1. | Regarding the knot point, we pre-specified the knot points according to realistic situations and personal judgments. Some pre-specified knot points turned out to be insignificant. A better approach would be using regression techniques to select significant yet unknown knots. |
2. | Regarding distance variables in each interval, using a quadratic or higher order polynomial functional form could better fit the relationship between distance and housing prices. |
3. | In a totally different direction, it would be interesting to look into Chinese culture and the Chinese traditional strong desire to own land or houses. Those cultural characteristics might help us better explain the housing market in Asian countries. |
4. | Are there any other variables besides typical neighborhood proxies that could be |
-255-