IN-CONFIDENCE
Treasury Report: Macroeconomic Effects of Housing Market
Executive Summary
You requested a report on the macroeconomic effects of housing. This report draws
together material to inform you about the different drivers influencing house prices and the
transmission mechanism of house prices through the economy with a particular focus on
interest rates and the exchange rate.
House price drivers
There are a range of both structural and cyclical factors that have been driving house prices
over the past 25 years. The key structural demand factors include household income growth,
a reduction in mortgage interest rates and financial deregulation. On the supply side, the
growth in land prices, increases in construction costs and declining productivity in the
construction sector have contributed to house price growth. The combination of these factors
suggests that house prices have experienced a permanent shift up.
Among the cyclical drivers, strong net permanent and long-term (PLT) international arrivals
(both New Zealanders and non-New Zealanders), ample global liquidity and easy credit
conditions drove up housing demand and house prices beyond the levels justified by pure
structural drivers during 2000-2007.
The interaction between the factors, particularly the slow adjustment of housing supply to
positive housing demand shocks amplified the house price cycles. A study estimated that it
took about eight years for housing supply to catch up with the higher housing demand from
migration inflows during 2000-2006. Further, the shortage in housing supply fuelled
expectations of higher house prices in future. It is likely these expectations played a role in
sustaining the house price boom during 2004-2007 even when mortgage interest rates
increased markedly in New Zealand.
Transmission of house prices through to the economy
The structural and cyclical factors discussed above would have had a direct positive effect on
overall economic activity. In addition, they are likely to have had an indirect effect via house
prices through two possible channels:
•
the wealth effect increasing consumption, and
•
additional residential investment.
Previous work suggested that the wealth effect was an important channel but more recent
work casts some doubt on the strength of that link. The residential investment channel is
strong and has been supporting economic activity in New Zealand in recent years. This
raises a potential tension: a more responsive housing supply reduces pressures on house
prices but simultaneously can build aggregate demand pressures through higher residential
investment.
Overall, the indirect effect of high house prices on economic activity is smaller than the direct
effects of the fundamental house price drivers. However, the indirect effects will add to
economic impulse and therefore will be a factor in decisions around monetary policy settings.
T2015/226 : Macroeconomic Effects of Housing Market
Page 2
IN-CONFIDENCE
IN-CONFIDENCE
A more responsive housing supply, particularly if achieved through an increase in
construction sector productivity, would ease pressure on aggregate demand. Smaller peak of
net PLT arrivals would also reduce pressure on consumption and residential investment.
Neither of these conditions, however, are completely determined by policy settings.
Effect of house price falls
The impact of a house price fall is dependent to some extent on what causes the change. In
general, we would expect demand pressures to fall through the same channels in the event
of a house price correction. For example, initial internal modelling work suggests that a fall of
50,000 net PLT arrivals would lead to a fall of up to 30 basis points in the policy interest rate
via weaker inflationary pressures.
Financial stability
There are also links between house prices, household debt and net external liabilities.
Household borrowing has increased as buyers tend to need larger mortgages to buy houses
when house prices rise. House price growth also increases the borrowing capacity of existing
homeowners. High household debt has kept New Zealand’s external liabilities elevated as
banks have financed household borrowing through raising offshore debt. Elevated household
debt and external debt levels pose risks to macroeconomic and financial stability particularly
from sharp corrections in house prices.
Recommended Action
We recommend that you:
a
note that both structural and cyclical factors have been driving house prices over the
past 25 years
b
note that the importance of the structural factors means we would expect some of the
house price increase observed to be a permanent change. But cyclical factors, the
interaction with supply responsiveness and the role of expectations have led to house
prices that are currently higher than structural factors would predict
c
note that these house price drivers will have direct effects on aggregate demand and
economic activity more generally as well as indirect effects (via the wealth effect on
consumption and via residential investment)
d
note that it was previously thought the wealth effect was most important but more
recent work suggests that the residential investment effect may be stronger than
previously thought
e
note that the indirect effect of the residential investment channel involves opposing
forces on aggregate demand as higher investment may reduce house price pressure
but will contribute directly to aggregate demand, and
f
note that initial internal modelling work suggests that weaker inflationary pressure
associated with a fall in house prices, based on a scenario using a 50,000 drop in net
PLT arrivals, would lead to a fall of up to 30 basis points in the policy interest rate.
Tim Ng
Hon Bill English
Director, Macroeconomic and Fiscal Policy
Minister of Finance
T2015/226 : Macroeconomic Effects of Housing Market
Page 3
IN-CONFIDENCE
IN-CONFIDENCE
Treasury Report: Macroeconomic Effects of Housing Market
Purpose of Report
1.
You recently requested a report on the macroeconomic effects of housing. This report
draws together material to inform you about the different drivers influencing house
prices and the transmission mechanism of house prices through the economy with a
particular focus on interest rates and the exchange rate.
Analysis
2.
The housing market can have three inter-related effects on the macroeconomy. House
price cycles can materially influence economic cycles, interest rates and the exchange
rate through its impact on consumption and investment decisions. Large house price
cycles can also incentivise banks and households to over-leverage towards the
housing market, increasing risks to financial and macroeconomic stability when house
prices do adjust abruptly. Finally, large house price cycles can have welfare effects by
redistributing housing wealth across cohorts. Our report focuses primarily on the first
two of these concerns.
3.
House prices have almost persistently risen since the early 2000s. In this context, we
examine the drivers behind the increase in house prices and the channels through
which house prices impact overall demand in the economy. We also examine
macroeconomic effects of a fall in house prices under two alternative scenarios: a fall in
net arrivals of permanent long-term migrants and a fall in global house prices as a
proxy for tightening global liquidity conditions. The links between house prices,
household debt and net external liabilities are also discussed.
House Price Cycles
4. The last house price cycle was exceptional in scale and length in New Zealand. Real
house prices (adjusted for CPI inflation) increased rapidly (90 percent) between 2000
and 2007, and the appreciation lasted for six and a half years – more than in any of the
previous cycles. After a temporary dip due to the Global Financial Crisis (GFC), the
current house price cycle that began at the end of 2011 is still ongoing. The
introduction of macro-prudential, monetary and housing supply measures from the last
quarter of 2013 have moderated the high growth in real house prices and contained the
appreciation in the current cycle below the intensity of the upturns in the 1990s and in
the early 2000s (Figures 1 and 2). Nonetheless, house prices continue to remain high
relative to the historical average since 1980.1
1 Notably, house price increases in some regions (Auckland during 2000-2007 and in both Auckland
and Christchurch after 2010) were far more pronounced compared to in the other regions mainly due
to shocks (net migration in Auckland and earthquakes in Christchurch).
T2015/226 : Macroeconomic Effects of Housing Market
Page 4
IN-CONFIDENCE
IN-CONFIDENCE
Figure 1: Index of real house prices
Figure 2: Real house price cycles
200
350
190
300
180
170
250
160
150
200
140
150
130
120
100
110
50
100
1
2
3
4
5
6
7
8
9
10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27
Quarter from the beginning of the cycle
0
1981Q2-1982Q2
1992Q2-1997Q2
2001Q1-2007Q2
2011Q3-2014Q2
1980
1984
1988
1992
1996
2000
2004
2008
2012
Sources: RBNZ, REINZ
Sources: RBNZ, REINZ
House price drivers
Structural Drivers
Rising incomes and increasing borrowing capacity of households have been pushing up
housing demand since the late 1980s
5.
During 2002, the average level of house prices was equal to the long-term average
between 1980 and 2014. The average level of house prices over 2014 was more than
twice the long-term average level. Both structural and cyclical factors have been
pushing up house prices over the last 25 years. Household income growth, falling
interest rates and financial deregulation have been found to be among the key
structural drivers of housing demand and the trend level of house prices since the late
1980s in the various studies.
6.
Housing demand has been growing strongly alongside a steady rise in household
disposable incomes (averaging 5 percent per year) since 1994. Growing incomes tend
to improve housing affordability. However, an empirical study based on an affordability
model found the trend house prices to be growing faster than household incomes
(Briggs and Ng, 2009). Over and above the income factor, the study found that the fall
in mortgage rates since the late 1980s increased the ‘borrowing capacity’ of
households and helped them to finance housing purchases at prices that exceeded the
growth in their incomes. A decline in CPI inflation since the late 1980s reinforced the
increase in borrowing capacity as effective mortgage rates reached a low by 2004. The
results of the study suggested that the increase in household income accounted for
around half of the increase in the trend house price to household income ratio between
1987 and 2008. The other half was accounted for by the fall in mortgage interest rates.
7.
Other structural factors that also seemed to be secularly pushing up housing demand
since the late 1980s include increased households’ access to credit following financial
deregulation, a higher propensity to borrow on the part of households and a steady
population growth (averaging one percent per year since 1991). Banks increased the
standard term for a table mortgage from 25 to 30 years and also offered interest only
mortgage loans. These developments helped in expanding the borrowing capacity of
households. Furthermore, the long-term trend of a falling persons-per-household ratio
also led to a rise in house price-to-income ratio because the average household
income has fallen with fewer income earners per household. Accordingly, housing
credit increased, and the share of mortgage payments in household income reached its
peak by 2007. Increased incomes, and the greater availability and cheaper cost of
T2015/226 : Macroeconomic Effects of Housing Market
Page 5
IN-CONFIDENCE
IN-CONFIDENCE
credit have acted as structural drivers pushing up housing demand and trend house
prices (Figures 3 and 4).
Figure 3: House prices and housing Figure 4: House prices, income and
demand determinants
mortgage payments
Index: March 1999=100
% of gross household income
Ratio
60
7
400
350
6
50
300
5
40
250
4
200
30
150
3
100
20
2
50
10
1
0
1999
2001
2003
2005
2007
2009
2011
2013
0
0
Household disposable income
House price index
Number of households
Housing credit
1999
2001
2003
2005
2007
2009
2011
2013
Average mortgage payment (left axis)
Median house price to median household income (right axis)
Sources: RBNZ, REINZ, Statistics New Sources: RBNZ, REINZ, Statistics New
Zealand
Zealand
Rising land prices and residential construction prices have been pushing up costs of housing
supply particularly since early 2000s
8.
Land prices have been growing over time in New Zealand, reflecting a sustained
shortage of residential sections, particularly in places where the population wants to
live. Residential section prices and construction prices have been growing strongly
since 2003 (Briggs and Ng,
op cit) (Figure 5). This is evidenced by the rise in the
implicit residential investment GDP deflator which has been higher than the other
construction deflators (Figure 6). This could be because buyers are offering a premium
on houses that are built to suit their preferences (bespoke/tailored houses) (NZIER,
2014). Studies show that productivity in the construction sector has been weak since
the 1990s due to a combination of factors – a low supply of large scale new land,
training and skill levels in the industry, and consumer preferences for individualised
construction (DPMC, 2008). Releasing constraints on land availability and increasing
productivity in the construction sector can help in limiting escalations in the costs of
supply of housing over time (Glaeser, Gyourko and Saiz, 2008).
Figure 5: Price indices of land and Figure 6: Implicit GDP deflators for
houses
construction groups
Index: March1993=100
Index: March 1993=100
500
250
450
400
200
350
300
150
250
200
100
150
100
50
50
0
0
1993
1995
1997
1999
2001
2003
2005
2007
2009
2011
2013
1993
1995
1997
1999
2001
2003
2005
2007
2009
2011
2013
Median house sale price
Median residential section price
Residential buildings
Non residential buildings
Other construction
Source: REINZ
Source: Statistics New Zealand
Cyclical Drivers
9.
The affordability model-based study discussed above (see paragraph 6) found the
actual value of house prices to be around 20 percent higher than the estimated trend
level in 2008. The estimated short-run equation in the study indicated that the gap
T2015/226 : Macroeconomic Effects of Housing Market
Page 6
IN-CONFIDENCE
IN-CONFIDENCE
between the actual house price level and the trend value was slow to close, thereby
suggesting that house price cycles – once they get underway – tend to be large and
long lasting. A large proportion of house price inflation in one quarter was found to
carry forward into the next suggesting that past house price inflation played a
significant role in shaping household expectations of future house price inflation. Other
factors accounting for the cyclical variation in house prices as per the study included
changes in population due to migration, housing credit cycles and the responsiveness
of housing supply.
Access to ample global liquidity contributed to housing credit and house price cycles in
New Zealand
10. New Zealand’s house price cycles during the last decade synchronised with similar
house price cycles experienced in other advanced economies. The OECD countries
experienced unprecedented house prices cycles in terms of magnitude and duration.
More than half of a sample of 18 OECD countries experienced house price increases
of more than 100 percent between 1995 and 2007 (André, 2010). After a modest
correction during the GFC, house prices have stabilised/started increasing in many of
these countries. While growing household incomes drove house prices, strong credit
growth played a key role in amplifying housing demand and house prices to
unprecedented levels in all these economies (financial accelerator channel).
11. The global factors contributing to high housing credit growth included a ‘glut’ of savings
supplied to world credit markets by Asian and oil exporting nations, the US-led low
world interest rates, rapid growth in financial innovation and progressive relaxation of
credit standards by banks in the Northern Hemisphere (Briggs and Ng,
op cit). While
banks in New Zealand remained sound, growth in real house prices remained more
pronounced than the OECD average during the house price cycle prior to the GFC
(Figure 7).
Figure 7: Real House Prices
* – New Zealand versus OECD countries
annual change (%)
25.0
20.0
15.0
10.0
5.0
0.0
-5.0
-10.0
-15.0
1998
2000
2002
2004
2006
2008
2010
2012
Australia
New Zealand
Total OECD
United States
* Nominal house prices deflated by private consumption deflator
Source: OECD Economic Outlook (Volume 2014, Issue 2), November 2014
12. Surplus global liquidity spilled over into New Zealand, which eased credit conditions
and acted as a major driver of sharp house price increases during 2000-2007. Figures
8 and 9 proxy surplus global liquidity through global house prices and show how New
T2015/226 : Macroeconomic Effects of Housing Market
Page 7
IN-CONFIDENCE
IN-CONFIDENCE
Zealand house prices caught up and exceeded the global house price boom during the
2000s after the economy recovered from the impacts of drought and the Asian
Financial Crisis during the late 1990s. The availability of foreign savings helped to fund
housing credit and drove house price booms beyond the levels that could be justified
from financial deregulation, lower interest rates and income growth. The borrowing
capacity of households increased significantly which encouraged some to ‘trade up’
their dwelling by buying a better quality and larger house. This added to housing
demand and lifted house prices (DPMC, 2008).
Figure 8: Global and New Zealand Figure 9: Contributions to House Price
House Prices and Net Migration
Growth
Quarterly % change
250
Net Migration (right axis)
14000
8
12000
Global Real House Price Index (exc.NZ)(left axis)
6
200
10000
NZ Real House Price Index (left axis)
4
8000
2
150
6000
4000
0
100
2000
-2
0
-4
50
-2000
Net Migration (Estimated)
-4000
-6
Global Real House Price (Estimated)
Actual House Prices (qpc)
0
-6000
-8
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
3
3
3
3
3
3
3
3
3
3
3
3
3
3
3
3
3
3
3
3
3
:Q
:Q
:Q
:Q
:Q
:Q
:Q
:Q
:Q
:Q
:Q
:Q
:Q
:Q
:Q
:Q
:Q
:Q
:Q
:Q
Q
Q
Q
Q
Q
Q
Q
Q
Q
Q
Q
Q
Q
Q
Q
Q
Q
Q
Q
Q
Q
5
7
9
1
3
5
7
9
1
3
5
7
9
1
3
5
7
9
1
3
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
7
7
7
8
8
8
8
8
9
9
9
9
9
0
0
0
0
0
1
1
9
9
9
9
9
9
9
9
0
0
0
0
0
0
0
0
0
0
1
1
1
9
9
9
9
9
9
9
9
9
9
9
9
9
0
0
0
0
0
0
0
9
9
9
9
9
9
9
9
0
0
0
0
0
0
0
0
0
0
0
0
0
1
1
1
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
2
2
2
Sources: Federal Reserve Board, Sources: Federal Reserve Board, Statistics
Statistics New Zealand
New Zealand, Treasury
Migrant inflows has contributed to cyclical pressures on housing demand
13. Increases in real house prices across OECD countries have been positively correlated
with population growth during the last two decades (André,
op.cit). In New Zealand,
while housing demand has grown alongside the steady growth rate in the population,
house price cycles have also been closely associated with permanent migrant inflows
and outflows. Real house prices declined during the late 1990s when there were
permanent migrant outflows and increased sharply during early 2000s alongside strong
migrant inflows.
14. The results from a study on housing markets over 1962-2006 in New Zealand found
positive net permanent migration shocks to be associated with large increases in house
prices in the short-term (Coleman and Langdon-Lane, 2007). The study cautioned that
the estimated impact on house prices turned out to be larger than can be explained by
purely from increases in population, and suggested that this could have been due to
the role played by other drivers. For instance, migrant inflows have occurred at times
when local residents were also increasing their housing demand owing to positive
expectations of future income growth. Further, strong migrant inflows may also be
destabilising agents’ expectations about the fundamental value of houses. Also, the
positive contribution of migration was particularly evident during the 1990-1997 and
2000-2008 episodes with the latter also being exacerbated by the surge in global
liquidity. These factors seemed to be magnifying the standard impact of migrant inflows
on house prices.
15. During the last couple of years, however, the impact of net migrant inflows on house
prices has been relatively muted. One possible explanation is that the net migrant
inflow position has increased more on account of fewer New Zealanders departing
overseas rather than arrivals of non-New Zealanders. Studies have found migrant
arrivals to have had bigger housing demand effects than departures (RBNZ, 2013). At
the same time, net migrant inflows have largely been concentrated amongst younger
age groups (less than 35 years) who tend to be less ‘capital rich’. Overall, this has
T2015/226 : Macroeconomic Effects of Housing Market
Page 8
IN-CONFIDENCE
IN-CONFIDENCE
likely to have resulted in a weaker impact on housing demand than has been observed
historically.
Slow housing supply response to rising housing demand has led to large house price
cycles
16. Housing supply typically responds to unexpected changes in housing demand with a
lag because it takes time for land to be rezoned under district plans, housing consents
to be approved, and the right mix of land, materials and labour to be assembled for
construction. While the average growth in dwelling stock per year (1.3 percent)
matched the average growth in population (1.1 percent) and the average growth in the
number of households (1.3 percent) over 1994-2014, housing supply has responded to
changes in housing demand with a lag which resulted in phases of under-shooting
(under-supply) and over-shooting (over-supply) of growth in the dwelling stock relative
to the projected growth in the number of households in the short term (Figure 10). The
growth in dwelling stock has been lower than the projected growth in the number of
households based on population estimates since 2003 which has led to large increases
in house prices. Also, higher sales of existing houses compared to housing consents
particularly during mid-2000s could have acted as an added impetus on house prices
(Figure 11).
Figure 10: Growth rates in private Figure 11: House sales and consents
dwellings and households
2.5
Index : 1993Q1=100
Over-supply
)
200
k
c 2.0
180
to
Under-supply
l s
160
itia
140
1.5
f in
o
120
(%
e
100
s
a 1.0
e
80
r
c
Under-supply
60
l in
a
u 0.5
40
n
A
20
0
0.0
1993 1995 1997 1999 2001 2003 2005 2007 2009 2011 2013
1993 1995 1997 1999 2001 2003 2005 2007 2009 2011 2013
Number of private dwellings
Number of households
House consents
House sales
Source: Statistics New Zealand
Sources: Statistics New Zealand, REINZ
Expected growth in house prices pushed up housing demand
17. Expected growth in house prices could have been pushing up housing demand in the
OECD countries (André
op cit). Home buyers in these countries have been primarily
influenced by motives of owning a house, saving on rental payments and avoiding
being priced out of the market rather than the prospects of capital gains. In New
Zealand, the landlords have been investing in property eyeing benefits of capital gain,
regular income stream, retirement investment/income, tax advantage and owning rental
property for paying off mortgage payments from rental receipts (DPMC, op cit).
18. Notably, house price expectations have been found to be backward looking as
suggested by an empirical study which found that expected growth in house prices has
been driven by the past three-year average increase in house prices in New Zealand
(Grimes and Hyland,
op cit). This suggested that expectations of future house price
growth may have played a role in sustaining housing demand and the house price
boom during 2004 to 2007 when mortgage interest rates increased. Moderating the
house price cycle through policy intervention during that period could have anchored
the agents’ expectations about future growth in house prices.
T2015/226 : Macroeconomic Effects of Housing Market
Page 9
IN-CONFIDENCE
IN-CONFIDENCE
Interaction between housing supply and housing demand
19. Housing supply and housing demand dynamics have been formally captured by a New
Zealand Regional Housing Model (NRHM) used by
Motu to model the interactions
between housing supply, housing demand, land prices and rents across all the local
territories of New Zealand over 1990-2011(Grimes and Hyland, 2013) 2 (See Figure 12
for the determinants of the four housing market variables and the directions of their
influences identified in the model). The model’s overarching conclusion was that
housing markets have been slow to adjust, such that exogenous shocks have had very
long lasting effects. The model found that an increase in population owing to a
migration surge during 2000-2006 led to a prolonged period of upward pressure on
prices (house, land and rents), which continued until the dwelling stock had adjusted to
restore dwellings per capita. The model estimates showed that it took eight years for
housing supply to catch-up with demand, and till that time house prices remained
higher than their control (counterfactual) level.
Figure 12: Determinants of housing market indicators
(+)
Rents
I. Housing demand
(-)
(+)
(or real house prices)
Regional
Dwelling stock
economic activity
(-)
(+)
per capita
(-)
(+)
Credit
Population
User costof capital
Accommodation
restrictions
(net of expected
supplement for
increase in house
own home
prices)
owners
(-)
II. Housing supply
(-)
(or dwelling stock)
Interest rates
Construction
(+)
costs
(-)
(-)
House prices
Lot prices
Credit restrictions
(+)
III.Lot prices
(+)
Population
House prices
(+)
(+)
(+)
Farm prices
Development
Finance
(+)
contributions
contributions
Construction
costs
IV. Rental
(+)
Yields
(-)
Accommodation
(+)
supplement for
renters
Interest rates
Expected increase
in house prices
2 As per the model, (i) housing demand increases when per capita dwelling stock and interest rate
(adjusted for capital gains) decrease and accommodation supplement and per capita regional income
increase. Based on profitability considerations, (ii) new construction supply responds positively to
increasing house prices, but is negatively affected by rises in residential lot prices, (iii) residential lot
prices increase with increases in farm prices, development contributions paid to local council, house
prices and population pressures on land, and rents are determined such that rental yield (ratio to rents
to house prices) equals financial market yield adjusted for expected capital gains on rental property.
T2015/226 : Macroeconomic Effects of Housing Market
Page 10
IN-CONFIDENCE
IN-CONFIDENCE
20. House prices fell during 2008 and 2009 driven by tight global liquidity and credit
restrictions. The results from
Motu’s regional housing model study showed that credit
restrictions implemented by banks after the global financial crisis in 2008 reduced
housing demand and house prices, but the fall in house prices was found to be
temporary as a lower availability of credit to builders reduced housing supply and
raised house prices again after a period of four years.
21. House prices resumed their growth from 2012 reflecting a strong increase in housing
demand particularly in Auckland on account of large migrant inflows and historically low
interest rates. The shortage in housing supply in Christchurch following the
earthquakes also fuelled house price growth. The imposition of a speed limit on high
loan-to-value restrictions (LVR) on housing credit from banks, monetary tightening and
increases in housing supply worked in slowing down house price growth over 2014
despite record increases in migrant inflows. The Reserve Bank of New Zealand
(RBNZ) exempted new construction loans from the high LVR restrictions, which helped
to reduce the adverse impacts of the restrictions on housing supply.
22. House price inflation (from 9.4% to 5.0%) and housing credit growth (from 5.8% to
4.7%) have eased between September 2013 and September 2014, and have now
moved closer to the average growth rate in household incomes (4.0%) over the past
five years. The RBNZ, however, has decided to continue with its high LVR restrictions
in light of the continued strong inflows of migrants, muted increase in mortgage interest
rates and continued lag in the response of housing supply to demand, particularly in
Auckland (RBNZ, 2014). Given the ongoing shortfall in housing supply, house price
inflation is likely to increase again particularly in Auckland. The RBNZ would continue
to monitor the growth in house prices vis-a-vis the financial capacity of the households,
which is related to growth in their incomes.
Transmission of house prices through to the economy
Rising house prices have boosted demand pressures in the economy
23. House price cycles have closely correlated with economic cycles in New Zealand.
Growth in house prices peaked in 2003 and troughed in 2008, slightly ahead of the
broader economic growth upturn and downturn in 2005 and 2010 respectively (Figure
13). The transmission of housing market disturbances to the broader economy has
been subject to some debate, with empirical literature identifying three possible
channels:
•
Wealth effects: rising house prices appreciate the value of housing assets and
lead home owners to increase consumption as they feel wealthier,
•
Liquidity effects: rising housing wealth eases liquidity constraints and enables
home owners to withdraw their housing equity to consume more (or reinvest in
housing), and
•
Investment effects: rising house prices relative to costs of building houses
boost profitability
prospects on construction activity and lead to higher
residential investment.
T2015/226 : Macroeconomic Effects of Housing Market
Page 11
IN-CONFIDENCE
IN-CONFIDENCE
Figure 13: House Prices and Gross National Expenditure
annual % change
average annual change (%)
10.0
25
20
8.0
15
6.0
10
4.0
5
2.0
0
0.0
-5
1994
1996
1998
2000
2002
2004
2006
2008
2010
2012
2014
-2.0
-10
Gross National Expenditure (nominal)-left axis
House Prices Index (Stratified)-right axis
Sources: Statistics New Zealand, REINZ
24. In a study of OECD countries, rising house prices were found to have significant effects
on private consumption through the wealth and liquidity channels (Girouard and
Blondal, 2001). The Supplementary Stabilisation Instruments report in 2006 (prepared
by the Reserve Bank of New Zealand and the Treasury) concluded that house price
increases and expectations of future increases occurred alongside a rise in private
consumption which suggested the importance of wealth and liquidity effects in New
Zealand. The increase in house prices during the last decade seemed to have
effectively eased credit constraints allowing households to withdraw housing equity to
increase their consumption (Briggs, 2007). Recent thinking has been somewhat
sceptical about the materiality of the wealth effects.
25. Present value models suggest that house prices are anchored to the growth of rents
over the long run. By this relative metric, house prices remained significantly
overvalued between 2003 and 2008 (Figures 14 and 15). IMF estimates that took into
account income, demographics and interest rates suggested an overvaluation of 15-25
percent in New Zealand’s house prices (Igan and Loungani, 2010). Despite significantly
overvalued house prices, recent studies have found little evidence to suggest sustained
effects from higher house prices on household consumption through the wealth
channel at the macro level. While increase in house prices may positively impact an
individual’s consumption, wealth effects across households tend to cancel each other
out and the net effect on aggregate wealth of household sector as a whole could be
minimal (Buiter, 2008).
T2015/226 : Macroeconomic Effects of Housing Market
Page 12
IN-CONFIDENCE
IN-CONFIDENCE
Figure 14: House prices-to-rent ratio
Figure 15: House prices growth and rental
yields
170
0.20
160
0.15
)
n
150
ia
d
II. Mega house price boom
0.10
e
0
III. Partial correction
0 140
t (m
1
n
=
0.05
e
1
ts
r
Q 130
in
/
o
e
3
p
9
e
ic
0.00
g
r
9 120
ta
n
p
: 1
e
x
rc
le
e
e
a
-0.05
d 110
P
s
e
In
s
u
o
100
-0.10
I Moderate increase
H
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
Q
Q
Q
Q
Q
Q
Q
Q
Q
Q
Q
Q
Q
Q
Q
Q
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
9
0
0
0
0
0
0
0
0
0
0
1
1
1
1
1
90
9
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
1
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
Rental yield less floating mortgage rate
Rental yield less 2-year fixed mortgage rate
80
Past three-year average house price growth
Equilibrium expected house price growth*
1993 1995 1997 1999 2001 2003 2005 2007 2009 2011 2013
Quarters
* Implied from rental yield less 2-year fixed mortage rates
Sources: REINZ, MBIE
Sources: REINZ, MBIE and RBNZ
26. A study found that a one percent increase in real house prices to be associated with an
increase in real per capita consumption in the following quarter of 0.07 to 0.1 percent
and a total increase of 0.2 to 0.3 percent in the long-run in New Zealand (De Veirman
and Dunstan, 2008). Another study using more micro data found the impact on
consumption to be larger for older households (Smith, 2010). However, at the macro
level, the share of household consumption in GDP remained stable compared to the
strong increase in the share of residential investment in GDP during the house price
boom years from 2002 to 2008. Therefore, a more recent study suggested that house
prices may not have played a large part in explaining household consumption
behaviour at the macro level (De Veirman and Reddell, 2011). On the other hand, the
stronger increase in the share of residential investment in GDP suggested that house
prices may have been impacting aggregate demand more prominently through the
investment channel (Figure 16).
Figure 16: Household Final Consumption and Residential Investment
% of GDP
% of GDP
60
8.0
7.0
59
6.0
58
5.0
57
4.0
56
3.0
55
2.0
Household consumption (left axis)
Residential investment (right axis)
54
1.0
53
0.0
1987
1989
1991
1993
1995
1997
1999
2001
2003
2005
2007
2009
2011
2013
March years
Source: Statistics New Zealand
T2015/226 : Macroeconomic Effects of Housing Market
Page 13
IN-CONFIDENCE
IN-CONFIDENCE
27. Regardless of the relative importance of the different channels, the exaggerated
increases in house prices prior to 2008 seemed to have accentuated overall demand
pressures thereby keeping interest rates and the exchange rate at levels higher than
they would have been otherwise (DPMC, 2008). Lower interest rates, compared with
New Zealand’s historical experience, particularly from 2001 to 2004 contributed to lifts
in house prices. In turn, higher house prices encouraged household spending and
additional construction that lifted inflationary pressures and eventually led to higher
interest rates during 2004-2007. This put upward pressures on the exchange rate.
Direct impacts of housing supply and housing demand drivers on aggregate demand
have been found to be stronger compared to indirect effects through house prices
28. The nature of the influence of house prices on the business cycle depends on the
underlying housing supply and housing demand drivers. Also, the total impact of the
underlying housing market drivers includes their direct impacts on aggregate demand
alongside the indirect channel through house prices. For instance, a more responsive
housing supply reduces pressures on house prices but contributes directly to
aggregate demand pressures through higher levels of domestic residential investment.
In a similar vein, shocks from higher migration increase aggregate demand directly
through consumption alongside indirect influences (i.e., rising housing demand and
house prices). Apart from demand effects, migration also contributes to higher labour
supply.
29. In an internal Treasury modelling exercise conducted in 2011 we had examined the
direct and indirect macroeconomic impacts in terms of two indicative scenarios,
namely, (a) of a more responsive housing supply (an increase in the elasticity of
housing supply by 0.5 percentage points vis-a-vis house prices) and (b) a reduction in
peak migrant inflows (by 10,000 at the height of housing cycle) using the New Zealand
Treasury model. The results showed that the direct impacts of these shocks on
domestic demand were stronger than the indirect effects through house prices.
•
Under the first scenario3, an increase in housing supply and improved productivity in
the construction sector were found to reduce non-tradable inflation, lower interest rates
and increase private consumption directly relative to the baseline. This positive effect
on private consumption was found to be stronger than the negative effect arising from
lower levels of house prices and housing wealth as housing supply increases.
However, the overall impact on the demand side, namely, consumption, the interest
rate and exchange rate, was found to be marginal compared to the overall impact on
the supply side in terms of increases in land supply boosting construction productivity
and long-run growth.
•
Under the second scenario4, a lower net migration peak was found to drive down
domestic demand directly through lower levels of residential investment and private
consumption demand on account of fewer domestic consumers and indirectly through
a reduction in housing demand, housing wealth and house prices. Again the direct
effects on domestic demand were found to be stronger than the indirect effects.
3 In this scenario, we modelled the immediate effects of a 3.5 percent increase in house prices by
considering that in response the elasticity of supply of housing goes up by 0.5 percentage point for
one percentage point increase in house prices. The increase in supply responsiveness increased
housing construction by 1.75 percent relative to the baseline. The extra housing output was assumed
to come from higher productivity rather than extra resources, which reduced non-tradable inflation.
The policy rate was estimated to be lower by around 25 basis points, which eased pressure on the
exchange rate. The increase in housing supply was estimated to reduce house prices by 1.6 percent
below the baseline.
4 In this scenario, we assumed a reduction in peak inflows by about 10 percent of net migration at the
height of the housing cycle. This assumed that overall net migration remained unchanged but with less
migrants arriving at the peak of the cycle. A lower peak in net migration reduced residential investment
and consumption demand due to lower population growth, with the policy rate estimated to decrease
by 30 basis points.
T2015/226 : Macroeconomic Effects of Housing Market
Page 14
IN-CONFIDENCE
IN-CONFIDENCE
Notably, in this scenario overall net migration was kept unchanged although the peak
level was reduced. This implied no change on potential growth as labour supply
remained the same as in the base case.
30. The results show that in principle migration policy settings that reduce the cyclicality of
net migration inflows could produce wider benefits on macroeconomic stability by
reducing both the direct and indirect influences on aggregate demand. While flows of
New Zealanders across the border are not amenable to policy influence, more material
improvement could be achieved by avoiding abrupt changes in policy strategy that lead
to a simultaneous increase in permanent residency, student visas and temporary
workers that occurred in the early part of 2000s. Similarly, the results show that policy
settings that increase housing supply (better infrastructure planning, housing
intensification in Auckland, increasing longer-term land supply for development and
building regulations) will be important for boosting construction sector productivity and
long-term growth. Overall, policies aimed at dampening peaks of house price cycles
would help in keeping interest rates and the exchange rate lower, promoting the
competitiveness of the tradable sector in the economy.
Effect of house price falls
31. Continued high level of house prices in New Zealand poses risks to the economy if
house prices were to fall sharply due to global or domestic shocks. The impact of a
house price fall is dependent to some extent on what causes the change. In general,
we would expect demand pressure to reverse through the same channels in the event
of a house price correction. Some initial modelling work tests two scenarios: a fall in net
arrivals of permanent and long-term migrants, and a drop in global house prices as a
proxy for tightening global credit conditions. Our model results are still preliminary and
subject to a large degree of uncertainty but seem to corroborate the underlying
macroeconomic transmission discussed in the existing literature.
Scenario one – 50,000 fewer permanent and long-term migrants in a year
32. In this scenario, we found that a fall of 50,000 net arrivals of permanent and long-term
migrants in a year (ignoring compositional effects) was associated with a decrease in
domestic demand through weaker consumption and more prominently through a fall in
residential investment by the end of the first year. As a result, real GDP was found to
be lower in comparison to the baseline. The OCR was estimated to be about 30 basis
points lower by the end of the first year in response to weaker inflationary pressures,
which drove down the exchange rate. Our house price model found that the fall in
migrant inflows was associated with a significantly lower house price level relative to
the baseline. This could be reflective of the impact of other drivers of house prices (as
suggested by Coleman and Langdon-Lane, 2007).
Scenario two – Tight global credit conditions
33. In this scenario, we looked at the impact of tight global credit conditions on New
Zealand’s house prices. Global house prices act as a close proxy of global liquidity
conditions. Tighter global liquidity drives down global house prices and tend to increase
capital outflows from New Zealand thereby tightening domestic liquidity, which in turn
drives down domestic house prices. Accordingly, the results from our housing model
showed that domestic house prices fell in response to a fall in global house prices,
though the magnitude of the fall was found to be lower than the fall in global house
prices.
T2015/226 : Macroeconomic Effects of Housing Market
Page 15
IN-CONFIDENCE
IN-CONFIDENCE
Financial stability
House prices, household debt and a build up of net foreign liabilities
34. Household debt has increased in New Zealand with the rise in house prices since the
early 2000s primarily because of two factors. First, household borrowing has increased
as buyers tend to need larger mortgages to buy houses when house prices rise.
Second, house price growth also increases the borrowing capacity of existing
homeowners owing to the rapid accumulation of household wealth. Household net
wealth more than doubled over the period from 2000 to 2007, with the share of housing
therein rising from 82 percent to 97 percent. Banks eased lending standards during the
housing boom years which further added to the actual borrowing capacity of the
households (De Veirman and Reddell, op cit). As a result, the household sector’s debt
as a share of nominal disposable income rose from 113 percent to 159 percent over
the period from 2002Q1 to 2007Q4. After the GFC, the share declined to 151 percent
in 2012Q1 but increased thereafter to reach 156 percent by 2014Q3.
35. High household debt has kept New Zealand’s external liabilities high as banks have
intermediated to finance household borrowing through raising offshore debt. The ratio
of housing loans in GDP increased sharply between 2000 and 2009 and led to a fall in
household sector’s net financial wealth, which has been one of the major drivers of
New Zealand’s high level of net external liabilities (Figures 17 and 18). The ratio of
housing loans in GDP has come down after 2009, thereby leading to an increase in net
financial wealth of the household sector. The NIIP has also improved after 2009.
Nevertheless, elevated household debt and external debt levels continue to pose risks
to macroeconomic and financial stability particularly from sharp corrections in house
prices (André, 2011).
Figure 17: Housing debt and household Figure 18: Net International Investment
sector net financial wealth
Position
% of GDP
% of GDP
100
-60
90
80
-65
70
60
-70
50
-75
40
30
-80
20
-85
10
0
-90
2001
2003
2005
2007
2009
2011
2013
2001
2003
2005
2007
2009
2011
2013
Housing Loans
Household net financial wealth
NIIP
NIIP (excluding Canterbury reinsurance assets)
Source: Reserve Bank of New Zealand
Source: Statistics New Zealand
Concluding remarks
36. We find that high house prices in New Zealand have been driven by both structural and
cyclical factors. The trend level of house prices has been growing alongside rising
incomes and increasing borrowing capacity since the 1980s. This suggests that house
prices may remain higher than the long-term historical average even after they revert to
their trend growth. We also find that cyclical drivers such as migrant inflows, global
surplus liquidity and easy domestic credit conditions, and expected future growth in
house prices alongside a slow catch-up of housing supply to housing demand
contributed to pronounced upturns in house prices. The housing stock has been
growing at a lower rate in comparison to the number of households since 2003. This
T2015/226 : Macroeconomic Effects of Housing Market
Page 16
IN-CONFIDENCE
IN-CONFIDENCE
has been increasing the shortfall of housing supply which has kept house prices at high
levels, particularly in Auckland.
37. Housing market disturbances which increase house prices have been found to transmit
to the broader economy more prominently through an increase in residential
investment (investment channel) rather than through the wealth and liquidity channels.
In this context, it could be useful to track productivity in the construction sector. An
increase in productivity in the construction sector would help restrain pressures on
aggregate demand while also boosting long-term economic growth.
T2015/226 : Macroeconomic Effects of Housing Market
Page 17
IN-CONFIDENCE
IN-CONFIDENCE
References:
André C. (2010), ‘A bird’s eye view of housing markets’,
OECD Economics Department
Working Papers No.746, OECD, Paris.
André, J-P (2011), ‘Economic Imbalances: New Zealand’s Structural Change’,
New Zealand
Treasury Working Paper 11/03.
Briggs, P (2007), ‘Lessons from the Economics Department’s work on household balance
sheets and related issues’,
Reserve Bank of New Zealand Bulletin, Vol. 70, No. 4,
Wellington.
Briggs, P. and T. Ng, (2009), ‘Trends and cycles in New Zealand house prices’, CHRANZ
Workshop, Wellington.
Coleman, A. and J. Landon Lane (2007), ‘Housing and migration in New Zealand, 1962-
2006’,
RBNZ Discussion Paper 2007/12, Reserve Bank of New Zealand, Wellington.
De Veirman, E., and M. Reddell (2011), ‘Towards understanding what and when households
spent’,
RBNZ Bulletin, Vol. 74, No. 4, RBNZ, Wellington.
Department of Prime Minister and Cabinet (DPMC) (2008),
Final report of the House Prices
Unit: House price Increases and housing in New Zealand, Wellington.
Girouard N. and S. Blondal (2001), ‘House prices and economic activity’,
OECD Economics
Department Working Paper 279, Paris.
Glaeser, E.L., J. Gyourko and A. Saiz (2008), ‘Housing supply and housing bubbles’,
NBER
Working Paper 14193, National Bureau of Economic Research, Cambridge.
Grimes, A, and S. Hyland (2013), ‘A New Zealand Regional Housing Model’,
Motu Working
Paper 13-02, Motu Economic and Policy Research, Wellington.
Mc Donald, C (2013),
‘Migration and the housing market’,
RBNZ Analytical Note, Reserve
Bank of New Zealand, Wellington.
New Zealand Institute of Economic Research (2014), ‘Bespoke residential housing demand
and construction innovation’,
NZIER Report to BRANZ and the Building and Construction
Productivity Partnership, NZIER, Wellington.
Reserve Bank of New Zealand (2014),
Financial Stability Report, Wellington.
Smith, M., (2010) ‘Evaluation household expenditure and their relationship with house prices
at the microeconomic level’,
RBNZ Discussion Papers Series DP 2010/01, Wellington.
T2015/226 : Macroeconomic Effects of Housing Market
Page 18
IN-CONFIDENCE
