Table 6 the results of regression household Total expenditure by OLS and Quantile regression.
Table6: OLS and Quantile Regression of Household Total expenditure - ATP (‘000 VND)
2012 2010
Quantile regression Quantile regression
OLS QR_25 QR_50 QR_75 OLS QR_25 QR_50 QR_75
Explanatory Variables b/se b/se b/se b/se b/se b/se b/se b/se
Income of household 0.189*** 0.273*** 0.401*** 0.538*** 0.084** 0.190*** 0.282*** 0.375***
(0.06) (0.00) (0.00) (0.00) (0.04) (0.00) (0.00) (0.00)
Gender of householder, Male=1 5320*** 6287*** 4749*** 3660*** 2750*** 3660*** 2744*** 2596***
(1,481) (791) (890) (1,003) (763) (357) (421) (409)
Age of householder (35.00) -176*** -104*** (11.00) 20 -59*** (20) 23*
(37) (25) (27) (30) (19) (11) (13) (12)
Household in urban areas, urban=1 20890*** 10166*** 12436*** 13120*** 13359*** 4996*** 5389*** 5652***
(2,261) (776) (883) (993) (1,651) (365) (416) (403)
Kinh/Hoa =1; Ethnic minority = 0 4257** 3742*** 2325* -269 4042*** 2165*** 1950*** 1245**
(2,019) (1,132) (1,244) (1,354) (720) (496) (581) (573)
Not completed Primary school, yes=1 7740*** 570.00 2,588.00 3605* 5534*** 1,099 2369*** 4088***
(1,914) (1,480) (1,667) (1,881) (906) (682) (787) (769)
Completed Primary school, yes=1 13067*** 3111** 4470*** 5145*** 7981*** 2039*** 3120*** 4208***
(2,373) (1,487) (1,676) (1,893) (1,146) (685) (786) (763)
Completed Lower-secondary school, yes=1 15116*** 4298*** 5359*** 6311*** 9602*** 3537*** 4295*** 5067***
(2,617) (1,551) (1,723) (1,928) (1,327) (694) (803) (777)
Completed Upper-secondary school, yes=1 21918*** 5028*** 8273*** 10922*** 12654*** 4061*** 4648*** 6311***
(3,588) (1,680) (1,886) (2,118) (2,013) (768) (879) (850)
Completed College/University, yes=1 40984*** 15487*** 17951*** 18226*** 26384*** 7118*** 7903*** 9886***
(7,030) (2,041) (2,287) (2,581) (4,106) (902) (1,037) (999)
Household in Red River Delta 879.00 2389** 2712** 2903* 1,598 (791) 329 689
(2,594) (1,218) (1,350) (1,498) (1,082) (535) (625) (602)
Household in Central Highland and North Moutain -9086*** -2566**
(2,285) (1,001)
Household in North Central Coast -4363* 1139 1560 5127*** -2451** 16 -79 631
(2,403) (1,195) (1,284) (1,368) (982) (515) (599) (575)
Household in East Highland 0 3011** 3195** 4876*** 0 -123 1092 1046
(.) (1,405) (1,594) (1,779) (.) (662) (772) (740)
Household in South East 1972 3546*** 2310 3125* 4817*** 1369** 1334* 3274***
(3,281) (1,370) (1,549) (1,722) (1,789) (611) (717) (687)
Household in Mekong River Delta -1648 4338*** 3449*** 6580*** 1073 745 891 1818***
(2,344) (1,183) (1,306) (1,444) (971) (533) (620) (596)
Constant 32890*** 19305*** 20510*** 22409*** 13059*** 8978*** 8965*** 8894***
(3,362) (2,057) (2,296) (2,601) (1,627) (924) (1,059) (996)
N 7161 7161 7161 7161 9176 9176 9176 9176
*significant at 10%, **significant at 5%, ***significant. Standard errors in brackets. Source: Estimation from panel data of VHLSSs 2012 and 2010
4.4.2. Average Per household Health Finance and Shares of Total Financing Table7 shows household health financing and total expenditure by quintile, with households ranked in ascending order of total expenditure. For each quintile, the first column displays the average household total expenditure including health care payments. The other columns show the same information for each source of health finance along with total health financing. All financing and consumption variables are expressed in terms of values per household in order to take economies of scale into account. The last line provides information for the whole population.
In year 2012, the first column inTable7 shows that the poorest quintile consumes, on average. Lowest quintile consumes 26,321 and the richest consumes 157,400.
When the population is taken as a whole (last line of the Table), equivalent gross consumption amounts to 74,944. The average financing increases with quintile for all other sources of financing,
The second part ofTable7 shows that the poorest quintile consumes, on average, 6.8 percent of total expenditure, whereas this amounts to 40.9 percent for the richest.
Inpatient expenditure appears to be borne mostly by the richest, as the first three quintiles contribute only 5, 9.1, and 14.2 percent, on average, whereas the last two contribute 22.6 and 49.1 percent, respectively. The financing share increases by quintile for all other sources of financing, but differences are, in general, less marked than for inpatient expenditure. In the case of insurance, the richest quintile (32.5 percent) contributes more than five times as much as the poorest one (5.8 percent).
In year 2010, the first column inTable 7 shows that the poorest quintile consumes, on average. Lowest quintile consumes 14,373and the richest consumes 80,396.
When the population is taken as a whole (last line of the Table), equivalent gross consumption amounts to 38,094. The average financing increases with quintile for all other sources of financing,
The second part of Table 6 shows that the poorest quintile consumes, on average, 7.3 percent of total expenditure, whereas this amounts to 40.6 percent for the richest.
Inpatient expenditure appears to be borne mostly by the richest, as the first three quintiles contribute only 4.3, 7.8, and 12.7 percent, on average, whereas the last two contribute 22.1 and 53.1 percent, respectively. The financing share increases by quintile for all other sources of financing, but differences are, in general, less marked than for inpatient expenditure. In the case of insurance, the richest quintile (35.3 percent) contributes about five times as much as the poorest one (7.3 percent).
The optional VHLSS weights were applied in this research. The Table thus displays the entire sources of financing, irrespective of their final contribution to the health system.
Table7: Average Per household Health Finance (‘000 VND) and Shares of Total Financing (%)
2012
Total household expenditure
Total expenditure
for health
Inpatient Outpatient OOP payments
Insurance premiums
Total expenditure
for daily activity
Food payments
Non-Food payments
Lowest quintile 26,321 1,530 472 626 1,451 79 24,790 19,779 3,491 2 46,583 2,496 864 944 2,305 191 44,087 34,555 6,738 3 64,727 3,784 1,350 1,462 3,497 287 60,943 47,618 9,397 4 89,411 5,444 2,143 2,086 5,077 367 83,967 65,863 12,666 Highest quintile 157,400 9,755 4,652 3,374 9,309 446 147,644 117,477 19,948 Total 74,944 4,477 1,832 1,657 4,208 269 70,467 55,606 10,210
Lowest quintile 6.8% 6.7% 5.0% 7.4% 6.7% 5.8% 6.9% 6.9% 6.7%
2 12.1% 10.8% 9.1% 11.1% 10.7% 14.0% 12.2% 12.1% 12.9%
3 16.8% 16.4% 14.2% 17.2% 16.2% 20.9% 16.9% 16.7% 18.0%
4 23.3% 23.7% 22.6% 24.6% 23.5% 26.8% 23.2% 23.1% 24.2%
Highest quintile 40.9% 42.4% 49.1% 39.7% 43.0% 32.5% 40.8% 41.2% 38.2%
Total 100% 100% 100% 100% 100% 100% 100% 100% 100%
2010
Total household expenditure
Total expenditure
for health
Inpatient Outpatient OOP payments
Insurance premiums
Total expenditure
for daily activity
Food payments
Non-Food payments
Lowest quintile 14,373 1,038 334 399 988 49 13,336 10,715 2,620 2 24,434 1,679 602 674 1,588 90 22,756 18,040 4,716 3 33,282 2,547 989 1,000 2,414 132 30,736 24,143 6,593 4 45,485 3,975 1,716 1,510 3,808 167 41,510 32,263 9,247 Highest quintile 80,396 9,077 4,121 3,456 8,837 240 71,320 54,472 16,848 Total 38,094 3,472 1,462 1,336 3,340 132 34,622 26,942 7,680
Lowest quintile 7.3% 5.7% 4.3% 5.7% 5.6% 7.3% 7.4% 7.7% 6.5%
2 12.3% 9.2% 7.8% 9.6% 9.0% 13.3% 12.7% 12.9% 11.8%
3 16.8% 13.9% 12.7% 14.2% 13.7% 19.5% 17.1% 17.3% 16.5%
4 23.0% 21.7% 22.1% 21.5% 21.6% 24.6% 23.1% 23.1% 23.1%
Highest quintile 40.6% 49.6% 53.1% 49.1% 50.1% 35.3% 39.7% 39.0% 42.1%
Total 100% 100% 100% 100% 100% 100% 100% 100% 100%
Source: Author
4.4.3. Distributional Incidence of Sources of Household Health Finance
Tables 8 analyze the progressivity of health financing. The first part of Table8 gives the average expenditure and financing share, by quintile, with households ranked in ascending order of total expenditure (ATP: ability to pay). Information related to expenditure gives an idea about income inequality: the greater the share of the richest quintiles, the greater is the inequality. The sources of financing show which part of the income distribution bears the financing: the greater the share of the richest, the more concentrated is the financing among the rich or, put differently, the more pro-rich the financing.
The second of Table8 provides measures of financing concentration and progressivity. The first line shows the financing concentration index, which indicates how financing is related to ATP. A positive value indicates that the rich bear a greater share of the financing than the poor (that is, pro-poor financing), whereas a negative value indicates the opposite. A concentration index not significantly different from 0 reflects no relationship between income and financing.
The Kakwani index is the key information in Tables 8. This index measures financing progressivity as the difference between the concentration index and the gross consumption Gini index. A positive value reveals that financing is more concentrated among the rich than the poor, which indicates progressivity. A simpler way to think about progressivity is that the financing budget share (that is, financing divided by ATP) increases with ATP.
In year 2012, all concentration indexes are positive, indicating that the better off contribute absolutely more to the financing of health care than do the poor. The concentration index is largest for inpatients (0.4122) and smallest for social insurance contributions (0.2696), suggesting that inpatients are the most progressive and social insurance contributions are the least so.
The Kakwani indexes for both inpatients (0.0681) and Out-of-pocket payment (0.0054) are positive, indicating progressivity. The Kakwani index is very close to 0
for food payments , total expenditure for daily activity and total expenditure for health, and is moderately negative for outpatient (-0.0205), social insurance contributions (-0.0745), non-food payments (-0.0102), indicating regressivity.
In year 2010, all concentration indexes also are positive, indicating that the better off contribute absolutely more to the financing of health care than do the poor. The concentration index is also largest for inpatients (0.4421) and smallest for social insurance contributions (0.3021), suggesting that inpatients are the most progressive and social insurance contributions are the least so.
The Kakwani indexes for both inpatients (0.1025) and Out-of-pocket payment (0.0715) are positive, indicating progressivity. The Kakwani index is very close to 0 for food payments. Like the results as 2012, in 2010 the social insurance contributions (-0.0375), food payments (-0.0182), indicating regressivity. But moderately possitive for outpatient (0.0701), indicating progressivity
Table 8: Distributional Incidence of Sources of Household Health Finance in Vietnam, 2012and 2010
2012
Equivalent household expenditure quintile
Equivalent household expenditure
Total expenditure for health
Inpatient Outpatient OOP payments
Insurance premiums
Total expenditure for daily activity
Food payments
Non-Food payments
Poorest 20% 6.8% 7.0% 5.5% 7.7% 7.10% 6.1% 6.7% 6.8% 6.4%
(standard error) (0.085) (0.303) (0.522) (0.447) (0.321) (0.426) (0.090) (0.092) (0.131)
Poorest 40% 18.8% 18.5% 15.3% 19.2% 18.36% 20.7% 18.8% 18.9% 18.9%
(0.163) (0.604) (1.048) (0.811) (0.637) (0.778) (0.167) (0.173) (0.256)
Poorest 60% 35.6% 35.2% 29.9% 36.6% 34.76% 41.5% 35.6% 35.5% 36.5%
(0.239) (0.960) (1.726) (1.253) (1.013) (0.974) (0.246) (0.255) (0.385)
Poorest 80% 58.7% 58.9% 52.9% 61.0% 58.27% 68.0% 58.7% 58.6% 60.3%
(0.301) (1.318) (2.532) (1.615) (1.394) (1.021) (0.313) (0.326) (0.502) Test Dominance
Against 45 line - - - - - - - -
Against Lorenz curve - + +/-
Concentration index (CI) 0.3440 0.3446 0.4122 0.3236 0.3495 0.2696 0.3441 0.3442 0.3339 (standard error) (0.013) (0.025) (0.016) (0.014) (0.013) (0.004) (0.004) (0.006)
Katwani index (K) 0.0006 0.0681 -0.0205 0.0054 -0.0745 0.0000 0.0001 -0.0102
(standard error) (0.020) (0.041) (0.023) (0.021) (0.017) (0.001) (0.003) (0.008)
2010 Equivalent household
expenditure quintile
Equivalent household expenditure
Total expenditure for health
Inpatient Outpatient OOP payments
Insurance premiums
Total expenditure for daily activity
Food payments
Non-Food payments
Poorest 20% 7.12% 6.16% 4.92% 5.98% 6.11% 7.40% 7.22% 7.50% 6.25%
(standard error) (0.074) (0.239) (0.388) (0.305) (0.245) (0.645) (0.080) (0.082) (0.112)
Poorest 40% 19.32% 15.91% 13.51% 15.97% 15.71% 20.78% 19.65% 20.24% 17.64%
(0.146) (0.483) (0.801) (0.627) (0.494) (1.282) (0.153) (0.156) (0.226)
Poorest 60% 35.93% 30.47% 27.14% 30.68% 30.08% 39.83% 36.46% 37.30% 33.58%
(0.218) (0.802) (1.392) (1.028) (0.821) (2.100) (0.227) (0.230) (0.357)
Poorest 80% 58.68% 52.96% 51.42% 52.05% 52.54% 63.32% 59.23% 60.14% 56.12%
(0.283) (1.211) (2.227) (1.543) (1.242) (3.035) (0.295) (0.295) (0.505) Test Dominance
Against 45 line - - - - - - -
Against Lorenz curve - - - + + -
Concentration index (CI) 0.3396 0.4068 0.4421 0.4097 0.4111 0.3021 0.3331 0.3214 0.3728 (standard error) (0.012) (0.021) (0.016) (0.012) (0.045) (0.004) (0.003) (0.006)
Katwani index (K) 0.0672 0.1025 0.0701 0.0715 -0.0375 -0.0065 -0.0182 0.0332
(standard error) (0.020) (0.037) (0.026) (0.020) (0.063) (0.002) (0.003) (0.007) Dominance tests: – indicates the 45-degree line/Lorenz curve dominates the concentration curve
+ indicates concentration curve dominates 45-degree line/Lorenz curve blank: non- dominate
a. Gini index for equivalent household expenditure.
4.4.4. Decomposition inequality of Household Total expenditure
Author of this study use equation (11) suggested by Wagstaff, van Doorslaer, and Watanabe (2003) to decompose ATP (total expenditure) inequality of households Vietnam in 2012 and 2010. A summary of the results is presented in Table9. The entries in each column are derived from equation (11) and give the elasticity of total expenditures (ATP) with respect to each factor, the concentration index for each factor, and the total contribution of each factor to the ATP concentration index.
The results are presented in Table9, both year 2012 and 2010.The large elasticities of ATP with respect to these factors are responsible for their large contribution to the ATP concentration index.
In year 2012, food and non-food expenditure have the largest elaticities, elasticities 0.772 and 0.162, contribute most inequality in ATP, 0.266 and 0.054, make highest share of contributions, 77.25% and 15.72%. Outpatient and inpatient have share of contributions 2.41% and 3.24% eventhough they have high concentration indices because they have very low elasticities
In year 2010, food and non-food expenditure have the largest elaticities, elasticities 0.7049 and 0.2068, contribute most inequality in ATP, 0.2265 and 0.0771, make highest share of contributions, 66.72% and 22.70%. Outpatient and inpatient have share of contributions 4.22% and 4.63% eventhough they have high concentration indices because they have very low elasticities
4.4.5. Decomposition inequality of Health Care
To decompose health care expenditure inequality of households Vietnam.
Author only uses financing varibles in health care to analyze, not including food and non-food expenditure. A summary of the results is presented in Table 9. To analyze more detailed, drugs, healthtools expense also used in this study.
The resultsare presented in Table 10, both year 2012 and 2010.
In year 2012,outpatient and inpatient have highest elasticities, 0.3756 and 0.3992. Next is drugs expenditure, 0.1505. So, contributions of inpatient and outpatient in inequality are too high. The last column is the share of contribution in inequality of healthcare expenditure.
In year 2010,outpatient and inpatient also have highest elasticities, 0.3961 and 0.4025. Next is drugs expenditure, 0.1506. So, contributions of inpatient and outpatient in inequality are too high. The last column is also the share of
contribution in inequality of healthcare expenditure
Table 10: Decomposition inequality of Health Carepayments
2012 2010
Elast CIs Contribution Share of
Contributions Elast CIs Contribution Share of Contributions
Outpatient expense 0.376 0.324 0.122 35.3% 0.396 0.410 0.162 39.89%
Inpatient expense 0.399 0.412 0.165 47.7% 0.403 0.442 0.178 43.74%
Insurance expense 0.061 0.270 0.016 4.8% 0.040 0.302 0.012 2.94%
Drugs expense 0.150 0.249 0.037 10.9% 0.151 0.330 0.050 12.22%
Healthtools expense 0.014 0.340 0.005 1.4% 0.011 0.441 0.005 1.21%
" Residuals" 0.000
0.000
Total 0.345 0.407
Source: Author
Table9: Decomposition inequality of Household Total expenditure
2012 2010
Elast CIs Contribution Share of
Contributions Elast CIs Contribution Share of Contributions
Outpatient expense 0.026 0.324 0.008 2.41% 0.035 0.410 0.014 4.22%
Inpatient expense 0.027 0.412 0.011 3.24% 0.036 0.442 0.016 4.63%
Insurance expense 0.002 0.270 0.000 0.13% 0.003 0.302 0.001 0.31%
Drugs expense 0.010 0.249 0.002 0.69% 0.013 0.330 0.004 1.29%
Healthtools expense 0.001 0.340 0.000 0.09% 0.001 0.441 0.000 0.13%
Food expense 0.772 0.344 0.266 77.25% 0.705 0.321 0.227 66.72%
Non-Food expense 0.162 0.334 0.054 15.72% 0.207 0.373 0.077 22.70%
" Residuals" 0.003 0.000
Total 0.345 0.340
Source: Author
4.4.6. Concentration Curves
Figure 9presents the Lorenz curve for household total expenditure gross of health payments along with the concentration curve for each source of household health financing, year 2012. It shows household Social Insurance Contribution, Inpatient and Outpatient payments, Out-of-pocket for health care
The Lorenz curve shows the cumulative share of consumption according to the cumulative share of population ranked in ascending order of consumption. For instance, only 20 percent of total consumption might come from the poorest 30 percent of the population. This curve provides us with a visual representation of household inequality: the farther the curve is from the 45° line, the greater is the inequality.
Concentration Curves for Health Payments and Lorenz Curve for Household Expenditure
Figure 9:Social Insurance Contribution, Inpatient and Outpatient payments, Out-of-pocket for health care
Source: Author
Insurance
inpatient
outpatient Lorenz
The concentration curves represent the cumulative share of health payments according to the cumulative share of population, again ranked in ascending order of consumption. For instance, the poorest 30 percentmight contribute only 10 percent to social insurance. These curves show how health financing varies according to consumption: the farther a curve is from the 45° line, the more the corresponding source of financing is borne by the richest households. For some sources of financing, the concentration curve might lie above the 45° line. In such cases, payments are more concentrated among the poorest households.
Furthermore, these graphs offer a powerful means of representing the effect of health financing on the distribution of household living standards. Indeed, whenever a concentration curve lies outside the Lorenz curve, this indicates progressivity.
However, a formal test of statistical dominance is required to conclude this definitively (see O’Donnell and others 2008, ch. 7).
The Results are presented in graph Figure 9. In graph Figure 9, the concentration curves for inpatients appear to lie outside the Lorenz curve, suggesting that this is progressive sources of finance. The curve for the social insurance appears to lie inside the Lorenz curve suggests regressivity for the richest households. However, the gap between the concentration and Lorenz curves is never wide.
4.4.7. Distribution of Health Payments
Figure 10 shows the average budget share of out-of-pocket health payments (that is, health payments divided by total expenditure) by quintile of total consumptions. This figure is a direct representation of the progressivity of health payments. These are progressive if their share ofhousehold consumption increases with consumption and are regressive in the opposite case. Finally, if their budget share does not vary with consumption, health payments are proportional to income.