The Effects of Public Pension on Elderly Life

The Effects of Public Pension on Elderly Life Taeil Kim & Jihye Kim Abstract In this study, we have attempted to clarify a variety of the effects of public pensions on elderly economic life. A quasi-experimental situation due to the time difference between the introductions of the two pensions was used to enable more logical causal reasoning compared to existing research by performing a comparison between experimental and control groups. Additionally, interesting evidence of elderly income-security behavior which had not been revealed by previous research could be uncovered. 1. Elderly pensions are very important for elderly income security. 2. Although people make more independent preparations for old age income in the absence of pension benefits, the amount is small. 3. The existence of an elderly pension somewhat reduces the work motivation in old age. 4. Old age income security followed by the existence of elderly pension has an important effect on the quality of elderly life. Ⅰ Purpose The purpose of this paper is to analyze the effect of old age pension on elderly life by using a quasi-experimental method. Old age pension plans in Korea are classified into two types one is the national pension(np) plan targeting the general public, the other is the special occupational pension(sop) plan targeting particular occupational groups such as public sector employees and teachers. The special occupation pension plan was introduced in 1960 and qualified retired employees have been receiving pension benefits for a considerable period. The national pension plan was put into effect for employees of enterprises of 10 or more members in 1988. The plan was expanded to cover the entire elderly population(excluding SOP beneficiaries) by including the self-employed in 1999. In general, the NP plan requires a minimum of 20 years of subscription to become a full pension recipient. Therefore, there are presently few recipients in the NP plan. The most important effect of a pension plan for the elderly would be that of income security. How could this be measured? One general method is to investigate the income of elderly pension - 1 -

recipients and determine what proportion of total income is pension benefits. However, this method presents the risk of overestimating the income security effect of the elderly pension because the remainder of the income after subtracting pension from total income would not be identical to the recipient s income if the plan were not to exist. Individuals would have consolidated greater preparations for their elderly life if the pension plan didn't exist, such as increased savings or subscription to private pension plans. Analysis of the current income of pension recipients is not the proper method for estimating the income security effect of an old age pension. The correct method is to compare the real income of recipients with the hypothetical income of recipients had the plan not existed. In other words, this must be accomplished through a comparison between the actual state and a counter-factual state of recipients' income. To uncover a causal relationship in the social sciences, the method of contrasting an experimental group to a comparison group is commonly used. In this case, the state of the comparison group is considered against the counter-factual state of the experimental group. Therefore, if a comparison group which could be considered to be in a counter-factual state to recipients -a group which does not receive an old age pension despite other conditions being similar to recipients- existed, the income security effect of the elderly pension could be estimated by comparing the incomes of the recipients and the comparison group. The problem lies in that such a comparison group is elusive because old age pension plans set their objectives as covering all citizens. Due to this problem, existing studies merely estimate the amount of elderly pension benefits as a portion of the total income of a recipient or estimate the effects through quantitative models based on hypothetical assumptions. However, unlike in previous studies, a comparison group can be set in this paper and we are attempting to reveal a range of effects of the elderly pension through contrasting experimental and comparison groups using a quasi-experimental method. Such a method is made possible by the uniqueness of South Korean elderly pension plans. As described above, the elderly pension plan in South Korea is classified into the national pension plan and the special occupational pension plan. In the case of the special occupational pension plan, the majority of - 2 -

qualified retired employees receive pension benefits due to its introduction in the relatively distant past. However, the national pension plan boasts few recipients due to its brief period of operation period. 1) While special occupational pension subscribers among seniors 60 or older are receiving an elderly pension, the large majority of senior citizens who do not belong to this group are unqualified for elderly pension benefits. The fundamental idea of this research is to set elderly pension recipients of the special occupational pension plan as an experimental group while reserving the large majority of the elderly as a non-recipient control group. The special occupational pension candidates are public sector employees, including teachers. 2) Therefore, the experimental group is made up of public sector employees. On the other hand, national pension candidates are private sector employees or the self-employed. It is desirable for as many conditions as possible other than the status of receipt of elderly pension benefits to be shared between experimental and control groups. However, the self-employed and employees of small companies are expected to express different characteristics than public sector employees in economic term. Accordingly, we have excluded the self-employed and employees of companies which have less than 100 employees, setting only employees of middle or large sized private companies as the control group. The research questions to be explored through an comparison of the experimental comparison groups are as follows. First, what differences does the status of receiving pension benefits bring to the incomes of the two groups? Second, what differences does the status of receiving pension benefits bring to the assets of the two groups? Third, what differences does the status of receiving pension benefits bring to the work behavior of the two groups? Fourth, what effects does the status of receiving pension benefits have on changes in satisfaction before and after the retirement of the two groups? 1) In order to become full beneficiaries of pensions, the subscription period must be 20 years or longer for both the national pension and the special occupational pension. While national pension beneficiaries granted early pensions and other special cases do exist, these are exceptions. 2) In the case of South Korea, there are two types of teachers; public teachers and private teachers. Although public teachers are considered public sector employees, private teachers do not possess such status. Because they are similar in many aspects, however, such as salary, pension plan and guarantee of status, private share a similar status as public sector employees. - 3 -

Ⅱ Analytical Methods and Data 1. Analytical Methods The analytical method of this research is a 'pretest-posttest comparison design of experimental and control groups'. This is illustrated in the figure below. <Figure 1> Pretest-posttest Comparison Design of Experimental and Control groups' Before retirement After retirement Public Sector Public Pension Plan Private Company For example, if satisfaction before retirement is A 1 and B 1, and satisfaction after retirement is A 2 and B 2, difference in satisfaction between public and private sector employees before retirement is (A 1 -B 1 ) and difference in satisfaction between public and private sector employees after retirement is (A 2 -B 2 ). Any difference in satisfaction before retirement between public and private sector employees is triggered by the difference in individual characteristics such as income or educational background. The difference in satisfaction after retirement is influenced by the status of receiving pension benefits as well as variations in individual characteristics between the groups. Accordingly, (A2-B2)-(A 1 -B 1 ) shows a difference in satisfaction influenced by status of receipt of pension benefits. In such a comparison, if there can be identified pre- and post-retirement data on the same individual, the effects of time-invariant individual characteristics can be completely controlled. However, as will be explained later, longitudinal data on a particular individual has not been established because the data for this research was gathered for the first time in 2006. Therefore, four groups consisting of public and private sector employees - 4 -

before and after retirement are compared at one point in time. In this case, the application of a regression model is more reasonable than the method of difference-in-difference on the mean values of four groups because the regression model may control some of the effects due to differences in individual characteristics. The regression model is as below: (1) In equation (1), is individual characteristic variables such as age, gender, and years of schooling. is a dummy variable which becomes 1 for the experimental group and 0 for the control group. is a dummy variable which becomes 1 for pre-retirement and 0 for post-retirement. is an error term. is a dependent variable and becomes degree of satisfaction in the example above. In this equation, the effect of receiving pension benefits becomes which is a coefficient of. In other words, presents the value of (A 2 -B 2 )-(A 1 -B 1 ) in the state in which the individual characteristics of each group are controlled. 2. Analytical Data The data used in this research is from the '2006 Korean Longitudinal Study of Ageing' by the Korea Labor Institute. While this survey was initially performed in 2006, it is expected to build panel data by being performed biannually. The subjects of the survey are citizens of the Republic of Korea who are 45 years of age or older. A sample of 10,254 people was established for the first-year data. The details of survey consist of individual characteristics(age, gender, years of schooling, etc), family, health, employment, income, assets, expectation and satisfaction with life, and more. Through this data, variables on age, current and past occupations, and status of receiving special occupational pension benefits are available. Using this information, four groups are constructed; public sector employees under 60 years of age, private sector employees under 60 years of age, special occupational pension recipients 60 years of age or older, and people 60 or older who worked for a private company before retirement and are - 5 -

non-recipients of the national pension plan. Ⅲ Analytical Results 1. Comparison of Individual Characteristics In <Table 1> and <Table 2>, individual characteristics of the experimental and control groups are presented. <Table 1> Comparison of Individual Characteristics of Experimental and Control Groups before Retirement Public Sector Private Company Age 51.6(4.67) 50.2(4.76) Male 0.79(0.41) 0.79(0.41) Years of Schooling 15.4(2.29) 13.9(2.81) Tenure 23.7(8.79) 16.4(10.37) Monthly Earnings (Thousand Won) 3,305(781) 3,351(2,048) Sample Size 82 227 standard deviations are in parentheses In <Table 1>, the earnings of private sector employees are higher than public sector employees. In particular, considering that public sector employees have longer schooling and employment periods, it can be seen that the earnings of private sector workers are considerably higher than those of public sector employees. <Table 2> Comparison of Individual Characteristics of Experimental and Control Groups after Retirement Public Sector Private Company Age 68.8(7.47) 64.7(10.43) Male 0.82(0.39) 0.71(0.45) Years of Schooling 11.0(4.82) 11.9(3.93) Tenure 27.0(12.15) 18.6(11.52) Retired Period 10.8(8.27) 12.5(9.35) Earnings before Retirement(Thousand Won) 1,671(1,444) 1,802(1,510) Working 0.04(0.19) 0.09(0.29) Sample Size 184 223 standard deviations are in parentheses - 6 -

In <Table 2>, the retired employees of private companies also show higher earnings immediately prior to retirement than those of retired public sector employees. This is consistent with the results in <Table 1>. 2. Comparison of Income before and after Retirement In <Table 3>, the income of public sector and private sector employee groups before and after retirement are compared in order to determine the extent to which the pension plan augments elderly income. <Table 3> Comparison of Income by Public Sector Employees and Private Company Employees before and after Retirement ( unit: thousand Won) Public Sector Private Company Total Income 3,438.76(923.47) 3,556.10(2321.86) Non Pension Income 3,438.76(923.47) 3,556.10(2321.86) Labor Income 3,305.28(781.82) 3,351.38(2047.63) Before Assets Income 124.02(399.19) 185.31(845.59) Social Security Income 0.00(0.00) 2.69(39.83) Private Transfer 9.45(38.71) 15.97(71.27) Other Income 0.00(0.00) 0.75(8.12) Sample Size 82 227 Total Income 1,862.37(1339.53) 779.91(2021.78) Pension Benefits 1,415.97(911.14) 76.41(250.08) Non Pension Income 446.40(878.60) 703.50(2022.44) Labor Income 97.28(498.00) 285.70(1022.64) After Assets Income 111.66(495.31) 203.95(1489.76) Social Security Income 26.94(122.08) 52.84(268.87) Private Transfer 208.26(565.77) 151.62(373.07) Other Income 2.26(30.72) 9.39(112.94) Sample Size 184 223 standard deviations are in parentheses In terms of total pre-retirement income, the income of private sector employees is somewhat higher: 3,556 thousand won compared to 3,489 thousand won for public sector employees. However, income after retirement for public sector employees is more than double that of the other group: 1,862 thousand won compared to 780 thousand won. If post-retirement income is classified into pension income and non-pension income, pension income for retired public sector employees is far higher, 1,416 thousand won for the public sector group and 76 thousand - 7 -

won for retired private sector employees. On the other hand, non-pension income of retired private sector employees, 704 thousand won as opposed to the 446 thousand won of retired public sector employees. It is also apparent that the elevated non-pension income of retired private sector employees is drawn from their greater labor and assets income. 3) In order to uncover how much the receipt of pension benefits increases elderly income, the regression analysis demonstrated in equation (1) was performed. Dependent variables are total income, labor income and assets income. The results of regression analysis are shown in <Table 4>, demonstrating that the receipt of pension benefits increases elderly income more than double. <Table 4> Regression Results of Income Total Income Labor Income Assets Income Constant 2717.49(0.000) 3323.00(0.000) 480.21(0.136) Retirement -2180.07(0.000) -2568.27(0.000) 163.73(0.180) Public -244.67(0.299) -84.89(0.611) -64.88(0.618) R P 1447.74(0.000) -40.51(0.848) 17.65(0.915) Age -22.45(0.021) -23.74(0.001) -8.54(0.112) Male 573.09(0.002) 702.97(0.000) -20.51(0.837) Schooling 108.57(0.000) 47.74(0.001) 10.74(0.319) R 2 0.35 0.61 0.01 Sample Size 716 P-values are in parentheses These results make clear that although those who do not receive pensions do make greater spontaneous preparations for old age compared to recipients of pensions, the amount is minuscule. Therefore, it may be concluded that the income of the elderly is very insufficient in the absence of pension benefits. 3. Comparison of Property before and after Retirement As previously shown, the earnings of private sector employees surpass those of workers in the public sector. Because private company employees receive little or no pension income after retirement, in contrast to public sector employees, their need for independent preparation for old age is 3) Because the standard deviations of labor income and assets income are very high, it can be presumed that labor income and assets income are disproportionately distributed. - 8 -

much greater than that of public sector employees. In this situation, a private sector employee would reserve more funds than a similar public sector employee and consequently would possess greater assets. This can be confirmed in part by the fact that assets income of private sector employees was shown to be higher than that of public sector employees, as evidenced in <Table 3>. However, most assets do not create income while in possession; income is generated only subsequent to disposal. Accordingly, it is difficult to determine differences in assets owned by the two groups purely by assets income, making it necessary to directly ascertain the amount of assets in possession. In the case of South Korea, the retirement grants received upon retirement are significantly larger in the private sector. In other words, although the private sector does not receive occupational pension benefits after retirement, retirees are presented with legally mandated grants. Therefore, retired private sector employees would see a higher increase in assets after retirement than public sector employees. The amount of assets possessed by each group is shown in <Table 5> <Table 5> Comparison of Assets by public and Private Sector Retirees (unit: Thousand Won) Public Sector Private Sector Net Assets (Total assets-total debt) 21,307(28,974) 22,283(36,845) Total assets 22,440(29,547) 24,798(35,717) Real estate 20,665(29,047) 22,556(33,455) Financial assets 1,775(4,869) 2,241(6,677) Total debt 1,133(2,911) 2,515(14,022) Sample Size 184 223 The private sector retirees are shown to hold more substantial net assets than public sector retirees, but the difference is small. Accordingly, evidence of private sector employees possessing greater assets due to retirement grants upon retirement is not found in the data. 4. Comparison of Intention to Work In order to examine the effect of public pension plan on the work incentive, the work motivation of public and private sector retirees have been analyzed. - 9 -

<Table 6> Comparison of Intention to Work between Public and Private Sector Retirees Intention to Work Public Sector Private Sector want to work 16(10.9) 31(15.6) don't want to work 131(89.1) 168(84.4) Total 147(100.0) 199(100.0) percentages are in parentheses The ratio of people who wish to work is 15.6% for private company retirees and 10.9% for public sector retirees. The reasons of wanting to work for 47 retired people who have responded that they want to participate in the labor are shown as the table below. <Table 9> Comparison of the reasons of wanting to work between Public and Private Sector Retirees Reasons Public Sector Private Sector I want to work as long as health allows 10(62.5) 15(48.4) it is helpful for living/i need money 2(12.5) 13(41.9) society still needs my ability 1(6.3) - to maintain health physical and mental 2(12.5) - staying at home is boring 1(6.3) 3(9.7) percentages are in parentheses The response 'because I want to work as long as health allows it' is shown as the highest in both groups with 62.5% for retired public sector employees and 48.4% for retired private company employees. The one showing largest difference is the response 'because it is helpful for living /because I need money'. There are much more cases of private company retirees hoping to participate in labor in order to earn 'living cost' than the public sector retirees. The ratio is 41.9% for private company retirees and 12.5% for public sector retirees. 4. Comparison of Changes in Satisfaction before and after Retirement The '2006 Korean Longitudinal Study of Ageing' measures satisfaction with life. This is made up of five categories, including 'health', 'economic status', 'relationship with spouse', 'relationship with children' and 'quality of - 10 -

life compared with people of the same age'. A score of 0 to 100 was given in each category, a higher score indicating a higher degree of satisfaction. The average degree of satisfaction in the four groups is shown in <Table 6>. <Table 6> Comparison of Satisfaction by Public Sector and Private sector Employee Groups before and after Retirement Before After 변수 Public Sector Private Sector Health 71.7(17.83) 73.58(15.85) Economic Status 69.3(17.34) 64.2(19.66) Relationship with spouse 80.5(16.58) 77.8(15.97) Relationship with children 83.3(13.15) 81.4(13.86) Quality of life compared with people of the same age 77.3(12.96) 73.9(15.34) Sample Size 82 227 Health 65.3(23.02) 61.1(24.38) Economic Status 63.0(21.38) 52.8(23.62) Relationship with spouse 78.7(17.21) 73.9(19.70) Relationship with children 78.4(17.35) 76.2(18.71) Quality of life compared with people of the same age 72.1(18.27) 65.5(20.37) Sample Size 184 223 Although private sector employees express somewhat higher satisfaction before retirement toward health than public sector employees, public sector employees scored higher than private sector employees in all remaining areas. After retirement, the satisfaction of public sector employees is higher than private sector employees in all five categories. Also, it is evident that the reported satisfaction of both groups decreases after retirement. Although the satisfaction of both groups lessened, the amount of decrease between the two groups varies. The rate of decline in satisfaction by the public sector employee group is lower than that of private sector employees. However, this result is drawn from simply subtracting the mean value - it is necessary to perform regression analysis in order to view this more clearly. The result of regression analysis is shown in <Table 7>. - 11 -

<Table 7> Regression Results of Life Satisfaction Health Econ. Status Spouse Children QoL Constant 67.44(0.00) 53.08(0.00) 58.81(0.00) 83.52(0.00) 64.20(0.00) Retirement -7.25(0.00) -6.51(0.01) -3.97(0.07) -2.68(0.18) -6.13(0.00) Public -3.01(0.26) 3.66(0.17) 1.63(0.47) 1.53(0.47) 2.28(0.31) R P 8.79(0.01) 7.75(0.02) 1.72(0.56) 1.68(0.54) 5.25(0.07) Age -0.20(0.07) -0.15(0.18) 0.11(0.30) -0.14(0.13) -0.04(0.63) Male 2.02(0.32) 3.17(0.12) 6.75(0.00) -0.59(0.72) 0.62(0.72) Schooling 1.05(0.00) 1.15(0.00) 0.60(0.00) 0.37(0.04) 0.82(0.00) R 2 0.11 0.12 0.06 0.04 0.09 Sample Size 716 Econ. Status: economic status, Spouse: relationship with spouse Children: relationship with children, QoL:Quality of life compared with people of the same age P-values are in parentheses As mentioned before, the coefficient of R P indicates the difference in change after retirement by public sector and private sector employee groups. It is apparent that the rate of decrease in satisfaction with 'health', 'economic status' and 'quality of life compared with people of the same age' is significantly lower among the public sector employee group than the private sector employee group. Although the rate of decrease is also lower for the public sector employee group compared to private sector employee group in the case of relationship with spouse and relationship with children, the difference is statistically insignificant. Ⅳ Summary of Results The summary of results analyzed to this point is as follows. 1. Elderly pensions are critical for elderly income security. 2. Although people make more independent preparations for old age income in the absence of pension benefits, the amount is small. 3. The existence of an elderly pension somewhat reduces the work motivation in old age. 4. Old age income security followed by the existence of elderly pension has an important effect on the quality of elderly life. - 12 -

In this study, we have attempted to clarify a variety of the effects of public pensions on elderly economic life. A quasi-experimental situation due to the time difference between the introductions of the two pensions was used to enable more logical causal reasoning compared to existing research by performing a comparison between experimental and control groups. Additionally, interesting evidence of elderly income-security behavior which had not been revealed by previous research could be uncovered. One of the important contributions of this study is to confirm the necessity of public pension plan. Some liberal economists insist on abolition of public pension plan. They argue that the preparation for elderly life can be made voluntarily through private savings or pension in case of middle class or above while the income is supported by the public assistance system for the low income class. However, the result of this study shows the voluntary preparation for elderly life is far from enough even in case of the middle class. In Korea, most private company employees belong to middle class. The average salary of these people are higher than public sector employees. However, the elderly income of retired private company employees is far less than the public sector employees receiving pension benefits and it is not sufficient to maintain a stable elderly life. World Bank recommends to maintain an income of about 70% of the average income before the retirement for a stable elderly life. But we can see that the elderly income of such standard cannot be accomplished without the existence of public pension. References - References will be added later - - 13 -