Transcript Document

Differential Movements in the Prices
of Houses and Units
in Melbourne
Kelvin Wong
Hao Wu
University of Hong Kong
University of Melbourne
Objectives
• Many macro studies have explored different determinants of
housing price movements
• We try to ask a different question:
– Do the price of houses and units move in tandem?
– If not, what might explain their differential movements?
• In Australia, esp. Melbourne & Sydney, development density has
undergone some structural transformation
– Past 20 years: Melbourne population grew by >30%; similar growth is
expected to continue in next 20 years
– Melbourne 2030 Plan (Melbourne @ 5million): impose Urban Growth
Boundaries to promote higher density development within the city
• A study still in progress, so comments are much welcome…
Dwelling types
• Australia Bureau of Statistics
– Separate house
– Semi-attached house / terrace house / townhouse
– Flat / unit / apartment
• Our definitions based on REIV and other studies
(e.g. Kupke et al, 2011)
– House: single family, low density
• Stock: 68% (1996) down to 67% (2006)
– Unit: multi-family, medium density
• Stock: 18% (1996) up to 21% (2006)
House v Unit
Melbourne (inner)
2x
1.5x
Source: REIV
Theoretical arguments
• Houses are generally more expensive – not an argument here
• Can the price of houses grow faster than the price of units?
• Yes, the two markets are completely segmented
• No, space is perfectly substitutable
– Buyers may shift to buy units
– Unit developers may become house builders
• Reality: somewhere in between, because of …
– transaction cost (e.g. search and moving) – short-run effect
– inelastic supply (e.g. construction lag; planning control) – medium-run
effect
– Structural differences in demand or supply drivers?
Motivated from a HK study
More non-local investors (speculators),
who want liquidity
Better
location
Source: RVD
Internal demand driver?
1. Local purchasing power
• As real income increases, the demand for houses
increases more than that for units
– Move from units to houses – better quality, better
neighborhood, greater control of the property
• Early 1990s: income growth driven by economic
recovery (e.g. Tu 2000)
• Post-2000: income growth stimulated by capital inflow
External demand driver?
2. Capital inflow
• Immigration increases the demand for houses more
than that for units
– Immigration policy in favor of wealthy people who
can invest substantially in local economy
• Foreign investment increases the demand for houses
more than that for units?
– Not necessarily, because of foreign investment
restrictions on landed property (e.g. houses)
New factor from real option theory
• Clapp, Salavei & Wong (2012)
– Estimated that up to 30% of the house price can be attributed to an
option premium in the US
– The premium is larger for towns with higher redevelopment potential
and greater house price volatility
– Supply, not just demand, is also substitutable
• Implication for the house-unit price gap
– Consider two otherwise identical pieces of land, one with a house and
the other with a higher density development (units)
– The latter, having fully utilized the land, has a higher overall value (while
each unit has a lower value than the house)
– The house owner can realize the higher overall value if it is redeveloped
into units (at some cost)
– If option theory is correct, the flexibility to redevelop an existing house is
valuable, and this should be translated into a higher house price (relative
to unit price which carries no option value)
Redevelopment in Melbourne
• Evidence on redeveloping houses into units
– Planning policy and home-starter subsidies encourage urban
consolidation, leading to gentrification and inner city
intensification
– Common practice: Purchase run-down existing homes and
convert them into higher density units
– More recently, builders actively seek to buy suburban backyard
to build units (i.e. subdivision)
• Main hypothesis: A higher volatility increases the option
premium embodied in house prices, hence a larger gap
between house and unit prices
– Volatility of unit prices
– Volatility of construction costs…
A reduced-form model
log(House price / Unit price)
log real income (deflated by CPI)
log(net migration)
Log(exchange rate of
AUD against USD)
𝐺𝐴𝑃
= 𝑏0 + 𝑏1 𝐼𝑁𝐶𝑂𝑀𝐸 + 𝑏2 𝑀𝐼𝐺𝑅𝐴𝑇𝐸𝑡−1 + 𝑏3 𝐸𝑋𝑅𝐴𝑇𝐸
+ 𝑏4 𝑉𝑂𝐿𝐴𝑇 + 𝑏5 𝑆𝑈𝑃𝑃𝐿𝑌𝐻𝑆𝐸 + 𝑏6 𝑆𝑈𝑃𝑃𝐿𝑌𝑈𝑁𝑇 + 𝜀
Return volatility of property
stocks listed on ASX
Lagged supply of houses and
units
Sample & descriptive statistics
• 1993-2011, quarterly data
• Inner Melbourne
INCOME MIGRATE
(weekly, (no. of
AUD)
persons)
GAP
Mean
EXRATE
VOLAT
SUPPLY_HS SUPPLY_U
E (no. of
NIT (no. of
units)
units)
45%
704
8,798
0.727
1.0%
7,265
2,322
Std. Dev.
7%
136
5,929
0.127
0.7%
1,442
1,215
Minimum
31%
520-
1,766
0.486
0.4%
3,999
380
Maximum
64%
972
25,872
1.072
4.1%
10,702
5,998
Results for inner Melbourne
Variable
Coefficient
Std. Error
t-Statistic
Prob.
-1.438
0.265
-5.418
0.0%
INCOME
1.256
0.135
9.305
0.0%
MIGRATE
0.025
0.008
3.234
0.2%
EXRATE
0.071
0.079
0.906
36.8%
VOLAT
2.750
0.789
3.485
0.1%
SUPPLY_HSE
-0.100
0.027
-3.760
0.0%
SUPPLY_UNIT
-0.030
0.010
-3.020
0.4%
C
N: 75, Adj. R2: 75%, DW-stat: 1.7
Results for outer Melbourne
Variable
Coefficient
Std. Error
t-Statistic
Prob.
1.435
0.321
4.465
0.0%
INCOME
-0.457
0.163
-2.796
0.7%
MIGRATE
0.023
0.009
2.433
1.8%
EXRATE
-0.039
0.095
-0.405
68.6%
VOLAT
-1.663
0.955
-1.742
8.6%
SUPPLY_HSE
-0.028
0.032
-0.881
38.2%
SUPPLY_UNIT
-0.017
0.012
-1.391
16.9%
C
N: 75, Adj. R2: 33%, DW-stat: 1.8
Conclusion
• Our preliminary results support the real option
theory
– Higher volatility increases house prices more than unit
prices
– Such a relationship is significant in inner Melbourne
but not outer Melbourne
• But we need more micro-level data for further
tests:
– Suburb-level: building age, planning parameters
– Transaction-level: hedonic pricing + option value