Mobile Edition
Print
Get the Mag
Weekly UpdatesSome markets have a negative effect on other markets when positive changes exist in the movement trend in the market. Table 4 is generated from the results of the eight cities' impulse responses. It shows the eight cities' total impulse responses within 20 lagged quarters. The bold numbers show a negative value of impulse response. The Darwin housing market exhibits a reversed impact on other markets, especially in the long run. However, all the markets (including Darwin itself) impose a positive influence on Darwin, except Sydney. Therefore Darwin is not one of the main engines but would play the role of a buffer in the aggregate Australian market during a price boom.
Except Darwin, the other seven markets have at least one negative influential factor. The Melbourne housing market has three negative external influential factors. Adelaide, Brisbane, Perth, Melbourne and Sydney, present an acceleration effect on the aggregate Australian market in the long run. These five markets can be regarded as the secondary level epicentres. The numbers on diagonal line in Table 4 describe the impulse response of each market to the innovation of itself. These numbers are the largest one in each row. The results suggest that the most important factor to each regional housing market is from its own individual performance. The values of response to innovation perform kinds of behaviour which first converges to zero (not exactly at zero) and then swinging around zero in the long run. The speed of convergence to zero can scale the sensitivity of one market to the influence from other markets. In this study, 0.05 and 0.01 of absolute value are set up as two standards to measure this speed.
6.2. Lagged effect of regional house price diffusion
Table 5 shows the numbers of lagged terms when first reaching a value of impulse response of less than 0.05. Most of the numbers in Canberra columns are greater than the numbers located in the same row. This indicates that the influence from Canberra on other markets will persist over a longer period. This proves again that the Canberra housing markets are two important factors in the aggregate Australian housing market. The smaller number of the lagged terms indicates that the speed is larger. This convergence speeds in the Sydney market and Melbourne market are from 11 to 4 and 10 to 4 respectively, while Adelaide is from 13 to 6, Brisbane is from 14 to 8 and Perth is from 15 to 7. This suggests that the Sydney and the Melbourne markets are safer than others. The impacts, either from themselves or from exogenous markets, can not persist for a long time. However, the markets of Adelaide, Brisbane and Perth are more sensitive to the change in external markets. Furthermore, the numbers of lagged terms shown on the diagonal line in Table 5 explains the duration of the time interval by each market is affected itself. Except for Brisbane and Perth, these numbers are not always the smallest or largest in each row. The number for Brisbane in the row "response of Brisbane to" is the smallest one. It indicates that the impacts of exogenous markets exist longer in Brisbane than the impact from the home market. The number for Perth in the row "response of Perth to" is the largest one. It indicates that the Perth market is more sensitive to itself than changes in external markets.
Figure 3 shows the numbers of lagged term when first reaching a value of impulse response of less than 0.05. There are eight octagons with the same centre (0 of lagged term) in Figure 3. Each octagon shows the sensitivity of one market to the others. The greater the area of the octagon is, the more sensitive the market is. There are eight semidiameter lines with eight intersection dots (on each line) where the eight octagons cross through each line. Each line stands for the duration of the time interval that one market affects the others (including influence from itself). The eight intersection dots on each line indicate eight time intervals at which each market affects the others. Figure 3 shows the octagons of Melbourne and Sydney are the smallest and the eight intersection dots on the lines of Melbourne and Sydney are closer to the centre (zero) than the others. The results show that the convergence speeds of Melbourne and Sydney are larger than the others. It demonstrates that the Melbourne and Sydney housing markets can absorb the shock from other markets more efficiently. The octagon for Darwin is the largest, which indicates that the Darwin market is the most inefficient.
[FIGURE 3 OMITTED]
The eight lagged terms of convergence speed indicates eight quarters. It does not support the notion in previous research that housing markets are inefficient. However, it can measure the sensitivity of response to the innovation of external markets. Podlodowski and Ray (1997) estimated the significance of the vector autoregression (VAR) model constructed with contiguous regions or noncontiguous regions, to test housing market efficiency. If the VAR model with a small number of lag order such as VAR(1) or VAR(2) is significant, then the market is efficient. Podlodowski and Ray suggested the market is inefficient. In our study, the VAR(1), VAR(2), VAR(3) and VAR(4) all constructed with time series of level form fail to satisfy the stable condition of VAR. In the context of Podlodowski and Ray, the market should be inefficient. However, the VAR(1) constructed with first difference of the data satisfied the stable condition and the impulse response function results show that most of the largest response values from external market shock occurred at the first or the second lagged term. It means the biggest reaction of one market to other markets' shocks performed very quickly. Therefore, the results suggest the housing market is efficient in the spatial dimension.
Table 6 shows the number of the lagged term when first reaching a value of impulse response of less than 0.01. Similar findings were detected as above.
Figure 4 is generated from Table 6. It displays the number of the lagged term when first reaching a value of impulse response of less than 0.01. Eight expanding octagons are found in the figure. It demonstrates similar finding as in Figure 3. The differences between Tables 5 and 6 show that the speed of convergence to zero slows down when the value of impulse response is getting close to zero.
[FIGURE 4 OMITTED]
7. CONCLUSIONS
This study first estimated the dynamics of house price diffusion within Australia's state capital cities. Using the impulse response function deriving from a VAR(1) model, this study examined the response of one market to the innovation of other markets and determined the lagged terms for the maximum absolute value of the response, from the December quarter 1989 to the June quarter 2007. The findings highlight a number of issues which are summarised below.
Numerical results of this research indicated that house price diffusion exists in all capital cities of Australia. The impulse response results suggest that Canberra and Hobart are the two main epicentres in the Australian housing market. Canberra is the key engine of the area of Adelaide, Brisbane, Canberra, Melbourne and Sydney, while Hobart is another key engine of the area of Hobart and Perth. Darwin played the role of a buffer in the latest housing boom. The other five housing markets in Adelaide, Brisbane, Melbourne, Perth and Sydney would be regarded as having secondary level impetus in Australian housing market.
The impulse responses of eight state capital cities in Australia were found to converge to zero with various speeds. The speed of con vergence to zero suggests that Melbourne and Sydney are safer markets than other markets while Adelaide, Brisbane and Perth are more sensitive to the changes from external markets. The results also suggest that the Australian housing market is efficient; and this influence from other markets would last in a long term.
Received 3 July 2008; accepted 2 October 2008
REFERENCES
Abelson, P. and Chung, D. (2004) Housing prices in Australia: 1970 to 2003. Department of Economics, Macquarie University, Sydney.
Abelson, P., Joyeux, R., Milunovich, G. and Chung, D. (2005) Explaining house price in Australia: 1970-2003, Economic Record, 81(s1), pp. S96-S103.
ABS (2008) House price indexes: eight capital cities. [Online] Australian Bureau of Statistics. Available at: www.abs.gov.au/ausstats/abs@. nsf/productsbytitle/A50B6B8CF85FO474CA2 5722900179E3F?OpenDocument [accessed 13 September 2008]
Alexander, C. and Barrow, M. (1994) Seasonality and cointegration of regional house prices in the UK, Urban Studies, 31(10), pp. 1667-1689.
Ashworth, J. and Parker, S. C. (1997) Modelling regional house prices in the UK, Scottish Journal of Political Economy, 44(3), pp. 225-246.
Asteriou, D. (2005) Applied Econometrics: A Modern Approach Using Eviews and Microfit. Palgrave Macmillan, New York.
Chen, M. C. and Patel, K. (1998) House price dynamics and Granger causality: An analysis of Taipei new dwelling market, Journal of the Asian Real Estate Society, 1(1), pp. 101-126.
Clapp, J. M. and Tirtiroglu, D. (1994) Positive feedback trading and diffusion of asset price change: Evidence from housing transaction, Journal of Econometric Behaviour and Organization, 24(3), pp. 337-355.
Cook, S. (2005) Detecting long-run relationships in regional house prices in the UK, International Review of Applied Economics, 19(1), pp. 107-118.
DeJong, D. N., Nankervis, J.C., Savin, N.E. and Whiteman, C.H. (1992) The power problems of unit root test in time series with autoregressive errors, Journal of Econometrics, 53(1-3), pp. 323-343.
Dicky, D. A. and Fuller, W.A. (1979) Distribution of the estimators for autoregressive time series with a unit root, Journal of the American Statistical Association, 74(336), pp. 427-431.
FEED