From the subprime to the terrigenous: Recession begins at home
United States, Denmark, Ireland, Greece, Latvia, South Korea, Hungary, United Kingdom, New Zealand, France, Lithuania, Estonia, Finland, South Africa, Luxembourg, Australia, Spain, Sweden, Hong Kong, Iceland, Singapore, Canada, Slovenia, Belgium, Ukraine, Czech Republic, Norway, Slovakia, Russia, Bulgaria, Malta
The American word “subprime” refers to the credit-worthiness of borrowers. But that wasn't supposed to matter, because if the borrowers defaulted the lenders could simply sell the collateral to repay the loans. The scheme came unstuck because the collateral was residential property, which was overvalued due to a speculative bubble. When the bubble popped, borrowers owed more than the collateral was worth, so that lenders couldn't collect and therefore couldn't re-lend — for housing or anything else. Thus the financial system communicated the housing crash to the rest of the U.S. economy.
What is commonly called a “property” bubble or “housing” bubble is actually a land bubble. A building is worth no more than the cost of producing an equivalent building, and loses value due to deterioration and obsolescence. But land does not have a production cost, and its locational value tends to increase due to improvements in location-dependent services. Moreover, from the viewpoint of private entities, the overall supply of land is fixed, as is the supply that can be legally used for any particular purpose, and the supply within acceptable distance of any particular services, infrastructure, or markets. Yet access to suitably located land is essential to economic participation. So land values are competed upward until they absorb the economy's capacity to pay. In a growing economy, one should therefore expect land values to rise. But rational expectations periodically give way to belief in the greater fool: rising prices attract buyers who expect to sell at still higher prices, and that expectation becomes the sole support for current prices. Banks create credit against speculatively inflated prices and, in so doing, facilitate further inflation.
Eventually the illusion becomes unsustainable: the bubble bursts. At first, the “burst” is manifested as a fall in turnover while prices flatten out, as prospective sellers refuse to take losses. This in itself can be enough to cause a recession, not only because builders and developers slow down while waiting for unsold stock to clear, but also because slower sales cause a credit contraction by frustrating the sellers' plans to repay loans. If sellers cannot wait for the market to meet their expectations, prices are forced down, leaving other borrowers with negative equity, in which case the credit crunch and recession are obviously worse.
The word “subprime” is therefore a distraction. Yes, loose lending helped to pump up the U.S. housing bubble; but rising collateral values came first, and loose lending was the response. Moreover, not all of the loose lending was to subprime borrowers (to say nothing of those “subprime” borrowers who met all the criteria of higher-rated borrowers, except the unwritten one concerning skin colour).
Describing the present global disaster as a financial crisis is a further distraction. Yes, British and European banks were exposed to repackaged U.S. subprime loans. But more importantly, they were exposed to their respective domestic property markets. As the following survey shows, most of the current recessions in European and Asian countries were preceded by bursting domestic property bubbles.
Strictly speaking, the mechanism by which a collapse in the property market spreads to the wider economy does not depend on the reason why the market collapses, provided that the collapse comes as a surprise to market participants. Hence, when assessing the correlation between falling property values and recessions, it is not necessary to establish in every case (although I shall venture an opinion in each case) that the fall in property values represents a bursting bubble. But when property values rise rapidly and then fall, a bursting bubble is the simplest explanation and satisfies the criterion of surprise. Moreover, a genuine bubble tends to involve high turnover and high leverage, which amplify the effects of the surprise.
President Obama's much-reported statement that “the crisis began in the U.S.” was ambiguous as to whether the beginning was causal or merely chronological. He is generally assumed to have meant the former. But whatever he meant, I argue for the latter. If the bursting of the U.S. housing bubble caused the U.S. recession — as seems to be acknowledged by all — and if other countries had similar problems with their domestic property markets before they too fell into recession, the claim that the latter recessions were fully imported from the USA, although convenient for certain political and pecuniary interests, strains credulity.
1. Parallel recessions
Changes in GDP, unless otherwise specified, are real, seasonally adjusted, relative to the previous quarter, and not annualized.
The bubble: The S&P/Case-Shiller inflation-adjusted home price index rose continuously, at an average rate of 7.1%/year, from the trough in Q4 of 1996 to the peak in Q4 of 2005.
The burst: Established home sales (by number, seasonally adjusted) peaked in Jun.'05, fell in Jul.'05, partly recovered in August, then fell for five consecutive months, and were more than 38% below the peak by Dec.'08. The aforesaid S&P/Case-Shiller index fell in Q1 of 2006, and by late 2008 was more than 30% below its peak.
The recession: The National Bureau of Economic Research (NBER) dates the onset of recession at Dec.'07, although this corresponded to an isolated quarter of negative growth. Some critics would date the onset earlier, claiming that official GDP figures underestimate inflation and/or wrongly include some repatriated profits generated offshore. The usual definition would give a later onset: real GDP fell about 0.1% in Q3 of 2008, 1.6% in Q4, and (according to advance estimates) 1.5% in Q1 of 2009. Hence the NBER dating may be taken not only as “official”, but also as a compromise between competing claims.
The bubble: Real year-on-year changes in home prices hovered between 8% and 10% (already unsustainable) between mid 2004 and mid 2005, then rose to 18% in early 2006 [source: see the endnote].
The burst: The number of existing homes for sale bottomed in Q1 or Q2, 2006 and immediately shot up, almost doubling by Sep.'07. Home prices fell in Q3 of 2006, and again in Q4. The real year-on-year change in prices went negative in Q3 of 2007.
The recession: Real GDP growth was initially reported as negative in Q4 of 2007 and Q1 of '08, and a recession was officially announced. The estimated growth for Q4 of '07 was later revised to a slightly positive value. The next negative figures were 0.8% for Q3 of '08 and 1.9% for Q4. However, year-on-year growth was negative for Q1 of '08 and even for Q2 of '07. Again the official announcement may be taken as a compromise.
The bubble: Nominal home prices rose 300% from early 1996 to early 2007 (average rate > 13%/year). P/E ratios reached all-time highs (approaching 100 in some suburbs in early 2006).
The burst: Home-loan approvals* (by number) peaked in Q2 of 2006 for both new and 2nd-hand homes. In Q3 of 2006 they fell 34% for new homes and 26% for 2nd-hand homes. While both falls were apparently assisted by seasonal factors, subsequent trends were clearly downward. Nominal prices stopped rising, hence real prices began to fall, in Q1 of 2007. Nominal prices started falling in Feb.'07 and declined in every subsequent quarter (to the end of 2008).
The recession: Real GDP fell 1% in Q1 of 2008 and 0.4% in Q2, rose 1.1% in Q3, but plunged 7.1% in Q4.
The bubble: The index of urban home prices (excluding Athens) rose from 1993 to 2007. The rise was mostly slow, but was 15% through 2005 and 12% through 2006. In mid 2008, gross rental yields were 2.5% to 4.2%.
The burst: Outside Athens, real prices started falling in Q1 of 2007, and real year-on-year changes went negative in Q4 of 2007.
The recession? After a year of slowing growth, real GDP contracted 1.2% in Q1 of 2009.
The bubble: Nominal prices per square metre for apartments in Riga almost quadrupled between early 2004 and mid 2007. The real year-on-year increase in home prices peaked at 62% in the 2nd half of 2006. By early 2007 it was clear than many borrowers were overcommitted; some had borrowed up to ten times their annual salary.* In mid 2008, rental yields ranged from 4.2% to 6.5%, compared with variable mortgage interest rates around 9% (higher in late 2007).
The burst: As a proxy for activity in the mortgage market, we may cite the monthly change in total loans in the local currency [bottom graph],* which indicates that deleveraging began in Q1 of 2007. Standard apartment prices peaked in Apr.'07, declined in May '07, and were 45% below the peak by Dec.'08. In 2008 the number of sales* was 65% below that in 2007, and commencements of new projects “practically stopped”. By March 2009, 21% of mortgages were in arrears.*
The depression: Real GDP fell 5% in Q1 of 2008, 1.4% in Q2, 0.4% in Q3, 3.9% in Q4, and an estimated 18% in the year to Q1 of 2009.
It was barely a bubble: Nominal home prices rose 68% from Q3 of 1998 to Q1 of 2008 — an average rate of only 5.6%/year. Real home prices fell 1.9% through 2005 but rose 3.8% through 2006, 6.4% through 2007, and 9.8% in the year to Q1 of 2007. For all property, capital growth was 19.2% in the year to Dec.'07, but rental yields were 6.6% for 2007 and 6.4% for 2008.
But it still popped: Buyers became conscious of a glut “early” in 2007. Real home prices flatlined from about Q2 of 2007, declining 0.4% in the year to Q2 of 2008. The capital loss across all property was 2.3% in the year to Dec.'08.
The recession? For Q4 of 2008, real GDP fell 5.1% (revised from 5.6%). For Q1 of 2009, the central bank claims positive growth of 0.1% (also subject to revision).
The artificial bubble: Average nominal prices of old condos in Budapest doubled between 2000 and 2006. This corresponds to a real increase of about 8%/year, which is still unsustainable when compared with GDP growth (about 4.4%/year over the same period). The price rise was partly driven by home buyers' grants and interest subsidies introduced in 2001.
The artificial burst: The grants and subsidies were reduced in 2003-4. Prices of those old condos fell in real and nominal terms between 2006 and 2007; for want of better data, the fall will be dated mid 2007. But rental yields were still described as “moderate to good” in mid 2008. Lending in Swiss Francs by foreign-owned banks ended in late 2008, causing the price slump to accelerate.
The recession: Real GDP fell 0.3% in Q2 of 2008, 0.8% in Q3, 1.5% in Q4, and 2.3% in Q1 of 2009.
The bubble: Real home prices rose 170% from Jan.'96 to Jul.'07 (average rate: 8.7%/year).
The burst: Transactions (home purchase approvals) peaked in Q4 of 2006, fell in Q1 of 2007, and continued falling through 2007 and into 2008. Real prices started falling in Q3 of 2007, and by Oct.'08 had fallen nearly 20% from the peak.
The recession: Real GDP stood still in Q2 of 2008, then fell 0.7% in Q3 of 2008, 1.6% in Q4, and 1.9% in Q1 of 2009.
The bubble: Nominal home prices doubled between early 2002 and mid 2007 (average rate of increase > 13%/year).
The burst: Home sales (by number) peaked in about Mar.'07, started falling in about Apr.'07, and fell more than 40% in 6 months. Real prices started falling in Q3 of 2007. Real year-on-year changes were negative from late 2007 onward.
The recession: Real GDP fell 0.3% in Q1 of 2008, 0.2% in Q2, 0.5% in Q3, and 0.9% in Q4.
The bubble: For existing apartments in Paris, nominal prices per sq.m. almost tripled from late 1997 to mid 2008. Residential rental yields got down to 3.4%.
The burst: Nationwide, real prices fell in Q3 of 2007, recovered partially in Q4, and declined through 2008. Capital losses for all property to Dec.'08 were 7.4% for the half-year and 6% for the year.
The recession: Real GDP fell in Q2 of 2008, rose 0.1% in Q3 according to initial figures, then fell 1.5% in Q4 and a further 1.2% in Q1 of 2009. But the revised figures on which recession was announced showed the economy contracting 0.2% in Q3, in which case the recession should be dated from Q2 of 2008.
The bubble: From early 2002 to late 2006, nominal prices of existing apartments in Vilnius rose 275%, while rental yields for one-bedroom apartments declined from nearly 8% to just over 3%.
The burst: Housing loan approvals peaked in Mar.'07, declined in Apr.'07, and were about 90% below the peak by early 2008. Nominal prices of new and existing apartments in Vilnius flatlined, and real prices fell, from Q3 of 2007. Nominal prices fell about 20% in the second half of 2008.
The depression: Real GDP fell 0.3% in Q3 of 2008, 1.4% in Q4, and 9.5% in Q1 of 2009 (12.6% year-on-year).
The bubble: Nominal home prices in Tallinn rose 460% from early 2001 to Q3 of 2007. As prices fell slightly in Q2 of 2006, in Q1 of 2007, and (barely) in the six months to Q2 of 2007, the apparent timing of the peak may be influenced by seasonal or sampling effects.
The burst: The number of notarized real-estate transactions peaked in Q4 of 2005 and declined in Q1 of 2006, and was less than a third of its peak value by Q1 of 2009. Nominal home prices peaked in Q3 of 2007 and fell more than 13% in Q4 of 2007. Real prices for the same quarter were down not only quarter-on-quarter but also year-on-year.
The depression: Real GDP fell 1.5% in Q1 of 2008, 1.7% in Q2, 2.2% in Q3, 4.3% in Q4, and an estimated 15.6% in the year to Q1 of 2009.
The bubble: From 1995, real home prices rose 52% to 1999, fell 10.5% to 2001, then rose 43% to 2007. In 2008, rental yields in Helsinki were low (3.7%) for large properties but moderate for smaller properties.
The burst: The percentage of unsold homes outside the capital more than doubled in 8 months, spanning (at least) Q2 of 2007. Real prices of existing apartments and terrace houses peaked in Q3 of 2007, fell in Q4 of 2007, and continued falling at an increasing rate through 2008.
The recession: Real GDP fell 0.3% in Q3 of 2008 and 1.3% in Q4.
The bubble: Nominal year-on-year “capital growth” for all property was 16.5% to Dec.'06 and 17.7% to Dec.'07. Prices rose even faster at the affordable end of the housing market. Nominal home prices in rural areas more than doubled from early 2003 to early 2007. The rental yield for all property in the half-year to Jun.'08 was 4.1%. Yields for larger houses and apartments in Cape Town were similar or lower, although yields in Johannesburg were much higher.
The burst: Real home prices roughly plateaued in Q3 of 2007 and fell in Q4 of 2007. The real year-on-year change in home prices went negative in early 2008.
The recession: Real GDP fell 0.5% in Q4 of 2008 and 1.6% in Q1 of 2009.
The bubble? Real home prices rose 10% through the year for 2004, 0.8% for 2005, 5.5% for 2006, and −1% for 2007. Note the decelerating trend. On the other hand, rental yields for Luxembourg City apartments in March 2009 ranged from 3.6% to 5.2% and were considered low. It would be useful to have data covering a longer time span.
The burst? In real terms, advertised prices of homes (not actual sale prices) roughly flatlined from Q3 of 2006. This timing coincides with a fall in the building industry's duration of assured activity,* and with a year-on-year fall (albeit an isolated one) in the number of mortgages granted, consistent with the theory that buyers were reacting to a perceived glut. In real and nominal terms, advertised prices have been declining since Q1 of 2008.
The recession: Real GDP fell 0.3% in Q3 of 2008, and 4.5% in Q4.
The bubble: The ABS home-price index for 8 capital cities rose about 76% from Q1 of 2002 to Q1 of 2008. Prices are high relative to household income and per-capita GDP. While rental vacancy rates are apparently low, the figures do not account for speculative vacancies — that is, unoccupied dwellings which are not available “to let”, but which may become available if expectations of capital gains are dashed.
The burst: Residential land sales (by number of lots) peaked in Q2 of 2007 and declined in Q3 of 2007 and (at least) the next four quarters. This timing is supported by housing loan approvals, which peaked (by value, seasonally adjusted) in Jun.'07 and fell in Jul.'07 [RBA, Statement on Monetary Policy, May 2009, graph 57; cf. ABS 5609.0]. It is further supported by the total number of established house transfers in Australia's eight capital cities [calculated from ABS 6416.0, tables 7 and 8], which began falling in Q3 of 2007, although the decisive fall (23%) came in Q1 of 2008. Real home prices [ABS 6416.0] fell in Q1 of 2008 and the next three quarters. Nominal prices peaked one quarter later than real prices, then fell 6.7% to Q1 of 2009. In Q4 of '08 and Q1 of '09, prices recovered in less expensive suburbs only [RBA, Statement on Monetary Policy, May 2009, graph 32], showing the temporary effect of the Federally-funded First Home Owners' Boost. Commercial property prices peaked and fell in synchronism with home prices [ibid., graph 39]. In the year to Mar.'09, the Property Council / IPD index for all property showed capital “growth” of −7.6% and income of 6.6%, giving the first negative total return since 1993.
The recession? Real GDP fell 0.5% in Q4 of 2008. The unemployment rate increased by 1.2 percentage points between Dec.'08 and Mar.'09 [ABS 6202.0], consistent with a second quarterly fall in GDP. The bounce in employment in Apr.'09 probably represents a pull-forward of activity due to the anticipated expiry of the First Home Owners' Boost — in which case the unexpected extension of the boost (at the full rate to Sep.30, then half that rate until Dec.31) will presumably spread the activity into the second half of the year. On account of this stimulus, positive GDP growth for Q2 of 2009 is not out of the question. But a contraction in Q1 looks inevitable.
The bubble: Home prices per unit area doubled between early 2002 and early 2007, as gross rental yields declined to about 2%/year. Real year-on-year increases in home prices exceeded 10% from early 2002 to early 2005 and declined to about 5% in early 2007.
The burst: Home sales (by number) peaked in the 1st half of 2006 and declined significantly (by comparison with seasonal and sampling effects) in the 2nd half of 2007. Real home prices fell slightly in the quarter and the year to Q1 of 2008. Real and nominal prices declined through the rest of 2008.
The recession: Real GDP fell 0.3% in Q3 of 2008, 1% in Q4, and 1.9% in Q1 of 2009.
The bubble: Nominal “capital growth” for all property in the year to December was about 13% for 2005, 16% for 2006, and 12% for 2007. In real terms, home prices more than doubled in the decade to Q1 of 2008.
The burst: Purchases of one- and two-dwelling buildings peaked in Nov.'07 and began falling in Dec.'07. Real prices plateaued in Q4 of 2007 and fell in Q1 of 2008. Nominal capital “growth” for all property in the year to Dec.'08 was −7.9%.
The recession: Real GDP fell 0.6% in Q1 of 2008, 0.5% in Q2, 1% in Q3, and 2.4% in Q4. But the figures on which recession was first announced had the economy contracting in Q2 of 2008 and again in Q3, but not in Q1. Year-on-year figures show growth of 3% in Q2, suggesting a still later onset of recession. The initial announcement may be taken as a compromise.
The bubble: Nominal prices per sq.m. for “class-C” apartments (70–100 sq.m.) rose 200% between Apr.'02 and Feb.'08 (average rate: 20%/year, in a period of otherwise low inflation).
The burst: The number of sale-and-purchase agreements, for all properties and for residential properties alone, peaked in Nov.'07, declined in Dec.'07, partly recovered in January, declined for four consecutive months, and eventually bottomed in Nov.'08, 79% below the peak. Residential sales by value behaved similarly, falling 87% from Nov.'07 to Nov.'08. Nominal prices of class-C apartments peaked in Feb.'08, fell in Mar.'08, recovered partially in May (although this may be due to statistical “noise”), and were 25% below the peak by the end of 2008. Real home prices peaked in Q1 of 2008 and declined at an increasing rate through the next three quarters.
The recession: Real GDP fell 1% in Q2 of 2008, 0.8% in Q3, 1.9% in Q4, and 4.3% in Q1 of 2009.
The bubble: Nominal home prices increased 150% between early 2000 and early 2008.
The burst: Property market turnover (by value) peaked in Oct.'07, started falling in Nov.'07, and was down about 65% from the peak by Jan.'08. Real home prices fell in Q1 of 2008. Nominal and real prices fell in the next three quarters. Note that the bank nationalizations/receiverships did not start until late Sep.'08.
The recession: Real GDP fell 3.4% in Q3 of 2008 and 0.9% in Q4. OECD and Eurostat figures disagree, showing negative growth from Q4 of 2007 through Q2 of 2008; but the corresponding year-on-year figures show healthy growth into 2008. The issue is further complicated by the difficulty of correcting for inflation when the exchange rate is changing rapidly. In any case, recession was not announced until the release of figures for Q4 of 2008.
The bubble? The real year-on-year change in home prices rose from about 6% in Q3 of 2006 to about 26% in Q4 of 2007, and remained positive until mid 2008. Rental yields in early 2008 were considered very poor, ranging from 2.7% to 4.2%. That much is consistent with a bubble. On the other hand, rental yields increased from 2004 to 2008, and the price index did not quite reach the peak of the mid '90s.
The recession/depression: Real GDP shrank in Q2 of 2008 and the next three quarters, and was down 10.1% in the year to Q1 of 2009.
It was barely a bubble: Nationwide, real year-on-year changes in home prices briefly nudged 11%, but exceeded 5% for only about 15 months. Rental yields remained within reason. In the western provinces, price rises were much faster.
But it still popped: Monthly home sales (by value) peaked in Dec.'07, began to fall in Jan.'08, and declined about 45% to Jan.'09. Monthly average home-sale prices peaked in May '08, began to fall in Jun.'08, and declined about 12% to Jan.'09. Real year-on-year changes in home prices went negative in mid 2008. Nominal capital losses in the year to Dec.'08 were 2.3% for all property, 4% for industrial property, and 5.7% for retail.
The recession? Real GDP fell 0.8% in Q4 of 2008. Monthly GDP figures continued to decline in January and February, making 7 consecutive monthly declines.
The bubble: Real year-on-year changes in home prices averaged about 10% from mid 2003 to mid 2007. In mid 2008, rental yields in central Ljubljana ranged from 4.3% to 5.5%.
The burst: Real prices roughly plateaued from early 2007 and have been falling since Q2 of 2008.
The recession? Real GDP fell 4.1% in Q4 of 2008.
The bubble? In real terms, average home prices rose 26% through 2005, 10.9% through 2006, and 2.5% through 2007. From 2000 to 2007, the average real price rise was 70% for houses and 59% for apartments, while rents declined in real terms and rose only about 4% in nominal terms. That said, rental yields in 2008 were considered moderate rather than low.
The burst? The nominal “capital loss” for all property in the year to Dec.'08 was 1.2%. In real terms the loss would have been greater. The timing of the downturn is more uncertain than for most other countries considered here, but will be taken as mid 2008 for want of better data.
The recession: Real GDP was static in Q3 of 2008, then contracted 1.7% in Q4 of 2008 and 1.6% in Q1 of 2009.
The bubble: Nominal prices per sq.m. for flats in Kiev grew more than eightfold from early 2002 to mid 2008.
The burst: Real prices probably fell in Q3 of 2008. Nominal prices fell 9% in Q4.
The depression: Real GDP fell about 8% in the year to Q4 of 2008, and about 20% in the year to Q1 of 2009. The timing of the recession was heavily influenced by a precipitous fall in the demand for exports, especially steel.
The double-headed bubble: Apartment prices rose 90% from early 1999 to mid 2003, retreated about 6%, then resumed rising under the influence of improving terms of trade, eventually surpassing the earlier peak.
The burst: An anticipated increase in the VAT on construction (from Jan.'08) was credited with causing a construction binge in 2007. But the high rate of construction continued unabated into 2008. We must therefore seek some other cause for the fall in sales of flats in 2008: sales for the first three quarters of 2008 were 38% lower than for the same period of 2007. In the absence of finer detail, we may tentatively date the downturn in sales from Q1 or Q2 of 2008. Home prices started falling in Q3 of 2008.
The recession: Real GDP fell 0.9% in Q4 of 2008 and 3.5% in Q1 of 2009.
The bubble: Nominal prices per sq.m. for new detached homes rose about 290% from Q2 of 1993 to Q3 of 2007 (average rate 10%/year) while inflation was otherwise low. By mid 2006, rental yields were below 5% (3.6% for large homes in Oslo). Yet real prices rose more than 13% through 2006 and 11% through 2007.
The burst: According to Norges Bank (Financial Stability 2/08, Chart 3.9), the stock of homes for sale bottomed in mid 2006 and started rising rapidly in Q3 of 2006, whereas the turnover of existing housing peaked in early 2008 and began a steep decline in Q2 of 2008. The same report, at Chart 3.12, has prices peaking in early 2007, but the methodology is not clear. According to Global Property Guide, nominal prices per sq.m. of new detached homes fell more than 4% in Q3 of 2008 and more than 6% in Q4, earlier smaller price falls being apparently due to seasonal factors. Statistics Norway, the apparent source of the figures for detached homes, provides an “All dwellings” table which supports the same timing, although prices for other types of houses peaked earlier. All sources agree that prices peaked well after the upturn in unsold stock.
The recession: Norwegian statistics distinguish between “total” and “mainland” GDP. On the former measure, the economy contracted 0.8% in Q3 of 2008, rebounded 0.8% in Q4, and contracted 0.4% in Q1 of 2009. On the latter, which excludes North Sea oil exports, the respective growth figures were 0.1%, −0.8%, and −1%, implying a recession beginning in Q4 of 2008. The relationship (if there is one) between the property market and the wider economy should be more evident in the mainland figures.
The bubble: Nominal home prices rose more than 80% from Q2 of 2005 to Q2 of 2008. Real prices rose 12.9% through 2006 and 28% through 2007. The price of an average apartment in Bratislava reached 8 years' household income. Rental yields in Bratislava fell from 10.1% to 5.2% in the two years to mid 2008; but, as only 0.1% of the housing stock is owned by private landlords, the significance of published rental yields is open to dispute.
The burst: For Q1–Q2 of 2008, Lexxus a.s. reported* that “the number of projects that are experiencing trouble selling flats is increasing.” Their next report stated that “the May-October 2008 period can be characterised as experiencing a sharp fall in demand for new housing. Almost half of projects on the Bratislava market have experienced no sales, and in some cases agreements have been cancelled.” Combining these two statements, we make take Q2 of 2008 as the onset of a decline in turnover. Nominal and real prices declined in Q3 of 2008 and again in Q4.
The depression? Real GDP grew rapidly and steadily to the end of 2008, then plummeted 11.2% in Q1 of 2009. Slovakia's adoption of the euro from Jan.1, 2009, may divert attention from the excesses of the property market.
The bubble: Real year-on-year increases in home prices averaged about 8% from 2001 to 2005, shot up to 40% in 2006, and retreated to about 11% in 2007. Price rises continued in defiance of downturns in other countries. The ratio of per-sq.m. home prices to per-capita GDP became very high by global standards.
The burst: The Russian stock-market crash of Sep.'08 frightened prospective buyers. Sales fell. Prices dropped 8% to 10% between Q3 and Q4 of 2008 and continued falling into 2009. The time difference between falling sales and falling prices is uncertain, but clearly small.
The recession: Quarter-on-quarter GDP figures are apparently not available, but a deputy minister admitted in December 2008 that “A recession has started.” As the housing crash was precipitated by a stock-market crash, which was caused by falling export prices, which in themselves are recessionary, the property crash cannot confidently be blamed for the onset of recession, but will obviously affect its severity.
The bubble: In a high-inflation environment, the real year-on-year increase in average home prices (per sq.m.) exceeded 10% from Q4 of 2006 to Q3 of 2008, and exceeded 15% from Q1 of '07 to Q2 of '08. Meanwhile rents hardly changed; rental yields in Sofia in Aug.'08 ranged from 4.7% to 6.1% (low by comparison with local interest rates). In 2007, 35% of foreign direct investment (FDI) went into real estate. Prices in provinces along the Danube rose threefold to fourfold in the last year of the upswing. Prices in Black Sea resorts rose much less, due to overbuilding leading to overloading of local infrastructure.
The burst: FDI in real estate fell 39% between 2007 and 2008. Property transactions in Sofia started falling in Q2 of 2008; in May and June, supply exceeded demand by 50%. In Q3, transactions were down 18% year-on-year. Real prices almost plateaued in Q3 of 2008. Real and nominal prices fell in Q4 of 2008.
The recession: After several years of growth in excess of 6% per year, real GDP shrank 3.5% in the year to Q1 of 2009. A year-on-year contraction of such magnitude is hard to classify as anything less than a recession.
The bubble: The real year-on-year increase in home prices exceeded 10% from late 2002 to early 2005 and spiked at 33% in early 2004. Meanwhile rents have declined due to pro-construction (and pro-density) planning rules; in March 2009, rental yields ranged from 2.4% to 3.55%.
The slow-motion burst: From p.18 of the Central Bank's 2008 Financial Stability Report, we may infer that sales fell in Q3 of 2008 as buyers became hesitant. The real year-on-year increase in home prices first went negative in Q3 of 2006 and has been continuously negative since Q3 of 2007. In each quarter mentioned there was a conspicuous reversal in quarter-on-quarter figures. But the underlying trend is smooth rather than precipitous, so that one cannot (yet) identify a “burst” in prices. The gradual slowdown in price appreciation may be explained by two peculiarities of the Maltese market, namely the pro-construction planning regime (already mentioned) and the complete absence of recurrent property taxes. In particular, the non-taxation of buildings further encourages construction, while the non-taxation even of land makes it easy for landlords to keep properties vacant, forgoing the rental income that would otherwise be needed to cover the tax; in July 2008 it was reported that “A third of the property stock is second homes, and 21% of all properties are empty.” The larger the oversupply that the market is willing to tolerate, the longer the time frame over which the oversupply builds up and the market responds. Had there been a land tax (not a building tax), at least some of the vacant properties would have been offered to let, and at least some of those that could not be let would have been offered for sale; thus the self-correction of the market would have been more timely and the unrealistic rise in prices would have been at least partly suppressed.
The recession? Real GDP grew at a falling rate through 2008 and eventually contracted 0.1% in Q4 of 2008.
2. Recessions out of sequence
The bubble: Late in 2001, in response to the long decline in home prices since 1994, the legislature passed a bill providing for low-interest loans, cutting the rate of the Land Value Increment Tax (effectively a capital gains tax on land alone), and allowing foreign acquisition of Taiwanese real estate. Prices then rose from Q2 of 2003 to Q2 of 2008. Although real prices did not quite reach 1994 levels, rental yields fell to 2.3% — the lowest in Asia.
The burst: The number of housing transactions peaked in Jun.'08, and declined in Jul.'08 and each of the next 7 months. Nominal home prices fell 5.3% in Q3 of 2008.
The recession/depression: Year on year, real GDP shrank about 1% to Q3 of 2008, more than 8% to Q4, and more than 10% to Q1 of 2009. Falling exports accounted for 3/4 of the last figure.
It was barely a bubble: Nominal home prices per unit area doubled between 1998 and 2007. Real year-on-year changes in home prices averaged just over 6% from 2001 to 2005. This was unsustainable in view of slow GDP growth (<1%/year from 2001 to 2007). Rental yields in historic Rome ranged from 3.2% for the largest apartments to 4.5% for the smallest.
But it still popped: Home sales (by number) declined 14% from the 1st half of 2007 to the 1st half of 2008. The real year-on-year change in home prices was down to 1% by early 2008. Nominal capital “growth” for all property was 3.8% in the year to Dec.'08, but −2.5% in the 6 months to Dec.'08.
The recession: Real GDP fell 0.6% in Q2 of 2008, 0.8% in Q3, 2.1% in Q4, and 2.4% in Q1 of 2009.
3. Recessions without domestic property bubbles
No bubble: Real home prices have been declining very slowly since 1991. Total returns for all property rose from 2005 to 2007, but not into bubble territory. Nominal “capital losses” for all property in the year to December were 3.1% for 2006, 0.2% for 2007, and 1.4% for 2008.
The recession: Real GDP fell 0.5% in Q2 of 2008, 0.5% in Q3, 2.2% in Q4, and 3.8% in Q1 of 2009.
No bubble: Real and nominal home prices have been slowly falling since the recession of the early '90s. Land-price figures and IPD figures [see sources] suggest a modest recovery through 2006 and 2007. For all property in the year to Dec.'08, capital losses were about 4.4%, and the rental yield was about 4.9% (high by comparison with interest rates).
The recession: Real GDP fell 0.9% in Q2 of 2008, 0.6% in Q3, 3.8% in Q4, and 4% in Q1 of 2009.
No bubble: Growth in home prices has been very slow since 2002. Nominal “capital growth” for all property in the year to December was +5.8% for 2007, but −1.7% for 2008 (−3.5% for the last quarter). Rental yields remained adequate (e.g. 7.7% for average apartments in Amsterdam).
The recession: Real GDP fell 0.1% in Q2 of 2008, 0.5% in Q3, 1.2% in Q4, and 2.8% in Q1 of 2009.
No bubble: The few available figures indicate no unusual rise in prices. There was a sharp upswing in mortgage lending, but from a low base, so that the value of the mortgage market remained a relatively small fraction of GDP (10% in 2007).
The recession: Real GDP fell 0.6% in Q3 of 2008, 2.5% in Q4, and 5.9% in Q1 of 2009.
No bubble: From 2002 through 2006, growth in home values was moderate and decelerating.
The recession: Real GDP fell 0.4% in Q3 of 2008, 1.9% in Q4, and 1.5% in Q1 of 2009.
No bubble: Real year-on-year changes in home prices were negative from 1992 to 2000, and thereafter averaged about 2%, never exceeding 4%.
The recession: Real GDP fell 0.1% in Q3 of 2008, and 0.3% in Q4.
No bubble: Nominal home prices in Vienna declined about 5% from 1992 though 2004, leaving plenty of scope for growth, then indeed started growing about 7.5% per year. Outside Vienna, the growth spurt was short-lived, registering as a single-quarter spike on a year-on-year graph. By 2008, the mortgage market was worth only 25% of GDP, compared with an EU average of 50%.
The recession: Real GDP stood still in Q3 of 2008, then shrank 0.4% in Q4 of 2008 and 2.8% in Q1 of 2009.
Definitions of symbols
T (for Turnover) means the fall from the peak in property turnover or related loans (direct figures on turnover being hard to find).
t means a reported or apparent glut of property for sale, suggesting that sales have slowed (another proxy for the elusive figures on turnover — not because the proxies are any easier to find, but because any alternative is better than none).
V (for Value) means the fall from the peak in property prices, in real terms.
R means the start of the official Recession.
r means one quarter of negative growth to date, perhaps indicating the onset of recession.
Group 1: Parallel recessions
The following table is in chronological order, sorted first by the fall in property values (V), then by the fall in turnover (T or t), then by the onset of recession (R or r).Year | 2005 | 2006 | 2007 | 2008 | 2009 Quarter |1 2 3 4|1 2 3 4|1 2 3 4|1 2 3 4|1 ------------|-------|-------|-------|-------|----- USA | T |V | R| | Denmark | | t V | R| | Ireland | | T |V |R | Greece | | |V | |r Latvia | | |T V |R | S. Korea | | | tV | r| Hungary | | | V | R | UK | | |T V | R | New Zealand | | | T V |R | France | | | V | R | Lithuania | | | T V | R | Estonia | |T | V|R | Finland | | | t V| R | S. Africa | | | V| R| Luxembourg | | t | |V R | Australia | | | T |V r| Spain | | | T |V R | Sweden | | | T|V R | Hong Kong | | | T|V R | Iceland | | | T|V R | Singapore** | | | T | VR | Canada | | | |T V r| Slovenia | | | | V r| Belgium | | | | V R| Ukraine | | | | V R| Czech Rep. | | | | T V R| Norway | | t | | T V R| Slovakia | | | | T V |r Russia | | | | TVR| Bulgaria | | | | T V|R Malta | | | | T r|
Group 2: Recessions out of sequenceYear | 2005 | 2006 | 2007 | 2008 | 2009 Quarter |1 2 3 4|1 2 3 4|1 2 3 4|1 2 3 4|1 ------------|-------|-------|-------|-------|----- Taiwan** | | | | TVR | Italy | | | | TR V |
Group 3: Recessions without domestic property bubblesYear | 2005 | 2006 | 2007 | 2008 | 2009 Quarter |1 2 3 4|1 2 3 4|1 2 3 4|1 2 3 4|1 ------------|-------|-------|-------|-------|----- Germany | | | | R | Japan | | | | R | Netherlands | | | | R | Mexico | | | | R | Portugal | | | | R | Switzerland | | | | R | Austria | | | | R|
** For Singapore and Taiwan, the adjacent symbols should be in the same quarter. In other cases, a symbol placed between two quarterly columns indicates a half-yearly or yearly figure or another uncertainty in timing.
With the aid of the tables in the summary, one can discern the following patterns:
- A downturn in the property market, especially in turnover (sales) of properties, is a leading indicator of recession, with a lead time of up to 9 quarters for turnover, or up to 8 quarters for values. Of all the countries in which a conspicuous fall in turnover was documented, there was no case in which the onset of recession preceded the fall in turnover, and only one case (Taiwan) in which the onset of recession seems to have been in the same quarter as the fall in turnover; and in the one case (Italy) in which recession preceded the downturn in property values, it did not precede the downturn in turnover.
- In the property market, a fall in turnover is a leading indicator of a fall in prices, and the lead time is usually one to two quarters. In no case is there persuasive evidence of the fall in prices coming first, although there are three cases (Taiwan, Russia, South Korea) in which the two falls may have been almost simultaneous, and one case (Norway) in which the sources disagree.
- Recessions are mostly home-grown. In the seven countries that entered recession without any resemblance of a bursting domestic property bubble, and in Taiwan, where the onset of recession was almost simultaneous with the property downturn, the recession came comparatively late, leaving plenty of time for the domestic economy to be affected by recessions in other countries, or bursting property bubbles in other countries. Such cases, however, were in the minority; in most countries the recession was preceded by a downturn in the domestic property market.
Concerning the last point, I do not deny that recessions in themselves are somewhat contagious; when people and firms in one country fall on lean times, they tend to import less, in which case they directly affect the GDPs of other countries. Neither do I deny that, due to the partial globalization of the financial industry, devalued collateral (chiefly property) in one country can affect the availability of credit in other countries. I do not even deny that a recession following a domestic property crash may occasionally have an external proximate cause; there is evidence of that in Ukraine and Russia. But clearly the first country to suffer a property crash and recession cannot blame the rest of the world. And because economic interactions are made easier by geographic and jurisdictional proximity, one would expect the economic health of one country to be more exposed to its domestic property market than to external events. The above tables support that expectation.
If, as I contend, recessions come mostly from domestic property markets, then the real significance of globalization lies in international arbitrage by property investors, which causes property bubbles and bursts to form global waves: recessions are global chiefly because property bubbles are global.
In the event of any dispute about the data on which these conclusions are based, it will be pertinent to note the use of conservative words such as indicator, usually, minority, and most. In view of the large amount of data and the eclectic range of sources, it is highly likely that there are a few errors in the data, but highly unlikely that there are enough errors to invalidate conclusions expressed in such qualified language.
6. Policy implications
To the extent that recessions are caused by bursting property bubbles, they can be prevented by preventing property bubbles. In particular, they can be prevented by policies that stabilize the growth of land values around the long-term trend. For this purpose it does not matter if the long-term trend becomes faster, as long as wild fluctuations around that trend are eliminated.
Wild swings in land values occur because of contradictory price signals. When a bubble is inflating, high prices by themselves deter buyers; but rising prices attract buyers by (seemingly) promising “capital gains”, negating the correction that should be provided by high prices. After the bubble bursts, low prices by themselves attract buyers, but falling prices deter buyers by threatening “capital losses”, negating the correction that should be provided by low prices. So the market overshoots in both directions. To prevent this, one must attenuate or counteract the signal sent by the rate of change of prices.
One solution is to reduce taxes on productive activities and increase taxes on unearned increases in land values. As noted on the home page, this policy would increase the long-term rate of growth in land values by enabling and encouraging governments to do things that increase land values — such as investing in infrastructure. This of course would be highly beneficial to property owners. But the same policy would reduce the attractiveness of “capital gains” (which would be taxed more) relative to income from assets (which would be taxed less), and thereby attenuate the effect of the rate of change of prices, as required.
The stabilization of growth in land values would be particularly effective if the taxation of unearned “capital gains” were accomplished by means of a periodic tax (i.e. a “holding tax”) proportional to the increase in the land value since a certain reference date. Then rising prices would mean rising holding costs, which would deter purchases and encourage sales, while falling prices would mean falling holding costs, which would encourage purchases and deter sales. Thus the signal sent by the holding tax, like that sent by the actual prices, and unlike that sent by the rate of change of prices, would be in such a direction as to stabilize the market.
This is probably not the first global recession in which most of the affected countries have independently taken the same road to destruction. But it is surely the first in which a researcher of modest means can discover or verify the recurring pattern without leaving the office or, as the case may be, the bedroom: this article merely identifies patterns in existing data gleaned from various online sources.
The most important source was Global Property Guide, especially the “House Prices” index page, from which one can follow the links to graphs and articles on the housing markets of numerous countries, including nominal prices, year-on-year changes in real and nominal prices, rental yields and quarterly price changes for most of the countries covered, and sometimes other information such as turnover or loan approvals.
The second-most important source was Investment Property Databank (“IPD”), especially the “Indices” page, from which one can find graphs of income and “capital growth” in each country for various classes of property, including “all property”. Most of the figures are yearly; some are more frequent. Email updates are available.
Other sources of property data are individually acknowledged by links in the text.
GDP figures, being widely reported in the media, are easy to find via search engines or news aggregators. But one can save time by consulting compilations of quarterly figures for multiple countries, e.g. at Eurostat and at OECD Stat Extracts. Note that official statistics are subject to revision.
* Sources marked thus, most of which concern market turnover or proxies thereto, were located by Mr Anh Pham, BBus MAppFin (Monash), an Earthsharing Australia volunteer. Here again it should be noted that information on turnover was more elusive than information on prices.
[Last modified June 2, 2009.]