## c3 ## Source details **Canonical URL:** [c3](https://www.imf.org/-/media/files/publications/gfsr/2018/april/chapter-3/doc/c3.pdf) ## Other formats - [Markdown version](/-/media/files/publications/gfsr/2018/april/chapter-3/doc/c3.pdf.md) - [Structured JSON version](/-/media/files/publications/gfsr/2018/april/chapter-3/doc/c3.pdf.json) --- ### Overview and context - Rising house prices have been a feature of the economic recovery in many countries since the global financial crisis. - During 2017, there was a pickup in growth in 120 economies, accounting for three-quarters of world GDP. - Recent increases in house prices occurred in an environment of easy financial conditions in major advanced economies characterized by low policy rates, compressed spreads, and low volatility that spread globally. - House price gains have been widespread, with simultaneous growth in many countries and cities in advanced and emerging market economies paralleling the coordinated run-up seen before the crisis. ### Primary drivers of house price synchronization (mechanisms and channels) - Comovement in economic fundamentals: - Synchronous supply and demand factors: construction costs, land acquisition, demographics, tax policy, depreciation and maintenance. - Slower-moving factors (construction costs, taxes, demographics) influence long-horizon synchronization; faster-moving fundamentals (rent, income, inflation) drive shorter-term comovement. - Business cycle comovement, trade, and financial links contribute to coincident recessions and housing downturns. - Global financial channels: - Changes in global financial conditions transmit via capital flows, affecting credit availability and mortgage rates even if flows do not go directly into housing. - An increase in global demand for safe assets can compress sovereign bond rates and hold down mortgage rates across many countries. - Portfolio channels and investor behavior: - Common lenders/investors create interdependence—global institutions may pull back mortgage lending across countries or liquidate leveraged housing investments across markets. - Growing institutional investor participation: trends in REIT market capitalization (index, 2005:Q1 = 100) and weighted average target allocation to real estate (percent) suggest increased institutional presence. - Expectations and risk premia: - Expected return to investing in housing varies over time and is predictable in the long term; variations in the risk premium influence housing demand. - Exchange rate flexibility: - Greater exchange rate flexibility may dampen the impact of global financial conditions by increasing the bite of monetary policies. ### Key empirical findings - Synchronization trends and scope: - Synchronization in house prices across countries and major cities has increased over past decades in advanced and emerging market economies. - Short-term comovement in house prices sharply increases around global recessions in advanced economies; spikes are much larger among major cities than at the country level. - Between 1991 and 2016, synchronicity is lower among major cities in advanced economies than among their countries but has gradually moved closer to country-level synchronicity. - Global factor and exposures: - The share of variance in house price growth explained by an estimated global factor increases from about 10 percent to 30 percent over 1971 to 2016 (dynamic factor model; window = 15 years). - Countries and cities differ in exposure; larger contribution observed in Europe. Advanced economies are more exposed to global factors than emerging market economies. - Interconnectedness and spillovers: - Many large advanced economies’ housing markets are closely interconnected; many emerging market economies show weaker connectivity. - Cities can be highly interconnected even if their countries are not; financial centers and global-investor-attractive cities (for example, Tokyo, London, Stockholm) lie more centrally in city-level networks. - Average spillovers (share of house price variance in a country accounted for by changes in another single country) increased from 1.4 percent in 1990–2006 to 2.1 percent in 2007–16. - Spillovers increased most for advanced→EME (about 60 percent increase) and then EME→advanced (about 40 percent increase). - Role of bilateral financial links: - Countries with deeper bilateral banking linkages exhibit more synchronization—an effect independent of comovement in output and other fundamentals. - Institutional investor participation: - Institutional investor participation in real estate investments—including residential real estate—appears to have grown in recent years. - Predictive risk signal: - 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. ### Risks from synchronized house price declines - Booms can lead households to excessive risk taking, financial institutions to relax lending standards, and overbuilding; declines can threaten macroeconomic and financial stability. - Declines in house prices can cause falls in consumption and household deleveraging, dragging on growth. - Banks’ exposure to house prices can cause financial difficulties and curtailment of lending, lowering employment. - Localized fire sales create blight and crime costs; foreclosure transfers can impose costs on households and financial institutions. - Synchronization across countries amplifies risks because pullbacks in domestic demand coincide with declines in external demand. - A pullback among global investors could lead to fire sales across asset classes, capital flight, and tighter mortgage market conditions. ### Measurement approach and econometric frameworks - Primary measures: - Cyclical component of real house prices (“house price gap”) and quarterly growth rate in real house prices. - Synch1: negative of the absolute difference in house price gaps; value closer to zero implies greater synchronization. - QCORR: instantaneous quasi correlation of house price gaps; higher value implies simultaneous deviations above or below historical averages. - Econometric frameworks: - Dynamic factor model estimating a common global factor (factor loadings and VAR parameters estimated via two-step procedure of Koop and Korobilis (2013) for 19 advanced economies, 1971:Q2–2016:Q4). - Diebold and Yilmaz (2014) approach to measure interconnectedness via quarterly house price growth correlations and spillovers. - Bilateral panel data approach at country- and city-pair level using quarterly data 1990–2016 for 40 countries and major city pairs; controls for global factors, output comovement, and other fundamentals. - Synchronicity decomposition and measures: - Instantaneous quasi correlation (QCORR), bilateral absolute difference (Synch1), and relative contribution of the global factor (r_ci) are defined and used to capture short-term comovement, long-term trend via γ, and comprehensive long-term synchronization respectively. - Filters and robustness: - House price gaps extracted using band-pass filter of Christiano and Fitzgerald (2003), maximum length 20 years; robustness with Hodrick and Prescott (1997) filter with lambda = 400,000. - Rolling window analyses use a 15-year window (robustness checks with 20 years). ### Quantitative indicators and statistics (preserve original numeric values) - Dynamic factor model sample period: 1971:Q2–2016:Q4 (19 advanced economies). - City-level VAR sample period: 2004:Q1 to 2017:Q2. - Country-level sample comparisons: 1990–2006 and 2007–16. - Rolling estimation window used: 15-year window (some robustness with 20 years). - Share of variance explained by global factor: increases from about 10 percent to 30 percent over 1971 to 2016. - Average spillovers: 1.4 percent in 1990–2006; 2.1 percent in 2007–16. - Standard deviation for business cycle synchronization used in a specification: 0.0124. - Standard deviation for bilateral bank integration used in a specification: 1.040. - Standard deviation of the country-level dependent variable (Figure 3.15 analysis): approximately 0.85. - Standard deviation of the city-level dependent variable (Figure 3.15 analysis): approximately 0.97. - Quantile regression coefficients (Figure 3.16): range from –0.7 to 0.0. - When leverage is high, the negative impact of house price synchronicity on downside growth risk is about twice as large; magnitude of relationship between house price synchronization and future growth is about two-thirds that of financial conditions (price of risk). - Macroprudential analysis sample: 41 countries spanning 1990:Q2–2016:Q4. - Total number of demand-side events in macroprudential analysis: 47. - Annex sample observation counts (selected): Observations by specification: 65,450; 65,343; 49,384; 43,871; 46,708; 47,353. - Annex R^2 by specification (selected): 0.353; 0.498; 0.386; 0.361; 0.360; 0.360. - Annex findings (coefficients, selected): - Business Cycle Synchronization of ij: coefficients include 0.766***, 0.675**, 0.733***, 0.657**, 0.658**, 0.746***, 0.725***, 0.725***, 0.675**, 0.706** (standard errors reported, e.g., (0.254), (0.293), (0.243)). - Bilateral Bank Integration of ij: coefficients include 0.006*, 0.007**, 0.012, 0.009*, 0.007**, 0.007*, 0.007**, 0.004 (standard errors reported, e.g., (0.003), (0.003), (0.007)). - Global Factor (global liquidity): coefficients near –0.001 to 0.001 (standard errors (0.001)). - GFC Dummy = 0.048*** (0.011). Post-GFC Dummy = 0.042*** (0.009). - In quasi correlation specifications, Global Factor coefficients: 0.016**, 0.016**, 0.020**, 0.019***, 0.019**, 0.018**, 0.022* (standard errors e.g., (0.006), (0.008), (0.008)). - Annex Table 3.3.1 regression coefficients (Chinn-Ito Index and other controls): - Chinn-Ito Index: - House Price Synchronicity (15 years): 0.06691** (0.02387) - House Price Synchronicity (20 years): 0.06220*** (0.01585) - Equity Price Synchronicity (15 years): 0.13516** (0.04697) - Equity Price Synchronicity (20 years): 0.12603*** (0.02585) - Exports plus Imports (over GDP): - House Price Synchronicity (15 years): 0.00911** (0.00394) - House Price Synchronicity (20 years): 0.01096*** (0.00351) - Observations: - House Price Synchronicity (15 years) = 1,861; (20 years) = 1,645. - Equity Price Synchronicity (15 years) = 1,296; (20 years) = 1,140. - R^2: - House Price Synchronicity (15 years) = 0.38823; (20 years) = 0.47414. - Equity Price Synchronicity (15 years) = 0.71709; (20 years) = 0.88296. ### Macroprudential policies and effectiveness - Main findings: - Tighter macroprudential tools targeting bank capital and credit conditions are associated with lower house price synchronicity. - Demand-side macroprudential measures (for example, loan-to-value limits) typically reduce house price growth; decline is larger and more persistent in countries with low house price synchronicity. - Capital-based measures (including countercyclical capital buffers) show the most negative relationship with synchronicity. - Loan-targeted and supply-side (loans) tools (including limits on foreign currency loans) are associated with lessening correlations with global house price cycle. - Fiscal-based measures (ad valorem and buyers’ stamp duty taxes) may lower synchronicity but less so than loan-targeted measures. - Magnitude: - Relative magnitude of macroprudential effects averages about one-half of the effect of global factors and about one-third of the effect of bilateral financial integration. - Results are qualitatively and quantitatively similar when only periods with credit booms are considered (relationships slightly less significant). - Interpretation: - Macroprudential policies can reduce synchronization by constraining local financial intermediaries and domestic demand; high synchronicity does not render such policies ineffective. - Definitions and classification: - Demand side: limits to debt-service-to-income and loan-to-value (LTV) ratios. - Supply side (loans): limits on credit growth, loan loss provisions, loan restrictions, limits on foreign currency loans. - Supply side (capital): capital requirements, conservation buffers, leverage ratio, countercyclical capital buffer. - Fiscal-based measures include ad valorem, sellers’ and buyers’ stamp duty, or other taxes. - Empirical specification for macroprudential effect: - Estimated panel regressions use data for 41 countries, 1990:Q2–2016:Q4; regressors lagged one quarter; standardized coefficients reported; statistical significance at the 10 percent confidence level noted for panel 2 results. ### Transmission channels, predictability, and amplification - Financial frictions: - Contagion and sudden capital flow stops can transmit shocks across countries, comparable in magnitude to business cycle synchronization. - Bank retrenchment from abroad after negative shocks can trigger credit crunches elsewhere and lower asset prices. - Global investors and openness: - Demand from global investors searching for yield or safe assets can influence city-level house price dynamics and push up prices in large cities. - Predictability of housing returns is greater in countries with high capital account openness; implication that global investor risk sentiment contributes more to house price dynamics when capital account openness is high. - Chinn-Ito Index is positively associated with house price synchronicity and equity price synchronicity (see quantitative indicators). - Exchange rate flexibility: - Exchange rate flexibility appears to dampen the impact of global liquidity on house price synchronization; flexible exchange rates may give central banks greater influence on financing conditions. - Amplification with leverage: - Quantile regressions and augmented forecasting specification include interaction HP_t × Agg_t (or p_t) to capture amplification; coefficient ς_{q}^{h} represents amplification when leverage increases or financial conditions tighten. - When leverage is high, the negative impact of house price synchronicity on downside growth risk is about twice as large. ### Boxes, case studies, and sectoral notes (selected) - Global investors, US city house price dispersion (Box 3.1, prepared by Anil Ari): - House price dispersion measured for 40 largest US cities by ratio of top to bottom deciles using Zillow ZIP-code level data; US house price dispersion has increased sharply. - Regression analysis finds statistically significant positive relationship between US house price dispersion and house prices in major foreign cities (Beijing, Dublin, Hong Kong SAR, London, Seoul, Shanghai, Singapore, Tokyo, Toronto, Vancouver), controlling for domestic determinants. - Housing as a financial asset (prepared by Alan Xiaochen Feng): - Housing wealth on average accounts for roughly one-half of total national wealth in a typical economy. - Average annual real return on housing assets between 1950 and 2015 in many advanced economies lies between 5 percent and 8 percent. - High current house-price-to-rent ratio strongly predicts low future housing return; predictive power increases with forecasting horizon (horizons 1 to 10 years). - Predictability stronger in countries with high capital account openness. - Globalization of farmland (prepared by Christian Bogmans): - Between 2000 and 2016 commercial investors negotiated more than 2,100 large-scale land acquisitions in 88 countries, cumulative size almost 59 million hectares (roughly equal to 15 percent of remaining global stock of unused and unforested arable land). - Regional deal counts (approximate): Sub-Saharan Africa about 900 deals; East Asia about 600 deals; Latin America about 350 deals. - Only 49 percent of land deals has been cultivated to some extent. - Rent examples: Africa $3–$12 a hectare; European Union €100–€240 a hectare; United States $200 a hectare. ### Policy implications and recommendations (surveillance, macroprudential, fiscal, and resilience) - Surveillance and data: - Monitor synchronization of house prices across countries and major cities in addition to within-country over- or undervaluation. - Increase granularity, timeliness, and coverage of house price data within countries for richer bilateral and multilateral surveillance. - Collect more comprehensive data on participation of global and institutional investors in housing markets. - Macroprudential and fiscal measures: - Macroprudential policies retain ability to influence local house price developments in highly synchronized markets, albeit less than in less synchronized countries. - Macroprudential measures implemented to tame rising vulnerabilities are followed by a decline in that country’s house price synchronization, suggesting synchronization drivers operate through local financial intermediaries. - Fiscal-based policies (ad valorem and buyers’ stamp duty taxes) may lower house price synchronization, but less so than measures such as limits on loan-to-value ratios. - Consider unintended effects on synchronization when evaluating policy trade-offs. - Policymakers aiming to deter foreign buyers face data limitations and potential circumvention; national or international coordination may be required to prevent displacement effects. - Other resilience-enhancing policies: - Exchange rate flexibility appears to damp house price synchronicity. - Policies that deepen domestic real estate markets or consumer financial protections that discourage excessive or predatory lending to households may help reduce synchronization and increase resilience to global financial shocks. _International Monetary Fund | April 2018_ ### Introduction ### c3 - Introduction ### Overview and context - Rising house prices have been a feature of the economic recovery in many countries since the global financial crisis. - House price gains have been widespread and, in some markets, brisk, with simultaneous growth in house prices in many countries and cities located in advanced and emerging market economies paralleling the coordinated run-up seen before the crisis. - During 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 occurred 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. ### Primary drivers of house price synchronization (analysis) - Comovement in economic fundamentals: - Synchronous supply and demand factors (construction costs, land acquisition, demographics, tax policy, depreciation and maintenance). - Slower-moving factors (construction costs, taxes, demographics) tend to influence long-horizon synchronization; faster-moving fundamentals (rent, income, inflation) can drive shorter-term comovement. - Business cycle comovement, trade, and financial links can contribute to the coincidence of recessions and housing downturns. - Global financial channels: - Changes in global financial conditions can transmit via capital flows, affecting credit availability and mortgage rates even if flows do not go directly into housing. - An increase in global demand for safe assets can compress sovereign bond rates and hold down mortgage rates across many countries. - Portfolio channels and investor behavior: - Common lenders or investors create interdependence in housing markets in crisis and normal times—global institutions may pull back mortgage lending across countries or liquidate leveraged housing investments across markets. - Growing participation of institutional investors: trends in REIT market capitalization (index, 2005:Q1 = 100) and the weighted average target allocation to real estate (percent) suggest increased institutional investor presence—figures show rising indices and allocations (figure references in chapter). - Expectations and risk premia: - Expected return to investing in housing varies over time and is predictable in the long term; variations in the risk premium demanded by investors can influence demand for housing. ### Key empirical 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. - 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 fixed and time-varying fundamentals. This contribution is particularly strong in major cities in advanced economies where global financial integration and local supply constraints may be more binding. - Institutional investor participation in real estate investments—including residential real estate—appears to have grown in recent years (referenced figures show rising REIT market capitalization and higher target allocations by institutions). - 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, suggesting comovement in house prices helps predict tail risk of an economic downturn. ### Policy implications and recommendations - Surveillance and data: - Policymakers may wish to monitor synchronization of house prices with respect to other countries in addition to over- or undervaluation within a country. - Increasing the granularity, timeliness, and coverage of data on house prices within countries would provide richer indicators for bilateral and multilateral surveillance. - More comprehensive data on participation of global and institutional investors in housing markets would strengthen surveillance efforts. - Macroprudential and fiscal measures: - Macroprudential policies seem to retain some ability to influence local house price developments in countries with highly synchronized housing markets, albeit to a lesser extent than in less synchronized countries. - Macroprudential policy measures implemented to tame rising vulnerabilities in a country’s financial sector are followed by a decline in that country’s house price synchronization, suggesting drivers of synchronization operate through local financial intermediaries. - Fiscal-based policies, such as ad valorem and buyers’ stamp duty taxes, may also lower house price synchronization, but less so than measures such as limits on loan-to-value ratios. - Policymakers should consider unintended effects on synchronization when evaluating trade-offs of macroprudential and other policies. - Other resilience-enhancing policies: - Exchange rate flexibility appears to play a role in damping house price synchronicity. - Policies that deepen domestic real estate markets or consumer financial protections that discourage excessive or predatory lending to households may also help reduce synchronization and increase resilience to global financial shocks. ### Conceptual framework (summary) - House prices can synchronize across countries and major cities through: - Supply constraints and construction/land costs. - Demand-side factors (demographics, taxes, depreciation, maintenance). - Financial factors (mortgage interest rates, risk premiums, household leverage, expected nominal house price appreciation). - Global financial conditions, portfolio channels, and expectation shifts (diagrammatic representation provided in chapter). - The chapter notes that simultaneous shifts in user cost can occur (for example, through coordinated tax reforms), but does not focus on those issues here. ### Chapter structure and focus - The chapter addresses: - Trends in synchronization of house prices across countries and major cities, including whether synchronization increased recently and before the global financial crisis. - Factors that contribute to or dampen synchronicity, including roles for financial factors and bilateral versus global influences. - Policy relevance: whether policymakers should monitor synchronicity to better understand financial vulnerabilities and risks. - The chapter complements country-level surveillance and detailed valuation analysis rather than replacing them. *Prepared by a staff team consisting of Jane Dokko (team leader), Adrian Alter, Mitsuru Katagiri, Romain Lafarguette, and Dulani Seneviratne, with contributions from Anil Ari, Christian Bogmans, and Alan Xiaochen Feng, under the general guidance of Claudio Raddatz and Dong He.* ### CHAPTER 3 hOuSE PRICE SYNChRONIzATION: WhAT ROLE FOR FINANCIAL FACTORS? ### CHAPTER 3 hOuSE PRICE SYNChRONIzATION: WhAT ROLE FOR FINANCIAL FACTORS? ### Mechanisms linking financial factors to house price synchronization - Global and institutional investors contribute to house price dynamics in select major cities; geographic concentration of investors can make synchronization more apparent among cities than among countries. - Changes in expected capital gains across countries or cities—driven by rational views of fundamentals or by overoptimism, psychological factors, and speculation—can produce synchronized movements in house prices. - Financial integration can expose mortgage markets to global financial conditions and sudden stops in capital flows, enabling global investors to purchase housing directly and allowing local shocks to spread more widely. - Greater exchange rate flexibility may dampen the impact of global financial conditions by increasing the bite of monetary policies. - Institutional characteristics (for example, participation of foreign investors with long horizons) may either amplify or dampen synchronization; countercyclical long-horizon foreign investors could stabilize prices, though evidence is limited. ### Risks from synchronized house price declines - Booming house prices can lead households to excessive risk taking, financial institutions to relax lending standards, and overbuilding; when booms end, declines can threaten macroeconomic and financial stability. - Declines in house prices can cause falls in consumption (housing is often the largest component of household wealth) and household deleveraging, dragging on growth. - Banks’ exposure to house prices can cause financial difficulties and curtailment of lending, which can lower employment. - Localized fire sales create blight and crime costs because distressed homes often sit vacant before sale; costs may be borne by households and financial institutions if foreclosure transfers properties to banks. - Synchronization across countries amplifies risks because pullbacks in domestic demand coincide with declines in external demand; synchronized declines have been associated with periods of financial instability across many countries at once. - A pullback among global investors could lead to fire sales across asset classes, capital flight, and tighter mortgage market conditions. ### Measurement approach and data - Two broad measures are used: the cyclical component of real house prices (“house price gap”) and the quarterly growth rate in real house prices. - Synchronicity measures discussed include: - Synch1: negative of the absolute difference in house price gaps; a value closer to zero implies greater synchronization. - QCORR: instantaneous quasi correlation of house price gaps; a higher value implies simultaneous deviations above or below historical averages. - Econometric frameworks used: - Dynamic factor model estimating a common global factor (factor loadings and VAR parameters estimated via the two-step procedure of Koop and Korobilis (2013) for 19 advanced economies, 1971:Q2–2016:Q4). - Diebold and Yilmaz (2014) approach to measure interconnectedness via quarterly house price growth correlations and spillovers. - Bilateral panel data approach at country- and city-pair level using quarterly data from 1990–2016 for 40 countries and major city pairs; this removes hard-to-observe bilateral characteristics and controls for global factors, output comovement, and other fundamentals. ### Main empirical findings - Synchronicity trends: - Synchronicity in the house price gap has become more synchronized in countries and cities in advanced and emerging market economies. - Between 1991 and 2016, synchronicity is lower among major cities in advanced economies than among their countries, but has gradually moved closer to country-level synchronicity—suggesting factors driving synchronization have become disproportionately important for cities. - Contribution of a global factor: - The share of the variance in house price growth explained by an estimated global factor increases from about 10 percent to 30 percent over the period from 1971 to 2016 (dynamic factor model; window = 15 years). - Countries and cities differ in exposure to the common global factor; a larger contribution is observed in Europe than in other regions. - Advanced economies are more exposed to global factors than emerging market economies; the relative importance of the global factor has increased over time but not uniformly across advanced economies. - Short-term comovement: - Short-term comovement in house prices increases sharply around the time of global recessions in advanced economies (instantaneous quasi correlation of house price gaps shows pronounced spikes around global downturns). - The increase in short-term synchronicity before recessions is much larger between major cities than between countries among advanced economies. - Interconnectedness and spillovers: - Network analysis shows many large advanced economies’ housing markets are closely interconnected; many emerging market economies show weaker connectivity. - Cities can be highly interconnected even if their countries are not; financial centers and attractive global-investor cities (for example, Tokyo, London, Stockholm) lie more centrally in city-level networks. - Average spillovers (the share of house price variance in a country accounted for by changes in another single country) 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). - Role of bilateral financial links: - Countries with deeper financial links, as captured by bilateral banking linkages, exhibit more synchronization—an effect independent of comovement in output and other fundamentals. ### Notable regional and city-level observations - Europe and “Europe and other” group (European countries, South Africa, and Israel) have a larger share of house price variation attributable to the global factor relative to other regions. - At the city level, many financial centers and global-investor-attractive cities are centrally positioned and influential in cross-border house price transmission. - Examples from network maps: Japan is peripheral at the country level but Tokyo is centrally located at the city level, closer to cities such as London and Stockholm. *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? ### Key empirical findings - House price synchronization has increased over the past three decades, especially among major cities. - City-level VAR analysis covers 2004:Q1 to 2017:Q2 and shows heterogeneous interconnectedness among cities; node size is based on a city’s total outward spillovers and only links above the 66th percentile are considered. - Average country-level housing market spillovers increased between the periods 1990–2006 and 2007–16 (Figure 3.11). - Bilateral links between countries are associated with house price synchronization; business cycle synchronization and bilateral banking integration are both positively related to house price synchronicity (Figure 3.12). - Financial openness, proxied by the Chinn-Ito index, is associated with higher exposure to global factors and higher house price synchronization (Figure 3.13). - Rolling 15-year window estimates show the global factor for house prices has increased along with that for equities (Window = 15 years; percent) (Figure 3.14). - Global liquidity and loose global financial conditions are strongly associated with higher short-term comovement in house prices (Figure 3.15). - The contribution of global financial conditions to house price synchronization is larger for cities than for countries; advanced-economy cities show greater responsiveness to global liquidity. ### Transmission channels and mechanisms - Financial frictions (for example, contagion and sudden capital flow stops) may play an important role in transmitting shocks across countries, similar in magnitude to the role of business cycle synchronization. - Bank retrenchment: when a negative shock affects a country, banks may retrench from activity abroad, triggering credit crunches elsewhere and contributing to lower asset prices. - Global investors: demand from global investors searching for yield or safe assets can influence city-level house price dynamics and exert upward pressure on higher-priced homes in large cities. - Predictability of housing returns is greater in countries with high capital account openness, implying global investor risk sentiment can contribute more to house price dynamics when capital account openness is high. - Exchange rate flexibility appears to dampen the impact of global liquidity on house price synchronization; flexible exchange rates may give central banks greater influence on financing conditions. ### Quantitative indicators and statistics (preserve original numeric values) - City-level VAR sample period: 2004:Q1 to 2017:Q2. - Country-level sample periods compared: 1990–2006 and 2007–16. - Rolling estimation window used in some analyses: 15-year window. - Standard deviation for business cycle synchronization used in a specification: 0.0124. - Standard deviation for bilateral bank integration used in a specification: 1.040. - Standard deviation of the country-level dependent variable (in Figure 3.15 analysis): approximately 0.85. - Standard deviation of the city-level dependent variable (in Figure 3.15 analysis): approximately 0.97. - House price synchronization negatively affects downside growth risks up to one year ahead; quantile regression coefficients shown range from –0.7 to 0.0 (Figure 3.16). - When leverage is high, the negative impact of house price synchronicity on downside growth risk is about twice as large; the magnitude of the relationship between house price synchronization and future growth is about two-thirds that of financial conditions (price of risk). ### Risks to growth - Higher house price synchronization corresponds to increased downside risks to growth at horizons of up to one year. - The negative effect on the lower tail of the growth distribution from house price synchronization persists after controlling for price of risk, leverage, and external conditions. - Interaction with leverage: at short horizons, the negative impact of house price synchronization on downside growth risk is amplified when leverage is high; the negative impact is about twice as large under high leverage. ### Policy implications and recommendations - Monitoring synchronization, in addition to over- or undervaluation, can help policymakers assess trade-offs from greater global links in housing markets. - Increasing the granularity, timeliness, and coverage of house price data would improve bilateral and multilateral surveillance of housing market linkages. - More comprehensive data on the participation of global investors in housing markets would strengthen surveillance efforts. - Macroprudential demand-side measures (for example, loan-to-value limits) typically reduce house price growth; the decline is larger and more persistent in countries with low house price synchronicity: - Demand-based macroprudential measures can reduce a country’s house price synchronization by dampening domestic financial vulnerabilities. - Fiscal-based measures (for example, ad valorem and buyers’ stamp duty taxes) may also lower house price synchronization, but to a lesser extent than demand-based measures. - Policymakers aiming to deter foreign buyers will face challenges due to data limitations and potential circumvention; national or international policy coordination may be required to prevent displacement effects. - Policies that enhance resilience to global financial shocks (for example, exchange rate flexibility, deeper domestic real estate markets, consumer financial protections) may also dampen house price synchronicity and limit fallout from household deleveraging. *Source: CHAPTER 3 hOuSE PRICE SYNChRONIzATION: WhAT ROLE FOR FINANCIAL FACTORS?, 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 investors, house price dispersion, and synchronicity - House price dispersion in the 40 largest US cities is measured by the ratio of the top and bottom deciles of house prices (equivalent to the interpercentile range at log scale). Percentiles are determined by pooling house price estimates from Zillow at the granularity of individual ZIP codes; cities are ranked according to 2015 population estimates from the US Census Bureau. - House price dispersion in the United States has increased sharply over recent decades. - There is substantial comovement between real house prices and house price dispersion (Figure 3.1.1). - Regression analysis shows a statistically significant positive relationship between US house price dispersion and house prices in major cities outside the United States: - Cities included: Beijing, Dublin, Hong Kong SAR, London, Seoul, Shanghai, Singapore, Tokyo, Toronto, and Vancouver. - The coefficient associated with the foreign city index is positive and significant in all specifications considered, including those that control for domestic determinants of housing demand. - Control variables in regressions include the unemployment rate; the Chicago Board Options Exchange Volatility Index (VIX); the effective federal funds rate; 30-year fixed-rate average mortgage interest rates; and the mortgage-backed security holdings of large domestically chartered commercial banks (excluding mortgage-backed securities with government guarantees). Specifications with a time trend and a dummy variable for the global financial crisis are also considered. - Interpretation: - Findings are consistent with common shocks to global investor demand contributing to house price synchronicity across cities and countries. - An alternative interpretation—luxury houses located in areas with tighter supply—cannot account for the positive significant relationship between US house price dispersion and house prices in foreign cities after controlling for domestic determinants (see Annex Table 3.3.2). *This box was prepared by Anil Ari.* ### Housing as a financial asset — time-varying expected returns and openness - Housing is both a residential good and an investment good; in a typical economy housing wealth, on average, accounts for roughly one-half of total national wealth. - Real returns: - In many advanced economies, the average annual real return on housing assets between 1950 and 2015 lies between 5 percent and 8 percent. - These returns are comparable in magnitude to equity investment but with a lower standard deviation (Jordà and others 2017). - Expected returns on housing assets vary over time and are predictable in the medium and long term: - 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). The forecasting equation uses the current price-to-rent ratio to predict future capital gains in housing assets; the plotted y-axis shows R2 from the forecasting equation for horizons 1 year to 10 years. - Predictability is stronger in countries with high capital account openness: - Figure 3.2.2 shows that the R2 from the housing return forecasting equation (predicting housing return nine years ahead using current price-to-rent) is higher with greater capital account openness. - Implication: In an integrated global financial system, global financial conditions and the risk sentiment of global investors can cause domestic house prices to deviate from rental fundamentals and exhibit excess volatility. - Sample note: Analysis based on a sample of 20 advanced economies with long time series for the price-to-rent ratio; relationships may differ when including emerging market economies. *This box was prepared by Alan Xiaochen Feng.* ### The globalization of farmland — deals, rents, and implications - Scale and scope of large-scale land acquisitions: - Between 2000 and 2016 commercial investors negotiated more than 2,100 large-scale land acquisitions in 88 countries, with a cumulative size of almost 59 million hectares. - The cumulative size is roughly equal to 15 percent of the remaining global stock of unused and unforested arable land. - Regional deal counts (approximate): Sub-Saharan Africa about 900 deals; East Asia about 600 deals; Latin America about 350 deals. - Cultivation and convergence: - To date, only 49 percent of the land deals has been cultivated to some extent. - Agricultural land rent historically low in developing economies compared with developed economies: - Rent on land in Africa: $3–$12 a hectare. - Rent in the European Union: €100–€240 a hectare. - Rent in the United States: $200 a hectare. - Implication: Rent across regions could converge, but the convergence process is likely to be very slow given current facts. - Drivers of farmland globalization and synchronization: - In the run-up to the 2007–08 global financial crisis, demand for farmland increased in tandem in Sub-Saharan Africa and East Asia and the Pacific, peaking shortly after the crisis (Figure 3.3.1). - Factors driving synchronization include lower returns on conventional assets after the crisis, lower interest rates, higher biofuel subsidies and agricultural commodity prices, and growing long-term demand for food due to population growth and rising incomes. - Global investors have directed much investment to remote developing economies that previously participated little in global agricultural trade, signaling flows of capital, technology, and agronomic knowledge to those countries and potentially driving convergence of global farmland prices. - Policy implications: - Much acquired land has been held idle for speculative purposes, implying high domestic opportunity costs depending on prior land use. - Land deals in countries with weak land rights for existing users can allow investors to obtain land at lower cost. - Host-country policy responses suggested: - Invest in monitoring capacity to ensure land is leased to responsible investors. - Set strict rules for compensation to displaced land users. *This box was prepared by Christian Bogmans.* ### Macroprudential policies and house price synchronicity - Question addressed: Relationship between macroprudential policies aimed at dampening domestic financial and housing vulnerabilities and house price synchronicity with global cycles. - Key empirical findings: - Tighter macroprudential tools targeting bank capital and credit conditions are associated with lower house price synchronicity. - Before adoption of demand-side macroprudential policies (for example, loan-to-value limits), house prices grow similarly in countries with high or low synchronicity. After adoption, house price growth declines in both groups, with a stronger and more sustained decline in low-synchronicity countries. - Macroprudential tools are associated with a reduction in house price synchronicity (Figure 3.4.1, panel 2). These tools mostly affect local financial intermediaries and domestic demand, implying that drivers of house price comovement operate at least partially through these channels. - Relative effectiveness by tool: - Capital-based measures (including countercyclical capital buffers): relationship with synchronicity is the most negative. - Loan-targeted measures (including loan-to-value limits) and supply-side loan-targeted tools (such as limits on foreign currency loans): associated with lessening correlations with the global house price cycle. - Fiscal-based measures (such as ad valorem and buyers’ stamp duty taxes) that could deter global investors are associated with a decline in synchronicity, but to a lesser extent than other macroprudential policies. - Magnitude of effects: - The relative magnitude of the effect of macroprudential measures averages about one-half of the effect of global factors and about one-third of the effect of bilateral financial integration. - When only periods with credit booms are considered, results are qualitatively and quantitatively similar, though relationships are slightly less significant. - Interpretation: - Macroeconomic and financial integration can amplify house price synchronicity, but domestic macroprudential measures can help reduce synchronization by constraining local financial intermediaries and demand. - A high degree of synchronicity does not render macroprudential policies ineffective; financial factors behind synchronization partly operate through domestic channels that these policies influence. *This box was prepared by Adrian Alter and Dulani Seneviratne.* *International Monetary Fund | April 2018* ### Box 3.4. House Price Gap Synchronicity and Macroprudential Policies ### Box 3.4. House Price Gap Synchronicity and Macroprudential Policies ### Key findings - Macroprudential tools indirectly reduce house price synchronicity with the global cycle. - On average, house prices are affected more by demand-side macroprudential policies in low-synchronicity countries. - Supply-side measures targeting bank capital and loan-specific measures, including loan-to-value limits, seem effective in reducing synchronicity with the global cycle. - The total number of demand-side events in the analysis is 47. - Panel 2 results: solid bars show statistically significant standardized coefficients at the 10 percent confidence level. ### Details on macroprudential tools and classifications - Demand side: limits to debt-service-to-income and loan-to-value (LTV) ratios. - Supply side (loans): limits on credit growth, loan loss provisions, loan restrictions, and limits on foreign currency loans. - Supply side (capital): capital requirements, conservation buffers, the leverage ratio, and the countercyclical capital buffer. - Supply side (general): reserve requirements, liquidity requirements, and limits on foreign exchange positions. - All loan measures include demand side and supply side (loans). - Fiscal-based measures include taxes such as ad valorem, sellers’ and buyers’ stamp duty, or other taxes. - A country is classified in the high-synchronicity group when its average synchronicity (over the sample period) with the global cycle is above the 50th percentile in the sample, and vice versa for low-synchronicity. ### Empirical specification, sample, and estimation - Estimated panel regressions use data for 41 countries spanning the period 1990:Q2–2016:Q4. - Regressions control for business cycle synchronicity, financial integration, and global financial conditions. All regressors are lagged one quarter. - Panel regressions report standardized coefficients; statistical significance indicated at the 10 percent confidence level. - Country-pair econometric baseline (quarterly, 1990–2016 for 40 countries) is: HPsynch_ijt = α_ij + β1 BCS_ijt−1 + β2 FININT_ijt−1 + β3 GLOBAL_t−1 + β4 INST_ijt−1 x GLOBAL_t−1 + β5 OTHER_ijt−1 + tr + ε_ijt - HPsynch_ijt is synchronization of house price gaps between country-pairs i and j at quarter t. - BCS_ij denotes business cycle synchronization between countries i and j. - FININT_ij refers to bilateral financial integration between countries i and j (measured using bilateral locational banking statistics on residency basis; bilateral banking integration is measured as the logarithm of the sum of bilateral claims of country i vis-à-vis country j and bilateral claims of country j vis-à-vis country i as a ratio of the sum of the GDPs of countries i and j). - GLOBAL_t is proxied by changes in global liquidity. - INST_ij denotes dummies equal to 1 if both countries have a high level of an institutional characteristic (for example, economic development level, capital account openness, exchange rate flexibility, or financial development). - OTHER_ij includes other controls (for example, institutional factors). - α_ij are country-pair fixed effects; tr denotes linear and quadratic time trends. - Standard errors are multiway clustered (at country i, country j, and time level, where appropriate). ### Definitions and measurement of synchronicity - Instantaneous quasi correlation (QCORR) in house price gaps: HPsynch_ijt = QCORR_ijt = (HPgap_it − ̄HPgap_i)(HPgap_jt − ̄HPgap_j) / (σ_i_gap σ_j_gap) - HPgap_it and HPgap_jt are house price gaps of countries i and j at quarter t. - ̄HPgap_i and ̄HPgap_j are average house price gaps; σ_i_gap and σ_j_gap are standard deviations of house price gaps. - Alternative synchronization metric (negative absolute difference): HPsynch_ijt = Synch1_ijt = − | HPgap_it − HPgap_jt | - Country-level synchronization measure based on a dynamic factor model: synch_L,i,t = var_L(λ_i,t g_t) + var_L(λ_r,i r_k,t) / var_L(h_i,t) or synch_L,i,t = var_L(λ_i,t g_t) / var_L(h_i,t) - var_L(⋅) is realized variance from period t − L to t. - λ_i,t and λ_r,i are factor loadings to the global factor g_t and regional factor r_k,t (regions: Europe, Asia, and the Americas). - The quarterly growth rate of house prices for country i in period t, h_i,t, consists of the global factor g_t, the regional factor r_k,t, and the country-specific idiosyncratic component c_i,t. - House price gaps are measured by extracting the cyclical component of real house prices using the band-pass filter of Christiano and Fitzgerald (2003), with a maximum length of 20 years to capture medium-term financial cycles. As a robustness check, gaps are also constructed using a Hodrick and Prescott (1997) filter with a lambda of 400,000. ### Robustness checks and sensitivity - Alternative proxies for the global factor include the US financial conditions index (FCI), global FCI, Chicago Board Options Exchange Volatility Index (VIX), and US shadow interest rates. - Interest rate synchronization is found to be a statistically significant driver of house price synchronization on its own when either synchronicity measure is used (either Synch1 or quasi correlation). However, statistical significance of interest rate synchronicity above and beyond other financial factors (such as global liquidity and bilateral banking links) is robust only to a less stringent manner of standard error clustering. - Robustness checks include alternative percentile thresholds for defining high institutional characteristics (for example, 75th or 66th percentiles instead of the 80th percentile) and alternative specifications (including city-level analyses with similar explanatory variables). *Source: IMF staff estimates and methodology as presented in Box 3.4 of the GLOBAL FINANCIAL STABILITY REPORT: A BuMPY ROAd AhEAd (April 2018).* ### CHAPTER 3 hOuSE PRICE SYNChRONIzATION: WhAT ROLE FOR FINANCIAL FACTORS? ### CHAPTER 3 hOuSE PRICE SYNChRONIzATION: WhAT ROLE FOR FINANCIAL FACTORS? ### Annex Table 3.2.1 — Country-Level and Bilateral Linkages (Dependent variable: House Price Gap Synchronization of Country Pair i and j (Synch1)) - Business Cycle Synchronization of ij: coefficients across specifications: 0.766***, 0.675**, 0.733***, 0.657**, 0.658**, 0.746***, 0.725***, 0.725***, 0.675**, 0.706**. - Standard errors reported (selected): (0.254), (0.293), (0.243), (0.254), (0.253), (0.262), (0.261), (0.262), (0.253), (0.337). - Bilateral Bank Integration of ij: coefficients reported in specifications: 0.006*, 0.007**, 0.012, 0.009*, 0.007**, 0.007*, 0.007**, 0.004. - Standard errors (selected): (0.003), (0.003), (0.007), (0.004), (0.003), (0.004), (0.003), (0.005). - Global Factor (global liquidity): coefficients reported: –0.001, –0.001, –0.001, –0.001, –0.001, –0.001, 0.001. - Standard errors (selected): (0.001) across specifications shown. - Interaction: Bilateral Bank Integration × EMEs-EMEs Dummy = –0.016* (0.009). - Interaction: Bilateral Bank Integration × EMEs-AEs Dummy = –0.009 (0.010). - Interaction: Bilateral Bank Integration × High Capital Account Openness with the World = –0.005 (0.003). - Interaction: Bilateral Bank Integration × High Exchange Rate Regime (15 categories; high = more flexible) = –0.005 (0.004). - Interaction: Bilateral Bank Integration × High Exchange Rate Regime (6 categories; high = more flexible) = –0.001 (0.004). - Interaction: Bilateral Bank Integration × High Financial Openness with the World (ij) = –0.019*** (0.004). - GFC Period Dummy Interacted with Bilateral Bank Integration of ij = 0.008** (0.004). - GFC Period Dummy Interacted with Business Cycle Synchronization of ij = –0.080 (0.516). - GFC Period Dummy Interacted with Global Factor = 0.001 (0.001). - Post-GFC Period Dummy Interacted with Business Cycle Synchronization of ij = 0.380 (0.456). - Post-GFC Period Dummy Interacted with Bilateral Bank Integration of ij = 0.007 (0.005). - Post-GFC Period Dummy Interacted with Global Factor = 0.004 (0.003). - GFC Dummy = 0.048*** (0.011). - Post-GFC Dummy = 0.042*** (0.009). - Observations by specification: 65,450; 65,343; 49,384; 49,384; 49,384; 43,871; 46,708; 46,708; 47,353; 49,384. - R^2 by specification: 0.353; 0.498; 0.386; 0.356; 0.356; 0.361; 0.356; 0.356; 0.360; 0.360. - Standard deviation notes: standard deviation for business cycle synchronization is 0.0124 and 1.040 for bilateral bank integration. - Estimation details: multiway clustering mostly used; exceptions noted (regression 10 two-way). Time FE and Country-Pair FE included in many specs; some include country*time FE and quadratic trends. Institutional characteristic dummies included in specifications 5 through 9 (not shown). ### Annex Table 3.2.2 — Country-Level and Global Factors (Dependent variable: House Price Gap Synchronization of Country Pair i and j (Quasi correlation)) - Business Cycle Synchronization of ij: coefficients across specifications: 0.025*, 0.030**, 0.022, 0.026*, 0.026*, 0.025*, 0.026*, 0.026*, 0.026**, 0.042. - Standard errors (selected): (0.013), (0.014), (0.014), (0.013), (0.013), (0.015), (0.014), (0.014), (0.013), (0.033). - Bilateral Bank Integration of ij: coefficients reported: –0.011, 0.012, 0.012, 0.011, 0.022, 0.022, 0.012, –0.016. - Standard errors (selected): (0.033), (0.031), (0.031), (0.036), (0.036), (0.035), (0.032), (0.034). - Global Factor (global liquidity): coefficients: 0.016**, 0.016**, 0.020**, 0.019***, 0.019**, 0.018**, 0.022*. - Standard errors (selected): (0.006), (0.008), (0.008), (0.007), (0.007), (0.007), (0.013). - Global Factor interacted with High Exchange Rate Regime (15 categories; high = more flexible) = –0.023*** (0.008). - Global Factor interacted with EMEs-EMEs Dummy = –0.001 (0.009). - Global Factor interacted with EMEs-AEs Dummy = 0.000 (0.006). - Global Factor interacted with High Capital Account Openness with the World = –0.002 (0.005). - Global Factor interacted with High Exchange Rate Regime (6 categories; high = more flexible) = –0.009 (0.007). - Global Factor interacted with High Financial Openness with the World (ij) = 0.003 (0.006). - GFC Period Dummy Interacted with Global Factor = –0.025* (0.012). - GFC Period Dummy Interacted with Business Cycle Synchronization of ij = –0.032 (0.038). - GFC Period Dummy Interacted with Bilateral Bank Integration of ij = –0.022 (0.035). - Post-GFC Period Dummy Interacted with Global Factor = –0.029 (0.018). - Post-GFC Period Dummy Interacted with Business Cycle Synchronization of ij = –0.039 (0.035). - Post-GFC Period Dummy Interacted with Bilateral Bank Integration of ij = 0.010 (0.033). - GFC Dummy = –0.137** (0.060). - Post-GFC Dummy = –0.044 (0.052). - Observations by specification: 65,450; 65,343; 49,384; 49,384; 49,384; 43,871; 46,708; 46,708; 47,353; 49,384. - R^2 by specification: 0.227; 0.354; 0.251; 0.230; 0.230; 0.233; 0.224; 0.223; 0.241; 0.232. - Estimation details: multiway clustering mostly used; exception regression 10 two-way. Time FE and Country-Pair FE included in many specs; some include country*time FE and quadratic trends. ### Robustness and Additional Findings (from annex text) - Trade integration was included as an additional control and found not to be statistically significant. - Including equity price synchronization as a control leaves results in Annex Tables 3.2.1 and 3.2.2 broadly unchanged; equity price synchronization itself does not consistently have a statistically significant relationship with house price synchronization. - Various standard error clustering alternatives were tested (country-pair; two-way country i and country j; two-way country-pair and time; robust), with significance levels generally improving under less restrictive clustering. - Additional time controls (year fixed effects and linear time trends) produce little change to main conclusions. - Robustness checks included dropping one country pair at a time. - Regressions using a panel of three nonoverlapping seven-year periods (house price and business cycle synchronization captured by bilateral Pearson correlation coefficients) show the interaction term of the global factor and foreign exchange regime remains statistically significant, and the global factor itself remains significant. - Using the Jordà, Schularick, and Taylor (2017) data set (starts in 1870 for 17 advanced economies at annual frequency), the relationship between house price gap synchronicity and business cycle synchronization is positive and statistically significant. Further analysis limited by data availability. ### Measuring Synchronicity: Conceptual Framework and Measures - Framework: two-country decomposition (Doyle and Faust 2005) with country house prices h_i and h_j decomposed into common factor ε_c and idiosyncratic factors ε_i and ε_j, and interconnectedness parameter γ with 0 ≤ γ < 1. - Three synchronization measures defined: - Instantaneous quasi correlation (q_cijt): formula shown (A3.3.3). - Interpretation: suitable for identifying short-term comovement caused by the common shock; sharp movements driven by (1 + γ) ε_c^2 when γ not very large. - Bilateral absolute difference (a_dijt): formula shown (A3.3.4) equal to –|h_it − h_jt| = –(1/(1 + γ)) |ε_it − ε_jt|. - Interpretation: independent of common shock; suitable for assessing long-term trend in synchronicity via changes in γ. An increasing trend in a_dijt implies γ has been increasing. - Relative contribution of the global factor in country i (r_ci): formula shown (A3.3.5) r_ci = σ_c^2 / [ (1 + γ^2)/((1 + γ)^2) σ_i^2 + σ_c^2 ] (as shown). - Interpretation: comprehensive measure for long-term synchronization; increasing r_ci could reflect larger σ_c, smaller σ_i, and/or larger γ, but the three are difficult to separate empirically. ### Estimation of a Dynamic Factor Model (house price dynamics) - Model specification: quarterly house price growth h_i,t = λ_g,i g_t + λ_r,i r_k,t + c_i,t (A3.3.6), where g_t = global factor, r_k,t = regional factor (Europe, Asia, Americas), and c_i,t = country-specific idiosyncratic component. - Factor loadings λ_g,i and λ_r,i estimated; regional factor extracted from residuals after extracting global factor. - Global and regional factors assumed to follow a vector autoregression 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 (synch_L,i,t) to openness and controls (A3.3.7): synch_L,i,t = α_i + δ_t + β1 kaopen_i,t + β2 tr_i,t + γ Z_i,t + ε_i,t. - Financial openness measure: Chinn-Ito index (kaopen_i,t). - Trade openness measure: ratio of gross trade volume to GDP (tr_i,t). - Controls Z_i,t include level of real GDP and CPI inflation. - Window length: baseline 15 years; robustness check with 20 years. - Weighted averages for kaopen_i,t and tr_i,t use weights that assign greater weight to periods close to the beginning of the window. - Findings: Annex Table 3.3.1 shows β1 and β2 are positive and statistically significant among 19 advanced economies observed for longer periods — implying increases in financial and trade openness partly account for the rise in exposure to the global factor. - Financial openness also explains the increase in equity comovement; overall interpretation: house price synchronization is part of asset price synchronization induced by progress in financial openness. ### Growth at Risk — Data Partitioning and Quantile Regressions - Financial data aggregated into three ad hoc groups: price of risk, leverage, and external factors. - Data-reduction technique: linear discriminant analysis (LDA) used to project into lower-dimensional space while maximizing discrimination between classes defined by a dummy whether future GDP growth at a one-year horizon is below the 20th percentile. - LDA chosen because loadings maximize contribution to discriminating low GDP growth periods from normal periods; differs from PCA by using the categorical variable in aggregation. - Quantile regression framework (panel quantile regressions) to capture nonlinear interplay between financial variables, house price synchronicity, and GDP growth: - Model (A3.3.8): 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}. - Variables: p = aggregated price of risk; Agg = aggregated credit aggregates; f = global and foreign variables; HP = house price synchronicity. - Estimation spans different quantiles and horizons (near, medium, long term). - Panel quantile regressions estimated using methodology proposed by Koenker (2004). - Justification for quantile regressions: optimality of conditional quantile estimator as predictor of true future quantile; robustness to outliers and non-normality; flexibility for time-varying parameters and predictor weighting; avoidance of overfitting relative to more complex models. *Source: IMF staff estimates, CHAPTER 3 hOuSE PRICE SYNChRONIzATION: WhAT ROLE FOR FINANCIAL FACTORS?* ### Annex Table 3.3.1. Capital Account Openness and Synchronicity ### Annex Table 3.3.1. Capital Account Openness and Synchronicity ### Regression results: House Price Synchronicity and Equity Price Synchronicity - Dependent variables: House Price Synchronicity (15 years, 20 years) and Equity Price Synchronicity (15 years, 20 years). - Coefficients (robust standard errors in parentheses): - Chinn-Ito Index - House Price Synchronicity (15 years): 0.06691** (0.02387) - House Price Synchronicity (20 years): 0.06220*** (0.01585) - Equity Price Synchronicity (15 years): 0.13516** (0.04697) - Equity Price Synchronicity (20 years): 0.12603*** (0.02585) - Exports plus Imports (over GDP) - House Price Synchronicity (15 years): 0.00911** (0.00394) - House Price Synchronicity (20 years): 0.01096*** (0.00351) - Equity Price Synchronicity (15 years): –0.00160 (0.00416) - Equity Price Synchronicity (20 years): –0.00715* (0.00346) - Log of Output - House Price Synchronicity (15 years): 0.22121 (0.13475) - House Price Synchronicity (20 years): 0.27416** (0.11590) - Equity Price Synchronicity (15 years): 0.80820* (0.43479) - Equity Price Synchronicity (20 years): 0.86895*** (0.21138) - Inflation - House Price Synchronicity (15 years): 0.02439 (0.02069) - House Price Synchronicity (20 years): 0.00052 (0.00643) - Equity Price Synchronicity (15 years): 0.02031 (0.02969) - Equity Price Synchronicity (20 years): 0.01830** (0.00775) - Sample and fit statistics: - Observations: House Price Synchronicity (15 years) = 1,861; (20 years) = 1,645. Equity Price Synchronicity (15 years) = 1,296; (20 years) = 1,140. - R^2: House Price Synchronicity (15 years) = 0.38823; (20 years) = 0.47414. Equity Price Synchronicity (15 years) = 0.71709; (20 years) = 0.88296. - Number of Countries: House Price Synchronicity models use 19 countries; Equity Price Synchronicity models use 12 countries. - Significance notation: - ***p < 0.01; **p < 0.05; *p < 0.1. - Note: "15 years" and "20 years" correspond to the window for variance decomposition. Robust standard errors are in parentheses. ### Interpretation and substantive findings - Capital account openness, as measured by the Chinn-Ito Index, is positively associated with both house price synchronicity and equity price synchronicity across 15-year and 20-year variance decomposition windows, with coefficients statistically significant at conventional levels (see stars for significance). - Trade openness (Exports plus Imports over GDP) is positively associated with house price synchronicity (both 15-year and 20-year windows) and statistically significant, but shows a weakly negative association with equity price synchronicity (20 years) with significance at the 10 percent level. - Larger economies (higher Log of Output) tend to exhibit higher equity price synchronicity (20 years significant at the 1 percent level; 15 years significant at the 10 percent level) and some positive association with house price synchronicity (20 years significant at the 5 percent level). - Inflation shows a statistically significant positive association with equity price synchronicity at the 20-year window (0.01830**). ### Amplification mechanism (equation A3.3.9) - Augmented forecasting specification: - 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}. - Role of ς_{q}^{h}: - The coefficient ς_{q}^{h} represents the amplification effect of the impact of house price synchronicity when leverage increases or when financial conditions tighten. - Purpose: - This specification disentangles the contribution of changes in house price synchronicity from the evolving price of risk, credit aggregates, and external shocks in forecasting risks to GDP growth, thereby providing insight into variables that signal growth tail risks over different time horizons. ### Methodological notes from associated boxes - Box 3.1 (equation A3.3.10): Main specification for analyzing the role of global investors in U.S. city house price dispersion: - HPD_{t} = β_{0} + β_{1} FC_{t} + β_{2} X_{t} + β_{3} γ_{t} + β_{4} GFC_{t} + ε_{t}. - Dependent variable HPD_{t}: ratio of the 90th percentile of house prices to the 10th percentile in the 40 largest US cities by population. - FC_{t}: unweighted real US$ average of house prices in non-US destinations for global investors. - X_{t}: domestic controls including unemployment rate, VIX, effective federal funds rate, 30-year fixed-rate average mortgage interest rates, and mortgage-backed security holdings of large domestically chartered commercial banks (excluding government-guaranteed MBS). - GF C_{t} equals 1 during 2008 and 2009. - Four specifications: (1) HPD_{t} on FC_{t} and time trend; (2) adds controls; (3) and (4) use first differences to remove common trends; (4) includes GFC_{t}. - Box 3.4 (equation A3.3.11): Panel regression assessing macroprudential tools' effectiveness in reducing house price synchronicity (41 countries, 1990:Q2–2016:Q4): - HP S_{i,t} = ρ BCS_{i,t−1} + β MPP_{i,t−1} + γ X_{i,t−1} + α_{i} + ε_{i,t}. - HP S: house price cycle synchronicity (instantaneous quasi correlation) with the global cycle. - BCS: business cycle synchronicity with the rest of the world. - MPP: a macroprudential tool (e.g., limits on loan-to-value ratios or debt-to-income ratios; fiscal-based measures such as sellers’ and buyers’ stamp duty taxes) or a macroprudential group index (loan-targeted, supply-side [capital, general, loans], demand-side tools). - X: controls including a global factor, financial integration with the world, and institutional characteristics. - α_{i}: country fixed effects. *Source: IMF staff estimates.* ### CHAPTER 3 hOuSE PRICE SYNChRONIzATION: WhAT ROLE FOR FINANCIAL FACTORS? ### CHAPTER 3 hOuSE PRICE SYNChRONIzATION: WhAT ROLE FOR FINANCIAL FACTORS? ### Selected references cited in the chapter - 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 Macro-economic 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/. *International Monetary Fund | April 2018* --- _Source: https://www.imf.org/-/media/files/publications/gfsr/2018/april/chapter-3/doc/c3.pdf_