DETECTING FINANCIAL VULNERABILITY FROM MACROPRUDENTIAL INDICATORS (MPIs): AN OVERVIEW OF METHODOLOGIES AND SIMULATION WITH NIGERIAN DATA
By Dr. Victor Ekpu Managing Consultant, Mindset Resource Consulting UK
Lecture Presented at a Regional Course for Economic Regulators in the West African Monetary Zone (WAMZ), organised by the West African Institute for Financial and Economic Management (WAIFEM) Accra, Ghana
Sept 2015
DETECTING FINANCIAL VULNERABILITY FROM MACROPRUDENTIAL INDICATORS (MPIs): OVERVIEW OF METHODOLOGIES AND SIMULATION WITH NIGERIAN DATA By Victor U. Ekpu*
*
Managing Consultant, Economic Consulting Division, Mindset Resource Consulting, UK E:
[email protected]
Abstract Macroprudential analysis is a method of economic analysis that evaluates the health, soundness and vulnerabilities of a financial system. The analysis involves the assessment and monitoring of the strengths and vulnerabilities of financial systems using quantitative information, largely in terms of indicators that provide a broader picture of economic and financial circumstances. This paper reviews the various methodologies used in monitoring financial vulnerabilities and performs some simulation with Nigerian data. The methods explored in this paper include early warning systems, FSIs, stress tests (micro and macro stress tests), and the impulse response function. The analysis provided in this paper with Nigerian data is only preliminary, selective and is incomprehensive due to the paucity of data on some macroprudential indicators, particularly aggregated microprudential variables or financial soundness indicators from the Nigerian banking sector. Key words: financial vulnerability, macroprudential indicators, detection tools, early warning systems, FSIs, stress testing, impulse response function. 1. Introduction Financial stability is necessary for sustained long term economic growth. Economic growth cannot be achieved without strong financial systems. Even with sound macroeconomic fundamentals, weak financial systems can destabilize economies, making them more susceptible to external shocks. The interaction of financial markets and the real economy needs close monitoring since the social externalities and knock on effects of instability of the financial system can be very costly. A smoothly operating, stable and efficient financial system is a major pillar for growth, output and employment, which are the three core mandates of the Central bank of Nigeria (CBN). Central banks everywhere have adopted macro prudential analysis as a method of detecting vulnerabilities in the financial system. This involves, among other things, the identification of financial soundness indicators (FSIs) and the methods used to analyze them. Observing potential signs of heightened risks in the financial system is important for central banks, as they rely on such insights to be able to take both preventive measures and appropriate action in crisis management.
Detecting Financial Vulnerability from MPIs
2
The analysis of the banking sector is a central component of a broader financial stability monitoring which covers also non-bank financial institutions, major financial markets and market infrastructures. While the stability of the banking sector continues to be key to financial stability as a whole, wider analysis is essential to cover increasing potential risks to financial stability stemming from non-bank financial institutions, financial markets and payment and settlement systems and other central parts of the infrastructure. Ex ante assessment of the resilience of financial institutions to withstand potential major disturbances and adverse developments stemming from external macroeconomic, financial market or endogenous financial sector developments is important in preventing financial instability. Macro-prudential analysis is distinct from micro-prudential supervision, which focuses on the financial condition of individual institutions, their risks and risk management. Macro-prudential analysis assesses the banking and financial system as a whole and covers the threats to financial stability stemming from common shocks affecting all (or a large part of) institutions or contagion of individual problems to the rest of the system. One of the lessons that policy makers in Nigeria and other emerging African economies have learnt from the recent global crisis is the need for an overarching policy framework to address the stability of the financial system as a whole - a macro prudential policy framework. Macro prudential analysis is a method of economic analysis that evaluates the health, soundness and vulnerabilities of a financial system. The analysis involves the assessment and monitoring of the strengths and vulnerabilities of financial systems using quantitative information, largely in terms of economic indicators that provide a broader picture of economic and financial circumstances. The assessment takes into account both the risks facing the financial system and the system’s capacity to withstand shocks. It also involves establishing or investigating the linkages between the financial system and the real sector of the economy. Scenario analysis and stress tests are major components of this analysis and they help in determining the system’s sensitivity to economic shocks and its resilience to such shocks. This paper reviews the various methodologies used in monitoring financial vulnerabilities and performs some simulation with Nigerian data1. The methods explored in this paper include early warning systems, FSIs, stress tests (micro and macro stress tests), and the impulse response function, dynamic multiplier functions and forecast error variance decomposition (FEVD). Banking activities in Nigeria have widened and deepened substantially over the past years, particularly since the consolidation of the banking industry in 2004, which has now produced very robust and well capitalized banks, offering a much broader range of products and services and acting as major players in the African financial system. The implications of the recent developments in the Nigerian banking system are analogous to those in other emerging market economies. First, the rapid increase in banks’ financial market activities has heightened their exposure to market risks and earnings volatility. Second, the greater links between banks and non-bank financial institutions may have increased the likelihood that shocks emanating from non-banks (such as Pension Fund Administrators PFAs, Insurance companies and Mutual funds) become systemic and spread to banks. Third, owing to changes in banks funding and investment patterns, liquidity conditions in money and other financial markets and contagion risks may play an increasingly relevant role, rather than traditional liquidity crises due to runs by retail depositors (who are protected through deposit insurance). In order to capture and monitor all relevant risks, both the traditional ones as well as those induced by the above (and other) structural changes, the framework for macro-prudential analysis in Nigeria needs to have a wide scope and be dynamic in nature. When developing the analytical framework and indicators for macro-prudential analysis of the banking sector, the main objective is to sufficiently cover all major sources of risk. This requires due attention to the structural changes in the economy, financial markets, and banks’ business activities as these changes influence banks’ risk profiles.
The analysis provided in this paper with Nigerian data is only preliminary, selective and is incomprehensive due to the paucity of data on some macroprudential indicators, particularly aggregated microprudential variables or financial soundness indicators from the Nigerian banking sector.
1
Detecting Financial Vulnerability from MPIs
3
2. The Role of Macroprudential Analysis in Detecting Financial Vulnerability The role of macroprudential analysis or financial stability analysis is to identify risks or threats to financial system stability and to design appropriate policy responses. It focuses on exposures, buffers, and linkages to assess the soundness and vulnerabilities of the financial system, as well as the economic, regulatory, and institutional determinants of financial soundness and stability. It considers whether the financial sector exhibits vulnerabilities that could trigger a liquidity or solvency crisis, amplify macroeconomic shocks, or impede policy responses to shocks. According to Nier (2011), the sources of systemic vulnerabilities in the financial system include but are not limited to the following factors: (i) the build-up of macroeconomic and financial imbalances accompanied by favourable economic conditions (e.g. rapid credit growth, large capital inflows, etc), (ii) financial innovation, (iii) low funding liquidity, and (iv) rise in asset prices also fuelled by rapid credit expansion. These factors were evident prior to the global financial crisis. Faced with these factors, the macro-prudential regulator thus has the responsibility of measuring, monitoring and mitigating systemic risk vulnerabilities over the life of a financial cycle. 2.1. Elements of Macroprudential Analysis The monitoring and analysis of financial soundness involves an assessment of macroeconomic conditions, soundness of financial institutions and markets, financial system supervision, and the financial infrastructure to determine what the vulnerabilities are in the financial system and how they are being managed. Depending on this assessment of the extent of the financial system’s soundness, policy prescriptions may include continuing prevention (when the financial system is inside the stable corridor), remedial action (when it is approaching instability), and resolution (when it is experiencing instability). The analytical framework to monitor financial soundness is centered on macro prudential surveillance and is complemented by surveillance of financial markets, analysis of macro financial linkages, and surveillance of macroeconomic conditions. According to the Financial Sector Assessment Handbook (World Bank, IMF, 2005), these four key elements play distinct roles in financial stability analysis. I.
Surveillance of financial markets helps to assess the risk that a particular shock or a combination of shocks will hit the financial sector. Models used in this area of surveillance include early warning systems (EWSs). Indicators used in the analysis include financial market data and macro-data, as well as other variables that can be used for constructing early warning indicators.
II.
Macroprudential surveillance tries to assess the health of the financial system and its vulnerability to potential shocks. The key quantitative analytical tools used for macro prudential surveillance are the monitoring of financial soundness indicators (FSIs) and the conducting of stress tests. Those tools are used to map the conditions of non-financial sectors into financial sector vulnerabilities. The analysis also draws on qualitative data such as the results of assessments of quality of supervision and the robustness of financial infrastructure.
III.
Analysis of macro financial linkages attempts to understand the exposures that can cause shocks to be transmitted through the financial system to the macro economy. This analysis looks at data such as (a) balance sheets of the various sectors in the economy and (b) indicators of access to financing by the private sector (to assess the extent to which private owners would be able to inject new capital to cover the potential losses identified through macro prudential surveillance).
IV.
Surveillance of macroeconomic conditions then monitors the effect of the financial system on macroeconomic conditions in general and on debt sustainability in particular. In general, assessing financial stability is a complex process. In practice, the assessment requires several iterations. For example, the effects of the financial system on macroeconomic conditions may produce feedback effects on the financial system. The profile of risks and vulnerabilities (ascertained through macro prudential surveillance) could feed into qualitative assessments of effectiveness of supervision, and those effects, in turn, might influence the analysis of vulnerabilities and overall assessment of financial stability.
Detecting Financial Vulnerability from MPIs
4
2.2. Analysis of Macro-financial Linkages A clear understanding of the transmission channels that exist between the financial and real sectors of the economy is crucially important when assessing financial stability. A strong financial system can be seen as one that does not adversely induce the propagation and magnification of disturbances that affect the financial system and those that are capable of withstanding shocks and limiting disruptions in the allocation of saving to profitable investment opportunities. Given the importance of this topic to regulatory authorities, the Basel Committee on Banking Supervision (BIS, 2011) has identified three important transmission channels as; I. II. III.
The borrower balance sheet channel; The bank balance sheet channel; and The liquidity channel The first two channels are often referred to as the financial accelerator channel; the third channel emphasizes the liquidity position of banks’ balance sheets. The borrower balance sheet channel – applies to both firms and households; it comes from the inability of lenders to (a) assess fully borrowers’ risks and solvency, (b) monitor fully their investments, and/or (c) enforce fully their repayment of debt. The bank balance sheet channel, predicts that adverse shocks to financial institutions’ balance sheets can entail sharp contractions in credit and result in such shocks having magnified effects on economic activity. Two conditions are necessary for such amplified effects to occur: (a) the inability of banks to fully insulate their supply of lending in response to such shocks and (b) borrowers to be highly dependent on banks for credit. The liquidity channel, which is the third theoretical transmission channel, emphasizes the importance of a liquidity channel as a determinant of banks’ ability to extend credit and in turn to affect real economic variables, either in influencing the strength of the traditional bank-lending channel or in creating additional transmission channels. High leverage ratios, large maturity mismatches in banks’ balance sheets and mark to market accounting have been highlighted as critical elements in the propagation of funding liquidity shocks to bank lending and the real economy2. 2.3. Monetary Policy and Macroprudential Analysis The global financial crisis has also brought to fore the strong complementarity between monetary and prudential policies. According to Borio and Shim (2007), “a sound financial system is a prerequisite for an effective monetary policy; just as a sound monetary environment is a prerequisite for an effective prudential policy”. A weak financial system undermines the effectiveness of monetary policy measures and can overburden the monetary authorities; a disorderly monetary environment can easily trigger financial instability and render void the efforts of prudential authorities. From the perspective of the build-up of financial imbalances, the key question is how best to calibrate tools to address the potential excessive procyclicality of the financial system. There is, at least theoretically, a wide range of tools. This paper considers a sub-set of those most typically regarded as being of a prudential nature. Besides efforts to promote a better risk management culture, these include: loan provisioning rules, capital standards, loan-tovalue (LTV) ratios, measures to address currency mismatches and, more generally, the intensity of the supervisory review process. It should be noted, though, that a range of instruments considered of a monetary nature, such as reserve requirements and restraints on lending, could in fact perform a very similar function. Indeed, they have often been operated alongside, or as an alternative to, prudential tools3.
3. Detecting Financial Vulnerability using Early Warning Systems This section takes a look at the process involves in measuring and monitoring vulnerabilities that build up in the upswing. Risk should be measured and monitored continually over the length of time that the indicators
The empirical analysis of macro-financial linkages is outside the scope of this paper. For a more detailed exposition on the literature on transmission channels between the financial and real sectors, see BIS (2011). 3 The empirical analysis of the interactions between of monetary policy variables and macroprudential indicators is outside the scope of this paper, but is a consideration for future research. 2
Detecting Financial Vulnerability from MPIs
5
of financial distress or signs of vulnerabilities hold sway. Experts say this is usually over the horizon of 2-4 years (Borio, 2006). There are a number of indicators of financial imbalances that serve as ‘early warning signals’ or predictors of system-wide distress, such as banking crises. We now take a look at how to construct early warning systems (EWS). Constructing EWS Using Macroprudential Indicators (MPIs): Regulators and private market participants have over time developed a system of early warning indicators by which an attempt is made to predict the likelihood or otherwise of occurrence of financial crisis. Early warning systems (EWS) help policy makers to know at an early stage when a country is heading for a crisis in order to take preventive measures. These measures are often referred to as ‘macro-prudential indicators’ (MPIs). There are at least three steps in the construction of EWS: Step 1: Identify potential leading macroprudential indicators (MPIs) and collate large pool of data. Some of these MPIs include both macroeconomic indicators and bank performance indicators: [A] Macroeconomic Indicators: Asset Price Misalignments - These have to do with measures that involve the coexistence of excessive asset price increases or misalignments with a limited capacity of the system to absorb the asset price reversal (Borio, 2006). Asset prices here include real equity prices and real property prices. Misalignments are usually captured by unusually large deviations of asset prices from their long-term trend. Figure 1 shows Nigeria’s all share index from 1998 to 2013. The all share index tracks the general market movement of all equities listed on the stock exchange. The banking industry consolidation of 2004 was the beginning of the boom period for the capital market as banks that couldn't meet up with the CBN’s minimum capital requirements of N25 billion resorted to the capital market. This recapitalization process led to increased capital market deepening, which paved way for the capital boom experienced between 2004 and 2008. Figure 1: NSE All Share Index (1998-2013)
NSE All Share Index 70,000.00
Banking Industry Consolidation started in 2004
60,000.00 50,000.00 40,000.00
ASI
30,000.00
2 per. Mov. Avg. (ASI)
20,000.00
120 per. Mov. Avg. (ASI)
0.00
1998M1 M10 M7 M4 2001M1 M10 M7 M4 2004M1 M10 M7 M4 2007M1 M10 M7 M4 2010M1 M10 M7 M4 2013M1
10,000.00
Source: NSE Credit-to-GDP-gap - This measures deviations of the credit to private sector/GDP ratio from the long run trend. This indicator has been recently developed as a guide variable for taking capital buffer decisions. However, the use of the credit–GDP-gap as an indicator of systemic risk has been criticised on several grounds: (a) it lags the business cycle (i.e. when GDP growth is low or negative, the gap still tends to be high
Detecting Financial Vulnerability from MPIs
6
and vice versa (for example, in UK – see Repullo and Saurina, 2011). This is partly because much of the new borrowing by business and households is drawn on pre-negotiated lines of credit; (b) Variations in the denominator can cause spurious movements in the ratio (Gordy, 2011); (c) the indicator is unreliable because it is sensitive to the choice of detrending method used (e.g. in U.S – see Edge and Meisenzahl, 2011). Figure 2 shows the ratio of Credit to Private Sector to GDP in Nigeria. The credit gap is estimated as deviation from a 15 year moving average (when data is annualised). Using the Hodrick Prescott (HP) filter as a trend removal technique, we can derive the credit-GDP gap (see appendix 1 for details). Figure 2: Credit to Private Sector/GDP (1981-2014)
CPS/GDP (%) 40.00
CreditGDP gap
35.00 30.00
(CPS/GDP) (%)
25.00 20.00
15 per. Mov. Avg. ((CPS/GDP) (%))
15.00 10.00
2 per. Mov. Avg. ((CPS/GDP) (%))
0.00
1981 1983 1985 1987 1989 1991 1993 1995 1997 1999 2001 2003 2005 2007 2009 2011 2013
5.00
Source: CBN Credit growth - Traditionally, credit growth is a measure of asset quality as it measures the riskiness of banks. For example, using Bankscope data for 16,000 banks in 16 major countries, Foos et al. (2010) finds that abnormal loan growth is associated with an increase in loan loss provisions in the subsequent three years, a reduction in relative interest income and lower capital ratios. Following from more recent studies on the unreliability of the credit gap, credit growth has been found to be a much better common reference point for countercyclical capital buffer (e.g. Repullo and Saurina, 2011). Figure 3: Growth in Credit to Private Sector (1998-2013)
CPS Growth 18,000,000.00 16,000,000.00 14,000,000.00 12,000,000.00
Build up to the global Ninancial crisis 2007-09
10,000,000.00
CPS/GDP
8,000,000.00
2 per. Mov. Avg. (CPS/GDP)
6,000,000.00
120 per. Mov. Avg. (CPS/GDP)
4,000,000.00 0.00
1998M1 M11 M9 M7 M5 M3 2003M1 M11 M9 M7 M5 M3 2008M1 M11 M9 M7 M5 M3 2013M1
2,000,000.00
Source: CBN
Detecting Financial Vulnerability from MPIs
7
Figure 3 shows credit growth in Nigeria from 1998 to 2013. From 2006, after the recapitalization process in the Nigerian banking industry, there was a massive build up of credit to the private sector, owing mainly to the robust capital base and strength of Nigerian banks, who were rated among the largest in Africa. As Foos et al. (2010) predicts, the exponential growth rate of credit may have been associated with a benign credit risk management culture and the subsequent accumulation of toxic assets in many Nigerian banks the wake of the financial crisis. Short-term capital inflows/GDP - A high incidence of short-term capital inflows may introduce fragility risk into the financial system to the extent that the reversal of such inflows (especially during a recession) can have huge consequences on banks’ solvency. If the ratio of short-term capital flows to GDP rises above a predetermined threshold, it could give signals of an impending distress. There are several other leading indicators of financial system vulnerabilities. For example, Krugman (1979) suggested a set of typical indicators of crisis in emerging markets. They include, among other factors: Budget Deficit/GDP - Rising budget deficit as a % of GDP may imply increasing central bank financing of fiscal activities and hence loss of reserve money. A good example is the impact of the current debt crisis in the euro zone and the U.S, which has seen rising fiscal deficits, dwindle the banking system reserves. Central Bank Credit to the Public Sector/GDP - This also measures the central bank’s direct financing of the public sector and reduction in the base money. Reduction in base money implies that the central bank has limited reserves to rescue banks in a period of bank distress. Foreign Exchange reserves/GDP - A continuous fall in international reserves (e.g. via a fall in exports or rise in fiscal deficit) may signal domestic currency depreciation and hence a potential future speculative attack (i.e. investors supply more of the local currency by purchasing and keeping more of foreign currency).
[B] Bank Performance Indicators Laeven and Valencia (2008) and other banking crisis researchers identify some bank performance indicators, which are crucial for observing the pathology of a banking distress: Non-performing loans (NPL) - An increasing trend in the ratio of nonperforming loans to total loans signals a deterioration in the quality of credit portfolios and, consequently, in financial institutions’ cash flows, net income, and solvency. It is often helpful to supplement this information with information on nonperforming loans net of provisions, and on the ratio of provisions plus interest suspension on impaired loans to total loans—particularly if impaired loans have not yet been classified as nonperforming. Deposits as a share of monetary aggregates- A decline in the ratio of deposits to M2, for example, may signal a loss of confidence and liquidity problems in the banking system. It could also indicate that nonbank financial institutions are more efficient in that they offer an array of other financial products, or they are acting as banks in all but in name, or they may have set up pyramid schemes. Interest rate on deposits- Higher interest rates on deposits and lower lending rates may signal increasing competition, which may lead to greater risk taking. Though, competition amongst banks is good, it often leads to interest volatility. However, higher real interest rates could be a function of the central bank’s inflation stabilization policies. Figure 4 shows retail credit spread for the Nigerian banking system between 1998 and 2013. Interest spreads are a measure of bank competition. Narrower gaps between PLR and 3MDR signals increasing competition, which may lead to greater risk taking.
Detecting Financial Vulnerability from MPIs
8
5
10
15
20
25
Figure 4: Retail Spread (PLR-3MDR) - (1998 – 2013)
1998m1 2000m1 2002m1 2004m1 2006m1 2008m1 2010m1 2012m1 2014m1 time PLR
3MDR
Source: CBN In addition, there are some other determinants of banking crises, which relate to the structure of the banking system and more generally of financial markets. They include: capital adequacy, degree of loan concentration, the liquidity of the interbank market and of the bond market, the ownership structure of banks (public versus private; foreign versus domestic, etc) and the quality of regulatory supervision. All these play an important role in breeding banking crises, but are usually neglected because of lack of data (Demirguc- Kunt and Detragiache, 2001). The IMF (2000) has specified a comprehensive list of macroprudential indicators for analysing the health and stability of the financial system. These MPIs comprise both aggregated microprudential indicators of the health of individual financial institutions and macroeconomic variables associated with financial system soundness. Aggregated microprudential indicators are primarily contemporaneous or lagging indicators of soundness; macroeconomic variables can signal imbalances that affect financial systems and are, therefore, leading indicators. Financial crises usually occur when both types of indicators point to vulnerabilities, that is, when financial institutions are weak and face macroeconomic shocks. Table 1 shows these MPIs at a glance. Table 1: Indicators for Macroprudential Surveillance Aggregated Microprudential Indicators
Macroeconomic Indicators
Capital adequacy Aggregate capital ratios Frequency distribution of capital ratios
Economic growth Aggregate growth rates Sectoral slumps
Asset quality Lending institution Sectoral credit concentration Foreign currency-denominated lending Nonperforming loans and provisions
Balance of payments Current account deficit Foreign exchange reserve adequacy External debt (including maturity structure) Terms of trade
Detecting Financial Vulnerability from MPIs
9
Loans to loss-making public sector entities Risk profile of assets Connected lending Leverage ratios
Composition and maturity of capital flows
Borrowing entity Debt-equity ratios Corporate profitability Other indicators of corporate conditions Household indebtedness
Interest and exchange rates Volatility in interest and exchange rates Level of domestic real interest rates Exchange rate sustainability Exchange rate guarantees
Management soundness Expense ratios Earnings per employee Growth in the number of financial institutions
Lending and asset price booms Lending booms Asset price booms
Earnings and profitability Return on assets Return on equity Income and expense ratios Structural profitability indicators Liquidity Central bank credit to financial institutions Segmentation of interbank rates Deposits in relation to monetary aggregates Loans-to-deposits ratios Maturity structure of assets and liabilities (liquid asset ratios) Measures of secondary market liquidity
Inflation Volatility in inflation
Contagion effects Trade spillovers Financial market correlation Other factors Directed lending and investment Government recourse to the banking system Arrears in the economy
Sensitivity to market risk Foreign exchange risk Interest rate risk Equity price risk Commodity price risk Market-based indicators Market prices of financial instruments, including equity Indicators of excess yields Credit ratings Sovereign yield spreads Source: IMF (2000)
Step 2: Define a crisis episode or scenario: According to Laeven and Valencia (2008), a systemic banking crisis is characterised by a large amount of defaults by financial institutions and corporations who face huge difficulties in settling contracts on time. Consequently, non-performing loans increase markedly and all or most of the total banking system capital is exhausted. This situation may be accompanied by falling asset prices (such as equity or real estate prices) on the build-up to the crisis, rapid increases in real interest rates and a reduction or reversal of capital flows. In
Detecting Financial Vulnerability from MPIs
10
some cases, a banking crisis is triggered by depositor run4 on banks. Off course, withdrawal of deposits can no longer be used to date banking crises as deposit insurance and all forms of government liquidity support now exists. However, following recent crisis episodes, one can argue that modern financial crises stem from the asset side of the balance sheet- especially from poor asset quality and low liquidity. The poor performance of banking stocks relative to the overall equity market is also a typical indicator of a crisis episode. Step 3: Analyse the indicators before a crisis: Once a typical crisis episode can be defined, the third step is to analyse the indicators before a crisis. Goldstein, Kaminsky and Reinhart (2000) provide an exposition on the analysis of indicators and crisis signals. A crisis signal indicates a departure from normal behaviour of the variable in consideration. It raises an alarm of probable incidence of future crisis. An alarm is defined as a predicted probability of crisis above some threshold level (the cut off point). For example, if the government’s fiscal deficit as a percentage of GDP rises beyond a certain threshold, it could signal an impending crisis. If an indicator sends a signal that is followed by a crisis within a reasonable time frame (known as the signalling window)5, it is called a good signal. If the signal is not followed by a crisis within that interval, it is called a false signal or noise. A threshold is defined as a certain percentile of the frequency distribution of the indicator variable, below or above which a variable sends a signal.
4. Detecting Financial Vulnerability Using Financial Soundness Indicators (FSIs) 4.1. The FSI Methodology In 2006, the IMF developed a comprehensive set of financial soundness indicators (FSIs) to help national supervisors assess the soundness of the aggregate financial system. The methodology is contained in the IMF’s Compilation Guide: Financial Soundness Indicators (IMF, 2006). The guide combines elements of macroeconomic frameworks (including monetary statistics), bank supervisory frameworks (as embodied in the work of the Basel Committee on Banking Supervision), and international financial accounting standards (to a great extent referring to the International Financial Reporting Standards (IFRS)). See Table 2 for a detailed breakdown of the FSIs, which consists of 12 core FSIs for deposit takers and 27 encouraged FSIs. These FSIs are used in detecting vulnerability in the financial system as financial innovation and risk taking deepens. Table 2: Financial Soundness Indicators: The Core and Encouraged Set Core set Deposit Takers Capital Adequacy
Regulatory capital to risk-weighted assets Regulatory Tier 1 capital to risk-weighted assets Nonperforming loans net of provisions to capital
Asset Quality
Nonperforming loans to total gross loans Sectoral distribution of loans to total loans
Earnings and Profitability
Return on assets Return on equity
Bank runs can be said to occur when there is a monthly percentage decline in deposits in excess of 5%. Deposits here refer to the sum of demand deposits and time, savings and foreign currency deposits for total deposits in national currencies. (See Laeven and Valencia, 2008). 5 This time frame is called a signalling window and it is usually about 12 months for a banking crisis and 24 months for a currency crisis. 4
Detecting Financial Vulnerability from MPIs
11
Interest margin to gross income Noninterest expenses to gross income Liquidity
Liquid assets to total assets (liquid asset ratio) Liquid assets to short-term liabilities
Sensitivity to Market Risk
Net open position in foreign exchange to capital Encouraged set
Deposit Takers
Capital to assets Large exposures to capital Geographical distribution of loans to total loans Gross asset position in financial derivatives to capital Gross liability position in financial derivatives to capital Trading income to total income Personnel expenses to noninterest expenses Spread between reference lending and deposit rates Spread between highest and lowest interbank rate Customer deposits to total (non-interbank) loans Foreign-currency-denominated loans to total loans Foreign-currency-denominated liabilities to total liabilities Net open position in equities to capital
Other financial corporations
Assets to total financial system assets Assets to GDP
Nonfinancial corporations sector
Total debt to equity Return on equity Earnings to interest and principal expenses Net foreign exchange exposure to equity Number of applications for protection from creditors
Households
Household debt to GDP Household debt service and principal payments to income
Market Liquidity
Average bid-ask spread in the securities market/foreign exchange market Average daily turnover ratio in the securities market/foreign exchange market
Real estate markets
Real estate prices Residential real estate loans to total loans Commercial real estate loans to total loans
Source: Compilation Guide: Financial Soundness Indicators (IMF, 2006)
4.2. FSIs and their Signaling Properties Gadanecz and Jayaram (2009) have put together a number of quantitative measures used in the literature (including the FSIs) to assess financial vulnerability. Their study summarizes some of these measures, their frequency, what they measure, as well as their signalling properties. The focus is on six main sectors. Firstly, the real sector is described by GDP growth, the fiscal position of the government and inflation. GDP growth reflects the ability of the economy to create wealth and its risk of overheating. The fiscal position of
Detecting Financial Vulnerability from MPIs
12
the government mirrors its ability to find financing for its expenses above its revenue (and the associated vulnerability of the country to the unavailability of financing). Inflation may indicate structural problems in the economy, and public dissatisfaction with it may in turn lead to political instability. Secondly, the corporate sector’s riskiness can be assessed by its leverage and expense ratios, its net foreign exchange exposure to equity and the number of applications for protection against creditors. Thirdly, the household sector’s health can be gauged through its net assets (assets minus liabilities) and net disposable income (earnings minus consumption minus debt service and principal payments). Net assets and net disposable earnings can measure households’ ability to weather (unexpected) downturns. Fourthly, the conditions in the external sector are reflected by real exchange rates, foreign exchange reserves, the current account, capital flows and maturity/currency mismatches. These variables can be reflective of sudden changes in the direction of capital inflows, of loss of export competitiveness, and of the sustainability of the foreign financing of domestic debt. Fifthly, the financial sector is characterized by monetary aggregates, real interest rates, risk measures for the banking sector, banks’ capital and liquidity ratios, the quality of their loan book, standalone credit ratings and the concentration/systemic focus of their lending activities. All these proxies can be reflective of problems in the banking or financial sector and, if a crisis occurs, they can gauge the cost of such a crisis to the real economy. Lastly, variables relevant to describe conditions on financial markets are equity indices, corporate spreads, liquidity premia and volatility. High levels of risk spreads can indicate a loss of investors’ risk appetite and possibly financing problems for the rest of the economy. Liquidity disruptions may be a materialization of the market’s ability to efficiently allocate surplus funds to investment opportunities within the economy. Typically financial stability analysis would use several sectoral variables either individually or in combinations. The use of such measures including the financial soundness indicators as key indicators of financial stability depends on the benchmarks and thresholds, which would characterise their behaviour in normal times and during periods of stress. In the absence of benchmarks, the analysis of these measures would depend on identifying changes in trend, major disturbances and other outliers (Worrel, 2004).
5. Detecting Financial Vulnerability Using Stress Testing Techniques Another important strategy to managing financial stability is the use of stress testing and scenario analysis. Stress tests cover a range of methodologies. Complexity can vary, ranging from simple sensitivity tests to complex stress tests, which aim to assess the impact of a severe macroeconomic stress event on measures like earnings and economic capital. Stress tests may be performed at varying degrees of aggregation, from the level of an individual instrument up to the institutional level. Stress tests are performed for different risk types including market, credit, operational and liquidity risk.
5.1. Micro-Stress Testing Micro stress tests are stress tests conducted at the level of individual financial institutions. Stress testing is an important risk management tool that is used by banks as part of their internal risk management and through the Basel II capital adequacy framework, is promoted by supervisors. Stress testing alerts bank management to adverse unexpected outcomes related to a variety of risks and provides an indication of how much capital might be needed to absorb losses should large shocks occur. While stress tests provide an indication of the appropriate level of capital necessary to endure deteriorating economic conditions, a bank alternatively may
Detecting Financial Vulnerability from MPIs
13
employ other actions in order to help mitigate increasing levels of risk. It is a tool that supplements other risk management approaches and measures. A stress test is commonly described as the evaluation of a bank’s financial position under a severe but plausible scenario to assist in decision making within the bank. The term is also used to refer not only to the mechanics of applying specific individual tests, but also to the wider environment within which the tests are developed, evaluated and used within the decision-making process. Stress testing involves the use of various techniques to assess a financial institution’s potential vulnerability to “stressed” business conditions. Typically, this relates, among other things, to the impact on the institution’s profitability and capital adequacy. Stress testing has played an increasingly important role in the risk management of banks in recent years. Stress scenarios deals with the assessment of the resilience of financial institutions and the financial system to severe but plausible scenarios. Past events have shown that extreme market movements or crises can have an adverse impact on a bank’s business; therefore, it is inadequate to manage risks only on the basis of “normal” business conditions. When a bank is affected by a severe market shock, it may incur substantial losses as a result of the following: §
Assumptions of how markets behave during normal conditions no longer hold true;
§
New concentrations of risk emerge through unexpected linkages of different markets;
§
Rapid price movements and squeeze in liquidity across multiple markets;
§
Sudden deterioration in the economic conditions of affected countries; and difficulties in unwinding or hedging positions during a crisis amongst others
The above illustrates the importance of employing stress-testing techniques to estimate a bank’s likely losses under adverse conditions so that it can be better prepared for such situations. Apart from considering the effects of exceptional events, a bank may vary the level of adversity (e.g. by including mildly stressed scenarios) to assess its vulnerability under different situations.
Micro-stress Tests: Some Examples with Liquidity Stress Tests Liquidity Measurement: Liquidity is the ability of a bank to fund increases in assets and meet obligations as they come due, without incurring unacceptable losses. According to Principle 5 of the BIS (2008) principles for sound liquidity risk management and supervision, ”a bank should have a sound process for identifying, measuring, monitoring and controlling liquidity risk”. A bank should identify: its liquidity needs; sources of liquidity; the interactions between funding liquidity risk and market risk; and the interactions between liquidity risk and other types of risk. Liquidity measurement involves assessing a bank’s cash inflows against its outflows and the liquidity value of its assets to identify the potential for future net funding shortfalls. A bank should be able to measure and forecast its prospective cash flows for assets, liabilities, offbalance sheet commitments and derivatives over a variety of time horizons, under normal conditions and a range of stress scenarios, including scenarios of severe stress. Time Horizons: A bank should identify and measure changes in liquidity needs and funding capacity: (a) on an intra-day basis, (b) on a day-to-day basis, (c) over short and medium term horizons up to one year, and (d) on a longer term basis over one year. It should also measure vulnerability to events, activities and strategies that can put a significant strain on internal cash generation capability. Areas to manage liquidity risk: A bank should identify, measure and control its liquidity risk positions for: (a) future cash flows of assets and liabilities, (b) sources of contingent liquidity demand and related triggers associated with off-balance sheet positions; (c) currencies in which a bank is active; and (d) correspondent, custody and settlement activities.
Detecting Financial Vulnerability from MPIs
14
Using Maturity Ladder to Calculate a Bank’s Cumulative Net funding Requirement One plausible way a bank can monitor and control its liquidity is to set limits to constrain the size of cumulative contractual cash flow mismatches (e.g. cumulative net funding requirement as a percentage of total liabilities). The analysis of a bank’s net funding requirements involves the construction of a maturity ladder and the calculation of a cumulative net excess or deficit of funds at selected maturity dates. Constructing a maturity ladder: Maturity ladder should be used to compare a bank’s future cash inflows to its outflows over a series of specified time periods. Cash inflows arise from (a) maturing assets, (b) saleable non-maturing assets, (c) established credit lines that can be tapped. Cash outflows arise from (a) liabilities falling due, and (b) contingent liabilities especially committed lines of credit that can be drawn down. The construction of a maturity ladder is based on some sort of prediction of the future behaviour of assets, liabilities and off-balance sheet items. For example, cash inflows can be ranked by the date on which assets mature or by a conservative estimate of when credit lines can be drawn down. Similarly, cash outflows can be ranked by the date on which liabilities fall due, the earliest date a liability holder could exercise an early repayment option, or the earliest date contingencies can be called. Table 3 illustrates a hypothetical use of maturity ladder for monitoring cash inflows and outflows.
Table 3: Maturity Ladder based on Contractual Maturities – Day 1-30 DAY$1:$ CASH$INFLOWS Maturing)assets
CASH$OUTFLOWS 100 Maturing)liabilities)with)contractual)maturities
Excess/(Shortfall) 50
Interest)receivable
20 Interest)payable
10
Asset)Sales
50 Other)deposit)runoffs
30
Draw)downs)on)Commitment)Lines Total DAY$2:$ CASH$INFLOWS Maturing)assets
10 Draw)downs)on)commitment)lines 180 Total
10 140
CASH$OUTFLOWS 100 Maturing)liabilities)with)contractual)maturities
Excess/(Shortfall) 70
Interest)receivable
25 Interest)payable
20
Asset)Sales
55 Other)deposit)runoffs
40
Draw)downs)on)Commitment)Lines
10 Draw)downs)on)commitment)lines
50
Total DAY$3C15:$ CASH$INFLOWS Maturing)assets
190 Total
180
CASH$OUTFLOWS 130 Maturing)liabilities)with)contractual)maturities
90
50 Interest)payable
30
Asset)Sales
60 Other)deposit)runoffs
40
Total
20 Draw)downs)on)commitment)lines 260 Total
10 Excess/(Shortfall)
Interest)receivable Draw)downs)on)Commitment)Lines
40
60 220
40
DAY$16C$30:$ CASH$INFLOWS Maturing)assets
CASH$OUTFLOWS 160 Maturing)liabilities)with)contractual)maturities
Excess/(Shortfall) 130
Interest)receivable
80 Interest)payable
60
Asset)Sales
90 Other)deposit)runoffs
80
Draw)downs)on)Commitment)Lines
40 Draw)downs)on)commitment)lines
80
Total
370 Total
350
Detecting Financial Vulnerability from MPIs
20
15
Maturity Ladder under Alternative Scenarios We can also analyze the liquidity position of a bank using “what if” scenarios, which is usually categorized under three: § Going concern condition – assumes normal behaviour of balance sheet related cash flows in the ordinary course of banking business. § Bank-specific crisis – assumes a liquidity crisis at the individual bank level. This scenario assumes many of the bank’s liabilities cannot be rolled over and are repayable on maturity. It can also stem from the inability to honour deposit maturities. These could lead to implications for winding down a bank’s book to some extent. § General market crisis: Here, all banks suffer liquidity problems in one or more markets. Even if banks believe the central bank would intervene in the markets, it is impossible to rule out the possibility of a severe market condition (e.g. the recent global financial crisis). A macro-prudential supervisor or the central bank may find the general market crisis scenario to be of particular interest when surveying the liquidity profile of the entire banking sector. The collective results would suggest the size of the total liquidity buffer in the banking system and the likely distribution of liquidity problems among large institutions if the banking system as a whole experiences a shortage of liquidity. Under each scenario, a bank should try to account for any significant positive or negative liquidity swings that could occur. A hypothetical example is provided in table 4. Table 4: Maturity Ladder under Alternative Scenarios CASH INFLOWS
NORMAL BANKBUSINESS SPECIFIC CONDITIONS CRISIS
GENERAL MARKET CRISIS
Maturing assets
100
100
90
Interest receivable
20
20
10
Asset sales
50
60
0
Draw downs
10
0
5
Total
180
180
105
Maturing liabilities
50
50
50
Interest payable
10
10
10
Deposit runoffs
30
100
60
Draw downs on lending commitments
50
60
75
Total
140
220
195
(40)
(90)
CASH OUTFLOWS
Liquidity Excess (Shortfall) 40
Early Warning Indicators of Liquidity Risk § Rapid asset growth, especially when funded by potentially volatile liabilities § Growing concentrations in assets or liabilities § Increases in currency mismatches § Rising loan delinquencies § Significant deterioration in the bank’s earnings, asset quality, and overall financial condition § Negative publicity § A credit downgrade § Stock price declines or rising debt costs § Widening wholesale or retail funding costs
Detecting Financial Vulnerability from MPIs
16
§ § § § §
Counterparties beginning to request additional collateral for credit exposures or that resist entering into new transactions Correspondent banks that eliminate or decrease their credit lines Increasing retail deposit outflows Difficulty assessing long term funding Difficulty placing short term debt (e.g. commercial papers)
5.2. Macro-Stress Testing The role of macro-stress tests in predicting the probability of system-wide distress is gaining increasing attention amongst macro-prudential policy makers. Macro-stress tests are analogous to micro-stress tests now routinely carried out by individual financial institutions to evaluate the risks hidden in their portfolios, but relate to the financial system as a whole or a large proportion thereof, such as the banking sector. The specific methods used range from simple sensitivity analyses to more complex scenario testing. Sensitivity analysis is generally intended to assess the output or outcome from quantitative models when certain inputs or parameters are stressed or shocked. In most cases, sensitivity analysis involves changing inputs or parameters without relating those changes to an underlying event or real world outcomes. For example, sensitivity test might explore the impact of varying declines in equity prices (such as by 10%, 20%, 30%, etc) or a range of increases in interest rates (such as 100, 200 or 300 basis points) on the financial system or the macroeconomy – See BIS (2009) for more details. Macro-stress tests could involve identifying new threats to the financial system or new sources of systemic risk. The underlying idea of macro stress tests and scenario analyses is to assess the vulnerability of the financial system to adverse shocks using reasonable but very tough circumstances (such as a major recession or an asset price collapse) and to evaluate the financial strength of institutions in withstanding such shocks. In emerging market countries in particular, the focus of macro-stress tests is usually on foreign currency and interest rate mismatches (Borio, 2006).
6.
Detecting Financial Vulnerability from MPIs using Impulse Response Function
Impulse response analysis in time series analysis is important in determining the effects of external shocks on the variables of the system. Simply put, an Impulse Response Function (IRF) shows how an unexpected change in one variable at the beginning affects another variable through time. It is so widely applicable that we can use a hypothetical analysis of the relationship between CPI and oil prices and the relationship between nominal exchange rate (NER) and oil prices in Nigeria. It should be emphasized that we are not looking at how one variable (oil prices, for example) affects another variable (CPI, for example). We can easily look at the coefficients to know that. What we are looking for is how unexpected changes that directly affect oil prices affect CPI or NER. In a sense, we are looking at shocks coming from the error term related to oil prices, and how such shocks change CPI or NER. Impulse Response Function and Structural VAR Models IRFs are estimated by first specifying a structural VAR model. A structural VAR model assumes that structural, economic shocks, which drive the dynamics of the economic variables, are assumed to be independent, which implies zero correlation between error terms as a desired property. This is helpful for separating out the effects of economically unrelated influences in the VAR. For example, there is no reason why an oil price shock (as an example of a supply shock) should be related to a shift in consumers’ preferences towards a style of clothing (as an example of a demand shock). Therefore one would expect these factors to be statistically independent. A structural VAR with p lags (sometimes abbreviated as SVAR) can be specified as follows:
B0 yt = c0 + B1 yt−1 + B2 yt−2 + ...+ Bp yt− p + ε t , Detecting Financial Vulnerability from MPIs
17
where:
c0 is a k x 1 vector of constants; Bi is a k x k matrix (for every i=0, …, p); yt−1 , yt−2 ,..., yt− p are lagged
values of the vector,
yt over time to the p-th order; and ε t is a k x 1 vector of error terms, which satisfies the
conditions that the error term has a mean of zero, and in particular that there is no correlation of error terms across time [i.e. no serial correlation in individual error terms or Cov(ε1 , ε 2 ,..., ε t ) = 0 ]. Hypothetical Example: Macroeconomic effects of oil price increases: Oil price increases are generally thought to increase inflation and reduce economic growth. However, this depends on whether the country in consideration is a major net producer or net consumer of oil. For Nigeria, which relies almost solely on oil revenues, an increase in oil pries is good news for the economy as the increase in revenue from windfall gains will fuel economic activities. In terms of inflation, oil prices directly affect the prices of goods made with petroleum products. From a microeconomic point of view, oil prices indirectly affect costs such as transportation, manufacturing, and heating. The increase in these costs can in turn affect the prices of a variety of goods and services, as producers may pass production costs on to consumers. The extent to which oil price increases lead to consumption price increases depends on how important oil is for the production of a given type of good or service. Oil price increases can also stifle the growth of the economy through their effect on the supply and demand for goods other than oil. Increases in oil prices can depress the supply of other goods because they increase the costs of producing them. In economics terminology, high oil prices can shift up the supply curve for the goods and services for which oil is an input. High oil prices also can reduce demand for other goods because they reduce wealth, as well as induce uncertainty about the future (Sill 2007). One way to analyze the effects of higher oil prices is to think about the higher prices as a tax on consumers (Fernald and Trehan 2005). The simplest example occurs in the case of imported oil. The extra payment that local consumers make to foreign oil producers can now no longer be spent on other kinds of consumption goods
Relationship between Oil prices and Key Macroeconomic Variables Despite these effects on supply and demand, the correlation between oil price increases and macroeconomic variables is not perfect. To understand the relationship between oil prices and macroeconomic variables, let’s graph the series for oil prices, headline CPI and nominal exchange rate in Nigeria to observe long-term trend (note that due to unavailability of monthly GDP data, I could not plot monthly oil prices with GDP).
0
50
100
150
Figure 5: Co-movement of Oil prices (Bonny Light) and Headline CPI (1998-2013)
1998m1 2000m1 2002m1 2004m1 2006m1 2008m1 2010m1 2012m1 2014m1 time OP
HCPI
Detecting Financial Vulnerability from MPIs
18
Note from figure 5 that there is actually no clear-cut relationship between oil prices and consumer prices for Nigeria. However, it can be observed that oil prices and consumer prices seem to move in the same direction in the long run, though oil prices tend to be more volatile than CPI inflation. The reason for the unclear relationship is not farfetched. First, due to the effect of petroleum subsidies, increases in crude oil price are not immediately translated to increases in retail oil prices, and hence on consumer prices. Second, not all increases in headline inflation is caused by oil price shocks. CPI increases could be driven by cost-push or wage inflation or even excessive increases in monetary aggregates. In addition, oil price increases in Nigeria always fuel expansionary fiscal policy and this could lead to inflationary pressures. This latter factor might also explain the long-term co-movement.
0
50
100
150
Figure 6: Co-movement of Oil prices and Exchange Rate (N/US$1)
1998m1 2000m1 2002m1 2004m1 2006m1 2008m1 2010m1 2012m1 2014m1 time OP
NER
Note from figure 6 that as oil prices increase, nominal exchange rates fall, indicating an appreciation of the naira, and vice versa. The reason for the appreciation of naira during an oil price increase is connected to the increase in foreign exchange reserves, which the windfall from oil price increases makes possible. With higher foreign exchange reserves, the value of the naira appreciates against the dollar, because the central bank has enough reserves to defend the exchange rate. So, while the relationship between oil price shocks and Nigeria’s nominal exchange rate can be understood, the relationship between oil price shocks and CPI inflation is still inconclusive. Therefore, to determine whether the relationship between oil prices and other variables has truly changed over time, one must go beyond casual observations and appeal to econometric analysis (which allows researchers to control for other developments in the economy when studying the link between oil prices and key macroeconomic variables). Using STATA to Analyse IRFs, Dynamic Multiplier functions and FEVDs As noted earlier, an IRF measures the effect of a shock to an endogenous variable on itself or on another endogenous variable. Of the many types of IRFs, irf create estimates the five most important: simple IRFs, orthogonalized IRFs, cumulative IRFs, cumulative orthogonalized IRFs, and structural IRFs. A dynamicmultiplier function, or transfer function, measures the impact of a unit increase in an exogenous variable on the endogenous variables over time. irf create estimates simple and cumulative dynamic-multiplier
Detecting Financial Vulnerability from MPIs
19
functions after var. The forecast-error variance decomposition (FEVD) measures the fraction of the forecasterror variance of an endogenous variable that can be attributed to orthogonalized shocks to itself or to another endogenous variable. In other words, the FEVD measures the contribution of each type of shock to the forecast error variance. Of the many types of FEVDs, irf create estimates the two most important: Cholesky and structural. To analyze IRFs and FEVDs in STATA, you first fit a model, then use irf create to estimate the IRFs and FEVDs and save them in a file, and finally use irf graph or any of the other irf analysis commands to examine results (See Appendix 2 for a simulation with Nigerian data on the effect of oil price shocks on headline CPI inflation and nominal exchange rate).
References [1] Allen, F. (2005) “Modelling Financial Instability”, National Institute Economic Review No 192, 57-67. [2] Bank for International Settlements (2008) “Sound Principles for Liquidity Risk Management and Supervision”, Basel Committee on Banking Supervision, BIS, Basel, Switzerland [3] Bank for International Settlements (2009), ‘Principles for Sound Stress Testing Practices and Supervision’, Basel Committee on Banking Supervision, BIS, Basel, Switzerland [4] Bank for International Settlements (2011), “The transmission channels between financial and real sectors: a critical survey of the literature”, Working Paper No 18, Basel Committee on Banking Supervision, Basel, Switzerland [5] Bardsen, G., Lindquist, K. and Tsomocos, D.P. (2006), “Evaluation of Macroeconomic Models for Financial Stability Analysis”, Norges Bank Working Paper, ANO 2006/1, Financial Markets Department, Oslo [6] Borio, C. (2006), ‘The Macro-prudential approach to regulation and supervision: where do we stand? Kapittel 7, Erfaringer og utfordringer Kredittilsynet 1986-2006 [7] Borio, C. and I. Shim (2007) “What can (macro-) prudential policy do to support monetary policy?”, BIS Working Papers No. 242, Bank for International Settlements [8] Crockett, A. (1997), ‘Maintaining financial stability in a global economy’, in Federal Reserve Bank of Kansas City’s Symposium, Jackson Hole, Wyoming, 28-30 August. [9] Demirguc-Kunt, A. and Detragiache, E. (2001),’Financial Liberalisation and Financial Fragility, in G. Caprio, P. Honohan and J. Stiglitz, eds., Financial Liberalization: How far, How fast? (Cambridge: Cambridge University Press) [10] Edge, R. and Meisenzahl, R. (2011), ‘The unreliability of credit-to-GDP ratio gaps in real-time and the implications for countercyclical capital buffers’, May 15, Federal Reserve Board [11] Espinoza, J.J. (2011), “Estimating Responses to Shocks in Germany’s Macroeconomy: Impulse Response Function (IRF)”, Quantitative and Applied Economics, Feb 20, 2011 [12] Fernald, J., and B. Trehan (2005) “Why Hasn’t the Jump in Oil Prices Led to a Recession?”, FRBSF Economic Letter 2005-31. [13] Foos, D., Norden, L., and M. Weber (2010), ‘Loan Growth and Riskiness of Banks’, Journal of Banking and Finance 34: 2929-2940 [14] Foot, M., (2003), ‘Protecting Financial Stability-How good are we at it?’, speech given at University of Birmingham, June 6, 2003. [15] Gadanecz. B. and K. Jayaram (2009) “Measures of financial stability – a review”, IFC Bulletin No. 31, pp. 365-380
Detecting Financial Vulnerability from MPIs
20
[16] Goldstein, M., Kaminsky, G., and Reinhart, C. (2000), Assessing Financial Vulnerability: An Early Warning System for Emerging Markets, Washington, DC, Institute for International Economics [17] Gordy, M. (2011), ‘Leaning against the Leverage Cycle: Why and How to Implement a Countercyclical Capital Buffer’, Paper delivered at the Credit Scoring and Credit Control XII Conference, Edinburgh, August 24-26 [18] Hodrick, R. J., and E. C. Prescott (1997) “Postwar U.S. business cycles: An empirical investigation” Journal of Money, Credit, and Banking 29: 1–16. [18] International Monetary Fund (2000) “Macroprudential Indicators of Financial System Soundness, Occasional Paper 192, IMF, Washington D.C. [19] International Monetary Fund (2006) “Financial Soundness Indicators Compilation Guide”, International Monetary Fund, available at: http://www.imf.org/external/pubs/ft/fsi/guide/2006/ [20] Issing, O. (2003), ‘Monetary and Financial Stability: Is there a Trade-off?’, paper delivered to Conference on ‘Monetary Stability, Financial Stability and the Business Cycle’, Bank for International Settlements, Basel, March 28-29, 2003. [21] King, R. G. and S. T. Rebelo (1993) “Low frequency filtering and real business cycles,” Journal of Economic Dynamics and Control, 17, 207–231. [22] Krugman, P. (1979) ‘A Model of Balance of Payments Crises’, Journal of Money, Credit and Banking, 11(3): 311-325 [23] Laeven, L. and Valencia, F. (2008) Systemic Banking Crises: A New Database, IMF Working Paper WP/08/224 [24] Mishkin, F.S. (1994), ‘Global financial instability: framework, events, issues’, Journal of Economic Perspectives, 13: 3-25. [25] Minsky, H. (1978), ‘The Financial Instability Hypothesis: A restatement’, Thames Papers on Political Economy. [26] Nier, E.W. (2011), Macro-prudential Policy - Taxonomy and Challenges, National Institute Economic Review No 216, April: R1-R15 [27] Ravn, M. O. and H. Uhlig (2002) “On adjusting the Hodrick–Prescott filter for the frequency of observations,” Review of Economics and Statistics, 84, 371–376. [28] Repullo, R. and J. Saurina (2011), ‘The Countercyclical Capital Buffer of Basel III: A Critical Assessment, Centre for Economic Policy Research, Discussion Paper Series No. 8304 [29] Sill, K. (2007), “The Macroeconomics of Oil Shocks”, FRB Philadelphia Business Review, 2007:Q1 [30] World Bank, IMF (2005) Financial Sector Assessment: A Handbook, joint publication of the World Bank and the IMF. [31] Worrell, D (2004): “Quantitative assessment of the financial Sector: an integrated approach”, International Monetary Fund Working Papers, no 153.
Detecting Financial Vulnerability from MPIs
21
Appendix 1: Detrending Macroeconomic Time Series: An Example with deriving the Credit GDP gap using HP Filter Hodrick and Prescott (1997) motivated the Hodrick–Prescott (HP) filter as a trend-removal technique that could be applied to data that came from a wide class of data-generating processes. In their view, the technique specified a trend in the data, and removing the trend could filter the data. The smoothness of the trend depends on a parameter λ. The trend becomes smoother as λ → ∞. Hodrick and Prescott (1997) recommended setting λ to 1,600 for quarterly data. For other frequencies, Ravn and Uhlig (2002) have shown that quite different values should be used: 6.25 for annual data, and 129,600 for monthly data. King and Rebelo (1993) showed that removing a trend estimated by the HP filter is equivalent to a high-pass filter. They derived the gain function of this high-pass filter and showed that the filter would make integrated processes of order 4 or less stationary. To derive the Credit GDP gap, we begin by applying the HP high-pass filter to cps and plotting the periodogram of the estimated business-cycle component. A periodogram is a starting tool for identifying the important frequencies (or periods) in the observed series. The periodogram graphs a measure of the relative importance of possible frequency values that might explain the oscillation pattern of the observed data. . tsfilter hp ip_hp = cps, gain (hpgcps ahp) . label variable hpgcps "HP filter" . pergram ip_hp, xline (0.03125)
6.00 -6.00 -4.00 -2.00 0.00
2.00
4.00
cps cyclical component from hp filter Log Periodogram -6.00 -4.00 -2.00 0.00 2.00 4.00 6.00
Sample spectral density function
0.00
0.10
0.20 0.30 Frequency
0.40
0.50
Evaluated at the natural frequencies
Because the HP filter is a high-pass filter, the high-frequency stochastic cycles corresponding to those periods below 6 remain in the estimated component. We can also graph the HP filter (hpgcps), the CPS cyclical component (ip_hp) and the angular frequency from the HP filter (ahp or the trend line) by specifying the STATA syntax . tsline hpgcps ip_hp ahp
Detecting Financial Vulnerability from MPIs
22
1000000 2000000 0 -2000000 -1000000
1998m1 2000m1 2002m1 2004m1 2006m1 2008m1 2010m1 2012m1 2014m1 time HP filter angular frequency from hp filter
cps cyclical component from hp filter
-1000000
0
cpsgap
1000000
2000000
The CPS gap is derived by subtracting the log of the actual CPS and the potential CPS, which is the angular frequency from HP filter. The Credit gap is shown in the figure below:
1998m1 2000m1 2002m1 2004m1 2006m1 2008m1 2010m1 2012m1 2014m1 time
Detecting Financial Vulnerability from MPIs
23
Appendix 2: Impulse Response Function (IRF) for the Effect of Oil Price Shocks on Consumer Prices and Exchange Rate Volatility Steps in performing IRF Step 1: Prepare dataset for VAR Analysis STATA Syntax use "/Users/victorekpu/Desktop/Nigerian Dataset.dta" Setting time variable format for monthly data: . generate time =tm(1998m2)+_n-1 . format time %tm . tsset time Create new variables, take the logs of op, ner and hcpi: . generate y=ln (op) . generate x1=ln (ner) . generate x2=ln (hcpi) Take the first difference of y, x1 and x2 . generate dy = D.y . generate dx1 = D.x1 . generate dx2 = D.x2 Rename dy, dx1 and dx2 as: . rename dy dln_op . rename dx1 dln_ner . rename dx2 dln_hcpi Step 2: Fit a structural form VAR model yt =(dln_op, dln_ner, dln_hcpi) Where: dln_op is the first difference of the natural log of oil prices dln_ner is the first difference of the natural log of nominal exchange rate, and dln_hcpi is the first difference of the natural log of headline CPI inflation STATA syntax . mat a = (., 0, 0\0,.,0\.,.,.) . . mat b = I(3) . svar dln_op dln_ner dln_hcpi, aeq(a) beq(b)
Detecting Financial Vulnerability from MPIs
24
SVAR Results: Iteration 0: log likelihood = -489.51027 Iteration 1: log likelihood = 446.08251 Iteration 2: log likelihood = 512.62755 Iteration 3: log likelihood = 666.93037 Iteration 4: log likelihood = 750.50618 Iteration 5: log likelihood = 854.86543 Iteration 6: log likelihood = 866.05438 Iteration 7: log likelihood = 866.68587 Iteration 8: log likelihood = 866.68768 Iteration 9: log likelihood = 866.68769 Structural vector autoregression ( 1) [a_1_2]_cons = 0 ( 2) [a_1_3]_cons = 0 ( 3) [a_2_1]_cons = 0 ( 4) [a_2_3]_cons = 0 ( 5) [b_1_1]_cons = 1 ( 6) [b_1_2]_cons = 0 ( 7) [b_1_3]_cons = 0 ( 8) [b_2_1]_cons = 0 ( 9) [b_2_2]_cons = 1 (10) [b_2_3]_cons = 0 (11) [b_3_1]_cons = 0 (12) [b_3_2]_cons = 0 (13) [b_3_3]_cons = 1 Sample: 1998m5 - 2013m10 Overidentified model
No. of obs = 186 Log likelihood = 866.6877
-----------------------------------------------------------------------------| Coef. Std. Err. z P>|z| [95% Conf. Interval] -------------+---------------------------------------------------------------/a_1_1 | 11.4704 .5947127 19.29 0.000 10.30479 12.63602 /a_2_1 | 0 (constrained) /a_3_1 | -.0667469 .8410579 -0.08 0.937 -1.71519 1.581696 /a_1_2 | 0 (constrained) /a_2_2 | 10.05799 .5214826 19.29 0.000 9.035905 11.08008 /a_3_2 | -.507805 .7379576 -0.69 0.491 -1.954175 .9385653 /a_1_3 | 0 (constrained) /a_2_3 | 0 (constrained) /a_3_3 | 64.60483 3.349604 19.29 0.000 58.03972 71.16993 -------------+---------------------------------------------------------------/b_1_1 | 1 (constrained) /b_2_1 | 0 (constrained) /b_3_1 | 0 (constrained) /b_1_2 | 0 (constrained) /b_2_2 | 1 (constrained) /b_3_2 | 0 (constrained) /b_1_3 | 0 (constrained) /b_2_3 | 0 (constrained) /b_3_3 | 1 (constrained) -----------------------------------------------------------------------------LR test of identifying restrictions: chi2( 1)= 1.078 Prob > chi2 = 0.299
Detecting Financial Vulnerability from MPIs
25
Step 3: Create IRF File . irf create order1, step(10) set(nigeriairf1) (file nigeriairf1.irf created) (file nigeriairf1.irf now active) (file nigeriairf1.irf updated) Step 4: Graph Orthogonalized IRF Note that the “oirf” command yield “orthogonal impulse response functions” which in this case correspond to selecting a Cholesky decomposition for the contemporaneous effects matrix (Espinoza, 2011). . irf graph oirf, impulse(dln_op) response(dln_hcpi) . irf graph oirf, impulse(dln_op) response(dln_ner)
order1, dln_op, dln_hcpi
.002
0
-.002
-.004 0
5
10
step 95% CI
orthogonalized irf
Graphs by irfname, impulse variable, and response variable
The blue line above represents the impulse response function and the grey band is the 95% confidence interval for the IRF. Notice that a shock to oil prices increases CPI inflation rate and the effect becomes statistically significant one 1 month after the shock and nominal exchange rates reduces for up to 2 months after the shock before rebounding as can be seen in the graph below. Notice that the response of CPI inflation to an unexpected increase in oil prices lasts initially for 1 month and then dies out in the subsequent month and then we see another spike in inflation in the third month, as some feedback effects of the initial shock reverberates through the system.
Detecting Financial Vulnerability from MPIs
26
order1, dln_op, dln_ner .02
0
-.02
-.04 0
5
10
step 95% CI
orthogonalized irf
Graphs by irfname, impulse variable, and response variable
Step 5: Compare Structural IRFs and Orthogonal IRFs To see whether the shapes of the structural IRFs and the structural FEVDs are similar in the two models, we type: . irf graph oirf sirf, impulse(dln_op) response(dln_hcpi) . irf graph oirf sirf, impulse(dln_op) response(dln_ner)
order1, dln_op, dln_ner
order1, dln_op, dln_hcpi .02 .002
0 0
-.02
-.002
-.004
-.04 0
5
10
step 95% CI for oirf orthogonalized irf Graphs by irfname, impulse variable, and response variable
0
5
10
step 95% CI for sirf structural irf
95% CI for oirf orthogonalized irf
95% CI for sirf structural irf
Graphs by irfname, impulse variable, and response variable
The graphs reveals that the oirf and the sirf estimates are not essentially the same for both models and that the shapes of the functions are somewhat different for the two models, implying that the structural shocks on oil prices will impact differently on consumer prices than it will on exchange rates. For example, a windfall increase in global oil prices will cause a reduction in the exchange rate (i.e. an appreciation of the naira), while causing an increase in inflationary pressures.
Detecting Financial Vulnerability from MPIs
27
Step 6: Compare Structural IRFs and FEVDs with the Cholesky Decomposition To see whether the structural IRFs and the structural FEVDs differ significantly from their Cholesky counterparts, we type: . irf graph fevd sfevd, impulse(dln_op) response(dln_hcpi) lstep(1) . irf graph fevd sfevd, impulse(dln_op) response(dln_ner) lstep(1) Model 1a
Model 1b
order1, dln_op, dln_hcpi
order1, dln_op, dln_ner .1
.02
.01
.05
0 0
-.01 -.05
0
5
10
0
5
step 95% CI for fevd fraction of mse due to impulse
10
step
95% CI for sfevd (structural) fraction of mse due to impulse
Graphs by irfname, impulse variable, and response variable
95% CI for fevd fraction of mse due to impulse
95% CI for sfevd (structural) fraction of mse due to impulse
Graphs by irfname, impulse variable, and response variable
This combined graph reveals that the shapes of these functions are also dissimilar for the two models. However, the graph illuminates one minor difference between them: In modela, the estimated structural FEVD is about the same with the Cholesky-based estimates, whereas in modelb the Cholesky-based estimates are slightly larger than the structural estimates. For both models, however, the structural estimates are close to the center of the wide confidence intervals for the two estimates.
Detecting Financial Vulnerability from MPIs
28
