1 Introduction
2 Literature Review
2.1 Conceptual model
2.2 Indicators of housing quality
2.3 Effects of indoor housing conditions on health
2.3.1 Dampness and cold homes
2.3.2 Indoor air pullant, and absence of indoor water tap
2.4 Effects of homeownership on health
2.5 Neighbourhood effects
3 Data
3.1 Dependent variable
3.2 Housing variables
Sturdy (S) or Flimsy (F) | Household gross income (’000 VND) | % of households | |
---|---|---|---|
Surrounding walls
| |||
Earth Leaves/branches/bamboo | F F | 36,099.69 41,112.14 | 6.85 3.39 |
Wood | S | 49,870.24 | 31.58 |
Galvanized iron | S | 66,873.13 | 0.83 |
Brick (fired or unfired) | S | 85,467.25 | 54.82 |
Concrete Others
Roof
Straw, leaves Sheeting/panel Canvas, tar paper Wood Tile Galvanised iron Concrete, cement Others | S S F F F S S S S S | 88,698.47 60,346.28 39,221.25 42,056.38 47,928.30 52,027.73 67,164.36 77,236.43 113,440.70 48,481.58 | 2.30 0.23 3.73 20.71 0.31 0.94 32.91 27.78 12.65 0.97 |
3.3 Control variables
Variables | Obs. | Mean | Std. Dev. |
---|---|---|---|
Number of sick days (SD) | |||
SD ≤ 60
| 73,369 | 4.49 | 8.41 |
Full sample
| 74,793 | 8.29 | 32.74 |
Flimsy outer wall (= 1 if true; =0 otherwise) | 73,828 | 0.10 | 0.30 |
Flimsy roof (= 1 if true; =0 otherwise) | 73,813 | 0.25 | 0.43 |
Harmful cooking source (= 1 if true; =0 otherwise) | 73,825 | 0.64 | 0.48 |
No flush toilet (= 1 if true; =0 otherwise) | 73,807 | 0.58 | 0.49 |
No water tap (= 1 if true; =0 otherwise) | 73,820 | 0.24 | 0.42 |
Housing tenure (= 1 for renters; =0 for homeowners) | 73,808 | 0.987 | 0.11 |
Housing area per capita (m2) | 73,806 | 16.56 | 13.56 |
Household gross annual income (‘000 VND) | 73,610 | 69,205 | 100,840 |
Working status | |||
Employed full-time by someone
| 79,381 | 0.24 | 0.43 |
Doing agricultural work of the household
| 79,381 | 0.40 | 0.49 |
Doing non-agricultural economic activities
| 79,381 | 0.03 | 0.16 |
Activities using common property resources
1
| 79,381 | 0.01 | 0.09 |
Doing domestic tasks
| 79,381 | 0.10 | 0.30 |
Pupil, student, retired & other
| 79,381 | 0.22 | 0.42 |
Educational attainment | |||
No degree
| 71,081 | 0.87 | 0.34 |
High school & vocational training
| 71,081 | 0.10 | 0.30 |
College & above
| 71,081 | 0.03 | 0.18 |
Marital status | |||
Single
| 71,199 | 0.41 | 0.49 |
Married
| 71,199 | 0.53 | 0.50 |
Widowed / Divorced
| 71,199 | 0.06 | 0.25 |
Gender (= 1 if male; =0 if female) | 79,198 | 0.43 | 0.49 |
Ethnicity (= 1 if Kinh majority; =0 if ethnic minorities) | 73,825 | 0.51 | 0.50 |
Age categories | |||
Age ≤ 20
| 79,381 | 0.39 | 0.49 |
21 ≤ Age ≤ 30
| 79,381 | 0.16 | 0.37 |
31 ≤ Age ≤ 40
| 79,381 | 0.13 | 0.33 |
41 ≤ Age ≤ 50
| 79,381 | 0.13 | 0.33 |
51 ≤ Age ≤ 60
| 79,381 | 0.09 | 0.29 |
Age ≥ 61
| 79,381 | 0.10 | 0.30 |
Currently smoker | |||
Yes
| 72,069 | 0.59 | 0.49 |
No, but used to
| 72,069 | 0.22 | 0.41 |
Never
| 72,069 | 0.20 | 0.40 |
Distance to public health centre (km) | 73,779 | 3.08 | 12.39 |
Distance to hospital (km) | 73,782 | 16.84 | 22.61 |
3.4 Preliminary analysis
Housing characteristics | Average sick days1 | |
---|---|---|
Yes
|
No
| |
Whether the house has flimsy outer walls | 4.450 | 4.486 |
Whether the house has a flimsy roof | 4.089 | 4.611 |
Uses harmful source for cooking | 4.633 | 4.216 |
Not having a flush toilet | 4.571 | 4.358 |
Not having a water tap | 4.548 | 3.838 |
Rented house | 6.091 | 4.462 |
4 Model specification
4.1 Econometric framework
4.2 Potential endogeneity
5 Results
5.1 Relationship between housing and health
Variablesa | Dependent variable: Sick Days Negative Binomial (hybrid model) | |
---|---|---|
(1)
|
(2)
| |
Coeff.
|
IRR
b
| |
Household-level
| ||
Flimsy outer walls | 0.119 (0.036)*** | 1.127 (0.040)*** |
Flimsy roof | 0.081 (0.025)*** | 1.084 (0.028)*** |
Harmful cooking source | 0.039 (0.022)* | 1.040 (0.023)* |
No water tap | 0.237 (0.033)*** | 1.268 (0.042)*** |
No flush toilet | -0.055 (0.022)** | 0.947 (0.020)** |
Renter | 0.387 (0.088)*** | 1.472 (0.130)*** |
Housing area per capita | 0.004 (0.001)*** | 1.004 (0.001)*** |
Income quintiles (Lagged) - Ref. cat.: Quintile 1 | ||
Quintile 2
| -0.006 (0.024) | 0.994 (0.024) |
Quintile 3
| 0.035 (0.024) | 1.036 (0.025) |
Quintile 4
| 0.120 (0.026)*** | 1.127 (0.029)*** |
Quintile 5
| 0.218 (0.028)*** | 1.243 (0.035)*** |
Individual-level
| ||
Working status (lagged) - Ref. cat.: Full-time worker | ||
Agricultural work of the household
| 0.003 (0.018) | 1.003 (0.018) |
Non-agricultural economic activities
| -0.038 (0.044) | 0.963 (0.043) |
Activities using common property resources
| 0.011 (0.077) | 1.011 (0.078) |
Domestic tasks
| 0.145 (0.027)*** | 1.156 (0.031)*** |
Pupil, student, retired & other inactive person
| 0.095 (0.024)*** | 1.100 (0.026)*** |
Education - Ref. cat.: No degree | ||
High school & vocational training
| 0.009 (0.024) | 1.009 (0.024) |
College & above
| -0.022 (0.043) | 0.978 (0.042) |
Marital status - Ref. cat.: Single | ||
Married
| 0.028 (0.028) | 1.028 (0.029) |
Widowed/divorced
| 0.176 (0.040)*** | 1.192 (0.047)*** |
Maled | -0.096 (0.013)*** | 0.909 (0.012)*** |
Kinh ethnic majorityd | -0.141 (0.020)*** | 0.869 (0.018)*** |
Age - Ref. cat.: Age ≤ 20 | ||
Age between 21 & 30
| 0.000013 (0.027) | 1.000 (0.027) |
Age between 31 & 40
| 0.016 (0.034) | 1.174 (0.040) |
Age between 41 & 50
| 0.359 (0.035)*** | 1.433 (0.050)*** |
Age between 51 & 60
| 0.661 (0.037)*** | 1.938 (0.071)*** |
Age 61 and above
| 1.070 (0.038)*** | 2.915 (0.110)*** |
Smoking status - Ref. cat.: Never smoked | ||
Yes
| 0.023 (0.021) | 1.023 (0.021) |
No, but used to
| 0.071 (0.023)*** | 1.073 (0.025)*** |
Neighbourhood
| ||
Distance to health care centre | -0.003 (0.002)* | 0.997 (0.002)* |
Distance to hospital | -0.006 (0.001)*** | 0.994 (0.001)*** |
Individual fixed effects | Yes | |
Year fixed effects | Yes | |
Constant | -1.282 (0.088)*** | 0.278 (0.024)*** |
Observations | 48,240 | |
Log likelihood | -109,725.3 | |
Dispersion \({{\chi }}^{2}\)c | 200,930.4 | |
p-value | 0.00 | |
Wald\({{\chi }}^{2}\) | 845.18 | |
p-value | 0.00 |
5.2 Rural vs. urban areas
Variables a | Sick days Negative Binomial (hybrid model) | |||
---|---|---|---|---|
Rural
|
Urban
| |||
(1) | (2) | (3) | (4) | |
Coeff. | IRR | Coeff. | IRR | |
Flimsy outer walls | 0.127*** | 1.135*** | 0.729** | 2.074** |
(0.036) | (0.041) | (0.288) | (0.598) | |
Flimsy roof | 0.079*** | 1.082*** | 0.064 | 1.066 |
(0.026) | (0.028) | (0.149) | (0.159) | |
Harmful cooking source | 0.044** | 1.045** | -0.081 | 0.922 |
(0.023) | (0.024) | (0.129) | (0.119) | |
No water tap | 0.195*** | 1.215*** | 0.970*** | 2.637*** |
(0.034) | (0.042) | (0.135) | (0.356) | |
No flush toilet | -0.054** | 0.948** | 0.010 | 1.010 |
(0.022) | (0.021) | (0.138) | (0.139) | |
Renter | 0.402*** | 1.495*** | -0.266 | 0.766 |
(0.091) | (0.136) | (0.37) | (0.285) | |
Housing area per capita | 0.004*** | 1.004*** | 0.006*** | 1.006*** |
(0.00056) | (0.00056) | (0.0022) | (0.0022) | |
Control variables included | Yes | Yes | ||
Individual fixed effects | Yes | Yes | ||
Year fixed effects | Yes | Yes | ||
Constant | -1.31*** | 0.27*** | -1.68*** | 0.18*** |
(0.091) | (0.025) | (0.45) | (0.084) | |
Observations | 46,456 | 1,778 |
5.3 Intensity of housing problems
Intensity (Number of dwelling problems) | Number of households | Frequency (%) |
---|---|---|
0 | 804 | 4.89 |
1 | 4,370 | 26.56 |
2 | 3,533 | 21.47 |
3 | 4,534 | 27.56 |
4 | 2,348 | 14.27 |
5 | 864 | 5.25 |
Total | 16,453 | 100% |
Variables a | Sick Days | |
---|---|---|
Negative Binomial (hybrid model)
| ||
(1) | (2) | |
Coeff. | IRR | |
Intensity (Ref. cat.: No housing problems) | ||
Any one problem
| 0.363*** | 1.437*** |
(0.044) | (0.064) | |
Any two problems
| 0.285*** | 1.329*** |
(0.047) | (0.062) | |
Any three problems
| 0.363*** | 1.438*** |
(0.050) | (0.071) | |
Any four problems
| 0.359*** | 1.431*** |
(0.055) | (0.079) | |
All five problems
| 0.596*** | 1.816*** |
(0.065) | (0.119) | |
Control variables included | Yes | |
Individual fixed effects | Yes | |
Year fixed effects | Yes | |
Constant | -1.155*** | 0.315*** |
(0.094) | (0.030) | |
Observations | 48,245 |
5.4 Robustness checks
Variables a | Dependent variable: Sick Days Negative Binomial (Hybrid model) | |||
---|---|---|---|---|
(1) | (2) | (3) | (4) | |
Coeff. | IRR | Coeff. | IRR | |
Panel A: Varying the cut-off points for SD
| ||||
Sick days - Full sample
|
Sick days ≤ 30
| |||
Flimsy outer walls | 0.100*** | 1.105*** | 0.097*** | 1.102*** |
(0.036) | (0.039) | (0.036) | (0.040) | |
Flimsy roof | 0.078*** | 1.081*** | 0.066** | 1.068** |
(0.025) | (0.028) | (0.026) | (0.028) | |
Harmful cooking source | 0.031 | 1.032 | 0.032 | 1.033 |
(0.022) | (0.022) | (0.023) | (0.023) | |
No water tap | 0.222*** | 1.48*** | 0.231*** | 1.259*** |
(0.033) | (0.041) | (0.034) | (0.043) | |
No flush toilet | -0.038* | 0.963* | -0.067*** | 0.935*** |
(0.021) | (0.021) | (0.022) | (0.021) | |
Renter | 0.360*** | 1.433*** | 0.376*** | 1.457*** |
(0.088) | (0.126) | (0.090) | (0.131) | |
Housing area per capita | 0.004*** | 1.004*** | 0.005*** | 1.005*** |
(0.001) | (0.001) | (0.001) | (0.001) | |
Panel B: Excluding Lao Cai and Dien Bien provinces
| ||||
Flimsy outer walls | 0.027 | 1.028 | ||
(0.046) | (0.048) | |||
Flimsy roof | 0.041 | 1.042 | ||
(0.030) | (0.031) | |||
Harmful cooking source | 0.058** | 1.060** | ||
(0.023) | (0.024) | |||
No water tap | 0.269*** | 1.308*** | ||
(0.036) | (0.047) | |||
No flush toilet | -0.084*** | 0.919*** | ||
(0.023) | (0.021) | |||
Renter | 0.339*** | 1.403*** | ||
(0.097) | (0.136) | |||
Housing area per capita | 0.005*** | 1.005*** | ||
(0.001) | (0.001) |