ensions, Bank of Japan Review, July. Hong Kong Moneary Auhoriy (8): Half-yearly Moneary and Financial Sabiliy Repor, December. IMF (9): Global Financial Sabiliy Repor, Sepember. Nelson, B (99): Condiional heeroskedasiciy in asse reurns: a new approach, Economerica, 59, pp Taylor, M (989): Covered ineres arbirage and marke urbulence, Economic Journal, 99, pp Taylor, J B and J C Williams (8): A black swan in he money marke, NBER Working Paper, no BIS Papers No 54 93