Survey
* Your assessment is very important for improving the workof artificial intelligence, which forms the content of this project
* Your assessment is very important for improving the workof artificial intelligence, which forms the content of this project
Aggregate risks Housing Risks Other Sectoral Risks Macro-prudential chartpack Reserve Bank of New Zealand Notes and data sources in appendix 27-Oct-2015 Appendix Aggregate risks Housing Risks Other Sectoral Risks Appendix Credit and asset prices Credit−to−GDP gaps and credit growth 40 % 30 % 40 One−sided HP with static forecasts with ARIMA forecasts Credit growth (annual) % 30 20 20 10 10 0 0 −10 −20 1988 200 Credit−to−GDP ratio by sector % 200 Household Business Agriculture 150 150 100 100 50 50 −10 1992 1996 2000 2004 50 2008 apc 2012 −20 0 1980 1985 Asset price growth by sector Housing Farm Commercial property 1990 1995 2000 2005 2010 0 2015 apc 50 40 40 30 30 20 20 10 10 0 0 −10 −10 −20 −20 −30 −30 −40 −40 1994 1997 2000 2003 2006 2009 2012 2015 1 Aggregate risks Housing Risks Other Sectoral Risks Appendix Bank leverage and risk-taking Bank capital ratios % Total Assets 20 % RWA 20 Tier 1 (RHS) Tier 1 regulatory min (RHS) Total (RHS) Leverage ratio 16 4.0 Bank profitability % ROA Net interest margin % 30 ROE (RHS) 3.5 16 12 12 25 3.0 20 2.5 2.0 8 8 4 4 0 1998 0 15 1.5 10 1.0 5 0.5 3.5 3.0 % 2001 2004 2007 2010 2013 Fixed term mortgage spreads Average across banks Median of best five % 3.5 0.0 1997 2000 2003 2.0 2.0 1.5 1.5 1.0 1.0 0.5 0.5 0.0 2000 2002 2004 2006 2008 2010 2012 2014 0.0 2012 0 2015 Change in lending standards net % net % 30 Business Household Agriculture 20 ↑ Tightening 20 2.5 2009 30 3.0 2.5 2006 10 10 0 0 −10 −20 −10 2010 2011 2012 2013 2014 2015 −20 2 Aggregate risks Housing Risks Other Sectoral Risks Appendix Funding and liquidity % GDP 100 External debt & current account Total Bank Current account (RHS) 80 % GDP % GDP Credit and deposit growth % GDP 20 25 15 20 25 10 15 15 5 10 10 0 5 5 −5 0 0 −10 −5 Credit Deposits as core funding 20 60 40 20 0 1988 % GDP 100 1992 1996 2000 2004 2008 Bank offshore funding < 90days ≥ 90days 2012 % GDP 100 80 80 60 60 40 40 20 20 0 0 2001 2003 2005 2007 2009 2011 2013 2015 100 2001 2003 2005 2007 2009 2011 2013 2015 Core funding ratio % −5 % 100 Core funding ratio Regulatory min 90 90 80 80 70 70 60 2003 2005 2007 2009 2011 2013 2015 60 3 Aggregate risks Housing Risks Other Sectoral Risks Appendix Financial market conditions 350 Bank funding spreads bp Retail Domestic wholesale bp 350 Offshore wholesale 300 Index 70 300 250 250 200 200 150 150 100 100 50 50 0 0 −50 −50 −100 −100 25 S&P 500 volatility Index 70 1998 % 2001 2004 2007 2010 2013 US corporate bond spreads 60 60 50 50 40 40 30 30 20 20 10 10 2001 2003 2005 2007 2009 2011 2013 2015 % 25 High yield Investment grade 20 20 15 15 10 10 5 5 9 10−year government bond yields % US UK NZ Japan Germany Aus 2007 2009 2011 2013 2015 0 9 8 7 7 6 6 5 5 4 4 3 3 2 2 1 0 2005 % 8 0 1999 1 2002 2005 2008 2011 2014 0 4 Aggregate risks Housing Risks Other Sectoral Risks Appendix Housing market imbalances (1) 35 House price inflation apc NZ AKL apc 35 Rest of NZ 25 15 5 −5 −15 ratio 10 NZ AKL Rest of NZ 9 25 House price−to−income ratios ratio 10 9 8 8 15 7 7 5 6 6 5 5 4 4 −5 1994 1997 2000 2003 2006 7 2009 % 2012 2015 −15 3 1998 Rental yields 2001 2004 2007 2010 2013 3 % NZ AKL 7 6 6 5 5 4 4 3 3 2 2 1994 1997 2000 2003 2006 2009 2012 2015 5 Aggregate risks Housing Risks Other Sectoral Risks Appendix Housing market imbalances (2) Housing purchases number 3000 Investor FHB Mover number New to Market Other 3000 2500 2500 2000 2000 1500 1500 1000 1000 500 500 0 0 Jul−13Oct−13Jan−14Apr−14Jul−14Oct−14Jan−15Apr−15 apc 14 12 number 450 Auckland investor purchases 10+ properties 450 250 250 150 150 50 Jul−13 House price expectations Jan−14 Jul−14 Jan−15 50 apc 14 12 10 8 8 6 6 4 4 2 2 2014 number 2 350 Household forecasts Professional forecasts Actual data 2013 3−4 350 10 0 5−9 2015 2016 0 6 Aggregate risks Housing Risks Other Sectoral Risks Appendix Household balance sheet stretch 180 % 160 Household debt−to−income ratios % 180 Total Owner occupier 160 140 140 60 Debt servicing ratios % 55 50 % 60 Aggregate New buyer − NZ New buyer − Auckland 55 50 45 45 40 40 120 120 35 35 100 100 30 30 25 25 80 80 20 20 15 15 60 60 40 1979 qpc 1984 1989 1994 1999 2004 2009 Gross and net mortgage lending 6 2014 40 qpc 15 Mortgage credit growth Commitments (RHS) Approvals (RHS) 12 5 10 5 1990 $bn 15 12 10 1994 1998 2002 2006 2010 Commitments by borrower type 2014 5 $bn 15 Investor Other Owner Occupier FHB 12 4 9 9 9 6 6 6 3 3 3 0 0 3 2 1 0 2003 2005 2007 2009 2011 2013 2015 14Q4 15Q1 15Q2 15Q3 0 7 Aggregate risks Housing Risks Other Sectoral Risks Appendix Lending standards LVR of new commitments (mean above median) % % 84 Total Investor 82 30 84 LVR>90 % 30 80<LVR≤90 25 25 20 20 15 15 10 10 5 82 80 80 78 78 76 76 5 74 74 0 5.0 Mortgage share of high−LVR % 13Q4 14Q1 14Q2 14Q3 14Q4 15Q1 15Q2 15Q3 % Mortgage share of NBLIs % 5.0 4.5 4.5 4.0 4.0 3.5 3.5 3.0 3.0 2.5 Jul−13 Jan−14 Jul−14 Jan−15 2.5 30 apc 2009 2010 2011 2012 2013 2014 2015 Housing and consumer lending 0 apc Housing lending Consumer lending 30 20 20 10 10 0 0 −10 2006 2008 2010 2012 2014 −10 8 Aggregate risks Housing Risks Other Sectoral Risks Appendix Farm market imbalances 60 Farm price inflation apc apc 40 300 number 300 Farm Dairy 250 40 20 Farm sales number 60 Farm Dairy 250 200 200 150 150 100 100 50 50 20 0 0 −20 −20 −40 −40 1998 2001 2004 2007 2010 1997 2000 Farm price−to−income ratio Index 3.0 2013 0 2003 2006 2009 2012 2015 0 % 900 Farm (index) Dairy (RHS) 800 2.5 700 2.0 600 500 1.5 400 1.0 1997 2000 2003 2006 2009 2012 300 2015 9 Aggregate risks Housing Risks Other Sectoral Risks Appendix Farm balance sheet stretch 30 Rural lending growth apc apc Rural debt−to−income ratio % % 30 400 25 25 350 350 20 20 300 300 15 15 250 250 200 200 150 100 Farm Dairy 10 10 5 5 0 0 150 −5 −5 100 −10 1981 −10 50 1980 1986 1991 1996 2001 2006 1985 Dairy farm cashflow forecast $/kgMS 8 2011 Farm Effective payout 400 Dairy 1990 6 6 5 5 4 4 2004 2007 2010 2005 2010 50 2015 $/kgMS 7 2001 2000 8 Breakeven 7 3 1995 2013 3 2016 10 Aggregate risks Housing Risks Other Sectoral Risks Appendix Commercial property imbalances 50 apc Commercial property price inflation Total Retail Office apc 50 Industrial 40 40 30 30 20 20 Price−to−rent by sector ratio 15 14 13 10 10 0 0 ratio 15 Total Retail Office Industrial 14 13 12 12 11 11 −10 −10 10 10 −20 −20 9 9 −30 1995 −30 8 1994 30 1998 2001 2004 2007 2010 2013 Vacancy rate by sector % % 25 30 Retail Office Industrial 25 20 20 15 15 10 10 5 5 0 1994 1997 2000 2003 2006 2009 2012 0 2015 16 1997 2000 2003 2006 2009 2012 Forecast supply % Retail Office 8 2015 % 16 Industrial 12 12 8 8 4 4 0 0 −4 1995 1999 2003 2007 2011 2015 2019 −4 11 Aggregate risks Housing Risks Other Sectoral Risks Appendix Business balance sheet stretch 25 apc Business lending growth apc 20 25 Property Non−property 20 15 15 10 10 5 5 0 0 −5 −5 −10 −10 −15 2006 2008 2010 2012 2014 −15 160 140 Debt−to−income ratio % % 18 Property Non−property (RHS) 17 16 120 15 100 14 13 80 12 60 40 11 1996 1999 2002 2005 2008 2011 2014 10 12 Indicators for Aggregate Financial Risks Indicator Credit cycle Intuition 1 Credit-to-GDP gap and growth Rapid credit booms are associated with rising systemic risk. This is particularly true if credit is persistently outstripping GDP (as indicated by an elevated creditto-GDP gap). 2 Credit-to-GDP level 3 Asset prices by sector Evidence Credit-to-GDP gap and credit growth are two of the best predictors for financial crises in the early warning literature. Methodology Source Credit growth is an annual percent change. Credit-to-GDP gap is computed in quasi real-time with a one-sided statistical filter (HP) and one-sided HP filter augumented by rolling-window ARIMA forecasts. RBNZ Standard Statistical Return (SSR), Statistics New Zealand. Rapid credit growth could be more concerning if the credit-to-GDP level is already elevated. Further work is required to understand the reasons for elevated credit-to-GDP, and whether it can be attributed to trend factors such as lower neutral interest rates. Ratio of household, business and agricultural lending to total GDP. RBNZ SSR, Statistics New Zealand. Synchronised asset prices across an economy could pose risks to the financial system, partciularly if this occurs alongside rapid credit growth. Early warning literature has Annual percent change in house, found that asset prices have commercial and farm prices. predictive power in addition to credit measures. REINZ, JLL. 4 Indicator Bank capital and leverage ratios Intuition A reduction in the quality and quantity of capital over the cycle would reduce lossabsorbing capacity and indicate increased risk appetite. Falling leverage (unweighted capital) ratios during credit booms could indicate pro-cyclical approaches to risk weighting. 5 Bank profitability Sustained periods of high or low profits require further investigation. Weak profitability could indicate increased competition with risks to lending standards, while high profits could be achieved by taking on riskier assets. Weak profitability during periods of stress reduces loss absorbing capacity. 6 Mortgage credit spreads Signs of reduced lending margins could reinforce credit/asset price cycles and weaken financial system resilience. This indicator helps identify supply driven increases in credit. 7 Lending standards Easier lending standards (through reduced margins or non-price conditions) could weaken the resilience of bank portfolios. Evidence Global bank failures during the GFC were strongly correlated with leverage ratio. Methodology Return on assets, return on equity and net interest margin. Macro-financial time series models suggest that credit spreads tend to amplify housing market cycles. Source Total capital/rwa, Tier 1 capital/rwa, Registered Bank Leverage ratio (Tier 1 Disclosure capital/assets). Statements . Registered Bank Disclosure Statements, RBNZ Income Statement . Spreads on fixed mortgage rates RBNZ SSR, are used as a proxy for overall credit interest.co.nz spreads, and are defined using the relevant benchmark swap rate. They are computed for the 7 largest mortgage lenders and across all fixed terms. Two methods are used to aggregate data (i) simple average of spreads (ii) median of the best 5 spreads. Specials for low-LVR customers are used where available. Banks are asked whether they have RBNZ Credit eased or tightened non-price and Conditions price lending standards. A net Survey. percentage reporting tighter standards is computed (weighted by market share of each bank). Indicator Funding and liquidity Intuition Evidence Methodology Source 8 Net foreign assets and current account deficit. A high stock of external liabilities creates Current account deficits an exposure to offshore market volatility. have some predictive power This vulnerability is exacerbated if the in early warning literature. current account deficit is also elevated, or if debt is concentrated in sectors with high leverage. Net international investment position, bank external debt, and current account deficit. All are scaled relative to annual GDP. 9 Credit and deposit growth Credit growing well in excess of deposits suggests banks are reliant on wholesale markets to fund lending. So an increase in wholesale funding costs/shut-down in markets could result in a significant tightening in credit supply. Deposit growth has outstripped lending growth in recent years, and appeared to explain banks being unaffected by market turbulence in early 2012. Credit and deposit growth, both RBNZ Liquidity scaled relative to annual GDP. Uses Survey, SSR. non-market funding from Liquidity Survey after 2010 and retail funding from SSR to backdate. Haircuts are applied to non-market funding in line with the liquidity policy. Availability of offshore funding can quickly deterioate during periods of market stress. Vulnerabilities are increased if most of the borrowing needs to be rolled over at relatively short maturities. Compression of CFR relative to regulatory minima could indicate increased risktaking. Many banks reliant on shortterm markets came under stress during the GFC. Some evidence in early warning literature that reliance on short-term wholesale funding has predictive power for financial crises. Bank offshore funding, and StatsNZ IIP, component that is due in less than RBNZ Liquidity 90 days. Scaled as a % GDP. Core Survey. funding ratio defined as in liquidity policy. 10 & 11 Offshore funding rollover and core funding ratio (CFR). StatsNZ IIP, BOP 12 Indicator Bank funding spreads Intuition Evidence Methodology Source Wholesale spreads proxy markets perception of risk of NZ banks. Retail deposit spreads will give an indication of whether trends in offshore markets are passing through to average funding costs. Long-term wholesale (domestic and RBNZ Liquidity offshore) and retail funding spreads. Survey, SSR. Wholesale spreads are a simple average of the landed cost of new issues by big-4 banks from the Liquidity Survey, at terms of between 4 and 7 years. The spread between the average 6-month deposit rate and the 180 day bank bill is the proxy for retail spreads. 13 Global corporate bond spreads and S&P volatility Very compressed spreads or volatility in global markets could suggest risk is underpriced and an increased probability of a market correction. Currently use high-yield and investment grade bond spreads from the US, and S&P volatility. Could be extended in future. 14 Long-term interest rates Low global long-term interest rates can encourage a 'search for yield'. NZ longterm interest rates will typically move in line with global, potentially resulting in lower long-term bank lending rates. 10-year government bond yields. Haver, FSIS. Indicators for Housing Risks Indicator Housing market imbalances Intuition Evidence Rapid house price growth is a useful leading indicator of crises, especiallly when combined with rapid lending. 1 House price growth by region Rapid house price growth can increase the risk of a correction further down the road. Increases in the value of collateral also result in feedback effects with housing borrowing. Regional data can detect concentrations of risk. 2 House price to income ratio by region Measures stretch in house prices relative to underlying household income. Higher house price-to-income ratios will typically imply new buyers need to take on more debt to enter the market. House price inflation is of most concern where prices are stretched. 3 Rental yield by region Metric of how stretched house prices are relative to underlying rental income. Very low and falling rental yields could indicate speculative behaviour (ie investors are buying properties mainly due to high expected capital gain). 4 House sales by buyer type When combined with other data, purchases by different buyer types can shed light on market risks. Investor share could be of concern if DTIs are high and rising and rental yields are falling. Buyer shares can also give insights into the distributional impact of LVR/DTI limits. 5 House price expectations Strong house price expectations are an indicator of speculative/bubble market. Methodology Annual three-month moving REINZ. average house price inflation for NZ, Auckland, and rest of NZ. Index for rest of NZ is constructed using share of sales for different regions. Regions correspond to above. Average house values are estimated by rescaling REINZ indices to match QV average values in each region. Regional income data for rest of NZ is aggregated using share of households. Ratio of regional house price to rent, for NZ and Auckland. Assumes investor purchase houses similar in value to average house value estimates discussed above. International evidence that investor lending is more risky, and often amplifies housing cycles. Source REINZ, New Zealand Income Survey. REINZ, Department of building and housing. Create buyer shares from the CoreLogic, CoreLogic unit records using REINZ. Stata. Aggregate to actual sales using REINZ data, adjusted for lags. More detailed data is charted for Multiple Property Owners in Auckland. All series are 3-month moving averages. Chart shows expectations of households and professional forecasts. Data is lagged to show expected annual percent change at the relevant date. RBNZ. Indicator Borrower balance sheets 6 Household debt-to-income ratio (DTI) Intuition Evidence Methodology Source High aggregate DTI makes the Aggregate DTIs have household sector more susceptible to a tended to increase prior deterioation in servicing ability or rise in to international crises. interest rates. Downturns can be amplified as high DTI households attempt to restore balance sheets. Total household debt (inclusive of rental property debt) as a ratio to annual sector income. Owneroccupied debt-to-income ratio is shown for comparison. RBNZ Household Asset and Liabilities Survey. For total sector, total principal and interest payments are estimated using debt stock, assumed average time to maturity, and disposable income. New buyer DSRs are estimated for a typical buyer in NZ and Auckland, assumed to have average income (for their region), and purchase an average house with a 20 percent downpayment. Net credit growth is change in mortgage debt over quarterly timeframe. Gross lending is new commitments over the quarter scaled by mortgage debt at start of period. Approvals are included for historical perspective. RBNZ SSR Survey, HHAL Survey, Statistics NZ. 3-monthly commitments to firsthome buyers, other owneroccupiers and investors. RBNZ New Residential Mortgage Commitment Survey. 7 Debt servicing ratios (DSRs) Elevated DSRs could indicate increased financial pressure. For new borrowers, high DSRs required to enter the housing market could indicate new cohorts of borrowers becoming increasingly vulnerable, and/or the housing market could be hitting affordability constraints. Early warning literature suggests DSRs have strong predictive power at short horizons (< 1 year). DSR increased prior to GFC and late 80's in NZ. 8 Gross and net mortgage lending Rapid mortgage credit growth adds to indebtedness and is often a sign of rising systemic risks in the housing market. Rapid gross lending could still pose risks alongside weak net credit growth. Gross lending is more closely related to housing market and gives a better indication of changing indebtedness of new customers. Mortgage lending is found to have strong predictive power for international crises, and account for most of the predictive power of the credit gap. 9 Lending by buyer type Distribution of lending by buyer type can shed light on credit risks. Eg if lending is disproportionately driven by more indebted or higher risk types. RBNZ SSR, New Residential Mortgage Commitment Survey, Approvals survey. Indicator Lending standards 10 Distribution of lending by LVR Intuition Evidence Methodology Source High-LVR borrowers are more exposed to a house price decline. Need to significantly cut back consumption to restore wealth and are more likely to default if servicing capacity declines. International evidence suggests that losses associated with high-LVR loans are larger, and downturns are worsened by a prevalence of highLVR lending. Mean LVR above estimated median. Total lending and security above median LVR is estimated by assuming average LVR within each bucket is the mid-point between bottom and top of the bucket. RBNZ LVR New Commitments Survey 11 Stock of high-LVR mortgage lending See above for risks around high-LVR lending. This indicator captures the stock exposure from previous lending. Also used to assess effectiveness of LVR policy. See evidence for 10 above. Total high-LVR lending (includes off-balance sheet) for big-5, divided by total mortgages. Disclosure statements. 12 Consumer vs housing lending Significant increase in consumer lending while restrictions on mortgage lending are in place could indicate regulatory leakage. Consumer lending growth from SSR, including min and max growth rates of big 5. RBNZ SSR. 13 Non-bank mortgage lending and consumer lending A rising share of lending by non-banks, who do not specialise in mortgage lending, could indicate higher risks or disintermediation due to bank regulation. Number of mortgages registered CoreLogic. by non-banks, as a share of total. Indicators for Other Sectoral Risks Indicator Rural sector Intuition Evidence Methodology Source 1 Rural land price inflation by sector Rapid farm price growth can increase the risk of a correction further down the road. Increases in the value of collateral also result in feedback effects with borrowing. Sectoral data can detect concentrations of risk. Booming farm prices were associated with rapid lending prior to the GFC, helping to explain sharp fall in subsequent years. Dairy and rural land prices, annual percent change of 3 month moving average. REINZ. 2 Rural price-to-income ratios by sector High and rising farm price-to-income ratios indicate an increased risk of a correction in farm prices in the future. High ratios indicate that farm prices are detatched from incomes, and purchases are likely being made on the basis of capital gain rather than income yields. Rise in price-to-income prior to GFC resulted in a significant farm price correction in the following years. Farm price index is used to proxy REINZ, Stats NZ, total value of farm land, while DairyNZ. value of dairy land is estimated using REINZ index, total hectares and average land value from DairyNZ. Agricultural GDP used to proxy farm income, and dairy income is constructed using effective payout*milk production. Due to volatility in farm incomes, the trend in both income measures is estimated using a HP filter with judgement imposed to ensure the trend payout is $6.25 in 2015-16. 3 Rural land sales by sector Farm market can become very illiquid, Liquidity in the farm 3-month moving average of dairy REINZ. particularly during periods of stress. market dried up in the and total farm land sales. Including the number of sales helps the wake of the GFC. reader to interpret the farm price index, and gives a broader indication of stress in the market. 4 Indicator Credit growth by agricultural sector Intuition Evidence Methodology Source Rapid credit growth could indicate deterioating lending standards and increased probability of a crash in coming years, particularly if it occurs alongside increasingly stretched farm prices. Credit growth can also be a sign of stress during periods of weak cashflow. Lending growth in pre- Annual percent change in dairy GFC years was and total agricultural debt. correlated with subsequent stress. RBNZ SSR, Annual Agri Survey. RBNZ SSR, Annual Agri Survey, StatsNZ, DairyNZ. 5 Debt-to-income ratio by sector High DTI ratios makes the sector more susceptible to a deterioation in farm income or rise in interest rates. This is likely to also impair the functioning of the farm market during periods of stress. Work with unit records Total debt in each sector divided shows farms with high by income measures discussed DTIs have much larger above. breakeven payouts. 6 Dairy farm cash flow Cash flow pressures in the indebted dairy sector could indicate release of SCR is appropriate, and suggests heightened risk of downward adjustment in farm prices. See above - work with unit records shows periods of negative cash flow will have significant implications for high debt farms. Cash flow data are averages DairyNZ Economic from survey data. Breakeven Survey. cash flow is working expenses plus drawings plus interest servicing minus livestock revenue. Milk revenue is effective milk revenue. Rapid credit growth could indicate deterioating lending standards and increased probability of a crash in coming years. Rapid growth in commercial property and/or property development lending is particularly concerning given the volatility in prices over history. Rapid increases in commercial property lending preceded periods of stress in late 80's and GFC. Evidence suggests commercial property losses play a significant role in most financial crises. Commercial property lending is proxied using lending to the property and business services ANZSIC code. Non-property lending is proxied using lending to manufacturing, utilities, construction, wholesale trade, retail trade, transport and communications. Other business sector 7 Credit growth by sector SSR ANZSIC Data 8 Indicator Debt-to-income ratios by sector Intuition Evidence High DTI makes the sector more susceptible to a deterioation in profitability or rise in interest rates. Increase in stressed sales could amplify a commercial property market downturn. Source Methodology Debt as above. Nominal GDP SSR ANZSIC (total output) for the relevant Data, Statistics NZ. ANZSIC code is used to proxy income. Real GDP and Producers Price Indices are used to bring these figures up to date. 9 Commercial property price inflation by sector Rapid price inflation increases the risk of a future correction. Commercial property prices tend to have particularly large cycles, due to high average values, large transaction costs, and long build times. Rapid increases in Annual percent change in capital JLL commercial property return index for office, retail and prices preceded industrial. periods of stress in the late 80's and GFC. 10 Commercial property priceto-rent Rising price-to-rent ratios indicate that investors are putting an increasing weight on expected capital gain in purchasing decisions, as opposed to income yield. Price-to-rent ratios in three JLL sectors above, defined using the inverse of yields on market transactions. 11 Commercial property vacancy rates Low vacancy rates suggest upward pressure on rents and prices. Vacancy rates also tend to increase sharply during periods of financial stress. Vacancy rate by sector. Possibly JLL include prime/secondary split for office and malls/non-malls split for retail. 12 Commercial property supply pipeline Commercial property projects take a long time to execute. By the time construction finishes, rapid increases in supply often exacerbate periods of falling prices. Both the GFC and late 80's period of stress were exacerbated by increased supply coming onstream as prices were falling. Supply pipeline for each sector, defined as a percent of total square metres of space in each sector. JLL