DISCUSSION PAPER MFM Global Practice No. 14 June 2016 Swati R. Ghosh MFM DISCUSSION PAPER NO. 14 Abstract Linkages between the real and financial sectors in an economy can lead to a buildup of balance sheet mismatches of key entities—corporates, financial institutions, households, and the public sector. Once such imbalances have built up, they can make the economy vulnerable to macroeconomic shocks, whether external or domestic in origin. This paper discusses the key mismatches that can make entities vulnerable to shocks and how such vulnerability can build up during the business cycle. Against this backdrop, the paper then discusses a framework and potential indicators that may be useful to monitor such developments. These indicators are being developed as part of the MFM macro-financial monitoring effort. Corresponding author: sghosh@worldbank.org Keywords: Macro and Financial Vulnerability JEL Classifications: E32, G10. ii This series is produced by the Macroeconomics and Fiscal Management (MFM) Global Practice of the World Bank. The papers in this series aim to provide a vehicle for publishing preliminary results on MFM topics to encourage discussion and debate. The findings, interpretations, and conclusions expressed in this paper are entirely those of the author(s) and should not be attributed in any manner to the World Bank, to its affiliated organizations or to members of its Board of Executive Directors or the countries they represent. Citation and the use of material presented in this series should take into account this provisional character. For information regarding the MFM Discussion Paper Series, please contact Ivana Ticha at iticha@worldbank.org iii Monitoring Macro-Financial Vulnerabilities Swati R. Ghosh I. Introduction This paper discusses how linkages between the real and financial sectors can lead to a buildup of balance sheet mismatches of key entities—corporates, financial institutions, households, and the public sector and how, once such imbalances have built up, they can make the economy vulnerable to macroeconomic shocks, whether external or domestic in origin 1. Against this backdrop it discusses potential indicators that may be useful to monitor such developments. These indicators are being developed as part of the MFM macro-financial monitoring effort2. The paper is organized as follows. Section II provides a brief discussion of the risks associated with these different balance sheet mismatches. Section III discusses how positive shocks in the real sector—such as an upturn in domestic business cycles (which in turn are often instigated or accompanied by external developments such as capital inflows)—can interact with the financial sector and lead to a build-up of balance sheet mismatches. Section IV then describes how, once such vulnerability has been built up, a negative shock can lead to a downward spiral of credit contraction and economic downturns. Finally, section V discusses a possible set of indicators for measuring the buildup of vulnerability. II. Vulnerability to different risks arising from balance sheet mismatches Although there are several types of risks that financial and non-financial entities are typically exposed to (and many different ways of categorizing such risks), for the purposes of analyzing the potential impact of macroeconomic shocks on balance sheets there are four key risks that are relevant. These are: systematic or market risk (risks associated with changes in systematic factors such as interest rates, exchange rates, equity or other asset prices), liquidity risks (or funding risks), capital structure risks and solvency risks3. The three former risks are closely related and may all lead to the risk of insolvency when the net worth of the entity becomes negative. (Note that debtors’ insolvency risk is credit risk from the perspective of creditors). These risks tend to arise because of balance sheet mismatches. Thus maturity mismatches in the balance sheets (where typically assets are long-term and liabilities are short-term) create interest risks for the debtor (the risk that the level and/or structure of interest rates that the debtor has to pay on its outstanding stock changes), as well as liquidity risks. Currency mismatches are caused by a disparity in the currencies in which the assets and liabilities are denominated and give rise to exchange rate risks. Capital structure mismatches result from relying excessively on debt financing rather than equity. Excessive reliance on debt financing—including on short-term debt— can give rise not only to maturity mismatch (hence interest rate risks) but also funding/liquidity risks. The absence of an equity buffer can lead to financial distress when a sector encounters a shock. While payments from equity are state contingent, with profits and dividends falling in bad times, debt service payments generally remain unchanged in bad times. Table 1 summarizes these risks and balance sheet mismatches. 1 Vulnerability can be defined as a susceptibility to loss resulting from a system’s exposure to negative shocks, internal cond itions and risk management—the latter being the process that involves confronting risk either ex ante (preparing for shocks) or ex post (coping with their effects). 2 I would like to thank Lara Lambert for her contribution in the implementation of the monitoring framework discussed in this paper. 3 Other risks include operational risk associated with problems of processing, settling or taking or making delivery on trades in exchange for cash; Counterparty risks from non-performance of trading partner; Legal risk endemic in financial contracting and are separate from the legal ramifications of credit, counterparty and operational risks. Table 1. Balance sheet mismatches and key risks as they apply to different sectors Sector Exchange rate risk Interest rate risk Liquidity risk Solvency risk/Credit risk for lender Currency mismatch Capital Maturity mismatch Liabilities versus structure Assets mismatch mismatch Government Government’s debt N/A Government’s short term Liabilities of denominated in hard currency debt government and foreign currency (domestic and external) central bank versus (domestic and versus government’s their assets. external) versus liquid assets (reserves)*. Assets include governments FX discounted value assets (reserves). Short term domestic of future primary currency denominated surpluses. debt versus liquid Liabilities may domestic currency assets include implicit of the government. liabilities from pensions as well as *not all CB reserves are contingent available for govt. debt liabilities service; some may be stemming from pledged to back currency, govt. guarantees. lent to banks etc. Banks Difference between Deposits to Short-term FX debts Bank liabilities FX assets (loans) and capital ratio. (domestic and external) versus bank assets FX liabilities versus banks liquid FX and capital. (deposits/interbank assets (and ability to lines). borrow from central Bank). Short-term domestic currency debts (often deposits) versus liquid assets. Firms Debts denominated in Debt to equity Short-term debt versus Firms’ liabilities FX (domestic and ratio. liquid assets. versus present external) versus FX value of firms’ generating assets. assets. Households FX assets (deposits) N/A Short-term debt versus Liabilities versus versus FX liabilities liquid household assets. future earnings (on (often mortgages). wages and assets). Country as Net FX denominated Net external debt Short-term external debt Stock of external a whole external debt* stock (external (residual maturity) versus debt relative to debt minus liquid FX reserves of govt. both external External FX external assets) and private sector*. financial assets denominated debt relative to net held by residents minus FX stock of FDI. *FX reserves of CB/govt. and the discounted denominated plus FX reserves of banks value of future currency. and firms trade surpluses (resources for future debt service) Note: adapted from Allen et al 2 III. How Macro Financial Linkages can lead to a buildup of balance sheet mismatches during an economic upturn While there may be many channels through which entities’ balance sheets can become vulnerable, a common mechanism is through the interaction of real business cycles and the financial sector. In fact, the financial sector is inherently pro-cyclical, and amplifies cycles in the real sector (business cycles). Thus, during an economic upturn, domestic credit to the private sector increases rapidly, while during an economic downturn, credit increases less rapidly or even decreases. Some of the variation in the size of banking assets can be accounted for by fluctuations in the size of the pool of positive Net Present Value (NPV) projects that banks can finance. But some of the fluctuations in credit extension are caused by shifts in banks’ willingness to take on risky positions over the cycle—i.e. due to changes in banks’ risk appetites. When the bank manages the size of its loan book so that its risk weighted assets are maintained to be equal to its capital, and the bank assesses that the risks of lending have declined, it can expand its lending without breaching its minimum capital requirements. Conversely, when the bank assesses that the risks of lending have risen, it will reduce its lending in order to maintain the level of risk weighted assets equal to its minimum capital requirements. Consider what happens during a domestic economic boom4. Such a boom could arise for instance with a surge of capital inflows and/or low global interest rates, or other external positive developments or due to positive domestic factors. If asset prices rise with the boom, as is generally the case, the bank’s capital is bolstered by increased profitability of the bank, or the capital gains implied by the increase in asset prices. This constitutes a positive shock to the bank’s equity. If the bank were to target a fixed leverage ratio, the positive shock to equity would cause the bank to expand the size of its balance sheet5. Generally though, banks actively manage their balance sheets, through for instance, the use of value-at-risk (VaR) models. In this case, banks will actually increase their leverage—and expand the size of their balance sheets even more—during a boom because measured risks decline during a boom. In addition there is the possibility of feedback— the higher demand for assets by banks and other financial intermediaries puts further upward pressure on asset prices, which in turn strengthen banks’ balance sheets once again and leads to even greater demand for assets. Conversely, during a downturn, banks’ desired leverage declines6, and they contract their balance sheets, pushing asset prices down further. Thus, unlike other investors, banks (and other financial intermediaries) have perverse portfolio choice behavior— because their asset holdings depend on their “balance sheet capacity” and balance sheet capacity increases during booms, banks’ demand for an asset tends to rise when the price of the asset rises (and falls when the price of the asset falls). The asset price increases also raise collateral values of households and corporates and increased the demand for credit. With the improved economic outlook, there is also likely to be increased consumption and investment (although much of this investment could be directed towards real estate and other less productive investments which would further contribute to asset 4 Note that this boom can arise either due to domestic business cycle conditions or due to external conditions (including capital inflows that are driven by global liquidity conditions and interest rate differentials). 5 For example, assume that initial balance sheet is $100 of assets in marketable securities financed by $90 of debt and $10 of equity, so leverage (Assets/Equity) is 100/10=10. If price of securities rises by 1% to $101, leverage falls to 9.18 (101/11). If bank targets a fixed leverage of 10, it must increase its debt by $9 so as to increase leverage back up to 10 ((101+D)/11)=10 which implies additional D of $9). 6 That is to say, banks’ leverage itself is pro-cyclical. See H. Shin (2013). “Adapting Macro Prudential Approaches to Emerging and Developing Economies” in “Dealing with the Challenges of Macro Financial Linkages in Emerging Markets” O. Canuto and S. Ghosh eds. 3 price increases). Output increases and expectations of future asset price increases can provide a further impetus to banks’ desire to increase the size of their balance sheet and extend credit. This cycle is illustrated in Figure 1 below. During the boom, the balance sheet of financial institutions, corporates and households can become vulnerable to different types of risks. First, as banks seek to expand their lending, they are likely to finance projects of increasingly lower net present value, i.e. of increasingly lower quality and higher degree of risk. This can heighten credit risks. Second, as banks’ balance sheets expand, the liabilities side can also become more vulnerable. When banks seek to extend credit rapidly, core funding—funding that banks draw on in normal times—is likely to be insufficient (retail deposits, for instance, tend to grow slowly and in line with income growth). So other sources of funds need to be tapped. These sources are generally other domestic or foreign financial intermediaries and the funds obtained are by nature “less sticky” or more volatile in the face of shocks. Defining banks’ core liabilities as those that are less volatile in the face of shocks or uncertainty (such as retail deposits) one finds that banks’ non-core to core liabilities ratio increases during economic booms. Often the non-core funding, especially from foreign creditors, is short- term (in part this is because, other things being equal, during an economic upturn, domestic interest rates tend to rise and short term capital flows are more sensitive to interest rate differentials). Thus banks and other financial institutions can become more susceptible to both market risks (risks associated with changes in systematic factors such as interest rates, exchange rates equity and other asset prices) and liquidity risks (which can be best described as funding risks) during the process. The fact that during cyclical upturns, lending among financial institutions also tends to rise, means that systemic cyclical risks (risks that arise when financial institutions behave similarly during cyclical upturns and downturns without talking into account the spillover effects of their decisions on the system as a whole) and systemic cross-sectional risks (systemic risks that arise due to interconnectedness across financial institutions) tend to increase in tandem. The other side of the coin is an increased indebtedness of corporates and households. For corporates, there can be an increase in leverage, and short-term borrowing, including short-term foreign borrowing. For households, there can also be a rapid increase in indebtedness, sometimes in FX as well. 4 Figure 1. How Macro financial linkages can lead to balance sheet vulnerabilities over time Source: Author 5 IV. How negative shocks can be amplified through macro financial linkages in the presence of balance sheet weaknesses Once the different sectors/entities have become vulnerable through balance sheet mismatches, a negative macroeconomic shock, or even a slight domestic economic downturn, can ultimately result in an economic “bust” (severe output decline) and even a financial crisis. This is because the interactions between the financial sector and real sectors described above, and that work in reverse in the face of a negative shock, are exacerbated in the presence of balance sheet vulnerabilities. Figures 2 and 3 illustrate how shocks such as capital flow reversals, global interest rate increases etc. can get transmitted and impact the domestic economy. For instance, a reversal of short-term capital flows that were funding banks could lead to funding and liquidity problems if financial institutions have significant funding mismatches, necessitating them to reduce their asset positions7. Negative externalities related to firesales can then come into play because a generalized sell-off of financial assets causes a decline in asset prices which in turn further impairs the balance sheets of the financial intermediaries. In addition to selling assets to regain liquidity and thus sustaining losses on existing positions in the face of declining prices, financial intermediaries may also face higher margins/haircuts8, which further adds to their funding problems. The financial intermediaries may then reduce new credit extension, or raise interest rates or other costs to borrowers (externalities related to credit crunches), contributing to an economic slowdown that can raise the probability of default more widely (credit losses for the financial institutions) and setting off a cycle of adverse effects further amplifying the initial losses sustained by financial institutions. Significant losses sustained by financial institutions could ultimately represent a risk to public sector balance sheets. Conversely, though, high sovereign indebtedness could also negatively affect banks. In principle, if banks are highly exposed—with sizable holdings of sovereign bonds on their asset side—a shock or negative impact that results in heightened sovereign risk will reduce the value of banks’ assets and ultimately affect their solvency (if the latter are already in distress). Higher sovereign risk will also reduce the value of any implicit or explicit “government guarantee” which in turn would reduce banks’ debt disproportionately relative to its equity 9 . And banks’ distress would have further feedback implications for the sovereign (Figure 4). 7 A funding shortage occurs when it is prohibitively expensive to a) borrow more funds (low funding liquidity) and b) sell of its assets (low market liquidity). In short, problems only arise if both funding liquidity (high margins/haircuts, restrained lending) and market liquidity evaporates (fire sale discounts). 8 When cash lent on repo trades is lower than the market value of the collateral security, the applicable discount is referred to as a haircut. There is no haircut when government bonds are used as collateral security. In securities lending, the market value of the collateral that is posted has to be higher than the value of the securities, and this overcollateralization is referred to as the margin. 9 See V. Acharya (2011) “A Pyrrhic Victory?-Bank Bailouts and Sovereign Credit Risk” NBER WP 17136. 6 Figure 2. Illustration of how shocks can be transmitted through the domestic economy in the presence of balance sheet mismatches Source: Author 7 Figure 3. Downward spiral in banks’ liquidity and leverage Source: Brunnermeier and Pedersen (2009) Figure 4. Negative feedback between bank solvency risk and sovereign risk Source: Author 8 V. Indicators of vulnerability The preceding sections have illustrated the role that macro financial linkages can play in leading to balance sheet mismatches and a build-up of vulnerability. Ideally, the appropriate indicators would be ones that measure balance sheet mismatches (asset and liabilities) that expose entities to the different types of risks and ones that capture changes in domestic and external macro and financial conditions. However the data needed to construct entities’ balance sheets and exposures are generally not available, even at the aggregate level. A further constraint is that we would like to use indicators that are more readily available for most emerging and developing economies. Thus a more pragmatic approach is adopted by looking at indicators that mainly capture the liabilities side (which in some cases can be broken down to look at exposure to interest or FX exposure). For each of key entities, a few indicators of “buffers” that attempt to proxy for the asset side, are also included10. Specifically, the approach taken here is that of Cervantes, Jeasakul, Maloney and Ong (2014), which we modify to provide more granularity in some instances and to allow for greater country customization. The individual indicators are standardized with respect to its historical mean and standard deviation and are weighted and aggregated to create sub-indicators. These sub- indicators are further aggregated into elements and into composite indicators. It is thus possible for each aggregated risk indicator/condition to be decomposed further into its different underlying components. The modified approach is illustrated in Figure 5. As can be seen, the composite/aggregate indicators are categorized into seven buckets that capture the risks and into two buckets that capture macro-financial conditions. The first bucket of spillover risks from external environment can be thought of as capturing a country’s net exposure to the external environment or international developments, where we take key global indicators such as commodity prices, partner country GDP growth, the VIX, and weight these by the importance of the relevant channels of transmission—commodity exports, partner country share of exports, and portfolio and bank flows to GDP—for the country. We also take account of the net impact of external developments by accounting for the buffers (reserves) and the impact as measured by the Exchange and Monetary pressure (EMP) index which is the sum of exchange rate depreciation, the change in gross international reserves and money market interest rates, each weighted by the inverse of its own volatility. The second bucket of macro risks captures risks in the domestic macro environment. The indicators are made up of three subgroupings—macro stability indicators, macro outlook indicators, and macro risk perception indicators. As noted in the preceding section, macroeconomic activity can influence financial sector developments, and the outlook and expectations for macroeconomic activity can influence the outlook for financial stability. The third bucket is risks to banks. Here we attempt to capture credit risks, FX risks (FX loans/total loans), maturity risks (short-term liabilities/GDP), and leverage (capital/total assets). 10 The public sector bucket does not include any buffers yet. In the future, the tool could include a measure of fiscal space. 9 For the credit risks we weight the public sector risks, corporate sector risks and household sector risks by the importance of each of these entities in banks’ portfolios. We also look at the growth of credit to the private sector—rapid credit growth seems to precede many episodes of banking crises in both advanced and emerging market economies—although not all episodes of high credit growth are followed by crises. We also look at asset quality (NPL ratio). As buffers, we use profitability indicators (ROA and ROE) and the Capital Adequacy Ratio (CAR). The fourth bucket captures risks emanating from the public sector’s balance sheets. Ideally we would capture the mismatches and the risks from interest rate, FX, liquidity and solvency risks. In reality we are only able to use two indicators on the liability side—namely fiscal position to GDP and public debt to GDP. Note that the fiscal position to GDP is also included in the macro conditions bucket as a variable that can influence macroeconomic stability. The fifth bucket comprises of the risks emanating from the corporate sector. The indicators (which capture the liabilities side only) are corporate debt/GDP, short-term debt/total debt and total debt/total assets. As buffers we include the rate of return to assets and rate of return to equity, and stock market returns. The sixth bucket captures household risks—where we use the indicators of household debt/GDP and the unemployment rate. For buffers we take property prices and stock market returns. The seventh bucket focuses on funding and liquidity risks (possibility of loss through funding/liquidity mismatches). Market and liquidity risks for financial institutions would manifest themselves in stresses in the secondary capital markets. And developments in these markets can be mutually reinforcing. Thus the two sets of indicators considered represent i) bank funding and liquidity which measures institutions’ vulnerability to a sudden pull bank in their funding and their ability to realize assets quickly and sufficiently to meet their short term obligations and comprises the following indicators: deposit to loan ratio; credit to private sector /resident deposits; liquid assets/short term liabilities and gross foreign liabilities/GDP; and ii) secondary market funding and liquidity where the following indicators are used: currency bid-ask spread; LIBOR-OIS spread (diff between LIBOR and overnight indexed swap); TED spread (diff between 3 month interbank/LIBOR and 3 month T bills) and stock market turnover. Finally, there are two groups of indicators that capture monetary and financial conditions (eighth bucket) and investor risk appetite (ninth bucket) respectively. Under monetary and financial conditions, we take the indicators of domestic bank credit growth, short-term real interest rates, and real broad money growth. Investor risk appetite is characterized by their pricing of risk associated with the country’s assets as well as their investment decisions and the actual volatility of market prices. Thus this bucket includes the equity risk premium, USD sovereign bond spreads, domestic currency sovereign bond yields, sovereign CDS spreads, as well as portfolio flows/GDP and FDI/GDP. For the volatilities, domestic currency government bond yields, stock market returns and exchange rate movements are included. 10 Figure 5. Vulnerability indicator framework Source: Author. Note: Some variables can enter more than one bucket/category of risk or conditions as they can feed into these categories in different ways. 11 As discussed below, the methodology follows Cervantes, Jeasakul, Maloney and Ong (2014), with modifications to include more granularity and greater country customization. The full list of indicators is presented in Appendix Table 1. Figure 6. Construction of the indicators Aggregated Indicator Element rankings weighted to derive aggregated indicator ranking—equally weighted X1 Elements Sub-indicators rankings weighted to derive element indicator ranking—equally weighted e11 e12 Sub-indicators Sub-indicators Z scores of variables converted into rankings which are weighted to derive sub indicator ranking-- a11 a12 a13 a14 a21 equally weighed except in case spillover risks and credit risks Variables v11 v12 v21 v31 v32 v41 v42 v51 v5 2 Each variable is made up of an individual or a combination of data series—e.g. public debt/GDP Source: Cervantes et al. IMF WP 14/99—modified to reflect differences in weighting To summarize:  Each category is represented by an aggregate indicator ( ).  Each aggregate indicator ( ) is developed from j elements ( ).  Each element ( ) is derived from k individual or several economic and/or market sub-indicators  Each sub-indicator uses l variables ( ) derived from m data series ( ) as input. 12 Calculation of indicators Variable. 1. Normalize the variable. Compute a standardized score for each variable—i.e. normalize. Each variable k at time t can be assessed against its historical mean or a pre-defined norm and standard deviation − , = , where is the mean or norm and is the standard deviation, both for a 5 year period. When comparing two specified time periods , t and t + s, the mean/norm and the standard deviation applied in the calculation of the z score at time t=s are the same for those for time t to ensure comparability of outcomes. Note that a variable may be one way, inverted or two way depending on the variable being considered. So a z score could also be one way, inverted or two ways. Fiscal balance is a one way variable—the more positive the balance (relative to GDP) vis a vis the medium term average, the more desirable it is for macro financial stability. On the other hand, government debt is an inverted one way variable—the smaller the balance (relative to GDP) the more desirable it is for macro financial stability. Other examples include government, household and corporate debt. In contrast inflation is a two way variable—the greater the change in prices from the mean or predetermined norm in either direction, (inflationary or deflationary pressures) the less desirable it is. Finally some variables could be both a one way and two way variable depending on the context. For instance this is the case with bank credit. Bank credit is a two way variable, under Macro risks as overly strong growth could lead to overheating while a sharp slowdown or contraction could significantly affect economic activity, On the other hand, it is a one way variable under credit risks to Banks as the stronger the growth in credit (a more positive z score) the greater the risks to banks. Conversely it is a one way inverted variable under Monetary and Financial conditions as stronger growth in credit and hence better credit availability is desirable. 2. Rank the normalized variable. For risks: zero captures the lowest first percentile of risks, rank 10 is the 99th percentile and rank 5 broadly corresponds to the long term average, calculated over the 5 year period t. For conditions: i) zero captures the most risk seeking behavior, rank 10 represents the greatest risk aversion and 5 corresponds to the long term average risk appetite all relative to the 5 year period to time t; ii) zero represents the loosest monetary and financial conditions, rank 10 represents the tightest and 5 corresponds to the long term average conditions all relative to the 5 year period at time t. 13 Sub-indicator: 1. Compute the numerical ranking for each sub indicator: The score assigned to each sub indicator is calculated as an equally weighted average of the rankings assigned to the related l variables. Weights vary with the number of selected variables such that: 1 ∑ ( ) =1 Element: 1. Compute the numerical ranking for each element ( ). Next the score assigned to each element is calculated as an equally weighted average of the rankings assigned to the related sub indicators. Weights vary with the number of sub indicators, such that: 1 ∑ =1 Aggregate indicator: 1. Compute the numerical rankings for each aggregated indicator ( ). The score assigned to each aggregated indicator representing a particular risk or condition is then calculated as an equally weighted average j associated elements, Weights vary with the number of elements such that: 1 ∑ =1 Data frequency: Use latest available data. And use quarterly data wherever possible and where necessary and possible merge quarterly and annual data. 14 Appendix Table 1. Proposed set of indicators Indicator Level 1. Spillovers risks from external environment Agg. Ind. 1.1. Exposure to external developments E 1.1.1 Global shocks through trade S* 1.1.1 Commodity prices V* 1.1.1 Partner country GDP growth V* 1.1.2 Global shocks through finance S* 1.1.2 VIX V* 1.1.2 LIBOR average (of three) V* 1.1.2 Libor-OIS Euro Area bps V 1.1.2 Libor-OIS US bps V 1.1.2 Libor-OIS Japan bps V 1.2 Impact from external shocks E 1.2.1 Pressure on exchange rate S 1.2.1 EMP index V 1.3 Buffers against shocks E 1.3.1 Reserve adequacy S 1.3.1 Gross reserves/ST debt % V 1.3.1 Gross reserves/imports (months) V 1.3.1 Gross reserves to broad money % V 2. Macroeconomic risks Agg. Ind. 2.1 Macro stability E 2.1.1 Output S 2.1.1 Output gap % V 2.1.2 Price S 2.1.2 Inflation y/y % V 2.1.3 Employment S 2.1.3 Unemployment rate % V 2.1.4 Fiscal position S 2.1.4 Budget balance to GDP % V 2.1.4 Government debt to GDP % V 2.1.5 External factors S 2.1.5 Current account balance/GDP % V 2.1.5 Real effective exchange rate chg y/y % V 2.1.6 Credit to the economy S 2.1.6 Domestic credit from banks (% deviation from trend) V 2.1.6 Change in ratio of domestic credit/GDP ppts y/y V 2.1.7 Property prices S 2.1.7 House price (% deviation from trend) V 2.2. Macro outlook E 2.2.1 Production S 2.2.1 Industrial production growth % y/y V 2.2.2 Investment S 2.2.2 Real investment growth % (y/y) V 2.2.3 Trade S 2.2.3 Trade growth (exports plus imports) % y/y V 15 2.3 Market perception of country risk E 2.3.1 Sovereign funding S 2.3.1 Sovereign CDS spread (5 years) bps V 2.3.1 USD sovereign bond spread bps V 2.3.1 Domestic currency sovereign bond yield % V 3. Bank risks Agg. Ind. 3.1 Credit risks to banks E 3.1.1 Risks to banks from corporates S* 3.1.1 Corporate risk indicator V 3.1.2 Risks to banks from households S* 3.1.2 Household risk indicator V 3.1.3 Risks from sovereign S* 3.1.3 Government risk indicator V 3.2 Other risks E 3.2.1 Bank credit S 3.2.1 Growth in domestic credit from banks % y/y V 3.2.1 Change in the ratio of domestic credit from banks/GDP % V 3.2.2 Bank asset quality S 3.2.2 NPL ratio V 3.2.3 Bank leverage S 3.2.3 Capital/assets % V 3.2.4 Bank FX exposure S 3.2.4 FX loans/total loans % V 3.2.5 Bank maturity exposure S 3.2.5 Short-term liabilities/GDP % V 3.3 Buffers E 3.3.1 Bank profitability S 3.3.1 ROA annualized % V 3.3.1 ROE annualized % V 3.3.2 Bank solvency S 3.3.2 CAR V 4. Public sector risks Agg. Ind. 4.1 Public sector risk E 4.1.1 Public sector position S 4.1.1 Fiscal position/GDP V 4.1.1 Public debt/GDP V 5. Corporate sector risks Agg. Ind. 5.1 Corporate risks E 5.1.1 Corporate risks S 5.1.1 Corporate debt/GDP1 V 5.1.1 Total debt/total assets2 V 5.1.1 Short term debt/total debt2 V 5.2 Corporate buffers E 5.2.1 Profitability S 5.2.1 Rate of Return to Assets2 V 5.2.1 Rate of Return on Equity2 V 5.2.1 Stock mkt return % yoy V 16 6. Household sector risks Agg. Ind 6.1 Household risk E 6.1.1 Household risk S 6.1.1 Household debt/GDP V 6.1.1 Unemployment rate % V 6.2 Household buffers E 6.2.1 Buffers S 6.2.1 Property price growth % V 6.2.1 Stock market return % yoy V 7. Market and liquidity risks Agg. Ind. 7.1 Exposure to stress in secondary markets E 7.1.1 Market funding and liquidity S 7.1.1 LIBOR-OIS spread bps V 7.1.1 Stockmarket turnover (trading volume/capitalization) % V 7.1.1 TED spread bps V 7.1.1 Currency bid-ask spread bps V 7.2 Exposure to stress in funding and mkt liquidity at banks E 7.2.1 Bank funding and liquidity S 7.2.1 Credit to private sector/resident deposits % V 7.2.1 Customer deposit/non-interbank loans % V 7.2.1 Liquid assets / short term liabilities % V 7.2.1 Gross foreign liabilities/ GDP % V 8. Monetary and financial conditions Agg. Ind. 8.1 Monetary policy stance E 8.1.1 Short term real interest rate S 8.1.1 Short term real interest rate % V 8.1.2 Real Money supply S 8.1.2 Real broad money growth y/y % V 8.2 Availability of bank credit E 8.2.1 Domestic bank credit S 8.2.1 Growth in domestic credit from banks y/y % V 9. Risk appetite conditions Agg. Ind. 9.1 Market price levels E 9.1.1 Market prices S 9.1.1 Equity risk premium V 9.1.1 Sovereign CD spreads V 9.1.1 US$ sovereign bond spread V 9.1.1 Domestic currency sovereign bond spread3 V 9.2 Actual volatilities of market prices E 9.2.1 Volatilities S 9.2.1 of Bond yields V 9.2.1 of Equity returns V 9.2.1 of Exchange rates V 9.3 Investment decisions E 9.3.1 Investment decisions S 9.3.1 Portfolio flows/GDP V 9.3.1 FDI flows/GDP V 17 Notes: Agg. Ind = Aggregate indicator ( ); E=Element ( ); S= Sub-indicator and V = Variables( ). In worksheet: P+ represents scores where higher equals greater risk or in the case of conditions, tighter conditions/less risk appetite. P – represents scores where higher equals lower risks, or in the case of conditions, looser conditions/higher risk appetite (inverted variable) and S represents a two way variable. * = Differences in weighting. 1. This is proxied by bank loans to corporates and so is an underestimate of total corporate debt 2. Listed companies only 3. Where domestic currency bonds are not available, US dollar or Euro denominated bonds are used. 18 References Allen, M., Rosenberg, C., Keller, C., Setser, B., and N. Roubini (2002). “A Balance Sheet Approach to Financial Crisis” IMF WP/02/210. Brunnermeier M. and L.H. Pedersen (2009) “Market Liquidity and Funding Liquidity” Review of Financial Studies (2009) 22 (6): 2201-2238. Cervantes, R., Jeasakul, P., Maloney, J., and L. Ong (2014). “Miss Muffet, the Spider(gram) and the Web of Macro Financial Linkages”. IMF WP/14/99. H. Shin (2013). “Adapting Macro Prudential Approaches to Emerging and Developing Economies” in “Dealing with the Challenges of Macro Financial Linkages in Emerging Markets” O. Canuto and S. Ghosh eds. World Bank. V. Acharya (2011) “A Pyrrhic Victory? -Bank Bailouts and Sovereign Credit Risk” NBER WP 17136. 19