Testing the weak form of the Property Market Efficiency

• Heteroscedasticity in the Repeat Sales Regression (Cont...)
  

The results confirm the heteroscedasticity of age in the estimators, and explain why heteroscedasticity still remains after the corrections following Case and Shiller’s method.

To correct for the heteroscedasticity in the property price index, the same process was followed, by weighting each of the observations by the fitted standard error but this time with age and holding period as explanatory variables. And the correct weighted “real” returns (by dividing by the quarterly increases in the Auckland CPI) are shown in the table below.







After the transformation there was no more significant relationship between holding period, age, or both, with the squared residuals of the model, thus confirming the elimination of this Heteroscedasticity.

To verify the performance of the constructed property index in the table above, the nominal weighted index is compared to the half yearly average house price index of Auckland, produced by the Massey University Real Estate analysis unit.



Since there are sampling and methodological differences between the two indexes, there will be no exact fit between the two. But according to the comparison above, the weighted repeat sales index exhibits a good explanatory power for the changes in capital gains, as the index follows the average index quite well. The average price index constructed by Massey University in fact seems to lag the WRS index constructed. We can conclude that the property pricing index fits well to the data given that the overall capital gains are well portrayed in the index, and inferences to the efficient market hypothesis can be made.