Financial projection of the basic pension under selective eligibility criteria

Kim, Dohun Article Financial projection of the basic pension under selective eligibility criteria KDI Journal of Economic Policy Provided in Cooperation with: Korea Development Institute (KDI), Sejong Suggested Citation: Kim, Dohun (2025) : Financial projection of the basic pension under selective eligibility criteria, KDI Journal of Economic Policy, ISSN 2586-4130, Korea Development Institute (KDI), Sejong, Vol. 47, Iss. 1, pp. 95-134, https://doi.org/10.23895/kdijep.2025.47.1.95 This Version is available at: https://hdl.handle.net/10419/314772 Standard-Nutzungsbedingungen: Die Dokumente auf EconStor dürfen zu eigenen wissenschaftlichen Zwecken und zum Privatgebrauch gespeichert und kopiert werden. Sie dürfen die Dokumente nicht für öffentliche oder kommerzielle Zwecke vervielfältigen, öffentlich ausstellen, öffentlich zugänglich machen, vertreiben oder anderweitig nutzen. Sofern die Verfasser die Dokumente unter Open-Content-Lizenzen (insbesondere CC-Lizenzen) zur Verfügung gestellt haben sollten, gelten abweichend von diesen Nutzungsbedingungen die in der dort genannten Lizenz gewährten Nutzungsrechte. Terms of use: Documents in EconStor may be saved and copied for your personal and scholarly purposes. You are not to copy documents for public or commercial purposes, to exhibit the documents publicly, to make them publicly available on the internet, or to distribute or otherwise use the documents in public. If the documents have been made available under an Open Content Licence (especially Creative Commons Licences), you may exercise further usage rights as specified in the indicated licence. https://creativecommons.org/licenses/by-sa/4.0/ KDI Journal of Economic Policy 2025, 47(1):95-134 https://dx.doi.org/10.23895/kdijep.2025.47.1.95 95 Financial Projection of the Basic Pension under Selective Eligibility Criteria† By D OHUN K IM * This study conducts a financial projection of the basic pension in Korea, which provides cash assistance to the bottom 70% of elderly individuals aged 65 and over. The projection is carried out under both expansion and selective eligibility criteria, with particular emphasis on the latter. Specifically, the study examines two well-discussed selective eligibility criteria: 1) fixing the eligibility threshold at the 2024 value and adjusting it according to the inflation rate, and 2) linking the eligibility threshold to the median household income figure. To estimate the number of recipients under these selective scenarios, the study projects the future income evolution of the elderly, assuming the continuation of past income trends. Using the financial model of the basic pension developed by Shin and Kim (2021), the study finds that total fiscal spending could be reduced by 22% under the first selective scenario and by 17% under the second selective scenario, relative to the current system, in real terms between 2024 and 2070. With these fiscal savings, the study concludes that the full benefit amount could be increased from the current level of 334,810 won to 435,000 won under the first scenario and to 405,000 won under the second scenario by 2025. Key Word: Basic Pension, Selective Eligibility Criteria, Pension Reform, Elderly Poverty, Financial Model of Basic Pension JEL Code: H53, H55, I31, J10 * Associate Fellow, Korea Development Institute (E-mail: [email protected]) * Received: 2024. 8. 30 * Referee Process Started: 2024. 9. 10 * Referee Reports Completed: 2024. 10. 22 † This paper is an extension of Chapters 2 from KangKoo Lee, DoHun Kim and Seung-Ryoung Shin, 2023, Refrom Measures for Enhancing the Sustainability of the Public Pension System, Policy Report 2023-08, Korea Development Institute (in Korean). I sincerely express my gratitude to YongOk Choi, YoungWook Lee, WooRim Kim and two anonymous referees for their helpful comments and suggestions. I am also grateful for SeokHyun Choo for his excellent research assistance. All remaining errors are solely my responsibility. 96 KDI Journal of Economic Policy FEBRUARY 2025 I. Introduction lthough Korea has experienced rapid economic growth over the past few decades, the poverty rate among the elderly ranks at the highest level among OECD countries. One of the main reasons for this high poverty rate can be attributed to the immature national pension system of Korea, which was introduced only in 1988. Due to the short history of the pension system, older generations did not have enough time for the national pension benefit to mature and were unable to prepare for an adequate amount of retirement income. To tackle the high poverty rate of the elderly, the government implemented a basic pension, a non-contributory pension system that provides monthly cash payments to elderly individuals aged 65 and above whose recognized amount of income (hereafter, RAI) falls below the top 70% of their age group.1 From 2014 to 2024, the maximum benefit amount of the basic pension has increased by nearly 62%, from 200,000 won to 334,810 won. Moreover, due to the aging population, the number of basic pension recipients has increased by 45%, from approximately 435,000 to nearly 624,000 recipients. As both the benefit amount and the size of the elderly population aged 65 and above have increased significantly, the expenditure on the basic pension has also risen substantially. As of 2023, the basic pension system has become one of the largest welfare programs in Korea, accounting for 1% of GDP, and it is projected to increase even further. Recently, the current selection criteria of the basic pension, which targets the poorest 70% of the elderly aged 65 and above, have received criticism for not being a cost-effective way to address the high poverty rate among the elderly. In particular, there is no clear rationale behind why the basic pension should target 70% of the elderly population despite the fact that the economic condition of the new elderly generation reaching age 65 has improved significantly. For instance, while the bottom 70% threshold income amount that determines the eligibility for the basic pension was only 56% of the median income in 2015, it increased to 94% of the median income by 2023, implying that the elderly covered by the basic pension system have become more affluent relative to the past. The OECD pension report on Korea suggested a narrowing down of basic pension recipients and an increase the benefit amount to tackle the poverty rate of the elderly more effectively (OECD, 2022). The Basic Pension Adequacy Evaluation Committee, an advisory body to the Minister of Health and Welfare, also suggested that the current selection criteria be transformed from a target population method to absolute criteria so that the number of recipients can gradually decline as the economic condition of the new elderly generation improves relative to the overall population. There have been several specific proposals by different researchers to make the selection criteria more selective. For instance, Yun (2023) suggested fixing the eligibility threshold amount at the year of 2023 and adjusting it according to the 1The “Recognized Amount of Income” is the weighted average of various monthly income sources and assets converted into a monthly income value, which is then used to determine eligibility for the basic pension. A VOL. 47 NO. 1 Financial Projection of the Basic Pension under Selective Eligibility Criteria 97 inflation rate. Others suggested setting the eligibility threshold line relative to the median income (Basic Pension Adequacy Evaluation Committee, 2023). Despite the importance of conducting financial projections based on selective scenarios, there has been limited research on this subject. While NABO (2018; 2023a) considers a selective scenario in which the benefit amount varies across different RAI levels among the elderly, it does not perform a financial projection for a scenario in which the target group is gradually narrowed. Although Ryu et al. (2022) does consider a scenario in which the target group is gradually reduced, the size of the target group is artificially adjusted from 70% to 30%. However, under the aforementioned selective eligibility criteria, the number of recipients is determined by the evolution of the RAI distribution among the elderly and the median household income. Therefore, it is crucial to compute the size of the recipient group within the model based on this mechanism. One approach is to forecast the evolution of both the RAI distribution and median household income. In this study, utilizing the financial projection model of basic pension developed by the Shin and Kim (2021), I conduct a financial projection under both expansion scenarios, in which the benefit amount and size of the recipient pool increase relative to the base case and selective scenarios in which the number of recipients is narrowed down. In particular, this study focuses on two well-discussed selective scenarios: 1) fixing the eligibility threshold amount at the year of 2024, and adjusting it according to the inflation rate (selective scenario 1), and 2) linking the eligibility threshold at the median household income (selective scenario 2).2 These selection scenarios have strength in that as the economic condition of future generations improves, the number of recipients will gradually decrease. Furthermore, existing recipients will continue to receive benefits, therefore facing less political resistance as opposed to if the number of recipients was dramatically reduced. This study makes a strong contribution to the literature by linking the selection criteria for basic pension eligibility to the evolution of the RAI distribution and median household income. To achieve this, it projects the future distribution of the RAI assuming that past trends will continue into the future. Specifically, it calculates the RAI for each percentile of the elderly population from 2014 to 2021 using data from the National Survey of Tax and Benefit (NaSTaB) and predicts the future evolution of the RAI for each percentile by means of linear projection. Similarly, the future evolution of median household income is forecasted based on the assumption that historical trends in the median income will persist. This information is then used to estimate the proportion of recipients under selective scenarios, serving as input for the basic pension model developed by Shin and Kim (2021) to make financial projections of expenditures. The findings demonstrate that under the current system, the total accumulated fiscal expenditure between 2024 and 2070 amounts to 1,891 trillion won in real terms. Under the expansion scenario, where the benefit amount increases from 334,810 won to 400,000 won, the total accumulated fiscal expenditure rises to 2,236 trillion won, which is 18% higher than that of the current system. Moreover, under 2The eligibility threshold for a single household is 2.13 million won, and for couple households, it is 3.408 million won in 2024. In this study, as will be explained in the main section, the eligibility threshold for couple households is used to project the size of recipients under selective scenarios. 98 KDI Journal of Economic Policy FEBRUARY 2025 the expansion scenario, where the eligibility criteria are expanded to include all elderly individuals on top of the increase in the benefit amount from 334,810 won to 400,000 won, the total accumulated fiscal expenditure rises to 3,191 trillion won, which is 69% higher than that of the current system. On the other hand, under the selective scenarios, as the proportion of recipients gradually declines from 70% to nearly 50% of elderly individuals aged 65 and above between 2024 and 2070, the total accumulated fiscal expenditure also decreases significantly. Specifically, the total accumulated fiscal expenditure declines by 22% relative to the current system under selective scenario 1 and by 17% under selective scenario 2. Utilizing the savings from each selective scenario, this study demonstrates that the full benefit amount of the basic pension can be increased to 435,000 won and to 405,000 won in 2025 under selective scenarios 1 and 2, respectively, without incurring additional fiscal spending compared to the current system. These results indicate that implementing selective eligibility criteria can provide the government with the financial flexibility to increase benefit amounts by specifically targeting the poorer group in relative terms among the elderly. The paper is divided as follows: section 2 provides an overview of the basic pension, including the history of the institution and detailed information pertaining to the selective eligibility criteria. Section 3 explains the forecast model of the basic pension and the method used to forecast the future distribution of the RAI using data from the National Survey of Tax and Benefit. Section 4 presents the financial projection results of the basic pension under both expansion and selective scenarios. Finally, section 5 discusses the results and concludes the paper. II. Overview of Basic Pension Income A. History of the Basic Pension Due to its relatively short history, the national pension system in Korea, which was only introduced in 1988 and later extended to cover the entire population, has not yet matured enough to serve fully as an old-age income security system. Given that a large portion of the elderly did not have sufficient time to contribute to the system and secure an adequate pension, the government introduced a cash assistance program in 1991 to supplement the income of the poor elderly in the form of an oldage allowance. However, this program was criticized for its limited effectiveness in reducing the high elderly poverty rate due to the small benefit amounts and restricted target group. In 2008, the government introduced what was termed the basic old-age pension system, the predecessor of the basic pension.3 It is a non-contributory public pension program that provides monthly cash assistance to elderly individuals whose RAI falls within the bottom 70% among those aged 65 and over. The benefit amount was set 3The major difference between the old-age pension and the basic pension in terms of system operation is that, under the former, government had the discretion to reduce the number of recipients based on budget constraints. VOL. 47 NO. 1 Financial Projection of the Basic Pension under Selective Eligibility Criteria 99 to the 5% of the A value, which is the average salary of national pension contributors over the last three years, and it included a 20% reduction rate for the married-couple households. As of 2014, the benefit amount was 99,100 won for single households and 158,600 won for married-couple households. It also introduced an incomereverse prevention-reduction scheme which reduces the basic amount to prevent cases in which a beneficiary’s income exceeds the income of non-beneficiaries after accounting for the basic pension. The discussion on introducing the basic pension system began due to growing social concern over elderly poverty and gained momentum during the 18th presidential election when candidates proposed introducing the system to expand income security for the elderly. As a result, the basic pension system was introduced following the enactment of the Basic Pension Act, which expanded the existing basic old-age pension system. While maintaining the eligibility criteria for elderly individuals aged 65 and over whose RAI falls within the bottom 70%, the basic pension system further enhanced old-age income security by increasing the full benefit amount from 5% to 10% of the A value. Consequently, the full benefit amount increased to 200,000 won for single households and 160,000 won for married-couple households. While retaining the income-reversal prevention-reduction scheme, the basic pension also introduced a reduction scheme in the benefit calculation for recipients who concurrently receive the national pension. The reduction rate varied according to the A value of the national pension. The history of the basic pension is summarized in Table 1. TABLE 1—HISTORY OF THE BASIC PENSION System Old-Age Allowance Senior Pension Basic Old-Age Pension Basic Pension Year of Implementation 1991 1998 2008 2014 Beneficiaries Low-income seniors aged 70 and above → seniors aged 65 and above who are eligible for livelihood protection Low-income seniors aged 65 and above Seniors aged 65 and above in the bottom 70% income bracket Seniors aged 65 and above in the lowest 70% income bracket Benefit Amount 35,000~50,000 won 20,000~50,000 won 5% of the A value of the National Pension 10% of the A value of the National Pension Reduction - 12.5% reduction for a couple two beneficiaries 20% reduction for a couple of two beneficiaries, income-reversal prevention adjustment 20% reduction for a couple of two beneficiaries, income-reversal prevention adjustment, reduction linked to the National Pension Benefit Adjustment - - Adjusted according to the National Pension A value Linked to the national consumer price index (CPI) fluctuation rate; adequacy evaluation conducted every five years Source: NABO, “2018-2027 Estimation of the Expenditure for the Basic Pension,” 2018, p.10. 100 KDI Journal of Economic Policy FEBRUARY 2025 TABLE 2—FORMULA FOR THE RAI CALCULATION RAI = Income Assessment Amount 1) + Assets Converted Income 2) 1) RAI = {0.7 * (Labor Income – Labor Income Deduction amount) + other income* * The labor income deduction amount is the amount of income that is deducted to ensure that work motivation is not hindered and it is based on the minimum income level (1.8 million Korean won as of 2023). ** Other income is the sum of business income, public transfer income, imputed rent income and property income, with an annual deduction of 40,000 won applied to interest income included in property income. 2) Asset-converted income = [{(general assets – basic asset amount*) + (financial assets – 20 million won) – liability}× income conversion rate for assets** ÷ 12] + P*** * The basic asset amount refers to the minimum amount required for housing maintenance (as of 2023, metropolitan cities: 135 million won, medium-sized cities: 85 million won, rural areas: 72.5 million won) ** 4% annual rate *** value of luxury automobiles (over 3,000cc or over 40 million won) and membership rights Source: Ministry of Health and Welfare, “2023 Basic Pension Program Guidelines,” 2023. The eligibility criteria for the basic pension are based on the elderly whose RAI falls below the 70th percentile among those aged 65 and over. However, civil servants and military pension recipients, as well as their spouses, are excluded from eligibility. The RAI, which serves as the basis for determining eligibility, is calculated by summing the assessed income amount, including labor, business, and asset income, as well as the income-converted amounts of assets. The exact process for calculating the RAI is detailed in Table 2. RAI amounts are calculated separately for single-person households and married-couple households. Meanwhile, as the economic conditions and national pension benefits for the new elderly generation have improved, the eligibility threshold, which is set at the 70th percentile of the RAI among those aged 65 and over, has continued to increase, as shown in Figure 1. For instance, in 2014, the eligibility threshold for single-person households was 870,000 won, and for married-couple households, it was 1,392,000 won. By 2024, these amounts had increased to 1,392,000 won and 3,230,000 won, respectively. The full basic pension amount refers to the maximum pension that recipients can receive if they are not subject to any reduction schemes, such as 1) a NationalPension-linked reduction, 2) the income-reversal prevention reduction, and 3) a spousal reduction. The National-Pension-linked reduction applies to recipients who receive a national pension benefit that exceeds 150% of the full basic pension amount. The proportion of national pension recipients among basic pension recipients increased from 30.4% in 2014 to 46.6% in 2022 (Kim, 2023). The incomereversal prevention reduction is applied when the sum of the benefit amount and pre-benefit RAI exceeds the eligibility threshold; this is done to minimize the income reversal between beneficiaries and non-beneficiaries. Approximately 3% of recipients are subject to the income-reversal prevention reduction. The spousal reduction applies to married couples, imposing a 20% reduction on each individual’s benefit. The full basic pension amount is adjusted relative to the inflation rate and has been gradually increasing due to a few jumps in 2018 and 2021. While the full basic pension amount is fixed at the 10% of the A value, which is the average salary of VOL. 47 NO. 1 Financial Projection of the Basic Pension under Selective Eligibility Criteria 101 national pension contributors in last three years, it increased to 11% of the A value (250,000 won) in 2018 and 11.8% of the A value (300,000 won) in 2021, as shown in Figure 2. The number of basic pension recipients increased gradually from 4.35 million in 2014 to 6.23 million in 2022 due to rapid aging, as shown in Table 3. As a result, the financial expenditure of the basic pension has more than doubled, rising from 9.68 trillion won in 2015 to 20.09 trillion won in 2022. (Unit: 10,000won) Single Households Couple Households FIGURE 1. ANNUAL ELIGIBILITY THRESHOLD OF THE BASIC PENSION Source: NABO, “2018-2027 Estimation of the Expenditure for the Basic Pension,” 2018, p.10; Ministry of Health and Welfare, “Basic Pension Program Guidelines,” 2014-2024. (Unit: won) FIGURE 2. ANNUAL FULL BENEFIT AMOUNT OF THE BASIC PENSION Source: NABO, “2018-2027 Estimation of the Expenditure for the Basic Pension,” 2018, p.12; Public Pension Reform and Long-term Projections I, 2023a, p.14. 102 KDI Journal of Economic Policy FEBRUARY 2025 TABLE 3—NUMBER OF BENEFICIARIES AND FINANCIAL STATUS OF THE BASIC PENSION EXPENDITURE (Unit: Thousands, 10 Billion won) Year 2014 2015 2016 2017 2018 2019 2020 2021 2022 Number of Beneficiaries 4,353 4,495 4,581 4,869 5,216 5,346 5,660 5,973 6,239 Expenditure 462 968 991 1,053 1,196 1,478 1,669 1,890 2,009 Source: NABO, 2018~2027 Estimation of the Expenditure for the Basic Pension, 2018, p.8; Public Pension Reform and Long-term Projections I, 2023a. pp.13~14. B. Selective Eligibility Criteria The purpose of the basic pension was to provide a secure income for poor elderly people in order to address the high elderly poverty rate in Korea, which is mainly due to the immature national pension system. However, the current eligibility criteria for the basic pension, which targets the bottom 70% of the RAI among those aged 65 and over, is not cost-effective with regard to how well it addresses the high poverty rate among the elderly. In that the basic pension targets 70% of the elderly population, increasing the benefit amount of the basic pension will result in large fiscal expenditures in the future. For example, fiscal expenditure on the basic pension is projected to reach 1.5% of GDP by 2050 and 1.2% of GDP by 2070, even if the benefit amount remains fixed in real terms at the current level due to Korea’s rapid aging population (NABO, 2023a). Although the basic pension system has a reduction mechanism that reduces the benefit amount for those whose national pension income exceeds 150% of the maximum basic pension benefit, this reduction system has become less effective in fulfilling its role, as the maximum basic pension benefit has increased substantially, outpacing the increase in national pension income such that only a small proportion of the elderly receiving a national pension are actually subject to the reduction system (Ryu et al., 2022). Moreover, as the national pension system has matured and economic conditions have improved, relatively economically well-off elderly households have become eligible for the basic pension. In particular, the eligibility threshold, which is the bottom 70% of the RAI, has been rising. For example, while the bottom 70% of the RAI among the elderly amounted to only 50% of the median household income in 2015, it increased to almost 93% of the median household income in 2023. At the current rate, the eligibility threshold is projected to exceed the median household income level. Therefore, in order to make the basic pension system more cost-effective in terms of its ability to address the high poverty rate in Korea, one approach is to make the eligibility criteria more selective by setting an absolute eligibility threshold. Currently, several proposals in relation to this exist in the literature. For example, Yun (2023) suggested setting the eligibility criteria at the level of 2023 and then increasing the benefit amount according to the inflation rate. The Basic Pension Adequacy Evaluation Committee suggested setting the eligibility line at 50% or 100% of the median household income. In both cases, the number of beneficiaries will decrease compared to the current system as the income of the elderly improves. VOL. 47 NO. 1 Financial Projection of the Basic Pension under Selective Eligibility Criteria 109 pension benefit – i.e., early national pension, special pension, and normal old-age pension with less than and at least 20 years of contributions - it is important to classify recipients into these four types. The information pertaining to the proportion of recipients and the average benefit amount for each type by age and gender is taken from the 2020 National Pension Annual Report. I then use the ratio of the A benefit to the national pension benefit amount to calculate the A benefit from the average benefit amount for each type by age and gender (Shin and Kim, 2021). Then, the benefit amount for concurrent recipients can be calculated by applying the average benefit amount by gender and age and the A benefit to the reduction formula of the basic pension benefit for each type of national pension. The fiscal expenditure of the basic pension is projected separately for each of the eight types of recipients under the assumption that the proportion of each type remains constant at the June 2021 level, as provided by Shin and Kim (2021). To predict future trends in the proportions of each type, administrative data from a source such as the National Pension Database, which is not publicly accessible, would be required. The specific procedures for the fiscal projection are included in the Appendix. C. Previous Financial Projection Scenarios and Results Previous studies presented financial projection results for the basic pension based on two scenarios: universal eligibility criteria, where the basic pension is expanded to all elderly, and selective eligibility criteria, where the target group gradually shrinks over time and the benefit amount varies according to the RAI. For example, NABO (2018; 2023a) projects that the financial expenditure of the basic pension will increase to 109.8 trillion won under the current system. In the universal eligibility scenario, where the criteria are expanded to include all elderly aged 65 and over and the benefit amount increases to 400,000 won as of 2024, the additional financial expenditure relative to the current system is projected to amount to 61 trillion won by 2090. While NABO (2018; 2023a) did not conduct a financial projection under selective criteria where the target group gradually shrinks, it did consider selective criteria where the benefit amount is differentially paid according to the RAI. For instance, while maintaining 70% of the elderly as the eligibility criteria, it considered a selective case in which the benefit amount increases to 450,000 won for those in the bottom 40% in terms of the RAI and 400,000 won for the remaining 30% of the elderly. These results show that by 2090, the expenditure declines by 15 trillion won compared to the universal criteria case. Ryu et al. (2022) conducted financial projections under selective criteria, where the size of the recipient pool shrinks only for the new generation of elderly who begin receiving the basic pension, while ensuring a minimum income level for the poor elderly by introducing a supplemental income program on top of the basic pension. Specifically, their study considered a scenario in which the basic pension amount is adjusted to 12% of the A value starting in 2023, equivalent to 326,000 won in 2023, while gradually lowering the eligibility criteria from the bottom 70% in terms of the RAI to the bottom 30% of the RAI. Under this scenario, the study shows that under 110 KDI Journal of Economic Policy FEBRUARY 2025 the current system, the projected expenditure amounts to 360 trillion won by 2092, whereas under the selective criteria, 195 trillion won could be saved compared to the current case by 2092. The selective criteria considered in earlier studies are limited in that they artificially reduce the size of the recipient pool over time, rather than linking it to the income distribution of future generations. The size of the recipient pool under the currently discussed selective criteria, which either fixes the threshold at the 2023 value in real terms or links it to median household income, depends on the future evolution of the income distribution among the elderly. Therefore, this study uses the basic pension financial projection model by Shin and Kim (2021) to conduct financial projections under selective criteria, where the size of the recipient pool is determined by the evolution of the RAI distribution over time. This projection is made under the strong assumption that the future linear trend of the RAI distribution will remain consistent with the trend observed between 2014 and 2021. D. Forecasting the Number of Recipients under Selective Criteria Previous studies conducted financial projections of the basic pension under selective criteria by artificially reducing the pool of basic pension recipients over time. However, this approach has a limitation in that the pool of basic pension recipients is exogenously determined at the discretion of the researcher rather than through the actual process involving selective eligibility criteria. Here, I explain the specific procedures for forecasting the proportion of basic pension recipients under the two aforementioned types of selective eligibility criteria: fixing the eligibility RAI threshold at the 2024 level and linking it to the inflation rate (selective scenario 1), and setting the eligibility RAI threshold to 100% of the median household income for two-person families (selective scenario 2). The critical information needed to predict the proportion of basic pension recipients under these selective scenarios is the future distribution of the RAI of the elderly, as the evolution of the number of basic pension recipients will depend on the evolution of the future distribution of the RAI. First, the RAI for each elderly household between 2014 and 2021 can be measured using National Survey of Tax and Benefit (NaSTaB) data from 2014 to 2021 based on the RAI formula for single and couple households, as described in the Appendix. Figure 3 shows the 70th percentile of the measured RAI using panel data and the 70th percentile of the actual RAI, representing the publicly announced eligibility criteria for the basic pension by the government between 2014 and 2023 for single and married-couple households. It turns out that the accuracy of the 70th percentile of the measured RAI is much higher for couple households compared to single households. For instance, in 2022 and 2023, the measured 70th percentile RAI corresponds to 2.93 million won and 3.13 million won for couple households, very similar to the actual 70th percentile of the RAI, 2.88 million won and 3.23 million won, respectively. However, for single-person households, the discrepancy between the 70th percentile of the measured RAI and the actual RAI is relatively large compared to that for married-couple households. Hence, for simplicity, throughout the analysis here, I focus on married-couple households when estimating the future VOL. 47 NO. 1 Financial Projection of the Basic Pension under Selective Eligibility Criteria 111 (Unit: 10,000won) Single-person household Couple household FIGURE 3. ELIGIBILITY THRESHOLD OF THE BASIC PENSION: 70TH PERCENTILE OF RAI FOR EACH HOUSEHOLD TYPE Source: Calculated using the data from the 7th to 14th waves of the NaSTaB survey by the Korea Institute of Public Finance. distribution of the RAI and the proportion of recipients. I assume that the future proportion of single-person households will follow that of couple households. Second, for each year, the measured RAI is divided into percentiles. Using this measured RAI for each percentile, I run the following linear regression to derive the predicted value of the future RAI for each percentile and year. , 0 1, , RAIst s st t ββ ε =++ In this equation, s represents the percentile and t represents the year. For each percentile of the measured RAI, I use eight samples from 2014 to 2021 to estimate 1 β . Table 5 reports the coefficient values, 1 β , corresponding to married-couple households in each tenth percentile of the RAI. As shown in the table, 1 β increases as the percentile of the RAI increases. This implies that the growth rate of the RAI increases with the percentile of the RAI. Using the estimated coefficients from the linear regression, I predict the future RAI for each percentile for 47 periods from 2024 to 2070. It should be noted that this analysis estimates the future distribution of the RAI under the strong assumption that each percentile of the RAI will grow linearly at a rate identical to the linear growth rate observed between 2014 and 2021. Thus, it TABLE 5—COEFFICIENT VALUES PER EACH 10TH PERCENTILE Percentile 10th 20th 30th 40th 50th 60th 70th 80th 90th 100th 1 β 4.774 5.049 6.167 9.0814 12.756 15.613 18.629 22.707 29.107 87.433 SE 0.626 0.444 0.324 0.603 0.682 1.137 1.725 3.804 5.679 42.993 Number of Obs. 8 8 8 8 8 8 8 8 8 8 Source: Calculated using data from the 7th to 14th waves of the NaSTaB survey by the Korea Institute of Public Finance. 112 KDI Journal of Economic Policy FEBRUARY 2025 (Unit: 1million won) <10th> <20th> <30th> <40th> <50th> <60th> <70th> <80th> <90th> FIGURE 4. MEASURED AND ESTIMATED RAI OF COUPLE FAMILIES PER 10TH PERCENTILE (2014-2021) Source: Calculated using the data from the 7th to 14th waves of the NaSTaB survey by the Korea Institute of Public Finance. implicitly assumes that the changes in the inflation rate, economic growth rate, and income distribution structure observed from 2014 to 2021 will continue into the future. Figure 4 shows both the estimated and predicted values of the RAI between 2014 and 2021 for each tenth percentile of the RAI. Because the estimated RAI increases linearly over the years, the predicted RAI, which was derived from the linear regression, matches the level and trend of the estimated RAI fairly well.6 Figure 5 illustrates the model fit by showing the official, measured, and estimated values for the 70th percentile of the RAI between 2014 and 2024. The official values 6 The estimated RAI at the 80th and 90th percentiles fluctuates somewhat over the years; therefore, the difference between the predicted and estimated RAI is greater than that at the lower percentiles. However, the predicted value of the RAI at the 80th and 90th percentiles is rarely used to predict the proportion of basic pension recipients under the selective eligibility criteria. VOL. 47 NO. 1 Financial Projection of the Basic Pension under Selective Eligibility Criteria 113 for the 70th percentile of the RAI, reported annually by the Ministry of Health and Welfare of South Korea, represent the figures used to determine the actual eligibility criteria for the basic pension. The measured values for the 70th percentile of the RAI refer to the values obtained from NaSTaB, while the estimated values are those predicted by means of an OLS regression. The left side of Figure 5 shows the yearly official, measured, and estimated values for the 70th percentile of the RAI between 2014 and 2024, while the right side of Figure 5 displays the percent difference between the estimated and official values. The model fit for the values from 2022 to 2024 demonstrate the performance of out-of-sample forecasts, as these values are not used to estimate the parameter values of the OLS. As shown on the left side of Figure 5, the yearly estimated and official values for the 70th percentile of the RAI exhibit a similar trend. The similarity in the level can be more accurately assessed by comparing the difference between the estimated and official values, as shown on the right side of Figure 5. The average difference between the estimated and official values from 2014 to 2024 is approximately 5.7%. When examining the yearly differences, the model’s prediction performance improves as the data points become more recent.7 Since 2021, the percentage difference has fallen below 5%. Notably, the differences are 1.8%, 3.5%, and 3.1% in 2022, 2023, and 2024, respectively, indicating relatively high forecast accuracy in out-of-sample periods. Figure 6 shows the predicted value of the RAI for married-couple households. Specifically, the figure on the left shows the 70th percentile of the predicted RAI between 2015 and 2060 and that on the right shows the distribution of the predicted RAI between the tenth and 90th percentiles in nominal value for the years 2023, 2033, (Unit: 10,000won, %) Eligibility Threshold: 70th Percentile of RAI Percent Difference between Official and Estimated Values FIGURE 5. OFFICIAL, MEASURED AND ESTIMATED ELIGIBILITY THRESHOLD AND CORRESPONDING PERCENT DIFFERENCES Source: Calculated using data from the 7th to 14th waves of the NaSTaB survey by the Korea Institute of Public Finance. 7 The source of the difference between the estimated and the official values can be decomposed into the difference between the measured and official values and the model misspecification from the fact that linear regression is used to derive the estimated values. 114 KDI Journal of Economic Policy FEBRUARY 2025 (Unit: 10,000won) 70th Percentile of Estimated RAI Distribution of Estimated RAI FIGURE 6. ESTIMATES OF ELIGIBILITY THRESHOLD AND DISTRIBUTION OF RAI Source: Calculated using data from the 7th to 14th waves of the NaSTaB survey by the Korea Institute of Public Finance. and 2043. The figure on the left shows that the eligibility threshold, the 70th percentile of the RAI, is predicted to increase to 13.4 million won for married-couple households in nominal terms in 2070, which is more than three times the 70th percentile of the RAI in 2024 at 3.408 million won. The figure on the right indicates that the distribution of the predicted RAI shifts to the right over time, which reflects the improving economic situation of future generations of the elderly. Thus far, I have explained the specific procedures used to estimate the RAI for each percentile using National Survey of Tax and Benefit data from 2014 to 2021 and the predict the future RAI sourced from linear regression analyses. In order to derive the proportion of elderly under selective scenario 1, I calculate the proportion of the elderly below the 70th percentile of the predicted RAI in 2024 value in nominal terms, which is 3.408 million won. Figure 7 shows the estimated proportion of recipients. Specifically, the figure on the left depicts the proportion of married-couple households below the eligibility criteria threshold in 2024 (3.408 million won), while the figure on the right shows the proportion of married-couple households below the median household income. As illustrated in the figure on the left, the proportion of married-couple households below the eligibility threshold of 3.408 million won in 2024—adjusted for 2% inflation annually—declines from 70% to 50% between 2024 and 2050 as the living standards of the new elderly generation improve. After 2050, this proportion stabilizes at 50%. This stabilization occurs because the estimated nominal value of the threshold line in 2023 follows a convex pattern due to the annual 2% increase, while the future RAI per percentile is projected under a linear trend. Consequently, my projection of recipients under selective scenario 1 may overestimate the number of recipients if the future inflation rate is lower than 2% or if the true trend of RAI is convex rather than linear.8 8The projection results from Lee, Kim, and Shin (2023) differ from those in this study because Lee, Kim, and Shin (2023) predict the 2023 threshold line by assuming a linear trend in the real-term values of the 2023 threshold VOL. 47 NO. 1 Financial Projection of the Basic Pension under Selective Eligibility Criteria 115 The right side of Figure 6 shows the proportion of recipients whose incomes are lower than the median income for two-person households. These results indicate that the proportion of households below the median household income gradually converges from 70% to 52% between 2024 and 2070. It is important to note that the proportion of recipients remains at 70% until 2026 because the original eligibility threshold, which is the 70th percentile of the RAI, is lower than the median income of two-person households until 2026. Therefore, selective scenario 2 only takes effect after 2026, when the 70th percentile of the RAI is projected to surpass the median income. Afterward, the rate declines gradually throughout the period between 2024 and 2070, as both the median income and the RAI are estimated under a linear trend. The proportion of recipients declines at a faster rate under selective scenario 1 compared to scenario 2 in the short term. Figure 6 shows that the proportion of recipients under selective scenario 1 reaches 0.5 much earlier than it does under selective scenario 2. This occurs because, in the short term, the growth rate of median income exceeds the inflation-adjusted value of the eligibility threshold set in 2024. Overall, the proportion of recipients under selective scenario 1 remains lower than that under selective scenario 2. Meanwhile, it is important to assess whether the inflation-adjusted value of the 2024 eligibility threshold ever falls below 50% of the median income for two-person households, which is the poverty line and which also serves as the lower boundary for selective scenario 1. The analysis shows that, throughout the period between 2024 and 2070, the inflation-adjusted threshold line from 2024 never falls below 50% of the median income. (Unit: %) Proportion under the 2024 Basic Pension Threshold Proportion under the Median Household Income FIGURE 7. PROPORTIONS OF RECIPIENTS UNDER SELECTIVE SCENARIOS Source: Calculated using data from the 7th to 14th waves of the NaSTaB survey by the Korea Institute of Public Finance. line in the past, rather than applying a 2% inflation rate to the eligibility threshold line. 116 KDI Journal of Economic Policy FEBRUARY 2025 IV. Financial Projection under Various Pension Reform Scenarios A. Explanation of Pension Reform Scenarios In this study, I conduct financial projections under different scenarios that vary in two aspects: eligibility criteria and benefit levels. The baseline scenario follows the current eligibility criteria, which target the bottom 70% of the RAI, and the benefit amount, which is set at 334,810 won in 2024 and evolves according to an inflation rate of 2% thereafter. For the alternative scenarios, I assume that the benefit amount increases to 400,000 won starting in 2025. I consider two alternative scenarios. The first is the expansion of old-age income security (hereafter the expansion scenario). These expansion scenarios are characterized by an increase in both the coverage and the level of benefits compared to the basic pension scheme. Specifically, I consider two expansion scenarios: in the first such scenario, the benefit amount is increased to 400,000 won in 2025 and adjusted according to the inflation rate (expansion scenario 1); in the second scenario, in addition to the benefit increase, the eligibility criteria are expanded from the bottom 70% to the bottom 100% of the population (expansion scenario 2). The second type of alternative scenario includes selective eligibility criteria (hereafter referred to as the selective scenario), designed to reduce the number of basic pension beneficiaries as the economic situation of the elderly improves. I consider two selective scenarios as well. The first sets the eligibility criteria threshold at the 70th percentile of the RAI among married-couple households in 2024, which is 3.408 million won, and then inflates this threshold in subsequent years (hereafter, selective scenario 1). The second selective scenario sets the RAI threshold at the median income for two-person families (hereafter, selective scenario 2). It should be noted that selective scenario 2 only takes effect when the original eligibility threshold, which is the 70th percentile of the RAI, surpasses the median income of two-person households. TABLE 6—PROJECTION SCENARIOS Scenario Full Benefit Amount Eligibility Criteria Threshold Baseline Scenario 334,810 won in 2024 inflation adjusted value (~2024) 70th percentile of RAI Expansion Scenario 1 334,810 won in 2024 400,000 won (2025) inflation adjusted value (~2025) 70th percentile of RAI Expansion Scenario 2 334,810 won in 2024 400,000 won (2025) inflation adjusted value (~2025) 100th percentile of RAI Selective Scenario 1 334,810 won in 2024 inflation adjusted value(~2024) Inflation Adjusted Value of 2024 Eligibility Criteria Threshold (3.408 million won) Selective Scenario 2 334,810 won in 2024 inflation adjusted value (~2024) Median Income of Two-Person Households VOL. 47 NO. 1 Financial Projection of the Basic Pension under Selective Eligibility Criteria 117 B. Financial Projection under the Baseline Scenario Before projecting the financial expenditure under the baseline case, first I present the population of elderly individuals aged 65 and over, basic pension recipients, and concurrent recipients, as doing so is essential for forecasting financial expenditure in Figure 8. Due to rapid aging, the population of elderly individuals aged 65 and over rises significantly, increasing by 1.9 times from ten million in 2023 to 19 million in 2050. However, due to the low fertility rate, the population size decreases slightly after 2050, reaching 17.5 million, which is still 1.75 times higher than the 2023 level. Similarly, the number of basic pension recipients, which is 70% of the elderly population, reaches its peak in 2050 and then begins to decline. Conversely, for concurrent beneficiaries, as the ratio of national pension recipients to the total population increases over time, there is an upward trend until 2070, although the growth rate slows somewhat after 2050. Figure 9 illustrates the financial projection results under the current system. Figure 9 shows that the financial expenditure for the basic pension amounts to 25 trillion won in 2024, 76 trillion won in 2050, and 102 trillion won in 2070. Relative to the nominal GDP, these expenditure levels reach 1% in 2024, 1.5% in 2050, and 1.3% in 2070. It is evident that while the number of recipients declines after 2050, the financial expenditure continues to rise even beyond 2050. This is due to the expenditure being presented in nominal terms. When converting the financial expenditure to real terms, as shown on the right side of Figure 10, there is a slight decline after 2050. Over time, the proportion of financial expenditure contributed by concurrent beneficiaries increases as their numbers grow. (Unit: 100,000) FIGURE 8. YEARLY POPULATION PROJECTIONS Source: Calculated based on population projections by the Department of Statistics and the basic pension formula. 118 KDI Journal of Economic Policy FEBRUARY 2025 (Unit: Trillion won, %) FIGURE 9. FINANCIAL PROJECTIONS UNDER THE BASELINE SCENARIO Note: Calculated based on the basic pension formula. (Unit: %) 2023~2032: Nominal Terms 2023~2070: Real Terms FIGURE 10. COMPARISON WITH FINANCIAL EXPENDITURE FROM NABO ( 2023a ) Note: The figure on the left is based on current prices and that on the right is based on 2023 constant prices. Source: Basic pension financial projection results calculated by the author and NABO (2023a). To assess the adequacy of the financial projections in this study, I compare them with the results from NABO (2023a). As shown on the left side of Figure 10, the result from NABO (2023a) and this study exhibit similar levels and trends between 2023 and 2032.9 For instance, NABO (2023a) predicts that the financial expenditure 9For a valid comparison with earlier studies that used projections from 2023, I make the financial projection assuming 2023 as the most recent year In the main results of this study, I use actual data from 2024 for the full VOL. 47 NO. 1 Financial Projection of the Basic Pension under Selective Eligibility Criteria 125 APPENDIX A. Calculation of RAI11 This section explains the procedure for calculating the RAI. To estimate the longterm proportion of basic pension recipients under the selective eligibility criteria, it is necessary to project the future distribution of the RAI. The future distribution is forecasted based on the linear trend of past RAI levels, under the assumption that the growth rate of this index will remain consistent. The RAI is calculated separately for single and married-couple households. When elderly individuals live with their parents, each household is considered to be a separate unit. In this study, household types are divided into three categories: those with a household head, those with the household head’s spouse, and those of the parents of either the household head or the household head’s spouse, each defined as a distinct household unit. It should be noted that while recipients of occupational pensions who meet exclusion criteria or special exceptions may still be eligible to receive the basic pension, this study did not consider these special cases and excluded individuals receiving occupational pensions. RAI refers to the combined value of the income assessment amount (IAA) and the converted value of assets (CVA). It is calculated based on the individual’s income and assets for single households and based on the combined income and assets of both the individual and their spouse for couple households. The formula for the RAI is given below. + RAI Income Assessment Amount Income Converted Asset= , where = {0.7 ( )} + + Income Assessment Amount Labour Income Earned Income Deduction Business Income Public Transfer Income Imputed Rent Income Property Income ×− +   +  and 11The RAI is calculated by Choo SeokHyeon, a research staff member at KDI. 126 KDI Journal of Economic Policy FEBRUARY 2025 ( ) ( 20 ) 12 Income Converted Asset General Property Basic Property Amount Property Income Conversion Rate Financial Asset million won Liability months Value of Luxury Cars and Membersh = − +   ×   −−       +ips 1. Method for Calculating the Income Assessment Amount To calculate the income assessment amount, one needs to compute labor income, business income, public transfer income, imputed rent, and property income. First, labor income refers to income from employment for three consecutive months with a fixed monthly salary (Ministry of Health and Welfare, 2023). In this context, income from public jobs or self-sufficiency work is excluded from labor income. However, in practice, NaSTaB lacks information with which one can identify whether the individual worked continuously for more than three months in the previous year. Therefore, in this study, the average monthly labor income is calculated by dividing labor income from the previous year by the number of months worked in the previous year. Business income is the sum of rental income and other business-related income. Annual rental income is converted into a monthly value by dividing the yearly rental income by 12. To calculate other business income on a monthly basis, the annual net business income is divided by the number of months worked. Property income consists of the sum of interest income and private pension income combined. Interest income is calculated using annual interest, dividends, and capital gains, with a deduction applied for amounts up to 40,000 KRW per month. Private pension income is derived from the total private pension insurance income. Public transfer income includes national pension income, industrial accident insurance benefits (such as sickness benefits, disability benefits, survivor benefits, occupational disease compensation pensions, pneumoconiosis compensation pensions, and pneumoconiosis survivor pensions), as well as benefits paid to veterans and other individuals eligible for national merit and honor. Therefore, public transfer income is calculated as the sum of national pension income (including oldage pensions, disability pensions, survivor pensions, and divided pensions), industrial accident insurance benefits (including occupational injury sickness benefits, occupational disability benefits, and occupational accident survivor pensions), and veteran benefits from NaSTaB. Imputed rent income refers to the amount calculated as the rental income equivalent for residing in a high-value property owned by a first-degree lineal descendant, such as a child. However, this is excluded from the current analysis. 2. Method for Calculating Imputed Income from Assets The value of each individual’s automobile assets is determined by applying the VOL. 47 NO. 1 Financial Projection of the Basic Pension under Selective Eligibility Criteria 127 type of car, engine displacement information, and the year of purchase by the residual value rates according to the vehicle age as provided by the National Pension Service. Vehicles with an engine displacement of less than 3,000 cc or valued below 40 million won are classified as general assets. Additionally, cars used by individuals with disabilities for living purposes should not be counted as automobile assets; however, this level of detail is not considered in this study. Differentiated basic asset values and imputed income conversion rates according to the metropolitan areas, for medium-sized cities, and for rural areas are derived from the official annual values provided in the Basic Pension Program Guidelines. General assets include land, buildings, housing, aircraft and ships, rental deposits, forest resources and fishing rights, as well as automobile assets (vehicles with an engine displacement of less than 3,000 cc or valued below 40 million won) (Ministry of Health and Welfare, 2023). General assets are derived from data on the market value of real estate, the market value of ships, rental deposit amounts, and automobile assets (vehicles with an engine displacement of less than 3,000 cc or valued below 40 million won). To distinguish the general assets of basic pension recipients accurately, it is necessary to identify the asset holder within the family. However, because it is difficult to distinguish these asset holders using NaSTaB data, this study utilizes family-level information on general assets. Financial assets include cash, checks, bonds, promissory notes, stocks, government and public bonds, deposits, savings accounts, insurance, investment securities, and pension insurance (Ministry of Health and Welfare, 2023). In this study, the value of financial assets is calculated by summing all of the financial asset variables provided by NaSTaB. However, national and private pensions received as a lump sum are not included in the calculation of financial assets for this study. Liabilities include loans obtained from financial institutions such as first-tier banks, second-tier banks, and lending companies (Ministry of Health and Welfare, 2023). This study calculates liabilities by summing the total amount of family liabilities from NaSTaB. B. Procedures for Projecting Fiscal Expenditure by Recipient Types Type 1: Non-pension recipients who are subject to both the income-reversal prevention reduction and the spousal reduction for married couples. The number of Type 1 recipients is calculated by multiplying the number of nonpension recipients (NonP)—derived by subtracting the number of concurrent beneficiaries from the total number of basic pension recipients—by the proportion of non-pension recipients who are subject to both the income-reversal prevention reduction and the spousal reduction for married couples. This proportion is based on the actual figures from 2021, as recorded in Table 4. 1 1 (2021, ) (, ) (, ) (2021, ) Type Type NonP NonP Ns N ts N ts Ns = × 128 KDI Journal of Economic Policy FEBRUARY 2025 The total fiscal expenditure for Type 1 can be derived by multiplying the number of Type 1 recipients by the full basic pension amount. Specifically, since Type 1 is subject to both the income-reversal prevention and spousal reductions, I multiply the actual ratio of the full basic pension benefit to the average basic pension amount paid to Type 1 recipients, using the 2021 values provided by Shin and Kim (2021). Although Shin and Kim (2021) provide age-specific ratios, given that there is little variation by age, I apply the ratio at age 65 uniformly across all ages in the calculation, which I denote as g(2021, 65) in the formula. 1 11 1 1 ( , ) ( ) (2021, ,65) 12 (2021, ,65) where, (2021, ,65) , () Type Type Type Type type Exp t s FBP t g s N BP s gs FBP t = × ×× = Since the full basic pension amount (FBP) is calculated on a monthly basis, it is multiplied by 12 to convert it into the annual expenditure. Type 2: Non-pension recipients who are subject to the spousal reduction for married couples. To calculate the number of recipients who are Type 2 recipients, we multiply the proportion of non-pension recipients who are subject to the spousal reduction for married couples but not subject to the income-reversal prevention reduction by the total number of non-pension recipients. 2 2 (2021, ) (, ) (, ) (2021, ) Type Type NonP NonP Ns N ts N ts Ns = × Since the spousal reduction rate is 20%, Type 2 recipients receive 80% of the pension amount. 22 ( , ) ( ) ( , ) 0.8 12 Type Type Exp t s FBP t N t s= × ×× Type 3: Non-pension recipients who are subject to income reverse prevention reduction. To calculate the number of recipients who are Type 3 recipients, we multiply the VOL. 47 NO. 1 Financial Projection of the Basic Pension under Selective Eligibility Criteria 129 proportion of non-pension recipients who are subject to the income-reversal prevention reduction but not subject to the spousal reduction for married couples by the total number of non-pension recipients. 3 3 (2021, ) (, ) (, ) (2021, ) Type Type NonP NonP Ns N ts N ts Ns = × The basic pension expenditure for Type 3 is then calculated by multiplying the number of Type 3 recipients by the product of the full basic pension amount and the actual ratio of the full basic pension benefit to the average basic pension amount paid to Type 3 recipients, based on the 2021 data provided by Shin and Kim (2021). 3 33 ( , ) ( ) ( , ) (2021, ,65) 12 Type Type Type Exp t s FBP t N t s g s =×× × Type 4: Non-pension recipients who are not subject to the spousal reduction for married couples or the income-reversal prevention reduction. To calculate the number of recipients who belong to Type 4, multiply the proportion of non-pension recipients who are neither subject to the income-reversal prevention reduction nor the spousal reduction for married couples by the total number of non-pension recipients. This proportion is based on actual 2021 data provided by Shin and Kim (2021). 4 4 (2021, ) (, ) (, ) (2021, ) Type Type NonP NonP Ns N ts N ts Ns = × The basic pension expenditure for Type 4 is calculated by multiplying the number of Type 4 recipients by the full benefit amount. This figure is then multiplied by 12 to convert it into an annual value. 44 (, ) () (, ) 12 Type Type Exp t s FBP t N t s=×× Type 5: Concurrent recipients who are subject to the income-reversal prevention reduction and the spousal reduction for married couples. Number of Type 5 recipients can be calculated by the product of the concurrent 130 KDI Journal of Economic Policy FEBRUARY 2025 recipients and the ratio of Type 5 recipients among the concurrent recipients. This proportion is based on the actual figures from 2021, as recorded in Table 4. 5 5 (2021, ) (, ) (, ) (2021, ) Type Type CR CR Ns N ts N ts Ns = × Then the basic pension expenditure for Type 5 is calculated by multiplying the number of Type 5 recipients by the product of the full basic pension amount and the actual ratio of the full basic pension benefit to the average basic pension amount paid to Type 5 recipients, based on the 2021 data provided by Shin and Kim (2021). 5 55 ( , ) ( ) (2021, ,65) 12 Type Type Type Exp t s FBP t g s N= × ×× Type 6: Concurrent recipients who are subject to the spousal reduction for married couples. Number of Type 6 recipients can be calculated by the product of the concurrent recipients and the ratio of Type 6 recipients among the concurrent recipients. This ratio is based on the actual figures from 2021, as recorded in Table 4. 6 6 (2021, ) (, ) (, ) (2021, ) Type Type CR CR Ns N ts N ts Ns = × Then the basic pension expenditure for Type 6 is calculated by multiplying the number of Type 6 recipients by the product of the basic pension amount and the 80% discount that is applied to the spousal reduction. Here, the basic pension amount is calculated based on the benefit formula for concurrent recipients as mentioned in section 3. 66 ( , ) ( ) ( , ) 0.8 12 Type Type Exp t s BP t N t s= × ×× Type 7: Concurrent recipients who are subject to the income-reversal prevention reduction. Number of Type 7 recipients can be calculated by the product of the concurrent recipients and the ratio of Type 7 recipients among the concurrent recipients. This VOL. 47 NO. 1 Financial Projection of the Basic Pension under Selective Eligibility Criteria 131 ratio is based on the actual figures from 2021, as recorded in Table 4. 7 7 (2021, ) (, ) (, ) (2021, ) Type Type CR CR Ns N ts N ts Ns = × The basic pension expenditure for Type 7 is calculated by multiplying the number of Type 7 recipients by the product of the basic pension amount and the actual ratio of the full basic pension benefit to the average basic pension amount paid to Type 7 recipients, based on the 2021 data provided by Shin and Kim (2021). 77 (, ) () (, ) 12 Type Type Exp t s FBP t N t s =×× Type 8: Concurrent recipients who are subject to neither the income-reversal prevention reduction nor the spousal reduction Number of Type 8 recipients can be calculated by the product of the concurrent recipients and the ratio of Type 8 recipients among the concurrent recipients which is based on the actual figures from 2021, as recorded in Table 4. 8 8 (2021, ) (, ) (, ) (2021, ) Type Type CR CR Ns N ts N ts Ns = × The basic pension expenditure for Type 8 is then calculated by multiplying the number of Type 8 recipients by the product of the basic pension amount, which is calculated based on the benefit formula for concurrent recipients, as mentioned in section 3. 88 (, ) () (, ) 12 Type Type Exp t s BP t N t s=×× C. Method for Estimating Median Income To convert the RAI derived in this study into a ratio relative to the median income, it is necessary to estimate future median household income trends. Assuming the trend in median household income remains consistent between 2015 and 2024, this study estimates median household income from 2025 onwards. First, the officially reported median household income data from 2015 to 2024 are used to predict future 132 KDI Journal of Economic Policy FEBRUARY 2025 (Unit: 10,000won) Median Income of Two-Person Households Projection of Median Income of Two-Person Households (~2070) FIGURE A1. OFFICIAL AND ESTIMATED VALUE OF MEDIAN INCOME OF TWO-PERSON HOUSEHOLDS Note: Official values for the median income are reported by the Ministry of Health and Welfare. values from 2025 onwards. Specifically, ordinary least squares (OLS) regression is applied to predict median household income from 2025 onwards using the official values for two-person households. The estimation results indicate that the median income for a two-person household is projected to increase from 3.61 million KRW in 2025 to 8.27 million KRW in 2070. D. Comparing the forecast performance between OLS and ARIMA In this section, I compare the forecast performance outcomes of linear regression and the autoregressive integrated moving average (ARIMA) model when used to estimate the RAI distribution. The ARIMA models are denoted as ARIMA (p,d,q), where p represents the number of lags, d stands for the degree of differencing, and q is the order of the moving-average model (Wikipedia Contributors, 2024). I selected ARIMA (1,1,0) because it has the lowest AIC and BIC compared to other ARIMA models, with p and q values that are less than or equal to 2, and with d equal to 1. I forecast the 70th percentile of the measured RAI using both ARIMA (1,1,0) and linear regression, and I report the yearly estimates for each model on the left side of Appendix Figure A2 and the percent difference between the official and estimated values for each model on the right side. The forecast performance outcomes of the OLS and ARIMA models vary in the earlier period. However, after 2021, the forecast performances of both models converge, with the percent difference between the official and estimated 70th percentile of the RAI falling below 5% for both models, as shown on the right side of the figure. When comparing future projections using the OLS and ARIMA models up to 2070, both exhibit a similar linear trend, as illustrated in Appendix Figure A3. This suggests that the long-term forecast performance capabilities of OLS and ARIMA do not differ significantly. Therefore, for the purposes of this paper, I use OLS regression to predict the RAI distribution. VOL. 47 NO. 1 Financial Projection of the Basic Pension under Selective Eligibility Criteria 133 Yearly Estimates: 2014-2024 Percent Difference FIGURE A2. YEARLY ESTIMATES AND PERCENT DIFFERENCE UNDER OLS AND ARIMA Source: Calculated using data from the 7th to 14th waves of the NaSTaB survey by the Korea Institute of Public Finance. FIGURE A3. YEARLY ESTIMATES UNDER OLS AND ARIMA: 2014-2070 Source: Predicted using data from the 7th to 14th waves of the NaSTaB survey by the Korea Institute of Public Finance. 134 KDI Journal of Economic Policy FEBRUARY 2025 REFERENCES Basic Pension Adequacy Evaluation Committee. 2023. “2023 Basic Pension Adequacy Evaluation Committee Report,” (in Korean). Kwon, Tae-Goo et al. 2022. “Ten Years of Social Insurance Contribution Support Programs: Evaluation and Improvement Tasks with a Focus on Local Government Initiatives,” Research Report 2022-10, Employment and Vocational Competency Development Center, Korea University of Technology and Education. Lee, KangKoo, DoHun Kim and Seung-Ryong, Shin. 2023. “Reform Measures for Enhancing the Sustainabiltiy of the Public Pension System,” Policy Study 2023-08, Korea Development Institute (in Korean). Ministry of Health and Welfare. 2014-2024. “Basic Pension Program Guidelines”. NABO. 2018. “2018-2027 Estimation of the Expenditure for the Basic Pension”. NABO. 2023a. Public Pension Reform and Long-term Projections I. NABO. 2023b. Public Pension Reform and Long-term Projections II. National Pension Service. 2020. “2020 National Pension Statistical Yearbook”. OECD. 2021. Pensions at a Glance 2021: OECD and G20 Indicators, OECD Publishing. OECD. 2022. OECD Economic Surveys: Korea 2022, OECD Publishing. Ryu, Jae-Rin et al. 2022. “A Study on the Restructuring Plans for Public Pensions,” Research Report 2022-01, Korea Institute for Health and Social Affairs (in Korean). Shin, Kyung-Hey and Hyung-Soo, Kim. 2021. “Development of a Fiscal Simulation Model for the Basic Pension,” Policy Report 2021-05, National Pension Research Institute, National Pension Service (in Korean). Wikipedia Contributors. 2024. Autoregressive Integrated Moving Average, Wikipedia, October, 9, 2024(https://en.wikipedia.org/wiki/Autoregressive_integrated_moving_average, last accessed October 10, 2024). Yun, SeokMyoung. 2023. “Politically and Fiscally Sustainable Pension Reform Directions,” Policy Forum on Reform Directions for Sustainable National Pension Finance, The Korean Association of Public Finance (in Korean). LITERATURE IN KOREAN 국민연금공단. 2020. 「2020년 국민연금 통계연보」. 국회예산정책처. 2018. 「2018~2027년 기초연금 재정소요 추계」. 국회예산정책처. 2023a. 『공적연금개혁과 재정전망 Ⅰ』. 국회예산정책처. 2023b. 『공적연금개혁과 재정전망 Ⅱ』. 권태구 외. 2022. 『사회보험료 지원사업 10년, 평가와 개선과제: 지자체 사업을 중심으로』, 연구보고서 2022-10, 한국기술교육대학교 고용직업능력개발센터. 기초연금 적정성 평가위원회. 2023. 『2023년 기초연금 적정성 평가위원회 보고서』. 류재린 외. 2022. 『공적연금 재구조화 방안 연구』, 연구보고서 2022-01, 한국보건사회연구원. 보건복지부. 2014-2024. 『기초연금 사업안내』. 보건복지부. 2023. 「제5차 국민연금 종합운영계획(안)」. 신경혜·김형수. 2021. 『기초연금 재정시뮬레이션모형 구축』, 정책보고서 2021-05, 국민연금공단 국민연금 연구원. 윤석명. 2023. 『정치적, 재정적으로 지속 가능한 연금개혁 방향』, 지속가능한 국민연금 재정을 위한 개혁 방향 정책토론회, 한국재정학회. 이강구·김도헌·신승룡. 2023. 『공적연금제도 지속가능성 제고를 위한 개혁방안』, KDI 정책 연구시리즈, 2022년 제7차 국민연금 전문가 포럼.