Table 5.
A comparison of the estimation results of various empirical methods (Variable = IWGE, tau = 0.5/mean).
| Region | Model | Estimate | Model | Estimate |
|---|---|---|---|---|
| Whole country | OLS | 0.0854 *** (0.0225) |
QR | 0.1043 ** (0.0154) |
| BLR | 0.0855 *** (0.0226) |
BQR | 0.1023 ** (0.0393) |
|
| High-income region | OLS | −0.0610 * (0.0413) |
QR | −0.0427 ** (0.0757) |
| BLR | −0.0607 *** (0.0416) |
BQR | 0.0028 ** (0.0661) |
|
| Middle-income region | OLS | −0.0191 * (0.0367) |
QR | −0.0282 ** (0.0456) |
| BLR | −0.0190 *** (0.0371) |
BQR | 0.0138 ** (0.0014) |
|
| Low-income region | OLS | −0.2671 *** (0.0480) |
QR | −0.3041 ** (0.0154) |
| BLR | −0.266 *** (0.0491) |
BQR | −0.0981 ** (0.0093) |
Note. “***” indicates statistical significance level at 1%. “**” indicates statistical significance level at 5%. “*” indicates statistical significance level at 10%. Both the outcomes of QR and BQR were 95% credible interval (**), In BQR, Number of burn in draws: 1000, Number of retained draws: 4000. OLS: Ordinary least squares; QR: Quantile regression; BLR: Bayesian linear regression; BQR: Bayesian quantile regression.