*5.1. Baseline Estimates*

The baseline estimates are shown in Table 3. To compare the results of models with/without dummies, column (1) in Table 3 shows the results of the model without regional dummies and yearly dummies, column (2) includes yearly dummies, and column (3) includes both regional dummies and yearly dummies.



Note: \*, \*\*, and \*\*\* indicate statistical significance at the 10%, 5%, and 1% levels, respectively. Standard errors are in parentheses.

The coefficient of LAWGOV in each column is statistically significantly positive, suggesting the law aspect of governance is positively associated with housing prices. Additionally, the signs of the coefficients of the control variables are basically as expected. Neither population density (POPDEN) nor green ratio (GRERAT) have significant effects on housing prices. However, the effects of other control variables on housing prices are all significant. Per capita disposable income of urban households (INCOME) and the ratio of tertiary industry's output value in GDP (INDSTR) have greater effects on housing prices than other control variables do. Not surprisingly, the effects of the per capita area of residential construction land launched by the governmen<sup>t</sup> (LANSUP), unemployment rate (UNEMRA), the shortest distance to the coastline (DISCOA), and the emission ratio of industrial soot and dust (DUSOOT) on housing prices are negative. Compared to columns (1) and (2), column (3) shows that the root MSE becomes lower and the R2 rises to a higher level. This indicates that when economic region submarket dummy variables are introduced into the OLS model, its explanatory power increases [61]. We use the model with both regional and yearly dummies to undertake further analysis.
