Housing Price in Chinese Property Market - Case Study Example

 Housing Price in Chinese property market under a certain degree of Housing Bubble Table of Contents Housing Price in Chinese property market under a certain degree of Housing Bubble 1 Table of Contents 1 Chapter 3 - Theoretical Background 2 Chapter 4 - Methodology and Data Collection 6 Chapter 5 - Empirical Findings and Data Interpretation 10 Chapter 6 - Conclusion 12 References 14 Census and Statistics Department of Hong Kong. ‘Housing in Figures’. Available at http://www.housingauthority.gov.hk/en/aboutus/resources/figure/0,,,00.html (Accessed: May 3, 2010) 14 Rating and Valuation Department of Hong Kong. Domestic Sales - Number of Sale and Purchase Agreements by Consideration Range (up to November 2009). Available at http://www.rvd.gov.hk/en/publications/pro-review.htm (Accessed: May 3, 2010) 14 Bibliography 14 Meen, G. (2001) Modelling spatial housing markets: theory, analysis, and policy. London, UK: Springer. 14 Wang, G. C. S. & Jain, C. L. (2003) Regression analysis: modeling & forecasting. USA: Graceway Publishing Company. 14 Appendix 15 Chapter 3 - Theoretical Background Price bubble is the result of a mismatch between the fundamentals and the outcome. For instance, if property price does not match with that of the core integrals of property say, the quality of the underlying property or that of the demand and supply situations in the market, then the resultant outcome is a price bubble. A price bubble is usually subject to an explosion after some point of time or another. An explosion indicates the situation when the commodity price falls suddenly without prior notice. The implications of such an explosion could be quite intense in case that the underlying sector is integrally related to that of the economic activities, i.e., it highly contributes to the aggregate income of the concerned nation. Example could be sought out from that of the Chinese example, which is structurally related to the domestic housing sector. Such a situation might arise only in case of a highly growing economy. Hence, the economic growth rate of a nation is one of the factors which could lead to a rise in the general price level. This is because a rise in the economic growth rate creates a certain level of optimism among the nationals and thus enhances their propensity to spend. Since investment is a factor that depends on the rate of growth in the previous period, a steep rise in one period results to a rise in the level of investment in the following year, which assures a steady flow of income for all the coming years. In fact, any nation concentrates on stabilizing their level of income over the long run and thus, the governments of such nations themselves come forward to assure the same. This stabilization could be ensured only through increasing the level of investment in the nation. But macroeconomic theory suggests that a rise in the demand for commodity leads to a rise in the market price of the same. Thus, the economic growth rate could be considered as one of the factors which might lead to a rise in property prices at least theoretically. With a rise in the level national income of the nation over time, there had been an increase in the demand for property in the economy. People had been running around with an objective of managing their wealth and increasing the quantity of their investment in the housing market. Possessing a house as a mode of investment is one of the ways by which a constant inflow of revenue or rather investment earnings could be assured. This is because, even a few years back, investing in houses had taken the front-seat in an investor’s preferences due to the certainty of the matter. Real estate is one of the assured contents of property which yields a steady flow of income that keeps on growing with time, as the level of income keeps on rising. In case of the Chinese example, the case is so since the nation had been witnessing a stable flow of income since a couple of years now; to be precise, right from the beginning of the era of liberalization or globalization. With rise in the per capita income of the nationals, there had been an increase in the aggregate demand for houses in the nation or rather among the nationals. Hence, rise in per capita income is one of the reasons why the nation had witnessed a rise in the level of income of the average citizen. However, there are many other factors as well, which already had been pointed out in the past researches made by various eminent observers. Some such factors are low rates of interest on bank deposits. The financial institutions had reduced their market rate of interest so that the nationals have a greater proportion of their income disposed for property investment. This initiative had mainly been focused to enhance the level of investment in the nation and thus ensure a robust economic growth figure as well as a steady flow of income prospects in the future. A fourth economic factor is that of the rate of inflation. A nation which had been subjected to a high rate of growth within a span of a few years will naturally be subjected to a certain level of inflation as well. A rise in the level of income will naturally lead to a rise in the demand at the aggregate level as well. This demand, however, might be for consumption as well as for investment purpose. With the level of supply in the economy remaining stagnant or increasing at a proportionately lower rate than the aggregate demand, there are rather slim chances of the general price level remaining at the similar level. However, people disapprove a rise in the general price level, i.e., inflation and rather prefer to keep themselves insured or assured against any such price hikes. This is because a rise in price implies a fall in the value of money and thus, a fall in the aggregate purchasing power of the people as well. Hence, people insure themselves through investment in property. With real estate yielding a constantly growing rate of income over time, unlike the equity shares of various companies, people often find the former option more lucrative than the previous one. Hence, there might also be a rise in the level of property prices. Finally, population accounts for another factor which could lead to a rise in property prices. China at present is at the middle stage of demographic changes, as is the case for most of the developing nations. In this intermediate stage, the birth rate of the population exceeds the mortality rate; to be precise the birth rate is quite high though the death rate is low. With the difference between the two rates increasing, the population of the concerned nation is bound to increase as well. But, a rise in population will put an upward pressure to the demand side of the economy, so that there will be an excess demand from each and every dimension of the economy. There will be a rise in consumption demand as well as demand for investments and so, the price of commodities as well as services in the economy will rise to cope up with the rise in aggregate demand; the case is especially so when the supply situation cannot be altered much since the level of national production has already reached the level of full employment. With a rise in population, there will be an economic inclination towards purchasing new houses for dwelling rather than investment. Moreover, as population pressure leads to a rise in the price level of the nation, people in their urge to keep their purchasing power intact, will increase their level of investment as well and more so in the real estate sector. Hence, from the discussion in the above paragraphs, it can be said that there are practically four types of factors, which have a primary impact on the house prices of an economy, especially like that of China. These four factors could be categorized into two classes, viz., economic factors and developmental factors. Economic factors include the per capita level of income, the economic growth rate of the nation, the rate of interest offered by the banks on demand deposits and the rate of inflation witnessed by the nation. On the other hand, the developmental factor includes the growing population in the nation (Tsatsaronis & Zhu, 2004). The present paper will try to figure out how far the aforesaid factors had been able to influence property prices. As already mentioned, a price bubble is something which is the result of a mismatch between the fundamentals and final outcome. If the rise in all the above elements is found to have had little effect on the house prices over the years, it could suggest a hint of a growing house price bubble in the economy. The following section deals with the methods through which the creation of a house price bubble in the Chinese economy could be assessed. Chapter 4 - Methodology and Data Collection The factors being mentioned above as the ones to be influencing the house prices of the Chinese economy will be portrayed in a regression equation, which will aim at figuring out the extent to which the house prices are being influenced by them. The objective of a regression equation is to find out the marginal impact as well as the significance of the same that an exogenous variable has on the dependent or endogenous variable. 4.1 Methodology Since house prices are assumed to be largely and considerable affected by all the aforementioned factors, a regression equation will be run on all exogenous variables over the endogenous variable, i.e., the growth in the house prices in China. The growth rate is preferred rather than the nominal value since the latter accounts for the rate of change in the variable and hence is a real term devoid of any influences accounting due to an external factor. On the other hand, a sheer consideration of the nominal house prices might lead to a potential misconception about the trend in house prices in the economy. Hence, the regression equation will assume the following form, H = constant + b1. G + b2. R + b3. Π + b4. P Where, H = House prices of China, G = Economic growth rate, i.e., the rate at which GDP grows over the years, R = Market rate of interest, Π = Rate of Inflation in the economy and P = Rate of Population growth over the years. To be more exact, the independent variables, ‘G’, ‘Π’ and ‘P’ being included in the model are expected to portray positive influences on the dependent variable, while the remaining variable, ‘R’ will impose a negative relation on the same. However, the extent of the influence or even the significance of the same will depend on the empirical findings. If the influence of the variables are found to be quite high or they are significant enough, that would imply an absence of bubble in house prices. In contrast, if the level of significance or the extent of the influence is low, the occurrence of a house price bubble will have to be admitted to have happened in the Chinese economy. OLS regression model will be used to conduct the regression equation in the present case. The regression will be entirely based on a collection of sample observations, and the behavior of a population of such data will be predicted on the basis of that. The estimated coefficients of the parameters in the sample regression will be accompanied by their respective estimated Student’s t-statistics. These Student's t-statistics will be responsible for concluding about the significance with which each of the variables can affect the endogenous variable or the dependent variable, i.e., growth in the house prices of China. Hence, Student’s t-statistics help to figure out the explanatory powers of each of the independent variables. On the other hand, the predicted F-statistic being included in the estimated model will conclude about the explanatory power of the estimated model as a whole; it will help to evaluate the significance of the effect that all variables taken together will have on the rate of growth in the house prices of China. The question which might arise in the present context is that about the concepts of Student’s t-statistics and F-statistics. As already mentioned, these statistics are used to figure out whether the estimated parameters could explain variations in the dependent variable or not. Null and alternative hypotheses in relation to testing the significance of the estimated values need to be framed first. In the present context, there will be, in all, five null and alternative hypotheses, which have been specified as under – H0i: The estimated population parameter ‘bi’ cannot explain variations in the dependent variable, i.e., bi = 0 against the alternative hypothesis, H1i: The estimated population parameter ‘bi’ can explain variations in the dependent variable, i.e., bi ≠ 0; ‘i’ = 1, 2, 3, 4 And, H05: None of the estimated population parameters can explain variations in the dependent variable, i.e., b1 = b2 = b3 = b4 = 0 against the alternative hypothesis, H15: All the estimated population parameters can explain variations in the dependent variable, i.e., b1 ≠ b2 ≠ b3 ≠ b4 ≠ 0 The first hypotheses deal with the explanatory powers of each of the individual estimated coefficients and thus, the ideal test for the purpose is Student’s t-statistic. On the other hand, the latter hypothesis which tells about the explanatory power of the regression model as a whole could be tested with the help of F-statistic. The rules to accept or reject the null hypotheses on the basis of the estimated values of the statistics have been illustrated as follows. However, before describing, it is essential to tell that the basis of comparison will be an assumed level of significance, which in the present context will be 0.05. There is no particular reason behind assuming 5% level of significance, but this is a universal benchmark which supposes that 5% of the total area under the cumulative frequency distribution curve falls under the rejection region. If the estimated level of significance (probability value or ‘p-value’) of the concerned variable is found to exceed the assumed level of significance, it implies that the value of the estimated population parameter falls under the acceptance region of the cumulative distribution curve. Hence, the null hypothesis in this case cannot be rejected at 5% level of significance. In contrast, if the estimated p-value of the population parameter is found to be lower than the assumed level of significance, it implies that the value falls under the rejection region and hence, needs to be rejected at 5% level of significance. As an endnote it could be said that even though there is an older method of comparing the estimated and the tabulated values of the estimated test statistics, the method of comparing the p-values is an easier one, given the use of advanced statistical software. Finally, another important statistic to be taken into account for the purpose of figuring out the reliability of the model will be the value of the multiple correlation coefficient, i.e., R Square. R Square assesses the goodness of fit of the model or rather the extent to which the estimated model complies with the actual one. If the value of R Square is quite high, it implies a high compliance or match between the actual and the estimated regression model. A graph will accompany the statistics for the goodness of fit. It will help to compare the actual population regression line with the fitted sample regression model. A high value of R Square will be represented by a close connection in the trends of the two lines, i.e., the two lines will closely follow the trends of each other. In contrast and quite obviously, a low value of R Square will be represented by a mismatch between the trends of the two lines. 4.2 Data Collection The data to be collected in the present instance are those of the growth in the house prices in the economy, the economic growth rate, the rate of inflation, the market rate of interest on demand deposits and the rate of growth in population. Since the incidence of a possible formation of a house price bubble occurred by the end of the twentieth century, i.e., during the commencement of the East Asian Financial Crisis of 1997, the data being included will be from the year 1988 and ending in the year 2007. It will be annual in nature so that there will be a total of 20 data points. Though 20 data points might not be statistically claimed as one that could result to a proper or a valid regression result, unavailability of data for the same imposes such restrictions on the model. Speaking of constraints in the analysis, there are actually a few of them which could be encompassed as follows – (1) The data being considered is that in the context of Hong Kong rather than of People’s Republic of China. Being a part of the latter, any changes taking place in the economy of China is assumed to be reflected in the trends that the former will take, though the reverse might not be true. Since Hong Kong accounts for a large proportion of the economic success of the Chinese, such an assumption could also be considered as a rather justified one. (2) Secondly, the data points being collected are 20 in all, each accounting for a year for 20 years between 1988 and 2007. Though the minimum number of data observations being considered to conduct a valid regression result is 30, the number has to be compromised to account for the lack of abundant information on the area. But, there is no reason to believe that the resultant output will merely be an estimate bearing no resemblance with the reality. The statistical significance tests being conducted as a part of the empirical analysis will assure the validity of the estimated outputs. Such assumptions had primarily been made due to unavailability of aggregate data in context of China. Hong Kong SAR region, though is an economy in itself, but is integrally related to the trends of China, implying that the former will mirror the developments taking place in its parent economy. Even real life observations could be used as an alibi to the assumption about Hong Kong being a miniature of Chinese trends. The data being collected are secondary in nature and are gathered from the archives of the Census and Statistics Department affiliated under the Government of the Hong Kong SAR region. The following section will brief about the empirical analysis in respect to the model being specified above. Use has been made of the statistical software application EViews in the present study. The data employed and the output being derived after running the regression according to the aforementioned lines have been provided in the appendix at the end of the paper. Chapter 5 - Empirical Findings and Data Interpretation The model has been estimated in the form being specified above. The predicted model is as under – H = - 0.035 + 0.031 G + 0.027 π – 0.044 P – 0.035 R (0.556) (0.000) (0.000) (0.232) (0.026) The expressions in parentheses are the estimated values of levels of significance for the respective population parameters. The values, both the estimated parameters as well as the statistics, are approximate in nature. The coefficient of the population growth rate, ‘P’ is found to be negative unlike the predicted direction of association. However, that for economic growth rate is ‘G’ as well as for that of the rate of inflation, ‘π’ are positive as was expected, while that for the rate of interest on savings deposits ‘R’ is found to be negative as previously predicted. Given that the assumed level of significance is 0.05, the null hypothesis is found to be rejected for most of the variables according to the lines of reasoning being described above. To be more exact, the null hypotheses in context of the explanatory powers of the variables ‘G’, ‘π’ and ‘R’ are found to be lower than 0.05, implying that all of them fall within the rejection region of the respective cumulative frequency distribution curves. On the other hand, the estimated p-value for the variable ‘P’ is found to be higher than 0.05, so that the particular parameter is assumed to fall within the acceptance zone. Hence, the explanatory powers of the former three variables are found to be quite strong or significant unlike the last one. In fact, the population parameter is found to portray a completely opposite association with the dependent variable than was expected. Hence, it could easily be said that the model being estimated is a valid one. In fact, the above statement could also be ensured from the value of the estimated F-statistic, which is approximately equal to 12.896, with the value of the corresponding p-value equal to 0.000. Hence, the explanatory power of the model cannot be denied as well, and so, the concerned null hypothesis (H15) cannot be rejected. Moreover, as far as the goodness of fit of the model is concerned, there are little doubts about the match between the estimated and the actual population regression line. The value of the multiple correlation coefficient, R Square is found to be approximately equal to 0.775, which is considered as a rather robust figure. This implies that the trends being followed by the actual regression line is explained up to approximately 77.5% by the estimated model. Chapter 6 - Conclusion The issue of the research paper had been about a potential occurrence of a house price bubble in the Chinese economy during the East Asian financial crisis of 1997. The adjoining diagram shows a sudden rise in the average price of houses per square meter by the end of the twentieth century, which surely suggests the occurrence of a house price bubble in the Chinese economy. The regression model being estimated is found to show significant explanatory powers of the population parameters, though the values of the estimated coefficients are found to be quite low implying that the independent variables being included do not bestow much influence on the dependent variable, i.e., house prices. However, theoretically there is almost no possibility of a little effect of the exogenous variables being considered on the endogenous one. Moreover, there is also no reason to believe that the model does not include all the relevant variables which might probably affect the dependent variable. This conclusion could be drawn from the significant value of the estimated F-statistic. Since F-statistic tells about the explanatory power of the estimated model, there are little chances about the omission of a relevant variable which could have enhanced explanatory power of the model. Moreover, R Square is quite high (0.775), implying a rather highly good model fitness, leaving no space to assume that the actual situation is different from the estimated one. Given the above interpretation of the estimated regression model, there are a number of reasons to believe that the house prices in China were involved in a house price bubble. The rise in the house prices of the nation was not supported by the economic and developmental indicators as had been specified and predicted theoretically. Thus, it could be said that the house price in the nation escalated without any support from the fundamentals, which is why there had been an unnatural rise in the level of prices, resulting to a price bubble in the sector. Though it must be mentioned that there had been oversimplification in some respects for instance, in considering Hong Kong as a proxy of China, but given that the latter is already a relevant part of PRC, there could be little doubts as to the estimated results. The figures of the estimated parameters might be different, but there is little scope of any misgivings about any difference in the ultimate conclusion. References Census and Statistics Department of Hong Kong. ‘Housing in Figures’. Available at http://www.housingauthority.gov.hk/en/aboutus/resources/figure/0,,,00.html (Accessed: May 3, 2010) Rating and Valuation Department of Hong Kong. Domestic Sales - Number of Sale and Purchase Agreements by Consideration Range (up to November 2009). Available at http://www.rvd.gov.hk/en/publications/pro-review.htm (Accessed: May 3, 2010) Tsatsaronis, K. & Zhu, H. (2004) “What drives housing price dynamics: cross-country evidence”, BIS Quarterly Review, March 2004 [Online]. Available at < http://www.bis.org/publ/qtrpdf/r_qt0403f.pdf> (Accessed: May 4, 2010). Bibliography Meen, G. (2001) Modelling spatial housing markets: theory, analysis, and policy. London, UK: Springer. Wang, G. C. S. & Jain, C. L. (2003) Regression analysis: modeling & forecasting. USA: Graceway Publishing Company. Appendix Data Set Growth in House prices GDP Growth rate Inflation rate Population Market rate of interest on savings 0.212697274 7.968149 9.44 0.8 5.19 0.228811684 2.560697 10.23 1.0 5.25 0.129915764 3.404704 10.90 0.3 5.5 0.255829781 5.05829 11.60 0.8 3.5 0.282892371 6.263075 9.34 1.35 1.5 0.092646545 6.131491 8.49 1.80 1.5 0.178206823 5.401593 7.76 2.24 3.75 -0.058490796 3.893572 9.03 2.61 4.21 0.096050582 4.3 6.30 2.67 3.75 0.211556801 5.1 5.84 1.63 4.75 -0.365359954 -5 2.84 0.75 4.46 -0.210840639 3.4 -3.96 0.84 3.75 -0.163613621 10.2 -3.77 0.90 4.75 -0.126965729 0.463235 -1.61 0.82 0.2 -0.100303323 1.938338 -3.03 0.73 0.03 -0.126854657 3.14705 -6.24 0.70 0.01 0.208667589 8.145263 -3.38 0.67 0.01 0.169135298 7.270713 0.75 0.64 2.22 0.0861397 7.01806 2.33 0.61 2.27 0.180582116 6.367597 2.02 0.58 1.35 Estimated Model Read More