Table 2. Estimations of country-level coverage with piped water.
Random Effects a | Fixed effects | |||
Simple model | Full model (all countries) b | Simple model | Full model (all countries) b | |
5-yr lagged ln GDP per capita | 11.2*** (1.96) | 9.9*** (1.83) | 8.1*** (2.2) | 6.1*** (2.1) |
% of GDP to lowest 80% of population | 0.11 (0.10) | 0.11 (0.09) | 0.12 (0.11) | 0.13 (0.10) |
% Urban population | 0.45*** (0.08) | 0.48*** (0.08) | 0.37** (0.17) | 0.45*** (0.16) |
Countries in LAC region | 27.9*** (5.0) | 27.4*** (4.9) | ||
Countries in MIDEAST region | 30.8*** (5.6) | 31.6*** (5.4) | ||
Countries in SOUTH ASIA region | 5.0* (2.7) | 5.6** (2.8) | ||
Countries in EAST ASIA/PACIFIC region | 8.2 (5.2) | 8.9* (5.3) | ||
Countries in EASTERN EUROPE region | 44.4*** (4.2) | 32.9*** (4.3) | ||
Developed countries | 37.5*** (6.2) | |||
1990 | −4.7*** (1.4) | −2.9** (1.2) | −6.4*** (1.8) | −4.7*** (1.7) |
1995 | −2.6** (1.1) | −0.64 (0.81) | −4.0*** (1.4) | −2.0* (1.2) |
2000 | −0.85 (0.78) | −0.78 (0.69) | −1.9** (0.97) | −1.7* (0.91) |
2005 | 0.21 (0.38) | −2.3*** (0.61) | −0.38 (0.49) | −2.6*** (0.65) |
Democracy-Autocracy Score | −0.16* (0.10) | −0.056 (0.09) | −0.23** (0.10) | −0.20* (0.10) |
Years since last regime change | −0.013 (0.05) | −0.050 (0.04) | −0.027 (0.05) | −0.066** (0.05) |
Coup | −1.3 (1.1) | −1.6 (1.1) | −0.93 (1.2) | −1.2 (1.1) |
Constant | −80.1*** (12.1) | −72.3*** (11.7) | −37.0** (17.7) | −19.9 (19.3) |
Number of observations | 470 | 634 | 470 | 634 |
Number of countries | 95 | 131 | 95 | 131 |
Adjusted R2 (overall)(within)(between) | 0.9010.5510.913 | 0.9070.4890.914 | 0.7740.5590.800 | 0.7910.5020.808 |
Hausman Test for simple model χ 2 (p-value) | 78.2 (0.000) |
Notes: *Significant at 90%, **Significant at 95%, ***Significant at 99%. Robust standard errors. The omitted region is SSA; the omitted year is 2010.
A random-effects tobit model with censoring at 0 and 100% coverage does not yield qualitatively different results.
Includes all countries (including developed and former Soviet republics dropped from the simple model) and full set of year-region interactions.