Inequity in Health Care Financing in Iran : Progressive or Regressive Mechanism ?

Objective: Having progressive health finance mechanism is very important to decrease inequity in health systems. Revenue collection is one of the aspects of health care financing. In this study, taxation system and health insurance contribution of Iranians were assessed. Materials and Methods: Data of 2012 household expenditures survey were used in this study, and payments of the families for health insurances and tax payments were extracted from the study. Kakwani index was calculated for assessing the progressivity of these payments. At the end, a model was designed to find the effective factors. Results: We found that taxation mechanism was progressive, but insurance contribution mechanism was very regressive. The portion of people living in urban regions was higher in the payments of insurance and tax. Less educated families had lower contribution in health insurance and families with more aging persons paid more for health insurance. Conclusion: Policy makers must pay more attention to the health insurance contribution and change the laws in favour of the poor.


Introduction
World Health Organization (WHO) emphasized financial protection for health as one of the three fundamental objectives of health systems.The main aim of financial protection is to ensure that all members of the society have similar access to both health and health care deliveries [1,2].Financial protection could not be provided unless the health system have financial resources and pool it in the society [3].Revenue collection is the process of collecting money from various sectors for health financing [4,5].Two of the main ways to collect revenues for health are health insurance premiums and taxes.To ensure that the society have an effective health financing system, it is important to pay attention to equity in health financing.Equity in revenue collection is as important as equity in health financing [6].Equity in revenue collection focuses on the fact that "Does each family take part in revenue collection in the proportion of its financial ability, or does the revenue collection not related with the families' income?".Similar to other definitions of equity, equity in revenue collection contains two aspects: vertical equity and horizontal equity.Vertical equity in revenue collection means that economically equal people must pay equally for health systems and horizontal equity means that unequal people must pay unequally for health system [7,8].The aim of financial protection is constructing a revenue collection system, which people pay for health insurances according to their ability to pay [9].It is important to emphasize that the ability to pay must be related to the payments for health insurances and taxes, not for health services.This relationship must not only be positive, but also people with higher ability to pay must pay higher proportion of their income for health [10].If this occurs in a health financing system, it is told that the health system financing is progressive.If rich and poor pay similar proportion of their income for health financing, the system is proportional and if the poor pay a greater proportion of their income for health financing, the system is regressive [11].
Iran's health financing system can be categorized as mixed financing system, where the government, health insurances, donors and other organizations collect revenues from people [12].Large percentages of Iran economy is based on oil revenues and the health system do not collect all of the health revenues from the people [13].Iran has two big health insurances: social security insurance and Iran health insurance organization.These two big insurances are subsets of the government and private health insurances do not have large contributions [14].The share of out of pocket health expenditures was very high in the country (52.5% in 2012, and the health financing system was not successful enough to protect people from financial risks.In the year of study (2012), nearly 8 million Iranians were not covered by health insurances.However, after the recent health system reform, the government tried to increase the coverage.In addition, tax and insurance revenues are completely separate from each other and government does not subsidize the health insurances by tax or oil revenues.Health budget is paid to the Ministry of Health directly and the insurances do not work as middle-organizations of tax revenues between government and patients.Despite several studies, which had analysed the inequality in the distribution of health deliveries, no studies were found about the equity in revenue collection for health system in Iran.Several studies assessed the equity in out of pocket health payments in Iran, but there were no evidence for other revenue collection sources such as insurance premiums and taxing.The main aim of this study was to assess the progressivity of revenue collection in health system of Iran.We assessed the inequity in tax and health insurance contributions and analysed the findings.Iran's new health reform, which is called "Iran Health evolution", has been implemented since 2014.Therefore, the results of this study are vital for health policy makers to know the situation of Iran before the reform.These findings could be compared with the results of future studies after the "Iran Health Evolution".

Materials and Methods
This was a descriptive-analytical study.Data of Iran household and expenditures survey in 2012 were used in this study.This study was approved ethically by ethics committee approval of Tehran University of medical Sciences, (TUMS) in January 2015.This survey is a random-clustered one, which is gathered scientifically all over the country by Iran Statistical Center (ISC).Household and expenditures data were one of the oldest surveys of Iran, which has been gathered since 1965 annually.After deleting 366 missing data, 12547 data remained.Missing data did not contain socioeconomic and income values and were not systematic.Some data were at household level and some others were at individual level.The individual level data contained age, sex and literacy.These data were merged and changed to household level.Age data were changed to the number of persons more than 65 years old and the number of persons less than five years old in the family.Sex changed to the number of females in the family and literacy changed to the number of illiterate persons in family.Data of tax, income and insurance contribution were at household level.Concentration index, Gini index and Kakwani index were calculated to show the inequality in the payments of families for health insurances (health insurance contribution) and tax.Since in Iran's taxation system, all of the tax revenues are collected together and are distributed to several sectors, there is no earmarked tax for health sector separately.Earmarked taxes contain such sale taxes that are levied for a special purpose (for example, taxing on cigarettes for health).Therefore, in this study, we assessed the inequality of whole tax payments, not the earmarked taxes.These taxes contained only direct taxes and no data for indirect taxes (such as value-added taxes, smoking levies...) were available.Insurance contribution contained all public, private and complementary health insurance payments (both voluntary and compulsory) and not contained out of pocket health expenditures or government payments for health and health insurances.
In a Lorenz curve, income is plotted from the poorest to richest.The Lorenz curve is compared with the equality line and a Lorenz curve placed nearer to equality line shows more equal distribution of income.GINI index is twice the value of distance between the Lorenz curve and the equality line.In concentration curve, tax and insurance contribution is plotted and ranked from the poorest to richest.Concentration index is twice the value of distance between the equality line and concentration curve.The Kakwani index is the difference between GINI index and concentration index.
The value of GINI index varies from zero to one.Zero indicates complete equality and one indicates complete inequality.The value of concentration index varies between -1 and +1, which -1 means that there is complete inequality in favour of the poor and +1 means that there is complete inequality in favour of the rich.The Kakwani index is the difference between concentration index and GINI index.It ranges between -2 and +1, which -2 means complete regressivity and +1 means complete progressivity.If the concentration is curve placed between the Gini curve and equality line, the value of Kakwani index will be negative and the financing mechanism will be regressive.This indicates that the percentage share of insurance contribution and tax payments is higher than the percentage share of income for the poor.At the end, a regression model was designed to show the effective factors of the inequality.R 2 statistics, coefficients and standard errors for the models were calculated in the Poisson regression.All statistical analysis were done using STATA SE.Version 13.1 (StataCorp LP, Texas, USA).

Results
In Table 1, the results of family income, insurance contribution and tax payments of each person are shown.The population of the study was divided into 5 quintiles.These quintiles ranked from the poorest to richest families.As shown in Table 1, family income and per person tax payments increased from poor to rich family quintiles, but insurance contribution decreased from poor to rich.In Figure 1, concentration curve and Lorenz curve of tax payments are shown.In the x-axis of Figure 1, the cumulative percentage of income ranked population are placed.In the Y-axis the cumulative percentage of income and tax payments are placed.As shown in Figure 1, the Lorenz curve lies between the equality line and concentration curve.Therefore, the tax mechanism was progressive.In Figure 2, the concentration and Lorenz curve for insurance contribution are shown.As shown in Figure 2, the concentration curve was above the equality  line and GINI curve was below it.These findings indicate that health insurance contribution is regressive.
In Table 2, the results of Gini index for income and concentration and Kakwani index and their standard errors for insurance contribution and tax payments are shown.In Table 2, the results for tax payments are placed on the right and the results for insurance contribution are placed on the left.The results of Table 2 confirmed the results of Figure 1.The Kakwani index for insurance contribution was negative, so revenue collection in health insurance had a regressive mechanism.The index for tax payments was positive and these results confirmed that the tax system of Iran had a progressive mechanism.
In Table 2, standard errors for Gini and concentration index are also presented.The standard errors of these variables showed that the findings were significant.In Table 3, the results of OLS models for tax payments and insurance contribution are shown.In Table 3, income was the annual income of each family, family was the number of family members, urb was the dummy variable of living in urban or rural regions, with the value of one for urban and zero for rural regions, fem was the number of females in each family, age5 was the number of children under 5 years old in each family, age65 was the number of persons more than 65 years old in each family, and illiterate was the number of illiterate persons in each family.The dependent variable for the tax payment model was tax payments per person, and in insurance contribution model, it was per person insurance contribution.As shown in Table 3, the relationship between income and tax payments was positive, but no relationship was found between the income and insurance contribution.The number of family members was not related to tax payments, but by increasing in the number of family members, the insurance contribution for each person increased.The coefficient of urb was significant and positive in both models and confirmed that people living in urban regions pay more for the health insurance and tax.The number of persons older than 65 years old had positive relationship with insurance contribution but the number

Discussion
This study has extended to a large reliable data to find equity health revenue collection in Iran.To the best of our knowledge, no similar studies were assessed the inequality of revenue collection in Iran.GINI index of Iran in 2012 was calculated as 0.4009.Estimations of the Central Bank of Iran showed that from 1997 to 2003, the GINI index of Iran varied between 0.40 and 0.43 [15].The World Bank estimations showed that GINI index for Iraq was 0.295, for Mexico it was 0.481, and 0.296 for Pakistan [16].Concentration index for tax payments was positive and indicated that people with higher income, had more tax payments.In addition, the Kakwani index was positive and showed that the tax payments of Iran were progressive.Progressivity of tax payments is vital for transferring money from the rich to the poor.Having progressive tax payment helps to improve the economical equity in the country [17].Taxation mechanism of Iran is in a way that higher income people must pay higher percentages of their income to the government.For insurance contribution, the concentration index was negative.This means that people with higher income, pay less for their health insurance.The Kakwani index was negative and indicated that the health insurance contribution was regressive.Low-income families not only face with more diseases compared to the richer, but they also face with financial barriers when they utilize health services.Findings of this study have shown that less income families pay more for health insurance and they not only bear the insurance contribution of themselves, but also pay the share of richer families.One of the reasons of this regressive insurance revenue collection arises from the type of the jobs of people.Lower income families are workers, government employees etc. [13].They are paid by the government and companies, and the insurance contribution of them are deducted from their salary.The poor and rich pay similar premiums to insurances and there is no law to make more contribution for the higher income peoples.
The results of the regression models have shown that urban population pay more money for their taxes and insurances.The awareness of the people living in urban regions is higher than others, so they are more willing to pay for health insurance [14].The negative coefficient of the number of illiterate persons in the family confirmed the rule of awareness on health insurance contributions.In addition, the number of mandatory payers for health insurance is higher in urban regions compared to rural population [12].The income of people living in urban regions is higher than others.Therefore, the levies that people pay in urban regions is higher than in rural regions [13].By increasing the number of persons older than 65 years old, the willingness of the people to buy health insurance would increase.In the insurance literature, it has been demonstrated that more risky people like to buy insurances.The coefficient of age 65 confirmed these findings.
In a study, Zare et al. [13] assessed Kakwani index in health care expenditures in Iran from 1984 to 2010.They used household expenditures survey data in the study and found that the Kakwani index was positive and health expenditures were progressive.It is important to note that in the household expenditures survey of Iran, only out of pocket expenditures were questioned from the households and the findings of Zare et al. [13] did not contain the overall health expenditures.In another study performed in four West African countries, authors assessed the progressivity in health care finance.They used the concentration curves and indices for health care payments and illness.They found that in all four countries, the payment pattern for health was regressive so that the lower income group bore greater burden of health payments.In their study performed in Uganda, Kwesiga et al. [18] assessed the taxes and out of pocket health expenditures.They found that the health financing system of this country was progressive and the rich paid a greater proportion of their income for health compared to the poor.Taxes were more progressive than out of pocket payments.In another study done in Kenya, Munge et al. found that the overall health financing mechanism was regressive [10].Crivelli et al. [19] evaluated Switzerland's health care financing progressivity.They used Kakwani index for this purpose.They found that the health financing system of this country remained regressive since the health reform of 1996.In his study, Ataguba [20] assessed the alcohol taxation in South Africa.He calculated the Kakwani index for overall alcohol tax as -0.353 and confirmed that the alcohol taxation mechanism was regressive.Baji et al. [17], in their study done in Hungary, assessed the changes in equity during the health care reform of 2007.They used household budget survey of the country.The Kakwani index was -0.22 for out of pocket expenditures, which was highly regressive.The informal payments, medical and pharmaceutical payments were also regressive.Akazili et al. [21] in Ghana found that the overall health care financing system of Ghana was progressive in 2005 and 2006.They found that the formal payments for health were progressive, but the informal payments were regressive.
In this study, we used income as the need indicator.We did not have data about the need to health services.For example, quality of life could be used as the need indicator.
Data used in this study, contained the family data level.We merged the individual level data to have more information about the factors affecting the inequality.We used tax payments as a financial resource for health.In fact, all of the tax payments were not used for health only a little share of them were used for health financing.In addition, tax revenues were used for all parts of economy and health system had a small share of it.The inequality in tax payments was extended to health system and calculated as an indicator for revenue collection of health.Out of pocket health payments were not analysed in this study, because it depends on the household income, not the need.The nature of out of pocket health expenditures is different from insurance payments and taxing.If insurance contribution and tax have progressive distributions, they are good for the society.This averment is not extendable for out of pocket (OOP) payments.The progressivity of OOPs is the result of the inability of the poor to pay for their health.
In conclusion in this study, some new important evidences about health care financing of Iran was assessed.Our results showed that taxation mechanism of Iran was progressive but insurance contribution mechanism was regressive.Iranian insurance policy makers must change the insurance law and increase the proportion of the rich.Using compulsory insurance with progressive imposition is vital in Iran's health insurance system.Increasing the proportion of the rich will increase the revenues of health insurances, and the financial barriers of health insurance system will be solved to some extent.For future studies, it is suggested to assess the progressivity of earmarked taxes for health in Iran.

Implications for policy makers
• The taxation mechanism of Iran is progressive and it is appropriate, but health insurance contribution is regressive.• Lack of appropriate law about health insurance payments of people led to the regressive mechanism.• Health insurance policy makers must change the law in the way that the insurance contribution becomes progressive.

Implications for public
• The community should be aware about the inequity of their payments for health and must force the policymakers to change the insurance law in favour of the poor.

Table 2 . Gini, concentration and Kakwani index for insurance contribution and tax pay- ments and their standard errors
*: significant at %95