The first regression includes all the variables mentioned earlier except for for the neighborhood effect variable. The results of this regression are reported in Table 4. The number of bedrooms, existence of a formal dining room, age of the house, whether the garage(s) are attached or not, and whether the house has a deck or patio are some of the variables that are not significant. I experimented with the functional form on the age variable, thinking it may be quadratic rather than linear but the results were the same. The variables that were significant at the 95% level are in bold. The insignificance of the attached garage variable is very surprising because I expected a homebuyer would be willing to pay a premium to not have to go outside on cold winter mornings. The fact that it get so cold in the area might explain why the fireplace variable was so high and statistically significant. The regression shows that each additional fireplace adds about $14,000.00 (almost 10% of the mean) to the price of the house. Other surprising results are that the number of bedrooms and the age of the house are statistically insignificant. The F-statistic is 178.1344, much higher than the critical value of F. The Adjusted R-Squared is 0.751468, so the equation explains approximately 75% of the variation in price.
The location variable turned out to be insignificant. The insignificance of this variable means that whether the house is in Fargo or Moorhead does not matter to homebuyers. These results did not support my hypothesis that there would be a premium to purchase a house in Fargo. The Law-of-One-Price holds and the differences in income and property taxes seem to offset each other