ASSET LIQUIDITY AND BANK PROFITABILITY IN SOUTH AFRICA

Transcription

1 ASSET LIQUIDITY AND BANK PROFITABILITY IN SOUTH AFRICA Godfrey Marozva* Abstract This paper empirically aalyses the relatioship betwee asset liquidity ad bak profitability for South Africa baks for the period betwee 1994 ad The study employs Ordiary Least Squares (OLS) ad the Autoregressive Distributed Lag (ARDL)-boud testig approach to examie the likage betwee retur o assets (ROA) ad liquidity, ad the exus betwee retur o equity (ROE) ad liquidity to capture the short-ru ad log-ru dyamics. The study observes that there is either a sigificat relatioship betwee ROE ad liquidity or a relatioship betwee ROE ad liquidity. These observatios hold for both the short-ru ad log-ru. Baks are recommeded to embrace the asset liability framework i their aalysis ad maagemet of liquidity as the asset oly approach is isufficiet ad misleadig. Keywords: South Africa, Liquidity, Bak Profitability, ARDL-Bouds Testig Approach, OLS. * Departmet of Fiace, Risk Maagemet ad Bakig, P.O.Box 392, UNISA,003, Pretoria, South Africa marozg@uisa.ac.za, godezhy@gmail.com 1. Itroductio Durig the last three decades the fiacial sector has experieced a sigificat chage i its operatioal eviromet. Chages i both the iteral ad exteral factors have affected fiacial istitutios structure ad profitability. Baks are required to hold a cosiderable positio i liquid assets while o the other had it is required to be profitable for it be sustaiable ad to remai as a goig cocer. Despite the icreased efficiecy i may baks resultig from holdig higher positios of liquid assets, profitability has severely suffered. Liquidity ad profitability are iversely related, whe liquidity icreases profitability decreases ad vice versa while o the other had there is a direct relatioship betwee higher risk ad higher retur, hece the dilemma i liquidity maagemet is fidig a balace betwee liquidity ad profitability. While it is geerally agreed that the is a egative relatioship betwee liquidity ad bak profitability the is a couter evidece that shows a eed to cosider the trade off betwee resiliece to liquidity shocks ad cost of holdig less profitable liquid assets as the later is assumed to impact bak s ability to take advatage of opportuities arisig i the market that may result i icrease i reveue, capital or ability to exted capital exted credit (Bordeleau ad Graham 2010). Baks o the asset side hold low yieldig securities such as treasury bills ad highly rated short term corporate bods i order to miimise a scramble for liquidity whe credit use icreases i time whe moey is tight (Holmstrom ad Tirole 1998). Thus i essece a liquid fiacial istitutio has a smaller portio of its assets i log term loas ad a greater proportio of its assets i short term securities that ca be quickly liquidated ito cash that ca the be loaed out, however a highly liquid bak may mea lack of profitable projects to ivest the moey. Give that liquid asset has a low liquidity premium ad therefore a lower retur relative to illiquid assets holdig them imposes a opportuity cost o a bak. Liquidity maagemet becomes a very importat part i fiacial maagemet decisios, where the liquidity maagemet efficiecy could be achieved by firms that maage a trade-off betwee liquidity ad profitability (Bhuia ad Kha 2011). The impact of bak asset liquidity o profitability has of late attracted the iterest of academic research, fiacial market aalysts, bak maagemet ad bak moitors. This paper ivestigates the effects of holdig liquid assets o profitability. The evidece is based o the pael of South Africa baks from 1994 to the ed of The Autoregressive Distributed Lag (ARDL) boud test approach ad Ordiary Least Squares (OLS) testig are utilised i a attempt to fid if the is a log-ru or short-ru relatioship betwee bak profits ad their asset liquidity ad results idicated that there is o sigificat relatioship betwee the variables uder cosideratio. This study differs from other studies i three mai respects. First, the article focuses maily o the ature of relatioship betwee asset liquidity ad South Africa Baks profitability with the aalysis icludig the coitergratio relatioship. Secod, the paper cosider all baks i South Africa as a represetative sample over a more recet period, thus 745

2 providig more appropriate ad recet empirical evidece. Lastly, our empirical aalysis does ot oly focus o the ature of relatioship of variables i questio but also looks at the implicatios of this itercoectedess i the cotext of Asset liability Maagemet (ALM). The paper is orgaised i the followig maer. Sectio 2 discusses facts o South Africa baks liquidity (Liquidity i this cotext refers to liquid assets of the bak that are defied as cash, iterbak deposits ad short term govermet ad corporate securities) ad also a brief discussio o the baks profitability. This is followed by sectio 3 that costitutes a brief discussio of literature ad the empirical framework as applied i this article. Sectio 4 presets the estimatio method ad empirical results. Fially, coclusios ad recommedatios are preseted ad policy implicatios are draw i sectio Asset liquidity ad bak profitability i South Africa As evideced i fig 1, there has bee a expoetial growth i the balace sheet of South Africa Baks. Total assets icreased udeterred util the fiacial crisis i The baks witessed a slump durig the crisis ad after some corrective measures the tred bega to be upward agai. Liquid asset had a steady growth over the etire period, ad this growth did ot correspod with the growth ad volatility i total assets. While the total assets sky rocketed, liquid assets maitaied its growth rate presumably because the baks were ivestig i less liquid asset i a bid to maximise profits i times whe fiacial markets were strog ad calm. Figure 1. Total assets ad liquid assets ZAR(000) Total assets Liquid assets Source: McGregor BFA database Total assets do ot iclude itagible assets like good will, while liquid assets refer to Moey Market Ivestmet Assets (Moey Market Ivestmet Assets represets the short-term ivestmet i fiacial assets by the bak as part of its bakig operatios) ad Liquid Ivestmet Assets (Liquid Ivestmet Assets represets the cash o had ad the balaces with other baks as per the otes to the aual fiacial statemets i respect of the bakig operatios of the Compay or Group). Fig 2 represets the percetage of liquid assets as a percetage of total assets. The graph shows that the ratio has bee icreasig over time but with a very high degree of volatility. 746

3 Figure 2. Liquid assets as a percetage of total assets Source: McGregor BFA data base. Sice 2005 util the begiig of fiacial crisis i 2008, baks have bee reducig their holdig of liquid assets relative to total assets. I reactio to fudig ad liquidity pressures durig global fiacial crisis baks i South Africa bega to hold cosiderable liquid assets relative to total assets. It is durig the crisis whe the stregth of a bak had to be measured i terms of how liquid the fiacial istitutio was rather tha o the basis of the size of the balace sheet or profitability. Previously baks took advatage of mismatches betwee assets ad liabilities; baks massively leveraged o these mismatches ad were their key compoet of extremely profitable busiess model a pheomeo that chaged after the recet fiacial crisis (Barua et all 2010). Durig the fiacial crisis, eed for liquidity became fudametally iheret to the fiacial sector. I fact, oe of the key fuctios of the bakig idustry i a moder ecoomic system is to allow the reallocatio of fiacial resources from the liquid sectors to the illiquid oes. The Basel lll framework, released i December 2010 also calls for sigificat chages i liquidity requiremets. The framework itroduces more striget liquidity requiremets which are expected to be phased i over a umber of years. Regardless of the fact that these chages has to be effected i over time may baks deems it prudet to maitai eve higher tha recommeded liquidity levels i the iterim. The bak profitability variables are better represeted by the retur o equity (ROE) ad retur o assets (ROA). ROE shows the retur that shareholders will get from their ivestmet i the equity of the bak. ROA idicates ability of a bak to geerate profits from utilisig its assets, thus it measures the efficiecy of maagemet i usig bak assets to geerate earigs. Figure 3. Retur o Equity Source: McGregor BFA data base. 747

4 A lot of developmets o the global area ad South Africa s local area i the past two decades resulted i a highly volatile ROE for South Africa baks. There was a steep icrease i ROE from 2002 to 2005 that was followed by a steep fall i 2006 util The fall i ROE could have bee a result of the coutry havig goe through a ecoomic recessio coupled with a spillover effect of the global ecoomic crisis i 2008/9 resulted i the people losig cofidece i fiacial istitutios. Sice the busiess eviromet i geeral was t coducive this meat tradig was theme o all spheres there by affectig Figure 4. Retur o assets cosumer affordability, resultig i cosumers beig more reluctat to take o more debt, thereby egatively affectig the profitability of the bakig sector. The fall i ROE was ot disastrous because most of the fiacial istitutios had very strog risk maagemet systems i place as the South Africa Baks were amogst the first to implemet Basel II recommedatios. Though stiff regulatios o fiacial istitutios is blamed for stiflig iovatio ad reducig growth 2008/9 global fiacial crisis meat otherwise Source: McGregor BFA data base. The retur o assets also has ot bee stable over the past decade. ROA has the same tred as ROE the oly major differece is that ROA is less tha ROE due to the mathematics behid their calculatios due to the differeces of the compositio of the deomiators, where ROA has a larger deomiator tha ROE. The steep rise i ROA durig the period of 2002 to 2006 was ehaced by the expasio of cosumer credit, the South Africa ecoomy experieced sigificat growth durig this period. By the ed of 2006 the ecoomy bega to slow dow as the effects of global depressio kicked i ad we saw the reversal of the gais from previous years. Figure 5. Retur o Equity ad Retur o Assets Retur o Equity Retur o Assets Source: McGregor BFA data base. 748

5 Fig 5 shows that South Africa baks some how relied heavily o liabilities to support assets ad cosequetly they have higher ROEs ad lower ROAs. This shows that the South Africa baks equity base is too small relative to the total balace sheet as sigified by total assets ad thus ca egatively impacts the baks ability to borrow ad ca eve be very disastrous if there is a ru o deposits especially i a eviromet associated with dwidlig cofidece ad a sudde icrease i iterest rates. 3. Literature review Liquidity though ot a ew pheomeo i fiace literature, there is o uiversally accepted defiitio of it. Adler (2012) asserts that the lack of agreed-upo defiitio emaates from the fact that the cocept of liquidity arises from differet ecoomic perspectives. Liquidity ca be defied i the cotext of how easy a security ca be traded ad i the cotext of how easy oe ca obtai fudig to trade a security, the former beig called market liquidity ad the latter beig fudig liquidity. The focus of this paper will be o both the fudig liquidity ad market liquidity sice the easier you ca trade a security meas the easier it is to get fuds to trade securities, ideally market liquidity ad fudig liquidity are complemetary. Most papers were writte o the sources of liquidity risk ad o how markets should be desiged ad regulated to cope with the effects of illiquidity. This literature review will attempt to summarise the impact of liquidity o bak profitability, hece eed to look at liquidity as a cost, ad as a risk ad their impact o ROA ad ROE. That is, ivestors eed to be rewarded for holdig illiquid assets ad for the sesitivity of the security to liquidity shocks. There are a very limited umber of studies that were specifically carried out to ivestigate the impact of liquidity o bak profitability. Surprisig most of these few studies were doe o maufacturig compaies. Therefore, most of the studies we draw the followig coclusios were maily focused i fidig determiats of bak profitability with liquidity beig oe of the determiats ad their empirical results were mixed. Some writers foud a positive relatioship; some foud a egative relatioship while others foud both results ad a few foud o relatioship at all. The debate is still rampat. Bourke (1989) i his study o performace of baks i twelve coutries i Europe, North America ad Australia foud evidece that there is a positive relatioship betwee liquid assets ad bak profitability. The results which seem couterituitive, as we expect that illiquid assets have higher liquidity premium ad hece higher retur tha liquid assets. Kosmidou ad Pasiouras (2005) realised that the ratio of liquid assets to customer ad short term fudig is positively related to ROA ad statistically sigificat. Also, they foud a sigificat positive relatioship betwee liquidity ad bak profits. Kosmidou (2008) examie the determiats of performace of Greek baks durig the period of EU fiacial itegratio ( ) usig a ubalaced pooled time series dataset of 23 baks foud that less liquid baks have lower ROA. This is cosistet with their previous fidigs like Bourke (1989) who foud out that there is a positive relatioship betwee liquidity risk ad bak profitability. Recetly, Olaguju, David ad Samuel (2011) foud out that there is a positive sigificat relatioship betwee liquidity ad profitability. They cocluded that there is a bidirectioal relatioship betwee liquidity ad profitability where the profitability i commercial baks is sigificatly iflueced by liquidity ad viceversa. O the cotrary, Molyeux ad Thorto (1992) recogized that there is iverse relatioship betwee bak profitability ad liquidity ad they attributed this to the fact that baks hold liquid assets as a obligatio to the requiremets imposed by the authorities. However, if we are to view this relatioship from the cotext that baks hold liquid assets as madated by the cetral bak or ay other authorities, the we may miss our argumet as baks also hold liquid assets for other reasos. Oe to assume that baks oly hold liquid assets as a requiremet is i its self-perfidious or a deliberate igorace of the kowledge of how baks fuctios, as Tobi (1958) way back suggested that liquidity is held for trasactio purposes ad for ivestmets reasos. Tobi s proposal was a simplificatio of Keyes liquidity preferece theory. Keyes (1936) argued moey is demaded for trasactio purpose, speculative purpose ad precautio purpose therefore we ca firmly say without ad prejudice say that liquid assets over ad above madatory requiremets they are held for trasactio, speculative ad precautioary purpose. Some authors foud mixed results of both egative ad positive relatioship. She et al (2009) asserts that i market-based fiacial system liquidity risk is positively related to et iterest margi a idicatio that baks with high levels of illiquid assets receives higher iterests icome. Coflictig to their earlier establishmet o the relatioship with et iterest margi, they realised that liquidity risk is egatively related to retur o average assets ad also iversely related to retur o average equity. They poited out that baks icurred higher fudig cost i the market if they have illiquid assets as they had to raise the moey i the market to meet the fudig gap. They also discovered that there is o relatioship betwee liquidity risk ad performace i a bakbased fiacial system as the baks play a major role i fiacig, therefore are ot affected by liquidity risk. Demirguc-Kut ad Huiziga (1999) had icoclusive results; they foud a positive relatioship betwee loas to total assets ad the et 749

6 iterest margis. The also established a iverse relatioship betwee the et iterest margi ad before tax profits. Naceur ad Kadil (2009) i their aalysis cost of itermediatio i the post-capital regulatio period which icluded; higher capital-toassets ratios, a icrease i maagemet efficiecy, a improvemet of liquidity ad a reductio i iflatio foud out that Baks liquidity does ot determie returs o assets or equity sigificatly. Therefore coclusios about the impact of baks liquidity o their profitability remai ambiguous ad further research is required. 4. Data, empirical model specificatio ad estimatio techiques. 4.1 Data sources ad defiitio of variables This study uses aual time series data for the period betwee 1994 ad 2011 all data used i this study were obtaied from McGregor data base ad the cetral bak of South Africa (SARB). Liquid assets were computed as a total of Moey Market Ivestmet Assets ad Liquid Ivestmet Assets. The ROE o the other had was measured by the ratio of et icome to total equity while the ROA measured as a ratio of et icome to total assets. Iitially the regressio model was ru to see the short ru relatioship of the profitability as measured by ROE/ROA ad idepedet variable of Liquidity. The, i a attempt to establish a log-ru coitergratio relatioship betwee liquidity ad profitability the auto regressive distributed lag (ARDL) - bouds testig approach by Perasa et al. (2001) model was adapted. 4.2 Regressio Model The regressio model was ru to ivestigate the short ru relatioship of the profitability as measured by ROE/ROA ad idepedet variable of Liquidity. We ca express the relatioship betwee liquidity ad profitability mathematically as follows: R (1) (2) R (3) (4) Equatios (1) ad (3) shows retur o asset ad retur o equity as fuctios of liquidity. The regressio models are idicated by the equatio (2) ad equatio (3). Profitability i this case is represeted by ROA ad ROE which are the depedet variables, Liquidity (L) is the explaatory/idepedet variable ad α ad β are coefficiets. There is a plethora of empirical evidece (Bourke (1989), Kosmidou et al.(2005) Kosmidou (2008), Olaguju et al. (2011), ad Molyeux et al. (1992)) that attest to the fact that there is statistically sigificat relatioship betwee liquid assets ad bak profitability. This evidece is coflictig ad therefore, it is icumbet upo the researcher to cotribute to the body of kowledge by further determiig the ature of relatioship betwee the variables i questio. The first step was to fid whether there is a determiistic or short ru relatioship betwee profitability ad liquidity. The results of a simple LOS are as follows. Depedet variable Fuctio P-Value F-Test statistic ROA ROA(Liquidity) ROE ROE(Liquidity) *** Deotes 1% level of sigificace, ** Deotes 5% level of sigificace, ad *Deotes 10% level of sigificace The results show above reveal that there is o sigificat determiistic or short-ru relatioship betwee the profitability ratios (ROA & ROE) ad liquidity (TL/TA). The F-Stats are well below the recommeded figure of 4 ad also the p-values are way above the threshold of Uit root tests The data sets of three variables (liquid assets, ROE ad ROA) were tested for statioarity usig Phillip- Perro ad Augmeted Dickey Fuller tests before they were tested for coitergratio usig ARDLbouds approach. The results of the statioarity tests o differeced variables are preseted i table

7 Table 1. Statioarity tests of variables o first differece Augmeted Dickey Fuller (ADF) test ad Phillips- Perro (PP) test Variable No tred Tred Itercept Statioary tests of variables o fist differece Augmeted Dickey Fuller (ADF) test DLiquidity (TL/TA) *** *** *** DROA *** *** *** DROE *** *** *** Statioary tests of variables o fist differece Phillips Perro (PP) test DLiquidity (TL/TA) *** *** *** DROA *** *** *** DROE *** *** *** *** Deotes 1% level of sigificace Give the result i the table above the hypothesis that first differece of ROA, ROE, ad Liquidity has uit roots ca be rejected. 4.4 Coitergratio test- ARDL bouds testig procedure To establish a log-ru coitergratio relatioship betwee liquidity ad profitability the auto regressive distributed lag (ARDL) - bouds testig approach by Perasa et al. (2001) model was adapted. The ARDL approach is uique ad superior i that it does ot require all the variables uder ivestigatio to be itegrated at the same order. Thus, the ARDL approach ca be used i situatio eve if the regressors are itegrated i ay order that is order oe (I (1)), order zero (I (0)) or partially itegrated (Pesara ad Pesara, 1997). Laureceso (2003) argue that usig the ARDL approach avoids problems resultig from o-statioary time series data. The ARDL framework for equatio 5, 6, 7 ad 8 are as follows: ROA Liquidity ROE Liquidity t t t t 0 1i ROA 2i Liquidity 3ROAt 1 4Liquidity t 1 t (5) 0 1i Liquidity 2i ROAt i 3Liquidity t 1 4ROAt 1 t (6) 0 1 i ROEt i 2i Liquidity 3ROEt 1 4Liquidity t 1 t (7) 0 1 i Liquidity 2i ROEt i 3Liquidity t 1 4ROEt 1 t (8) Where - first differece operator, ROA Retur o Assets, ROE- Retur o Equity ad Liquidity Liquid Assets divided by Total Assets (TL/TA). I the above equatios, the terms with the summatio sigs represet the error correctio dyamics while the secod part (terms with γ i equatio (5), ϕ i equatio (6), β i equatio (7), ad with ρ i equatio (8)) correspod to the log ru relatioship. The ull hypotheses i 5, 6, 7 ad 8 are, 0, 0, ad 0, respectively, which idicate the o-existece of the log ru relatioship. The first step of the ARDLbouds testig requires examiig the order of lags o the first differeced variables i equatio 5, 6, 7, ad 8 usig the Akaike iformatio criterio (AIC) ad the Schwartz-Bayesia criterio (SBC). The results of the AIC ad the SBC suggest that optimal lag of ROA ad liquidity is 3, while the optimal lag for ROE ad liquidity is 3. The secod step requires us to apply the bouds F-test to equatio 5, 6, 7, ad 8 i order to determie whether ay log ru relatioship betwee South Africa Bak profitability ad liquidity. 751

8 Depedet variable Fuctio F-Test statistic ROA ROA(Liquidity) Liquidity Liquidity(ROA) ROE ROE(Liquidity) Liquidity Liquidity(ROE) ** ** Deotes 5% level of sigificace Our results show that there is o evidece of log ru relatioship betwee profitability ad liquidity. All other thigs beig equal ROA ad ROE are ot i iflueced by liquidity i the log ru. To determie whether liquidity is drive by ROE i the log ru we used Table CI (v) o p.301 of Pesara et al. (2001) to determie the asymptotic critical value bouds for the F-statistic sice the models had ucostraied itercept ad urestricted tred. The lower ad upper bouds for the F-test statistic at the 10%, 5%, ad 1% sigificace levels are [5.59, 6.26], [6.56, 7.30], ad [8.74, 9.63] respectively. As the value of our F-statistic is blow the lower boud at the 5% sigificace level, we ca also coclude that there is o evidece of a log-ru relatioship betwee the two time-series at this level of sigificace or greater. 5 Summary ad Coclusio I cotext of Basel lll framework, released i December 2010 that calls for sigificat chages i liquidity requiremets as a mitigatig dyamic to the damage that resulted from 2009/10 fiacial crisis this paper attempted to determie empirically the relatioship betwee bak profitability ad liquidity. This was doe with the aim to establish the impact of chages i liquidity o the performace of baks. The results reported i this paper are cosistet with the view that there is o sigificat relatioship betwee profitability ad liquidity. This could be a idicatio that performace of baks ca be attributed to other macro-ecoomic factors ad other firm specific characteristics besides the compositio of its assets. Our fidigs are cosistet with the fidigs of Naceur et al. (2009) who foud out that Baks liquidity does ot determie returs o assets or equity sigificatly. These results came as a surprise especially for the bakig system whose busiess ad competitive edge is cetred o the size ad compositio of their balace sheet. By ature of their busiess baks act as itermediaries betwee deficit uits ad surplus uits, where the take mostly short term (highly liquid) moey from surplus uits ad pass it o to the deficit uit at a price makig a positive iterest rate spread, cosequetly makig the liquidity of bak assets a focal poit of ivestigatio. The liquidity issue outside capital adequacy is the most importat fudametal assumed to be directly attributed to the performace of baks i the recet past, particularly durig the global fiacial crises. Fially, liquidity seems to be quite persistet all over the world, which probably sigals eed for ew, efficiet ad effective maagemet of assets ad liabilities. A aalysis of assets oly without referece to liabilities could have bee the major drawback of this study hece for future studies emphasis should be withi a framework of asset liability maagemet (ALM). Refereces 1. ADLER, D., The New Field of Liquidity ad Fiacial Frictios. Research Foudatio Literature Reviews, 7(2), pp Barua R.; Battaglia F.; Jagaatha R.; Medis J.; Oorato M. (2010), Basel III: What s New? Busiess ad Techological Challeges 3. BCBS (2010): Basel III: Iteratioal framework for liquidity risk measuremet, stadards ad moitorig, Bak for Iteratioal Settlemets. December BEN NACEUR, S. ad KANDIL, M., The impact of capital requiremets o baks cost of itermediatio ad performace: The case of Egypt. Joural of ecoomics ad busiess, 61(1), pp BHUNIA, A. ad KHAN, I.U., Liquidity maagemet efficiecy of Idia Steel Compaies (a Case Study). Far East Joural of Psychology ad Busiess, 3(1), pp BORDELEAU, É. ad GRAHAM, C., The impact of liquidity o bak profitability,. 7. BOURKE, P., Cocetratio ad other determiats of bak profitability i Europe, North America ad Australia. Joural of Bakig & Fiace, 13(1), pp DEMIRGÜÇ-KUNT, A. ad HUIZINGA, H., Determiats of commercial bak iterest margis ad profitability: Some iteratioal evidece. The World Bak Ecoomic Review, 13(2), pp HOLMSTRÖM, B. ad TIROLE, J., Private ad Public Supply of Liquidity. Joural of Political Ecoomy, 106(1), pp KEYNES, J.M., The geeral theory of iterest, employmet ad moey. 11. KOSMIDOU, K., The determiats of baks' profits i Greece durig the period of EU fiacial itegratio. Maagerial Fiace, 34(3), pp KOSMIDOU, K., TANNA, S. ad PASIOURAS, F., Determiats of profitability of domestic UK commercial baks: pael evidece from the period , Moey Macro ad Fiace (MMF) Research Group Coferece LAURENCEASON, J. ad H CHAI, J.C., Fiacial reform ad ecoomic developmet i Chia. Edward Elgar Publishig. 14. MOLYNEUX, P. ad THORNTON, J., Determiats of Europea bak profitability: A ote. Joural of Bakig & Fiace, 16(6), pp

9 15. OLAGUNJU, A., DAVID, A.O. ad SAMUEL, O.O., Liquidity Maagemet ad Commercial Baks' Profitability i Nigeria. Research Joural of Fiace ad Accoutig, 2(7-8), pp PESARAN, M.H. ad PESARAN, B., Workig with Microfit 4.0: iteractive ecoometric aalysis. Oxford Uiversity Press. 17. PESARAN, M.H., SHIN, Y. ad SMITH, R.J., Bouds testig approaches to the aalysis of level relatioships. Joural of Applied Ecoometrics, 16(3), pp PESARAN, M.H., SHIN, Y. ad SMITH, R.J., Bouds testig approaches to the aalysis of level relatioships. Joural of Applied Ecoometrics, 16(3), pp SHEN, C., CHEN, Y., KAO, L. ad YEH, C., Bak liquidity risk ad performace. Iteratioal Moetary Fud, Workig Paper,. 20. TOBIN, J., Liquidity preferece as behavior towards risk. The Review of Ecoomic Studies, 25(2), pp