A Macro Stress Testing Framework for Assessing Financial Stability: Evidence from Malaysia

Throughout the 1990s global liquidity glut (The global savings glut was mainly fostered by the U.S. Federal Reserve’s expansive monetary policies (i.e. cheap dollar) in conjunction with the historically low interest rates in the United States and the Euro area (see Bernanke, 2005)) created an asset boom mania worldwide turning ordinary people into avid buyers. This in turn persuaded large banks (Prior to the GFC (2008), large internationally active banks had constant propensity to circumvent banking regulation and supervision via loopholes such as skirting, race to the bottom, de facto versus de jure and cherry picking) in advanced countries and emerging market economies to expand credit offering into very risky segment known as the sub-prime via lax and predatory lending (Schwartz, 2009). Subsequently, banks being at the epicenter of financial intermediation became contributors to imbalances and receivers of adverse impact due to severe pro-cyclicality, on-and off-balance sheet exposures, leverage, and shadow banking. After the failure of micro stress tests, used by large private banks and the supervisory community (Basel Committee on Banking Supervision (BCBS) monitors microprudential stress testing in over 100 countries) since the late 1990s, the most recent global financial crisis (GFC) marked the birth of macro stress testing (Micro and macro stress testing are not standalone tools nor do they supplement other tools in the macroeconomic toolkit) as a crisis management tool (see Fed, 2009a; b) (The foundation of this article is the research and conclusions contained in my PhD (e.g. Taskinsoy, 2018)).

Propagation of financial crises since the late 1990s has made macro stress testing a central focus for the Bank Negara Malaysia (BNM) (Bank Negara Malaysia (BNM) is Malaysia’s central bank) to test the resilience of Malaysia’s banking sector to extreme but plausible scenarios. Although main results of the macro stress testing exercise revealed only a modest change in capital ratios and bank profitability in the baseline scenario, macroeconomic events such as the taper tantrum of May followed by the August rout in 2013 proved that Malaysia’s banking sector and its currency (ringgit) (Bouts of shocks since 2013 rattled markets across the world and caused one of the fastest depreciation of ringgit against major currencies in recent memory (since 2012, ringgit has depreciated as much as 50% of its value against the US dollar)) were still not immune to exogenous shocks. The domineering lesson of the global financial crisis (GFC) revealed that banks worldwide gravely failed to differentiate various risk dynamics between structured products and bonds. Other important lessons of the GFC have proved that micro stress tests conducted by banks and supervisory authorities failed to capture risks related to short-term funding liquidity, securitization, interbank contagion and counterparty default (BCBS, 2009).

Stress testing is indispensable, but still not failsafe as a standalone tool to measure financial stability. It is effective when complemented by value-at-risk (VaR) or stressed VaR. Financial crises since the 1980s (i.e. Black Monday of 1987 in the U.S.) have unmistakably shown that stress testing is not an early-warning mechanism; to think of it as one would be ill-advised. Stress tests as part of a bank’s comprehensive risk management framework can aid bank executives in the decision-making process and central banks in monetary policy decisions.

Macro stress testing was an arcane topic in finance prior to the GFC which markedly unveiled the banking sector’s inability to absorb losses in the event of a high-magnitude crisis. As a result, the use of macro stress tests as a crisis management tool has developed rather rapidly since 2009 (The success of the Supervisory Capital Assessment Program (SCAP), undertaken for the first time by the U.S Federal Reserve (Fed) in 2009, spurred the use of macro stress testing as a crisis management tool (a similar program was taken in parallel by the CEBS and EBA in EU). For details, see Fed (2009), CEBS (2010), and EBA (2011)). The forward-looking aspect of macro stress tests has proved a complement to VaR models (VaR is not a standalone tool to measure potential vulnerabilities within the banking sector because VaR results are less reliable under extreme but plausible scenarios. Virtually all VaR models focus on the tail-risk which is the lower quantile of the distribution of the profit and loss (P&Ls) and answer the question of the largest loss over a specified time horizon at a given confidence interval. See Jorion (2001) for a survey of VaR models used) to help detect underlying risks and quantify innumerable vulnerabilities within the banking sector and assess its resilience to both endogenous and exogenous shocks under highly adverse market conditions mainly triggered by macroeconomic events. The utmost objective of the macro stress test constructed in this paper is twofold; to safeguard financial stability in Malaysia during both benign and malignant economic times, and to aid prudential policy decisions by its central bank (BNM). A top-down macro stress testing framework (e.g. Čihák, 2007) is constructed to assess the solvency of Malaysia’s banks and the resilience of banking sector in a forward-looking manner. Malaysia’s stress testing experience began with the IMF’s Financial Sector Assessment Program (FSAP) (The FSAP has been initially praised as a forward-looking process for making stress testing systematic and consistently applied in the IMF-member countries, but the repeated misleading results such as Iceland FSAP (IMF, 2008) caused loss of credibility and left a long lasting scuff on the unblemished reputation of the IMF and the World Bank) in response to the first systemic (Systemic crisis is not confined to a single country, the contagion of which spreads fast like an epidemic across multiple countries that the consequential prevalent financial instability impairs the normal functioning of a financial system to the point where many segments of the real economy would suffer materially and financial losses can be so unprecedented) Asian crisis of 1997-98 (see Blaschke et al., 2001 for the early examples of stress tests).

Against this background, the aim of this paper is to present the main elements of the stress testing framework by Čihák (2007); we extend on the model by adding macroeconomic shocks (i.e. unemployment) for conducting macro stress tests. A top-down (TD) stress testing approach is carried out where the aggregated actual bank data is employed to address two critically important questions: (1) is Malaysia’s banking sector as a whole able to withstand the assumed shocks under the two adverse scenarios during the stress testing horizon? (2) What would be the potential cost to Malaysia’s government for bailing out failed banks? The Malaysia-wide stress test (“MAST”) mainly investigates adverse effects of banks’ exposures to credit, interest rate, and foreign exchange risk under extreme but plausible scenarios. The MAST exercise primarily attempts to measure the impact on Malaysia’s banking sector using one baseline and two adverse scenarios. The outcomes of the analyses are expressed in terms of capital adequacy (CAR) or government intervention as a percent of GDP.

Our empirical work is based on solid foundations. The theoretical and scientific foundation of the study, empirical support, and the basis of research can be found in the macroprudential regulation and supervision of the banking system and the stress testing literature published by the industry participants and experts. More specific studies related to Malaysia have been published by the BNM and the multilateral organizations such as the IMF and World Bank. We have used reference points for the risk assessment analyses conducted by the U.S. Federal Reserve, Committee of European Banking Supervisors (CEBS), European Banking Authority (EBA), Financial Services Authority of the UK, Basel Committee on Banking Supervision (“the Basel Committee”) plus national supervisors for banking regulation and supervision.

The stress testing horizon of the MAST is three years based on actual bank data from 2013 to 2015, and covers the entire banking sector consisting of 56 entities as deposit takers. The benchmark CAR of 10.5% and the hurdle Tier 1 capital ratio of 6.0% were used. The aggregate CAR and Tier 1 capital ratio prior to the start of MAST were comfortably high; 15.04% and 13.30% respectively. A bank falling below either ratio is considered at the risk of insolvency, and would be required to raise fresh capital in order to comply with the regulatory capital minima. The results of the MAST are expressed in terms of independent variables (CAR and capital injection as a percent of GDP). The key results of the MAST indicate that Malaysia’s financial sector is well diversified, highly interconnected and its regulatory and supervisory frameworks are consistently resilient to shocks under acute financial stress.

The stress testing scenarios developed and employed in the MAST are hypothetical rather than forecasts. The baseline scenario follows a normal course of economic activity which is consistent with the latest IMF World Economic Outlook projections for Malaysia (i.e. 5% GDP growth). Both hypothetical but plausible adverse and severely adverse scenarios assume a recession, slower economic recovery, and high unemployment. The main results of the MAST are merely indications based on assumptions and trajectories for a number of key variables describing the nature of economic activity in Malaysia which may or may not reflect the views of the BNM, actual developments or market events domestically and globally.

Malaysia has a thriving equity market, a large bond market (the largest within the Association of Southeast Asian Nations – ASEAN-5), and growing private debt securities. Nevertheless, some restrictions still exist in the financial sector where the Malaysian government has a substantial controlling stake including explicit and implicit guarantees. The BNM must have the utmost responsibility as well as accountability to monitor the levels of private corporate and household indebtedness. Excessive leverage in these two segments put strain on banking sectors and cause unbearable burden on households. It is not alarmingly concerning yet, but Malaysia should work proactively towards reducing the level of foreign claims to avoid a deleveraging process in the event of an acute financial distress. Moreover, Malaysia’s banking operation abroad (about one-third of the GDP) should be subject to rigorous monitoring, this may require the BNM to initiate cooperation in the host countries to ensure no escaping from its banking regulation and supervision. Despite a growing criticism, some market distortions still exist (i.e. tax incentives are provided for the Islamic finance); the BNM needs to eliminate such deficiencies and on a regular basis improve both quality and quantity of capital.

Financial crises are not instant; on the contrary, they evolve over time via manifestations of protracted (nocuous at times) interactions with linkages between bank intermediation and the broader economy. Financial instability and ensuing crises can stem from a wide spectrum of sources involving an amalgam of high-magnitude events that are complex, multifaceted, and systemic in nature (Claessens & Kose, 2013). Crockett (1997) asserts that monetary and financial stability are the two integral components of a normal-functioning financial system, Borio (2003) calls it “twin stability”. Alternatively, Minsky (1975) strongly argues that cyclical excesses leads to financial instability (i.e. the Fed’s expansive monetary stand). Monetarists Friedman and Schwartz (1963), Eichengreen (2002), Bernanke and Gertler (1995) believe that both policy errors and ineffective policy decisions may foster financial instability.

It would be edifying to look at the past century. The world trade (1870-1910), underpinned by the classical gold standard (The gold standard was officially inaugurated in 1844 by the British Empire which marked the new beginning in trade where British pound sterling was fully backed by gold. By 1908, most countries adopted gold standard, but after Britain dropping gold standard in 1931, countries in Europe were off-gold by 1936 (see Kindleberger & Alibar 2005)), was hugely disrupted by World War-I (WWI) that changed the power structure in Europe and re-defined the pattern of world trade. At the turn of the 19th century (1890s) before the United States entered the scene of dominance, the power structure in Europe was shared by Germany, France, and Britain who were each other’s chief enemy. Prolongation of the protectionist policies, currency blocks, and restricted trade zones put in place by each of these major powers elicited constant rivalries among them which are regarded as causal contributors to the sudden eruption of WWI; and severe consequences emanated from the unresolved issues of WWI eventually led to the outbreak of World War II.

The adequate supply of gold, on-and-off, had been a problem in the classical gold standard; as a solution, this was replaced by a gold exchange making the British sterling and the U.S. dollar as reserve currencies for gold shortage (Eichengreen & Irwin, 2010a; Estevadeordal et al., 2003; and Lopez-Cordova & Meissner, 2003). Unresolved issues of WWI culminated in the banking crisis triggered by speculative attacks on British pound subsequently forced Britain to leave the gold standard in 1931. Among many others, Bernanke and James (1991) and Eichengreen (2003) strongly argue that Britain’s decision to drop gold arguably contributed to the Great Depression (The Great Depression has been described as “…by far the worst catastrophe of the 20th century” (Wheelock, 2007)) which forced countries (the U.S. in particular) to pass isolationist “beggar-the-neighbor” (These isolationist policies were passed by countries to resolve their own economic problems at the expense of other nations. The United States passed more policies in this nature in the 1930s than any country (Eichengreen & Irwin, 2010b)) policies that severely disrupted the world trade in the 1930s (However, Monetarists Friedman and Schwartz (1963) provide a compelling argument that falling money stock in the U.S was the driving force behind the severe contraction in consumption that in turn caused a chain reaction for the steady state output to plummet (more than 30%), prices to free-fall (more than half), unemployment to soar (25% of Americans out of a job), and thousands of banks to fail. These unfolding events actually caused the Great Depression of the 1930s).

In the U.S., the dark and devious side of investment banking activities of commercial banks (e.g. Lardner, 2009) leading up to the Great Depression made investors and the American public lose trust in the entire financial system (Crawford, 2011; Mester, 1996). To restore confidence, the U.S. authorities quickly responded to the outcry by the American public with the passage of the Glass-Steagall Act (or the Banking Act of 1933) which was drafted by US Senator Carter Glass and Representative Henry Steagall (After the Federal Reserve failed to avoid the Great Depression; the Federal Deposit Insurance Corporation (FDIC) was established and Social Security System was initiated. Also, the first national minimum wage law was ratified, and the federal government spending rose from 3% of GDP then to 10%, now 20% (Wheelock, 2007)). The motivation behind the act was to make the banking system in the United States more resilient through two important objectives; the separation of commercial banking from investment banking to ensure that banks’ assets were not used inappropriately by diverting funds into some other speculative purposes (commercial banks were barred from underwriting securities); second, create an independent Federal Deposit Insurance Corp. (FDIC) under the Federal Reserve; previously, the Office of Comptroller of the Currency (OCC) was in charge of national bank liquidations.

After WWII broke out in September 1939, Harry Dexter White (the U.S.) and John Maynard Keynes (the UK) began working on respective monetary plans in the outset of 1940 (Keynes (“Baron Keynes”), the distinguished economist and leader of the British delegation at the Bretton Woods Conference in 1944, publicly criticized White’s plan for its narrow focus on unresolved issues and immediate postwar problems such as Germany’s reparations from WWI, undisrupted world trade, and unrestricted currencies). The Bretton Woods Conference (1944) gave birth to a new monetary system proposed by the U.S. and its decision to establish the International Monetary Fund (IMF) and the World Bank. Although Keynes strongly argued against the use of a fixed exchange rate regime pegged to the U.S. dollar (Helpman 1981), his proposal of a supranational currency and the creation of two watchdog institutions were overwhelmingly rejected (Schumacher, 1943). With the US dollar as the new reserve currency marked the beginning of U.S. hegemony, and many fixed currencies became free-floating (see Ruggie, 1982). In the aftermath of WWII, the U.S. was by far the most dominant (The U.S. nearly produced two-thirds of the world’s total manufacturing output helped by almost full employment (see Becker, 1964; Keesing, 1966). Extraordinary production capacity and output transformed the U.S. from a net importer (historically) to a net exporter in the late 1940s (Branson et al., 1980)) industrial power in the world, provided that no competition from the war-devastated Europe and Japan which were completely destroyed by the war.

While different monetary proposals of the Bretton Woods system (e.g. Moggridge, 1980) were being drafted at the onset of 1940, ASEAN-5 (ASEAN-5 stands for the Association of Southeast Asian Nations (the five founding member-nations in alphabetical order are; Indonesia, Malaysia, Philippines, Singapore, and Thailand)) countries were colonized or occupied by foreign powers. Furthermore, each founding member of the ASEAN-5 was invaded by Japan during WWII (1942-45). For instance, Indonesia was colonized by the Dutch (the 1800s- 1945), Malaysia by the British (1824-1946), and the Philippines by the Americans (1898- 1946). Interestingly, Thailand had never been colonized but instead faced several military dictatorships for four decades (1932-73). In the post WWII era, ASEAN-5 nations one after another declared independency starting with Indonesia in August 17, 1945; Philippines in July 4, 1946; Malaysia in August 31, 1957 (Malaysia was formed in 1963 consisting of one-third of Borneo, Sabah, Sarawak, and Singapore; however, Singapore was forced to withdraw from the federation in 1965, which in the same year declared its independency); and Singapore in August 9, 1965 (Singapore’s decision to separate from Malaysia to become an independent sovereign state was the result of a long lasting political tensions between the two countries’ ruling parties. There were also deep economic differences between them).

Economists and historians suggest that the Bretton Woods system was destined to fail due to unrealistic promises and many overlooked issues (governance) (See Siklos (2013) for issues with the international financial regulation). The foreseen collapse of the Bretton Woods system resulted in significant volatility in major foreign exchange markets in the 1970s (see Mundell & Swoboda, 1969 for monetary problems). This quickly turned into a worldwide financial turmoil in the wake of two other high-magnitude events causing enormous disruptions to the world trade and financial stability. First was the Yom Kippur War (This was a follow-up retaliation by the Egyptian and Syrian forces to win back territory lost to Israel during the Arab- Israeli war in 1967. Each side claimed victories but in reality the Yom Kippur war was a disaster for Syria) (1973) and ensuing oil crisis (1974-78); the second event was the closure and forced liquidation of Germany’s Cologne-based Bankhaus Herstatt (1974) (The sudden bankruptcy of a small privately-owned Bankhaus Herstatt in June 1974 is a famous as well as a shocking incident that clearly illustrated the settlement risk in foreign exchange payments, plus ignored regulatory issues). These macro events and their farfetched implications resulted in unprecedented spikes in commodity prices by the mid-1970s during which the prices of both gold and oil broke records (OPEC sharply raised oil prices during 1973-74 to show its strong disapproval of the U.S. aid to Israel during the war. By March 1974, the price of oil rose from $3/barrel to $12/barrel. The price of gold rose from $44/ounce in 1972 to $185 in 1975 (see Kindleberger & Alibar, 2005 for the reasons behind the huge spike)). Hazlitt (1984) argues that the instability episodes were the “…consequence of the inherently inflationary institutions set up in 1944” (i.e. the IMF and the World Bank) and the Bretton Woods system “…not only permits and encourages but almost compels world inflation”. Dooley et al. (2004) stress that “Bretton Woods system does not evolve; it just occasionally reloads a periphery”.

The severity of financial and political turbulence in the 1970s fostered by the aforementioned instability-inflicting events forced the G-10 Governors (From Belgium, Canada, France, Germany, Italy, Japan, Netherlands, Sweden, Switzerland, the UK, and the United States.) to engage in financial collaboration and cooperation, which gave an imminent birth to the creation of the Basel Committee on Banking Supervision in 1974. To converge the existing disparate capital standards across G- 10, the Basel Committee published the 1988 Capital Accord (known as Basel I), “International Convergence of Capital Measurement and Capital Standards” (BCBS, 1988). Elizalde (2007) and Jones (2000) criticize Basel I for not only ushering riskier lending but encouraging capital arbitrage. Ferguson (2003) emphasizes that “Basel I Accord is too simplistic to adequately address the activities of our most complex banking institutions”. Jackson et al. (1999) argue that Basel I is another undue regulatory taxation imposed on banks (Five arbitrary risk weighting categories and corresponding risk buckets (0%, 10%, 20%, 50%, and 100%) under Basel I for credit risk had adverse effects on financial sectors of the ASEAN-5 prior to the Asian crisis (see BCBS, 1999a;b)).

The drawback of Basel I was that it predominantly focused on credit risk while ignoring other important elements such as the supervisory review process and market discipline. To correct the overly criticized flaws of Basel I, and more specifically, to increase the sensitiveness to risk, the Basel Committee published a new revised framework in June 2004 (known as Basel II), fully implemented by the end of 2006: “International Convergence of Capital Measurement and Capital Standards: A Revised Framework” (see BCBS, 2004 for details). On the contrary to expectations, Basel II further augmented procyclicality and turned too-big-to-fail banks into bigger-and harder-to-fail banks. Moreover, Basel II allowed the creation of shadow banking by giving large banks incentives to move assets off their balance sheets. Probably the worst effect of Basel II was to make banks overly rely on the external credit assessment institutions (ECAIs) ratings for decisions on audit frequency, dividend payout, and deposit rates. This last point in turn made banks feel the least urgency to allocate funds to strengthen their existing risk-management frameworks or develop far better ones.

Each milestone in the stress testing evolution has often corresponded with financial crises since the late 1990s, and the introduction of Basel Accords (II and III) followed a similar path during the same period. Once an arcane topic but now a common household name, micro stress testing has gained prominence in the late 1990s, but macro stress testing has become indispensable as a crisis management tool on account of the 2008 credit crisis (GFC). Details of the earlier stress testing practices employed by banks and regulatory supervisors and this paper’s topic macro stress testing will be covered in the Literature Review section.

In the post Second War (WWII) era, financial markets were relatively small and the coverage for pure financial risk was in infancy stages. Insurance policy was predominantly used as a protection against financial losses resulting from work related incidents or environmental catastrophes. The publication of the five seminal papers (Markowitz’s (1952, 1959) portfolio theory (mean-variance criterion); Modigliani-Miller’s (1958) irrelevance and arbitrage-reasoning; Sharpe’s (1964) Capital Asset Pricing Model (CAPM); Black and Scholes’ (1973) option pricing formula; and Merton’s (1974) pricing of corporate debt (arbitrage-based pricing model)) instigated the evolution of finance to accelerate at a remarkable pace. Prior to the Asian crisis of 1997-98, intensifying financial turmoil by the mid-1990s prompted the Basel Committee to release the 1996 market risk Amendment to Basel I (BCBS 1996a; b) requiring large banks to use backtesting. However the deficiencies and imperfect signals generated by backtesting (Campbell, 2005), stress testing became a mainstay when banks with substantial trading volumes were required to use stress testing to confirm the accuracy of VaR, Jorion (2001) calls this process “a reality check” (Interested readers can further read studies such as CGFS (2000, 2001, and 2005) that surveyed major and large financial institutions for their uses of earlier stress testing (i.e. microprudential)).

The typology of stress testing is categorized along two dimensions: micro-prudential and macro-prudential. Microprudential stress testing (The Basel Committee required banks to use internally designed stress tests in the late 1990s and the use of micro stress testing became mandatory under Basel II; under Basel III, banks are required to use enhanced stress testing approaches) (BU: bottom-up) is employed by banks for internal risk management purposes and by regulatory supervisors for pillar II solvency. Macroprudential stress testing (both BU and TD: top-down) is employed by central banks and the supervisory community for assessing the entire financial system and by the IMF for country-level surveillance. Starting with the Supervisory Capital Assessment Program of the U.S. in 2009, macroprudential stress tests (BU) are used as a crisis management tool. There is also liquidity stress testing but it is still in infancy stage and not linked to bank solvency.

Greenspan (1998), the US Fed Chairman then, in his testimony to the U.S. Congress on January 26 warned that “…significant mistakes in macroeconomic policy also reverberate around the world at a prodigious pace’’. Asian crisis of 1997-98 may have been the inevitable outcome of such policies’ adverse effects; Alan Greenspan indicated in 1998 that foreign investors in Asian equities (excluding those in Japan) lost an estimated $700 billion-including $30 billion by American investors (Nanto, 1998). Throughout the 1990s, macro events’ massive costs to economies prompted the IMF and the World Bank jointly to initiate the FSAP in 1999 (The focus of FSAP assessments is twofold: to assess the resilience of the financial sector to endogenous and exogenous shocks and to measure the financial sector’s potential contribution to growth and development of non-financial sectors). Even though Malaysia now conducts its own micro and macro stress tests, it underwent its first macro stress testing experience via FSAP. The main results of Malaysia’s latest FSAP indicate that its economy is robust with a sound financial system helped by a continued flow of funds amid the GFC, but there has been a slow and persistent deterioration since 2013. High levels of public finances, overexpansion of credit and excessive household debt are disquieting, plus the subdued inflation has moved up after the minimum wage law in 2013 (IMF, 2014).

The IMF’s FSAP, as part of its surveillance program known as Article IV Consultation, was primarily established “...to help countries enhance their resilience to crises and foster growth by promoting financial stability and financial sector diversity” (IEO, 2004). Hilbers (2001) sees three components of an FSAP as critical; (1) an assessment of micro/macro prudential reforms and development needs; (2) an assessment of vulnerabilities to macroeconomic and financial factors; (3) an assessment of the existing, laws, rules, codes, regulation, and banking supervision. Although the IMF’s FSAP has been initially praised as a forward-looking process for making stress testing systematic and consistently applied in the IMF-member countries, but the misleading results of Iceland’s FSAP (IMF, 2008) caused not only loss of credibility but left a long-lasting scuff on the unblemished reputation of the IMF and the World Bank (see IMF & World Bank, 2003 for tools; IMF & World Bank, 2005a; b for the lessons learned) (An extensive review of the program’s processes was initiated after overhauling of banking systems (i.e. Irish bank system) and collapsing banks (i.e. Belgium’s Dexia) shortly after passing FSAPs and EU-wide stress tests in 2010 and 2011).

The contemporaneous financial crises in the new millennium have unmistakably proved that individual banks’ micro stress tests for the internal risk management purposes had serious deficiencies; they were both inadequate (not bank-wide) and insufficient (light scenarios generated results by design) to prevent a high-magnitude financial crisis similar to the GFC. The Committee on the Global Financial System (CGFS) (The CGFS is a central bank forum, established by the Governors of the G-10 central banks to monitor and examine broad issues relating to global financial markets) investigated the use of stress tests through 424 stress tests performed by 43 large internationally active banks and securities firms from ten countries. The overall conclusion of the CGFS survey reveals the importance of stress testing as a valuable tool to gauge and manage risks (CGFS, 2000). Most frequently used stress testing technique among these institutions in 2000 was the simple sensitivity test to measure the adverse impact of changes in a single risk factor on portfolios or business units. The second popular stress testing technique used was a scenario analysis (i.e. historical or/and hypothetical) to measure risk exposures under extreme but plausible market events. The domineering lesson of the latest global financial meltdown during 2007-08 revealed that banks gravely failed to differentiate common risk dynamics of the modern finance between structured products and bonds; further, micro stress tests conducted by banks and the supervisory authorities failed to capture key risks such as pipeline, securitization, short-term funding liquidity, interbank contagion and counterparty default (see BCBS, 2009 for details).

Another shortcoming of the earlier micro stress tests was that they had a narrow focus in scope and the health of the financial system as a whole or the buildup of systemic risk was never a concern. International Institute of Finance argues that “during the market turbulence, the magnitude of losses at many firms made it clear that their stress testing methodologies needed refinement – stress testing was not consistently applied, too rigidly defined, or inadequately developed (IIF, 2008). The stress testing literature vividly illustrates that micro stress tests prior to the GFC had weaknesses under extreme market conditions and the not have the sufficient capacity to cope with the subsequent unfolding events. According to the Basel Committee, this huge oversight on the part of financial authorities contributed to the intensity of the GFC (e.g. BCBS, 2009). In the post-GFC (since 2009), macro stress testing as a crisis management tool is used virtually by all central banks to measure financial stability and systemic risk. To date, IMF has completed about 130 of 188 member-FSAPs, but it is a slow-moving process. As an improvement, the IMF has expanded its initial goal of one FSAP per month to 24 FSAPs per year, but the current rate is still 17 to 19 per year (IMF, 2004).

The voluntary nature of FSAP has been criticized since the GFC. The proponents argue that countries voluntarily choosing to undergo FSAP or not may potentially impose instability on the global financial system. On that note, the US, Indonesia, and Malaysia did not undergo nor requested to participate in an FSAP prior to the GFC. Malaysia initially refused to participate in the IMF administered program due to its resentments towards the IMF’s policy responses handling the Asian crisis. A widely perceived belief suggests that interest rate hikes along with inappropriate fiscal measures of the IMF during and in the aftermath of the Asian crisis acted as a crisis’ intensifier role causing further turmoil in Malaysia’s economy. Many argue that, giving the state of economies and frail banking systems, the IMF should have at least raised interest rates moderately and fiscal targets should have been less rigid. IMF defended its actions by arguing that policy choices were limited without alternatives (The GFC has proved that the predecessors of Basel III (i.e. I & II) along with microprudential stress testing used by banks, supervisors, and multilateral organizations failed gravely to strengthen the resilience of the global financial system).

The success of the Supervisory Capital Assessment Program (SCAP), undertaken for the first time by the Fed in 2009 spurred the use of macro stress testing by central banks as a crisis management tool (see Fed, 2009a for design and implementation; 2009b for results details). In stark contrast to the US, Europe bungled with its first two macro stress tests designed and conducted by the CEBS (2010) and its successor EBA (2011). The SCAP was informative, which provided credible and market-demanded information regarding the projected post- stress losses (Bernanke, 2013). However, results of the EU-wide stress tests conducted by the CEBS and the EBA were widely perceived as insufficiently granular. According to Ellahie (2012), the stress test conducted by the CEBS, in most part, was uninformative and its partial disclosure caused a decline in equity values in Europe (CEBS, 2010). Beltratti (2011) believes that the EBA’s macro stress test in 2011 had issues as well but it was nonetheless informative in terms of methodologies and scenarios used. Tarullo (2010) contends that stress tests can stand a chance of succeeding if the rigor is absent in the design of exceptional but plausible scenarios that must be consistent and comparable. Wall (2013) argues that one of the success attributes of the SCAP was a backstop (i.e. temporary financial relief) provided by the U.S. Treasury which enabled the regulatory supervisors to identify the tail risks better.

Bernanke (2013) states that macro stress tests are “forward-looking” providing protection against “so-called tail risks”. The Basel Committee argues that adequately designed macro stress tests with extreme but plausible scenarios “…improve banking sectors’ ability to absorb shocks arising from an acute financial and economic stress” (BCBS, 2010). Two strands of macro stress testing models exist; (i) models that link macro variables to micro risk drivers (i.e. credit risk); (ii) and integrated models that take into account liquidity risk and feedback effects (Foglia, 2009). Macroeconomic models do not deal with financial sector variables, therefore satellite models such as reduced-form are employed to map macro shocks into measurable form (i.e. asset quality). In BU stress tests, banks compute loss distributions on a provided scenario and then send the results to the central bank for aggregation (Sorge, 2004). The Basel Committee stresses that both BU and TD approaches should be used to capture second-round liquidity and systemic effects (BCBS, 2013).

The successful initiatives of the ten-year Financial Sector Masterplan (2001–2010) led by the Bank Negara Malaysia (BNM) and the parallel-run Capital Market Masterplan (CMP1) led by the Securities Commission, Malaysia has escaped the severe consequences of the GFC with a minor dent in its economy. Malaysia’s financial sector went through great transformation since the Asian crisis. Asia’s systemically induced home-grown crisis in the late 1990s forced Malaysia to initiate widespread micro and macro-prudential reforms. The end result of the relentless efforts is a strengthened financial sector complemented by a rigorous regulatory and supervisory frameworks. The stress tests conducted by both BNM and IMF indicate that Malaysia’s banking system is well capitalized, therefore it is resilient to withstand economic and market shocks, but not absent from risks. Malaysian banks’ heavy reliance on demand deposits increases liquidity risk, rising household debt increases default risk (IMF, 2013).

Despite a decade-long highly praised micro and macroprudential reforms introduced by BNM and other government authorities, the farfetched implications of the GFC clearly proved that Malaysia is still disquietingly susceptible to developments in G-2 (the U.S. and EU). As such, policy normalization in the United States, the Federal Reserve’s interest rate lift-off by 25 bps in December of 2015, and the increased probability of another rate hike by the Fed in end- 2016 stirred Malaysia’s equity market and made its currency (ringgit) plummet against the dollar, in fact ringgit’s depreciation against dollar was the fastest in 17 years. Other macro events such as the slow-growth path of China, global imbalances, accommodative monetary policies and reemerged weaknesses in the Euro area contribute to the volatility.

Malaysia’s well-diversified economy is expected to experience a noticeable decline; low oil and gas prices will keep inflationary pressures subdued and cause a sizable reduction in the current account surplus. However, revenue loss due to lower commodity prices is expected to be offset by the elimination of the oil subsidies and manufacturing exports favored by a weaker exchange. Surging house prices and fast rising household debt in Malaysia remain to pose a threat to financial stability, this may even lead to the formation of a real estate bubble. Although rising real interest rates globally will curb the growth of financial risks eventually, enhanced stress tests along with macroprudential measures may be necessary (IMF, 2015).

Since the banking operation is at the epicenter of any financial intermediation, there is no perfect substitute for it in capital markets; thus, the broader economic activity is adversely affected during financial turmoil. In the aftermath of the most unprecedented global financial crisis of 2007-08, the BNM has used macro stress testing as a measure of financial crisis. Malaysia-wide stress test (“MAST”) is based on a top-down (TD) stress testing approach consisting of one baseline (not a real stress scenario, mostly used for adjustment purposes) and two adverse scenarios; the main results of all three stress scenarios are informative to central banks, supervisors, bank executives, and academia.

This paper focuses on the macro stress testing of Malaysia’s banking sector to assess its resilience to extreme but plausible scenarios (see Figure 1 for the stress testing framework). Some of the main research methods have been utilized; as such, the analysis and synthesis methods, induction, deduction, and analogy. Also, research methods such as factorial and comparative analysis, statistical and mathematical methods are used. For the purpose of an econometric study, specific macroprudential banking data was compiled from various sources including but not limited to bankscope, central bank databases, World Bank, the IMF FSAPs, Eurostat, the Basel Committee, and individual banks’ websites (Taskinsoy, 2018).

The paper follows the stress testing model developed by Čihák (2007); as an improvement, actual bank data is used in the Malaysia-wide stress test as opposed to hypothetical banks and banking data. A top-down (TD) macro stress testing approach is employed with the objective of assessing credit, interest rate, and exchange rate risks under one baseline and two adverse scenarios. The focus of measure is on the aggregate impact resulting from tighter capital and liquidity regulation of Basel III plus the assumed macro shocks, and how to express them in terms of capital adequacy and capital injection as percentage of GDP. The stress testing horizon of the MAST is three years based on annual data from 2013 to 2015 (Q4), and covers the entire banking sector consisting of 56 entities as deposit takers. As of end-2015, 31 domicile banks controlled MR1.82 trillion of the total financial assets (18 banks owned 74.6% or RM1.74 trillion and 13 investment banks owned 3.26% or RM84 billion). Domestically incorporated foreign-controlled banks (25) enjoyed 22.26% (or RM520 billion) of the consolidated banking assets. Malaysia’s five largest banks account over 70% of the total financial assets, consistent with that in advanced economies (IMF, 2013; 2015).

Stress testing is overly misunderstood and often confused with macroeconomic forecasting and early-warning indicators. Before explaining in detail the important elements of the macro stress test used in this thesis, it would be informative to outline the conceptual differences between them. The following is based on Sorge (2004), macroeconomic forecasting can be described as;

$\mathrm{E}\left(\tilde{\mathrm{x}}_{\mathrm{t}+1}\right)=\mathrm{g}_1\left\{\mathrm{X}^{\mathrm{t}}, \mathrm{Z}^{\mathrm{t}}\right\}$

Where t represents the history of past events of a random variable up to time (t); g₁ is the forecasting function that maps variables X and other factors Z into a vector of expected outcomes that take place in the future. An early-warning indicators model is described as;

$\mathrm{P}\left(\tilde{\mathrm{x}}_{\mathrm{t}+1} \geq \overline{\mathrm{x}}\right)=\mathrm{g}_2\left\{\mathrm{X}^{\mathrm{t}}, \mathrm{Z}^{\mathrm{t}}\right\}$

The problem with an early-warning mechanism is to identify subsets of X and Z as leading signals to predict the probability of a future crisis. When some critical macroeconomic variables (X) go over the pre-set thresholds, a crisis occurs (i.e. $\tilde{\mathrm{x}}_{\mathrm{t}+1} \geq \bar{x}$) and no crisis if $\widetilde{\mathrm{x}}_{\mathrm{t}+1} \leq \overline{\mathrm{x}}$.

The underlying difference between stress testing and the other two is that early-warning indicators and macroeconomic forecasting models use historical data as an input, but micro or macro stress tests use historical or/and hypothetical scenarios along with several assumptions as inputs to assess how individual banks or banking sectors can withstand extreme but plausible shocks which have not yet occurred. The consequences of $\tilde{\mathrm{x}}_{\mathrm{t}+1} \geq \bar{x}$ can be evaluated as;

$\Omega\left(\widetilde{\mathrm{Y}}_{\mathrm{t}+1} / \tilde{\mathrm{x}}_{\mathrm{t}+1} \geq \overline{\mathrm{x}}\right)=\mathrm{f}\left\{\mathrm{X}^{\mathrm{t}}, \mathrm{Z}^{\mathrm{t}}\right\}$

Where $\tilde{Y}_{t+1} / \tilde{x}_{t+1} \geq \bar{x}$ is the aggregate measure of the distress of the financial system ($\widetilde{Y}_{t+1}$) restricted to the tail events ($\tilde{\mathrm{x}}_{\mathrm{t}+1} \geq \overline{\mathrm{x}}$). Ω (.), denotes the metric used to compare the vulnerabilities of the financial system across portfolios and scenarios.

The macro-financial parameters must meet two conditions of properties (see Table 1); (1) in assessing the soundness of a banking sector, variables should be measurable/quantifiable, interpretable as well as comparable with other variables; (2) variables should be linked to a number of different risk factors so that econometrics, analytical, and statistical analyses can be performed. A list of commonly used variables is adopted from Čihák (2007).

Capital is a key measure of impact due to its close link as well as implications on solvency, but capital as a standalone variable is not a clear indication of vulnerability to shocks in an acute stress. To make capital a critical variable in the measurement of financial soundness, it needs to be viewed as a ratio (i.e. Tier 1 capital / RWAs or capital as percent of GDP).

Three important capital ratios used in stress tests as variables are the following:

Where CET1 represents common equity Tier 1 capital, RWA is the risk-weighted asset.

The capital adequacy ratio (CAR) is used as a key variable and calculated as;

Where RWA denotes risk-weighted assets (i.e. RWA = ∑ W_𝑖A_i); RC is the regulatory capital, CRWA is the risk-weighted credit risk, MRWA is the risk-weighted market risk, ORWA is the risk-weighted operational risk (see Table 2), and p is the minimum CAR used in the paper. The constituents of capital changed significantly from Basel II to Basel III (see Table 3 for the Basel III phase-in arrangements). Capital adequacy and Tier 1 capital ratios are critically important financial stability indicators (FSIs). The threshold rate of 10.50% CAR is 2.5% higher than the total capital requirement of 8% under Basel III. This hurdle rate is also significantly higher than all of the micro and macro stress tests conducted in the U.S., Europe, and Japan. Using the above equations, RWAs are computed as;