Table 3.
Obtaining a joint preferred specification for social care, healthcare and public health expenditure by combining full specifications with forward selection (second-stage results) (created by the authors)
| Variables | (1) | (2) | (3) | (4) |
| All causes | All causes | All causes | All causes | |
| 2013/2014 PB/GSS/PH spend | 2013/2014 PB/GSS/PH spend | 2013/2014 PB/GSS/PH spend | 2013/2014 PB/GSS/PH spend | |
| SYLLR 2013/2014/2015 | SYLLR 2013/2014/2015 | SYLLR 2013/2014/2015 | SYLLR 2013/2014/2015 | |
| Outcome model | Outcome model | Outcome model | Outcome model | |
| Instrument PB/GSS/PH spend | Instrument PB/GSS/PH spend | Instrument PB/GSS/PH spend | Instrument PB/GSS/PH spend | |
| Weighted | Weighted | Weighted | Weighted | |
| IV second stage | IV second stage | IV second stage | IV second stage | |
| Parsimonious specification | Specification_2 | Specification_3 | Specification_4 | |
| 11 instruments | 7 instruments | 5 instruments | 4 instruments | |
| Public health expenditure per person | −0.051 (0.039) |
−0.043 (0.047) |
−0.060 (0.044) |
−0.099** (0.045) |
| CCG (PB) healthcare spend per person | −0.862*** (0.223) |
−0.461 (0.313) |
−0.637* (0.369) |
−0.693** (0.333) |
| Social service (GSS) spend per person | −0.206 (0.133) |
−0.469*** (0.161) |
−0.370* (0.205) |
−0.471** (0.237) |
| % population aged 16–74 that are permanently sick | 0.649*** (0.046) |
0.651*** (0.054) |
0.672*** (0.052) |
0.528*** (0.073) |
| % population providing unpaid care | −0.381*** (0.086) |
−0.388*** (0.096) |
−0.400*** (0.097) |
−0.143 (0.118) |
| % population in white ethnic group | 0.163*** (0.042) |
0.195*** (0.043) |
0.180*** (0.046) |
0.299*** (0.078) |
| Older adult: social service need per person | 0.416*** (0.143) |
|||
| Constant | 13.352*** (1.733) |
10.204*** (2.389) |
11.655*** (2.873) |
12.245*** (2.546) |
| Observations | 150 | 150 | 150 | 150 |
| Endogeneity test statistic | 10.644 | 17.686 | 21.100 | 23.214 |
| Endogeneity p value | 0.014 | 0.001 | 0.000 | 0.000 |
| Hansen-Sargan test statistic | 25.690 | 10.432 | 2.173 | 0.080 |
| Hansen-Sargan p value | 0.001 | 0.034 | 0.337 | 0.778 |
| Pesaran-Taylor reset statistic | 0.052 | 0.372 | 0.080 | 0.001 |
| Pesaran-Taylor p value | 0.820 | 0.542 | 0.777 | 0.972 |
| SW_PB F-statistic | 5.977 | 6.833 | 7.878 | 13.534 |
| SW_PB p value | 0.000 | 0.000 | 0.000 | 0.000 |
| SW_GSS F-statistic | 4.939 | 5.521 | 6.135 | 9.722 |
| SW_GSS p value | 0.000 | 0.000 | 0.001 | 0.000 |
| SW_PH F-statistic | 18.451 | 22.092 | 30.944 | 46.946 |
| SW_PH p value | 0.000 | 0.000 | 0.000 | 0.000 |
Robust SEs in parentheses.
***p<0.01, **p<0.05, *p<0.1.
CCG, Clinical Commissioning Group; PH, public health; SW, Sanderson-Windmeijer.