## Preface ## Source details **Canonical URL:** [Preface](https://www.imf.org/-/media/files/publications/gfsr/2018/april/ch1/doc/text.pdf) ## Other formats - [Markdown version](/-/media/files/publications/gfsr/2018/april/ch1/doc/text.pdf.md) - [Structured JSON version](/-/media/files/publications/gfsr/2018/april/ch1/doc/text.pdf.json) --- ### Growth-at-Risk (GaR) assessment and outlook - Chapter 1 will regularly provide a quantitative assessment of downside risks to future GDP growth from financial vulnerabilities using a Growth-at-Risk (GaR) framework. - GaR links financial conditions to the distribution of future GDP growth outcomes and provides a framework for assessing the trade-off between supporting short-term growth and protecting financial stability and future growth over the medium term. - Current assessment findings: - Over the past six months, short-term downside risks to global financial stability have increased somewhat, reflecting somewhat tighter financial conditions amid investors’ concerns about newly announced trade measures. - Still-accommodative financial conditions continue to be supportive of economic growth. - Taking a longer view, downside risks as measured by GaR remain large: easy financial conditions continue to fuel financial vulnerabilities, leaving the global economy exposed to the risk of a sharp tightening in financial conditions. - GaR model specific projections and characterizations: - One-year-ahead severely adverse outcome (5th percentile) ≈ about 3 percent or less global growth. - Three-year-ahead severely adverse outcome: projects negative global growth under current conditions. - The three-year-ahead density exhibits a much fatter left tail than the one-year-ahead density. ### Monetary policy normalization and communication - Managing the gradual process of monetary policy normalization will be tricky given elevated medium-term risks. - Key policy recommendations: - Gradual removal of monetary accommodation and clear communications will help anchor market expectations and prevent undue volatility. - Central banks should maintain accommodation, as needed, to support the recovery and ensure inflation objectives are met; when normalizing policy, central banks should do so in a gradual and well-communicated manner and provide guidance on prospective changes to policy frameworks if warranted. - Careful communication from central banks and policymakers is required to reduce the risks from a sharp tightening of financial conditions. - Risks noted: - Spike in volatility in global equity markets in early February highlighted the risk of abrupt, adverse feedback loops during asset price adjustments. - An upside inflation surprise (for example, faster-than-expected U.S. inflation possibly owing to recent fiscal expansion) could prompt faster withdrawal of accommodation, sudden decompression of term premiums, rising risk premiums, and sharp tightening of global financial conditions with adverse global consequences. - Markets are not pricing in sharply higher inflation; term premiums have moved lower and uncertainty about future inflation has declined. ### Financial vulnerabilities, leverage, and credit allocation - Indicators point to vulnerabilities from financial leverage, a deterioration in underwriting standards, and pronounced reach-for-yield behavior by investors in corporate and sovereign debt markets globally. - New gauge of riskiness of credit allocation (Chapter 2): - Metric computes the difference in vulnerability between the firms with the largest and smallest expansions in debt. - Indicator exhibits strong forecasting power for downside risks to GDP growth; predictive horizons extend up to three years. - Currently at medium to elevated levels in several countries; globally rebounded to historical average by end-2016 with potential further rise as financial conditions loosened in 2017. - Empirical findings and numerical results: - Sample for credit-allocation measures: 55 economies (26 advanced economies and 29 emerging market economies) over 1991–2016 (Worldscope listed-firm sample). - Robustness sample: 50 economies from 2000 (Orbis, listed and unlisted firms). - One standard deviation change in the change of credit-to-GDP ratio = 5.5 percentage points. - A one standard deviation increase in ∆ credit-to-GDP associated with an increase in riskiness of credit allocation of 0.12–0.25 standard deviation (range depends on measure). - A one standard deviation increase in the riskiness measure increases the odds of a systemic banking crisis by a factor of about four (sample probability context: baseline crisis probability ≈ 5 percent). - For banking-sector equity stress, a one standard deviation increase in riskiness raises the odds by a factor of 1.3 to 2, depending on measure and horizon. - In quantile regressions, interaction terms such as Change in Credit-to-GDP Ratio × Riskiness show negative coefficients (examples preserved): - Change in Credit-to-GDP Ratio × Riskiness_Leverage: –0.0549** (0.0253) and –0.0820*** (0.0288) - Change in Credit-to-GDP Ratio × Riskiness_ICR: –0.237*** (0.0467) and –0.217*** (0.0360) - Change in Credit-to-GDP Ratio × Riskiness_Debt Overhang: –0.146*** (0.0297) and –0.204*** (0.0277) - Change in Credit-to-GDP Ratio × Riskiness_Expected Default Frequency: –0.0798 (0.0749) and –0.171*** (0.0199) - Policy-relevant institutional correlates that mitigate cyclicality: - Tightening of macroprudential policy reduces cyclicality. - Greater independence of the supervisory authority reduces cyclicality (Change in Credit-to-GDP Ratio × Independence of Supervisory Authority from Bank: –0.09*** (0.02)). - A smaller government footprint in the corporate sector and greater minority shareholder protection reduce cyclicality (interaction coefficients preserved). ### Housing markets and cross-border spillovers - House price synchronization findings (Chapter 3): - House price synchronization has increased over the past three decades for 40 countries and 44 major cities; the share of variation in house price growth explained by a common global factor increased from about 10 percent to 30 percent (1971–2016). - Short-term comovement in house price gaps increases sharply around global recessions; spikes larger among major cities than at country level. - Global financial conditions contribute to synchronization even after controlling for business cycle comovement; contribution stronger in major cities in advanced economies. - Higher house price synchronization corresponds to increased downside risks to growth at horizons of up to one year; effect amplified when leverage is high (negative relationship about twice as large when leverage is higher). - Methodology and quantitative markers: - Synchronization measures include instantaneous quasi correlation (QCORR) and dynamic factor model share (rolling window = 15 years). - Network spillovers: average share of house price variance in a country accounted for by another single country increased from 1.4 percent (1990–2006) to 2.1 percent (2007–16). - Bilateral banking integration and financial openness are significant drivers of synchronization (example standardized coefficients preserved in annex tables). - Policy implications: - Monitor synchronization in addition to country-level valuations; improve granularity, timeliness, and coverage of house price data. - Macroprudential policies can influence local house price developments and may reduce a country’s house price synchronization; effectiveness is weaker in highly synchronized markets. - Exchange rate flexibility, deeper domestic real estate markets, and consumer financial protections can dampen synchronization and enhance resilience. ### Crypto assets and emerging vulnerabilities - Observations on crypto assets (overview): - Over the past year, crypto assets trading emerged as a new potential vulnerability; price volatility much higher than that of commodities, currencies, or stocks. - Total market value of crypto assets is less than 3 percent of the combined G4 central bank balance sheets. - Market composition and concentration (March 2018 reported volumes and shares): - Bitcoin accounts for 47 percent of crypto assets’ market value. - Ethereum accounts for 15 percent. - Ripple accounts for 8 percent. - Reported fiat currency volume share among pairs with fiat on one side: US dollar 71 percent; Yen about 14 percent; Euro about 11 percent. - More than 180 dedicated crypto-asset exchanges (CEs) transact in thousands of coins, average daily volume of $30 billion. - Top 14 CEs account for more than 80 percent of reported volume; top 10 crypto assets account for 82 percent of total reported volume. - Reported volumes by cryptocurrency (March 2018): Bitcoin 39%; Tether 14%; Ethereum 11%; Ripple 4%; Litecoin 3%; EOS 3%; Bitcoin Cash 3%; Ethereum Classic 2%; TRON 2%; Storm 1%; Others 19%. - Bitcoin futures volumes are a small fraction of overall activity and equal 2.3 percent of reported trading in the Bitcoin cash market on CEs (December 2017 CME and CBOE introductions noted). - Risks from market structure and investor behavior: - CEs are opaque and often unregulated; security breaches and exchange failures have caused severe losses. - Leveraged trading limits reported include 15 times, 25 times, and even 100 times; industry contacts indicate actual average leverage tends to be between 3 and 8 times. - Futures clearing members bear risks through guarantee fund obligations. - Low unconditional correlation between Bitcoin and other asset classes between September 2015 and March 2018 (example covariance entry: Bitcoin–S&P 500 covariance 0.02). - Pairwise crypto correlations (September 2015–March 2018 examples): Bitcoin–Monero 0.36; Bitcoin–Ethereum 0.35; Bitcoin–Ripple 0.28; Bitcoin–Litecoin 0.49. - Monitoring priorities and policy recommendations: - Regulators need consensus on categorization of crypto assets (security or currency) and define their role in the financial system. - Monitor leveraged trading limits and margin practices on exchanges. - Monitor extent of integration of crypto assets into mainstream investment products and potential correlation increases with traditional assets. - Prudential regulation may be needed given fraud cases, operational failures, and infrastructure weaknesses. - At present, crypto assets do not appear to pose macrocritical financial stability risks, but dramatic growth warrants vigilance. ### Financial market structure, leveraged loans, ETFs, and nonbank vulnerabilities - Leveraged loan market signals of overheating: - Global leveraged loan issuance in 2017: $788 billion (record), surpassing precrisis high $762 billion in 2007. - US leveraged loan issuance in 2017: $564 billion. - US institutional leveraged loans outstanding: almost $1 trillion. - US high-yield bonds outstanding: $1.3 trillion. - In the United States, percent of new loan issuance rated single-B or lower: about 25 percent in 2007; 65 percent in 2017. - Covenant-lite loans made up 75 percent of new institutional loan issuance in 2017. - Average debt cushion of first-lien covenant-lite loans: 15 percent (current) vs. about 33 percent (before the financial crisis). - Average recovery rates for defaulted loans: 69 percent (current cycle) vs. 82 percent (precrisis average). - ETFs and liquidity-mismatch risks: - Assets under management of ETFs invested in less liquid assets (bank loans, high-yield, and emerging market bonds) have risen rapidly to more than $140 billion. - Share of high-yield bond and emerging market bond ETF assets: still less than 5 percent of the total market value of underlying bond markets; more than tripled from 2010 to 2017. - Monthly outflows during crisis: historical maximum monthly outflow from high-yield bond investment funds was 2.5 percent (of assets) in October 2008; monthly ETF outflows now often exceed 3 percent (of assets). - Key figure note: "3 percent (of NAV) outflows" reported in context of flows for high-yield bond ETFs and regulated investment funds. - Other leverage metrics: - Synthetic CDO issuance estimated between $80 billion and $100 billion in 2017 (up from about $20 billion a year in 2014–15). - Margin debt from stock borrowing in the United States: record $580 billion (end of 2017), about 2 percent of overall market capitalization (end of 2017). - Average derivatives leverage (gross notional exposure) of an asset-weighted sample of more than 200 US- and European-domiciled bond funds: rose from 215 percent to 268 percent of assets over the past four years. - Distribution of derivatives leverage: bottom 25th percentile 100–150 percent; top 25th percentile funds with average leverage between 300 and 2,800 percent. - Assets under management of large regulated bond investment funds that actively use derivatives increased to more than $1.5 trillion, about 17 percent of the world’s bond fund sector. - Data and disclosure gaps: - No disclosure requirements for detailed leverage information for regulated investment funds in the United States; partial requirements in some European countries. - Recommendation: endorse a clear and common definition of financial leverage in investment funds, strengthen supervisory frameworks for liquidity risk management, and coordinate harmonized macroprudential approaches across jurisdictions. ### Emerging markets, low-income countries, and debt composition - More than 45 percent of LICs are at high risk or in debt distress. - Observed shifts: - Increase in the share of private sector creditors, Eurobonds, and commercial loans with shorter maturities, raising rollover and interest rate risk. - Use of collateralized debt grants holders of collateralized claims favorable treatment and may lower recovery rates for unsecured creditors; cases cited include Chad, Republic of Congo, Venezuela. - Policy recommendations: - Reduce vulnerabilities related to debt structure and attract a stable investor base (for example, local bond market development). - Minimize rollover, foreign exchange mismatch, and collateralization risks. - Ensure transparency of contractual terms for new debt, including debt issued by entities related to the sovereign. - Official creditors should emphasize timely resolution, transparent creditor coordination, and consider sustainable lending rules. ### Banking sector resilience, dollar liquidity, and cross-border funding risks - Banks have raised capital and liquidity buffers since the global financial crisis, increasing resilience; a tail of weaker banks remains. - International US dollar balance-sheet vulnerabilities: - Non-US banks’ international US dollar balance sheets rely more on short-term or wholesale dollar funding than consolidated balance sheets. - Aggregate stable funding ratio is lower for US dollar international balance sheets than for consolidated positions; international US dollar liquidity ratio is lower than reported LCRs for consolidated positions. - Use of foreign exchange swap markets has increased; swap markets have been more volatile and may not be a reliable backstop in stress. - Policy recommendations: - Develop or maintain currency-specific liquidity risk frameworks, including stress tests, emergency funding strategies, and resolution planning; enhance coordination among regulators. - Central bank swap lines should be retained to provide foreign exchange liquidity in systemic stress. - Continue completing the postcrisis reform agenda, implement macroprudential tools, and ensure independence of supervision. *This GFSR reflects information available as of March 30, 2018.* ### Preface ### Preface ### Growth-at-Risk (GaR) assessment and outlook - Chapter 1 will regularly provide a quantitative assessment of downside risks to future GDP growth from financial vulnerabilities using a Growth-at-Risk (GaR) framework. - GaR links financial conditions to the distribution of future GDP growth outcomes and provides a framework for assessing the trade-off between supporting short-term growth and protecting financial stability and future growth over the medium term. - Current assessment: over the past six months, short-term downside risks to global financial stability have increased somewhat, reflecting somewhat tighter financial conditions amid investors’ concerns about newly announced trade measures. - Even so, still-accommodative financial conditions continue to be supportive of economic growth. - Taking a longer view, downside risks as measured by GaR remain large: easy financial conditions continue to fuel financial vulnerabilities, leaving the global economy exposed to the risk of a sharp tightening in financial conditions. ### Monetary policy normalization and communication - Managing the gradual process of monetary policy normalization will be tricky given elevated medium-term risks. - Careful communication from central banks and policymakers is required to reduce the risks from a sharp tightening of financial conditions. - The spike in volatility in global equity markets in early February highlighted the risk of abrupt, adverse feedback loops during asset price adjustments. - Recently increased trade tensions have raised investor jitters; a wider escalation of protectionist measures could harm the global economy and global financial stability. ### Financial vulnerabilities, leverage, and credit allocation - Indicators point to vulnerabilities from financial leverage, a deterioration in underwriting standards, and pronounced reach-for-yield behavior by investors in corporate and sovereign debt markets globally. - Chapter 2 introduces a new gauge of the riskiness of credit allocation: - The metric computes the difference in vulnerability between the firms with the largest and smallest expansions in debt. - This indicator exhibits strong forecasting power for downside risks to GDP growth. - It is currently at medium to elevated levels in several countries. - Conventional metrics of corporate debt vulnerability— including deterioration in nonprice terms and underwriting standards—suggest rising market risks, supported by easy financial conditions, high issuance, and strong global capital flows. - In low-income countries, the share of private and non–Paris Club creditors is increasing, and greater use of collateralized debt exposes borrowers to potentially costly debt restructurings in the future. ### Housing markets and cross-border spillovers - Asset price spillovers have important implications for the housing market. - Chapter 3 finds that house price correlations across countries and major cities have been trending up during the past 30 years, suggesting that spillovers via the housing sector may play a prominent role in a future crisis. ### Crypto assets and emerging vulnerabilities - Over the past year, crypto assets trading has emerged as a new potential vulnerability. - Observations: - Price volatility of crypto assets has been much higher than that of commodities, currencies, or stocks. - Financial stability risks could arise from leveraged positions taken by investors, infrastructure weaknesses of cryptocurrency exchanges, and fraud, in addition to elevated volatility. - Policy responses: regulators worldwide are responding through enforcement actions, indirect interventions via the banking system, and outright bans. - Opportunities: crypto assets also create opportunities; a number of central banks are considering issuance of central bank digital currency. ### Policy implications and recommendations - Policymakers face twin challenges: - Continue to support growth in the short term by keeping monetary policy accommodative. - Rein in rising financial stability risks in the medium term by deploying micro- and macroprudential policy tools. - Investors and policymakers should be cognizant of risks associated with rising interest rates after years of very easy financial conditions and take active steps to reduce these risks. - Central banks and policymakers should prioritize careful communication to manage the risks of sharp tightening in financial conditions. *This GFSR reflects information available as of March 30, 2018.* ### FOREWORD ### FOREWORD ### Global economic outlook and financial conditions - The global economic outlook has continued to improve, as discussed in the April 2018 World Economic Outlook, with the pace of economic growth picking up and the recovery becoming more synchronized around the world. - Global financial conditions have tightened somewhat since the October 2017 Global Financial Stability Report (GFSR), reflecting primarily the bout of equity volatility in early February and a decline in risky asset prices at the end of March following concerns about a wider escalation of protectionist measures. - Global financial conditions remain broadly accommodative relative to historical norms across both advanced and emerging market economies despite recent tightening. - Global financial conditions have tightened somewhat, on balance, since the October 2017 GFSR, reflecting the spike in equity market volatility in early February and investors’ jitters in late March about a wider escalation of trade tensions. ### Financial stability risks and vulnerabilities - Short-term risks to financial stability have increased somewhat relative to the previous GFSR, and medium-term risks continue to be elevated. - Financial vulnerabilities have accumulated during years of extremely low rates and volatility and could make the road ahead bumpy and put growth at risk. - Growth-at-Risk analysis (described in Chapter 3 of the October 2017 GFSR) shows that risks to medium-term economic growth, stemming from easy financial conditions, remain well above historical norms. - Valuations of risky assets are still stretched, with some late-stage credit cycle dynamics emerging, reminiscent of the precrisis period. - Liquidity mismatches and the use of financial leverage to boost returns could amplify the impact of asset price moves on the financial system. - Although no major disruptions were reported during the episode of volatility in early February, market participants should not take too much comfort. ### Monetary policy, inflation, and spillovers - In advanced economies, stronger growth momentum and the firming of inflation have eased to some extent the challenge of maintaining monetary accommodation required to support the recovery while addressing medium-term financial vulnerabilities. - Inflation may pick up faster than currently anticipated, possibly propelled by significant fiscal expansion enacted in the United States. - Central banks may respond more aggressively to higher inflation than currently expected, which could lead to a sharp tightening of financial conditions with potential spillovers to risky asset prices, bank dollar funding markets, emerging market economies, and low-income countries. - Recommendation: Central banks should continue to normalize monetary policy gradually and communicate their decisions clearly to support the economic recovery. ### Banking sector, nonbank financial sector, and regulatory agenda - The banking sector has become more resilient since the global financial crisis. - Some weaker banks in advanced economies still need to strengthen their balance sheets, and some internationally operating institutions run dollar liquidity mismatches that could be exposed by market turbulence. - Regulators should complete and implement the postcrisis regulatory reform agenda. - Policymakers should address financial vulnerabilities by using micro- and macroprudential tools more actively or by enhancing their toolkits as needed—for example, to address risks in the nonbank financial sector. - Authorities in China have taken steps to address risks stemming from the interconnectedness of the banking and shadow banking sectors, but vulnerabilities remain high; further regulatory actions are crucial. ### Emerging markets, low-income countries, and spillover considerations - A number of emerging market economies have improved fundamentals during benign external financial conditions but could be vulnerable to a sudden tightening of global financial conditions or spillovers from monetary policy normalization in advanced economies. - The severity of potential shocks will differ across countries, depending on economic fundamentals and policy responses. - Recommendation: Enhance resilience via higher-quality credit intermediation, avoiding credit booms that lead to excessive risk taking, and, where feasible, permitting exchange rate flexibility. ### Crypto assets and technology - The technology behind crypto assets has potential to make financial market infrastructure more efficient. - Crypto assets have been afflicted by fraud, security breaches, operational failures, and associations with illicit activities. - At present, crypto assets do not appear to pose financial stability risks, but they could do so should their use become more widespread without appropriate safeguards. ### Findings from analytical chapters (high level) - Chapter 2: Documents that firms obtaining more credit during strong credit expansion periods are relatively riskier, especially when lending standards are loose or financial conditions are easy; increased riskiness of credit allocation signals heightened downside risks to GDP growth and higher probability of banking stress. - Chapter 3: Documents a striking increase in house price synchronization among 40 countries and 44 major cities; heightened synchronicity of house prices can signal a higher probability of adverse scenarios for the real economy, especially when credit is high or rapidly expanding. ### IMF Executive Board perspectives and policy priorities - Executive Directors welcomed the broadbased recovery, noted global growth is expected to rise further in the near term, and observed inflation remains muted in many countries. - Directors judged risks around the short-term outlook broadly balanced, but beyond the next several quarters, risks are tilted to the downside. - Key downside risks include a sharp tightening of global financial conditions, escalating trade protectionism, record-high levels of global debt, geopolitical tensions, and climate events. - Priorities recommended by Directors: - Raise potential output and ensure gains are widely shared. - Enhance economic and financial resilience and safeguard debt sustainability. - Rebuild fiscal buffers where appropriate to create room for an eventual downturn. - Calibrate fiscal adjustment to avoid procyclicality and anchor it on fiscal reforms that increase productivity and promote human and physical capital. - Use fiscal policy to facilitate growth-enhancing structural reforms in countries with ample fiscal space. - Develop and deploy micro- and macroprudential tools and closely monitor risks related to credit allocation and synchronized house prices. - Enhance multilateral cooperation to reduce trade barriers, reduce incentives for cross-border profit shifting and tax evasion, implement the postcrisis financial regulatory reform agenda, and address shared challenges such as refugees, security threats, cyber risks, and climate change. - Directors concurred that monetary accommodation should continue in advanced economies with inflation below target; where output is close to potential and inflation is rising toward target, a gradual, data-dependent, and well-communicated withdrawal of monetary support is warranted. ### Outlook for financial stability and assessment tools - Still-easy financial conditions continue to support near-term economic growth but may contribute to buildup of financial imbalances, excessive risk taking, and mispricing of risks. - The growth-at-risk (GaR) approach links financial conditions to the distribution of future GDP growth outcomes and assesses the intertemporal trade-off between supporting near-term growth and risking future growth and financial stability. - Key methodological point: Severely adverse growth outcomes are defined as those that occur with a 5 percent probability (the “tail” of the distribution), and changes in financial conditions that deteriorate these tail outcomes can signal vulnerabilities moving toward macrocritical levels. - Figure and data references in the chapter illustrate Global and regional Financial Conditions Indices (FCIs) and their components, including comparisons to a “Price of Risk” FCI and historical quintiles relative to 2017:Q3. *Source: FOREWORD, Global Financial Stability Report: A Bumpy Road Ahead, April 2018.* ### 1. Global Financial Conditions Index ### 1. Global Financial Conditions Index ### Short-term risks and medium-term vulnerabilities - Short-term financial stability risks have increased somewhat since the previous GFSR, despite broadly accommodative financial conditions dampening near-term growth risks relative to a few years ago. - Under current financial conditions, the GaR model forecasts that the severely adverse one-year-ahead outcome is for global growth to fall to about 3 percent or less (the 5th percentile). - The estimated three-year-ahead growth distribution has a much fatter left tail than the one-year-ahead distribution; under the GaR severely adverse scenario, global growth is forecast to be negative three years from now. - Continued easing of financial conditions over the past two years has improved near-term prospects while worsening medium-term growth prospects; current medium-term risks are well above historical norms. - High leverage in the nonfinancial sector remains elevated and could cause aggregate debt-service ratios to deteriorate quickly once financial conditions tighten. - Banks have raised capital and liquidity buffers since the global financial crisis, increasing resilience, but they may still be vulnerable to funding shocks. - Use of financial leverage outside the banking sector is on the rise amid prolonged low interest rates and compressed market risk measures. ### Growth-at-Risk (GaR) model findings - GaR focuses on outcomes in the 5th percentile of conditional forecast densities of global growth. - One-year-ahead: severely adverse outcome (5th percentile) ≈ about 3 percent or less global growth. - Three-year-ahead: severely adverse outcome projects negative global growth under current conditions. - The three-year-ahead density exhibits a much fatter left tail than the one-year-ahead density. - Comparison across horizons and vintages shows the current medium-term downside risk (5th percentile) is higher than at end-2016 and well above historical norms. ### Monetary policy normalization in advanced economies - Central banks must balance gradual withdrawal of accommodation with avoiding disruptive volatility; clarity in central bank communications is important. - Financial markets have adjusted relatively smoothly to gradual normalization, aided by clear communication and historically large central bank asset holdings. - Expected path of policy interest rates in the United States points to a faster pace of tightening relative to other advanced economies, consistent with gradual removal of accommodation. - As central banks raise short-term interest rates and shrink balance sheets, decompression of term premiums may cause abrupt tightening of financial conditions; a smooth normalization process is important to avoid sudden volatility and disruptions. ### Term premiums: levels, drivers, and vulnerabilities - Term premiums have remained compressed in major economies and are near historic lows. - Despite six federal funds rate increases since December 2015 by the Federal Reserve, the term premium in the United States remains near historical lows and financial conditions have continued to ease. - Factors potentially explaining muted responses of term premiums to balance sheet reduction: - Asymmetric liquidity responses: initial asset purchases may have had larger effects by ameliorating illiquidity; initial withdrawal may not have opposite effect on liquidity premiums. - Structural lowering of term premiums by central bank purchase programs, especially given lower equilibrium policy rates and expectations that asset purchases remain in the policy toolkit. - Muted signaling channel from balance sheet reduction compared with asset purchase implementation; unwinding is mapped out with a “high hurdle” for revision and is less data dependent, providing less signaling about future short-rate paths. - Model estimates suggest term premiums are broadly in line with investors’ expectations for growth, inflation, and the current stance of monetary policy; estimated term premium has remained near the lower bound of fitted model values. - However, term premiums are significantly vulnerable to revisions in expectations about inflation, growth, or the path for monetary policy. ### Financial markets vulnerability to an inflation surprise and spillovers - Markets continue to price in gradual monetary tightening; uncertainty about future inflation outcomes has diminished alongside declining term premiums. - Market participants are not pricing in a risk of sharply higher inflation over the next few years. - An upside inflation surprise (for example, faster-than-expected U.S. inflation possibly owing to recent fiscal expansion) could prompt faster withdrawal of accommodation, sudden decompression of term premiums, rising risk premiums, and sharp tightening of global financial conditions with adverse global consequences. - Emerging markets are vulnerable to spillovers from an abrupt tightening in global financial conditions: - Gradual normalization has allowed weaker issuers to access markets and broadened creditor bases toward investors more inclined to turn over portfolios. - Under realistic assumptions, a tightening cycle accompanied by higher investor risk aversion could reduce portfolio flows to emerging markets by at least one-quarter, increasing rollover risks and funding costs. - Low-income, small, non-investment-grade borrowers are particularly exposed given rising debt vulnerabilities. - Sovereign term premiums among major economies (Canada, Germany, Japan, United Kingdom, United States) have moved closely together, increasing the risk that rapid decompression of term premiums could spill over internationally. ### Nonfinancial private sector debt and house prices - Nonfinancial sector leverage has been rising in many major economies and remains high. - Economies with already-high nonfinancial sector debt are often experiencing faster house price growth. - Figures referenced (from 2017:Q3 or latest available) indicate credit-to-GDP ratios and three-year annualized real house price growth; countries with gray dots have credit-to-GDP above the historical median for their peer group. *International Monetary Fund | April 2018 — Chapter 1: A BuMPY ROAd AhEAd* ### 1. US Term Premium and Dispersion of 10-Year-Ahead ### 1. US Term Premium and Dispersion of 10-Year-Ahead ### US inflation expectations and term premium - Figure title: "US Inflation Expectations and Term Premium" - Market indications: - "Market-Implied Probability of High Inflation (Percent chance of CPI Inflation > 3 percent over five years)" (figure heading present; no numeric probability provided in text). - Ten-year term premium (left scale) and dispersion of 10-year-ahead inflation forecasts (three-quarter moving average, right scale) are shown for the United States and the Euro area. - Key qualitative finding: - "Term premiums have moved lower, and uncertainty about future inflation has declined. Markets are not pricing in sharply higher inflation." - Sources noted for the figure: "Bloomberg Finance L.P.; Federal Reserve Bank of Minneapolis; and IMF staff estimates." - Note in figure caption: "In panel 2, euro area options are illiquid. CPI = consumer price index." ### Term premium correlations and spillovers across G4 - Figure title: "Term Premium Correlations, Spillovers, and Exchange Rate Relationships" - Findings: - "Term premiums in major advanced economies move very closely together even as market expectations of policy rate paths diverge." - "Spillovers between G4 term premiums are elevated, with the United States dominating the direction." - Net pairwise spillovers: a positive (negative) value indicates that the US term premium is a shock transmitter (receiver) to German, Japanese, and UK term premiums. - Methodology: Diebold and Yilmaz (2012) time-varying spillover index using rolling generalized forecast error variance decompositions in a generalized vector autoregression model; directional spillovers use normalized elements of the variance decomposition matrix; net pairwise spillovers are total spillovers transmitted from market i to all markets j minus spillovers transmitted from all markets j to market i. - Policy-relevant point: - "Spillovers from a faster withdrawal of US Federal Reserve monetary policy accommodation in the wake of an inflation surprise and associated repricing of inflation risk and term premiums could rapidly tighten US and global financial conditions." - Such tightening "could challenge major central banks, such as the European Central Bank, that are not as far along in the normalization process, perhaps forcing them to respond through additional accommodation." - Exchange-rate transmission: - "Short rates dominate movements in exchange rates between the euro-US dollar ... and the sterling-US dollar cross rates." - "The sensitivity of currencies to expected short rate differentials has remained elevated in recent years (Figure 1.6, panels 3 and 4). This finding holds both on average over the past 20 years and for estimates for the latest sensitivity." ### Monetary policy communication and recommended approach - Core recommendations: - "Gradual removal of monetary accommodation and clear communications will help anchor market expectations and prevent undue volatility." - "Monetary authorities should maintain accommodation, as needed, to support the recovery and ensure inflation objectives are met." - "When normalizing policy, central banks should do so in a gradual and well-communicated manner." - "They should also provide guidance on prospective changes to policy frameworks if such changes are warranted." - "Gradualism and clear communications are crucial given the confluence of still relatively low inflation, easy global financial conditions, and rising financial vulnerabilities." - "To address the buildup in financial vulnerabilities and avoid putting growth at risk, policymakers should also deploy and develop appropriate micro- and macroprudential tools." ### Financial vulnerabilities: asset valuations and leverage - Overall framing: - "Against a backdrop of mounting vulnerabilities, risky asset valuations appear overstretched, albeit to varying degrees across markets, ranging from global equities and credit markets, including leveraged loans, to rapidly expanding crypto assets." - "The increasing use of financial leverage to boost returns and the growing influence of some passive investment vehicles, particularly exchange-traded funds (ETFs) in less liquid underlying markets, could amplify the impact of asset price moves on the financial system." - "Rather than a reach for yield prompted by central bank accommodation, there may be outright speculative overreach in some risky assets." - Key policy dichotomy: if valuations are not significantly out of line, gradual policy normalization and macroprudential measures suffice; if misalignments are significant and may put growth at risk, a more forceful policy response may be needed. ### Equity valuations - Evidence and metrics: - "In the United States, and through the spate of volatility beginning in early February, equity market capitalization has risen from 95 percent of GDP in 2011 to 155 percent of GDP in March 2018." - "Standard price-to-earnings and price-to-book valuation metrics remain elevated in most regions." - "Cyclically adjusted price-to-earnings" (CAPE) and other longer-term earnings-based measures continue to indicate high valuations even after volatility in February and March. - On equity risk premium: some measures conditional on interest rates suggest shares closer to fair value because "strong near-term earnings expectations, as well as historically low interest rates, sustain comparatively wide equity risk premiums." - But: "Equity valuations deteriorate under alternative, less sanguine proxies for earnings, such as longer-term averages or nominal GDP growth. Also, higher projected paths for interest rates similarly narrow the equity premium and imply richer valuations." ### Corporate bond valuations and credit quality - Key observations: - "With central banks in advanced economies continuing to lift policy rates ... the share of negative-yielding global bonds has dipped lower since the October 2017 GFSR. This ratio, however, remains significant." - "Against a backdrop of low default rates, corporate spreads remain at very low levels, even in the riskiest segments." - "Issuance of riskier bonds has surged, and the share of lower-grade bonds (BBB-rated) in the investment-grade universe has been rising." - Corporate sector metrics: - "Debt ratios—while still high—have edged lower, especially in China and other emerging markets." - "Effective interest rates paid by the corporate sector moved higher, particularly outside the United States." - "Interest coverage ratios have dipped everywhere except China and the United States." - Tax reform implications: - "Recent US tax reform will have important implications for the corporate sector. ... historical experience in the United States in the 1980s and with the repatriation tax holiday in 2004 suggests that financial risk taking often follows tax policy changes, as evidenced by heightened purchases of financial assets, mergers and acquisitions, dividends, and share buybacks." - "The cap on the tax deductibility of interest expense will reduce incentives for debt financing, which tends to affect highly leveraged companies disproportionately." - "These firms may face funding pressures because of higher interest expenses, more volatile earnings, and a more compressed schedule for adapting their funding structure to the new tax code." ### Leveraged loan market: signs of overheating - Market size and issuance: - "Global leveraged loan issuance hit a record high in 2017 of $788 billion, surpassing the precrisis high of $762 billion in 2007." - "Most issuance occurred in the United States, amounting to $564 billion." - "Since 2007, US institutional leveraged loans outstanding have doubled to almost $1 trillion, compared with $1.3 trillion in US high-yield bonds outstanding." - Use of proceeds and leverage: - "Borrowing to fund mergers and acquisitions, leveraged buyouts, dividends, and share buybacks still accounts for half of total issuance amid improving global growth." - Covenant and quality trends: - "Covenant protections have weakened over time ... leading to potentially lower recovery rates in the next default cycle." - "Covenant-lite percent of new issuance" has increased (figure panels referenced). - "Average annual first lien US loan implied recovery rates" and "Moody’s loan covenant quality index score" discussed in figures. - Market structure: - "Highly leveraged loan deals have increasingly been arranged by nonbank lenders." - "Adjustments to earnings expectations have led to less conservative leverage calculations." - Additional figure values shown: - "69% average" and "82% average" (panel annotation present in figure sequence). - Panel labels and metrics such as "Greater than six times total leverage", "All loan issuance", and "Fraction of US deals with a nonbank entity as lead agent (Percent)" are presented in the figures. ### Select exact numeric figures and labels preserved from the content - US equity market capitalization: "95 percent of GDP in 2011" and "155 percent of GDP in March 2018" - Global leveraged loan issuance: "$788 billion" in 2017; precrisis high "$762 billion" in 2007 - US leveraged loan issuance in 2017: "$564 billion" - US institutional leveraged loans outstanding: "almost $1 trillion" - US high-yield bonds outstanding: "$1.3 trillion" - Figure labels showing monetary amounts: "$7.6 trillion" and "$10.6 trillion" (appearing in figure context) - Covenant-related averages noted in figures: "69% average" and "82% average" *Sources: Bloomberg Finance L.P.; Federal Reserve Bank of Minneapolis; and IMF staff estimates.* ### CHAPTER 1 A BuMPY ROAd AhEAd ### CHAPTER 1 A BuMPY ROAd AhEAd ### Leveraged loan market: issuance, terms, and credit quality - Strong issuance and lofty valuations have coincided with weakening of nonprice terms and credit quality. - Percentage of new loan issuance rated single-B or lower in the United States: - about 25 percent in 2007 - 65 percent in 2017 - Covenant-lite loans made up 75 percent of new institutional loan issuance in 2017. - Average debt cushion of first-lien covenant-lite loans: - 15 percent (current) - about 33 percent (before the financial crisis) - Average recovery rates for defaulted loans: - 69 percent (current cycle) - 82 percent (precrisis average) - Observed behaviors and market structure changes: - Institutional leveraged loans outstanding have grown rapidly, with institutional investors increasingly important in highly leveraged loan deals. - New loans with EBITDA add-backs or adjustments that conceal deteriorated leverage metrics have reached new highs. - Regulatory actions and consequences: - US federal banking agencies guidance (March 2013) and ECB supervisory guidance (May 2017) aimed to reduce risk in leveraged transactions. - One unintended consequence: migration of activity away from banks toward institutional investors such as collateralized loan obligations, bank loan mutual funds, private equity firms, and other private funds. - Policy recommendation / supervisory gap noted in text: - Implementing comprehensive and globally consistent reporting standards across the asset management industry would give regulators better data with which to locate leverage risks; reporting standards should include enough information on derivatives to show funds’ sensitivity to large moves in underlying rate and credit markets. ### The price of volatility and option-implied measures - During turmoil in global equity markets in early February, implied volatilities derived from equity options spiked sharply from subdued levels. - The VIX term structure not only shifted higher but also inverted briefly. - Realized volatility and underlying forecasts of future volatility also increased. - Near- and longer-term equity options now appear close to, if not below, levels consistent with volatility forecasts, implying little net change in the premium investors require to compensate for volatility risk. - Similar patterns observed across other asset classes, including US dollar swaptions. ### Correlations, interconnectedness, and market turnover - Correlations have rebounded from subdued levels relative to historical norms. - Within the US stock market: - Correlations between individual stocks and across sectors picked up after major tax legislation and increased further after the February 2018 volatility spike. - Global equity market correlations also rebounded in recent months, even before the drop in global share prices. - Broader correlations across asset classes have increased, suggesting global diversification has become somewhat more difficult. - Market turnover has been relatively low, especially for high-yield bonds, which may compound price discovery distortions and illiquidity in the future. - Structural changes in the investment management industry: - Broker dealers’ intermediation role has declined, increasing the role of the non-bank sector. - Institutional investors include high-frequency trading firms and insurance companies and pension funds that may use less procyclical strategies. - These new market structures have not been tested during a significant market downturn. ### Increasing use of financial leverage and related measures - Repo activity has declined relative to market capitalization since the crisis, reflecting less use of that leverage channel. - Other forms of financial leverage appearing to rise: - Synthetic CDO issuance estimated at between $80 billion and $100 billion in 2017 (up from about $20 billion a year in 2014–15). - Margin debt from stock borrowing in the United States: - record $580 billion (end of 2017) - about 2 percent of overall market capitalization (end of 2017) - Current net exposure of investors involved in stock margin borrowing is at record negative highs relative to overall market capitalization compared with the past 25 years. - Assets under management of large regulated bond investment funds that actively use derivatives: - increased to more than $1.5 trillion - about 17 percent of the world’s bond fund sector - Average derivatives leverage (gross notional exposure) of an asset-weighted sample of more than 200 US- and European-domiciled bond funds: - rose from 215 percent to 268 percent of assets over the past four years - Distribution of derivatives leverage across funds: - bottom 25th percentile: embedded derivatives leverage in the 100 to 150 percent range - top 25th percentile: funds with average leverage between 300 and 2,800 percent - Data and disclosure gaps: - No disclosure requirements for detailed leverage information for regulated investment funds in the United States; requirements exist only on a selected basis in some European countries. - Lack of sufficient data collection and oversight by regulators compounds the risks. ### Growth of less liquid bond ETFs and liquidity-mismatch risks - Assets under management of ETFs invested in less liquid assets—bank loans and high-yield and emerging market bonds—have risen rapidly to more than $140 billion. - Share of high-yield bond and emerging market bond ETF assets: - still less than 5 percent of the total market value of underlying bond markets - more than tripled from 2010 to 2017 - Benefits of ETFs noted: - enhance price discovery, provide an alternative source of liquidity through exchange trading, facilitate hedging and diversification, charge lower fees, and offer intraday liquidity. - Risks and concerns: - Liquidity mismatches between ETFs (which offer intraday liquidity) and underlying less liquid bond assets. - Frequent trading and volatile ETF flows: - Monthly outflows from high-yield bond investment funds during the financial crisis were limited, with a maximum monthly outflow of 2.5 percent (of assets) in October 2008. - Monthly ETF outflows now often exceed 3 percent (of assets). - Sensitivity to changes in risky asset prices: - High-yield and emerging market bond ETFs’ sensitivity to S&P 500 returns is higher than the sensitivity of their underlying indices to S&P 500 returns, suggesting potential for increased contagion risk and amplification of price moves across asset markets during stress. - Greater investment in passive strategies, such as ETFs, may be related to the rise in cross-asset correlations during periods of stress, increasing contagion risk. - Empirical note: - Although some evidence shows largest holdings of high-yield bond ETFs underperforming more during days of large outflows (top 5th percentile of daily shares destroyed) in 2015–17 versus 2010–11, there is no evidence of large redemptions from these ETFs having a significant impact on pricing of the broader underlying market; their share of the underlying markets remains less than 5 percent. *International Monetary Fund | April 2018* ### 1. Assets under Management of ETFs Invested in Global High-Yield, ### 1. Assets under Management of ETFs Invested in Global High-Yield, Bank Loan, and Emerging Market Bonds (Billions of US dollars) ### ETFs and less liquid bond markets - Finding: ETFs invested in less liquid bond markets are receiving strong inflows and are owning a growing share of the underlying markets. - Finding: The investor base in these ETFs is significantly more flight prone. - Finding: Although ETFs can provide additional liquidity to less liquid bond markets, their greater sensitivity to major liquid markets increases contagion risks. - Key statistic: "3 percent (of NAV) outflows" is reported in the context of flows as a percentage of net asset values for high-yield bond ETFs and regulated investment funds. ### Market functioning, risks, and supervisory recommendations - Finding: Financial markets functioned well during the turbulence in early February, but the episode was largely confined to global equity markets. - Finding: Asset valuations remain stretched, and rising interest rates may be accompanied by a repricing of risky assets and further spikes in volatility. - Recommendation: Regulators should ensure financial institutions maintain robust risk management standards, including close monitoring and assessment of exposures to asset classes deemed to be overvalued. - Recommendation: Financial market participants should remain attuned to risks associated with rising interest rates and monetary policy normalization. - Recommendation: Given signs of late-stage credit cycle dynamics, policymakers should use macroprudential tools more actively, including standard capital- and borrower-based instruments, improved credit risk monitoring (including deterioration of nonprice terms), and strengthened investor protection. - Recommendation: Regulators should be mindful of unintended consequences of regulatory measures (for example, migration of activity toward more opaque segments of the financial system). - Recommendation: Expand the macroprudential toolkit to address risks in the nonbank financial sector, with specific actions including: - Endorse a clear and common definition of financial leverage in investment funds to improve transparency, particularly for derivatives positions. - Continue to strengthen supervisory frameworks for liquidity risk management in investment funds, monitor effectiveness of existing liquidity risk management tools, and coordinate a harmonized macroprudential approach across jurisdictions (including stress test exercises). ### Macroprudential focus on investment funds - Finding: IOSCO’s latest report on liquidity risk management for collective investment funds provides welcome guidance, but country authorities should further monitor effectiveness. - Recommendation: Harmonized and coherent macroprudential approach across jurisdictions to financial stability risks stemming from investment fund activities. --- ### Crypto Assets: New Coin on the Block, Reach for Yield, or Asset Price Bubble? ### Role, technological features, and current limitations - Finding: Crypto assets have the potential to combine benefits of traditional currencies and commodities: they can potentially be exchanged for other currencies, be used for payments, and store value. - Finding: As investment products, crypto assets may offer portfolio diversification, but this is limited by short track record, regulatory uncertainty, and primitive market infrastructure. - Finding: Distributed ledger technology (DLT) could increase efficiency of market infrastructure; DLT differs from traditional payment systems by using multiple copies of a central ledger validated and recorded as blocks (blockchain). - Finding: New units are supplied by “miners” who solve cryptographic puzzles; mining is costly in terms of energy and time. - Finding: Supply processes differ across crypto assets; some have an upper limit (for example, Bitcoin) while others may not, allowing design choices that mimic fiat money dynamics. ### Size, market share, and concentration - Finding: Even after recent price corrections, crypto assets experienced spectacular appreciation over the past year, spurred by the global reach for yield. - Finding: Crypto assets represent only a small share of the global financial system; total market value is less than 3 percent of the combined G4 central bank balance sheets. - Key statistics on market composition: - Bitcoin accounts for 47 percent of crypto assets’ market value. - Ethereum accounts for 15 percent. - Ripple accounts for 8 percent. - Finding: More than 180 dedicated crypto-asset exchanges (CEs) transact in thousands of coins, adding up to an average daily volume of $30 billion. - Concentration: - The top 14 CEs account for more than 80 percent of reported volume. - The top 10 crypto assets account for 82 percent of the total reported volume. - Fiat currency volume share among pairs with fiat on one side: - US dollar: 71 percent of volume. - Yen: about 14 percent. - Euro: about 11 percent. - Finding: In December 2017, CME and CBOE introduced Bitcoin futures; futures volumes are a small fraction of overall trading activity on these exchanges and equal 2.3 percent of reported trading in the Bitcoin cash market on CEs. ### Volatility, returns, and correlations - Finding: Bitcoin’s realized volatility is much higher than that of other asset classes. - Finding: After accounting for price volatility, risk-adjusted returns of crypto assets have not dramatically exceeded those of mainstream assets over the medium term, though they have in the most recent year. - Example: The Sharpe ratio of crypto assets was relatively close to the risk-reward ratio of the S&P 500 over the past three years, and below the return from FANG stocks (Facebook, Amazon, Netflix, Google) over that period. - Finding: The unconditional correlation between Bitcoin and other asset classes was close to zero between September 2015 and March 2018. - Covariance / correlation indicators reported (September 2015–March 2018): - Bitcoin–S&P 500 covariance entry: 0.02 (from the unconditional covariance matrix). - Pairwise correlations among selected crypto assets remain low (example entries from the crypto-asset covariance matrix): - Bitcoin–Monero: 0.36 - Bitcoin–Ethereum: 0.35 - Bitcoin–Ripple: 0.28 - Bitcoin–Litecoin: 0.49 ### Risks from market structure and investor behavior - Finding: CEs are a major source of risk due to opaque and often unregulated nature; security breaches and exchange failures have caused short-lived periods of high volatility and severe losses. - Finding: Data on trading volumes can be unreliable because CEs operate under heterogeneous rules, fee structures, investor bases, and regulatory oversight. - Finding: Leveraged trading on CEs has been set at generous limits in some cases; reported leverage limits include 15 times, 25 times, and even 100 times, though industry contacts indicate actual average leverage tends to be between 3 and 8 times. - Finding: Futures contracts on CME and CBOE raise concerns because clearing members bear risks associated with these contracts through obligations to the guarantee fund. - Finding: The combination of low asset return correlations and crypto assets’ small footprint suggests limited risk of spillovers from idiosyncratic crypto price moves to the wider market at present. - Risk channel: Integration into mainstream financial products (investment funds, ETFs, futures) could broaden the investor base and increase correlation between crypto assets and traditional assets over time, raising potential for transmission of shocks, especially during episodes of risk aversion. - Risk channel: A broad shift from fiat money toward crypto assets could partially disintermediate the banking system, curtailing prudential and safety-net functions and the central bank’s ability to act as lender of last resort. - Risk channel: Cross-border market disruptions could transmit rapidly across national boundaries given the borderless nature of transaction mechanisms and rapid growth with limited transparency. ### Monitoring priorities and policy implications - Recommendation: Regulators and the global regulatory community need to reach a consensus on what crypto assets are (for example, a security or a currency) and define their roles in the financial system. - Recommendation: Increased prudential regulation may be needed given noted fraud cases and operational failures. - Monitoring priorities highlighted: - Leveraged trading limits and margin practices on exchanges. - The extent of integration of crypto assets into mainstream investment products and potential correlation increases with traditional assets. - Risks to banking business models from partial disintermediation. - Cross-border transmission risks arising from opaque market structures and rapid growth. - Conclusion: At present, crypto assets do not appear to pose macrocritical financial stability risks, but the sector’s dramatic growth warrants vigilance by regulators. *International Monetary Fund | April 2018* ### 2. Reported Volumes by Cryptocurrency, March 2018 ### 2. Reported Volumes by Cryptocurrency, March 2018 ### Composition of Reported Volumes by Cryptocurrency - Bitcoin: 39% - Tether: 14% - Ethereum: 11% - Ripple: 4% - Litecoin: 3% - EOS: 3% - Bitcoin Cash: 3% - Ethereum Classic: 2% - TRON: 2% - Storm: 1% - Others: 19% ### Market Concentration across Cryptoexchanges - 95 percent volume = 36 exchanges (20 percent of total) - 80 percent volume = 14 exchanges (8 percent of total) - 60 percent volume = 7 exchanges (4 percent of total) - Note: Composition of reported volumes has shifted away from the Chinese exchanges. ### Bitcoin Futures versus Cash Volume - Bitcoin futures volumes remain low. - Charted metric: Ratio of volumes traded in futures versus cash (time series points include Dec. 28, 2017; Jan. 1, 18; Jan. 15, 18; Jan. 29, 18; Feb. 12, 18; Feb. 26, 18; Mar. 12, 18; Mar. 26, 18) ### Bitcoin Reported Volumes by Fiat Currency - Reported fiat currency labels referenced: US dollar, Japanese yen, Chinese renminbi, Korean won, Euro, Others - Policy/research note in figure caption: “People’s Bank of China crackdown on cryptoexchanges” ### Key Observations (textual context from surrounding chapter) - “Composition of reported volumes has shifted away from the Chinese exchanges.” - “Bitcoin futures volumes remain low.” *International Monetary Fund | Global Financial Stability Report: A BuMPY ROAd AhEAd (April 2018), chapter figure captions and accompanying text.* ### CHAPTER 1 A BuMPY ROAd AhEAd ### CHAPTER 1 A BuMPY ROAd AhEAd ### Rising Debt Vulnerabilities and More Complex Creditor Composition - More than 45 percent of LICs are at high risk or in debt distress. - Key observed changes and risks: - Increase in the share of private sector creditors, Eurobonds, and commercial loans with shorter maturities, raising rollover and interest rate risk. - Use of collateralized debt (for example, senior loans through state-owned enterprises or pledging commodity shipments) can grant holders of collateralized debt favorable treatment and may lower recovery rates for unsecured creditors; recent debt distress cases cited include Chad, Republic of Congo, Venezuela. - Details on collateralized deals remain scant. - Sovereigns typically enjoy significant protections from seizure of assets; most creditors rely on good faith negotiations for recovery in distress. - Collateralized claims could impair the sovereign’s ability to offer more generous renegotiation terms on unsecured debt and may require a more significant haircut on remaining debt to ensure debt sustainability. - New creditors who do not specialize in sovereign debt may reallocate funds rapidly if higher-yielding opportunities arise elsewhere, potentially triggering rapid market repricing at early signs of sovereign stress. ### Empirical Indicators and Figures (as presented) - Panel and sample notes (verbatim from source): - Panel 1 debt sustainability assessment ratings for LICs are based on the Debt Sustainability Framework for Low-Income Countries. - Panel 2 sample includes 35 non-investment-grade relatively small issuers that are part of JP Morgan’s EMBIG index. Each has a weight of less than 2 percent of the index. - Panel 3 is based on 37 LIDCs where continuous data are available from 2007 to 2016. - Panel 4 countries included are Cameroon, Chad, Republic of Congo, Ethiopia, Ghana, Mauritania, Mozambique, and Zambia. - Figures are in simple averages and may be overly influenced by the experience of countries with very high levels of debt. ### Policy Recommendations on Debt Management and Official Creditor Conduct - To ensure a sustainable debt burden, policymakers should: - Reduce vulnerabilities related to the structure of their debt and attract a stable investor base, including through local bond market development. - Minimize risks emanating from rollovers, potential foreign exchange mismatches, and collateralization. - Explore state contingent debt instruments that may offer protection against unforeseen shocks such as natural disasters, assuming these instruments are priced at reasonable cost for the issuer by investors (IMF 2017b). - Ensure transparency of contractual terms for new debt, including debt issued by entities related to the sovereign. - Official creditors, when needed, should: - Emphasize timely resolution of debt distress cases to avoid potential spillovers and to minimize costs for both issuer and creditors. - Encourage transparent and broad creditor coordination, especially when the set of lenders is diverse. - New official creditors should consider benefits of adopting sustainable lending rules, such as those endorsed by the Group of 20. ### Shadow Banking Reform and Risks in China — Overview - System scale and linkages: - China’s banking system: RMB 250 trillion (300 percent of GDP). - Investment products outstanding: RMB 75 trillion. - Investment products are largely funded through issuance, with roughly half sold to multiple investors as high-yielding alternatives to bank deposits and half held by single investors, including banks. - Investment vehicles invest in bonds, bank deposits, nonstandard credit assets, and other investment products; insurance companies have considerable exposure because they invest in these products and use them as funding. - Investment vehicles are largely little-regulated and create a complex web of exposure between financial institutions. - Banks’ roles and exposures: - Banks are exposed to investment vehicles as investors, creditors, borrowers, guarantors, and managers. - Small and medium-sized banking institutions and insurance companies: investment products account for one-fifth and one-third of their assets, respectively. - About one-quarter of investment vehicle assets are invested in other vehicles, creating opaque cross-holding and leverage structures. - Banks are seen as implicitly guaranteeing the RMB 25 trillion in investment products they manage. - Regulatory actions: - Since summer of 2016, regulators have incorporated bank-sponsored investment vehicles in the macroprudential framework and taken steps to curb financial sector leverage and interconnectedness. - Proposed asset management rules beginning in 2018 would limit investment vehicle leverage and complexity and gradually restrict banks from investing in these vehicles or providing financial support, effectively converting roughly half of the market from deposit-like products into mutual funds. - Insurance regulator has clamped down on sale of short-term investment products by life insurers. ### Chinese Banks: Deleveraging Progress and Remaining Risks - Reported developments: - Growth of banks’ exposure to other financial institutions fell from about 80 percent on an annual basis in 2016 to less than 20 percent at the end of 2017. - Banks’ holdings of investment products issued by other banks have declined sharply. - Ongoing vulnerabilities: - Bank buffers continue to thin at many commercial banks. - Core Tier 1 capital ratios are declining and remain near minimum levels for many small and medium-sized banks. - Preprovision profitability continues to weaken. - Reliance on short-term nondeposit funding remains high for small and medium-sized banks; short-term wholesale liabilities are still more than double available liquidity buffers at smaller banks. - Market indicators: - Money market rates have risen sharply, leading to wider corporate bond spreads, particularly for weaker borrowers. ### Reforming Investment Product Market — Risks and Adjustment Challenges - Key features and concerns: - Investment vehicles are the largest net borrower in China’s repurchase market, often with relatively illiquid collateral. - Direct lending by large banks to their sponsored vehicles amounts to about 10 percent of their investment product liabilities, on average. - Allocations to safe, liquid assets by bank-sponsored investment vehicles decreased to one-third in 2017 from about half in 2015. - Investment vehicles have bought nearly all the net increase in corporate and financial bond issuance in the past three years and hold 70 percent of such bonds outstanding. - Without bank-guaranteed fixed yields on investment products, retail investors are likely to shift toward less risky instruments, reducing net demand for illiquid corporate bonds. - Potential consequences for bank lending and capital: - Banks will need to recognize some portion of corporate credit exposure held through investment vehicles as loans or bonds, requiring capital and provisioning costs that will reduce loan growth capacity. - For small and medium-sized banks, absorbing half of these exposures over two years would reduce net new loan growth from 17 percent to 6 percent, unless banks raise new capital. - Quantitative notes (verbatim from source): - Eight banks (including four of the Big Five lenders) disclose active direct lending to their investment vehicles, which account for nearly half of the bank-managed investment product market (more than RMB 10 trillion in non-principal-guaranteed wealth management products). This lending was equivalent to 15 percent of these banks’ core Tier 1 capital as of mid-2017. - Shadow credit definition used in analysis: 100 percent of banks’ investments in third-party unconsolidated structured products and 20 percent of their sponsored non-principal-guaranteed wealth management products, based on a sample of 25 listed banks. ### China’s Insurance Sector: Growth, Risk Profile, and Linkages - Growth and market behavior: - Insurers’ assets have more than tripled in size over the past seven years. - Growth fueled by “universal life insurance” and flexible savings products in 2015 and 2016, and more traditional life policies in 2017. - Insurers’ share prices have risen sharply with increased volatility, reflecting perceived elevated risks. - Regulator recently took control of a large insurance group that financed rapid expansion into other business areas with short-term high-guarantee investment products. - Asset allocation and vulnerabilities: - Insurers pursue high guaranteed returns on long-term policies (4 percent, in many cases). - To attain guarantees amid a relatively small and illiquid corporate bond market, insurers have shifted investments from bonds and deposits to equity, funds, and “other assets.” - “Other assets” include asset and wealth management products, debt and equity products, and participations in joint ventures. - Large investments in infrastructure, real estate, and loan portfolios concentrate credit risks, including for insurers with limited expertise in credit assessment. - Industry labeling and risk perception: - More than 80 percent of outstanding wealth management products are billed as low risk, rated as 1 or 2 on an industry group–defined scale to 5 (with 5 being riskiest). *International Monetary Fund | April 2018* ### 3. China Bond Market: Corporate and Non-Policy-Bank Financial Bonds ### 3. China Bond Market: Corporate and Non-Policy-Bank Financial Bonds ### Insurers’ growth, asset allocation, and vulnerabilities - Insurers have grown rapidly, "fueled by life insurance sales." - Insurers’ shares have risen sharply, accompanied by high volatility. - Increased revenues have been invested in higher-risk assets but "capital has not been raised." - "Other assets are mainly portfolios of infrastructure projects, real estate, and loans provided by asset managers." - Variation of alternative investments and capital buffers within the sector is large. - Asset allocation of bank-issued investment products and wealth management products (selected categories, percent of product portfolios shown in figure): - Asset management products and wealth management products: 30% - Debt schemes: 27% - Equity funds and schemes: 25% - Associates and joint ventures: 12% - Real estate: 6% - Decomposition and concentration: - The "Largest 15 life insurers" cover two-thirds of the total assets of the Chinese insurance sector. - Equity performance and volatility: - Equity performance indexed to 2014 = 100 (figure reference). - Volatility measured as annualized standard deviation of relative price change for the 60 most recent trading days’ closing price. ### Risks from asset-liability mismatches and illiquidity - Uncertain and volatile returns on insurers’ alternative assets "may not match the minimum yields promised to policyholders." - Increased illiquid assets covered by deposit-like insurance products raise exposure to redemptions at short notice. - When faced with net cash outflows, insurers may need to sell illiquid assets, "potentially adding to market volatility." - Insurers are in some instances part of financial conglomerates; one-third of the consolidated balance sheets of the five largest insurance groups consists of banking, asset management, or other activities. - Whether all insurers have sufficient resilience against these vulnerabilities is uncertain. - Current regulations require relatively low capital charges for infrastructure investments, joint ventures, and real estate compared with, for instance, corporate bonds. - Capital requirements for investments in funds are fixed and not based on the risks of the underlying assets. - Despite elevated risks, "capital levels have remained unchanged." - Medium-sized and smaller insurers have invested more heavily in alternative assets and have weaker capability to manage related risks. - Risk assessments are clouded by complex and opaque company structures and uncertainty about the exact nature and credit quality of the underlying investments, including implicit guarantees. ### Quantified exposures and regulatory risk weights - About "one-fifth of life insurers’ liabilities are deposits and policyholders’ investments, which are presumed to be more easily withdrawn by policyholders than traditional life insurance products." - The risk factor applied to: - Infrastructure equity plans: 12 percent - Real estate: 8 to12 percent - 10-year AA-rated corporate bonds: 15 percent - Bond funds: 6 percent ### Policy recommendations for investment product market and insurance supervision - Authorities should continue to reform the investment product market and enhance the insurance supervisory regime. - Specific recommendations: - Further limit leverage for lower-risk products. - Eventually require that implicitly guaranteed off-balance-sheet business carry the same capital and liquidity buffers as on-balance-sheet business. - Careful sequencing of reforms is critical. - Prioritize strengthening policy frameworks and financial institutions’ liquidity and capital buffers to prevent the dismantling of implicit guarantees from inadvertently bringing forward stability risks (as recommended by the IMF’s recent Financial Sector Stability Assessment). - Address nonregulatory factors driving proliferation of risky investment products and excessive demand for credit more broadly; for instance, GDP growth targets. - Move the insurance supervisory regime toward a transparent, market- and risk-based regime that includes close cooperation with other authorities. - Increase transparency on the nature, credit quality, and valuation of "other assets." - Conduct a thorough review of prudential treatment to adequately reflect the risks of the underlying assets. - Closely monitor liability profiles—including duration and surrenders—and consider further action to curb unusual liquidity risks. - Enhance group supervision, strong cross-sector coordination, and develop a framework for recovery and resolution for the largest life insurers. - Recent supervisory actions noted: - Authorities have curtailed the sale of "universal life" policies and addressed duration mismatches. - Introduction of the China Risk-Oriented Solvency System in 2016 as a stronger prudential standard. - The recently announced merger of the China Insurance Regulatory Commission and the China Banking Regulatory Commission should facilitate closer cooperation with respect to insurance and banking supervision. ### Funding challenges for internationally active banks and dollar liquidity risks - Although banks have strengthened their consolidated balance sheets, "dollar balance sheet liquidity remains a source of vulnerability." - International dollar lending continues to increase, dominated by non-US banks operating through international branch networks. - Most rely heavily on short-term wholesale dollar funding and, at the margin, on volatile foreign exchange swap markets. - A sharp tightening of financial conditions could expose structurally vulnerable liquidity positions and trigger forced asset sales or even defaults, amplifying and transmitting market turbulence. - Non-US banks’ branches in the United States have been dollar borrowers from overseas, on net, since 2011, but gross flows in each direction remain considerable. - US subsidiaries of foreign banks gather retail dollar deposits but play little role in the international dollar system because they are limited in flexibility to transfer funds intragroup across borders or legal entities. ### Structural vulnerability metrics for international US dollar balance sheets - Non-US banks’ international US dollar balance sheets rely more on short-term or wholesale dollar funding than their consolidated balance sheets. - Short-term wholesale instruments (interbank deposits, commercial paper, certificates of deposit) and relatively unstable corporate, nontransactional, and uninsured deposits are prone to outflows and can generate refinancing risk under stressed conditions. - The use of short-term funding makes international US dollar balance sheets structurally vulnerable to liquidity risks. - Two indicators used to assess vulnerability: - A liquidity ratio that approximates the Basel Liquidity Coverage Ratio (LCR). Note: the liquidity ratio estimated here may be somewhat overstated compared with the Bank for International Settlements LCR. - A stable funding ratio (stable funding divided by loans) broadly analogous to the Basel net stable funding ratio but likely generates higher estimates since it does not apply available stable funding haircuts to wholesale deposits. - Aggregate findings: - The aggregate stable funding ratio is lower for US dollar international balance sheets than for consolidated (aggregate position in all currencies) balance sheets. - The international US dollar liquidity ratio is lower than reported LCRs for banks’ consolidated positions. - US dollar liquidity ratios vary widely between banking systems. - Composition note for stable funding calculations: - For Japan, 70 percent of swap funding is greater than one year in duration and is treated as stable, based on Bank of Japan data. - For other countries, 50 percent of swap funding is included in stable funding. ### Broader banking-sector observations relevant to global stability - Bank-level adjustments and resilience: - Markets provide mixed signals about bank health; equity price-to-book ratios vary across banks. - Banks’ consolidated financial positions have been fortified over the past decade through increased capital and liquidity, higher provisions, and improved funding profiles. - In 2007 almost 40 percent of the sample, by assets, had weak buffers and high loan-to-deposit ratios; this proportion is now less than 10 percent. - Despite aggregate improvement, a tail of weaker banks representing about 20 percent of sample assets has lower levels of capital and provisions against nonperforming loans (NPLs). - Nonperforming loans and regional concentration: - Weaker banks are mainly concentrated in Europe (inside and outside the euro area). - NPL levels remain high at some banks despite recent declines due to economic pickup, NPL reduction actions, and policy measures. - Funding profiles and liquidity: - About one-third of sample banks, by assets, still have loan-to-deposit ratios in excess of 100 percent. - Attention should continue to liquidity risks, particularly regarding dollar-funding profiles of internationally operating banks. *International Monetary Fund | April 2018 — Chapter excerpt* ### 4. US Subsidiaries of Non-US Banks: Intragroup Borrowing and Lending ### 4. US Subsidiaries of Non-US Banks: Intragroup Borrowing and Lending ### Role of non-US banks and intragroup structures - Non-US banks’ international branches are key dollar intermediation channels, while subsidiaries play a very limited role. - Dollar bonds outstanding have increased rapidly, but loans remain the largest form of credit. - Non-US banks tend to rely on short-term or wholesale US dollar funding; their US dollar liquidity is usually weaker than their overall positions. ### Developments in dollar liquidity and stable funding (2006–17) - Overall US dollar liquidity ratios have improved since the global financial crisis, largely driven by large increases in High Quality Liquid Assets (HQLA, reserves at central banks and holdings of official sector bonds). - Only the Japanese banking system’s liquidity ratio declined over the same period; it currently stands at about 100 percent. - Aggregate US dollar stable funding ratios are largely unchanged over 2006–17. - Individual banking systems have shown little progress in strengthening stable funding ratios; in some the ratio has fallen, reflecting rapid growth in dollar loans—particularly in the Canadian, French, and Japanese banking systems—that has exceeded banks’ ability or willingness to source deposits. - Systems whose stable funding ratios have improved (UK and German banking systems) accomplished this only by shrinking dollar loans. ### Use of foreign exchange swaps and market vulnerabilities - Non-US banks use foreign exchange swap markets to meet short-term currency needs; use has increased overall over the past decade. - Japanese banks rely relatively heavily on these instruments. - Cross-currency basis swap spreads have moved sharply in the past, and swap markets have been more volatile than other short-term funding sources such as repo and interbank markets—suggesting swap markets may not be a reliable backstop in periods of stress. - Demand to hedge currency risk by Asian life insurers has driven a surge in demand for swaps; US banks’ dollar swap supply has not kept up with this growing demand. - Non-traditional lenders, such as hedge funds and sovereign wealth funds, now account for about 70 percent of the supply of foreign currency derivatives to Japanese financial institutions. - About 85 percent of short-term Japanese government bills are now held by non-Japanese investors and the Bank of Japan. ### Signs of tightening in dollar funding markets - Market participants attribute tighter US dollar funding conditions to factors including an expected rise in Treasury bill issuance, US companies changing investment patterns ahead of repatriating offshore assets, and continued central bank normalization. - Dollar LIBOR-OIS spreads have widened recently, illustrating tightening. ### Risks and potential amplification channels - The combination of balance sheet vulnerabilities and market tightening could trigger funding problems in the event of market strains. - Market turbulence may make it more difficult for banks to manage currency gaps in volatile swap markets, possibly rendering some banks unable to roll over short-term dollar funding. - Funding pressures could compel banks to sell assets in turbulent markets or shrink dollar lending to non-US borrowers, reducing credit availability and potentially causing defaults on dollar obligations. ### Policy recommendations and regulatory implications - Banks should ensure that currency-specific mismatches within individual entities in their banking groups continue to be managed effectively to reduce the risk of funding strains. - Consideration should be given to enhancing disclosure of foreign currency funding risks to help investors and analysts better assess international liquidity and maturity mismatches. - Regulators should develop or maintain currency-specific liquidity risk frameworks, including stress tests, emergency funding strategies, and resolution planning; coordination and sharing of information among regulators are crucial to reduce unintended cross-border spillovers from jurisdiction-specific liquidity requirements. - Central bank swap lines should be retained to provide foreign exchange liquidity in periods of systemic stress to help prevent foreign currency funding difficulties from spilling over to other parts of the financial system. - Completing the postcrisis reform agenda (including Basel III implementation) remains vital; ensuring the independence of supervision and addressing new challenges posed by technology are important elements of this effort. *Source: IMF Global Financial Stability Report: A BuMPY ROAd AhEAd (April 2018), Chapter 1 — "US Subsidiaries of Non-US Banks: Intragroup Borrowing and Lending"* ### 1. United States ### 1. United States ### Term premiums and drivers (Box 1.2) - Estimated term premiums on 10-year German bunds are close to historical lows; the latest fitted value is about −15 basis points, which is less than the observed estimate of about 15 basis points. - Estimated term premiums across Canada, France, Japan, and the United Kingdom are similarly close to their fitted values. - Key statistical drivers of low fitted term premiums: - Low survey-based uncertainty about near-term GDP growth and inflation. - Subdued volatility of US Treasury returns. - A persistently lower correlation between Treasury and risky asset returns. - Model limitations and risks: - The models do not forecast future directions of underlying factors; increases in uncertainty about inflation, growth, or the path for monetary policy would imply significant increases in term premiums. - Other important influences are hard to capture statistically, including regulatory restrictions affecting investor demand for government paper and debt-management considerations. - Policy-relevant implication: - Statistical results are consistent with the view that the overall level of longer-dated yields is appropriate given the stance of monetary policy, which “should remain largely accommodative to support growth and to bring inflation closer to central banks’ targets.” ### US leveraged loan market and changing investor base (Box 1.3) - Structural shift in buyer base: - Since 2014, CLOs have purchased more than half of total issuance of leveraged loans. - US CLOs accounted for 57 percent of leveraged loans outstanding in 2017, with $495 billion in assets under management. - CLO issuance reached $118 billion in 2017, above precrisis levels. - Loan mutual funds (including ETFs) grew from roughly $20 billion in 2007 to $170 billion in assets in 2017, and now account for more than 20 percent of the institutional loan market. - Market-concentration and investor composition: - Asset managers, insurance companies, and pension funds now account for 45 percent of AAA CLO market share. - Risks from the shift to institutional investors: - Migration of loan assets to open-end loan mutual funds offering daily liquidity may exacerbate price moves during large investor redemptions under distress. - In the precrisis period, AAA CLO tranches were routinely funded in the repo market using financial leverage; unwinding such leverage amplified loan price moves when investor confidence fell. - Current mitigants: - The use of financial leverage to fund CLO positions appears to be limited at this point. - Investors do not seem to be widely using total return swaps to gain leveraged exposure to the loan market (common in 2006–07). ### Central Bank Digital Currencies (CBDC) (Box 1.4) - Definition and variants: - A CBDC could be defined as a digital form of central bank money that can be exchanged, peer to peer, in a decentralized manner; a token representation of, or an addition to, cash and/or electronic deposits. - It could be issued directly to commercial banks and other payment services providers or to individuals, and exchanged at par with the central bank’s other monetary liabilities. - Potential benefits and use cases: - Faster clearing and settlement without an intermediary; anonymity for transactions; geographic flexibility; flexible denomination structures—attributes making cryptocoins attractive for cross-border payments and micropayments. - CBDCs could counter monopoly power of private payment networks and address stability/safety issues of private cryptocoins. - Retail benefits include potential savings on maintaining physical notes and coins, reduced transaction costs for individuals and small enterprises, and facilitation of financial inclusion. - Central banks could tailor anonymity (cash-like anonymity for small-value payments; more tailored compliance for larger-value payments). - From a monetary policy perspective, CBDCs could help maintain demand for central bank money; central bank seigniorage would continue. - CBDCs, along with abolition of cash, might allow central banks to overcome the zero lower bound, facilitating truly negative interest rates when necessary. - Risks and trade-offs: - Competition between CBDCs and commercial bank deposits could lead to volatility in fund flows and potential bank runs toward CBDCs, potentially hampering financial stability. - Recommended approach: - A gradual and cautious approach that builds on experience and takes into account evolving financial technologies seems warranted. - Financial-stability risks could be reduced if CBDC design respects the current two-tier banking system (separation of commercial banking from central banking) and merely creates a digital form of cash. ### Regulatory reform and outstanding agenda (Box 1.5) - Progress: - Postcrisis regulatory reforms, primarily via Basel III, have enhanced major banks’ resilience. - To address gaming of internal models, the Basel Committee proposed enhancements in 2014 and agreement was reached in December 2017. - The measures limit risk-weighted assets, based on the internal-ratings-based approach, to a minimum of 72.5 percent of the amount calculated using the simpler standardized approach. - The standardized approach to credit risk was revised to be more risk sensitive (for example, varying risk weights for real estate exposures using loan-to-value ratios). - Implementation timing and concessions: - Implementation of the Fundamental Review of the Trading Book was postponed to 2022 in response to practical challenges. - The implementation timeline for these reforms was extended to 2022–27. - The final agreement included: a less conservative RWA floor than the initially proposed 80 percent; an annual cap on any increase in risk-weighted assets resulting from the measures; and lowering some minimum risk weights in the standardized approach. - Remaining challenges and priorities: - Full, timely, and consistent implementation is essential; implementation has been delayed and is lagging in areas such as cross-border resolution frameworks for banks. - A major challenge is shortcomings in the operational independence of supervisors from political and market influence; IMF Financial Sector Assessment Programs found that only a handful of the nearly 40 countries assessed since the global financial crisis are in full compliance with the Basel Core Principles on independence and accountability. - Policymakers must ensure supervisors have resources and power to take timely, preemptive, and corrective actions. - Outstanding items on the agenda include: translating Financial Stability Board recommendations to transform shadow banking into resilient market-based finance into operational guidance; resolution efforts for nonbanks, including central counterparties; insurer reform; reforming governance, addressing misconduct, reinforcing individual accountability, and creating supportive institutional culture to tackle incentives for excessive risk taking. - The decision on better incorporating sovereign risks into the regulatory framework has been shelved for the time being. - Policy implication: - Given calls for rolling back reforms, completing and implementing the postcrisis agenda is vital to allow supervisors to focus on emerging challenges, including rapid developments in financial technology and cyberattack threats. *International Monetary Fund | April 2018* ### CHAPTER 1 A BuMPY ROAd AhEAd ### CHAPTER 1 A BuMPY ROAd AhEAd ### Summary and purpose - Examines the evolution of the riskiness of corporate credit allocation (the extent to which riskier firms receive credit relative to less risky ones) across advanced and emerging market economies since 1991. - Focuses on allocation of credit across firms rather than aggregate volume of credit or credit growth. - Constructs several cross-country measures of the riskiness of credit allocation using firm-level financial statement data for 55 economies since 1991. - Emphasizes monitoring these measures as part of macro-financial surveillance; measures are described as simple to compute and readily replicable. ### Key empirical findings - The riskiness of credit allocation rises during periods of fast credit expansion, especially when loose lending standards or easy financial conditions occur concurrently. - Globally, the riskiness of credit allocation: - Increased in the years preceding the global financial crisis and peaked shortly before its onset. - Declined sharply after the crisis and rebounded to its historical average in 2016 (latest comparable year). - Might have risen further as financial conditions loosened in 2017. - An increase in the riskiness of credit allocation signals: - Heightened downside risks to GDP growth. - A higher probability of banking crises and banking sector stress, over and above signals provided by credit growth. - The riskiness of credit allocation is an independent source of financial vulnerability. - At the country level, the riskiness of credit allocation is more strongly associated with credit growth when: - Lending standards are easier. - Domestic financial conditions are looser. - Credit spreads are lower. - Global risk appetite is higher. - The riskiness of credit allocation at the global level followed a cyclical pattern over the past 25 years and was slightly below its historical average at end-2016. ### Predictive power and horizons - Taking the riskiness of credit allocation into account helps better predict: - Full-blown banking crises. - Financial sector stress. - Downside risks to GDP growth. - Predictive horizons extend up to three years. - A period of high credit growth is more likely to be followed by a severe medium-term downturn if accompanied by an increase in the riskiness of credit allocation. - When credit is stagnant or falling, the riskiness of credit allocation has a negligible effect on downside risks to GDP growth. ### Conceptual framework (mechanisms) - Theoretical channels link the riskiness of credit allocation to financial conditions: - In canonical business-cycle models with financial frictions, access to credit for riskier firms is procyclical. - Positive macroeconomic shocks or lower interest rates raise firms’ short-term prospects and net worth, relax information asymmetries, and increase access to credit for higher-leverage firms. - Negative shocks or higher interest rates tighten access for firms with weak balance sheets. - References to related empirical findings: - U.S. evidence since mid-1990s that allocation risk increases during expansions and declines during recessions. - Studies (Greenwood and Hanson 2013; Jiménez and others 2014; Dell’Ariccia, Laeven, and Suarez 2017) linking riskier credit allocation to stronger credit growth, lower short-term Treasury yields, lower term spreads, and higher risk appetite. ### Policy-relevant institutional and policy correlates - Policy and institutional settings associated with a smaller increase in the riskiness of credit allocation during credit expansions: - Tightening of the macroprudential policy stance. - Greater independence of the supervisory authority from banks. - A smaller government footprint in the corporate sector. - Greater minority shareholder protection. - Implication: Macroprudential, supervisory, and legal-institutional frameworks can mitigate procyclicality of credit allocation risk. ### Data and measurement notes - No existing cross-country measures capture riskiness of total credit flows across firms; the chapter constructs several measures mapping credit flows across firms to distributions of firm-level vulnerability indicators. - Four alternative measures of the riskiness of credit allocation are discussed to accommodate different market and data environments. - The sample and figures referenced: - Measures constructed for 55 economies since 1991. - Figure 2.1 notes a sample comprising 41 advanced and emerging market economies and shows various percentiles of a Financial Conditions Index (FCI) over 1991–2017. - Figure 2.2 presents the share of low-rated (high-yield and BBB-rated) nonfinancial corporate bond issuance in selected advanced economies over 1997–2017 (three-year moving average). ### Chapter organization (topics covered) - Stylized conceptual framework for macro-financial shocks and credit-allocation risk. - Construction and cross-country evolution of new measures. - Cyclical properties and relationship to indicators of financial conditions. - Empirical analysis linking indicators to future financial instability and downside GDP risks. - Determinants of credit-allocation risk cyclicality, including macroprudential policies and supervisory, legal, and institutional frameworks. - Conclusions and policy implications. *Prepared by a staff team led by Jérôme Vandenbussche; International Monetary Fund | April 2018.* ### CHAPTER 2 ThE RISkINESS OF CREdIT ALLOCATION: A SOuRCE OF FINANCIAL VuLNERABILITY? ### CHAPTER 2 ThE RISkINESS OF CREdIT ALLOCATION: A SOuRCE OF FINANCIAL VuLNERABILITY? ### Mechanisms driving procyclicality and riskiness of credit allocation - Fluctuations in credit quantity and the riskiness of credit allocation can be driven by variations over time in investor beliefs, risk appetite, or perceptions of economic uncertainty, which directly affect credit spreads and expected volatility. - In booms: - Optimistic investors borrow extensively to acquire assets, pushing up asset prices; bad news raises uncertainty and volatility, leading lenders to require higher margins, triggering deleveraging and fire sales. - Optimism correlated with risk can generate procyclical increases in the riskiness of credit allocation. - Investors may form unduly optimistic beliefs and extend credit to more vulnerable firms, allowing excessive borrower leverage (Minsky 1977; Kindleberger 1978; Bordalo, Gennaioli, and Shleifer 2018). - Intermediaries with long-term liabilities and short-term assets search for yield when monetary conditions are loose, easing access for riskier firms (Rajan 2006). - Bank screening and incentives: - Screening capacity deteriorates during prolonged credit expansions due to loss of institutional memory and the need to intermediate larger volumes of credit, reducing profitability of careful screening (Berger and Udell 2004; Dell’Ariccia and Marquez 2006). - Role of bank capital: - Banks’ role in gathering borrower information depends on their capital levels; an increase in bank capital may lead to expansion of credit to firms with poorer fundamentals (Holmstrom and Tirole 1997). - Theoretical relationship among short-term interest rates, bank leverage, and bank risk taking is ambiguous because multiple opposing effects operate (Dell’Ariccia, Laeven, and Marquez 2014; Dell’Ariccia, Laeven, and Suarez 2017). ### Measuring the riskiness of credit allocation: data and indicators - Sample and period: - Measures constructed for a set of 55 economies (26 advanced economies and 29 emerging market economies) over the 1991–2016 period using data for listed firms. - Firm-level vulnerability indicators used (four): - Leverage ratio. - Interest coverage ratio (ICR). - Debt-to-profit ratio (debt overhang). - Expected default frequency (EDF) — a market-based indicator of credit risk. - Construction of the riskiness measure: - Raw measure = average of the vulnerability indicator among firms whose debt (loans and bonds) increases the most minus average among firms whose debt increases the least or declines the most. - Final measure = raw measure minus country-specific mean to remove influence of sectoral composition and ensure cross-country and cross-measure comparability. - Interpretation: an increase signals that vulnerability of firms getting relatively more credit has risen relative to firms getting relatively less credit; positive (negative) values indicate above (below) country sample average riskiness. - Robustness check: - Similar measures constructed using data covering both listed and unlisted firms (Orbis database) for a smaller set of countries and from 2000 onward (50 economies), confirming main patterns. ### Stylized facts and cross-country evolution - Global pattern (all four indicators broadly similar): - Elevated levels in the late 1990s. - Fell in 2000–04, reached historical low in 2004. - Rose steeply during 2004–08, peaked at onset of the global financial crisis. - Declined sharply over the next two years and was slightly below precrisis level at end of 2016. - Country-level nuances (1995–2016 shown for selected economies): - United States and Japan: very similar cyclicality and magnitude through much of the sample; divergence in 2014–16 with U.S. riskiness decreasing to a relatively low level and Japan remaining relatively high. - Spain: experienced a credit boom from late 1990s to mid-2000s with very high riskiness until 2008 crisis, followed by a large fall in the indicator. - Germany: no credit boom over the 20-year period; measures remained within a narrower range but moved into positive territory in recent years. - India: broadly followed global patterns; measure at a relatively low level in 2016. - China: weaker synchronization with global developments; peaks and troughs often lag global by two to three years, with a peak in 2009–10 consistent with large stimulus beginning end-2008 and misallocation of credit. - Korea and United Kingdom: instances where accounting-based and EDF-based measures diverged, indicating complementarity and occasional contrasts between market-based and accounting-based signals. ### Empirical findings on cyclicality and amplifying factors - Procyclicality: - Regression analysis indicates the riskiness of credit allocation is procyclical: it increases when GDP growth or changes in the domestic credit-to-GDP ratio are stronger. - The association of credit expansion with greater riskiness is statistically significant for all four measures. - Magnitudes: - A one standard deviation increase in the change of the credit-to-GDP ratio (equivalent to an increase of 5.5 percentage points) is associated with an increase in the riskiness of credit allocation of 0.12–0.25 standard deviation, depending on the exact measure. - Financial conditions amplify the effect: - Credit expansions accompanied by loose financial conditions or loose lending standards are more likely driven by shifts in credit supply and higher risk appetite, producing riskier allocations. - Low corporate credit spreads (or high global risk appetite proxied by the Chicago Board Options Exchange Volatility Index [VIX]) during credit expansions result in riskier allocations than expansions accompanied by high credit spreads (or low global risk appetite). - A higher stock market price-to-book ratio is associated with a higher level of the riskiness of credit allocation. - Panel vector autoregression analysis confirms a significant effect of financial conditions on the riskiness of credit allocation. - Heterogeneity: - Results are similar for advanced and emerging market economies, though dispersion of estimated relationship is larger in emerging markets (likely due to smaller sample size). - Robustness: - Similar patterns and properties are confirmed using measures constructed from a broader firm sample (listed and unlisted firms, Orbis database covering 50 economies from 2000). ### Key statistics and indices preserved from the analysis - Sample: 55 economies (26 advanced economies and 29 emerging market economies), period 1991–2016 (listed firms dataset). - Robustness sample: 50 economies from 2000 (Orbis database, listed and unlisted firms). - Credit-to-GDP change: one standard deviation = 5.5 percentage points. - Associated increase in riskiness: 0.12–0.25 standard deviation (range depends on measure). - Financial condition indicators explicitly referenced: corporate credit spreads; Chicago Board Options Exchange Volatility Index (VIX); stock market price-to-book ratio. - Reference years and episodes noted: late 1990s elevated riskiness; 2000–04 decline; historical low in 2004; steep rise 2004–08; peak at onset of global financial crisis; decline over next two years; slightly below precrisis level at end of 2016. *Source: IMF staff estimates based on Worldscope data and related analyses in Chapter 2, GLOBAL FINANCIAL STABILITY REPORT: A BuMPY ROAd AhEAd (April 2018).* ### CHAPTER 2 ThE RISkINESS OF CREdIT ALLOCATION: A SOuRCE OF FINANCIAL VuLNERABILITY? ### CHAPTER 2 ThE RISkINESS OF CREdIT ALLOCATION: A SOuRCE OF FINANCIAL VuLNERABILITY? ### Overview and main empirical approach - Purpose: assess whether the riskiness of credit allocation helps predict episodes of financial instability and downside risks to growth. - Method: cross-country regressions augmenting standard specifications with a riskiness-of-credit-allocation measure; econometric framework in Annex 2.3. - Firm-level measures used to construct the riskiness metric include: leverage, interest coverage ratio (ICR), expected default frequency (EDF), and debt overhang (EBITDA-to-debt, inverted as negative of median). ### Dynamics around systemic banking crises - The riskiness of credit allocation follows a clear inverted-U pattern around systemic financial crises: - It rises gradually during the five years preceding the crisis, - Reaches a relatively high level at crisis onset (time 0), - Falls following the onset of the crisis. - This dynamic is consistent across all four firm-level indicators (Figure 2.8). - Traditional corporate vulnerability indicators tend to pick up only after the crisis has already struck. ### Predictive power for systemic banking crises and banking-sector stress - Regression evidence: - A one standard deviation increase in the riskiness measure increases the odds of a systemic banking crisis by a factor of about four. - Contextual calibration: in the sample used, the probability of observing a crisis is about 5 percent; the probability of not observing a crisis is about 95 percent; the odds of a crisis are 5.3 percent (100*5/95). A fourfold increase would raise the odds to 21 percent. - Adding the riskiness variable increases explanatory power by between 11 and 25 percentage points. - Banking-sector equity stress model (objective timing, more events): - The riskiness measure adds predictive power for horizons from t to t + 3 years. - A one standard deviation increase in riskiness raises the odds of banking-sector stress by a factor of 1.3 to 2, depending on the measure and horizon. - Banking sector equity stress defined as annual excess return of the banking sector lower than the country-specific mean by at least one standard deviation. ### Signals for GDP growth downside risks (growth-at-risk) - The riskiness of credit allocation predicts downside risks to cumulative real GDP growth at horizons one to three years ahead. - Analysis examines the 20th and 50th percentiles of cumulative real GDP growth from year t to t + h, with h = 1, 2, 3. - Findings: - The riskiness measure is strongly related to both the median (50th percentile) and left tail (20th percentile) of the growth distribution over all horizons (Figure 2.11). - Effects are in addition to those of changes in the credit-to-GDP ratio and financial conditions. - The effect on downside risks to growth is significant when riskiness measures are constructed using samples that include unlisted as well as listed firms. - Interaction with credit expansions: - Credit booms accompanied by rising riskiness of credit allocation signal elevated downside risks to growth two and three years ahead. - Conversely, during credit contractions or relatively soft credit expansions, a higher riskiness of credit allocation does not increase downside risks to future GDP growth; at a three-year horizon and low credit-to-GDP growth, an increase in risk taking has no significant impact on downside risks to growth (Figure 2.12). ### Role of policy and structural determinants - Categories analyzed: banking-sector soundness, macroprudential policies, supervisory/legal/institutional frameworks. - Robustness: determinants discussed are those whose effects are robust across the four riskiness measures (leverage-, ICR-, debt overhang–, and EDF-based). See Annex 2.1 and 2.2 for variable definitions and methodology. Key quantitative and qualitative findings on determinants: - Bank capital: - Bank capital appears to have little significant effect on the cyclicality of the riskiness of credit allocation. - Evidence in the literature is contrasting; effects likely depend on country circumstances. - Greater buffers are generally associated with greater cyclicality, but this result is not robust. - Macroprudential policies: - Tightening macroprudential policy reduces the cyclicality of the riskiness measure. - Specific instruments with observed effects: - Changes in minimum leverage ratio and changes in ceilings and penalties related to credit growth curb the increase in riskiness associated with faster credit growth. - Increases in capital conservation buffers reduce the level of the riskiness of credit allocation. - Interpretation: postcrisis regulatory tightening has helped limit rebounds in the riskiness measure. - Note: tightening of minimum capital requirements shows a nonrobust association with increased riskiness (suggesting possible reverse causality). Loan provisioning requirements show no significant effects. - Supervisory independence and institutional settings: - Greater supervisory independence is associated with reduced cyclicality of the riskiness of credit allocation: when supervisory authority enjoys greater legal protection from the banking industry, credit allocation quality is less sensitive to domestic credit growth. - Lower government footprint in the nonfinancial corporate sector reduces cyclicality. - Greater protection of minority shareholders reduces cyclicality, underscoring the role of corporate governance for financial stability. - Figure 2.13 summarizes the quantitative impact (in standard deviations of the riskiness measure) of a contemporaneous one standard deviation increase in the change in the credit-to-GDP ratio under “lower” and “higher” settings of policy and institutional variables. ### Robustness and caveats - Results are robust to inclusion of standard corporate vulnerability indicators (median firm leverage, high-yield share of bond issuance), though weaker if the post-2008 period is excluded. - Predictive performance analysis is in-sample (all available observations used to estimate the models). - Some policy variable effects are nonrobust or may reflect reverse causality; interpretations are conditional on the empirical specifications described in Annex 2.3. *International Monetary Fund | April 2018* ### Chapter 3 of the October 2016 GFSR. ### Chapter 3 of the October 2016 GFSR ### Conclusions and Policy Implications - A riskier credit allocation is a source of vulnerability that may threaten financial stability; both the volume and allocation of credit matter for financial stability. - Periods of high credit growth are more likely to be followed by a severe downturn or financial sector stress over the medium term if accompanied by an increase in the riskiness of credit allocation. - Supervisory and policy priorities: - Pay close attention to the evolution of credit allocation riskiness, especially when rapid credit expansion and rising allocation riskiness coincide. - Monitor credit origination standards and the riskiness of credit allocation on a continuous basis. - Intensify supervisory scrutiny during episodes of large credit expansion and loose financial conditions, and require corrective action if needed.29 - Data and surveillance: - Firm-level financial statement data can measure the riskiness of credit allocation and are available in many countries; the usefulness for surveillance depends on the speed with which these data become available.30 - Policymakers should engage in efforts to collect these granular data as swiftly as possible. - Institutional and policy settings to mitigate rising allocation riskiness: - A more independent banking supervisor can better exert control over lending and origination standards during good times. - Promote sounder corporate governance to reduce managers’ ability to “gamble for resurrection.” - Macroprudential measures (for example, tightening some regulatory capital requirements) may reduce banks’ ability or willingness to lend to vulnerable firms.31 - Consider increased provisioning requirements and thicker countercyclical capital buffers when allocation riskiness rises; the calibration of capital buffers should consider the riskiness of credit allocation.32 - Discourage policies that direct credit to certain firms or sectors without due consideration of underlying credit risk during periods of strong credit growth. ### Measurement and Methodology - The chapter uses the Greenwood and Hanson (2013) approach to construct measures of the riskiness of credit allocation for four firm-level vulnerability indicators: - Leverage (total debt to total assets) - Debt overhang (total debt to EBITDA) - Interest coverage ratio (ICR; EBITDA to interest expenses) - Expected default frequency - Construction steps (exact procedure preserved): - For every year, each firm is assigned the value (from 1 to 10) of its decile in the distribution of the indicator in the country where it is located; a higher decile represents a larger value of the underlying vulnerability. - Firms are sorted by the changes in net debt to lagged total assets into five equal-size bins. Firms in the bin with the largest increases in debt are “top issuers,” and firms in the bin with the largest decreases in debt are “bottom issuers.” - The measure is computed as the difference between the average vulnerability decile for the top issuers and the corresponding average for the bottom issuers. - Interpretation and properties: - The measure focuses on relative rankings; it abstracts from changes in the mean and shape of the vulnerability indicator distribution. - Using deciles minimizes outlier influence, avoids picking up secular trends, normalizes across countries, and facilitates comparison across indicators; a downside is the loss of information about changes in cross-sectional dispersion. - For debt overhang, deciles of EBITDA to debt (instead of debt to EBITDA) are used to avoid classifying firms with negative earnings as low-vulnerability firms. - Data coverage and sample: - The histograms and analysis cover 55 economies for the period 1991–2016; data are demeaned at the country level. - The distributions have the shape of a bell curve and have a standard deviation of about one. - The measure reflects broad debt (including loan and bond financing) and continuous firm vulnerability/default risk. ### Empirical Findings — Global Patterns and Dynamics - Global-level observation: - The riskiness of credit allocation rebounded since its post-global-financial-crisis trough and was back to its historical average at the end of 2016. - The relatively mild credit expansion in recent years, combined with postcrisis regulatory tightening, contributed to a softer rebound in allocation riskiness than might be expected given very loose financial conditions. - Country-level heterogeneity is important; some countries experienced a more pronounced rise in allocation riskiness. - As financial conditions loosened further in 2017, the riskiness of credit allocation might have continued to rise, warranting close monitoring and heightened vigilance. - Sectoral and household considerations: - Relatively low credit allocation riskiness can coexist with a large increase in conventional corporate vulnerability indicators (for example, average leverage) in some major economies. - The chapter focuses on the corporate sector; riskiness of credit allocation to households may also be relevant and need not follow the same patterns. Monitoring household allocation riskiness is difficult across many countries, but selected household surveys reported in the October 2017 GFSR suggest indebtedness of lower-income, more vulnerable households has increased in recent years in various countries. ### China: Profitability and Allocation of Credit (Box 2.2) - Motivation: - Nonfinancial corporate debt in China has continued to expand at a brisk pace; understanding credit allocation helps assess whether vulnerabilities are building.1 - Profitability of credit allocation measure: - Constructed analogously to the allocation riskiness measure but comparing profitability of firms whose credit is growing fastest to profitability of firms whose credit is growing slowest. - Key findings for China: - The riskiness of credit allocation declined markedly in China since 2012, but the profitability of credit allocation experienced only a mild recovery and remained relatively low at the end of 2016. - The profitability of credit allocation rose in the early 2000s following SOE reforms but began declining just before the global financial crisis as the credit-to-GDP ratio accelerated; it continued declining during and after the crisis amid the 2009–10 stimulus. - The decline in profitability of credit allocation over the past decade has been stronger among state-owned enterprises (SOEs) and firms in traditional sectors. - From 2007 to 2011, the decline occurred within both SOEs and private firms; since then, the decline has continued within SOEs while profitability of credit allocation has improved among private firms. - Within sectors identified as new engines of Chinese growth (information and communication technology; technology hardware and equipment; health care equipment and services; pharmaceuticals, biotechnology, and life sciences), profitability of credit allocation has stabilized or improved over the past 10 years; by contrast, traditional drivers (automobiles and components; energy; materials) show sharper falls and severe overcapacity issues. - Numerical context: - The outstanding stock of corporate debt in China reached about 163 percent of GDP at the end of 2017.2 ### Dynamics: Financial Conditions, Credit Growth, and Allocation Riskiness (Box 2.3) - Method: - A panel vector autoregression (VAR) using annual data for 41 countries from 1991 to 2016 examines joint dynamics of the leverage-based allocation-risk measure, a financial conditions index (FCI), credit growth, and GDP growth. The VAR includes country fixed effects and one lag. - Main empirical result: - Loosening financial conditions lead to: - Riskier credit allocation over a two- to three-year horizon, - Credit expansion, and - Higher GDP growth (Figure 2.3.1 summarizes impulse responses). - Measurement notes: - The VAR responses: the FCI and allocation-risk responses are in standard deviations; credit growth responses are in percent of GDP; a rise in the FCI indicates a loosening of financial conditions. ### Policy Recommendations (summarized) - Surveillance and data: - Build and use firm-level vulnerability measures (leverage, debt overhang, ICR, expected default frequency) based on decile rankings and issuer quintiles to track allocation riskiness. - Collect granular firm-level data with minimal lag to improve macro-financial surveillance. - Prudential and macroprudential action: - Intensify supervisory scrutiny and require corrective action when allocation riskiness rises amid credit expansions and loose conditions. - Strengthen supervisory independence and corporate governance. - Consider tightening regulatory capital requirements, increasing provisioning, and deploying thicker countercyclical capital buffers calibrated to allocation riskiness. - Caution against directed credit policies that ignore underlying credit risk during strong credit growth. *Chapter 3 of the October 2016 GFSR.* ### 2. Response of Credit Growth to Riskiness Shock ### 2. Response of Credit Growth to Riskiness Shock ### Joint dynamics: riskiness of credit allocation, credit growth, GDP, and FCI - Panel VAR setup: - Sample average: 19 years per country. - Responses measured using a simple Cholesky decomposition. - Variable ordering: riskiness of credit allocation first, followed by credit growth and GDP growth, and the FCI last. - Assumptions: the FCI responds contemporaneously to all other variables; the riskiness of credit allocation does not respond contemporaneously to credit growth, GDP growth, or financial conditions (it responds only with a lag). - Changing the ordering of the other variables or including more lags does not materially affect results. - Main empirical findings: - An increase in the riskiness of credit allocation is followed by a tightening (decline) in financial conditions. - Credit growth increases significantly after an increase in the riskiness of credit allocation. - The likely mechanism: an unobserved loosening of credit standards that leads to a more immediate deterioration in credit quality and higher credit growth. - When lending standards are included (smaller sample), results seem to support the loosening-standards hypothesis. - The augmented panel VAR with lending standards shows GDP first rises, but then declines after an increase in the riskiness of credit allocation, consistent with a trade-off between current economic conditions and future financial vulnerabilities. - Higher-frequency data are probably needed to fully identify timing and causal channels. ### High-Yield (HY) share during credit booms and output growth (Box 2.4) - HY share measure and sample: - HY share constructed for issuance by nonfinancial corporations and governments. - HY share can be constructed for a sample of 38 countries, with coverage for some starting in 1980. - For the credit-boom analysis: sample comprises 25 advanced economies. - Properties and interpretation: - The HY share is procyclical: it rises when recent economic performance has been good and falls when recent economic performance has been bad. - The HY share moves in line with survey measures of bank lending standards and provides a complementary, bond-market–based measure of issuance quality. - Empirical specification: - Credit booms defined as episodes in which the change in the credit-to-GDP ratio over the previous five years is high relative to recent international experience. - Local projection specifications interact dummies for credit booms with the average change in the HY share over the course of the boom. - Key result: - A one standard deviation increase in the HY share during a credit boom lowers cumulative GDP growth over the next three years by 2 percentage points. - The HY share helps separate “good” from “bad” credit booms: the probability of low growth following a credit boom is very low given a “good” HY indicator and substantially higher given a “bad” HY indicator. - Policy implication: - Issuance quality (proxied by the HY share) during a credit boom contains information about growth out to three or four years and warrants special attention from policymakers. ### Data, variable definitions, and samples - Worldscope-based sample and cleaning: - Financial-sector firms dropped (except real estate). - Observations dropped if market capitalization, total assets, total debt, total liability, or interest expenses are strictly negative or if operating profit margin or ratio of short-term debt to total debt exceeds 100 percent. - Observations kept only if full information on net debt issuance; leverage; EBITDA; and market capitalization is available. - Economy-year pairs retained only if no fewer than 40 firms and available information on aggregate credit to the private sector. - Final Worldscope sample: about 500,000 nonfinancial firm-year observations from 55 economies during 1991 to 2016. - Orbis robustness sample: - Covers listed and unlisted firms; cleaned following Kalemli-Özcan and others (2015). - Observations require full information on net debt issuance, leverage, EBIT, loans, and long-term debt. - Economy-year pairs retained only if at least 50 nonfinancial (including real estate) firms. - Final Orbis sample covers 50 economies; start date chosen as 2000 for Orbis-based analysis. - Data limitations: EBIT used instead of EBITDA for debt overhang when EBITDA unavailable; ICR not used in Orbis robustness due to poor availability of interest expenses; EDF cannot be computed for unlisted firms. - WIND database (Box 2.2 analysis): - Covers listed Chinese firms with annual ownership information. - Observations dropped if key financial variables missing or invalid; only one observation/year for firms listed on several markets; years with at least 50 nonfinancial (including real estate) firms retained. - Final WIND sample: about 37,000 firm-year pairs from 1995 to 2016. Ownership information available for most firms only from 2004. - Variable definitions (selected): - Leverage ratio = total debt / total assets. - Interest coverage ratio (ICR) = interest expenses / EBITDA. - Debt overhang = total debt / EBITDA. - Expected default frequency (EDF) computed using the Black-Scholes-Merton model as in Vassalou and Xing (2004). - Return on assets (used in Box 2.2) = EBITDA / total assets. - HY share construction: - Based on bond-market issuance by nonfinancial corporations and governments. ### Empirical approaches and determinants of riskiness of credit allocation (Annex 2.2) - Cyclicality specification: - Riskiness_{i,t}^X = α_i^X + γ_t^X + β_1^X ∆Credit_{i,t} + β_2^X ∆GDP_{i,t} + β_3^X Appreciation_{i,t} + ε_{i,t}^X - X ∈ {leverage, interest coverage ratio, debt overhang, expected default frequency} - ∆Credit = change in the ratio of bank credit to the nonfinancial private sector to nominal GDP. - ∆GDP = real GDP growth. - Domestic currency appreciation against the US dollar included to control valuation effects from foreign-currency debt. - Country (α_i^X) and year (γ_t^X) fixed effects included; standard errors clustered at the country level. - Results robust to controlling for general government structural balance and to instrumenting GDP growth and ∆Credit by lagged values. - Financial conditions, lending standards, and interaction specification: - Riskiness_{i,t}^X = α_i^X + γ_t^X + β^X Controls_{i,t} + δ^X FC_{i,t} + θ^X × FC_{i,t} × ∆Credit_{i,t} + ε_{i,t}^X - Controls_{i,t} include change in the credit-to-GDP ratio, real GDP growth, and domestic currency appreciation. - FC_{i,t} represents the financial conditions index (FCI), financial variables representing components of the FCI, or a measure of lending standards. - ∆Credit and FC demeaned at the country level. - δ̂^X measures the level effect of FC on riskiness when demeaned ∆Credit is 0. - θ̂^X captures the marginal effect on the credit cyclicality of riskiness caused by a change in the FCI, financial variables, or lending standards. - Robustness and additional variables: - Results are robust to alternative credit data (e.g., BIS total credit to the nonfinancial private sector; credit to nonfinancial corporate sector), different cycle measures (output gap; real credit growth), and two-way clustered standard errors by country and year. - Investigated additional financial variables (stock market volatility, credit boom dummy, length and phases of credit booms, cross-border bank flows-to-GDP, housing price inflation) but none had a robust significant impact on the riskiness of credit allocation. ### Sample lists and data notes (selected) - HY share construction coverage: 38 countries (coverage for some starting in 1980). - HY credit-boom analysis sample: 25 advanced economies. - Worldscope end year: 2016; Orbis end year: 2015. - Minimum observation requirements for indicators: - Minimum 40 observations for interest expenses to construct ICR-based indicator (exception: Ireland with 38 or 39 in some years). - Minimum 40 observations for non-zero debt for debt overhang-based indicator. - Minimum 40 observations for EDF-based indicator. *International Monetary Fund | April 2018* ### Annex Table 2.2.1. Cyclicality of the Riskiness of Credit Allocation ### Annex Table 2.2.1. Cyclicality of the Riskiness of Credit Allocation ### Cyclicality regressions (leverage-based riskiness) - Dependent variable: Riskiness of Credit Allocation Based on Leverage (columns (1)–(3)) - Change in Credit-to-GDP Ratio: - Column (1): 0.05*** (0.01) - Column (2): 0.04*** (0.01) - Column (3): 0.06*** (0.01) - Real GDP Growth: - Column (1): 0.08*** (0.02) - Column (2): 0.12*** (0.02) - Column (3): 0.05* (0.03) - Appreciation against the US Dollar: - Column (1): –0.04*** (0.01) - Column (2): –0.02* (0.01) - Column (3): –0.05*** (0.01) - Country Group: All / AE / EM (columns correspond) - Country Cluster: Yes (all) - Country Fixed Effect: Yes (all) - Year Fixed Effect: Yes (all) - Observations: 986 / 563 / 423 (columns (1)/(2)/(3)) - Number of Countries: 55 / 26 / 29 - R 2 : 0.31 / 0.34 / 0.37 - Robustness reporting (columns (4) and (5)): number of alternative risk measures with same sign and significant at 10 percent reported as 4 and 4 (table header indicates robustness but specific entries shown as 44) ### Note on inference and robustness - Standard errors clustered at the country level; standard errors reported in parentheses. - Significance codes: ***p < 0.01; **p < 0.05; *p < 0.1. - Robustness exercise investigates cyclicality using three other risk measures: interest coverage ratio, debt overhang, expected default frequency; columns (4)–(5) report counts of measures with same sign and significance. --- ### Annex Table 2.2.2. Impact of Financial Conditions and Lending Standards on the Riskiness of Credit Allocation ### Baseline and interaction findings (leverage-based riskiness) - Change in Credit-to-GDP Ratio (columns (1)–(5)): - Column (1): 0.05*** (0.02) - Column (2): 0.05*** (0.01) - Column (3): 0.05*** (0.01) - Column (4): 0.05*** (0.01) - Column (5): 0.05*** (0.01) - Bank Lending Standards: - Coefficient: –0.10 (0.07) (increase means stricter bank lending standards) - Change in Credit-to-GDP Ratio × Bank Lending Standards: - Coefficient: –0.03* (0.02) - Financial Conditions Index (FCI): - Coefficient: –0.05 (0.07) (increase means tighter financial conditions) - Change in Credit-to-GDP Ratio × FCI: - Coefficient: –0.01** (0.00) - Corporate Credit Spreads: - Coefficient: –0.07 (0.06) - Change in Credit-to-GDP Ratio × Corporate Credit Spreads: - Coefficient: –0.02** (0.01) - Stock Price-to-Book Ratio: - Coefficient: 0.20*** (0.06) - Change in Credit-to-GDP Ratio × Stock Price-to-Book Ratio: - Coefficient: 0.01 (0.01) - Change in Credit-to-GDP Ratio × Log (VIX): - Coefficient: –0.04** (0.02) - Controls: Yes (real GDP growth and domestic currency appreciation against the US dollar controlled in all regressions) - Country Cluster: Yes; Country Fixed Effect: Yes; Year Fixed Effect: Yes - Observations by column: 2668 / 2464 / 396 / 394 / 9986 (table shows multiple numbers across columns; header lists Observations 2668 2464 396 394 9986) - Number of Countries: 21 / 41 / 37 / 51 / 55 (columns vary) - R 2 : 0.39 / 0.34 / 0.33 / 0.33 / 0.31 - Robustness columns (6)–(7) report counts of alternative measures with same sign and significance (entries displayed as 41 and 44 in header area) ### Note on interpretation - Increase in financial conditions index means tighter conditions. - VIX = Chicago Board Options Exchange Volatility Index. - Standard errors clustered at country level; standard errors in parentheses. - Significance codes: ***p < 0.01; **p < 0.05; *p < 0.1. --- ### Annex Table 2.2.3. Impact of Policy and Institutional Settings on the Riskiness of Credit Allocation ### Overview of regressions by policy/institutional category - Dependent variable: Riskiness of Credit Allocation Based on Leverage - Columns cover Financial Soundness, Macroprudential Policy, Supervision Quality, Legal and Institution Aspects, plus robustness/significance summaries. - Change in Credit-to-GDP Ratio (selected columns shown): - Columns (1)–(5): 0.05*** (0.01) - Column (6): 0.12*** (0.02) - Column (7): –0.00 (0.02) - Column (8): 0.15*** (0.03) - Specific policy/institution interactions (selected significant coefficients): - Lag Buffers from Banking Default: 0.01 (0.01) - Change in Credit-to-GDP Ratio × Lag Buffers from Banking Default: 0.005** (0.002) - Net Tightening of Capital Conservation Buffers: –0.45** (0.21) - Change in Credit-to-GDP Ratio × Net Tightening of Capital Conservation Buffers: –0.09*** (0.03) - Net Tightening of Minimum Leverage Ratio: –0.29 (0.20) and –0.30 (0.20) in alternate columns - Change in Credit-to-GDP Ratio × Net Tightening of Minimum Leverage Ratio: –0.09* (0.05) and –0.09* (0.05) - Net Tightening on Ceilings and Penalties on Bank Credit Growth: –0.57 (0.54) (two columns) - Change in Credit-to-GDP Ratio × Net Tightening on Ceilings and Penalties on Bank Credit Growth: –0.07** (0.03) (two columns) - Change in Credit-to-GDP Ratio × Independence of Supervisory Authority from Bank: –0.09*** (0.02) - Change in Credit-to-GDP Ratio × Rareness of State-Owned Enterprises: –0.01* (0.01) - Change in Credit-to-GDP Ratio × Minority Shareholder Protection Index: –0.02*** (0.01) - Controls: Yes in all columns; Country Cluster/Fixed Effects and Year Fixed Effect: Yes in all columns - Observations by column: 861 / 976 / 976 / 976 / 976 / 929 / 739 / 898 - Number of Countries by column: 55 / 54 / 54 / 54 / 54 / 52 / 37 / 46 - R 2 by column: 0.33 / 0.31 / 0.31 / 0.31 / 0.32 / 0.32 / 0.34 / 0.34 - Column (4) is a horse race among macroprudential measures; robustness counts reported in columns (9) and (10) (entries shown as 3/0 etc in table for individual rows) ### Interpretative notes - Real GDP growth and domestic currency appreciation vis-à-vis the US dollar are controlled in all regressions. - For macroprudential policies, robustness information is based on the horse race in column (4). - Legal and supervisory variables averaged at country level and enter only as interaction terms due to limited time-series variation. - Standard errors clustered at country level; standard errors in parentheses. - Significance codes: ***p < 0.01; **p < 0.05; *p < 0.1. --- ### Annex 2.3. The Riskiness of Credit Allocation and Macro-Financial Outcomes ### Empirical frameworks (methods) - Systemic banking crisis risk (panel logit model, equation A2.3.1): - Dependent variable: Crisisstart (dummy = 1 at start of systemic banking crisis) - Key regressors: ∆ Credit (change in ratio of bank credit to nonfinancial private sector to nominal GDP), Riskiness (leverage, interest coverage ratio, debt overhang, expected default frequency) - Explanatory variables: lag of simple three-year moving average, demeaned at country level - Controls: change in current-account-balance-to-GDP ratio, real GDP growth, financial conditions index - Change in credit-to-GDP ratio winsorized at 1 percent. - Banking sector equity stress risk (panel logit, equation A2.3.2): - Dependent variable: stress (dummy = 1 if banking-sector annual excess equity return below country mean by more than one standard deviation in window t to t + h, h = 0,...,3) - Key regressors and controls similar to crisis model; financial conditions index included. - Downside risks to GDP growth (equation A2.3.3): - Dependent variable: ∆ y i,t,t+h (cumulative real GDP growth over future h years, h = 1,...,3) - Regressors: ∆ Credit, Riskiness, interaction ∆ Credit × Riskiness, controls include real GDP growth and financial conditions index (which includes sovereign spread) - Estimation using quantile regressions with nonadditive fixed effects for 20th and 50th percentiles; also examined via logit with low-growth dummy (below 20th percentile). ### Panel logit results: Probability of start of systemic banking crisis (Annex Table 2.3.1) - Change in Credit-to-GDP Ratio coefficients: - Column (1): 0.202*** (0.0699) - Column (2): 0.141* (0.0737) - Column (3): 0.0565 (0.0849) - Column (4): 0.0745 (0.100) - Column (5): 0.0808 (0.108) - Column (6): –0.0902 (0.131) - Financial Conditions Index: - Column (1): –1.742** (0.682) - Column (2): –2.536*** (0.611) - Column (3): –2.686*** (0.604) - Column (4): –2.907*** (0.724) - Column (5): –4.441*** (0.854) - Riskiness measures (included individually in columns): - Riskiness_Leverage: 1.924*** (0.674) - Riskiness_Interest Coverage Ratio: 2.533*** (0.861) - Riskiness_Debt Overhang: 2.087*** (0.461) - Riskiness_Expected Default Frequency: 2.113*** (0.734) - Observations: 443 / 443 / 443 / 443 / 431 / 361 (columns (1)–(6)) - Number of Countries: 21 / 21 / 21 / 21 / 20 / 17 - Pseudo R 2 : 0.243 / 0.353 / 0.465 / 0.487 / 0.515 / 0.606 - Standard errors in parentheses; explanatory variables enter as lag of three-year moving average and are demeaned; change in credit winsorized at 1 percent. - Significance codes: ***p < 0.01; **p < 0.05; *p < 0.1. ### Panel logit results: Banking sector equity stress risk (Annex Table 2.3.2) - Change in Credit-to-GDP Ratio: coefficients vary by horizon; examples: - t + 1 (column 1): –0.000975 (0.0381) - t + 3 (column 2): 0.01290 (0.0433) - other columns show small coefficients up to 0.03450 (0.0437) and 0.0253 (0.0464) - Riskiness measures (selected significant coefficients): - Riskiness_Leverage: 0.898*** (0.246) and 0.727*** (0.246) in alternate horizons - Riskiness_Interest Coverage Ratio: 0.690*** (0.256) and 0.717** (0.320) - Riskiness_Debt Overhang: 0.569** (0.223) and 0.440 (0.271) - Riskiness_Expected Default Frequency: 0.451* (0.274) and 0.321 (0.296) - Observations vary by specification: e.g., 573 / 573 / 573 / 573 / 552 / 552 / 505 / 505 - Number of Countries vary: 36 / 36 / 36 / 36 / 34 / 33 / 33 / 33 - Pseudo R 2 range: 0.0388 to 0.1300 - Standard errors in parentheses; explanatory variables lagged three-year moving averages and demeaned; change in credit winsorized at 1 percent. - Significance codes: ***p < 0.01; **p < 0.05; *p < 0.1. ### Quantile regression results: Impact on downside risks to growth (Annex Table 2.3.3) - Dependent: Cumulative Real GDP Growth Rate over Future Three Years (t, t + 3); columns present 20th percentile (20 pt) and 50th percentile (50 pt) estimates for different risk measures. - Change in Credit-to-GDP Ratio (consistent negative impact across columns; examples): - Column (1) 20 pt: –0.232*** (0.0335) - Column (2) 50 pt: –0.268*** (0.0358) - Column (3) 20 pt: –0.239*** (0.0364) - Column (4) 50 pt: –0.254*** (0.0471) - (Further columns show similar negative coefficients ranging from –0.172*** to –0.291***) - Riskiness_Leverage: - Coefficients: –0.468*** (0.144) and –0.480*** (0.107) in two specifications - Change in Credit-to-GDP Ratio × Riskiness_Leverage: - –0.0549** (0.0253) - –0.0820*** (0.0288) - Riskiness_Interest Coverage Ratio (ICR): - –0.927*** (0.207); –0.421*** (0.118); –1.306*** (0.237); –0.391*** (0.0948) across specifications - Change in Credit-to-GDP Ratio × Riskiness_ICR: - –0.237*** (0.0467) - –0.217*** (0.0360) - Riskiness_Debt Overhang: - –0.406* (0.237); –0.328** (0.140); –0.522*** (0.161); –0.229* (0.132) - Change in Credit-to-GDP Ratio × Riskiness_Debt Overhang: - –0.146*** (0.0297) - –0.204*** (0.0277) - Riskiness_Expected Default Frequency: - –0.879*** (0.243); –0.383** (0.161); –0.942*** (0.233); –0.397** (0.190) - Change in Credit-to-GDP Ratio × Riskiness_Expected Default Frequency: - –0.0798 (0.0749) - –0.171*** (0.0199) - Observations across columns: typically 602 (some columns 592 or 532) - Number of Countries: 41 (some specifications 39) - Controls include real GDP growth and a financial conditions index; explanatory variables are lagged three-year moving averages and demeaned; change in credit winsorized at 1 percent. - Significance codes: ***p < 0.01; **p < 0.05; *p < 0.1. --- *Source: IMF staff estimates (Annex Table 2.2.1; Annex Table 2.2.2; Annex Table 2.2.3; Annex 2.3 and Annex Tables 2.3.1–2.3.3).* ### CHAPTER 2 ThE RISkINESS OF CREdIT ALLOCATION: A SOuRCE OF FINANCIAL VuLNERABILITY? ### CHAPTER 2 ThE RISkINESS OF CREdIT ALLOCATION: A SOuRCE OF FINANCIAL VuLNERABILITY? ### Summary findings - Rising house prices have been a feature of the economic recovery in many countries since the global financial crisis. - The chapter finds an increase in house price synchronization, on balance, for 40 countries and 44 major cities in advanced and emerging market economies. - Countries’ and cities’ exposure to global financial conditions may provide an explanation for the increase in house price synchronization. - Cities in advanced economies may be particularly exposed to global financial conditions, perhaps owing to their integration with global financial markets or to their attractiveness for global investors searching for yield or safe assets. - Macroprudential policies seem to retain some ability to influence local house price developments even in countries with highly synchronized housing markets. - Macroprudential policy measures put in place to tame rising vulnerabilities in a country’s financial sector may have the additional effect of reducing a country’s house price synchronization with the rest of the world. ### Evidence on synchronization and its drivers - In 2017, there was a pickup in growth in 120 economies, accounting for three-quarters of world GDP, described as the broadest synchronized global growth upsurge since 2010 (IMF 2018a). - Recent increases in house prices have been occurring in an environment of easy financial conditions in major advanced economies characterized by low policy rates, compressed spreads, and low volatility that has spread globally. - Real estate investments—including in residential real estate—by private equity firms, real estate investment trusts (REITs), and institutional investors appear to have grown in recent years, and anecdotes point to increasing investor participation in select housing markets such as Amsterdam, Melbourne, Sydney, Toronto, and Vancouver. - Data notes: Latest available data as of 2017:Q2 for most economies; fewer than 15 economies have data through 2017:Q3. ### Risks from synchronized house prices - Higher house price synchronization can signal stronger transmission of external shocks to local housing markets and may propagate local economic and financial shocks. - A sharp reversal of the prevailing accommodative global financial conditions could challenge policymakers’ ability to stabilize household balance sheets, financial markets, and economic activity if house prices across many countries decline at once. - While synchronization may bring benefits—more liquidity in housing and mortgage markets, higher capital flows from abroad, and enhanced risk-sharing—it may also increase exposure of local housing markets to global financial conditions or to shocks affecting foreign investors active in local markets. ### Policy implications and considerations - Policymakers cannot ignore the possibility that shocks to house prices elsewhere may affect domestic markets. - House price synchronization in and of itself may not warrant policy intervention, but heightened synchronicity can signal a downside tail risk to real economic activity, especially when taking place in a buoyant credit environment. - Macroprudential policies have some ability to influence local house price developments and may reduce a country’s house price synchronization with the rest of the world; these unintended effects should be considered when evaluating the trade-offs of implementing macroprudential and other policies. *International Monetary Fund | April 2018* ### 1. Advanced Economies: Country Level2. Advanced Economies: City Level ### 1. Advanced Economies: Country Level2. Advanced Economies: City Level ### Main questions addressed - Trends in synchronization of house prices across countries and major cities; whether synchronization has increased in recent years and before the global financial crisis. - Factors that contribute to or dampen synchronicity: role of financial factors versus comovement in economic activity; importance of bilateral/two-way links versus global factors. - Policy relevance: whether policymakers should monitor house price synchronicity to better understand financial vulnerabilities and risks. ### Key findings - On balance, synchronization in house prices across countries and major cities has increased over the past several decades in advanced and emerging market economies. - The short-term comovement in house prices sharply increases around the time of global recessions in advanced economies; these spikes are much larger among major cities than at the country level. - Global financial conditions contribute to synchronization in house prices across pairs of countries and cities even after accounting for comovement in economic activity and other fundamentals; contribution is particularly strong in major cities in advanced economies. - The dynamics of house prices resemble those of other financial assets: the expected return to investing in housing varies over time and is predictable in the long term, a pattern associated with variations in the risk premium demanded by investors. - Higher house price synchronization corresponds to increased downside risks to growth at horizons of up to one year, controlling for other financial and macroeconomic conditions. ### Institutional investor participation (selected figures from the chapter) - Market capitalization index of REITs normalized by the total market capitalization (index, 2005:Q1 = 100): figure axes indicate values in the 50–300 range with time series from 2005:Q1 through 2018. - Weighted average target allocation to real estate: All Institutions (Percent): selected tick values shown include 8.5, 9.0, 9.5, 10.0, 10.5. - Developed markets: aggregate REITs series compiled by Thomson Reuters Datastream (country classification aligned with Morgan Stanley Capital International and Dow Jones). ### Conceptual framework: supply, demand, financial channels - Supply considerations: costs of construction and land acquisition. - Demand considerations: demographics, tax and policy considerations, depreciation and maintenance. - Financial demand-side factors: mortgage interest rate, risk premium on assets with similar risk characteristics as housing, household leverage, expected nominal house price appreciation rate. - Mechanisms leading to synchronization: - Changes in global financial conditions (international transmission via capital flows; global demand for safe assets compressing sovereign bond rates and mortgage rates). - Portfolio channels (common lenders or investors causing interdependence in mortgage lending or investment behavior; global institutions pulling back lending or liquidating holdings across countries). - Changes in expected capital gains (coordinated changes in households’ or investors’ views across countries; propagation through social networks and investor behavior). ### Empirical measures and patterns - Synchronization measures used: cyclical component of real house prices (house price gap) and quarterly growth rate in real house prices. - The house price gap has become more synchronized in countries and cities in advanced and emerging market economies. - Synchronization in the house price gap reflects medium-term changes in how shocks propagate across markets; quarterly growth measures capture higher-frequency comovement. ### Policy implications and recommendations - Policymakers may wish to monitor synchronization of house prices relative to other countries, in addition to valuation within a country; improving granularity, timeliness, and coverage of house price data within countries would aid bilateral and multilateral surveillance. - More comprehensive data on participation of global and institutional investors in housing markets would strengthen surveillance efforts. - Macroprudential policies can influence local house price developments even in highly synchronized markets, though their effectiveness may be weaker compared with less synchronized markets. Evidence indicates macroprudential measures put in place to tame rising vulnerabilities are followed by a decline in a country’s house price synchronization. - Fiscal-based policies (for example, ad valorem and buyers’ stamp duty taxes) may lower house price synchronization but tend to do so less than measures such as limits on loan-to-value ratios. - Other policies that enhance resilience to global financial shocks may dampen house price synchronicity: exchange rate flexibility, policies that deepen domestic real estate markets, and consumer financial protections that discourage excessive or predatory lending. ### Risks from synchronized house price declines - Simultaneous house price declines across countries compound macro-financial stability risks: coincident household deleveraging and contraction in external demand leave little room for the current account to offset domestic demand contraction. - Large and widespread house price swings have historically been associated with periods of financial instability across many countries at once. - A pullback among global investors could trigger fire sales across asset classes, capital flight, and tighter mortgage market conditions, amplifying downturns. *Source: IMF staff calculations.* ### 1. Within Advanced Economies ### 1. Within Advanced Economies ### Trends in house price synchronization - Between 1991 and 2016, synchronicity is lower among major cities in advanced economies than among the countries where they are located, but it has gradually moved closer to country-level synchronicity. - The share of the variation in house price growth explained by a common global factor (dynamic factor model) increases from about 10 percent to 30 percent over the period from 1971 to 2016. - Short-term comovement (instantaneous quasi correlation) in house price gaps increases sharply around the time of global recessions in advanced economies; this increase before recessions is much larger between major cities than between countries. - The instantaneous quasi correlation measure is constructed not to have a trend; sharp increases around global economic downturns suggest common shocks affecting housing markets in many advanced economies. ### Regional and cross-country/city differences - Exposure to the common global factor varies across countries and cities, with a larger contribution of this factor to house prices in countries and cities in Europe than in other regions. - Advanced economies are more exposed than emerging market economies to global factors, and the relative importance of the global factor has increased over time but not uniformly across advanced economies. - Cities may be highly interconnected even when their countries are not; some financial centers (cities) are more centrally positioned and influential, implying city-level house price dynamics may transmit across borders (example: Tokyo more central among cities than Japan is among countries). ### Interconnectedness and spillovers - Network analysis shows average share of house price variance in a country accounted for by changes in another single country (“spillovers”) increased from 1.4 percent in 1990–2006 to 2.1 percent in 2007–16. - Spillovers are particularly strong among advanced economies. The proportional increase is largest for: - spillovers from advanced economies to emerging market economies (about 60 percent increase), - and then from emerging market economies to advanced economies (about 40 percent increase). - Many large advanced economies’ housing markets are closely interconnected (central nodes in network maps); many emerging market economies show weaker connectivity. ### Financial factors and drivers of synchronization - Countries with deeper bilateral banking linkages exhibit more house price synchronization; this relationship is nearly as large as that between business cycle synchronization and house price synchronization. - A country’s financial openness (proxied by the Chinn-Ito index) contributes to house price synchronicity: countries with greater capital account openness are more exposed to global factors. - Past increases in global liquidity, good market sentiment, and loose global financial conditions are strongly associated with higher short-term comovement in house prices (instantaneous quasi correlation), even after accounting for business cycle comovement. - Greater exchange rate flexibility appears to dampen the importance of global financial conditions; the impact of global liquidity is lower in countries with high exchange rate flexibility. ### Methodology and measurement notes (as reported) - Synchronicity measures used: - synch1: negative value of the absolute difference in house price gaps (value closer to zero suggests differences have declined), - QCORR: instantaneous quasi correlation of house price gaps (higher value implies simultaneous deviations above or below historical averages). - Dynamic factor model estimation: factor loadings and vector autoregression parameters estimated by a two-step procedure (Koop and Korobilis 2013) using data for 19 advanced economies from 1971:Q2 to 2016:Q4. - Vector autoregression network analyses control for global factors and are based on quarterly house price growth rates; sample periods vary (country-level 1990:Q1–2016:Q4; city-level 2004:Q1–2017:Q2). - Standard-deviation-based presentation: statistically significant standardized coefficients are calculated using coefficients in specification 7 in Annex Table 3.2.1 and their standard deviations; the standard deviation for business cycle synchronization is 0.012 and 1.00 for bilateral banking integration. ### Key quantitative points (preserved exactly) - Share of variance explained by global factor: increases from about 10 percent to 30 percent (1971–2016). - Rolling window reported for dynamic factor model: Window = 15 years. - Spillovers: increased from 1.4 percent (1990–2006) to 2.1 percent (2007–16). - Standard deviation for business cycle synchronization: 0.012. - Standard deviation for bilateral banking integration: 1.00. _Italic: Source: IMF staff estimates; text excerpt from Chapter 3, "HOUSE PRICE SYNCHRONIZATION: WHAT ROLE FOR FINANCIAL FACTORS?", Global Financial Stability Report: A Bumpy Road Ahead, International Monetary Fund | April 2018._ ### CHAPTER 3 hOuSE PRICE SYNChRONIzATION: WhAT ROLE FOR FINANCIAL FACTORS? ### CHAPTER 3 hOuSE PRICE SYNCHRONIzATION: WhAT ROLE FOR FINANCIAL FACTORS? ### Global financial conditions and house price synchronization - House price synchronization has increased over the past three decades, especially among major cities. - The contribution of global financial conditions to house price synchronization in cities is somewhat larger than for countries. - Cities in advanced economies show greater responsiveness to global financial conditions, using global liquidity as a proxy. - Possible mechanisms: - Global investors searching for yield or safe assets in residential real estate can induce comovement across markets. - Higher-priced homes are more responsive to changes in house prices of non-US cities (granular analysis within the United States). - Expected returns on housing assets are predictable in the long term and this predictability is greater in countries with high capital account openness, suggesting risk sentiment of global investors matters more when capital account openness is high. - Empirical notes and measures: - Synchronicity is measured by the share of the variation in house price growth from 2002 to 2016 explained by a common global factor in the dynamic factor model. - Rolling estimations use a 15-year window for the share of variation explained by a common global factor (Figure 3.14; Window = 15 years). - Global liquidity is used as a proxy for global financial conditions (Figure 3.15). ### House price synchronization and risks to growth - Higher house price synchronization corresponds to increased downside risks to growth at horizons of up to one year. - In a growth-at-risk framework, house price synchronization—as measured by the instantaneous quasi correlation between a country’s house price growth and the global factor—negatively affects the lower tail of the growth distribution, over and above the risks associated with the price of risk, leverage, and external conditions. - Interaction with leverage: - At short horizons, the negative relationship between house price synchronization and risks to future growth is amplified when leverage is high. - The negative impact of house price synchronicity is about twice as large when leverage is higher. - When leverage is high, the magnitude of the relationship between house price synchronization and future growth is about two-thirds that of financial conditions, which measure the price of risk. - Quantitative estimation notes: - Coefficients are standardized. - Solid bars denote statistically significant quantile regression coefficients at a 10 percent confidence interval (Figure 3.16). - For city level, separate bars correspond to city pairs in country pairs that are AEs, EMEs, or AE-EMEs. - Standard deviation of the country-level dependent variable is approximately 0.85, and standard deviation of the city-level dependent variable is approximately 0.97. ### Behavior of housing as a financial asset - Housing is a major asset class: in a typical economy, housing wealth, on average, accounts for roughly one-half of total national wealth. - Average annual real return on housing assets between 1950 and 2015 lies between 5 percent and 8 percent in many advanced economies. - Housing return predictability: - A high current house-price-to-rent ratio strongly predicts low housing return in the future and vice versa. - Predictive power increases with the forecasting horizon (Figure 3.2.1); forecasting horizon ranges from 1 year to 10 years. - The y-axis in the predictability figure shows the R2 from the forecasting equation (proportion of variance in future housing return explained by current price-to-rent ratio). - Predictability of housing returns nine years ahead (R2) is shown in relation to capital account openness (Figure 3.2.2). ### Evidence on global investors and house price dispersion (Box 3.1) - House price dispersion can proxy demand from high-net-worth foreign investors preferring luxury housing. - Measure constructed for 40 largest US cities: ratio of the top and bottom deciles of house prices (equivalent to the interpercentile range at log scale, using Zillow data at ZIP code granularity). - Empirical findings: - House price dispersion in the United States has increased sharply over recent decades (Figure 3.1.1). - Substantial comovement exists between real house prices and house price dispersion. - Regression analysis finds a statistically significant positive relationship between US house price dispersion and house prices in major non-US cities (Beijing, Dublin, Hong Kong SAR, London, Seoul, Shanghai, Singapore, Tokyo, Toronto, Vancouver). - The coefficient associated with the foreign city index is positive and significant in all specifications considered, including those controlling for domestic determinants (unemployment rate, VIX, effective federal funds rate, 30-year fixed-rate average mortgage interest rates, mortgage-backed security holdings of large domestically chartered commercial banks). - Interpretation caveats: - Alternative interpretation: luxury houses may be located in areas with tighter supply constraints, amplifying price responses; however, the positive relationship between US house price dispersion and foreign city prices remains after controlling for domestic determinants. ### Policy implications and recommendations - Monitoring: - Policymakers should monitor synchronization in addition to over- or undervaluation of house prices to understand trade-offs associated with greater global links in housing markets. - Increasing the granularity, timeliness, and coverage of data on house prices would provide richer indicators for bilateral and multilateral surveillance. - More comprehensive data on participation of global investors in housing markets would strengthen surveillance efforts. - Macroprudential measures: - Demand-side macroprudential policy measures (for example, loan-to-value limits) are typically followed by declines in house price growth, but the decline is larger and more persistent in countries with low house price synchronicity. - Macroprudential measures aimed at dampening domestic financial vulnerabilities may also reduce a country’s house price synchronization (Box 3.4). - Fiscal-based measures (ad valorem and buyers’ stamp duty taxes) may lower house price synchronization but to a lesser extent than demand-based measures. - Policies to deter foreign buyers: - Systematically identifying the impact of foreign buyers on housing affordability is difficult because of data limitations, creating uncertainty in timing and method of interventions. - A range of instruments—tax policy, land-use regulation, and macroprudential policy—may be contemplated, but effectiveness is uncertain and policies may be circumvented; limiting purchases in one area may steer buyers elsewhere, suggesting a role for national or international coordination. - Enhancing resilience: - Policies that enhance resilience to global financial shocks may also dampen house price synchronicity. - Exchange rate flexibility appears important by giving monetary authorities more room to influence domestic conditions. - Deeper domestic real estate markets and consumer financial protections to limit excessive or predatory lending may help insulate households and limit fallout from deleveraging. *International Monetary Fund | April 2018* ### Box 3.2. Housing as a Financial Asset ### Box 3.2. Housing as a Financial Asset ### Key findings on housing return predictability - A high degree of housing return predictability indicates that house price variation is driven mostly by time-varying risk premiums on housing assets as opposed to shocks to rental income growth. - As a result, volatility of house prices is generally much higher than suggested by the volatility of rent growth. - Empirical evidence suggests that housing return predictability is particularly strong in countries with high capital account openness. - In an integrated global financial system, global financial conditions can significantly affect domestic house price variation because domestic prices are more likely to be affected by the risk sentiment of global investors. - Consequently, house prices in these countries are more prone to temporary deviations from their domestic rental market fundamentals and are likely to exhibit excess volatility. ### Data and sample note - The analysis is based on a sample of 20 advanced economies that have long time series for the price-to-rent ratio. - The estimated relationships may or may not be the same when emerging market economies are also considered. *Source: Box 3.2, Global Financial Stability Report: A BuMPY ROAd AhEAd, International Monetary Fund | April 2018* ### CHAPTER 3 hOuSE PRICE SYNChRONIzATION: WhAT ROLE FOR FINANCIAL FACTORS? ### CHAPTER 3 hOuSE PRICE SYNChRONIzATION: WhAT ROLE FOR FINANCIAL FACTORS? ### Measuring Synchronization - Instantaneous quasi correlation (QCORR) in house price gaps (HPsynchijt = QCORRijt) defined as: - HPsynchijt = QCORRijt = ((HPgapit − ̄HPgapi)(HPgapjt − ̄HPgapj)) / (σigap σjgap) (A3.2.1) - HPgapit and HPgapjt are house price gaps of countries i and j at quarter t. - ̄HPgapi and ̄HPgapj are average house price gaps of countries i and j. - σigap and σjgap are standard deviations of house price gaps of countries i and j. - Alternative synchronization measure (Synch1): - HPsynchijt = Synch1ijt = − |HPgapit − HPgapjt| (A3.2.2) - Dynamic factor model measure (synchL,i,t) defined as: - synchL,i,t = varL(λi,t gt) + varL(λr,i r k,t) / varL(hi,t) - or synchL,i,t = varL(λi,t gt) / varL(hi,t) (A3.2.3) - varL(⋅) is realized variance from t − L to t. - λi,t and λr,i are factor loadings to global (gt) and regional (rk,t) factors. - hi,t is quarterly growth rate of house prices for country i in period t, decomposed into global factor gt, regional factor rk,t (k = Europe, Asia, the Americas), and country-specific idiosyncratic component ci,t. - House price gaps construction: - Cyclical component of real house prices extracted using the Christiano and Fitzgerald (2003) band-pass filter with a maximum length of 20 years to capture medium-term financial cycles. - Cyclical components taken as a ratio of house price levels to obtain house price gaps. - Robustness check: Hodrick and Prescott (1997) filter with lambda = 400,000 yields broadly consistent gaps. - CF filter chosen to avoid tail bias and compute cyclical component for all observations. ### Country-Pair Analysis (baseline specification) - Econometric baseline (quarterly frequency, 1990–2016, 40 countries): - HPsynchijt = αij + β1 BCSijt−1 + β2 FININTijt−1 + β3 GLOBALt−1 + β4 INSTijt−1 × GLOBALt−1 + β5 OTHERijt−1 + tr + εijt (A3.2.4) - HPsynchijt: synchronization of house price gaps between country-pairs i and j at quarter t. - BCSij: business cycle synchronization between countries i and j. - FININTij: bilateral financial integration between countries i and j. - GLOBALt: global factor proxied by changes in global liquidity. - INSTij: dummies = 1 if both countries have a high level of an institutional characteristic (economic development level, capital account openness, exchange rate flexibility, or financial development). - OTHERij: other controls (for example, institutional factors). - All regressors lagged by one quarter. Linear and quadratic time trends (tr) included. αij = country-pair fixed effects. εijt = error term. - Data notes: - Analysis restricted to series beginning in 1990 because availability of data on bilateral banking links improves that year. - Bilateral banking integration measured using BIS locational banking statistics (residency basis); measured as logarithm of sum of bilateral claims of country i vis-à-vis country j and of country j vis-à-vis country i as a ratio of sum of GDPs of i and j. - Additional bilateral financial integration measures (portfolio links, direct investment) not used due to lower frequency and shorter span. - Standard errors multiway clustered (country i, country j, time level), with alternatives used for robustness. ### Key Empirical Results (Annex Table 3.2.1 — Dependent variable: Synch1) - Coefficients and standard errors (parentheses) across specifications (selected): - Business Cycle Synchronization of ij: - 0.766*** (0.254) - 0.675** (0.293) - 0.733*** (0.243) - 0.657** (0.254) - 0.658** (0.253) - 0.746*** (0.262) - 0.725*** (0.261) - 0.725*** (0.262) - 0.675** (0.253) - 0.706** (0.337) - Bilateral Bank Integration of ij: - 0.006* (0.003) - 0.007** (0.003) - 0.012 (0.007) - 0.009* (0.004) - 0.007** (0.003) - 0.007* (0.004) - 0.007** (0.003) - 0.004 (0.005) - Global Factor (global liquidity): - −0.001 (0.001) in multiple specifications; 0.001 (0.001) in one specification. - Interaction: Bilateral Bank Integration × High Financial Openness with the World (ij): - −0.019*** (0.004) - GFC Dummy: - 0.048*** (0.011) - Post-GFC Dummy: - 0.042*** (0.009) - Model fit and samples: - Observations reported: 65,450; 65,343; 49,384; 49,384; 49,384; 43,871; 46,708; 46,708; 47,353; 49,384 (varies by specification). - R2 values: 0.353; 0.498; 0.386; 0.356; 0.356; 0.361; 0.356; 0.356; 0.360; 0.360. - Additional notes: - Standard deviation for business cycle synchronization = 0.0124. - Standard deviation for bilateral bank integration = 1.040. - Multiway clustering used except regression 10 where two-way clustering at country i and country j used. - Significance codes: ***p < 0.01; **p < 0.05; *p < 0.1. ### Key Empirical Results (Annex Table 3.2.2 — Dependent variable: Quasi correlation) - Coefficients and standard errors (parentheses) across specifications (selected): - Business Cycle Synchronization of ij: - 0.025* (0.013) - 0.030** (0.014) - 0.022 (0.014) - 0.026* (0.013) - 0.026* (0.013) - 0.025* (0.015) - 0.026* (0.014) - 0.026* (0.014) - 0.026** (0.013) - 0.042 (0.033) - Bilateral Bank Integration of ij: - −0.011 (0.033) - 0.012 (0.031) - 0.012 (0.031) - 0.011 (0.036) - 0.022 (0.036) - 0.022 (0.035) - 0.012 (0.032) - −0.016 (0.034) - Global Factor (global liquidity): - 0.016** (0.006) - 0.016** (0.008) - 0.020** (0.008) - 0.019*** (0.007) - 0.019** (0.007) - 0.018** (0.007) - 0.022* (0.013) - Interaction: Global Factor × High Exchange Rate Regime (15 categories; high = more flexible): - −0.023*** (0.008) - GFC Dummy: - −0.137** (0.060) - Post-GFC Dummy: - −0.044 (0.052) - Model fit and samples: - Observations reported: 65,450; 65,343; 49,384; 49,384; 49,384; 43,871; 46,708; 46,708; 47,353; 49,384. - R2 values: 0.227; 0.354; 0.251; 0.230; 0.230; 0.233; 0.224; 0.223; 0.241; 0.232. - Significance codes: ***p < 0.01; **p < 0.05; *p < 0.1. ### Robustness Checks and Additional Findings - Alternative proxies for the global factor include: - US financial conditions index (FCI), global FCI, Chicago Board Options Exchange Volatility Index (VIX), US shadow interest rates (in the spirit of Wu and Xia 2016; and Krippner 2013). - Interest rate synchronization: - Interest rate synchronization is a statistically significant driver of house price synchronization on its own when either Synch1 or quasi correlation used. - Interest rate synchronicity retains statistical significance above and beyond other financial factors (global liquidity, bilateral banking links) only under less stringent standard error clustering. - Trade integration: - Included as additional control; found not to be statistically significant. - Equity price synchronization: - Including equity price synchronization leaves main results broadly unchanged. - Equity price synchronization itself does not consistently have a statistically significant relationship with house price synchronization. - Clustering alternatives and time controls: - Various clustering alternatives used (country-pair level; two-way at country i and country j; two-way at country-pair and time levels; without clustering robust). - Level of significance improves under less restrictive clustering. - Year fixed effects and linear time trends included with little change to main conclusions. - Additional robustness: dropping one country pair at a time; panel of three nonoverlapping seven-year periods using bilateral Pearson correlation coefficients — interaction of global factor and foreign exchange regime remains statistically significant. - Long-run historical check: - Relationship between house price gap synchronicity and business cycle synchronization positive and statistically significant when using Jordà, Schularick, and Taylor (2017) data set (starts in 1870 for 17 advanced economies at annual frequency). - Further analysis limited by data availability. ### Economies and Cities Included (selection) - Economies included (sample): Australia; Austria; Belgium; Canada; Chile; China; Colombia; Cyprus; Czech Republic; Denmark; Estonia; Euro area; Finland; France; Germany; Greece; Hong Kong SAR; Hungary; India; Indonesia; Ireland; Israel; Italy; Japan; Korea; Malaysia; Mexico; Netherlands; New Zealand; Norway; Philippines not listed in economies but cities include Manila; Poland not listed; Portugal; Russia; Singapore; Slovenia; South Africa; Spain; Sweden; Switzerland; Taiwan Province of China; Thailand; Turkey; United Kingdom; United States; Serbia. - Cities included (selection of largest population centers / top global investor cities): Amsterdam; Athens; Auckland; Bangkok; Belgrade; Berlin; Bogotá; Brussels; Budapest; Buenos Aires; Copenhagen; Dubai; Dublin; Finland metro area; Greater Stockholm; Hong Kong SAR (urban areas); Inner Paris; Istanbul; Jakarta; Kuala Lumpur; Lima; Lisbon; Ljubljana; London; Madrid; Manila; Mexico City; Mexico City repeated as sample; Metro areas including Moscow; Mumbai; New York City; Oslo; Prague; Rome; Santiago; Seoul (Southern); Singapore (core central region); Shanghai; Taipei City; Tallinn; Tokyo; Toronto; Vienna; Zurich; São Paulo; Sydney. - Note: Cities selected are the largest cities based on population owing to data availability, and overlap with the top 50 cities for global investors identified by Cushman & Wakefield (2017). An additional sample of 76 cities based on top 30 cities for global investors used in robustness checks. *Source: IMF staff estimates (Annex 3.2, CHAPTER 3 hOuSE PRICE SYNChRONIzATION: WhAT ROLE FOR FINANCIAL FACTORS?, International Monetary Fund | April 2018).* ### Annex 3.3. Technical Annex ### Annex 3.3. Technical Annex ### Measuring Synchronicity: Conceptual Issues - Framework (two-country example, based on Doyle and Faust 2005): - House prices decomposed into a common factor εc and idiosyncratic factors εi and εj: - hi = εc + εi + γ hj, and hj = εc + εj + γ hi. (A3.3.1) - Interconnectedness parameter: 0 ≤ γ < 1. - Closed-form expressions: - hi = [1 / (1 − γ^2)] [εi + γ εj + (1 + γ) εc]. (A3.3.2) - hj = [1 / (1 − γ^2)] [εj + γ εi + (1 + γ) εc]. (A3.3.2) - Assumptions: σi = σj; σij = σic = σjc = 0; mean zero house prices. - Three measures of synchronization: 1. Instantaneous quasi correlation qc_ijt: - qc_ijt = (hi_t hj_t) / (σ_hi σ_hj) = [1 / (1 − γ^2)^2 σ_hi σ_hj] × [γ(εit^2 + εjt^2) + (1 + γ) εit εjt + (1 + γ)(εit εct + εct εjt + εct^2)]. (A3.3.3) - Interpretation: - When γ is not very large, idiosyncratic squared terms εit^2 + εjt^2 have limited effect. - Interaction terms εi εj, εi εc, εc εj fluctuate around zero. - Systematic movements of qc_ijt are driven by (1 + γ) εc^2. - Suitable for identifying short-term comovement caused by the common shock; sharp movements observed around global recessions in advanced economies. 2. Bilateral absolute difference ad_ijt: - ad_ijt = −|hi_t − hj_t| = −[1 / (1 + γ)] |εit − εjt|. (A3.3.4) - Interpretation: - Independent of the common shock because it cancels out. - Captures long-term trend in synchronicity driven by changes in γ. - An increasing trend in ad_ijt implies rising interconnectedness γ over the long term. 3. Relative contribution of the global factor rci: - rci = var(((1 + γ) / (1 − γ^2)) εc) / var(hi) = σc^2 / [((1 + γ^2) / (1 + γ)^2) σi^2 + σc^2]. (A3.3.5) - Interpretation: - Estimated over a long-term window (for example, 15 years) to identify long-term synchronization trends. - An increasing rci can reflect: (1) σc has risen; (2) σi has declined; (3) γ has risen. - Comprehensive measure but empirically difficult to disentangle these three channels. ### Estimation of a Dynamic Factor Model - Model decomposition (quarterly growth rate hi,t): - hi,t = λg,i gt + λr,i rk,t + ci,t. (A3.3.6) - λg,i and λr,i are factor loadings on global and regional factors. - Regional factor extracted from residuals after extracting the global factor; regions: Europe, Asia, Americas. - Global and regional factors follow a VAR jointly with global output, global inflation, and the global interest rate (first principal components across countries). - Time-varying factor loadings and VAR parameters estimated by the two-step procedure in Koop and Korobilis (2013). - Panel regression linking synchronization to openness: - synchL,i,t = αi + δt + β1 kaopeni,t + β2 tri,t + γ Zi,t + εi,t. (A3.3.7) - αi: country fixed effect; δt: time dummy. - Financial openness: kaopeni,t measured by the Chinn-Ito index. - Trade openness: tri,t = exports + imports over GDP. - Controls Z include real GDP level and CPI inflation. - Baseline uses L = 15 years fixed window; robustness check uses 20 years. - Weighted averages of kaopen and tr over the window use weights that assign greater weight to periods close to the beginning of the window. - Empirical finding (Annex Table 3.3.1 summary): - For house price synchronicity (15-year and 20-year windows) among 19 advanced economies: - Chinn-Ito Index: positive and statistically significant coefficients (15 years: 0.06691**, 20 years: 0.06220***). - Exports plus Imports (over GDP): positive and statistically significant for house prices (15 years: 0.00911**, 20 years: 0.01096***). - Log of Output: coefficients (15 years: 0.22121; 20 years: 0.27416**). - Inflation: coefficients (15 years: 0.02439; 20 years: 0.00052). - Observations: 15 years = 1,861; 20 years = 1,645. - R^2: 15 years = 0.38823; 20 years = 0.47414. - Number of Countries: 19 (both windows). - For equity price synchronicity (15-year and 20-year windows): - Chinn-Ito Index: positive and significant (15 years: 0.13516**, 20 years: 0.12603***). - Exports plus Imports: not significant or negative at 20 years (20 years: −0.00715*). - Log of Output: positive and significant (15 years: 0.80820*, 20 years: 0.86895***). - Inflation: 20 years significant (0.01830**). - Observations: 15 years = 1,296; 20 years = 1,140. - R^2: 15 years = 0.71709; 20 years = 0.88296. - Number of Countries: 12 (both windows). - Note: Robust standard errors in parentheses. ***p < 0.01; **p < 0.05; *p < 0.1. - Interpretation: - Increases in financial and trade openness over time partly account for the rise in exposure to the global factor in house prices among the observed advanced economies. - Financial openness also explains the increase in equity comovement. - House price synchronization is part of broader asset price synchronization induced by financial openness. ### Growth at Risk - Data partitioning and data reduction: - Financial data reduced into three ad hoc groups: price of risk, leverage, and external factors. - Linear discriminant analysis (LDA) used to project data onto a lower-dimensional space while maximizing discrimination between classes defined by a dummy variable: - Dummy = 1 if future GDP growth at a one-year horizon is below the 20th percentile of historical outcomes; = 0 otherwise. - LDA loadings maximize contribution to discriminating between low GDP growth and normal growth periods. - LDA differs from PCA by ensuring discrimination across classes; PCA only aggregates common trends. - Quantile regression framework for linking financial variables, house price synchronicity, and GDP growth: - y_{t+h,q} = α^q_h p_t + β^q_h Agg_t + γ^q_h y_t + φ^q_h f_t + θ^q_h HP_t + ε_{t,q}^h. (A3.3.8) - p: aggregated price of risk (asset prices and risk spreads). - Agg: credit aggregates (leverage). - f: global and foreign variables (commodity prices, exchange rates, global risk sentiment). - HP: house price synchronicity. - Estimation across quantiles spans the full GDP growth distribution at different horizons (near, medium, long term). - Augmented specification to capture amplification effects: - y_{t+h,q} = α^q_h p_t + β^q_h Agg_t + γ^q_h y_t + φ^q_h f_t + θ^q_h HP_t + ς^q_h HP_t × Agg_t (or p_t) + ε_{t,q}^h. (A3.3.9) - ς^q_h captures amplification of house price synchronicity effects when leverage increases or financial conditions tighten. - Purpose: disentangle contribution of house price synchronicity from price of risk, credit aggregates, and external shocks in forecasting risks to GDP growth; inform surveillance and policy design across frequencies. ### Methodology for Boxes - Box 3.1 (global investors and US house price dispersion): - Main regression: - HP_D_t = β0 + β1 FC_t + β2 X_t + β3 γ_t + β4 GFC_t + ε_t. (A3.3.10) - HP_D_t: ratio of the 90th percentile to the 10th percentile of house prices in the 40 largest US cities. - FC_t: unweighted real US$ average of house prices in non-US destinations for global investors. - X_t: domestic controls (unemployment rate, VIX, effective federal funds rate, 30-year mortgage rate, mortgage-backed security holdings of large domestically chartered commercial banks excluding government-guaranteed MBS). - γ_t: time trend. - GFC_t: global financial crisis dummy (equals 1 during 2008 and 2009). - Four specifications: - (1) HP_D_t on FC_t and time trend. - (2) adds control variables. - (3) and (4) use first differences to eliminate common trends; (4) also includes GFC_t. - Empirical results (Annex Table 3.3.2 summary): - Foreign City House Price Index (FC_t): coefficients - (1) Levels: 0.600*** (p-value 0.000) - (2) Levels with controls: 0.339** (p-value 0.031) - (3) Differences: 0.019** (p-value 0.039) - (4) Differences with GFC: 0.019** (p-value 0.039) - VIX Index: - (2) −0.002** (p-value 0.041) - (3) 0.000** (p-value 0.021) - (4) 0.000** (p-value 0.022) - Federal Funds Rate (effective): - (2) 0.002 (p-value 0.952) - (3) 0.007*** (p-value 0.009) - (4) 0.007*** (p-value 0.009) - Mortgage Interest Rates: - (2) −0.011 (p-value 0.654) - (3) 0.005** (p-value 0.017) - (4) 0.005** (p-value 0.017) - Bank MBS Holdings: - (1) 0.434* (p-value 0.053) - (2) −0.002 (p-value 0.883) - (3) −0.002 (p-value 0.881) - Time Trend: - (1) and (2): 0.025*** and 0.024*** (p-values 0.000) - Observations: (1) 256; (2) 250; (3) 250; (4) 250. - Adjusted R^2: (1) 0.9040; (2) 0.9110; (3) 0.0600; (4) 0.083. - Notes: - Monthly data from 1996:Q4 to 2017:Q9. - Robust (Newey-West, 12 lags) p-values in parentheses. - Dependent variables are lagged one month. - FC, GFC, MBS, and other definitions as specified in text. - ***p < 0.01; **p < 0.05; *p < 0.1. - Box 3.4 (macroprudential tools and house price synchronicity): - Panel regression: - HPS_{i,t} = ρ BCS_{i,t−1} + β MPP_{i,t−1} + γ X_{i,t−1} + α_i + ε_{i,t}. (A3.3.11) - HPS: house price cycle synchronicity (instantaneous quasi correlation) with the global cycle. - BCS: business cycle synchronicity with the rest of the world. - MPP: macroprudential tool or macroprudential group index (loan-to-value limits, debt-to-income limits, seller/buyer stamp duty taxes, loan-targeted, supply-side, demand-side indices). - X: controls (including a global factor, financial integration with the world, institutional characteristics). - α_i: country fixed effects. - Sample: 41 countries, 1990:Q2–2016:Q4. - Purpose: gauge effectiveness of macroprudential tools in reducing house price synchronicity. *Source: Annex 3.3. Technical Annex, Global Financial Stability Report: A BuMPY ROAd AhEAd, International Monetary Fund, April 2018.* ### CHAPTER 3 hOuSE PRICE SYNChRONIzATION: WhAT ROLE FOR FINANCIAL FACTORS? ### CHAPTER 3 hOuSE PRICE SYNChRONIzATION: WhAT ROLE FOR FINANCIAL FACTORS? ### References cited - Cetorelli, Nicola, and Linda S. Goldberg. 2011. “Global Banks and International Shock Transmission: Evidence from the Crisis.” IMF Economic Review 59 (1): 41–76. - Cheng, Ing-Haw, Sahil Raina, and Wei Xiong. 2014. “Wall Street and the Housing Bubble.” American Economic Review 104 (9): 2797–829. - Chinco, Alex, and Christopher Mayer. 2015. “Misinformed Speculators and Mispricing in the Housing Market.” Review of Financial Studies 29 (2): 486–522. - Chinn, Menzie D., and Hiro Ito. 2006. “What Matters for Financial Development? Capital Controls, Institutions, and Interactions.” Journal of Development Economics 81 (1): 163–92. - Christiano, Lawrence J., and Terry J. Fitzgerald. 2003. “The Band Pass Filter.” International Economic Review 44 (2): 435–65. - Claessens, Stijn, M. Ayhan Kose, and Marco E. Terrones. 2008. “What Happens during Recessions, Crunches and Busts?” IMF Working Paper 08/274, International Monetary Fund, Washington, DC. - Claessens, Stijn, M. Ayhan Kose, and Marco E. Terrones. 2011. “How Do Business and Financial Cycles Inter-act?” IMF Working Paper 11/88, International Monetary Fund, Washington, DC. - Collier, Paul, and Anthony J. Venables. 2012. “Land Deals in Africa: Pioneers and Speculators.” Journal of Globalization and Development 3 (1): 1–22. - Cushman & Wakefield. 2017. “Winning in Growth Cities.” December 19. https:// cushmanwakefield .turtl .co/ story/ 59d48cf5d1c76a6148cabb4f. - Dell’Ariccia, Giovanni, Deniz Igan, and Luc Laeven. 2012. “Credit Booms and Lending Standards: Evidence from the Subprime Mortgage Market.” Journal of Money, Credit and Banking 44 (2–3): 367–84. - Del Negro, Marco, and Christopher Otrok. 2007. “99 Luftballons: Monetary Policy and the Housing Price Boom across U.S. States.” Journal of Monetary Economics 54 (7): 1962–85. - Demyanyk, Yuliya, and Otto Van Hemert. 2009. “Understanding the Subprime Mortgage Crisis.” Review of Financial Studies 24 (6): 1848–80. - Diebold, Francis X., and Kamil Yilmaz. 2014. “On the Network Topology of Variance Decompositions: Measuring the Connectedness of Financial Firms.” Journal of Econometrics 182 (1): 119–34. - Dornbusch, Rudiger, Yung Chul Park, and Stijn Claessens. 2000. “Contagion: How It Spreads and How It Can Be Stopped.” World Bank Research Observer 15 (2): 177–97. - Doyle, Brian, and Jon Faust. 2005. “Breaks in the Variability and Comovement of G-7 Economic Growth.” Review of Economics and Statistics 87: 721–40. - Duda, Richard O., Peter E. Hart, and David G. Stork. 2001. Pattern Classification. New York: Wiley-Interscience. - Duval, Romain, Nan Li, Richa Saraf, and Dulani Seneviratne. 2016. “Value-Added Trade and Business Cycle Synchroniza- tion.” Journal of International Economics 99 (March): 251–62. - Fama, Eugene, and Robert Bliss. 1987. “The Information in Long-Maturity Forward Rates.” American Economic Review 77 (4): 680–92. - Fama, Eugene, and Kenneth French. 1988. “Dividend Yields and Expected Stock Returns.” Journal of Financial Economics 22 (1): 3–25. - Forbes, Kristin. 2012. “The ‘Big C’: Identifying Contagion.” NBER Working Paper 18465, National Bureau of Economic Research, Cambridge, MA. - Fruchterman, Thomas M. J., and Edward M. Reingold. 1991. “Graph Drawing by Force‐Directed Placement.” Software: Practice and Experience 21 (11): 1129–64. - Glaeser, Edward L., Giacomo A. M. Ponzetto, and Andrei Shleifer. 2016. “Securing Property Rights.” NBER Working Paper 22701, National Bureau of Economic Research, Cambridge, MA. - Glancy, David. 2017. “Housing Bust, Bank Lending, and Employment: Evidence from Multimarket Banks.” Finance and Economics Discussion Series 2017–118, Federal Reserve Board, Washington, DC. - Goldstein, Morris. 1998. “The Asian Financial Crisis.” Policy Brief 98–01, Peterson Institute for International Economics, Washington, DC. - Goodhart, Charles, and Philip Turner. 2014. “Rate Rise Pattern Is Different This Time.” Financial Times, April 2. https:// www .ft .com/ content/ 2e28e3fc -b984 -11e3 -b74f -00144feabdc0. - Gruss, Bertrand. 2014. “After the Boom—Commodity Prices and Economic Growth in Latin America and the Caribbean.” IMF Working Paper 14/154, International Monetary Fund, Washington, DC. - Hansen, Lars, and Robert Hodrick. 1980. “Forward Exchange Rates as Optimal Predictors of Future Spot Rates: An Econometric Analysis.” Journal of Political Economy 88 (5): 829–53. - Haughwout, Andrew, Donghoon Lee, Joseph Tracy, and Wilbert van der Klaauw. 2011. “Real Estate Investors, the Leverage Cycle, and the Housing Market Crisis.” Federal Reserve Bank of New York Staff Report 514, New York. - Hekwolter of Hekhuis, Melanie, Rob Nijskens, and Willem Heeringa. 2017. “The Housing Market in Major Dutch Cities.” DNB Occasional Studies 1501, Netherlands Central Bank, Research Department. - Himmelberg, Charles, Christopher Mayer, and Todd Sinai. 2005. “Assessing High House Prices: Bubbles, Fundamentals and Misperceptions.” Journal of Economic Perspectives 19 (4): 67–92. - Hirata, Hideaki, M. Ayhan Kose, Christopher Otrok, and Marco E. Terrones. 2012. “Global House Price Fluctua- tions: Synchronization and Determinants.” NBER Work- - Hodrick, Robert J., and Edward Prescott. 1997. “Postwar U.S. Business Cycles: An Empirical Investigation.” Journal of Money, Credit, and Banking 29 (1): 1–16. - International Monetary Fund (IMF). 2013. “Key Aspects of Macroprudential Policy.” IMF Policy Paper, Washington, DC. - International Monetary Fund (IMF). 2018a. “Brighter Prospects, Optimistic Markets, Challenges Ahead.” World Economic Outlook Update, January, Washington, DC. - International Monetary Fund (IMF). 2018b. “People’s Republic of China––Hong Kong Special Administrative Region: Selected Issues.” IMF Country Report 18/17, International Monetary Fund, Washington, DC. - Ilzetzki, Ethan, Carmen M. Reinhart, and Kenneth S. Rogoff. 2017. “Exchange Rate Arrangements Entering the 21st Century: Which Anchor Will Hold?” NBER Working Paper 23134, National Bureau of Economic Research, Cambridge, MA. - Izenman, Alan Julian. 2009. Modern Multivariate Statistical Techniques: Regression, Classification, and Manifold Learning. New York: Springer. - Jara, Alejandro, and Nestor Romero. 2016. “International Syn- chronicity of Housing Prices.” Revista de Análisis Económico 31 (2): 115–34. - Jones, Dustin, and Douglas Weill. 2017. 2017 Institutional Real Estate Allocations Monitor. Ithaca, NY: Cornell University’s Baker Program in Real Estate and Hodes Weill & Associates, LP. - Jordà, Òscar, Moritz Schularick, and Alan. M. Taylor. 2017. “Macrofinancial History and the New Business Cycle Facts.” In NBER Macroeconomics Annual 2016, volume 31, edited by Martin Eichenbaum and Jonathan A. Parker. Chicago: University of Chicago Press. - Jordà, Òscar, Moritz Schularick, and Alan. M. Taylor. 2017. “Global Financial Cycles and Risk Premiums.” IMF Working Paper, International Monetary Fund, Washington, DC. - Jordà, Òsca, Katharina Knoll, Dmitry Kuvshinov, Moritz Schularick, and Alan Taylor. 2017. “The Rate of Return on Everything, 1870–2015.” NBER Working Paper 24112, National Bureau of Economic Research, Cambridge, MA. - Kalemli-Özcan, Sebnem, Elias Papaioannou, and Fabrizio Perri. 2013. “Global Banks and Crisis Transmission.” Journal of International Economics 89 (2): 495–510. - Kalemli-Özcan, Sebnem, Elias Papaioannou, and José-Luis Peydró. 2013. “Financial Regulation, Financial Globalization, and the Synchronization of Economic Activity.” Journal of Finance 68 (3): 1179–228. - Kaminsky, Graciela L., and Carmen M. Reinhart. 2000. “On Crises, Contagion, and Confusion.” Journal of International Economics 51 (1): 145–68. - Kodres, Laura E., and Matthew Pritsker. 2002. “A Rational Expectations Model of Financial Contagion.” Journal of Finance 57 (2): 769–99. - Koenker, Roger. 2004. “Quantile Regression for Longitudinal Data.” Journal of Multivariate Analysis 91 (1): 74–89. - Koenker, Roger. 2005. Quantile Regression. New York: Cambridge University Press. - Komunjer, Ivana. 2013. “Quantile Prediction.” In Handbook of Economic Forecasting, Volume 2, Part B. Amsterdam: Else- vier, 961–94. - Koop, Gary, and Dimitris Korobilis. 2013. “Large Time-Varying Parameter VARs.” Journal of Econometrics 177 (2): 185–98. - Kose, M. Ayhan, Christopher Otrok, and Eswar Prasad. 2012. “Global Business Cycles: Convergence or Decoupling?” International Economic Review 53 (2): 511–38. - Kose, M. Ayhan, Eswar S. Prasad, and Marco E. Terrones. 2003. “Financial Integration and Macroeconomic Volatility.” IMF Staff Papers 50: 119–142. - Krippner, L. 2013. “Measuring the Stance of Monetary Policy in Zero Lower Bound Environments.” Economics Letters 118 (1): 135–38. - Kuttner, Kenneth N., and Ilhyock Shim. 2016. “Can Non-Interest Rate Policies Stabilize Housing Markets? Evidence from a Panel of 57 Economies.” Journal of Financial Stability 26 (October): 31–44. - Landier, Augustin, David Sraer, and David Thesmar. 2017. “Banking Integration and House Price Co-movement.” Journal of Financial Economics 125 (1): 1–25. - Leamer, Edward E. 2015. “Housing Really Is the Business Cycle: What Survives the Lessons of 2008–09?” Journal of Money, Credit and Banking 47 (S1): 43–50. - Mian, Atif, Kamalesh Rao, and Amir Sufi. 2013. “Household Balance Sheets, Consumption, and the Economic Slump.” Quarterly Journal of Economics 128 (4): 1687–726. - Mian, Atif, and Amir Sufi. 2009. “The Consequences of Mortgage Credit Expansion: Evidence from the US Mortgage Default Crisis.” Quarterly Journal of Economics 124 (4): 1449–96. - Miyajima, Ken, Madhusudan S. Mohanty, and Tracy Chan. 2015. “Emerging Market Local Currency Bonds: Diversification and Stability.” Emerging Markets Review 22 (March): 126–39. - Morgan, Donald, Bertrand Rime, and Philip Strahan. 2004. “Bank Integration and State Business Cycles.” Quarterly Journal of Economics 119 (4): 1555–85. - Obstfeld, Maurice. 2015. “Trilemmas and Trade-offs: Living with Financial Globalization.” BIS Working Paper 480, Bank for International Settlements, Basel. - Obstfeld, Maurice, Jonathan D. Ostry, and Mahvash S. Qureshi. 2017. “A Tie That Binds: Revisiting the Trilemma in Emerging Market Economies.” IMF Working Paper 17/130, International Monetary Fund, Washington, DC. - Perri, Fabrizio, and Vincenzo Quadrini. 2011. “International Recessions.” NBER Working Paper 17201, National Bureau of Economic Research, Cambridge, MA. - Piketty, Thomas. 2014. Capital in the Twenty-First Century. Cambridge, MA: Harvard University Press. - Poterba, James M. 1984. “Tax Subsidies to Owner-Occupied Housing: An Asset-Market Approach.” Quarterly Journal of Economics 99 (4): 729–52. - Poterba, James M., David N. Weil, and Robert Shiller. 1991. “House Price Dynamics: The Role of Tax Policy and Demography.” Brookings Papers on Economic Activity 2: 143–203. - Reinhart, Carmen M., and Kenneth S. Rogoff. 2008. “This Time Is Different: A Panoramic View of Eight Centuries of Financial Crises.” NBER Working Paper 13882, National Bureau of Economic Research, Cambridge, MA. - Rey, Helene. 2015. “Dilemma Not Trilemma: The Global Financial Cycle and Monetary Policy Independence.” NBER Working Paper 21162, National Bureau of Economic Research, Cambridge, MA. - Sá, Filipa. 2016. “The Effect of Foreign Investors on Local Housing Markets: Evidence from the UK.” CEPR Discussion Paper 11658, Centre for Economic Policy Research, London. - Sá, Filipa, Tomasz Wieladek, and Pascal Towbin. 2014. “Capital Inflows, Financial Structure, and Housing Booms.” Journal of the European Economic Association 12 (2): 522–46. - Shiller, Robert C. 2015. Irrational Exuberance: Revised and Expanded, third edition. Princeton, NJ: Princeton University Press. - US Department of Housing and Urban Development. 2000. “Summary: The US Cost of Home Ownership.” In US Housing Market Conditions, 2nd quarter. Office of Policy Development and Research, US Department of Housing and Urban Development, Washington, DC. https:// www .huduser .gov/ periodicals/ ushmc/ summer2000/ summary -2 .html. - Vandenbussche, Jerome, Ursula Vogel, and Enrica Detragiache. 2015. “Macroprudential Policies and Housing Prices: A New Database and Empirical Evidence for Central, Eastern, and Southeastern Europe.” Journal of Money, Credit and Banking 47 (S1): 343–77. - Wu, Jing Cynthia, and Fan Dora Xia. 2016. “Measuring the Macroeconomic Impact of Monetary Policy at the Zero Lower Bound.” Journal of Money, Credit and Banking 48 (2–3): 253–91. - Zillow Research. 2017. “Are Home Buyers from China Competing with Americans?” Accessed on February 1, 2018 from https://www.zillow.com/research/chinese-home-buyers -competition-14915/. ### Publication highlights and listings (as presented) - To order, visit bookstore.imf.org/gf418 - Highlights from IMF Publications - Unleashing Growth and Strengthening Resilience in the Caribbean $25. ©2017. Paperback ISBN 978-1-48431-519-4. 378pp. - Digital Revolutions in Public Finance $25. ©2017. Paperback ISBN 978-1-48431-522-4. 362pp.. - Fiscal Policies and Gender Equality $25. ©2017. Paperback ISBN 978-1-51359-036-3. 250pp. - Fiscal Politics $40. ©2017. Paperback ISBN 978-1-47554-790-0. 548pp. - Financial Integration in Latin America: A New Strategy for a New Normal $25. ©2017. Paperback ISBN 978-1-51352-024-7. 177pp. - Shifting Commodity Markets in a Globalized World $25. ©2017. Paperback ISBN 978-1-48431-032-8. 112pp. *Global Financial Stability Report, April 2018* --- _Source: https://www.imf.org/-/media/files/publications/gfsr/2018/april/ch1/doc/text.pdf_