exposicion sab

Upload: grenderff

Post on 07-Apr-2018

243 views

Category:

Documents


0 download

TRANSCRIPT

  • 8/6/2019 exposicion sab

    1/57

    1

    Macroeconomic Stress Testing forIndonesian Banking System

    Muliaman D. Hadad1

    , Wimboh Santoso2

    , Bagus Santoso3

    , Dwityapoetra S. Besar4

    , ItaRulina 5

    Abstract

    The aim of this paper is to perform macro stress test (top down approach) forIndonesian banking system using banks monthly data from December 1996 untilMay 2005. We compare the behavior of big banks (15 banks) and small banks(116 banks) by using univariate and multivariate approach. The variables are banks Loan Loss Provisions, GDP, Inflation, growth of Money, SBIs rate, foreignexchange, gasoline prices, diesel fuel prices 6. We employed pooled least square -fixed effects technique. The result of the univariate regressions show that theprice stability indicators, i.e. the growth of money and the inflation, play greatrole in large credit risk behavior. In most regressions, those variables havesignificant long-run impact on the proxies of credit risk used. In addition, the riseof Premium price and Solar price have large long-run impact on some dependent variables which reflect credit risk. Overall, the result of the multivariateregression suggests the importance of price stability in order to maintainfinancial stability in term of credit quality.

    Keywords : Banks, Macroeconomic, Business Fluctuations and Cycles JEL Classification : G21, C32

    1 Director of Directorate for Banking Research and Regulation, Bank Indonesia;[email protected] 2 Executive Bank Researcher at Financial System Stability Bureau Directorate for BankingResearch and Regulation, Bank Indonesia ; e-mail address : [email protected] 3 Researcher at Gajah Mada University; [email protected] 4 Senior Bank Researcher at Financial System Stability Bureau Directorate for BankingResearch and Regulation, Bank Indonesia; [email protected] 5 Bank Researcher at Financial System Stability Bureau Directorate for Banking Research and

    Regulation, Bank Indonesia ; email address: [email protected] 6 Government differentiates the oil prices into two separate prices namely: gasoline price anddiesel fuel price.

  • 8/6/2019 exposicion sab

    2/57

    2

    1. Background

    The linkage between the financial system and the business cycle has been thesubject of a lot of study. It is important for central bankers to understand how banks are affected by the development of business cycle.

    During the downturn of economy, consumers profitability is going down and willdeteriorate banks loan quality, then, often causing losses in banks balancesheets. In addition, the bad economy condition usually accompanied by the riseof unemployment will reduce households income and their repayment capacity.To overcome the risks, bank creates larger provisions and higher levels of capital, just when it is hard to have one. Sometimes, banks may respond by reducinglending, particularly if their capital is in critical point. This will worsen theeconomic downturn (procyclical).

    Berger and Udell (2003) mentioned that the Ideas of procyclicality of financialsystem is procyclical are quite consistent with economic events, such as creditcrunch in the US during the early 1990s, Asian crisis in the late 1990s and thelarge corporate bankruptcies in the early 2000s. Many theories focus on theincrease and decrease in the supply of bank credit over the business cycle in orderto explain two stylized facts widely observed by regulators, practitioners andresearchers. First, lending often increases significantly during business cycleexpansions and then falls significantly during subsequent downturns, sometimesdramatically enough to be labeled a credit crunch. These changes in lending aregenerally more than proportional to the changes in economic activity, suggestingthat they are changes in bank loan supply that tend to highlight the businesscycle.

    The second fact about bank loan procyclicality is that observed measures of loanperformance problems appear to follow a distinct pattern over the business cycle.Past due, nonaccrual, provisions and charge-offs are generally very low duringmost of the expansion, start to appear at the end of the expansion, then risedramatically during the downturn. This suggests that banks may takesignificantly more risks during the expansion, but these risks are revealed only later because it takes time for loan performance problems to appear (Berger andUdell, 2003).

    The analysis of banking sector is a central component of financial stability monitoring since it is still a key of financial stability as a whole. The bankingcrisis in several emerging countries in the late of 1990s has proven costly.Empirical studies show that the costs of crisis in developing countries, on

    average, reached 15-20 percent of GDP annually, for example Hoggarth andSaporta (2001). Costs of banking crisis were generally higher in case of asimultaneous currency crisis (Mrttinen et al., 2005). In Indonesia, the cost of banking crisis in 1997 was 51% of GDP.

    In response to the increases financial instability in many countries in the 1990s,central bankers have become interested in better understanding the vulnerabilities in financial systems and measures that could help preventfinancial crises. One of the key techniques for quantifying the linkage between

  • 8/6/2019 exposicion sab

    3/57

    3

    macroeconomic and financial sector is stress testing. Stress tests are generally arisk management tool banks use to determine how the value of their portfolio would change in the event of unexpected crisis. Stress test may produce usefulresult not only for individual banks. Central banks who maintain financial systemare challenge with the questions to what level a certain unexpected scenario would affect financial system and the banking system as a whole. Macro stresstest will explore what implications general macroeconomic crisis scenario such asoil shocks would have for the financial system.

    The following macro stress test for the Indonesian banking sector, which is using banks individual loan loss data, draws on the approach outlined in BanksPerformance Over the Business Cycle: A Panel Analysis on Italian Intermediaries, by Mario Quagliariello (2004). The best way to approach this is probably to start by simply replicating what they have done for Indonesia. To comprehendIndonesian condition, we employ other macro variables influencing most banksin Indonesia. The purpose of this paper is to perform macro stress test forIndonesian banking system using banks monthly data from December 1996 untilMay 2005. We compare the behavior of big banks (15 banks) and small banks(116) by using univariate and multivariate approach.

    This paper is structured as follows: section 2 provides literature study on macrostress testing. The methodology underlying stress test and data analysis arepresented in section 3. Section 4 explains the simulation and stress testingprocedures and the result. Finally, section 5 provides a summary.

    2. Literature Studies

    Stress tests are increasingly used by the supervisory authorities to assess the

    resilience of the financial system to adverse macroeconomic disturbances, thusenhancing their action. Kupiec (1998) points out that a common practice tocompute stress exposure measures may be imperfect. He thought that thecommonly computed stress exposure measures, based on the variance-covarianceapproach for calculating VaR, are usually obtained by setting to zero the changeson the "remaining" risk factors. Basically, stress tests are aimed to complementthe internal model approach, i.e. Value at Risk (VaR), which is fairly well-known.Guy et al. (1999) shows the fact in addition to VaR calculation, stress tests should be used to measure the risk of financial transactions.

    The exact definition of stress test is surprisingly unclear. Berkowitz (1999) citesseveral definition, for example 1999 BIS document, Framework for Supervising Information about Derivatives and Trading Activities, which says that stressscenarios need to cover a range of factors that can create extraordinary losses orgains in trading portfolios and that they should provide insights into the impactof such event on positions. Despite the lack of formal definition, Berkowitz(1999) suggests that stress test means choosing scenario that are costly and rare,and then putting them to a valuation model.

  • 8/6/2019 exposicion sab

    4/57

    4

    Other researchers, Fender et al (2001), argue that stress test then is consideredcomplements firms other risk management tool, particularly VaR. It has limitedability to capture the risks of exceptional market events, especially those in whichasset prices move in ways that differ sharply from historical norms. VaR calculation should be combined with stress tests because a VaR model does notshed light on the dimension of heavy losses and somewhat skeptical attitudetowards the assumption on which most VaR calculations are based.Oesterreichische Nationalbank (2001), stated that stress tests are not based onstatistical assumptions on how the changes in risk factors are distributed.Therefore, the results of stress tests are not distorted by fat tails.

    Additionally, Blaschke et al. (2001) define the term stress test as a range of techniques used to assess the vulnerability of a portfolio to major changes in themacroeconomic environment or to exceptional, but plausible events. Bank forInternational Settlement (2000) identifies stress test as the examination of thepotential effects on a firms financial condition of a set of specified changes in risk factors, corresponding to exceptional but plausible events.

    The concept of stress testing is generally considered a way to identify scenariosthat could cause significant loss, outside regular market events (Kuo, et al .,2002). This test is valuable in assessing the stability of banking system (DeutscheBundesbank, 2003). On a broader view, stress testing the vulnerability of financial institutions to adverse macroeconomic events is an important tool inassessing financial stability (Hoggarth et al., 2003). Some historical events withparticular relevance to bank and security firm portfolios are Black Monday (1987), the Asian financial crisis (1997) and Russian default (1998). The terroristattack in the United States (2001) has also formed the basis for many historicaland hypothetical scenarios.

    The International Monetary Fund (IMF) and the World Bank consider put stresstest as an important component of their Financial Sector Assessment Program(FSAP). Stress tests carried out in the context of the FSAP aim to assess key risksand vulnerabilities arising from macro-financial linkages by assessing the impactof exceptional but plausible shocks to key macroeconomic variables on thesoundness of the financial system (IMF and World Bank, 2003).

    Moreover, Bank regulators consider stress tests to be an effective and necessary tool that complements statistical model for quantifying and monitoring risk, forinstance Basle Committee on Banking Supervision and EU Directive(Oesterreichische Nationalbank, 2001). Basle Committee on Banking Supervision(1995) notes that stress testing to identify events or influences that could greatly impact banks is a key component of a banks assessment of its capital position.Therefore, banks that use internal model approaches for meeting market risk capital requirement must have in place a rigorous and comprehensive stresstesting program. The Basel Committee underlined the need for stress testing when it published the Amendment to the Capital Accord to Incorporate MarketRisks in 1996; banking supervisors have then established the use of stress testsas an important component of the intermediaries internal-models approach tomarket risk monitoring (IAIS, 2003).

  • 8/6/2019 exposicion sab

    5/57

    5

    Essentially, there are two major goals of stress testing. First, is to evaluate thecapacity of the banks capital to absorb potential large losses and second, is toidentify steps the bank can take to reduce its risk and conserve capital(Berkowitz, 1999). Therefore, during the past few years stress tests have becomemore important as a part of banks risk management.

    Top Down vs Bottom Up

    In principle, two solutions are available for the aggregation rule: supervisors candefine the macroeconomic shock, let the intermediaries evaluate its impact ontheir balance sheets and then aggregate the bank-level outcomes to get the overalleffect (bottom-up approach) or, conversely, they can directly apply the shock tosome sort of banking system-level portfolio and analyze the aggregate effect (top-down approach). Of course, in the bottom-up methodology, the issue of comparability is a relevant one since each intermediary may employ differentmethodologies and modeling assumptions, making the aggregation less reliable.Conversely, the top-down approach enhances the comparability of the results, but it is typically based on historical relations (IAIS, 2003).

    The bottom-up approach is suited mainly to the field of market risk as many banks possess comparable market risk models. In the field of credit risk and inthe case of macro stress tests, the heterogeneity of the models means that a top-down tends to be called for (Deutsche Bundesbank, 2003).

    Hoggarth et al. (2003) applies both bottom-up and top-down approach in macrostress testing of United Kingdom banks. Four scenarios are constructed based ondomestic and global events, and shift in the demand for and supply of goods andservices in the economy. The bottom-up result shows that in overall, the impacts

    those events were quite small in all scenarios. The top-down simulation wasconducted using single equation econometric model, in which the effects of key macroeconomic and bank-specific variables on the provision made againstaggregate credit losses by major UK-owned commercial banks was estimated.These simulations suggest that the likely increases in credit losses arising underall scenarios are quite small. The results tend to be sensitive to the nature andspecification of the macroeconomic stress tests. The size of shocks is basedlargely on historical experience averaged over normal times and period of stress,rather than taken from stress period alone.

    Types of Stress TestStress test involved one or more of the following types of analysis

    1. Sensitivity test

    Sensitivity test estimates the impact of one or more moves in a particular risk factor on the financial condition such as exchange rates, economic growth andinterest rates (IAIS, 2003). It is also called univariate stress test. In sensitivity

  • 8/6/2019 exposicion sab

    6/57

    6

    test, only a single risk factor was changed. In this case, it is assumed that thereis no correlation between the risk factors (Deutsche Bundesbank, 2003).

    The advantage of this test is that it can isolate the specific influence of individual risk factors from that of other factors. Based on this test, credit

    institutions can identify the weakneses of their portfolio structure relatively accurately. The drawback, however, is that it ignores the interaction of variousrisk factors (Deutsche Bundesbank, 2004). Therefore there should bemultivariate stress-test or scenario analysis as a supplement, in which morethan one risk factor is changed at one time (Deutsche Bundesbank, 2003).

    2. Scenario Analysis

    Scenario analysis is more comprehensive test. It includes simultaneous movesin a number of risk factors and is often linked to explicit changes in the view of the world. There are two basic types of scenario analysis; historical andhypothetical scenario. Historical scenarios reflect changes in risk factors thatoccurred in specific historical episodes. Hypothetical scenarios use a structureof shocks that is thought to be plausible, but has not yet occurred (IAIS,2003).

    Historical scenarios is the most intuitive approach despite that it is backwardlooking and may lose relevance over time as markets and institutionalstructures change (Blaschke et al ., 2001). BIS (2000) discusses twoadvantages and disadvantages of historical scenario. The advantages are itstransparency and that the structure of market factor changes is historicalrather than arbitrary. Regarding the disadvantages, the firms may structuretheir risk-taking to avoid losing money on shocks that have occurred in thepast, rather than anticipating future risk that do not have a precise historicalparallel. It may also be difficult to apply to products which do not exist at thetime of the historical event being discussed or to risk factors whose behaviorhas changed significantly since that event.

    Regrettably, the number of usable historical scenario is limited because they are rare by definition and it is difficult to obtain reliable data for many instruments further back than just a few years. If historical event is includedin a stress test, the start and end date for the scenario must be specified.However, even Basle Committee on Banking Supervision did not state whichdate to use for their prescribed scenarios (Schachter, 1998).

    Hypothetical scenario is a forward looking one. It can allow a more flexibleformulation of potential events, as well as encouraging risk manager to bemore forward looking. The main disadvantage of this approach is thedifficulty to determine the likelihood of an event occurring since it is beyondthe range of expectation (Blaschke et al ., 2001).

    It is important to determine whether the exercise should be based onhistorical scenarios, assuming that past shocks may happen again, or ratheron hypothetical scenarios, that is on extreme but plausible changes in the

  • 8/6/2019 exposicion sab

    7/57

  • 8/6/2019 exposicion sab

    8/57

    8

    situation that could lead to high losses of the bank. On the other hand, insystematic search for worst-case scenarios, banks use their computers tosystematically search for worst-case scenarios (OesterreichischeNationalbank, 2001).

    4. Extreme Value Theory This is the statistical theory concerned with the behaviour of the tails, i.e., the very high and low potential values) of a distribution of market returns. It is a better means to capture the risk of loss in extreme, but possible,circumstance. Due to focusing on the tail of a probability distribution, it can be more flexible. However, it produces a problem if it adapts to a situation where many risk factors drive the underlying return distribution.Furthermore, the usually unstated assumption that extreme events are notcorrelated over time is questionable (BIS, 2000).

    5. Contagion analysis

    Contagion analysis seeks to assess the impact of transmitting from anindividual financial institution to the rest of the financial system (IMF and World Bank, 2003). The values used by financial institutions for contagioneffects are generally based on judgment and historical experiences rather thanon formal models of market behaviour (BIS, 2000).

    Basle Committee on Banking Supervision (1996) requires that banks stress testsscenario should cover a range of factors that can create extraordinary losses orgains in trading portfolios. These factors include low-probability events in allmajor types of risk, including the various components of market, credit andoperational risks (Oesterreichische Nationalbank, 2001). Basle Committee onBanking Supervision (2000) states that stress testing should involve identifying

    possible events or future changes in economic conditions that could haveunfavourable effects on a banks credit exposures and assessing the banks ability to withstand such changes. Economic or industry downturns, market-risk eventsand liquidity conditions are three areas banks could usefully examine.

    The concept of stress test is relatively straightforward. However, the applicationof this technique in practice is more complicated (IAIS, 2003). Some of thedifficulties are:

    determining what risk factors to stress

    establishing how such factors should be stressed

    establishing what range of values should be used

    determining the time horizon that such tests should consider

    meaningfully analyzing the results and making informed judgments.

    Regarding risk factors, there were two risk factors commonly employed in stresstests, i.e. credit and market risk. The credit risk refers to the risk that borrowers will not fulfill their payment obligations or not fulfill them on time (default risk).In a broader sense, it denotes the risk of deterioration in a borrowers

  • 8/6/2019 exposicion sab

    9/57

    9

    creditworthiness (Deutsche Bundesbank, 2004). Basle Committee on BankingSupervision (2000) simply defines credit risk as the potential that a bank borrower or counterpart will fail to meet its obligations in accordance with agreedterms.

    Market risk denotes the risk of a change in the market value of long or shortposition due to fluctuation in the underlying market prices. It comprises interestrate, equity price, commodity price, exchange rate and volatility risk (DeutscheBundesbank, 2004).

    It is important to note that the choice of risk factors depends on the portfolio of banks due to not all portfolios are influenced by the same risk factors. However,the procedure for selecting risk factors is not clearly defined (OesterreichischeNationalbank, 2001). The surveys conducted by Deutsche Bundesbank (2004)confirm that stress tests are employed mainly in the area of market risk,especially interest rate risk. On the other hand, only about 30% cases analyzedcredit risk. Blaschke et al. (2001) discusses deeply the type of risks utilized in

    stress test.

    Univariate vs Multivariate

    Most of previous stress tests have involved single-factor sensitivity analysis basedon historical extreme values. Growing number of researches have also appliedscenario analysis.

    Gabon, Mexico and Sweden are among countries conduct FSAP IMF and WorldBank, 2003). FSAP to Gabon was based on scenario analysis that focus to explorethe linkages between macroeconomic developments and the financial sector by examining the effect on commercial banks solvency or liquidity of various typesof risks. Contagion analysis was also applied to measure the risk stemming from areduction in oil production by tracing its effect on the government's ability toservice its debt (which is a function of oil income) and the corresponding impacton commercial banks income. Main findings of this test suggest that minimumregulatory solvency ratios were low in some institutions and relatively smallshocks to macroeconomic variables could place individual institutions at risk. Inaddition, there was no indication of contagion risk across banks.

    The stress test in Mexico was conducted through scenario analysis performed by using Generalized Vector Auto-regression model to examine the impact of macroshocks on credit quality. The result of the test showed that slowdown in the U.S.

    economy and depreciation in the peso would have a large affect on the bankingsystems capital and profitability.

    Stress testing of the four major financial groups in Sweden indicated that they were robust to equity and real estate price shocks, as well as to shocks toexchange and interest rate shocks, and to a temporary economic slow down. Theresults of the tests were useful in identifying risk exposures of major banks.

    In 2001, a committee on the Global Financial System (CGFS) of the Central banksof the Group of the ten countries provides a comprehensive international survey

  • 8/6/2019 exposicion sab

    10/57

    10

    on stress testing. The census revealed that banks place particular emphasis onstress-testing equity (see Bank for International Settlements, 2001).

    Quagliariello (2004) provides a comprehensive investigation on banksperformance over the business cycle using an unbalance panel of 207 Italian

    intermediaries over the period 1985-2002. Quagliariello estimates both static anddynamic model in order to carry out simple stress test. The test aims to assess theeffect of macroeconomic shocks on banks balance sheets. His study shows that banks loan loss provision, non performing loans, and return on assets areaffected by the changes of general economic condition. Quagliariello also findsthat several banks level indicators are also relevant in explaining the changes inthe evolution of riskiness and profitability. This supports the idea that the overallperformance of the intermediaries is the result of interaction between the generaleconomic condition and banks management.

    3. Methodology of Study

    Data

    This study uses monthly data over the period of December 1995 to May 2005.Initially we select 17 the biggest banks in Indonesia, based on their asset value.The next step we refine our sample by dropping all five foreign banks from oursample and add three more banks. The new set of sample consists of 15 large banks. Data for financial bank variables (NPL and LLP) are obtained from Bank Indonesia. Whereas data for macroeconomic variables are downloaded from IMFInternational Financial Statistic and Blomberg. Data for oil prices are fromPertamina.

    Model Specifications

    We employ linier regression model for panel data. We start with a GeneralUnrestricted Panel Model (GUPM). Our GUPM is a varian of dynamic fixedeffects model as follows

    it

    n

    mmit i

    k

    j j it i i it X Y t Y ++++=

    =

    =

    11

    (1)

    where

    Y it: credit risks (LLP or NPL)

    Xit : macroeconomic variable i at time t a i : individual effect from each bank it : residual, where e it~N(0, 2)

    In order to estimate model (1), we employ general to spesific approach usingPcGets. Here the GUPM is estimated by two strategies, first fixing i dan Y it-1 toexist at all the times and by setting =0. Second, we treat all variables can be

  • 8/6/2019 exposicion sab

    11/57

    11

    candicate for reduction. We label the first approach as fixed approach and thesecond one as the flexible approach.

    As mentioned earlier, the dependent variable is the credit risk, and here wepropose six variant of NPL and LLP, these are

    1. LLP/TL = LLP/Total Loan

    2. NPL/TL = NPL/Total Loan

    3. D(LLP/TL) = (LLP/Total Loan)

    4. D(NPL/TL) = (NPL/Total Loan)

    5. G(LLP) = dlog (LLP)

    6. G(LLP/TL) = dlog (LLP/Total Loan)

    Whereas the independent variables are macroeconomic variables in which from

    various literature reviews show correlation with credit risk. Here examines ninemacroeconomic variables. These variables are grouped into five categories, i.e.cyclical indicator, price stability indicator, financial market indicator and central bank indicator.

    1. Cyclical indicator

    Loan quality is sensitive to economic cycle therefore we include cyclical variable in our study. Kalirai and Scheicher (2002) provide the transmissionlink between economic activity and loan quality. Deterioration in economic

    activity leads to falling income and rising payment difficulties. Then, more business failures that will result in higher default risk. Therefore the cyclical variable, the growth of real (GDP), is expected to negatively related to creditrisk.

    2. Price Stability indicators

    As price stability indicators we include consumer price index inflation andmoney growth (M1 and M2. We included money growth since they havepotential connection to inflation. High inflation forces borrowers to pay higher material prices then have less money to meet their obligations. Thenumber of loans in actual default will rise therefore banks will increase theirprovisions 7.

    3. Financial Market Indicator

    7 http://moneycentral.msn.com/content/P127636.asp

  • 8/6/2019 exposicion sab

    12/57

    12

    We include Jakarta Stock Composite Index ( Indeks Harga SahamGabungan -IHSG) as the proxy for financial market indicator. It has similarpattern with the cyclical trend of economy. The high IHSG denotes the highreturns for investor and hence lower the credit risk. IHSG is expected to havepositive impact on credit risk.

    4. External Indicators

    External indicators examined in this study are Rupiah exchange rate to USDollar and domestic oil prices, i.e. Premium and Solar prices. Therelationship between exchange rate and credit risk is ambiguous. It dependsto the international trade and capital account of a country. Some of Indonesia borrowing is in foreign fund hence the depreciation of Rupiah lead to highcredit risk since the debt repayment increase.

    The oil price hike is a negative demand shock to Indonesian economy. Itraises household and business energy cost and cause deterioration ineconomy. Therefore, it produces greater credit risk. Based on the aboverationalization, the depreciation of Rupiah and the increase of oil price areexpected to positively related to credit risk.

    5. Central Bank Indicators

    We examined the impact of the rise of Central Bank Bond ( Sertifikat Bank Indonesia -SBI) one-month rate as the central bank indicator. Banks in

    Indonesia mainly refer to SBI rate in setting their interest rates. The higherSBI rate is usually followed by higher interest rate, which reflects higher costof borrowing. This condition leads to higher loan default since firms andhouseholds face difficulties to service their debt. Therefore, the increase of SBI rate is expected to have positive impact on credit risk.

    These nine variables are transformed into first difference and to ensurestationarity of the series.

    1. G(GDP): real GDP growth rate, measured as d log (GDP)

    2. G(PREMIUMDN) = rate of change of premium fuel price, measured asdlog (premium fuel price)

    3. G(SOLARDN) = rate of change of diesel fuel price, measured as d log (dieselfuel price)

    4. G(M1) = rate of money growth of M1, measured as d log (M1)

    5. G(M2) = rate of money growth of M2, measured as d log (M2)

  • 8/6/2019 exposicion sab

    13/57

    13

    6. G(IHSG) = rate of growth of Jakarta Stock Composite Index (IHSG), ,measured as d log (IHSG)

    7. INF = inflation rates, measured as d log (CPI)

    8. SBI = Central Bank bond rate (1 month SBI), measured in decimal

    9. G(EXRATE) = rate of depreciation of rupiah, measured as d log (Rupiahexchange rate, Rp/$)

    Procedures of Estimation

    In our procedure, the time series are assumed stationary since most of them arein the form of first difference. At the beginning, the group of banks analyzedconsists of 17 banks, five of which are foreign banks. For this group, we only

    conduct stress test for two dependent variables, i.e. NPL/TL and G(LLP) and only for fixed effect model.

    Those five foreign banks are then eliminated and we include three domestic banks into the sample. For this new group, all six dependent variables areestimated. For each regression, we apply both fixed effect and flexible form of GUM.

    General-to-Specific Method is utilized in selecting the optimal lag for eachregression. Here, the short-run and long-run coefficients of each variable areidentified.

    In the next step, both univariate (sensitivity test) and multivariate (scenarioanalysis) regression are employed to estimate the impact of shock of macroeconomic variables on the credit risk of large banks in Indonesia. The long-run coefficient is then utilized in this scenario analysis. We refer to historicalscenario rather than hypothetical scenarion to assess the impact of the largestshock experienced in the time series of those macroeconomic variables (Table 1).These historical scenarios are considered plausible enough due to they actually happened.

    Over parameterization becomes an issue in this analysis, especially due to thesimultaneous inclusion of growth of M1 and growth of M2 in multivariateregression. We then also carry out multivariate regression for eightmacroeconomic variables, in which only growth of M1 or growth of M2 isincluded.

  • 8/6/2019 exposicion sab

    14/57

    14

    Loan Loas Provision and Non-Performing Loan as the Measures of Credit Risk

    Credit risk is still the main source of instability for most banks (Mrttinen et al.,2005). In this case, credits are the largest source of credit risk (Basle Committeeon Banking Supervision, 2000). Therefore credit quality often becomes thedependent variable. Loan loss provision (LLP) and bad debts or non performingloans (NPL) are among measures of credit quality. LLP and NPL have beengenerally considered the transmission channels of the macroeconomic shocksto banks balance sheets (Quagliariello, 2004).

    The quality of credit has a relation with economic cycle. Salas and Saurina (2002)study the relation between the condition of loan and economic cycle in Spain overthe period 1985-1997. They find that during the recession phases, non-

    performing loans (NPLs) are increasing. In such deteriorating economiccondition, the borrowers lack the capacity repay their loans. Basle Committee onBanking Supervision (1999) states that in case of credit quality deterioration, banks should make loan loss provision against profit.

    Banks make loan loss provisions against profits when they believe that borrowers will default; this is the tool they can use for adjusting the (historical) value of loans to reflect their true value. Provisions affect both banks profitability, sincethey represent a cost for the intermediary, and capital, since they reduce the book value of the assets (Quagliariello, 2004). Therefore, the increasing (decreasing)LLP is interpreted as the improvement (deterioration) of credit performance.

    Basle Committee on Banking Supervision (1999) classifies LLP into twocategories, i.e. specific and general provision. Specific provision is reserved inorder to cover a loss that is identified in an individual loans. On the other hand,general provision is reserved for latents losses that are known exist but cannot beascribed to individual loans yet. Further, general provisions are the interim steppending the identification of losses on individual loans that are impaired.

    Kearns (2004) recommends banks to adjust the level of general provision basedon their past experiences, current economic condition, the current rate of impairment in the loan book, changes in lending policies, loan growth andconcentration of credits. Furthermore, the provision should be adequate absorbestimated credit losses associated with the loan portfolio.

  • 8/6/2019 exposicion sab

    15/57

    15

    Figure 1

    Banks Loan Loss Provision and Indonesian Economic Condition

    An important aspect of provisioning is its timing with respect to business cycleand the related issues of procyclicality. The common view is that an economicupswing and rising incomes indicate improving condition for firms and reducethe likelihood of loan defaults, whereas a recession will have the opposite effect(Bikker and Metzemakers, 2002).

    Based on Figure 1, the ratio of LLPs towards total credit significantly increased

    since the beginning of 1998 until April 1999. It is due to the high NPLs, some of which were then transferred to Indonesian Banking Restructuring Agancy (IBRA)to be managed. During 1999, the growth of LLPs was higher than that of credit. Itindicates that banks were too prudent in supplying credit. In addition, banks alsoincreased LLPs to improve their worsening balance sheets due to crisis.

    On the other hand, real GDP experienced negative growth in the same period dueto economic crisis in the late 1997. The improving economy since middle 1999 was then responded by banks by decreasing their ratio of LLP to total credits.

    One attractive finding is provided by Bikker and Hu (2001) who evaluate theprocyclicality of banks provision using unbalanced panel of 26 OECD countriesfrom 1979 to 1999. They find that both contemporaneous and lag coefficient of GDP growth have significant and positive impact on banks provision. Some otherresearches have provided the same result, for example Gambacorta, Gobbi andPanetta (2001). The possible explanation of this ambiguity is that banks policy to

    -30

    -20

    -10

    0

    10

    20

    30

    40

    1996 1997 1998 1999 2000 2001 2002 2003 2004

    RASIO LLP/TOTLOAN

    GDP GROWTH

  • 8/6/2019 exposicion sab

    16/57

    16

    smooth their income. Tax deductible, capital management and signallinghypothesis also provide explanationm for this behaviour (Kearns, 2004).

    Based on income-smoothing hypothesis, banks increase (decrease) their LLPs when their income is inceasing (decreasing). In this case, they make use certainexpenditures, e.g. LLP, to smooth the reported income. The motivation behindincome smooting can be negative or positive. Greenawalt and Sinkey (1988)propose an idea that banks may save in the good times, when they obtain highincome, to anticipate the bad times. On the other hand, , Kanagaretnam et al.(2000) observe that banks smoot their income when their performance, in termof income, is over or under the industry average. They expect that government will pay extra attention to banks that deviate from the industry average.

    The evidence of income-smoothing hypothesis is mixed. Some studies find

    evidence in favor of income smoothing hypothesis (Wahlen, 1994; Scholas et al.,1990) while some do not (Ahmed et al., 1999; Kearns, 2004).

    The term capital management is used to decribed a situation where an institutionmanage its capital to asset ratio by varying the level of provision (Kearns, 2004).This behaviour is documented by Kim and Kross (1998) and Ahmed et al. (1999).This hypothesis assumes that banks with lower Tier 1 capital ratios tend to makemore general provision in order to keep their capital ratios adequate (Bikker andMetzemakers, 2002).

    Some studies find evidence in favor of capital management (Ahmed et al., 1999;Lobo and Yang, 2001) while Collins (1995) finds against iti. Kearns (2004) alsofinds that there is a very little correlation betweeen the rate of provisioning andthe capital to asset ratio in Irish case.

    In most countries, the general provisions are tax deductible. It is an incentives for banks to shift Tier 1 (pure) capital to general provision in order to lower the tax burden (Cortavarria et al., 2000). Kearns (2004) also mentions that a creditinstitution has a clear and unambigous incentive to increase its provisions inorder to reduce its taxable income.

    The last possible explanantion that weaken the relationship between provisionand economic cycle is signalling hypothesis. It is suggested that banks that wantto signal that they are optimistic about the future incomes will reduce theircurrent incomes by increasing provisions and vise versa. Ahmed et al. (1999) andLobo and Yang (2000) provide qualified conclution for this hypothesis.

  • 8/6/2019 exposicion sab

    17/57

    17

    Descriptive Analysis of Macroeconomic Variables

    Table 1 shows the expected sign and the worst extreme values of macroeconomic variables in this study while the descriptive statistic of those variables ispresented in Table 2. The shocks in those variables mainly took place in crisesperiod, in 1997 and 1998, except for solar price. We estimate the hystorical worstimpact of the macroeconomic variables based on their worst extreme values inthe period of this study.

    Table 1Expected Sign and Worst Extreme Values of Macroeconomic Variabels

    VariableExpected Sign

    Effect to CreditRisk

    WorstExtreme

    Values (%)Dates of Event

    Growth of Real GDP - -6.91 April 1998

    Growth of M1 + 25.42 February 1998Growth of M2 + 24.01 Februariy1998Growth of IHSG - -37.86 September 1997Inflation + 12.00 March 1998SBI rate + 70.44 September 1998Depreciation of Rupiah + 83.88 February 1998Price of Premium + 45.20 June 1998Price of Solar + 50.08 May 2001

  • 8/6/2019 exposicion sab

    18/57

    Table 2Descriptive Statistic of Macroeconomic Variables

    2004:01-2005:05 1995:12-2005:05 MacroeconomicVariables Mean Max Min Std Deviasi Mean Max Min Std Deviasi Me

    Growth of Real GDP 0.009 0.022 -0.016 0.009 0.004 0.085 -0.069 0.022 0.011Growth of M1 0.005 0.038 -0.020 0.016 0.013 0.254 -0.056 0.035 0.023Growth of M2 0.006 0.033 -0.018 0.014 0.014 0.240 -0.049 0.032 0.020Growth of IHSG 0.027 0.128 -0.067 0.049 0.006 0.250 -0.379 0.099 0.035Inflation 0.006 0.018 -0.002 0.005 0.011 0.120 -0.011 0.019 0.005SBI rate 0.075 0.080 0.073 0.002 0.179 0.704 0.073 0.138 0.134Depreciation of Rupiah 0.007 0.063 -0.029 0.021 0.013 0.839 -0.343 0.123 0.003Price of Premium 0.017 0.282 0.000 0.067 0.012 0.452 -0.095 0.059 0.000Price of Solar 0.014 0.241 0.000 0.057 0.016 0.501 -0.220 0.078 0.000

    1997 1998Macroeconomic

    Variables Mean Max Min Std Deviasi Mean Max Min Std Deviasi

    Growth of Real GDP 0.009 0.041 -0.042 0.023 -0.017 0.060 -0.069 0.035Growth of M1 0.018 0.062 -0.048 0.027 0.045 0.254 -0.056 0.084sGrowth of M2 0.019 0.072 -0.031 0.023 0.041 0.240 -0.049 0.080Growth of IHSG -0.039 0.101 -0.379 0.132 -0.001 0.250 -0.341 0.166Inflation 0.008 0.017 0.000 0.006 0.048 0.120 -0.003 0.033SBI rate 0.157 0.300 0.106 0.061 0.495 0.704 0.220 0.154Depreciation of Rupiah 0.069 0.394 -0.001 0.108 0.033 0.839 -0.343 0.314Price of Premium 0.000 0.000 0.000 0.000 0.030 0.452 -0.095 0.130Price of Solar 0.000 0.000 0.000 0.000 0.031 0.414 -0.045 0.117

  • 8/6/2019 exposicion sab

    19/57

    19

    4. Result of Stress Test

    Univariate Regression

    17 Banks

    For the group of 17 banks, the impact of each macroeconomic variable on the ratioof NPL to Total Loan (NPL/TL) is significant and in line with the expected signs(Table 3). We find that the growth of GDP, inflation and the growth of M2 havethe largest long-run coefficient, respectively. In term of the worst impact, theincrease of Premium price, the depreciation of Rupiah and the growth of M2historically produce the greatest impact on the rise of NPL/TL ratio (Table 7).

    The increase of Premium price consistently has the greatest impact on the rise of credit risk of these 17 banks due to it has the largest long-run coefficient (Table 3)and the greatest impact on the growth of LLP (Table 8). Table 3 also shows that

    the growth of M1 and the growth of M2 have quite large long-run coefficients. Thegrowth of GDP, inflation, SBI rate and the rise of Solar price are foundinsignificant. This model suggests that the depreciation of Rupiah have greatimpact on the growth of LLP as shown in Table 8.

    15 Banks

    The long-run coefficients of the macroeconomic variables based on univariatefixed effect and flexible form are shown in Table 5 and Table 6, respectively. Thesensitivity analysis of those estimations is shown respectively in Table 9, Table 10,Table 11, Table 12, Table 13 and Table 14.

    Fixed Effect

    From the six dependent variables, the fixed effect regressions with LLP/TL,NPL/TL and D(LLP/TL) as dependent variables provide better result. Forregression with LLP/TL as dependent variables, only the depreciation of Rupiahhas insignificant long-run coefficient. Inflation, money growth and GDP growthare found having rather large long-run coefficients as shown in Table 5. In termof historical impact, money growth and the increase of oil price have provenproduce the largest impact on LLP/TL (Table 9).

    For NPL/TL as dependent variables, only SBI rate has insignificant long-runcoefficient. The growth of GDP has the largest long-run coefficient, followed by inflation and oil price (Table 5). In term of historical impact, the rise of Premiumprice has the largest impact, followed by the depreciation of Rupiah, the rise of Solar price and the growth of GDP (Table 10).

  • 8/6/2019 exposicion sab

    20/57

    20

    When we estimate D(LLP/TL) as dependent variable, only the growth of IHSG hasinsignificant long-run coefficient (Table 5). The money growth, inflation and thegrowth of GDP are the most significant variables, while the depreciation of Rupiah, the rise of Premium price and the money growth have the greatesthistorical impact (Table 11). For regression using D(NPL/TL) as dependent variable, the oil prices show insignificant long-run coefficient. The growth of M2,the inflation, the growth of M1 and the growth of GDP have significant long-runcoefficient (Table 5) while the depreciation of Rupiah, the money growth and theinflation have the greatest historical impacts (Table 12).

    The result of our study on the group of 15 banks suggests that the ratio of growthof LLP to total loan (G(LLP)/TL) is not an appropriate variable to measure theimpact of macroeconomic variables on credit risk. For fixed effect model, none of the macroeconomic variables has long-run impact on the G(LLP)/TL, even thegrowth of GDP has no long-run impact (Table 14). The regression using G(LLP) asdependent variable is not appropriate also. Only the growth of M2, the inflationand the depreciation of Rupiah that have significant long-run impact (Table 13).

    Flexible Form

    The result of flexible form regression provides quite similar results. The modelusing G(LLP) and G(LLP)/TL provides unfavorable results. For model usingG(LLP) as dependent variable, the inflation has no long-term coefficient. Thegrowth of GDP, the growth of IHSG, the SBI rate and the rise of Premium priceare found insignificant (Table 6 and Table 19). For model using G(LLP)/TL asdependent variable, only the depreciation of Rupiah and the rise of Premium pricethat show significant long-run impact (Table 6 and Table 20).

    Based on the flexible form regression, the model using D(LLP/TL) and D(NPL)provide relatively favorable result in term of expected sign. For both model, therise of Solar price has insignificant long-run coefficient while the growth of GDP,the growth of money and the inflation show the large long-run coefficient (Table6). In term of historical impact, the depreciation of Rupiah, the growth of money and the the rise of Premium price produce great impact in model usingD(LLP/TL) as dependent variable (Table 17). In model using D(NPL/TL) asdependent variable, the growth of money, the depreciation of Rupiah and theinflation have proven produce the worst impacts.

    For both model with LLP/TL and NPL/TL as dependent variable, we find that thegrowth of GDP, the growth of money and the inflation are the most significant variables (Table 6). In term of the worst impact, the growth of M1, the growth of GDP, the growth of M2 and the inflation have the worst impact on LLP/TL in the

  • 8/6/2019 exposicion sab

    21/57

    21

    period of estimation (Table 15). The growth of M1, followed by the growth of M2and the depreciation of Rupiah as well as the inflation also produce worst impacton NPL/TL (Table 16).

    Multivariate Regression We consider that single macroeconomic variable shock as we estimate in theunivariate regressions are not likely to occur in isolation without being related toother changes in other macroeconomic variables. Therefore, we conduct morecomprehensive, multivariate stress test on the credit risk of the group of 15 large banks.

    The result of the multivariate regression suggests that none of the macroeconomic variables has long-run impact on D(LLP/TL) and G(LLP)/TL for both fixed effectand flexible model (Table 21, Table 22, Table 25, Table 28, table 31 and Table 34).

    To some extent, the model using LLP/TL, NPL/TL and G(LLP) as dependent variables provides more favorable result despite insignificance of somemacroeconomic variables (Table 21, Table 22, Table 23, Table 24, Table 27, Table29, Table 30 and Table 33).

    Fixed Effect

    For fixed effect model, the growth of GDP, the growth of IHSG and thedepreciation of Rupiah have insignificant impact on credit risk, and even have nolong-run impact for several dependent variables. Variables that have significant

    long-run impact on LLP/TL are the growth of M1, the rise of Solar price and theSBI rate. For model using NPL/TL as dependent variable, only the inflation andSBI rate have significant long-run impact.

    The inflation also has the most significant long-run impact on D(NPL/TL),followed by the rise of Premium price. For model with G(LLP) as dependent variable, the growth of M2 is the most significant variable, followed by theinflation, the rise of Solar price and the rise of Premium price while other variables are not significant.

    Flexible

    In flexible form regression, the growth of M2 has the largest long-run impact forfour models, i.e. model using LLP/TL. NPL/TL, D(NPL/TL) and G(LLP) asdependent variables. The inflation also has significant impact on LLP/TL andNPL/TL.

  • 8/6/2019 exposicion sab

    22/57

    22

    The other significant variables in this multivariate flexible model are the growth of GDP, the growth of IHSG, the SBI rate, the growth of Premium price and the riseof Solar price. The growth of GDP and the growth of IHSG have significant long-run impact on NPL/TL and G(LLP). The rise of Solar price has significant long-run on LLP/TL and G(LLP) while the SBI rate only have significant impact onD(NPL/TL) and the rise of Premium price on NPL/TL.

    5. Conclusion

    The result of this study is sensitive to the construction of dependent variablesused as the proxy of credit risk. Some of dependent variables are not appropriateto be applied in analyzing the relationship between macroeconomic shocks andthe credit risk of large banks. The regressions using G(LLP)/TL as dependent variable provide unfavorable result. For multivariate regression, D(LLP/TL) is

    also not appropriate to employ as a proxy of credit risk. None of themacroeconomic variables has significant long-run impact on it.

    The univariate regressions show that the price stability indicators, i.e. the growthof money and the inflation, play great role in large banks financial stability, interm of credit risk, in Indonesia. In most regressions, those variables havesignificant long-run impact on the proxies of credit risk used. The condition of economy also significantly affects the credit risk of large banks, hence it isimportant to foster the growth of GDP as a cyclical indicator. The rise of oil pricesshould be taken into consideration in maintaining the banks financial stability.Based on the result of univariate regression, the rise of Premium price and Solar price have large long-run impact on some dependent variables which reflect creditrisk.

    To summarize, the inflation is the most-often significant variable in themultivariate fixed effect model. It implies that the price stability indicator hassignificant relation with credit risk. The result of the multivariate flexible modelregression also suggests the same conclusion, except that the most significant variable is not the inflation but the growth of M2. Overall, the result of themultivariate regression suggests the importance of price stability in order tomaintain financial stability in term of credit quality.

  • 8/6/2019 exposicion sab

    23/57

    23

    APPENDICES

    I. UNIVARIATE REGRESSION

    a. Long-Run Coefficient

    Table 3Long Run Impact of Macroeconomic Variables Change to NPL/TL

    (17 Banks - Fixed Effect)Independent Variables Coeff Std Error t-value 1-tail ProbG(GDP) -23.14 3.18 -7.27 0.00G(M1) 14.26 1.79 7.99 0.00G(M2) 15.39 1.76 8.76 0.00G(IHSG) -5.50 0.78 -7.02 0.00INF 17.31 1.90 9.13 0.00SBI 1.55 0.15 10.11 0.00G(EXRATE) 4.54 0.77 5.88 0.00G(PREMIUMDN) 11.21 1.95 5.73 0.00G(SOLARDN) 1.94 0.95 2.04 0.02

    Table 4Long Run Impact of Macroeconomic Variables Change to G(LLP)

    (17 Banks - Fixed Effect )Independent Variables Coeff Std Error t-value 1-tail ProbG(GDP) 0.31 1.18 0.26 0.40G(M1) 4.07 0.99 4.12 0.00G(M2) 3.77 0.74 5.10 0.00G(IHSG) -3.05 0.56 -5.49 0.00INF 1.76 1.45 1.22 0.11

    SBI -0.16 0.20 -0.78 0.22G(EXRATE) 3.05 0.48 6.30 0.00G(PREMIUMDN) 6.22 1.04 6.00 0.00G(SOLARDN) -0.62 0.21 -2.89 0.00

  • 8/6/2019 exposicion sab

    24/57

    24

    Table 5Long Run Impact of Macroeconomic Variables Change to Credit Risk

    (15 Banks - Fixed Effect)LLP/TL NPL/TL D (LLP/TL)

    Independent Variables Coeff StdError t-value Coeff StdError t-value Coeff StdError t-value

    G(GDP) -1.88 0.25 -7.62 -20.76 2.23 -9.29 -0.53 0.11 -4.83G(M1) 3.47 0.54 6.48 1.35 0.12 11.54 0.34 0.07 4.97G(M2) 3.81 0.61 6.25 1.44 0.12 11.66 0.38 0.08 5.00G(IHSG) -0.43 0.12 -3.72 -0.31 0.05 -6.64 -0.01 0.02 -0.69INF 4.22 0.63 6.69 11.80 1.18 9.97 0.31 0.06 5.57SBI 0.39 0.05 7.73 0.88 0.88 0.88 0.02 0.01 2.45G(EXRATE) -0.12 0.18 -0.65 3.57 0.60 5.96 0.17 0.03 6.52G(PREMIUMDN) 1.39 0.38 3.64 8.30 1.81 4.58 0.30 0.06 4.87G(SOLARDN) 1.48 0.40 3.65 4.82 1.36 3.55 0.06 0.04 1.67

    D (NPL/TL) G (LLP) G (LLP/TL)Independent

    Variables Coeff StdError t-value Coeff StdError t-value Coeff StdError t-valueG(GDP) -0.52 0.19 -2.78 0.47 1.30 0.37 - - -G(M1) 0.59 0.09 6.74 0.17 0.75 0.23 -2.58 1.15 -2.25G(M2) 0.79 0.11 7.33 1.99 1.04 1.90 -2.91 1.26 -2.3G(IHSG) -0.20 0.03 -7.03 -0.53 0.35 -1.49 2.76 0.64 4.33INF 0.69 0.10 6.85 3.33 1.03 3.24 1.14 2.11 0.54SBI 0.03 0.01 2.07 -0.23 0.14 -1.61 0.01 0.23 0.02G(EXRATE) 0.23 0.03 7.88 2.16 0.51 4.22 -0.15 0.45 -0.34G(PREMIUMDN) 0.07 0.06 1.09 -0.46 0.45 -1.01 -0.97 0.62 -1.57G(SOLARDN) -0.05 0.05 -1.11 -0.70 0.24 -2.87 -0.67 0.54 -1.25

  • 8/6/2019 exposicion sab

    25/57

    25

    Table 6Long Run Impact of Macroeconomic Variable Changes to Credit Risk

    (15 Banks - Flexible Form)

    LLP/TL NPL/TL D (LLP/TL)Independent

    Variables Coeff StdError t-value Coeff StdError t-value Coeff StdError t-valueG(GDP) -8.52 1.48 -5.75 -16.29 1.90 -8.58 -0.51 0.09 -5.36G(M1) 3.82 0.80 4.76 8.60 0.66 13.03 0.24 0.04 5.84G(M2) 2.11 0.57 3.68 8.89 0.68 13.01 0.32 0.05 5.88G(IHSG) -0.33 0.09 -3.75 -0.84 0.20 -4.14 -0.03 0.01 -3.04INF 2.28 0.71 3.20 10.98 0.93 11.83 0.22 0.04 6.12SBI 0.23 0.07 3.58 0.88 0.07 13.08 0.02 0.01 3.47G(EXRATE) -0.77 0.15 -5.21 2.04 0.47 4.33 0.14 0.02 6.30G(PREMIUMDN) 0.24 0.19 1.25 -0.91 0.42 -2.14 0.13 0.03 4.36G(SOLARDN)DN 0.40 0.22 1.81 0.43 0.45 0.95 -0.01 0.02 -0.62

    D (NPL/TL) G (LLP) G (LLP/TL)Independent

    Variables Coeff StdError t-value Coeff StdError t-value Coeff StdError t-valueG(GDP) -0.52 0.14 -3.64 -0.52 0.92 -0.57 0.30 0.46 0.66G(M1) 0.56 0.06 8.77 1.29 0.61 2.12 0.01 0.28 0.05G(M2) 0.64 0.07 8.89 2.16 0.59 3.67 0.04 0.31 0.13G(IHSG) -0.18 0.03 -7.13 -0.43 0.30 -1.43 0.19 0.08 2.54INF 0.60 0.07 9.13 - - - 0.33 0.38 0.85SBI 0.02 0.01 3.47 0.02 0.06 0.31 0.01 0.03 0.26G(EXRATE) 0.16 0.02 7.43 2.23 0.41 5.49 0.30 0.13 2.32G(PREMIUMDN) 0.09 0.04 1.98 0.28 0.40 0.70 0.49 0.22 2.22G(SOLARDN) 0.02 0.04 0.56 0.87 0.34 2.58 -0.01 0.14 -0.10

    b. Sensitivity TestTable 7

    Sensitivity Test NPL/TL 17 Banks (Fixed Effect)

    NPL/TL

    Independent Variables

    Base Line(2004:01-2005:05)

    WorstCase 1996 1997 1998

    G(GDP) -20.06 159.93 -25.86 -19.77 38.16G(M1) 7.68 362.44 33.00 25.01 64.76G(M2) 8.61 369.41 30.73 28.88 62.47

    G(IHSG) -14.66 208.33 -19.07 21.17 0.42INF 10.07 207.75 8.44 14.15 82.87SBI 11.59 109.08 20.68 24.38 76.59G(EXRATE) 3.28 381.04 1.24 31.31 14.86G(PREMIUMDN) 18.60 506.51 - - 33.31G(SOLARDN) 2.75 97.18 - - 5.98

  • 8/6/2019 exposicion sab

    26/57

    26

    Table 8Sensitivity Test G(LLP) 17 Banks (Fixed Effect)

    G(LLP)Independent Variables

    Base Line2004:01-2005:05)

    WorstCase 1996 1997 1998

    G(GDP) n.s. n.s n.s. n.s. n.s.G(M1) 2.19 103.49 9.42 7.14 18.49G(M2) 2.11 90.49 7.53 7.07 15.30G(IHSG) -8.13 115.59 -10.58 11.75 0.23INF n.s. n.s. n.s. n.s. n.s.SBI n.s. n.s n.s. n.s. n.s.G(EXRATE) 2.20 256.06 0.83 21.04 9.98G(PREMIUMDN) 10.33 281.24 - - 18.49G(SOLARDN) n.s. n.s. n.s. n.s. n.s.

    Table 9Sensitivity Test LLP/TL 15 Banks (Fixed Effect)

    LLP/TLIndependent Variables

    Base Line (2004:01-2005:05)

    WorstCase 1996 1997 1998

    G(GDP) -1.63 12.97 -2.10 -1.60 3.09

    G(M1) 1.87 88.18 8.03 6.08 15.76

    G(M2) 2.14 81.57 7.62 7.16 15.48

    G(IHSG) -1.16 16.44 -1.50 1.67 0.03

    INF 2.46 50.68 2.06 3.45 20.22

    SBI 2.96 27.82 5.27 6.22 19.53

    G(EXRATE) n.s. n.s. n.s. n.s. n.s.

    G(PREMIUMDN) 2.30 62.72 - - 4.12G(SOLARDN) 2.10 73.99 - - 4.55

    Table 10Sensitivity Test NPL/TL 15 Banks (Fixed Effect)

    NPL/TL

    Independent Variables

    Base Line(2004:01-2005:05)

    WorstCase 1996 1997 1998

    G(GDP) -18.00 143.51 -23.20 -17.74 34.24G(M1) 0.73 34.37 3.13 2.37 6.14G(M2) 0.81 34.54 2.87 2.70 5.84G(IHSG) -0.82 11.71 -1.07 1.19 0.02INF 6.87 141.64 5.75 9.65 56.5SBI n.s. n.s. n.s. n.s. n.s.G(EXRATE) 2.57 299.22 0.97 24.59 11.67G(PREMIUMDN) 13.77 375.12 - - 24.67G(SOLARDN) 6.84 241.44 - - 14.86

  • 8/6/2019 exposicion sab

    27/57

    27

    Table 11Sensitivity Test D(LLP/TL )15 Banks (Fixed Effect)

    D(LLP/TL)

    Independent

    Variables

    Base Line(2004:01-

    2005:05)

    Worst

    Case 1996 1997 1998G(GDP) -0.46 3.67 -0.59 -0.45 0.88G(M1) 0.18 8.58 0.78 0.59 1.53G(M2) 0.22 9.24 0.77 0.72 1.56G(IHSG) n.s. n.s. n.s. n.s. n.s.INF 0.18 3.70 0.15 0.25 1.48SBI 0.18 1.69 0.32 0.38 1.19G(EXRATE) 0.13 14.59 0.05 1.20 0.57G(PREMIUMDN) 0.50 13.72 - - 0.90G(SOLARDN) 0.09 3.23 - - 0.20

    Table 12Sensitivity Test D(NPL/TL )15 Banks (Fixed Effect)

    D(NPL/TL)

    Independent Variables

    Base Line(2004:01-2005:05)

    WorstCase 1996 1997 1998

    G(GDP) -0.45 3.62 -0.59 -0.45 0.86G(M1) 0.32 15.04 1.37 1.04 2.69G(M2) 0.44 19.01 1.58 1.49 3.22G(IHSG) -0.52 7.45 -0.68 0.76 0.02INF 0.40 8.28 0.34 0.56 3.30SBI 0.21 1.95 0.37 0.44 1.37G(EXRATE) 0.16 19.03 0.06 1.56 0.74G(PREMIUMDN) n.s. n.s. n.s. n.s. n.s.G(SOLARDN) n.s. n.s. n.s. n.s. n.s.

    Table 13Sensitivity Test G(LLP)15 Banks (Fixed Effect)

    G(LLP)

    Independent Variables

    Base Line

    (2004:01-2005:05) WorstCase 1996 1997 1998G(GDP) n.s. n.s. n.s. n.s. n.s.G(M1) n.s. n.s. n.s. n.s. n.s.G(M2) 1.11 47.72 3.97 3.73 8.07G(IHSG) n.s. n.s. n.s. n.s. n.s.INF 1.94 39.94 1.62 2.72 15.93SBI n.s. n.s. n.s. n.s. n.s.

  • 8/6/2019 exposicion sab

    28/57

    28

    G(LLP)

    Independent Variables

    Base Line(2004:01-2005:05)

    WorstCase 1996 1997 1998

    G(EXRATE) 1.56 181.43 0.59 14.91 7.07G(PREMIUMDN) n.s. n.s. n.s. n.s. n.s.G(SOLARDN) n.s. n.s. n.s. n.s. n.s.

    Table 14Sensitivity Test G(LLP)/TL 15 Banks (Fixed Effect)

    G(LLP)/TLIndependent Variables

    Base Line (2004:01-2005:05) Worst Case 1996 1997 1998

    G(GDP) - - - - -G(M1) n.s. n.s. n.s. n.s. n.s.G(M2) n.s. n.s. n.s. n.s. n.s.

    G(IHSG) n.s. n.s. n.s. n.s. n.s.INF n.s. n.s. n.s. n.s. n.s.SBI n.s. n.s. n.s. n.s. n.s.G(EXRATE) n.s. n.s. n.s. n.s. n.s.G(PREMIUMDN) n.s. n.s. n.s. n.s. n.s.G(SOLARDN) n.s. n.s. n.s. n.s. n.s.

    Table 15Sensitivity Test LLP/TL 15 Banks (Flexible Form)

    LLP/TL

    Independent Variables Base Line (2004:01-2005:05) WorstCase 1996 1997 1998G(GDP) -7.38 58.87 -9.52 -7.28 14.05G(M1) 2.06 97.09 8.84 6.70 17.35G(M2) 1.18 50.73 4.22 3.97 8.58G(IHSG) -0.87 12.40 -1.13 1.26 0.03INF 1.33 27.38 1.11 1.86 10.92SBI 1.75 16.50 3.13 3.69 11.59G(EXRATE) n.s. n.s. n.s. n.s. n.s.G(PREMIUMDN) n.s. n.s. n.s. n.s. n.s.G(SOLARDN) 0.56 19.91 - - 1.23

    Table 16Sensitivity Test NPL/TL 15 Bank (Flexible Form)

    NPL/TLIndependent Variables

    Base Line (2004:01-2005:05)

    WorstCase 1996 1997 1998

    G(GDP) -14.12 112.62 -18.21 -13.92 26.87

  • 8/6/2019 exposicion sab

    29/57

    29

    NPL/TLIndependent Variables

    Base Line (2004:01-2005:05)

    WorstCase 1996 1997 1998

    G(M1) 4.64 218.70 19.92 15.09 39.08G(M2) 4.98 213.47 17.76 16.69 36.1

    G(IHSG) -2.23 31.74 -2.91 3.23 0.06INF 6.39 131.81 5.35 8.98 52.58SBI 6.61 62.20 11.79 13.90 43.67G(EXRATE) 1.47 170.87 0.55 14.04 6.66G(PREMIUMDN) n.s. n.s. n.s. n.s. n.s.G(SOLARDN) n.s. n.s. n.s. n.s. n.s.

  • 8/6/2019 exposicion sab

    30/57

    30

    Table 17 Sensitivity Test D(LLP/TL) 15 Banks (Flexible Form)

    D(LLP/TL)

    Independent

    Variables

    Base Line(2004:01-

    2005:05)

    Worst

    Case 1996 1997 1998G(GDP) -0.44 3.51 -0.57 -0.43 0.84G(M1) 0.13 6.14 0.56 0.42 1.10G(M2) 0.18 7.58 0.63 0.59 1.28G(IHSG) -0.08 1.12 -0.10 0.11 0.00INF 0.13 2.66 0.11 0.18 1.06SBI 0.14 1.30 0.25 0.29 0.91G(EXRATE) 0.10 11.47 0.04 0.94 0.45G(PREMIUMDN) 0.22 6.07 - - 0.40G(SOLARDN) n.s. n.s. n.s. n.s. n.s.

    Table 18Sensitivity Test D(NPL/TL) 15 Banks (Flexible Form)

    DNPL/TL

    Independent Variables

    Base Line(2004:01-2005:05)

    WorstCase 1996 1997 1998

    G(GDP) -0.45 3.59 -0.58 -0.44 0.86G(M1) 0.30 14.24 1.30 0.98 2.54G(M2) 0.36 15.43 1.28 1.21 2.61G(IHSG) -0.48 6.80 -0.62 0.69 0.01

    INF 0.35 7.14 0.29 0.49 2.85SBI 0.14 1.30 0.25 0.29 0.91G(EXRATE) 0.11 13.34 0.04 1.10 0.52G(PREMIUMDN) 0.15 3.98 - - 0.26G(SOLARDN) n.s. n.s. n.s. n.s. n.s.

    Table 19Sensitivity Test G(LLP) 15 Banks (Flexible Form)

    G(LLP)

    Independent Variables

    Base Line(2004:01-2005:05)

    WorstCase 1996 1997 1998

    G(GDP) n.s. n.s. n.s. n.s. n.s.G(M1) 0.70 32.90 3.00 2.27 5.88G(M2) 1.21 51.94 4.32 4.06 8.78G(IHSG) n.s. n.s. n.s. n.s. n.s.INF - - - - -SBI n.s. n.s. n.s. n.s. n.s.

  • 8/6/2019 exposicion sab

    31/57

    31

    G(LLP)

    Independent Variables

    Base Line(2004:01-2005:05)

    WorstCase 1996 1997 1998

    G(EXRATE) 1.61 186.92 0.61 15.36 7.29

    G(PREMIUMDN) n.s. n.s. n.s. n.s. n.s.G(SOLARDN) 1.23 43.53 - - 2.68

    Table 20Sensitivity Test G(LLP)/TL 15 Banks (Flexible Form)

    G(LLP)/TL

    Independent Variables

    Base Line(2004:01-2005:05)

    WorstCase 1996 1997 1998

    G(GDP) n.s. n.s. n.s. n.s. n.s.G(M1) n.s. n.s. n.s. n.s. n.s.

    G(M2) n.s. n.s. n.s. n.s. n.s.G(IHSG) n.s. n.s. n.s. n.s. n.s.INF n.s. n.s. n.s. n.s. n.s.SBI n.s. n.s. n.s. n.s. n.s.G(EXRATE) 0.22 25.29 0.08 2.08 0.99G(PREMIUMDN) 0.82 22.25 - - 1.46G(SOLARDN) n.s. n.s. n.s. n.s. n.s.

  • 8/6/2019 exposicion sab

    32/57

    32

    II. MULTIVARIATE REGRESSION (9 Macroeconomic Variables)

    a. Long-Run Coefficient

    Table 21Long Run Impact of 9 Macroeconomic Variables Change to Credit Risk

    15 Banks - Fixed Effect

    LLP/TL NPL/TL D (LLP/TL)Independent Variables Coeff StdError t-value Coeff StdError t-value Coeff StdError t-value

    G(GDP) 1.43 0.47 3.06 2.44 1.97 1.24 - - -G(M1) 0.55 0.24 2.34 3.31 3.15 1.05 - - -G(M2) - - - -2.02 3.26 -0.62 - - -G(IHSG) - - - 0.12 0.20 0.58 - - -INF -1.17 0.52 -2.24 9.54 2.74 3.48 - - -SBI 0.48 0.07 6.57 0.44 0.23 1.97 - - -G(EXRATE) - - - -0.83 0.29 -2.90 - - -

    G(PREMIUMDN) - - - -0.78 0.32 -2.43 - - -G(SOLARDN) 0.53 0.14 3.92 - - - - - -

    D (NPL/TL) G (LLP) G (LLP/TL)Independent Variables Coeff StdError t-value Coeff StdError t-value Coeff StdError t-value

    G(GDP) - - - -0.17 4.07 -0.04 - - -G(M1) -0.22 0.10 -2.29 -112.33 12.60 -8.92 - - -G(M2) - - - 114.40 13.95 8.20 - - -G(IHSG) - - - -0.01 0.51 -0.01 - - -INF 1.45 0.18 7.86 22.29 9.27 2.40 - - -SBI -0.03 0.02 -1.49 -0.36 0.71 -0.50 - - -G(EXRATE) -0.01 0.03 -0.22 -4.75 1.57 -3.02 - - -G(PREMIUMDN) 0.10 0.03 3.12 1.08 0.56 1.92 - - -G(SOLARDN) - - - 2.74 0.69 4.00 - - -

    Table 22Long Run Impact of 9 Macroeconomic Variables Change to Credit Risk

    15 Bank - Flexible FormLLP/TL NPL/TL D(LLP/TL)

    Independent Variables Coeff StdError t-value Coeff StdError t-value Coeff StdError t-value

    G(GDP) - - - -3.08 0.95 -3.25 - - -G(M1) -5.52 1.91 -2.89 -3.90 1.80 -2.17 - - -G(M2) 8.36 2.35 3.56 9.75 2.34 4.17 - - -G(IHSG) - - - -0.42 0.21 -1.96 - - -INF 4.44 1.07 4.15 3.84 1.22 3.14 - - -SBI -0.15 0.14 -1.07 -0.03 0.15 -0.19 - - -G(EXRATE) - - - - - - - - -

  • 8/6/2019 exposicion sab

    33/57

    33

    LLP/TL NPL/TL D(LLP/TL)Independent

    Variables Coeff StdError t-value Coeff StdError t-value Coeff StdError t-value

    G(PREMIUMDN) - - - 0.79 0.25 3.21 - - -

    G(SOLARDN) 0.57 0.18 3.14 - - - - - -

    D(NPL/TL) G(LLP) G(LLP/TL)Independent

    Variables Coeff StdError t-value Coeff StdError t-value Coeff StdError t-value

    G(GDP) - - - -11.51 2.14 -5.38 - - -G(M1) -0.97 0.20 -4.84 -29.87 6.00 -4.98 - - -G(M2) 0.76 0.18 4.09 64.80 8.40 7.71 - - -G(IHSG) 0.04 0.01 2.77 -1.02 0.30 -3.41 - - -INF - - - -10.04 3.13 -3.21 - - -SBI 0.13 0.02 8.59 -1.60 0.30 -5.30 - - -G(EXRATE) - - - -3.81 0.70 -5.44 - - -G(PREMIUMDN) - - - 0.72 0.54 1.34 - - -

    G(SOLARDN) - - - 2.20 0.41 5.36 - - -

    b. Scenario Analysis

    Table 23Scenario Analysis LLP/TL 15 Banks (Fixed Effect)

    LLP/TL

    Independent Variables

    Base Line(2004:01-2005:05)

    WorstCase 1996 1997 1998

    G(GDP) n.s. n.s.n.s. n.s. n.s.

    G(M1) 0.30 14.05 1.28 0.97 2.51G(M2) - - - - -G(IHSG) - - - - -INF n.s. n.s. n.s. n.s. n.s.

    SBI 3.59 33.75 6.40 7.54 23.69G(EXRATE) - - - - -G(PREMIUMDN) - - - - -G(SOLARDN) 0.76 26.76 - - 1.65

    Table 24Scenario Analysis NPL/TL 15 Banks (Fixed Effect)

    NPL/TL

    Independent Variables

    Base Line(2004:01-2005:05)

    WorstCase 1996 1997 1998

    G(GDP) n.s. n.s. n.s. n.s. n.s.

    G(M1) n.s. n.s. n.s. n.s. n.s.

  • 8/6/2019 exposicion sab

    34/57

    34

    NPL/TL

    Independent Variables

    Base Line(2004:01-2005:05)

    WorstCase 1996 1997 1998

    G(M2) n.s. n.s. n.s. n.s. n.s.

    G(IHSG) n.s. n.s. n.s. n.s. n.s.INF 5.55 114.51 4.65 7.80 45.67SBI 3.33 31.28 5.93 6.99 21.96G(EXRATE) n.s. n.s. n.s. n.s. n.s.

    G(PREMIUMDN) n.s. n.s. n.s. n.s. n.s.

    G(SOLARDN) - - - - -

    Table 25Scenario Analysis D(LLP/TL) 15 Banks (Fixed Effect)

    D(LLP/TL)

    Independent Variables

    Base Line(2004:01-2005:05)

    WorstCase 1996 1997 1998

    G(GDP) - - - - -G(M1) - - - - -G(M2) - - - - -G(IHSG) - - - - -INF - - - - -SBI - - - - -G(EXRATE) - - - - -G(PREMIUMDN) - - - - -G(SOLARDN) - - - - -

    Table 26Scenario Analysis D(NPL/TL) 15 Banks (Fixed Effect)

    D(NPL/TL)

    Independent Variables

    Base Line(2004:01-2005:05)

    WorstCase 1996 1997 1998

    G(GDP) - - - - -G(M1) n.s. n.s. n.s. n.s. n.s.

    G(M2) - - - - -G(IHSG) - - - - -INF 0.84 17.38 0.71 1.18 6.93SBI n.s. n.s. n.s. n.s. n.s.

    G(EXRATE) n.s. n.s. n.s. n.s. n.s.

    G(PREMIUMDN) 0.17 4.50 - - 0.30G(SOLARDN) - - - - -

  • 8/6/2019 exposicion sab

    35/57

    35

    Table 27 Scenario Analysis G(LLP) 15 Banks (Fixed Effect)

    G(LLP)

    Independent

    Variables

    Base Line(2004:01-

    2005:05)

    Worst

    Case 1996 1997 1998G(GDP) n.s. n.s. n.s. n.s. n.s.

    G(M1) n.s. n.s. n.s. n.s. n.s.

    G(M2) 64.03 2746.40 228.48 214.68 464.41G(IHSG) n.s. n.s. n.s. n.s. n.s.

    INF 12.97 267.58 10.87 18.22 106.73SBI n.s. n.s. n.s. n.s. n.s.

    G(EXRATE) n.s. n.s. n.s. n.s. n.s.

    G(PREMIUMDN) 1.80 49.04 - - 3.22G(SOLARDN) 3.89 137.17 - - 8.44

    Table 28Scenario Analysis G(LLP)/TL 15 Banks (Fixed Effect)

    G(LLP)/TL

    Independent Variables

    Base Line(2004:01-2005:05)

    WorstCase 1996 1997 1998

    G(GDP) - - - - -G(M1) - - - - -G(M2) - - - - -G(IHSG) - - - - -INF - - - - -SBI - - - - -G(EXRATE) - - - - -G(PREMIUMDN) - - - - -G(SOLARDN) - - - - -

  • 8/6/2019 exposicion sab

    36/57

    36

    Table 29Scenario Analysis LLP/TL 15 Banks (Flexible Form)

    LLP/TLIndependent

    VariablesBase Line

    (2004:01-2005:05) Worst

    Case 1996 1997 1998

    G(GDP) - - - - -G(M1) n.s. n.s. n.s. n.s. n.s.

    G(M2) 4.68 200.73 16.70 15.69 33.94G(IHSG) - - - - -INF 2.59 53.34 2.17 3.63 21.28SBI n.s. n.s. n.s. n.s. n.s.

    G(EXRATE) - - - - -G(PREMIUMDN) - - - - -G(SOLARDN) 0.80 28.36 - - 1.74

    Table 30Scenario Analysis NPL/TL 15 Banks (Flexible Form)

    NPL/TL

    Independent Variables

    Base Line(2004:01-2005:05)

    WorstCase 1996 1997 1998

    G(GDP) -2.67 21.26 -3.44 -2.63 -2.67G(M1) n.s. n.s. n.s. n.s. n.s.

    G(M2) 5.46 233.98 19.46 18.29 39.56G(IHSG) -1.12 -10.51 -1.46 1.62 0.03

    INF 2.23 46.08 1.87 3.14 18.38SBI n.s. n.s. n.s. n.s. n.s.

    G(EXRATE) - - - - -G(PREMIUMDN) 1.32 35.84 - - 2.36G(SOLARDN) - - - - -

  • 8/6/2019 exposicion sab

    37/57

    37

    Table 31Scenario Analysis D(LLP/TL) 15 Banks (Flexible Form)

    D(LLP/TL)

    Independent

    Variables

    Base Line(2004:01-

    2005:05)

    Worst

    Case 1996 1997 1998G(GDP) - - - - -G(M1) - - - - -G(M2) - - - - -G(IHSG) - - - - -INF - - - - -SBI - - - - -G(EXRATE) - - - - -G(PREMIUMDN) - - - - -G(SOLARDN) - - - - -

    Table 32Scenario Analysis D(NPL/TL) 15 Banks (Flexible Form)

    D(NPL/TL)

    Independent Variables

    Base Line(2004:01-2005:05)

    WorstCase 1996 1997 1998

    G(GDP) - - - - -G(M1) n.s. n.s. n.s. n.s. n.s.

    G(M2) 0.42 18.13 1.51 1.42 3.07G(IHSG) n.s. n.s. n.s. n.s. n.s.

    INF - - - - -SBI 0.97 9.14 1.73 2.04 6.41G(EXRATE) - - - - -G(PREMIUMDN) - - - - -G(SOLARDN) - - - - -

  • 8/6/2019 exposicion sab

    38/57

    38

    Table 33Scenario Analysis G(LLP) 15 Banks (Flexible Form)

    G(LLP)

    Independent

    Variables

    Base Line(2004:01-

    2005:05)

    Worst

    Case 1996 1997 1998G(GDP) -9.98 79.55 -12.86 -9.83 18.98G(M1) n.s. n.s. n.s. n.s. n.s.

    G(M2) 36.27 1555.67 129.42 121.60 263.06G(IHSG) -2.71 38.48 -3.52 3.91 0.08INF n.s. n.s. n.s. n.s. n.s.

    SBI n.s. n.s. n.s. n.s. n.s.

    G(EXRATE) n.s. n.s. n.s. n.s. n.s.

    G(PREMIUMDN) n.s. n.s. n.s. n.s. n.s.

    G(SOLARDN) 3.12 109.99 - - 6.77

    Table 34Scenario Analysis G(LLP)/TL 15 Banks (Flexible Form)

    G(LLP)/TL

    Independent Variables

    Base Line(2004:01-2005:05)

    WorstCase 1996 1997 1998

    G(GDP) - - - - -G(M1) - - - - -G(M2) - - - - -G(IHSG) - - - - -INF - - - - -SBI - - - - -G(EXRATE) - - - - -G(PREMIUMDN) - - - - -G(SOLARDN) - - - - -

  • 8/6/2019 exposicion sab

    39/57

    39

    III. MULTIVARIATE REGRESSION (8 Macroeconomic Variables)

    a. Long-Run Coefficient

    Table 35Long Run Impact of 8 Macroeconomic Variables Change

    (Without M1) to Credit Risk (15 Banks - Fixed Effect)

    LLP/TL NPL/TL D (LLP/TL)Independent

    Variables Coeff StdError t-value Coeff StdError t-value Coeff StdErrort-

    valueG(GDP) - - - 0.95 0.64 4.79 - - -G(M2) 1.02 0.37 2.73 4.79 1.25 3.83 - - -G(IHSG) - - - 0.03 0.20 0.13 - - -

    INF-

    3.06 0.87 -3.52 14.23 3.34 4.27 - - -SBI 0.48 0.09 5.45 -0.25 0.28 -0.89 0.03 0.01 4.32G(EXRATE) 0.27 0.09 2.92 -1.36 0.36 -3.73 - - -

    G(PREMIUMDN)-

    0.66 0.24 -2.73 0.64 0.47 1.35 - - -G(SOLARDN) 0.93 0.22 4.24 - - - - - -

    D (NPL/TL) G (LLP) G (LLP/TL)Independent

    Variables Coeff StdError t-value Coeff StdError t-value Coeff StdErrort-

    value

    G(GDP)-

    0.26 0.09 -2.87 17.21 3.39 5.08 - - -G(M2) 0.37 0.12 3.04 1.62 4.63 0.35 - - -G(IHSG) - - - -0.20 0.51 -0.39 - - -INF 1.43 0.18 7.78 7.08 6.88 1.03 - - -

    SBI-

    0.07 0.02 -3.60 -0.31 0.57 -0.54 - - -G(EXRATE) -0.16 0.03 -5.07 -10.26 1.56 -6.56 - - -G(PREMIUMDN) 0.16 0.03 4.79 2.68 1.21 2.22 - - -G(SOLARDN) - - - -3.60 0.98 -3.66 - - -

  • 8/6/2019 exposicion sab

    40/57

    40

    Table 36Long Run Impact of 8 Macroeconomic Variables Change

    (Without GM1) to Credit Risk (15 Banks - Flexible Form)LLP/TL NPL/TL D (LLP/TL)

    Independent Variables Coeff StdError t-value Coeff StdError t-value Coeff StdError t-val

    G(GDP) 2.18 0.84 2.58 -1.21 0.76 -1.60 - - -G(M2) -4.98 1.48 -3.37 1.37 0.37 3.70 - - -G(IHSG) 0.68 0.23 3.02 0.07 0.17 0.44 - - -INF 5.45 1.48 3.67 -0.76 1.18 -0.65 - - -SBI 0.49 0.12 3.98 0.74 0.10 7.40 - - -G(EXRATE) 0.88 0.30 2.93 - - - 0.02 0.01 3.32G(PREMIUMDN) -1.41 0.37 -3.85 -0.63 0.21 -3.07 - - -G(SOLARDN) 0.75 0.35 2.13 -0.59 0.18 -3.34 0.03 0.01 1.97

    D (NPL/TL) G (LLP) G (LLP/TL)Independent Variables Coeff StdError t-value Coeff StdError t-value Coeff StdError t-value

    G(GDP) -0.26 0.06 -4.56 -13.17 3.10 -4.25 - - -G(M2) - - - -1.86 2.89 -0.64 - - -G(IHSG) -0.08 0.02 -4.74 -2.62 0.42 -6.22 - - -INF 0.51 0.12 4.26 29.71 5.35 5.55 - - -SBI -0.02 0.01 -1.94 -3.63 0.63 -5.75 - - -G(EXRATE) - - - 1.64 0.95 1.73 - - -G(PREMIUMDN) - - - -1.88 0.49 -3.88 - - -

    G(SOLARDN) - - - -0.91 0.23 -3.91 - - -

  • 8/6/2019 exposicion sab

    41/57

    41

    Table 37 Long Run Impact of 8 Macroeconomic Variables Change

    (Without GM2) to Credit Risk (15 Banks - Fixed Effect)

    LLP/TL NPL/TL D (LLP/TL)Independent

    Variables Coeff StdError t-value Coeff StdError t-value Coeff StdError t-valueG(GDP) - - - -2.43 2.27 -1.07 -0.21 0.62 -0.34G(M1) - - - 2.77 0.64 4.34 2.08 1.02 2.03G(IHSG) 0.29 0.12 2.33 -0.13 0.24 -0.54 0.00 0.09 -0.06INF 3.99 0.69 5.77 10.08 2.72 3.71 0.65 1.26 0.51SBI - - - -0.16 0.25 -0.63 -0.20 0.10 -1.95G(EXRATE) - - - - - - -0.25 0.20 -1.25G(PREMIUMDN) -2.02 0.39 -5.19 -0.20 0.49 -0.41 -0.09 0.19 -0.50

    G(SOLARDN) 1.02 0.27 3.74 - - - 0.31 0.15 2.07

    D (NPL/TL) G (LLP) G (LLP/TL)Independent

    Variables Coeff StdError t-value Coeff StdError t-value Coeff StdError t-valueG(GDP) -0.26 0.09 -2.88 19.98 4.85 4.12 - - -G(M1) 0.37 0.11 3.49 -27.36 4.37 -6.26 - - -G(IHSG) - - - 4.00 0.80 5.00 - - -INF 1.48 0.18 8.01 14.95 6.25 2.39 - - -SBI -0.08 0.02 -3.91 2.11 0.61 3.48 - - -G(EXRATE) -0.17 0.03 -5.51 -3.18 1.05 -3.02 - - -G(PREMIUMDN) 0.16 0.03 4.90 2.72 0.65 4.19 - - -

    G(SOLARDN) - - - 0.79 0.81 0.97 - - -

  • 8/6/2019 exposicion sab

    42/57

    42

    Table 38Long Run Impact of 8 Macroeconomic Variables Change

    (Without GM2) to Credit Risk (15 Banks - Flexible Form)

    LLP/TL NPL/TL D (LLP/TL)

    Independent Variables Coeff StdError t- value Coeff StdError t-value Coeff StdError vG(GDP) - - - 2.17 1.24 1.76 - - -G(M1) 0.85 0.27 3.16 - - - - - -G(IHSG) - - - 0.15 0.18 0.85 - - -INF - - - 0.29 1.36 0.21 - - -SBI 0.35 0.06 5.46 0.90 0.16 5.73 0.01 0.00 2.68G(EXRATE) 0.20 0.07 2.73 0.30 0.13 2.36 0.02 0.01 3.12G(PREMIUMDN) 1.17 0.28 4.14 -1.89 0.38 -4.95 - - -G(SOLARDN) 0.61 0.26 2.33 - - - - - -

    D (NPL/TL) G (LLP) G (LLP/TL)Independent Variables Coeff StdError

    t- value Coeff StdError t-value Coeff StdError v

    G(GDP) - - - 0.74 2.62 0.28 - - -G(M1) - - - -1.49 1.08 -1.38 - - -G(IHSG) - - - 1.49 0.38 3.91 - - -INF 0.60 0.07 8.16 9.62 3.07 3.13 - - -SBI - - - -1.14 0.36 -3.18 - - -G(EXRATE) - - - 0.35 0.19 1.84 - - -G(PREMIUMDN) -0.13 0.02 -6.90 -0.59 0.32 -1.81 - - -G(SOLARDN) - - - 1.23 0.22 5.68 - - -

  • 8/6/2019 exposicion sab

    43/57

    43

    b. Scenario AnalysisTable 39

    Scenario Analysis (Without GM1) LLP/TL 15 Banks (Fixed Effect)

    LLP/TLIndependent

    Variables

    Base Line

    (2004:01-2005:05)

    Worst

    Case 1996 1997 1998G(GDP) - - - - -G(M2) 0.57 24.42 2.03 1.91 4.13G(IHSG) - - - - -INF n.s. n.s. n.s. n.s. n.s.

    SBI 3.56 33.51 6.35 7.49 23.53G(EXRATE) 0.20 22.70 0.07 1.86 0.89G(PREMIUMDN) n.s. n.s. n.s. n.s. n.s.

    G(SOLARDN) 1.32 46.67 - - 2.87

    Table 40Scenario Analysis (Without GM1) NPL/TL 15 Banks (Fixed Effect)

    NPL/TLIndependent

    VariablesBase Line

    (2004:01-2005:05) Worst

    Case 1996 1997 1998G(GDP) n.s. n.s. n.s. n.s. n.s.

    G(M2) 2.68 115.11 9.58 9.00 19.46G(IHSG) n.s. n.s. n.s. n.s. n.s.

    INF 8.28 170.84 6.94 11.64 68.14SBI n.s. n.s. n.s. n.s. n.s.

    G(EXRATE) n.s.n.s. n.s. n.s. n.s.

    G(PREMIUMDN) n.s. n.s. n.s. n.s. n.s.

    G(SOLARDN) - - - - -

    Table 41Scenario Analysis (Without GM1) D(LLP/TL) 15 Banks (Fixed Effect)

    D(LLP/TL)Independent

    VariablesBase Line

    (2004:01-2005:05) Worst

    Case 1996 1997 1998G(GDP) - - - - -

    G(M2) - - - - -

    G(IHSG) - - - - -

    INF - - - - -

    SBI 0.25 2.31 0.44 0.52 1.62G(EXRATE) - - - - -

    G(PREMIUMDN) - - - - -

    G(SOLARDN) - - - - -

  • 8/6/2019 exposicion sab

    44/57

    44

    Table 42Scenario Analysis (Without GM1) D(NPL/TL) 15 Banks (Fixed Effect)

    D(NPL/TL)

    Independent

    Variables

    Base Line(2004:01-

    2005:05)

    Worst

    Case 1996 1997 1998G(GDP) -0.23 1.81 -0.29 -0.22 0.43G(M2) 0.21 8.85 0.74 0.69 1.50G(IHSG) - - - - -INF 0.83 17.13 0.70 1.17 6.83SBI n.s. n.s. n.s. n.s. n.s.

    G(EXRATE) n.s. n.s. n.s. n.s. n.s.

    G(PREMIUMDN) 0.26 7.12 - - 0.47G(SOLARDN) - - - - -

    Table 43Scenario Analysis (Without GM1) G(LLP) 15 Banks (Fixed Effect)

    G(LLP)Independent

    VariablesBase Line

    (2004:01-2005:05) Worst

    Case 1996 1997 1998G(GDP) n.s. n.s. n.s. n.s. n.s.

    G(M2) n.s. n.s. n.s. n.s. n.s.

    G(IHSG) n.s. n.s. n.s. n.s. n.s.

    INF n.s. n.s. n.s. n.s. n.s.

    SBI n.s. n.s. n.s. n.s. n.s.G(EXRATE) n.s. n.s. n.s. n.s. n.s.

    G(PREMIUMDN) 4.46 121.33 - - 7.98G(SOLARDN) n.s. n.s. n.s. n.s. n.s.

    Table 44Scenario Analysis (Without GM1) G(LLP)/TL 15 Banks (Fixed Effect)

    G(LLP)/TLIndependent

    VariablesBase Line

    (2004:01-2005:05) Worst

    Case 1996 1997 1998

    G(GDP) - - - - -G(M2) - - - - -

    G(IHSG) - - - - -

    INF - - - - -

    SBI - - - - -

    G(EXRATE) - - - - -

    G(PREMIUMDN) - - - - -

    G(SOLARDN) - - - - -

  • 8/6/2019 exposicion sab

    45/57

    45

    Table 45Scenario Analysis (Without GM1) LLP/TL 15 Banks (Flexible Form)

    LLP/TLIndependent

    VariablesBase Line (2004:01-

    2005:05) Worst

    Case 1996 1997 1998G(GDP) n.s. n.s. n.s. n.s. n.s.

    G(M2) n.s. n.s. n.s. n.s. n.s.

    G(IHSG) n.s. n.s. n.s. n.s. n.s.

    INF 3.17 65.43 2.66 4.46 26.10SBI 3.70 34.80 6.60 7.78 24.43G(EXRATE) 0.64 74.13 0.24 6.09 2.89G(PREMIUMDN) n.s. n.s. n.s. n.s. n.s.

    G(SOLARDN) 1.06 37.36 - - 2.30

    Table 46Scenario Analysis (Without GM1) NPL/TL 15 Banks (Flexible Form)

    NPL/TL

    Independent Variables

    Base Line (2004:01-2005:05)

    WorstCase 1996 1997 1998

    G(GDP) n.s. n.s. n.s. n.s. n.s.

    G(M2) 0.77 32.94 2.74 2.57 5.57G(IHSG) n.s. n.s. n.s. n.s. n.s.

    INF n.s. n.s. n.s. n.s. n.s.

    SBI 5.57 52.40 9.93 11.71 36.79G(EXRATE) - - - - -G(PREMIUMDN) n.s. n.s. n.s. n.s. n.s.

    G(SOLARDN) n.s. n.s. n.s. n.s. n.s.

  • 8/6/2019 exposicion sab

    46/57

    46

    Table 47 Scenario Analysis (Without GM1) D(LLP/TL) 15 Banks (Flexible Form)

    D(LLP/TL)

    Independent Variables

    Base Line (2004:01-2005:05)

    WorstCase 1996 1997 1998

    G(GDP) - - - - -G(M2) - - - - -

    G(IHSG) - - - - -

    INF - - - - -

    SBI - - - - -

    G(EXRATE) 0.02 2.09 0.01 0.17 0.08G(PREMIUMDN) - - - - -

    G(SOLARDN) 0.04 1.44 - - 0.09

    Table 48Scenario Analysis (Without GM1) D(NPL/TL) 15 Banks (Flexible Form)

    D(NPL/TL)Independent

    VariablesBase Line (2004:01-

    2005:05) Worst

    Case 1996 1997 1998G(GDP) -0.23 1.80 -0.29 -0.22 0.43G(M2) - - - - -

    G(IHSG) -0.20 2.85 -0.26 0.29 0.01INF 0.30 6.17 0.25 0.42 2.46SBI n.s. n.s. n.s. n.s. n.s.

    G(EXRATE) - - - - -

    G(PREMIUMDN) - - - - -G(SOLARDN) - - - - -

  • 8/6/2019 exposicion sab

    47/57

    47

    Table 49Scenario Analysis (Without GM1) G(LLP) 15 Banks (Flexible Form)

    G(LLP)

    Independent Variables

    Base Line (2004:01-2005:05)

    WorstCase 1996 1997 1998

    G(GDP) -11.41 91.02-

    14.72 -11.25 21.72G(M2) n.s. n.s. n.s. n.s. n.s.

    G(IHSG) -6.97 99.09 -9.07 10.07 0.20INF 17.29 356.65 14.49 24.29 142.26SBI n.s. n.s. n.s. n.s. n.s.

    G(EXRATE) 1.18 137.51 0.45 11.30 5.36G(PREMIUMDN) n.s. n.s. n.s. n.s. n.s.

    G(SOLARDN) n.s. n.s. n.s. n.s. n.s.

    Table 50Scenario Analysis (Without GM1) G(LLP)/TL 15 Banks (Flexible Form)

    G(LLP)/TLIndependent

    VariablesBase Line (2004:01-

    2005:05) Worst

    Case 1996 1997 1998G(GDP) - - - - -G(M2) - - - - -G(IHSG) - - - - -INF - - - - -SBI - - - - -

    G(EXRATE) - - - - -G(PREMIUMDN) - - - - -G(SOLARDN) - - - - -

    Table 51Scenario Analysis (Without GM2) LLP/TL 15 Banks (Fixed Effect)

    LLP/TLIndependent

    VariablesBase Line

    (2004:01-2005:05) Worst Case 1996 1997 1998G(GDP) - - - - -

    G(M1) - - - - -G(IHSG) n.s. n.s. n.s. n.s. n.s.

    INF 2.32 47.90 1.95 3.26 19.11SBI - - - - -G(EXRATE) - - - - -G(PREMIUMDN) n.s. n.s. n.s. n.s. n.s.

    G(SOLARDN) 1.44 50.89 - - 3.13

  • 8/6/2019 exposicion sab

    48/57

    48

    Table 52Scenario Analysis (Without GM2) NPL/TL 15 Banks (Fixed Effect)

    NPL/TL

    Independent

    Variables

    Base Line(2004:01-

    2005:05)

    Worst

    Case 1996 1997 1998G(GDP) -2.11 16.81 -2.72 -2.08 4.01G(M1) 1.49 70.32 6.40 4.85 12.57G(IHSG) n.s. n.s. n.s. n.s. n.s.

    INF 5.87 121.06 4.92 8.25 48.29SBI n.s. n.s. n.s. n.s. n.s.

    G(EXRATE) - - - - -G(PREMIUMDN) n.s. n.s. n.s. n.s. n.s.

    G(SOLARDN) - - - - -

    Table 53Scenario Analysis (Without GM2) D(LLP/TL) 15 Banks (Fixed Effect)

    D(LLP/TL)

    Independent Variables

    Base Line(2004:01-2005:05)

    WorstCase 1996 1997 1998

    G(GDP) n.s. n.s. n.s. n.s. n.s.

    G(M1) 1.12 52.79 4.81 3.64 9.43G(IHSG) n.s. n.s. n.s. n.s. n.s.

    INF n.s. n.s. n.s. n.s. n.s.

    SBI n.s. n.s. n.s. n.s. n.s.

    G(EXRATE) n.s. n.s. n.s. n.s. n.s.G(PREMIUMDN) n.s. n.s. n.s. n.s. n.s.

    G(SOLARDN) 0.43 15.30 - - 0.94

  • 8/6/2019 exposicion sab

    49/57

    49

    Table 54Scenario Analysis (Without GM2) D(NPL/TL)15 Banks (Fixed Effect)

    D(NPL/TL)

    Independent

    Variables

    Base Line(2004:01-

    2005:05)

    Worst

    Case 1996 1997 1998G(GDP) -0.23 1.80 -0.29 -0.22 0.43G(M1) 0.20 9.40 0.86 0.65 1.68G(IHSG) - - - - -INF 0.86 17.73 0.72 1.21 7.07SBI n.s. n.s. n.s. n.s. n.s.

    G(EXRATE) n.s. n.s. n.s. n.s. n.s.

    G(PREMIUMDN) 0.27 7.25 - - 0.48G(SOLARDN) - - - - -

    Table 55Scenario Analysis (Without GM2) G(LLP) 15 Banks (Fixed Effect)

    G(LLP)

    Independent Variables

    Base Line(2004:01-2005:05)

    WorstCase 1996 1997 1998

    G(GDP) n.s. n.s. n.s. n.s. n.s.

    G(M1) n.s. n.s. n.s. n.s. n.s.

    G(IHSG) n.s. n.s. n.s. n.s. n.s.

    INF 8.70 179.44 7.29 12.22 71.57SBI 15.80 148.64 28.18 33.21 104.36G(EXRATE) n.s. n.s. n.s. n.s. n.s.G(PREMIUMDN) 4.52 122.99 - - 8.09G(SOLARDN) n.s. n.s. n.s. n.s. n.s.

  • 8/6/2019 exposicion sab

    50/57

    50

    Table 56Scenario Analysis (Without GM2) G(LLP)/TL 15 Banks (Fixed Effect)

    G(LLP)/TLIndependent

    VariablesBase Line

    (2004:01-2005:05) Worst

    Case 1996 1997 1998

    G(GDP) - - - - -G(M1) - - - - -G(IHSG) - - - - -INF - - - - -SBI - - - - -G(EXRATE) - - - - -G(PREMIUMDN) - - - - -G(SOLARDN) - - - - -

    Table 57

    Scenario Analysis (Without GM2) LLP/TL 15 Banks (Flexible Form)LLP/TL

    Independent Variables

    Base Line (2004:01-2005:05)

    WorstCase 1996 1997 1998

    G(GDP) - - - - -G(M1) 0.46 21.53 1.96 1.49 3.85G(IHSG) - - - - -INF - - - - -SBI 2.64 24.84 4.71 5.55 17.44G(EXRATE) 0.14 16.51 0.05 1.36 0.64

    G(PREMIUMDN) 1.94 52.90 - - 3.48G(SOLARDN) 0.87 30.64 - - 1.89

    Table 58Scenario Analysis (Without GM2) NPL/TL 15 Banks (Flexible Form)

    NPL/TL

    Independent Variables

    Base Line(2004:01-2005:05)

    WorstCase 1996 1997 1998

    G(GDP) n.s. n.s. n.s. n.s. n.s.

    G(M1) - - - - -G(IHSG) n.s. n.s. n.s. n.s. n.s.

    INF n.s. n.s. n.s. n.s. n.s.

    SBI 6.75 63.52 12.04 14.20 44.60G(EXRATE) 0.22 25.42 0.08 2.09 0.99G(PREMIUMDN) n.s. n.s. n.s. n.s. n.s.

    G(SOLARDN) - - - - -

  • 8/6/2019 exposicion sab

    51/57

    51

    Table 59Scenario Analysis (Without GM2) D(LLP/TL ) 15 Banks (Flexible Form)

    D(LLP/TL)Independent

    VariablesBase Line (2004:01-

    2005:05) Worst

    Case 1996 1997 1998

    G(GDP) - - - - -G(M1) - - - - -G(IHSG) - - - - -INF - - - - -SBI 0.07 0.70 0.13 0.16 0.49G(EXRATE) 0.02 1.78 0.01 0.15 0.07G(PREMIUMDN) - - - - -G(SOLARDN) - - - - -

    Table 60Scenario Analysis (Without GM2) D(NPL/TL ) 15 Banks (Flexible Form)

    D(NPL/TL)Independent

    VariablesBase Line (2004:01-

    2005:05) Worst

    Case 1996 1997 1998G(GDP) - - - - -G(M1) - - - - -G(IHSG) - - - - -INF 0.35 7.19 0.29 0.49 2.87SBI - - - - -G(EXRATE) - - - - -G(PREMIUMDN) n.s. n.s. n.s. n.s. n.s.

    G(SOLARDN) - - - - -

    Table 61Scenario Analysis (Without GM2) G(LLP) 15 Banks (Flexible Form)

    G(LLP)Independent

    VariablesBase Line (2004:01-

    2005:05) Worst

    Case 1996 1997 1998G(GDP) n.s. n.s. n.s. n.s. n.s.

    G(M1) n.s. n.s. n.s. n.s. n.s.

    G(IHSG) n.s. n.s. n.s. n.s. n.s.

    INF 5.60 115.49 4.69 7.87 46.06SBI n.s. n.s. n.s. n.s. n.s.

    G(EXRATE) 0.25 29.62 0.10 2.43 1.15G(PREMIUMDN) n.s. n.s. n.s. n.s. n.s.

    G(SOLARDN) 1.74 61.47 - - 3.78

  • 8/6/2019 exposicion sab

    52/57

    52

    Table 62Scenario Analysis (Without GM2) G(LLP)/TL 15 Banks (Flexible Form)

    G(LLP)/TL

    Independent

    Variables

    Base Line(2004:01-

    2005:05)

    Worst

    Case 1996 1997 1998G(GDP) - - - - -G(M1) - - - - -G(IHSG) - - - - -INF - - - - -SBI - - - - -G(EXRATE) - - - - -G(PREMIUMDN) - - - - -G(SOLARDN) - - - - -

  • 8/6/2019 exposicion sab

    53/57

    53

    DAFTAR PUSTAKA

    Ahmed, A.S., C.T. Taheda ans S. Thomas. 1999. Bank Loan Loss Provisions: A Re-examination of Capital Management, Earnings Management andSignalling Effect. Journal of Accounting and Economics . 28. 1-25.

    Bangia et al. 2001. Rating Migration and the Business Cycle, With Application toCredit Portfolio Stress Testing.

    Bank for International Settlement (BIS). 2000. Stress Testing by Large FinancialInstitutions Current Practice and Aggregation Issues. Committee on theGlobal Financial System . Basel. Switzerland

    Basle Committee on Banking Supervision. 1999. Sound Practices for Loan Accounting and Disclosures. Bank of International Settlement.

    Bee, Marco. 2001. Mixture Models for VaR and Stress Testing. ALEA Tech Report

    Berkowitz, Jeremy. 1999. Coherent Framework for Stress Testing.

    Bikker, J.A. and P.A.J. Metzemakers. 2002. Bank Provisioning Behaviour andProcyclicality. Research Series Supervision. 50. Section Banking andSupervisory Strategies, Directorate Supervision, De Nederlandesche Bank.

    Blaschke et al. 2001. Stress Testing of Financial Systems; an Overview of Issues,Methodologies, and FSAP Experiences. IMF Working Paper WP/01/88/2001. International Monetary Fund.http://www.imf.org/external/pubs/ft/wp/2001/wp0188.pdf

    Breuer, Thomas dan Gerald Krenn. Identifying Stress Tests Scenarios.

    Breuer, Thomas et. al. 2002. Stress Tests, Maximum Loss, and Value at Risk. Liechtensteinisches Finanz-Dienstleistungssymposium

    Bulir, Ales. 2004. External and Fiscal Sustainability of the Chech Economy; A Quick Look through the IMFs Night-Vision Goggles. CBN Internal Research and Policy note No.4. Czech National Bank

    Calari, Cesare dan Stefan Ingves. 2002. Financial Sector Assessment Program-Review, Lessons, and Issues Going Forward. International Monetary Funddan World Bank

    Calari, Cesare dan Stefan Ingves. 2003. Analytical Tools of Financial Sector Assessment Program. International Monetary Fund dan World Bank

    Carson, Carol S. dan Stefan Ingves. 2001. Financial Soundness Indicators Policy Paper . International Monetary Fund

    Cherubini, U. dan G. Della Lunga. 1999. Stress Testing Technique and Value atRisk Measures: A Unified Approach. Vol. 8

    Christoffersen, Peter. 2003. Back Testing and Stress Testing Risk ManagementSystems. Canadian Investment Review Seminar . McGill University and

  • 8/6/2019 exposicion sab

    54/57

    54

    CIRANO. www.investmentreview.com/conferences/risk2003/pdfs/christoffersen.pdf

    Cihak, Martin. 2004. Designing Tests for the Czech Banking System. CNB Internal Research and Policy Note No. 3. Czech National Bank (CNB)

    Cihak, Martin. 2004. Stress Testing: A Review of Key Concepts. CNB Internal Research and Policy Note No. 2. Czech National Bank (CNB)

    Committee on the Global Financial System. 2005. Stress Testing at MajorFinancial Institutions: Survey Results and Practice.

    Crockett, Andrew. 2001. Prudential Supervision and Financial Stability. Financial Stability Forum

    de Bandt, Oliver dan Vinchett Oung. 2004. Assessment of Stress TestConducted on the Frech Banking System. Financial Stability Review No. 5.Banque de France

    Deutsche Bu