## 1. Demographics ## Source details **Canonical URL:** [1. Demographics](https://www.imf.org/-/media/websites/imf/imported-full-text-pdf/external/pubs/ft/wp/2007/_wp07109.pdf) ## Other formats - [Markdown version](/-/media/websites/imf/imported-full-text-pdf/external/pubs/ft/wp/2007/_wp07109.pdf.md) - [Structured JSON version](/-/media/websites/imf/imported-full-text-pdf/external/pubs/ft/wp/2007/_wp07109.pdf.json) --- ### Demographic pressures and projections - China has created over 350 million new jobs and is estimated to have lifted over 400 million people out of poverty. - The number of workers for every elderly person is projected to fall from 3 at present to 2 by 2015 and to only 1 by the mid-2030s (World Bank (2006a)). - The working-age population (15–60 years) as a proportion of total population is projected to peak as early as 2010; thereafter the share falls steadily while the share of the elderly (over 60 years) will double before 2030. - The median age is expected to rise from 32½ years in 2005 to 48 years in 2050. - Example cross-country comparison: - Korea reached median age 32½ years in 2000 with per capita income U.S.$16,200 (PPP basis), nearly two-and-a-half times China’s 2005 per capita income (U.S.$6,700). - Projections and figures are based on United Nations (2007) population projections and World Population Prospects. - Sensitivities noted: - Definitions of “working-age” and “elderly” affect the picture; workers often start after age 15 and actual retirement ages may be below 60 for many people (World Bank (2006a), Trinh (2006)). ### Implication for pensions - The demographic outlook indicates significant aging in coming decades, increasing urgency for pension system reform. --- ### II. Pension reform context and the 1997 design ### Historical context - Post-1949 pension system: “iron rice bowl” with guaranteed employment in SOEs or state institutions; defined benefits provided by enterprises financed by enterprise contributions topped up by fiscal allocations; no employee contributions. - SOE reforms from the late 1980s hardened SOE budget constraints, reduced ability of SOEs to provide promised benefits, and progressively reduced social obligations of SOEs. - 1991: authorities encouraged a multi-pillar pension system with a social pension and individual accounts; employee contributions introduced; pilot projects sought greater balance and pooling at higher government levels. - 1995: key operational details (benefits and eligibility) were specified and pilot projects conducted in Shanghai and Guangzhou. ### 1997 reform structure and parameters - Objective: replace enterprise-provided pensions with a countrywide social insurance system covering health, unemployment, disability, maternity, and pensions. - Basic structure: mandatory two-pillar pension comprising a defined benefit “social pension” plus defined contribution “individual accounts.” - Social pension: - Financed by enterprise (employer) contributions of 17 percent of wages. - Provides benefits equal to 20 percent of average local wages, provided the enrollee has contributed for at least 15 years. - Intended to be pay-as-you-go (PAYG). - Individual accounts: - Financed by enterprise contributions of 3 percent and employee contributions of 8 percent of wages. - Provide a monthly annuity equal to 1/120th of the accumulated contributions plus interest. - Intended to be fully funded. - Designed on the assumption that contributing for 35 years would yield a monthly annuity equivalent to 38.5 percent of wages. - Combined replacement example: - Replacement rate for an individual with wage equal to average local wage: 58.5 percent (20 percent from the social pension + 38.5 percent from the individual account). - Retirement age: - 60 years for men and 50–60 years for women, with the higher age applying to women who are professional and salaried. --- ### III. Experience and outcomes under the 1997 reform ### Coverage and participation (key statistics) - A decade after initiation, the revised system covers less than half of urban workers. - As of 2005: - Pension contributors in enterprises: 117 million people. - Total urban employment: 273 million (National Bureau of Statistics (2006a and 2006b)). - Implied coverage ratio: 43 percent of urban employment, or only 15 percent of total national employment. - Rural pension system covers only an estimated 12 percent of rural workers. - For the nation as a whole, less than a quarter of workers have any form of pension coverage under the old or revised system. ### Performance and implementation issues - Pilot municipalities: 3 municipalities have been relatively successful in developing pension systems for urban workers; pilot provinces’ experience has been mixed. - Reform process impeded by focus on resolving “legacy costs” from transition generations entitled to more generous benefits under the old system. --- ### IV. Remaining problems — legacy costs and fiscal exposure ### Fiscal dimensions and unfunded liabilities - Transition costs represent an unfunded liability. - IMF staff estimates: as of 2003, the unfunded liability amounted to 7 percent of GDP. - Actual costs could be higher if realized parameters differ (e.g., higher replacement rates or higher life expectancy after retirement). - Common definitions of legacy costs omit pensions to be paid to current enrollees and costs of covering those currently not covered. ### System responses and limits - Individual accounts partially used to fill gap between pension payouts and social pension contributions, making individual accounts partially funded or notional rather than fully funded. - National Social Security Fund (NSSF) established in 2000 to serve as a reserve and to cover cash flow deficits; NSSF entitled to receive 10 percent of proceeds from equity sales of SOEs in IPOs (practice suspended in 2002). - Lottery ticket sales provide an additional revenue source. ### Structural and administrative weaknesses - Decentralized and fragmented system (usually county or municipal level) prevents effective pooling. - Local authorities exercise substantial discretion over parameters and administration; surpluses in some areas not used to cover shortfalls in others, leading to central government financing of shortfalls. - Problematic features: - Low retirement age, implying a relatively long retirement period. - An implicit assumption in individual accounts that understates life expectancy after retirement. - Demographic profile of increasingly rapid aging and rising dependency ratios. ### Pensions, wage growth, and replacement rates (key statistics) - During 2000–2005: - Average (nominal) wages grew by 15 percent annually in state-owned units. - Average (nominal) wages grew by 12½ percent annually in the manufacturing sector. - Average (nominal) pensions grew by 8½ percent annually. - Effective replacement rate fell from 75 percent of SOE wages in 2000 to 55 percent in 2005. - Provincial/local variations in contribution rates and transition supplements have emerged due to decentralized administration. ### Demographic pressures (additional statistics) - Proportion of urban retirees relative to urban workers rose from under one-fifth in 1990 to over one-third in 2005. - Urban dependency ratio projections: - Around 50 percent before 2015. - 100 percent in the mid-2030s (Sin (2005)). - Retirement-period assumptions versus observed longevity: - System’s annuity rule assumes retirees live for 11 years after retirement (monthly annuity of 1/120th of accumulated savings). - Estimated actual life expectancy after retirement: 20 years for men and 27 years for women (Sin (2005)). - Life expectancy at birth between 1990 and 2000 increased to 70 years (men) and 73 years (women) from 67 years and 70 years, respectively (NBS (2006)). --- ### V. Provincial pilot projects — experience and limitations ### Liaoning pilot (initiated 2001) - Contribution structure: - Enterprise contribution (20 percent of wages) used for social pension benefits. - Employee contribution (8 percent) funds individual accounts. - Gap financed by central and local governments. - Achievements: - Funded individual accounts separate from social pension. - Improvements in pension administration (World Bank (2006b)). - Weaknesses: - Low coverage, high and non-uniform contribution rates across municipalities, pooling effectively only at municipal level, financial unsustainability. ### Extensions to Heilongjiang and Jilin (2004) - Modifications: - Basic pension based on individual (not local) salaries. - Payouts from individual accounts take into account life expectancy. - In Heilongjiang, pension funds intended to be pooled at provincial level. - Assessment limited by short experimental period and lack of data. --- ### VI. A new approach — principled separation of legacy and new system design ### Findings and rationale - Current approach not working; proposal: separate legacy problems from establishing a new national pension system covering the whole labor force. - Legacy problems should be set aside temporarily and solved later via negotiated agreements between local and central governments on parameters (retirement age, replacement rates) and burden sharing. - New national pension system should be set up quickly to address aging and extend coverage (including rural population). ### Recommended architecture and key design elements - Two-pillar system (consistent with State Council): - Pillar 1: defined-benefit social pension guaranteeing a minimum level of retirement income. - Pillar 2: mandatory defined-contribution individual account. - Funding approach: - Social pension: could be fully or largely funded; PAYG option is unrealistic given demographic outlook. At least partial funding advisable to smooth future tax burden. - Individual accounts: should be fully funded and independent of legacy costs. Do not adopt a “notional defined contribution” system for the new plan. - Coverage and inclusion considerations: - Inclusion of older workers raises fiscal questions: contributions from older entrants may not cover social pension costs or generate sufficient individual account balances; budget contribution may be needed for social pension costs of such cohorts. - Institutional arrangements: - Establish a single new national pension administration to receive all pension revenue, manage funds, and deliver pensions, set up independently of existing ministries and preferably under the State Council. - Governance: full and transparent accounting, annual audits, simplified annual reports to individual account holders. - Payroll tax to finance both social pension and individual accounts; State Administration of Taxation to manage collection for individual accounts on behalf of pension administration. - Parameter setting and reviews: - Retirement age, replacement ratio, contribution rates, benefits indexation should be set to reflect retirement needs, social objectives, and long-term financial viability. - Annual financial reports and periodic comprehensive reviews (perhaps every 5 years) to assess long-term viability and adjust parameters and implementation as needed. - Asset management and investment strategy: - Keep assets for social pension strictly separate from those for individual accounts. - Initially manage assets centrally by government due to underdeveloped domestic financial markets; individual accounts would resemble provident funds initially. - Consider offshore investment at least initially to achieve rates of return necessary for reasonable replacement ratios; move to domestic investment and more diversified portfolios as markets develop. - Over time, consider giving workers more control over investment decisions for their individual accounts and means-testing the social pension depending on developments and retirement needs. - Role of existing NSSF: - Current National Social Security Fund could be left to manage legacy costs associated with the old and revised pension systems. --- ### VII. Concluding remarks and policy urgency - Substantial need to restart pension reform in China now to address looming aging-related fiscal pressures and to aid in rebalancing China’s economy. - Major innovation proposed: temporarily sidestep legacy problems and proceed rapidly to implement a nationwide pension system covering all workers. - Rationale: broad consensus exists on the basic two-pillar structure; the core structure can remain stable while allowing adjustments in details over time to reflect economic development and social preferences. *Source: _wp07109 - 1. Demographics* ### 1. Demographics ....................................................................................................... ### 1. Demographics ### Introduction and demographic pressures - China has become the world’s fourth largest economy and a leading trading nation; it has created over 350 million new jobs and is estimated to have lifted over 400 million people out of poverty. - To sustain rapid growth, China needs to rebalance away from heavy reliance on investment and exports toward consumption; financial sector reform and improvements in social services, including pensions, are key. - The number of workers for every elderly person is projected to fall from 3 at present to 2 by 2015 and to only 1 by the mid-2030s (World Bank (2006a)). - China’s age dependency ratio has been rising since the mid-1980s as the working-age population (15–60 years) has increased slowly relative to the elderly population (over 60 years). - The working-age population as a proportion of the total population is projected to peak as early as 2010 (Figure 1); in subsequent decades the share of working-age population falls steadily while the share of the elderly will double before 2030. - The median age is expected to rise from 32½ years in 2005 to 48 years in 2050. - Many countries that have already experienced more marked population aging did so at higher levels of income. Example given: - Korea reached median age 32½ years in 2000 with per capita income U.S.$16,200 (PPP basis), nearly two-and-a-half times China’s 2005 per capita income (U.S.$6,700). ### Demographic projections and data sources - Projections and figures in the source are based on United Nations (2007) population projections and World Population Prospects. - The text notes sensitivity to definitions of “working-age” and “elderly” and highlights that a starker picture can emerge if one considers that workers typically start after age 15 and actual retirement ages may be below 60 for many people (see World Bank (2006a), Trinh (2006)). ### Implication for pensions - The demographic outlook clearly indicates significant aging in coming decades, increasing urgency for pension system reform. ### Pension reform context and history - Post-1949 pension system was part of the “iron rice bowl”: guaranteed employment in state-owned enterprises (SOEs) or state institutions with defined benefits provided by enterprises and financed by enterprise contributions topped up by fiscal allocations; there were no employee contributions. - SOE reforms beginning in the late 1980s hardened SOE budget constraints, reduced ability of SOEs to provide promised benefits, and progressively reduced the social obligations of SOEs. - In 1991 authorities encouraged a multi-pillar pension system with a social pension and individual accounts, and introduced employee contributions; pilot projects sought greater balance between employee and employer contributions and pooling at higher levels of government. - In 1995 key operational details (benefits and eligibility) were spelled out and pilot projects were conducted in Shanghai and Guangzhou. ### The 1997 reform design - 1997 revised pension policy aimed to cover China’s entire urban labor force by replacing enterprise-provided pensions with a countrywide social insurance system responsible for health, unemployment, disability, maternity, and pensions; pilot projects in 2000 expanded implementation in selected provinces/municipalities. - The revised system’s basic structure: a mandatory two-pillar pension comprising a defined benefit “social pension” plus defined contribution “individual accounts.” - Social pension: - Financed by enterprise (employer) contributions of 17 percent of wages. - Provides benefits equal to 20 percent of average local wages, provided the enrollee has contributed for at least 15 years. - Intended to be pay-as-you-go (PAYG). - Individual accounts: - Financed by enterprise contributions of 3 percent and employee contributions of 8 percent of wages. - Provide a monthly annuity equal to 1/120th of the accumulated contributions plus interest. - Intended to be fully funded. - Designed on the assumption that contributing for 35 years would yield a monthly annuity equivalent to 38.5 percent of wages. - Combined replacement example: - An individual with wage equal to average local wage could expect a replacement rate of 58.5 percent (20 percent from the social pension + 38.5 percent from the individual account). - The first pillar (social pension) pays benefits based on average local wages and thus represents redistribution from higher-than-average earners to lower-than-average earners. - Retirement age: - 60 years for men and 50–60 years for women, with the higher age applying to women who are professional and salaried. ### Experience and outcomes under the 1997 reform - The 1997 reform has fallen far short of objectives and meaningful coverage for the population as a whole. - Coverage status (as of data cited in the source): - A decade after initiation, the revised system covers less than half of urban workers. - As of 2005, pension contributors in enterprises (which in principle fall under the revised system) amounted to 117 million people, compared with total urban employment of 273 million (National Bureau of Statistics (2006a and 2006b)). - Implied coverage ratio is 43 percent of urban employment, or only 15 percent of total national employment. - This calculation omits non-enterprise contributors and thus under-estimates actual coverage; even if all non-enterprise contributors were assumed to fall under the new plan, coverage would still be only a fraction (text notes it would still be only ...). - The rural pension system covers only an estimated 12 percent of rural workers. - For the nation as a whole, less than a quarter of workers have any form of pension coverage whether under the old or revised system. - Pilot project outcomes: - The 3 municipalities (unnamed in this excerpt) have been relatively successful in developing pension systems for urban workers. - Experience in pilot provinces has been mixed. - The reform process has been bogged down by a preoccupation with resolving “legacy costs” associated with transition from the old system to the revised system. - “Legacy costs” refer to differences in pension entitlements for transition generations: workers who retired or who enrolled in the old pension system before 1997 are entitled to significantly more generous benefits than under the revised system (example figures begin to be cited but are truncated in the excerpt). *Source: _wp07109 - 1. Demographics* ### 58.5 percent for an average worker). ### _wp07109 - 58.5 percent for an average worker) ### III. Remaining problems in the pension system — legacy costs and fiscal exposure - The transition costs associated with the move from the old unfunded system to the revised system represent an unfunded liability. - IMF staff estimates suggest that, as of 2003, the unfunded liability amounted to 7 percent of GDP. - The actual cost of funding unfunded liabilities could be higher if different parameters are realized than assumed (for example, higher replacement rates or higher life expectancy after retirement). - The commonly used definition of legacy costs omits pensions to be paid to current enrollees and the cost of covering those currently not covered. ### System responses to legacy costs and their limits - The pension plan was modified so the gap between pension payouts and social pension contributions was filled with resources from individual accounts, making individual accounts partially funded or notional rather than fully funded. - A National Social Security Fund (NSSF) was established in 2000 to serve as a reserve and to cover the cash flow deficit in the social pension; the NSSF was entitled to receive 10 percent of the proceeds from equity sales of SOEs in initial public offerings (practice suspended in 2002). - Lottery ticket sales serve as an additional source of revenue. ### Structural and administrative weaknesses - The pension system is decentralized and fragmented (usually county or municipal level), preventing effective pooling. - Local authorities have substantial discretion over parameters and administration; surpluses in some areas have not been used to cover shortfalls in others, forcing central government financing of shortfalls. - Key problematic features: - Low retirement age, implying a relatively long retirement period. - An implicit assumption in individual accounts that understates life expectancy after retirement. - Demographic profile: increasingly rapid aging and rising dependency ratios. ### Pensions, wage growth, and replacement rates (key statistics) - During 2000–2005: - Average (nominal) wages grew by 15 percent annually in state-owned units. - Average (nominal) wages grew by 12½ percent annually in the manufacturing sector. - Average (nominal) pensions grew by 8½ percent annually. - Effective replacement rate fell from 75 percent of SOE wages in 2000 to 55 percent in 2005. - Provincial/local variations in contribution rates and transition supplements have emerged due to decentralized administration. ### Demographic pressures (key statistics and implications) - The proportion of urban retirees relative to urban workers rose from under one-fifth in 1990 to over one-third in 2005. - Urban dependency ratio projections: - Around 50 percent before 2015. - 100 percent in the mid-2030s (Sin (2005)). - Retirement-period and life expectancy tensions: - System’s annuity rule: monthly annuity of 1/120th of accumulated savings assumes retirees live for 11 years after retirement. - Estimated actual life expectancy after retirement: 20 years for men and 27 years for women (Sin (2005)). - Standard retirement ages: 60 years for men and 50–60 years for women (NBS (2006)); life expectancy at birth between 1990 and 2000 increased to 70 years (men) and 73 years (women) from 67 years and 70 years, respectively (NBS (2006)). ### Provincial pilot projects — experience and limitations - Liaoning pilot (initiated 2001): - Enterprise contribution (20 percent of wages) used for social pension benefits; employee contribution (8 percent) funds individual accounts. - Gap financed by central and local governments. - Achievements: funded individual accounts separate from social pension; improvements in pension administration (World Bank (2006b)). - Weaknesses: low coverage, high and non-uniform contribution rates across municipalities, pooling effectively only at municipal level, financial unsustainability. - Extensions to Heilongjiang and Jilin in 2004 with modifications: - Basic pension based on individual (not local) salaries. - Payouts from individual accounts take into account life expectancy. - In Heilongjiang, pension funds intended to be pooled at provincial level. - Assessment limited by short experimental period and lack of data. ### IV. A new approach — principled separation of legacy and new system design Findings and rationale: - The current approach is not working; the proposal is to separate legacy problems from establishing a new national pension system covering the whole labor force. - Legacy problems should be set aside temporarily and solved later via negotiated agreements between local and central governments on parameters (retirement age, replacement rates) and burden sharing. - The new national pension system should be set up quickly to address aging and extend coverage (including rural population). Recommended architecture and key design elements: - Two-pillar system (consistent with State Council): - Pillar 1: defined-benefit social pension guaranteeing a minimum level of retirement income. - Pillar 2: mandatory defined-contribution individual account. - Funding approach: - Social pension: could be fully or largely funded; PAYG option is unrealistic given demographic outlook. At least partial funding advisable to smooth future tax burden. - Individual accounts: should be fully funded and independent of legacy costs. Do not adopt a “notional defined contribution” system for the new plan. - Coverage and inclusion considerations: - Inclusion of older workers raises fiscal questions: contributions from older entrants may not cover social pension costs or generate sufficient individual account balances; budget contribution may be needed for social pension costs of such cohorts. - Institutional arrangements: - Establish a single new national pension administration to receive all pension revenue, manage funds, and deliver pensions, set up independently of existing ministries and preferably under the State Council. - Governance: full and transparent accounting, annual audits, simplified annual reports to individual account holders. - Payroll tax to finance both social pension and individual accounts; State Administration of Taxation to manage collection for individual accounts on behalf of pension administration. - Parameter setting and reviews: - Retirement age, replacement ratio, contribution rates, benefits indexation should be set to reflect retirement needs, social objectives, and long-term financial viability. - Annual financial reports and periodic comprehensive reviews (perhaps every 5 years) to assess long-term viability and adjust parameters and implementation as needed. - Asset management and investment strategy: - Keep assets for social pension strictly separate from those for individual accounts. - Initially manage assets centrally by government due to underdeveloped domestic financial markets; individual accounts would resemble provident funds initially. - Consider offshore investment at least initially to achieve rates of return necessary for reasonable replacement ratios; move to domestic investment and more diversified portfolios as markets develop. - Over time, consider giving workers more control over investment decisions for their individual accounts and means-testing the social pension depending on developments and retirement needs. - Role of existing NSSF: - The current National Social Security Fund could be left to manage legacy costs associated with the old and revised pension systems. ### V. Concluding remarks — policy urgency and strategic recommendation - There is a substantial need to restart pension reform in China now to address looming aging-related fiscal pressures and to aid in rebalancing China’s economy. - Major innovation proposed: temporarily sidestep legacy problems and proceed rapidly to implement a nationwide pension system covering all workers. - Rationale: broad consensus exists on the basic two-pillar structure; the core structure can remain stable while allowing adjustments in details over time to reflect economic development and social preferences. *Source: IMF working paper content provided in the supplied PDF excerpt.* --- _Source: https://www.imf.org/-/media/websites/imf/imported-full-text-pdf/external/pubs/ft/wp/2007/_wp07109.pdf_