Table 1.
Definitions and descriptive statistics of variables.
Table 2.
Correlation analysis between independent variables.
Table 3.
Collinearity and unit root tests.
Fig 1.
The changes in health workforce in Inner Mongolia from 2013 to 2022.
Fig 2.
Spatial distribution of health workforce density in counties of Inner Mongolia, 2013-2022.
Table 4.
Global Moran’s I of HW in Inner Mongolia.
Fig 3.
Moran’s I scatterplots for HW in 2013, 2016, 2019, and 2022 (In the Moran’s I plots, Quadrant I represents High–High (HH) clusters, Quadrant II Low–High (LH), Quadrant III Low–Low (LL), and Quadrant IV High–Low (HL)).
Table 5.
Local spatial autocorrelation analysis of HW in 2013, 2016, 2019, and 2022.
Fig 4.
The LISA cluster map of HW in 2013, 2016, 2019, and 2022 (Areas with significant local spatial autocorrelation are color-coded: High–High (HH) clusters in red, Low–Low (LL) clusters in blue, Low–High (LH) outliers in yellow, High–Low (HL) outliers in pink, and non-significant areas in gray).
Table 6.
LM test and Robust LM test.
Table 7.
LR test and Wald test.
Table 8.
Regression results of SLM, SEM, and SDM.
Table 9.
Spatial effect decomposition of SDM with a two-way fixed effects.
Table 10.
Instrumental variable regression results (two-stage least squares estimation).
Table 11.
Spatial IV estimation using GS2SLS.
Table 12.
Robustness checks using different specifications of the spatial weight matrix.
Table 13.
Robustness checks under alternative values of the decay parameter (β).
Table 14.
Estimation results of SDM with lagged independent variables.
Table 15.
Robustness checks of model estimates after excluding Hohhot (9 districts/counties).
Table 16.
Direct, indirect, and total effects of key determinants on health workforce allocation with interaction terms (CityDistrict = 0 as reference).