by elias oikarinen

Turun kauppakorkeakoulu Turku School of Economics

ERES Conference15-18 June, 2011, Eindhoven

The Adjustment of Housing Prices Towards the Housing Market No-Arbitrage Relation

By

Elias Oikarinen

Turun kauppakorkeakoulu Turku School of Economics

Background

• Housing market no-arbitrage relation gives the asset market equilibrium for housing prices

• In practice, long-lasting deviations from the no-arbitrage relation have been perceived in a number of countries

• The adjustment process towards the no-arbitrage relation is a central question regarding the dynamics and predictability of housing markets

of importance to households, construction companies, investors and to economic policy makers

• Nevertheless, empirical research on the adjustment towards the no-arbitrage relation is limited

Turun kauppakorkeakoulu Turku School of Economics

Aim of the Study

• To examine empirically the adjustment towards the no-arbitrage relation in the Helsinki Metropolitan Area (HMA) and in the rest of Finland

• To investigate the role of liquidity constraints in the adjustment process

• To estimate the impact of a user cost shock on the free-market housing prices, rents and supply

• To examine whether the dynamics notably differ between the regions

Turun kauppakorkeakoulu Turku School of Economics

Housing market four-quadrant model (1)

ASSET MARKET: Rent (€/m2) PROPERTY MARKET:Valuation Rent determination

P=R/u

D = S

Price (€/m2) Stock (m2)

S = C/dP=F(C)

ASSET MARKET: Construction (m2) PROPERTY MARKET:Construction Stock adjustment

Turun kauppakorkeakoulu Turku School of Economics

Asset market equilibrium – the no-arbitrage relation

E(u) = after-tax opportunity cost of capital (%) + depreciation/maintenance (%) – expected appreciation (%)

In the Finnish case, where the imputed rent is not taxed:

• Because of the notable frictions in the housing market, substantial and long-lasting deviations from the asset market equilibrium relation may emerge and the price adjustment towards the relation may be highly sluggish

Ut = Rt – TtitPt – TttPt

E(Ut) = Rt where E(Ut)= Pt E(ut)

Turun kauppakorkeakoulu Turku School of Economics

Rent (€/m2)

Price (€/m2) Stock (m2)

Construction (m2)

User cost change in the four-quadrant model

Turun kauppakorkeakoulu Turku School of Economics

• The adjustment dynamics and magnitudes after a user cost shock are of particular interest: via asset price level the shock affects supply, rental price level and the equilibrium price/rent-ratio

• The theory leaves the adjustment speeds and magnitudes open

To get information on the actual adjustment process, rigorous empirical analysis is needed

• Liquidity constraints may influence the adjustment speed

• Adjustment dynamics may differ between regions

Turun kauppakorkeakoulu Turku School of Economics

Closer look at the impact of a user cost shock

Rent (€/m2)

Price (€/m2) Stock (m2)

Construction (m2)

(e1S)

e1S

e1L

e0

Turun kauppakorkeakoulu Turku School of Economics

Adjustment paths of the equilibrium price level and the actual price level

Price

time

e1S

e1L

e0

t=0 t=1

Turun kauppakorkeakoulu Turku School of Economics

Drt = Rt –1Y + 2S

Dst = St –1P + 2CC

Econometric Model

• System of three error-correction models:

Where

(R = 0.72*Y – 2.4*S / 0.67*Y – 2.9*S)

(S = 0.23*P – 0.06*CC / 0.31*P – 0.03*CC)

Exact lag structure and variables not know a priori

Dt = ueqt / (Rt / Pt) -1, where ueq = (Rt – TtitPt – TttPt) / Pt

pt = p – pDpt-1 + ∑piθt-i + pt

rt = r – rDrt-1 + ∑rirt-i + ∑riφt-i + rt

st = s – sDst-i + ∑sist-i + ∑siΩt-i + st

Turun kauppakorkeakoulu Turku School of Economics

Potential Complication

• Comparability between the housing price and rental price series - different dwellings

• Privately finance flat market price data and square meter prices are used: diminishes the heterogeneity problem

• User cost measurement

• Expected appreciation

• Risk premium

• Liquidity constraints

• Other data complications• E.g. measurement of liquidity constraints

Turun kauppakorkeakoulu Turku School of Economics

Empirical Findings

• Asset prices appear to adjust towards the no-arbitrage relation significantly but slowly

• No evidence of asymmetric adjustment or liquidity constraints affecting the adjustment speed

• Housing price growth and supply changes are highly predictable

• Also the adjustment speeds of R and S towards the long-term equilbirium relations are low (but significant)

• Adjustment slower in HMA

• Rental price response greater in the rest of Finland

Turun kauppakorkeakoulu Turku School of Economics

Asset price does most of the adjustment in HMA

The estimated impact of a 10% increase in u on asset price level, rental price level and on housing stock, HMA

-1,60%

-1,40%

-1,20%

-1,00%

-0,80%

-0,60%

-0,40%

-0,20%

0,00%

-0,10-0,09-0,08-0,07-0,06-0,05-0,04-0,03-0,02-0,010,000,010,020,030,040,05

0 10

20

30

40

50

60

70

80

90

100

110

120

130

140

150

160

Quarters from the shock

price

rent

stock

Turun kauppakorkeakoulu Turku School of Economics

Outside HMA prices adjust less and rents more

The estimated impact of a 10% increase in u on asset price level, rental price level and on housing stock, rest of Finland

-1,80%

-1,60%

-1,40%

-1,20%

-1,00%

-0,80%

-0,60%

-0,40%

-0,20%

0,00%

-0,10

-0,08

-0,06

-0,04

-0,02

0,00

0,02

0,04

0,06

0,08

0,100 10

20

30

40

50

60

70

80

90

100

110

120

130

140

150

160

Quarters from the shock

price

rent

stock

Turun kauppakorkeakoulu Turku School of Economics

Concluding Remarks

• Theory does not give the adjustment speeds or magnitudes

• Housing prices adjust significantly but slowly towards the asset market equilbirium condition

• It appears that asset prices do the major part of the adjustment after a user cost shock in a highly supply restricted area (HMA)

• The role of rental price adjustment is notably greater in less supply restricted regions (other parts of Finland)

• The impact of changes in the tax code are more complicated: need for further research

Turun kauppakorkeakoulu Turku School of Economics

Asset market disequilibrium (€/m2, annual level) together with real housing price and rental price indices, HMA

1988 1991 1994 1997 2000 2003 2006 20094.4

4.6

4.8

5.0

5.2

5.4

-12.5

-10.0

-7.5

-5.0

-2.5

0.0

2.5

5.0

7.5

10.0

Real rental price index

Real housing price index

Disequilibrium (right scale)

Turun kauppakorkeakoulu Turku School of Economics

Computation of the user cost

• Maintenance costs from Statistics Finland

• A prediction model for expected appreciation• Prediction for nominal price growth based on an ECM (predictors:

one period lagged values of nominal housing appreciation, nominal aggregate income and of the deviation from a long-run relation between housing prices and aggregate income)

• Constant risk premium at 2% (following Himmelberg et al. 2007)

• Risk-free cost of capital is the average after-tax mortgage rate