Value-at-Risk Disclosures of Banks: The Case of Israel

Transcription

1 Value-a-Risk Disclosures of Banks: The Case of Israel Shilo Lifschuz, Roi Polanizer 2. Inroducion Banks in Israel are mandaed o manage marke risks and calculae heir capial adequacy hrough a Value a-risk (VaR) sysem (Supervisor of he Banks, Direcive 339 and Direcive 34). A deailed disclosure on VaR can be found in he banks' financial saemens. VaR is he maximum expeced loss on a given financial posiion over a given ime horizon and a a defined saisical confidence level. Regulaors in Israel proliferaed he implemenaion of VaR o oher indusries as well. As of he end of 2007, under he Israeli Securiies Regulaions, companies whose main business is in he area of finance mus repor according o he VaR model. I should be noed ha oher han hese companies, since 2008, providen funds and insurance companies mus also calculae VaR. In his paper, we examine he relaionship beween he VaR measuremens repored by he five larges banks in Israel and he subsequen variabiliy of heir rading revenues. The research quesion is: Does he VaR repored by banks over ime predic he variabiliy of he rading revenues in he Israeli banking sysem? A high VaR value (in given parameers of confidence inerval and ime horizon) indicaes a larger sandard deviaion of he changes in he prices of he risk facors. Therefore, if he measuremen has informaive value, we would expec his o also be Dr. Shilo Lifschuz, CPA, is Chairman, Deparmen of Accouning, Academic Cener of Law and Business, and an adjunc lecurer in he Faculy of Law of he Hebrew Universiy of Jerusalem. shilo@clb.ac.il 2 Mr. Roi Polanizer, MBA, is Risk Manager, Founder and Head of he Fair Value Deparmen of he consuling firm of Ogen Ld. - Acuarial, Financial and Business Consuling. roi@ogen.co.il

2 refleced in he changes in fuure rading revenues, which he VaR aemps o capure. In oher words, we expec o find a posiive correlaion beween he VaR daa of he banks and he subsequen variabiliy of heir rading revenues. Barh (2007) argues ha accouning researchers can help he accouning sandards seers boh a he heoreical and pracical level. Research is exac, unbiased and based on economic heory; i can herefore assis in measuremen and disclosure decisions, and can indicae he exen o which hese mehods provide useful informaion o users of he financial saemens. In a previous aricle wrien by he auhor (Barh, 2006), she relaes o his when discussing IFRS 7: "Informaion abou he variance of esimaes of he fuure can be imporan o users in assessing he riskiness of expeced benefis" (page 283). We believe ha examining his issue is essenial for several reasons: Israel is characerized by: () high concenraion of conrol by conrolling shareholders in publicly raded companies, (2) srong cenralizaion of he banking sysem, (3) he 5 larges banks in Israel hold a 94% marke segmen. These characerisics may impac he way banks are managed and he level of informaion derived from he financial saemens. Furhermore, over he pas decade, Israel has become a popular arge for foreign invesmens and is widely covered by analyss. Consequenly, i is imporan o examine informaiveness in he Israeli conex. As previously menioned, use of he VaR measuremen in Israel has increased over he years, and i is implemened in banks, insurance companies, providen funds and oher publicly raded companies whose primary business is in he financial area. IFRS 7 (IASB, 2005) has mandaed publicly raded companies in 2

3 Israel o repor on sensiiviy analysis (or alernaively on VaR) since I is imporan for he regulaory auhoriies and accouning sandards insiuions o undersand wheher he new informaion is informaive and wheher i can forecas he variance of profis in he fuure. The VaR included in he financial saemens of banks may be used by invesors and analyss - when analyzing he banks' risk-reurn profile. This also explains why i is imporan o examine he informaiveness of he VaR informaion. The srucure of he sudy is as follows: In Secion 2, we review he relevan lieraure: we describe he regulaion wih respec o he issue of VaR a he banks and also review aricles ha have examined he relevance of risk managemen in general, wih specific aenion o VaR daa. In Secion 3, we presen he research hypohesis, mehodology and sample. In Secion 4, we describe he daa and descripive saisics. In Secion 5, we presen he empirical resuls, and in he las secion we conclude and provide limiaions as well as recommendaions for furher research. 2. Review of Relevan Lieraure 2. Definiion of VaR VaR is he maximum expeced loss on a given financial posiion over a given ime horizon and a a defined saisical confidence level. For insance, a bank disclosing a daily VaR of 0 million dollars a 99% level means ha here is only % chance ha he bank will incur a loss of more han 0 million dollars over he nex day. Three mehods are used o calculae VaR: The normal disribuion, hisorical simulaion and Mone Carlo simulaion. The VaR model has become a sandard ool 3

4 used by banks o conrol risk (Dowd, 2000) and as a conrol and reporing ool used by managers, raders and direcors. VaR is repored o he regulaory agencies and invesors in he financial saemens (Rogachev, 2007). When VaR is compared o he oher mehod of risk conrol - sensiiviy analysis, we find ha he sensiiviy analysis mehod is simpler and more ransparen han VaR. However, sensiiviy analysis has is shorcomings. I akes each risk ino accoun separaely neiher considering he correlaion beween he risk facors, nor offseing risks aribuable o he diversificaion and exisence of negaive correlaions beween he risk facors. Addiionally, sensiiviy analysis uses a fixed change, 5% for example, and does no consider pas scenarios. The VaR model akes ino accoun he pas sandard deviaion and correlaion beween risk facors. I can be improved by back-esing, which examines he gap beween he loss forecas under he VaR model and he acual loss. Addiionally, a sress-es can be applied o capure he risk of he model wih exreme condiions and heavy-ailed disribuions. However, implemening VaR sysems is cosly in erms of sofware, inerfaces, recruimen of skilled employees, ec. 2.2 VaR-relaed Regulaion in Israel In 997, Direcive 339 o he Proper Conduc of Banking Business Regulaions (risk managemen) was published. I relaes o implemenaion of he VaR model. Laer, Direcive 34 o he Proper Conduc of Banking Business Regulaions, (capial allocaion for marke risk exposure) relaing o calculaion of capial adequacy in respec of he exposure o marke risks, was issued. 4

5 According o Direcive 339, all banks (excluding morgage banks) which ake posiions in foreign currency and ineres, inves in heir own accoun in securiies (nosro) or are marke makers in derivaive financial insrumens, mus manage he marke risks by means of a VaR sysem. The Direcive emphasizes he following requiremens: An inernal model for measuring marke risks, based on saisical echniques such as variancecovariance (normal disribuion), hisorical simulaions and Mone Carlo simulaions Ongoing measuremen of he bank's exposure o marke risks by esimaes of VaR based on he corporaion's inernal model Sress-es analyses of marke risks A model for assessing he profiabiliy of posiions A separae risk-conrol uni Direcive 339 marked a revision a he banks in maers relaed o organizaional and IT preparedness from he various aspecs of risk managemen. The Direcive defined new funcions in he organizaions (including a risk manager and a risk-conrol uni) and increased he involvemen of managemen and he board of direcors in all maers relaed o risk managemen and conrol. The Direcive also served o be an inspiraion for regulaors in oher fields, as i moivaed hem o ighen regulaion in erms of financial risks. We are now seeing a process in which banking regulaions rickle down o oher financial companies and from hem o publicly raded companies, hus srenghening corporae governance guidelines. Direcive 34 defines, iner alia qualiaive and quaniaive sandards for approving an inernal model used o calculae capial adequacy of marke risks. The qualiaive 5

6 sandards sress he following parameers: ) Involvemen of he risk-conrol uni and responsibiliy of he managemen and board of direcors, 2) independence of he risk conrol officer, 3) exisence of an inernal model ha is par of risk managemen, 4) performance of monioring and full conrol of he daabase and he possible changes o he models. The quaniaive sandards emphasize he various parameers and paricularly he following: ) VaR mus be compued on a daily basis wih a confidence inerval of 99% and for a ime horizon of 0 days, 2) he daabase is o be updaed every hree monhs, 3) recogniion of correlaions, 4) special esimaion of he risk in opions, 5) performance of sress ess on he parameers in he model. 2.3 Relevance of Informaion on Risk Managemen Many sudies find ha he sock reurns of financial insiuions are sensiive o ineres rae changes (Lloyd and Shick, 977; Chance and Lane, 980; Flannery and James, 984; Booh and Officer, 985; Sco and Peerson, 986; Kane and Unal, 988; Kwan, 99; Choi e al., 992). Schrand (997) examines wheher he repors on derivaives provide informaion ha is useful in esimaing he exposure o ineres among savings and loan insiuions. The researcher sudies his by measuring he sensiiviy of sock price relaive o flucuaions in he ineres rae. The findings of he sudy show ha aciviy in derivaives reduces exposure o ineres rae flucuaions. In addiion, Harris e al. (99) find ha sock reurns of commercial banks were sensiive o exchange rae movemens. Oher sudies repor ha he use of derivaives reduced he sensiiviy of he equiy reurns of financial insiuions o ineres raes (Choi and Elyasiani, 997; Chamberlain e al., 997; Carer and Sinkey, 997; Hirle, 997; Brewer e al., 200). Similarly, Rajgopal 6

7 (999) sudies wheher informaion on derivaives presened in a able or in sensiiviy analysis is useful when assessing exposure o changes in price. The researcher finds ha share prices of companies ha used derivaives as a hedge were less impaced by changes in he prices of gas and oil. Linsley and Shrives (2000, 2005) sugges ha he forward-looking risk informaion (and VaR is forward looking), would be especially useful o invesors. Hirle (2003) shows ha US banks quarerly marke risk daa conains valuable informaion abou fuure risk exposures. The resuls of oher researchers also demonsrae he usefulness of he marke risk and VaR informaion (see also Hirle, 2007; Bali e al., 2007; Taylor, 2005; Alexander and Sheedy, 2008). Jorion (2002) sudies he relaionship beween he VaR measuremens and he subsequen variabiliy of he rading revenue in a group of eigh large commercial banks in he US over he course of six years. Jorion finds a posiive and significan correlaion beween he VaR-based volailiy and fuure marke risk, concluding ha he VaR measuremens published by he banks effecively predic he bank's marke risks from rading aciviies. Therefore, hey can be used by analyss o compare risk profiles of differen banks. Liu e al. (2004) examine he correlaion beween VaR disclosures and he rading margins among 7 commercial banks beween 997 and The researchers find ha banks' rading VaRs have predicive power for rading income variabiliy ha increases wih bank echnical sophisicaion and over ime. They find ha rading VaRs have predicive power for differen measures of risk. 7

8 3. Research hypohesis, mehodology and sample 3. Research hypohesis In ligh of he lieraure review in he previous secion, we would expec ha he VaR daa published by he banks would have informaive value. The higher he VaR value (in given parameers of confidence inerval and ime horizon) he larger sandard deviaion of he changes in he prices of he risk facors ha consiue he financial posiion. Therefore, if he VaR measuremen has informaive value, we would expec i o be refleced in he changes of he fuure rading revenues, which he VaR aemps o esimae. We es his wih he following hypohesis: H : The VaR measuremen of a bank is posiively relaed o he subsequen variabiliy of is rading revenues. 3.2 Mehodology To examine he research hypohesis, we ill use he mehodology proposed by Jorion (2002). Firs, le us define: VaR VaR a he end of a day R rading revenue during day + + ( C) minus he confidence level (e.g., % when C = 99%) P [ R E( R ) < ] = % + + -VaR C = () 8

9 3.2.. Transforming VaR ino a Risk Measuremen: We assume ha he changes in he rading revenues have a normal disribuion, paricularly when he average of he changes is zero (a reasonable assumpion for large commercial banks whose rading porfolios include a wide variey of financial insrumens ha are exposed o risk facors). A a confidence level of 99%, in oher words α = 2.33, he equaion is wrien as follows: VaR = α S = S (2) % 99% 2.33 σ VaR E ( R + ) = 0 The expeced absolue value of he rading revenue is a linear funcion of he sandard deviaion, and he sandard deviaion is a proporionae o he VaR, as shown in Equaion (3): E 2 π ( R + ) = S = 0.80 S (3) 9

10 From his we can see ha a linear link exiss beween VaR and he absolue value of he rading revenue and herefore he predicive power of he VaR, can be examined using he following equaion: α β σ ε (4) R + = Where: σ = N (5) S N = 63 rade days in a quarer One of he assumpions is ha R + is disribued normally wih a mean of 0. However, because i is no acually so, Jorion (2002) measures he unexpeced rading revenues as he difference beween he quarerly rading revenue and is moving average over he previous four quarers. R + - E [ R ] = R - R (6) + -i Thus, Equaion (4) was implemened as follows [ R ] R + - E + = α + β σ + ε + (7) In his paper, we esimae Equaion (7) hrough he OLS regression model. If β is posiive and significan, we will conclude ha VaR is informaive for he risk of fuure rading revenues 0

11 Calculaing he Sandard Deviaion of Trading Revenues (σ ): The following are he sages in calculaion of he quarerly sandard deviaion (σ ): Sep : Transformaion, in order o sandardize he repored VaR o daily by dividing he repored VaR by n, as described in Equaion (8): Where: VaR(n days,99%) VaR(daily, 99%) = ( 8) n n - ime horizon according o which he VaR was calculaed. Sep 2: Placemen of he daily VaR in Equaion (8) o compue he daily sandard deviaion of revenues implied in he daily VaR (in NIS). VaR =α S (9) where: α - he number of sandard deviaions for a confidence level of 99%, which is Sep 3: Subsiuing S from Equaion (9) ino Equaion (0) o compue he quarerly sandard deviaion. Where: N = 63 rade days in a quarer = S N σ (0)

12 Calculaion of he Unexpeced Trading Revenues: The unexpeced rading revenues will be calculaed using Equaion (6): R + - E[ R + ] = R i Where: R will be calculaed as he oal of he following quarerly iems aken from he income saemens of he five banks: Shares: quarerly profi (loss) from invesmen in shares, ne. Derivaives: quarerly profi from derivaive insrumens, ineffecive par of ALM and ohers. Bonds: quarerly cumulaive revenues from rading bonds and unrealized profi and loss on rading bonds, ne. In his sudy we examine he relaion beween he VaR-based quarerly volailiy and he absolue value of he unexpeced rading revenue in he subsequen quarer using Ordinary Leas Squares (OLS) regression analysis. Equaion (7) is esimaed for he period December 3, 2003 o December 3, 2007 and for each of he five banks separaely (7 quarerly observaions for each bank). The regression model is: [ R ] R + - E + = α + β σ + ε + R 3.3 The sample The sample will include he VaR daa repored by he five larges banks in Israel: Bank Hapoalim Ld. Bank Leumi Le-Israel Ld. 2

13 Israel Discoun Bank Ld. Bank Mizrahi Tefaho Ld. Firs Inernaional Bank Ld. Oher banks are no included in he sample, as complee daa was no available. According o he Bank of Israel s 2008 Annual Review 2008 (page ), on December 3, 2008, he five larges banks hold a 94% share of oal asses of he banking indusry, and hey herefore represen he banking secor in Israel. 4. The daa and Descripive Saisics As aforemenioned, he daa was aken from he financial saemens of he five larges banks for he period December 3, 2003 o December 3, 2007, where he daa on rading revenues was aken from he financial saemens for he period December 3, 2002 o December 3, 2007, given ha he dependen variable also uses he four quarers preceding December Table presens an example of Bank Leumi's VaR reporing for 2007, exraced from he bank's financial saemens. Table : Sample of VaR reporing by Bank Leumi, December 2007 Toal a group level Toal a group level for revalued porfolios by marke value Limiaion December 3, average December 3, average

14 Source: Bank Leumi, 2007 annual financial saemens Bank Leumi uses he normal disribuion mehod o calculae VaR. The following are he parameers i uses: 99% confidence level and wo-week ime horizon. The able disinguishes beween oal banking aciviy and he revalued porfolios according o marke value and beween year-end VaR and average VaR. The able also shows a very ineresing process. The VaR sysem has become an inegral par of risk conrol. The board of direcors defined risk policy and limis in erms of VaR, inferring ha VaR is used for risk conrol by he board and no only as a regulaory and risk managemen ool. Table 2 specifies how VaR was calculaed by he five banks. I can be seen ha hree banks (Leumi, Discoun and Firs Inernaional) used he parameric approach ha is based on he assumpion of normal disribuion. A wo banks (Hapoalim and Mizrahi) he calculaion was done using he hisorical simulaion mehod and a imes also using several mehods. Table 2: Informaion abou VaR mehods a he five large banking groups Time Confidence Bank horizon level Calculaion mehod Bank Hapoalim The high resul from he hisorical simulaion 0 calculaion and he calculaion according o business 99% he Mone Carlo simulaion (on oal rade in days Israel) Bank Leumi Two weeks 99% The parameric mehod (for MTM revalued porfolios) Israel Discoun 0 99% The parameric mehod (for he Group's oal 4

15 business days Mizrahi-Tefaho Monh 99% 0 Firs business Inernaional days 99% risk) Hisorical simulaion on Group's oal asses (on he Group's oal risk) The parameric mehod (for he Group's oal risk) Source: Financial saemens of he five large banks Figure illusraes he developmen of he quarerly rading revenues for he five banks. The figure shows how over mos of he period, here is a posiive correlaion beween he rading revenues of he five banks. Figure : Developmen of he quarerly rading revenues of he five large banks - December 3, December 3, 2007 Source: Financial saemens of he five large banks and he auhors' processing. 5

16 Table 3 provides addiional informaion on he variables in he regression. Table 3: Average, sandard deviaion, maximum value and minimum value of he variables in he regression (NIS, housands) R S - E [ R ] + + Bank Maximum Minimum Sandard Deviaion Average Bank Hapoalim 67,436 4,004 6,55 33,027 Bank Leumi 297,320 3,240 65,858 69,83 Israel Discoun 84,209 29,732 35,67 63,665 Mizrahi-Tefaho 50,904 87,78 5,327 05,2 Firs Inernaional 33,578 33,395 25,43 77,055 Bank Hapoalim 787,688 6,25 208, ,77 Bank Leumi 800,000 8,375 28, ,898 Israel Discoun 366,500 4,875 03,3 07,47 Mizrahi-Tefaho 46,438 6,375 45,895 55,776 Firs Inernaional 232, ,235 05, Empirical resuls We es he associaion beween he VaR based sandard deviaion and he absolue value of unexpeced rading revenues in he subsequen quarer. The OLS regression resuls are repored in Table 4. As expeced, he impac of he sandard deviaion based on VaR is posiive and significan for hree of he five large banks (Bank Leumi, Israel Discoun Bank, Firs Inernaional Bank). This finding indicaes ha he VaR disclosures are posiively relaed o he variabiliy of he subsequen rading revenues of hree banks. I is ineresing o noe ha hese hree banks used he parameric approach for calculaing VaR. This is also he approach used o develop our mehodology in secion 3.2 above. 6

17 Table 4: Resuls of he regressions of Equaion 7 for each of he five large commercial banks, quarerly daa for December 3, December 3, 2007 Bank Inercep Slope 2 R Bank Hapoalim 24,647 (0.56) Bank Leumi 47,467 (.89)* Israel Discoun Bank 79,076 (2.08)** Mizrahi-Tefaho Bank 403,7 (.42) Firs Inernaional Bank -593 values in he second line. (-0.0) * Significan a 0.0 level; **significan a 0.05 level (0.73) 2.53 (2.)** 0.49 (.84)* -.82 (-0.69).56 (2.33)** 6.5% 27.9% 44.3% 73.8% 84.2% We also performed a regression analysis based on panel daa (polled sample) of he five banks over ime (unabulaed). The coefficien based on he VaR was found o be posiive, bu no saisically significan. If we summarize he resuls of he regressions for he five large Israeli banks, we find ha in mos cases hey suppor our hypohesis ha here is a posiive associaion beween he VaR based sandard deviaion and he absolue value of unexpeced rading revenues in he subsequen quarer. In oher words, he VaR daa generally have an informaive value. A similar sudy upon which we relied was conduced in he US by Jorion (2002). His es was conduced on a quarerly basis for , a oal of 23 observaions for each bank. Jorion showed ha VaR has an informaive value for some of he banks 7

18 and for he sysem as a whole. The relaion was found o be posiive and saisically significan for four of he eigh banks in his sudy. Addiionally, Jorion analyzed he pooled daa of he eigh banks and found a posiive and significan relaion. Similarly, we usually find a posiive and significan associaion beween he VaR based measure of volailiy and he absolue value of unexpeced rading revenues in he subsequen quarer. For hree of he five banks, a posiive and significan relaion is found beween he variables, supporing our hypohesis. However, in conras wih Jorion's sudy, we do no find a significan relaion for he pooled sample. 6. Summary and Conclusions In his paper, we presen he Value-a-Risk, VaR, which is an advanced ool for esimaing marke risks a banks, insurance companies and oher companies primarily engaged in he finance secor. We hen use a regression analysis o examine he associaion beween he VaR daa repored and he subsequen unexpeced rading revenues for he five large banks in Israel, from We assume ha in ligh of he fac ha VaR is forward-looking and is an acceped indicaor of exposure o marke risks, we can expec a posiive relaion beween he sandard deviaion based on VaR repored by he Israeli banks and he variabiliy of heir rading revenues. The resuls indicae ha here is a posiive and significan associaion beween he sandard deviaion based on VaR and he fuure volailiy of he rading revenues a hree of he five large banks examined. However, we should pracice a cauious approach in order no o generalize he resuls. Firs of all, he resuls are sensiive o he sample and period. Second, i is imporan o noe ha he mehodology we used o examine he research hypohesis 8

19 is based on he parameric approach (normal disribuion), while banks someimes use oher mehods o calculae VaR, leading o resuls ha differ from hose hrough he parameric approach (e.g. Bank Hapoalim and Mizrahi Bank in our sample). Third, i is imporan o remember ha he mehodology we used includes an assumpion of sabiliy of he posiion during he coming fuure quarer. In fac, changes can cerainly occur in a posiion ha will impac on he quarerly resuls. The resuls of he sudy may have several implicaions. We noed ha he providen funds and insurance companies have sared using VaR. The reporing on VaR may be helpful o he users of he financial saemens. Thus, for example, lenders can use his measuremen when deciding on wheher o deposi in an insurance company and a wha rae. Furhermore, he sudy may be imporan for regulaors and accouning sandards insiuions when seing banking regulaion and financial reporing sandards. We believe i is imporan o coninue o research he issue a banks, afer expanding he informaion on public financial saemens. Thus, for example, expanding he reporing so ha i include a breakdown of he VaR repored for each of he marke risks (CPI, exchange rae and ineres) and wih a quaniaive descripion of he impac of diversificaion could conribue o improving he informaiveness of he VaR daa. This ype of reporing can indicae he conribuion of exposure of each of he ypes of marke risks o he overall measuremen and show he offseing impac aribuable o he diversificaion among he risk facors. Addiionally, full disclosure regarding he changes in he fair value of he banking porfolio as a resul of he exposures may also be helpful o users of he financial saemens. Finally, his research should be expanded in he fuure o insurance companies and companies 9

20 primarily engaged in he financial secor afer hey adop he VaR model and repor on i in heir financial saemens. References Alexander, C. and Sheedy, E. (2008), "Developing a sress esing framework based on marke risk models", Journal of Banking & Finance, Vol. 32 No. 0, pp Bali, T.G., Gokcan, S. and Liang B.(2007), "Value a risk and he cross-secion of hedge fund reurns", Journal of Banking & Finance, Vol. 3 No. 4, pp Barh, M., (2006), "Including Esimaes of he Fuure in Today's Financial Saemens", Accouning Horizons, Vol. 20, No 3, pp, Barh, M., (2007)," Sandard Seing Measuremen Issues and he Relevance of Research", Accouning and Business Research, Special Issue, 7-5. Booh, J.R. and Officer, D.T. (985), "Expecaions, ineres raes, and commercial bank socks", Journal of Financial Research, Vol. 8, pp Brewer, E., Jackson W.E. and Moser J.T. (200), "The value of using ineres rae derivaives o manage risk of U.S. banking organizaions", Federal Reserve Bank of Chicago Economic Perspecives, No. Q, pp Carer, D.A. and Sinkey, J.J. (997), "The use of derivaives and he ineres-rae sensiiviy of bank sock reurns", working paper, Universiy of Georgia. Chamberlain, S.L., Howe J.S. and Popper H. (997), "The exchange rae exposure of U.S. and Japanese banking insiuions", Journal of Banking & Finance, Vol. 2 No. 6, pp Chance, D.M. and Lane, W.R. (980), "A re-examinaion of ineres rae sensiiviy in he common socks of financial insiuions", The Journal of Financial Research, Vol. 3, pp Choi, J.J. and Elyasiani, E. (997), "Derivaive exposure and he ineres rae and exchange rae risks of U.S. banks", Journal of Financial Services Research, Vol. 2 No. 2, pp Choi, J.J., Elyasiani, E. and Kopecky, K.J. (992), "The sensiiviy of bank sock reurns o marke, ineres and exchange rae risks", Journal of Banking & Finance, Vol. 6 No. 5, pp

21 Dowd, K. (2000), "Accouning for value a risk in financial insiuions' porfolios", Journal of Risk Finance, Vol. 2, No., pp Flannery, M.J. and James, C.M. (984), "The effec of ineres rae changes on he common sock reurns of financial insiuions", Journal of Finance, Vol. 39 No. 4, pp Harris, J.M., Marr, W.M. and Spivey, M.F. (99), "Exchange rae movemen and he sock reurns of U.S. commercial banks", Journal of Business Research, Vol. 22 No. 3, pp Hirle, B. (997), "Derivaives, porfolio composiion, and bank holding company ineres rae risk exposure", Journal of Financial Services Research, Vol. 2 No. 2, pp Hirle, B. (2003), "Wha marke risk capial reporing ells us abou bank risk?", Federal Reserve Bank of New York Economic Policy Review, Sepember, pp Hirle, B. (2007), "Public disclosure, risk, and performance a bank holding companies", Federal Reserve Bank of New York Saff Repor, No Inernaional Accouning Sandards Board (IASB), (2005), Financial Insrumens: Disclosures. Inernaional Financial Reporing Sandard No. 7. London, U.K.: IASB. Jorion, P. (2002), "How informaive are value-a-risk disclosures?", The Accouning Review, Vol. 77 No. 4, pp Kane, E.J. and Unal H. (988), "Change in marke assessmens of deposi-insiuion riskiness", Journal of Financial Services Research, Vol., pp Kwan, S.H. (99), "Re-examinaion of ineres rae sensiiviy of commercial bank sock reurns using a random coefficien model", Journal of Financial Services Research, Vol. 5 No., pp Linsley, P.M. and Shrives, P.J. (2000), "Risk managemen and reporing risk in he UK", The Journal of Risk, Vol. 3 No., pp Linsley, P.M. and Shrives, P.J. (2005), "Transparency and he disclosure of risk informaion in he banking secor", Journal of Financial Regulaion and Compliance, Vol. 3 No. 3, pp Liu, C.C., Ryan, S.G. and Tan H. (2004), "How banks' value a risk disclosures predic heir oal and priced risk: effecs of bank echnical sophisicaion and learning over ime", Review of Accouning Sudies, Vol. 9, pp Lloyd, W.P. and Shick, R.A. (977), "A es of sone s wo-index model of reurns", Journal of Financial and Quaniaive Analysis, Vol. 2 No. 3, pp

22 Rajgopal, S. (999), "Early evidence on he informaiveness of he SEC's marke risk disclosures: he case of commodiy price risk exposure of oil and gas producers", The Accouning Review, Vol. 74 No. 3, pp Rogachev, A. (2007), "Value-a-risk concep by Swiss privae banks", Journal of Risk Finance, Vol. 8, No., pp Taylor, J.W. (2005), "Generaing volailiy forecass from value a risk esimaes", Managemen Science, Vol. 5 No. 5, pp Schrand, C.M. (997), "The associaion beween sock-price ineres rae sensiiviy and disclosures abou derivaive insrumens", The Accouning Review, Vol. 72 No., pp Sco, W.L. and Peerson, R.L. (986), "Ineres rae risk and equiy values of hedged and unhedged financial inermediaries", The Journal of Financial Research, Vol. 9, pp Supervisor of he Banks, Direcive 339,Risk Managemen. Supervisor of he Banks, Direcive 34, Capial Allocaion for marke risk exposure. Auobiographical noe Shilo Lifschuz Academic Cener of Law and Business 26 Ben-Gurion S. Rama Gan, Israel Tel: shilo@clb.ac.il Shilo Lifschuz, is he chairman of he accouning deparmen and a faculy member of accouning and finance a he Academic Cener of Law and Business, Rama Gan, Israel. He holds a PhD from he Hebrew Universiy of Jerusalem and he is a CPA (Israel and Illinois, USA). His main research ineress are: risk reporing, banking and risk managemen 22

23 Auobiographical noe Roi Polanizer Address: 5 Branizky S. Rishon LeZion 75242, Israel Telefax: +972 (0) roi@ogen.co.il Mr. Polanizer had successfully passed all he Porfolio Manager License Exams adminisered by he Israel Securiies Auhoriy (ISA), holds a B.A. (cum laude) in Economics specializing in Finance and a M.B.A. (cum laude) in Business Adminisraion specializing in Finance, boh from he Ben-Gurion Universiy of he Negev. Mr. Polanizer has considerable experience as risk manager, founder and head of he Fair Value Deparmen of he consuling firm of Ogen Ld. - Acuarial, Financial and Business Consuling, research assisan for he Dr. Shilo Lifschuz, CPA, in he field of Risk Managemen in he Israeli Banking Sysem (paricipaion and leading empirical sudies abou he VaR Model and he Marke-o-Book Value Raio in he Israeli Banking Sysem, as well as wriing academic aricles for publicaion in professional journals), eaching assisan for he Dr. Shilo Lifschuz, CPA, for Finance (using Excel's advanced ools in finance) and Banking (Risk Managemen Regulaions in public companies and financial insiuions) courses, risk manager and chief modelis of an Invesmen Commiee a Ben-Gurion Universiy of he Negev, adjunc lecurer in he Managemen Deparmen a Achva Academic College for Derivaives and Risk Managemen courses, adjunc lecurer in he School of Economics a Ashkelon Academic College for Financial Saemen Analysis and Valuaions courses, and lecurer in a prive preparaion course for he professional exams adminisered by he Israel Securiies Auhoriy (ISA) for a Porfolio Manager License. 23