Attachment 8 - Annex 3.doc

DRAFT

Central Bank of the Republic of Armenia Pension reform preparation group Pension System Reform Program Main principles, strategic directions, tactical steps and actuarial forecasts for 2008-2080 Yerevan

25 August 2008 “Pension system reform program” was drafted in accordance with the list of measures stipulated by Government Order No. 796 dated 26 May 2006 "Approving the conceptual framework of pension security system reform in the Republic of Armenia” (Annex 2, Item 3).The draft is based on strategic papers adopted by the Government of the Republic of Armenia in reference to the following Government Orders: “Conceptual approaches to pensions system reform” (No. 666-N dated 28 April 2005), “Conceptual framework of pensions system reform” (No. 796-N dated 26 May 2006) and “Program of the government of Republic of Armenia” (No. 695-A dated 21 June 2007).

This draft was prepared through joint efforts of specialists from the Central Bank of Armenia (CB), Ministry of Finance (MF), Ministry of Economy (ME), Ministry of Labor and Social Issues (MLSI), as well as State Social Security Service (SSSS) functioning under the MLSI supervision. They were formed into three thematic groups by the Pensions system Reform (PSR) working group of the GoA , in order to take into account the viewpoints and expert assessments of different agencies in the drafting of the main provisions of the program.

The draft has received valuable support from international organizations, specially the World Bank and USAID.

The draft is a result of numerous discussions and consultations with experts and will form the basis for elaborating the Pensions system Act of the RoA and relevant regulations.

CONTENTS

5CHAPTER I. THE NECESSITY FOR PENSION REFORM

51. Current situation (2000-2006)

51.1 GDP, wage and pension

71.2 Pension, living standard and poverty

92. Pension system

92.1 Current legal framework

112.2 Description of the system

142.3 Types of pension

172.4 Expenditures and revenues

223. Population and labor market

223.1 Perspectives of the demographic situation

263.2 Employment and unemployment trends

334. Financial sustainability of the system

334.1 Social and pension expenditures

354.2 Financial sustainability and pension adequacy advocacy

424.3 Incentives and disincentives

515. Main conclusions

53Chapter II. Pension system reform

536. International experience in pension reform

536.1 Main international concepts

596.2 Cross-Country Comparisons: Main Challenges

626.3 Pension reforms: International Experience

706.4. Lessons learned

726.5 Summary and conclusions

7. Armenian model of pension reform

7.1 Brief description of the model

7.2 Key principles

827.3 Defined parameters

927.4 Baseline scenario: 2008-2012

947.5 Reform scenario: 2009-2080

106CHAPTER III. ARMENIA’S MULTI-PILLAR PENSION SYSTEM

1068. Pillar “0”of the Pension System

1068. “O” Pillar of the Pension System

1068.1 Pillar “0”: Social Pension

1078.2 Types of Social Pension

1098.3 General Provisions

1119. Pillar “1” of Pension System

1119.1 Pillar “1”: EMPLOYMENT PENSIONS

1129.2 Types of EMPLOYMENT Pension

1139.3 General Provisions

11410. “2” pillar of pension system

11410.1. “2” pillar –mandatory funded pension

11510.2 Mandatory cumulative contributions

11610.3 Use of funded pension

11910.4 Types of funded pensions

11910.5 Taxation of accumulated amounts

12010.6 Inheritance of accumulated amounts

12110.7 Asset management – description of the process

12110.7.1. The asset management process starts with the collection of contributions and information. Employers send information on employees making mandatory cumulative contributions to State Tax Service of the Government of Armenia (hereinafter STS). The information is entered into the STS database.

12610.8 Investment portfolios and assets investment

12910.9. Ensuring safety of investments

12911. Pillar "3" of the Pension System

12911. 1 Pillar 3 – Voluntary Funded Pension

12911.2 Tax Provisions

13011.3 Management of Voluntary Funded Pension

13111.4. Pension Insurance

13111.5. Assets management process

13211.6. Regulation of investments

13512 Insurance against risks of disability and survivorship

13613. Projections of Replacement Ratios

141Chapter IV. Governance of multi-pillar pensions system

14114. Institutional framework of the system

14114.1 Main principles

14114.2 Institutions and functions

20214.3 Management Schemes

17615. Regulation and supervision of nongovernmental pension institutions

17615.1. Central Depository of Armenia – CDA

17615.2 Pension Assets Managers – AM

17915.3 Custodians

17915.4 Private pension funds

17916. Possible risks and their management

17916.1 Risks of the program

18116.2 Market risks

18216.3 Operational risks

18217. List of 2008 measures to ensure project implementation

186Annex 1

186Annex 2 (Charts)

213Annex 4 (Figures)

CHAPTER I. THE NECESSITY FOR PENSION REFORM

The analysis of country’s socio-economic situation reveals serious problems in the pension system resulting from both the pension policy and shortcomings in the pension system. Experts’ demographic forecasts depict an unpromising picture with regard to the possibilities for resolving the problems of the pension system in the coming years as a result of the expected deterioration in the sex and age composition of the population and destruction of the system’s financial sustainability.

1. Current situation (2000-2006)

1.1 GDP, wage and pension

1.1.1 The current socio-economic situation in the country is characterized by improving macroeconomic stability and unprecedented high rates of annual GDP growth. In 2000-2006, the real average annual GDP growth constituted 11.4%, and the GDP per capita in 2006 amounted to USD1,983. Due to the improvement of this indicator, Armenia was reclassified from a “low income” to “middle income” country in 2004

1.1.2 The share of wages in population’s gross incomes is increasing again, after a sharp decline in 1990s (around 15% in 1996). Recent household surveys reveal that the share of wages in the total monetary incomes of Armenian families has increased 1.9 times, from 27-28% in 1999 to 51-52% in 2006. Such increase in the share of wages is a result of the high rate of their annual growth, averaging at 13.4% in 2000-2006.

1.1.3 The highest rate of wage growth was recorded in 2003-2006, when the nominal value of the average monthly wage almost doubled, reaching AMD62,293 (or USD150). In the same period, wages were increasing at accelerated rates, exceeding the GDP growth rate by 1.5-1.6 times in some years (in 2003 and 2004). Moreover, considering the appreciation of the national currency since 2004, real wages, in U.S. dollars terms, increased by 2.5 times, i.e. faster than the real GDP per capita (2.3 times in 2003-2006).

1.1.4 In the same period (2000-2006), average pension growth rates in the country did not coincide with those of the GDP and wages (see Figure 1). Remaining practically unchanged in 2000-2001 (only 2.7-0.4 percent increase), the average real pension growth rate increased dramatically in 2002-2003 (19.1-22.7%), and a steady decrease was recorded in the following three years, from 11.9% in 2004 8.4 percent in 2006. As a result, the 2.3 times increase in pensions in 2000-2006, lagged behind the per capita GDP and average real wage growth (3.0 and 2.7 times respectively).

Figure 1. Dynamics of real GDP, average wage and average pension, 2000-2006

1.1.5 The unprecedented 122.7% increase in pensions in 2003 (compared to the previous year), was linked to the parliamentary and presidential elections of that year. The issue of state pensions is an important part of election campaigns in any country, which aims to receive the votes of the most active group of the electorate, i.e. pensioners.Euro-barometer surveys conducted by Eurostat record that the group of 60-69 year-olds in the electorate of the European countries is two times larger than the group of 18-29 year-olds, which creates huge pressures, especially in western European countries, for policies of increasing pensions.

1.1.6 As a result of parliamentary elections held in Armenia in 2007 the newly appointed government increased the pensions for 60% effective as of 1 January 2008. The basic pension recorded increase from AMD4,250 to AMD 6,800 orfor 60% and the value of one year of contributory service increased from AMD 230 to AMD395 or for 71.1%.

1.1.7 Thus, the comparison between annual growth rates of the GDP, wages and pensions brings us to the following conclusion:

1.2 Pension, living standard and poverty

1.2.1 The significant gap between the average growth of annual GDP and wage rates and pensions had a negative impact on the living standard of pensioners. Household and vulnerable groups’ surveys reveal that pensioners are one of the most vulnerable groups of the population in the country (see Box 1). The Household Survey 2005 indicated that 14.7% of pensioners are poor, among whom 6.2% are extremely poor . Exclusively elderly households are particularly vulnerable: every fifth such household is poor.

1.2.2 Pensioners' low level of living standard is a result of small pensions. Calculations reveal that in 2006 the average insurance pension amounted to 75.4% of the extreme poverty line and 50.0% of the poverty line (see Figure 2). Social pensioners’ situation is even worse: their average pension does not reach even 1/3 of the extreme poverty line. Pensions are the main source of income for the elderly and undoubtedly their low levels are a matter of concern.

Figure 2. Ratio of insurance pension to the poverty line and extreme poverty line

2001-2006

1.2.3 Pension issues, as a matter of concern to the public, were discussed on numerous occasions during the drafting of the Poverty Reduction Strategy Paper. According to PRSP forecasts, the average insurance pension should have exceeded the extreme poverty line by 2005-2006, and the poverty line by 2008 . The important precondition of maintaining the financial sustainability of the current pension system, however, makes it impossible to increase pensions.

1.2.4 Yet another PRSP forecast did not materialize as a result of small pensions: the ratio of average pension to average wage, i.e. the replacement rate. In 2006 the latter amounted to 16.9% instead of the forecasted 26.4% for all pensions (including social pensions) (see Figure 3).

Figure 3. Dynamics of overall replacement rate for 2002-2006, actual (NSS data) and PRSP forecasts

1.2.5 In-countries comparisons reveal that in 2005 Armenia at the lower replacement rate than all other CIS countries (see Box 2), except Georgia. At the same time, Armenia’s annual GDP growth rate was only second to Azerbaijan (2005-2006).

This means that:

2. Pension system

2.1 Current legal framework

2.1.1 Armenia currently has PAYG pension system

, which is based on the principle of solidarity between generations. As in other countries, Armenia’s solidarity system is based on financing from contributions of employed citizens. This means that the currently employed provide pensions for individuals who are passed working-age.

2.1.2 The current pension system of Armenia is financed from mandatory social insurance contributions (see Box 3), which by 2008 were accumulated in an extra budgetary account by the State Social Insurance Fund (SSIF) , which was established for managing the accumulated contributions. The RoA SSIF was levying, recording and accumulating the contributions. Since 2004 the function of levying contributions was transferred to the RoA State Tax Service. From January 2008 contributions are directly paid to the state budget of Armenia.

2.1.3 During the transition years the legislation regulating the pension sector was undergoing improvement mainly in terms of clarifying the definitions of “state security” and “social insurance” and reducing the social groups supported from the financial resources of the SSIF directly related to it. The main goal of implemented legislative improvements was to ensure financial sustainability of the pension system through certain parametric reforms (see Boxes 4 and 5).

2.1.4 However, in 90s it was possible to address financial sustainability of the pension system only with the condition of paying low pensions. As it was described in the previous section, the average amount of pensions was more than 2 times lower than the value of extreme poverty line, i.e. it was not sufficient even for food.

2.1.5. Legislative changes, however, did not ensure the most essential principle characteristic to pension insurance, i.e. Guarantee for equivalent compensation of personal income depending on person’s contributions in case of lost work capacity (due to age or other insurable cases). The foundation for the mentioned guarantee was laid in the State Pension Act (hereinafter the Act ) effective from April 2003.. The following legislative innovations were noticeable achievements in terms of pension insurance:

· Individual (personified) records system: in order to established the database needed for pension insurance (chapter 9, articles 40-44);

· Voluntary pension insurance: stipulation of its legal basis (chapter 1, article 4);

· Mandatory accumulation pension insurance: creating needed preconditions (chapter 14, article 76).

2.1.6 The Act also defined new pension entitlements: persons reaching 63 years of age with 25 years of insurance rate have the right to retirement, with the exception of retirement with privileged conditions, for which 55 and 59 years of age limits were defined. Entitlement of women to insurance pension, including to privileged pension was differentiated according to the scale approved by age groups. At the same time, the act defined the right to partial pension for those reaching 50 and 55 years of age, in cases where certain conditions are met and also for long-term services in special fields of activities

2.1.7 The Government of Armenia issued a number of Orders in 2003 to ensure the enforcement of the Act. The Orders defined the values of the base pension and one year of contributory service; industries, jobs, specialties, positions, categories which create entitlement to pensions with privileged conditions and partial pensions; types of work and positions creating entitlement to long-term service pension (including partial) and the regulations for registering the length of such services; etc.

2.1.8 Developments of the legal framework regulating pensions undoubtedly aimed to transform pension security into pension insurance. In practice however:

2.2 Description of the system

2.2.1 The pension policy is currently indispensable from the state social policy. Therefore it is developed, regulated and supervised by the state authorized body managing the policy, i.e. the Ministry of Labor and Social Issues. This is the reason that the current pension programs, in essence, are included in the two most important areas of social policy:

· Pension (social) insurance;

· Pension (social) security.

2.2.2 Pension insurance pension programs guarantee the realization of every citizen’s right to pension: when reaching retirement age, in case of the loss of breadwinner, disability and other defined cases. The insurance rate is the basis for calculations of insurance pensions. The insurance rate is a mandatory attribute for Armenian citizens who wish to participate in this program. The availability and length of the insurance rate is an attestation that mandatory social insurance contributions were made. Social insurance programs are financed from the budget formed by mandatory social insurance contributions. All types of insurance pensions are paid by the SSIF

2.2.3 Pension security covers citizens who do not participate in insurance programs and are not entitled to insurance pension. For example those who have never worked and/or have not accumulated the defined contributory service and/or have not reached the defined age for insurance pension. They are entitled to social pension based on old age, disability, loss of breadwinner and other cases defined by the Act. Social pensions are financed from the state budget and are paid by the SSIF.

2.2.4 Special pension security program is implemented for the servicemen, whose pension rights are stipulated and regulated by the Military Personnel and Their Families Social Security Act 1998. These programs are financed from the state budget. Pensions of military personnel are paid by the SSIF.

2.2.5 As at 1 January 2007, the total number of citizens included in all pension programs was nearly 570,000 or 18% of Armenia’s population. The majority, or 60%, are old-age pensioners. The next largest group is the group of disability pensioners with around 120,000 citizens, or 21% of all pensioners. The majority of disability pensioners, or 73%, receive insurance pensions. The second important group of pensioners is the group of military personnel, with more than 52,000 persons, or 9% of the total. The distribution of pensioners by pension programs and average pension is presented in Box 6.

2.2.6 Pension programs are implemented in all marzes of Armenia. The SSIF has 51 territorial centers and 3 payment centers providing services to beneficiaries, calculating and paying pensions. The centers are located almost in all cities of the country and 10 districts communities of Yerevan (see Chart 1).

2.2.7 Analyses of the operations of the pension system reveal that individual (personified) registry institution, which was the most important concept introduced by the Act (see paragraphs 2.1.5), is not fully operational yet (see Box 7). Accordingly, this infrastructure is not presented in Scheme 1. And if the regional registry mechanisms are not in place, it is impossible to enact the key principle of pension insurance, i.e. establishing a direct link between individual’s contributions and the amount of his pension. Consequently, if the mentioned principle is not enacted, the “social insurance programs” component of the existing system, in effect, cannot be considered as full-fledged insurance. Thus, it can be concluded that:

2.3 Types of pension

2.3.1 Pensions currently paid in Armenia are defined in two laws: Military Personnel and Their Families Social Security Act 1998 and State Pension Act 2003. The mentioned laws define the following types of state pensions:

1. Insurance pension, which includes

1) retirement

2) privileged conditions

3) long-term service

4) disability

5) loss of breadwinner

6) partial

2. Social pension, which includes old-age

2) disability

3) loss of breadwinner

3. Military personnel pension, which includes

1) long-term service

2) disability

3) loss of breadwinner

2.3.2 The insurance pension, as already mentioned, is the most common type of pension paid to Armenian citizens (see Box 6). 89.8% of all Armenian pensioners receive insurance pension. It consists of: basic pension and b) insurance component. The size of the basic pension is defined by the legislation and cannot be smaller than its previously defined size . The insurance component is calculated by a formula which is described in detail in the Act (chapter 4, article 16).

2.3.3 For each component of the insurance component formula (see Box 8), the Act defines the rules for calculation and/or registration. Chapter 10 of the Act is devoted in its entirety to the insurance rate (n) and its calculation. Article 15 of the same chapter defines the calculation of personal coefficients for up to 25 years of insurance rate and the value of the coefficient for more than 25 years of insurance rate.

2.3.4 The key element of the insurance component is the insurance rate value (V). The value is defined by the legislation and, as in the case of basic pension, it cannot be smaller (naturally by its nominal value) than the previously defined one. Thus, the two elements of the insurance pension: basic pension and insurance component, are determined by the legislation. The main source of their financing is the extra budgetary fund formed by mandatory social contributions

2.3.5 All types of insurance pension are interconnected and have the same elements. They are different only by: a) eligibility criteria (age, length of service and other data of personal and investigative nature defined by the law); b) conditions for pensioner’s registration; and c) methodology used for pension calculations. It must be noted that insurance pension calculations currently do not include all the elements defined by the Act (see Box 8).

2.3.6 The formula currently used for insurance pension calculations is the following:

MP = B + n x V x C where:

MP is the monthly pension;

B is the basic pension;

n is the years of insurance rate;

V is the value of one year of insurance rate;

C is the personal coefficient of the pensioner.

It is defined that the personal coefficient for 5-25 years of insurance rate is calculated by the following formula:

C = 0.04 x n The personal coefficient varies between 0.2 (for five years of contributory service) and 1 (for 25 years of contributory service).

The formula for calculating a personal coefficient of individuals with 25 years and more years of insurance rate is somewhat different:

C = 1 + [0.02 (n – 25)] In this case, the personal coefficient varies between 1.02 (for 26 years of contributory service) and 1.5 (for 50 years of contributory service) and more.

2.3.7 Social pensions are also regulated by the State Pension Act, but the principles are different from the insurance pension in two main aspects. First, it does not have an insurance component and second, which results from the first, is that it is totally covered by the state budget. Thus, specialists prefer to call it social benefit. A 65 year old individual whose insurance rate is less than five years is entitled to old-age social pension. Disability and loss of breadwinner social pensions are also allocated based on the same principle. The amount of old-age social pension is equal to the basic pension (see paragraph 3.3.2), which is defined by the legislation. The disability pension is also linked to the value of the basic pension. It amounts to 140% of the basic pension for disabilities of the first category, 120% for disabilities of the second category and 100% for disabilities of the third category. In case of loss of breadwinner, social pension to the amount of the basic pension is allocated to members of the family who do not have insurance rate. The social pension of parentless children is five times with the basic pension.

2.3.8 As already mentioned, social security, including pensions, of military personnel and their family members, is regulated by a separate law. Individuals of at least 50 years of age at the time of discharge from military service, who have 25 or more years of length of service (insurance), including at least 8 years in military service, are entitled to military personnel pension. Conditions for allocating pension in case of long-term military service, sizes of pension, regulations for calculating length of service, as well as the legal basis for pension security allocation to family members in case of disability or death of military servicemen, are defined in the Military Personnel and Their Family Members Social Security Act. Pensions of military personnel are financed from the state budget.

2.3.9 The most common type of military personnel pension is the long-term service pension. The principles for its calculation are different for officers and privates. In case of officers, the pension has two components: a) monthly monetary allowance and b) food compensation. For privates, the basic pension is the basis for calculation of long-term service pension. All the mentioned components of military personnel pensions are defined by the legislation. According to legislation, the size of those pensions cannot be smaller than 150% of the basic pension.

2.3.10 Analysis of the types of pension allows us to conclude that:

2.4 Expenditures and revenues

2.4.1 By January 2008 expenditures on pension programs in Armenia were financed from two main sources. directly from the state budget (funding of pension security programs) and budget of contributory resources of the pension fund (Funding of pension insurance programs).

The latter was mainly formed from mandatory contributions (see Box 9). As a result of SSIF restructuring the pension fund budget was dissolved and starting from January 2008 the funding of the pension system was provided only from the state budget. In fact, as a result of system modifications contributions were also added to the state budget revenues. The Ministry of Labor and Social Issues is in charge of developing pension programs and their annual budgets.

2.4.2 Major part of the revenues of the pension budget, nearly 80% in 2007, comes from mandatory social insurance contributions. Rates and payment procedures, as well as the legal, financial and organizational bases for their collection are regulated by Mandatory Social Insurance Contributions Act 1998.

2.4.3 According to the mentioned law, social contributions are made by:

· Employers;

· Employees;

· Individual entrepreneurs and notaries public (chapter 1, article 3, paragraph 4) Social contributions of employees amounts to 3% of their wages and other tax-deductible incomes. Individual entrepreneurs should make a minimum social contribution of AMD5000 every month. Employers make social contributions for each employee according to the defined rates (see Box 10) from the payroll fund and other tax-deductible income. Employer’s social contribution rates are regressive and diminish as wages increase. For wages up to AMD20,000 it amounts to 28% vs. the wage fund and 20.1% - in case of a wage amounting to AMD250,000 (see Annex 3, Table 1).

2.4.4 The other source of revenue for the pension fund is allotments from the state budget. In 2007 they constitute more than 20% of total revenues. Usually, possibilities for financing pension programs are determined based on the size of income tax collected from employees. Rates and payments of income tax are regulated by the Income Tax Act 1998. The defined income tax rates are progressive and increase from 0% for AMD20,000 monthly income, to 13% and more for AMD250,000 and more incomes per month. Calculations of social contributions and income taxes for some levels of wages, including the average nominal wage in the country, are also presented in Annex 3, Table 1 of this document.

2.4.5 The structure of expenditures for implementing pension programs in much more divers. According to the approved budget of mandatory social insurance for 2007, there are, in general, 27 expenditure items. The largest item is the allocation to pension insurance, amounting to 72.1% of total expenditures (see Figure 4).

Figure 4. Structure of mandatory social insurance expenditures in 2007

2.4.6 In order to study the dynamics of the annual budget for pension programs in 2000-2005, certain expenditure and revenue items were grouped together. Analysis of the summary data (see Table 1) reveals the following interesting trends:

· All pension programs are included in one comprehensive budget program titled “The mandatory social insurance budget”. But the budget also includes pension security programs.

· Expenditures on pension insurance do not exceed the revenues from mandatory social contributions.

· Pension security expenditures increase 1.5 times faster than pension insurance expenditures.

· The volume of social contribution made by employers increases 1.5 times lower than that of the employees, and 4.3 times slower than contributions made by individual entrepreneurs.

· The positive balance of the budget is a result of restraining expenditures on insurance pensions in consideration for revenues received.

· Part of pension security programs are, either directly or indirectly, financed through savings in expenditures on pension insurance programs.

Table 1. Performance of mandatory social insurance budget, 2000-2005

Revenues, mil. AMD Including:

Expenditures, mil. AMD Including:

* The growth rate is calculated for 2000-2004.

** Those employed and self-employed in agriculture are exempted from making social contributions since 2003.

2.4.7 The mandatory social insurance budget of 2007 does not plan any significant changes in the trends described above (see Table 2). Mandatory social contributions will continue to surpass state budget allocations at higher rates, but the ratios of pension insurance and other indicators will remain unchanged.

Table 2. Comparative analysis of mandatory social insurance budgets 2005 (actual) and 2007 (approved).

times

2.4.8 Analysis of the finances of current pension programs allows us to conclude:

3. Population and labor market

3.1 Perspectives of the demographic situation

3.1.1 The policy of restraining insurance pensions it basically predicated on the unfavorable perspectives of the demographic situation. The population of Armenia has had a declining trend since 1990. The reasons were changes in the natural growth of the population and also its migratory movements. At the beginning of 2004, for the first time, an increase in the number of population was recorded. This trend is still continuing. As at 1 January 2006, Armenia’s population was 3219.2 thousand. This is an increase of 0.2% compared to 2001, but still 8.4% lower than the figure for 1990 (see Annex 3, Table 2)

3.1.2 Changes in the age composition of the population are much more important from the viewpoint of pension system reforms. Those changes in Armenia indicate the rapid aging of the population. Between 1990 and the beginning of 2006, the proportion of children up to 14 years of age decreased by 10 percentage points (from 30.6% to 20.6 percent) and the proportion of 65+ population – for 1.7 times (from 5.6% to 9.7%) (see Annex 3, Table 3).

3.1.3 According to actuarial calculations made by using the World Bank pension forecast computer model (PROST), the indicator for aging of the population, i.e. The proportion of the 63+ group in the entire population will amount to 15.6% in 2021 and will reach 26.7% in 2080 (see Figure 5).

Rates of aging of the population in Armenia are impacted by:

· Higher life expectancy, particularly starting from the retirement age;

· Rapid decline of birthrate;

· Emigration of working-age population from the country (see Annex 3, Table 4).

3.1.4 In 1990-2005, increases of average life expectancy were recorded among both women and men. In 2005, they amounted to 76.5 and 70.3 years respectively, compared to 73.4 and 67.9 years in 1990 (see Annex 3, Figure 1 and 2). The difference between average life expectancies of women and men is a result of higher mortality rate among men (see Annex 3, Figure 3).

Figure 5. Forecast of population’s distribution by age groups

(percentage of the total population)

3.1.5 The average life expectancy starting from the retirement age, which is important from the viewpoint of pension system reforms, is increasing steadily. According to forecasts using PROST, by 2021 this indicator will amount to 19.7 for women and 16.0 for men, and 24.1 and 19.6 respectively by 2080 (see Figure 6).

3.1.6 A declining trend of the birthrate, as another factor influencing the aging of the population, is recorded since early 1980s (see Box 12). The trend slightly slowed down after the 1989 earthquake as a result of the compensation phenomenon. From early 1990s, however, the birthrate started to decline again. In 2005, the total coefficient of birthrate (fertility) amounted to 1.37, compared to 2.62 in 1990, not ensuring even the simple reproduction of the population. According to forecasts using PROST, although the total coefficient of the birthrate is on the rise, it will reach the level of simple reproduction (2.15) only starting from the mid of 2030s (Figure 7).

Figure 6. Forecast of average life expectancy after retirement 2005-2080, years

Figure 7. Forecast of total coefficient of birthrate, 2005-2080

3.1.7 The third important factor with an impact of the number of population and its sex-age composition is migration. Since 1990s, the external migration balance of the country is negative (emigration exceeds immigration). This is a result of the political and socio-economic situation in the country. The current procedures for administrative registration of migration (see Box 14) does not allow for assessing the real size and composition of emigration. According to official statistics, the negative balance in 2005 was -7.8 thousand people, which is a 25% reduction compared to 2001.

3.1.8 Although the state registration of migratory movements does not have the capacity to report their real sizes, the analysis of its data enables us to describe the essence of the migration structure. In 2000-2005, there was an increase in the proportion of men among citizens who left the country and were removed from the registry. In 2005, the majority of people who left the country and were removed from the registry were 20-49 year old, among both men (50%) and women (67%). An increase in the proportion of men has also been recorded in the mentioned age group (see Annex 4, Table 3).

3.1.9 Some improvement of the demographic situation in 2003-2004 due to a small increase in birthrate coefficient, is not yet adequate for a sustainable demographic development. From the viewpoint of simple reproduction of the population, the quite low level of birthrate coefficient continues to deteriorate the age composition of Armenia’s population (see Box 13).

3.1.10 From the viewpoint of pension reforms, the deterioration of population’s composition is primarily reflected in population’s dependency ratio. Increase in the ratio shows how much more retired citizens are dependent on the number of the working-age population. The declining trend of this indicator recorded in 1990-2005 will come to a halt in 2012. Thereafter, in the entire forecast period (with the exception of a few years in 2008-2015), the dependency ratio will steadily increase. In 2080 it will exceed by 2.5 times the projected level of 2009.

3.1.11 Current and forecasted trends of Armenians demographic situation indicate that:

3.2 Employment and unemployment trends

`3.2.1 Official statistics record development trends in Armenia’s labor market with their own particular logic, which are characteristic to transition countries. As a result of the economic decline in 1990s, the country had negative GDP growth rates up to 1993 (see Figure 8). In spite of that, a rise in employment was recorded in 1991. And in 1993, with the start of the economic recovery, employment rates started to decline. The light does such a if if and In the following years (1994-2004), the link between GDP growth and overall employment rate was negative, since the GDP grew steadily

Figure 8. Dynamics of the GDP, overall employment rate and labor productivity (percentage, 1990=100)

* Calculated as the ratio of GDP and employment indexes.

3.2.2 Comparison of GDP and employment trends reveals that economic growth in Armenia was accompanied not by employment, but by increase in productivity. Productivity started to grow in 1995 and in 2002 already exceeded its level of 1990.

3.2.3 With regard to structural changes of employment in the same period, a significant flow of the employed from the public sector to the private sector of the economy was recorded. In 2005, the proportion of those employed in the private sector of the economy amounted to 79.9%, compared to 18.4% in 1990. As a result, proportion of those employed in the public sector reduced by nearly four times. In 2005 it constituted 19.8%, compared to 79.1% in 1990. Changes in the employment structure of the population by fields of economic activity are also significant. In the mentioned period, the number of employed in manufacturing and construction reduced by 500 thousand. In 2005, those two sectors accounted for 15.5% of the total employment (compared to 41.7% in 1990). At the same time, the number of employed in agriculture almost doubled. In 2005, 507.6 thousand people, or 46.1% of all the employed, were employed in agriculture

3.2.4 From the viewpoint of pension reforms, it is extremely important to analyze the structural changes of the employed population by employment status. Waged employees in 2005 constituted 48.2% of the employed, which is nearly half the level of 1995. On the contrary, the proportions of the self-employed and those employed in agriculture have increased significantly, respectively 6.0% and 45.8% in 2005 (see Annex 4, Figure 8).

3.2.5 The other important indicator from the viewpoint of pension reforms is the economic activity of the population or labor force participation indicator. This indicator is impacted by demographic trends in the country. The rate of economic activity of the population in Armenia (calculated as the ratio of the numbers of economically active population and labor resources) constituted 57.7% in 2005, compared to 80.3% in 1990. This 1.4 times decline of the indicator is a result of the following factors impacting the labor force supply:

· low birth rate

· population outflow

· informal employment.

3.2.6 Birthrate has a direct impact on the ratio of numbers of employed to the working-age population. Growth of the birthrate until early 1980s was reflected in the increase of the absolute number of working-age population in 2000s (see Annex 3, Table 2). At the same time, the decline of birthrates in 1980-2003 relatively increased the proportion of working-age population (see Annex 3, Table 3) from 63.9% in 1990 to 69.7% in 2006. In this demographic situation (when a labor force surplus was recorded due to high birthrates of 1960s-70s), the demand for labor force reduced in the labor market and the proportion of the employed in the working-age population declined to 53.5% in 2005 (compared to 83% in 1990). Of course, the decline of the overall employment rate had a negative impact on the indicator as well. However, the employment rate in 1990-2005 declined by 32.7% (see Figure 8), while the ratio of the employed to working-age population dropped by 64.8%.

Figure 9. Non-working-age populations support rates, 1990-2080

3.2.7 The ratio of the employed to working-age population would have been much lower if there was no laborforce outflow to other countries. As described in paragraphs 3.1.7-3.1.8 and box 14, the migration of the working-age population, which had the highest proportion in the total outflow of the population, somewhat mitigated the tension in the labor market. Population’s labor migration, having a positive impact on labor market tension, nonetheless, worsened another important socio-economic indicator, i.e. the dependency ratio of the employed to non-working-age population (0-14 ¨ 65+), the so-called overall support rate of the population.

3.2.8 The overall support rate of the population characterizes the social “load” of the country and is the basis for drafting state social policy. Higher values of the indicator reflect larger capacities of the population for social support. This rate was lower in 2005 compared to 1990, respectively 1.06 and 1.28 (see Figure 9). Although a small improvement of the indicator has been recorded in the last 4-5 years, but, according to forecasts, it will start to drop again from 2012. Unfavorable labor market and demographic trends limit not only the pension policy, but also the possibilities of the entire social policy, despite current economic growth rates.

3.2.9 Part of the economic growth has been achieved in the informal sector. The size of this sector, according to NSS assessment, is around 27%. Informal or unregistered employment, amounting to 140-180 thousand people according to various sources.

If we also take into account the hidden or so called “discouraged” unemployed, who do not refer to the employment centers, since they do not believe in their real support for finding jobs , then the gaps between survey data and official statistics regarding employment and unemployment rates and trends, as described in Box 15, will become understandable.

3.2.10 Within the context of pension reforms, there are certain problems related to the registration of those employed and self-employed in agriculture. They are exempted from taxes and conditionally can be included in the informal segment of the economy. According to the current legislation, they are not included in the list of entities subject to mandatory social contribution payments.

3.2.11 According to the assessment of the Central Bank

, nearly 53% of the employed are currently left out of the pension system registry, since they are employed in agriculture and/or have hidden employment and/or are self-employed. These are citizens who, by the “permission” of the legislation or due to the situation in the labor market, refrain from making income tax and social contribution payments.

But recording to State Pension Act, the citizens currently receive and in the future will also receive either insurance or social pension. Naturally, this is an obstacle to ensuring the sustainability of the current pension system.

3.2.12 There are registration issues also related to the assessment of the unemployment rate. For 2005 only, depending on the methodology for defining unemployment, its level fluctuated between 8.2% (officially registered unemployment) and 31.2% (unemployment calculated based on data from labor force sample survey according to ILO methodology) (see box 16). Unemployment was officially registered in the country for the first time in 1992 at 1.8%. The official unemployment rate reached its peak in 1999 at 11.2%. Later it gradually declined amounting to 7.5% in 2006.

3.2.13 Such disproportions linked with informal, including agricultural employment, and hidden employment, are characteristic to labor market in Armenia (see Figure 10). As a result, it has become impossible to forecast and take fully into account their data in labor market indicators. Thus, the actuarial calculations by PROST for forecasting employment rates were based on application of the pattern of the above mentioned trends.

Figure 10. Dynamics of the number of employed by employment status 1995-2005 (proportion of the total number of employed)

3.2.14 Analysis of the employment and unemployment trends in 1990-2005 allows us to conclude that:

4. Financial sustainability of the system

4.1 Social and pension expenditures

4.1.1 The Constitution

declares that “The Republic of Armenia is a sovereign, democratic state, based on social justice and the rule of law”, which obliges the Government of Armenia to pay particular attention to expenditures on the social sector. In 2000-2006, the physical volume of social expenditures in the consolidated budget increased by 2.6 times, amounting to 42.2% of total expenditures in 2006 (see Figure 11). The ratio of social expenditures to the GDP, however, is still under 9%, with some decline in 2006 compared the previous year (see Figure 12). The ratio of social expenditures to GDP in Armenia is a fraction of that of the European Union countries and two times smaller compared to the three Baltic republics (see Figure 17).

Figure 11. The proportion of social expenditures in the consolidated budget, 2000-2006

Figure 12. Ratio of social expenditures to GDP , 2000-2006

4.1.2 Expenditures on pensions constitute the largest component of total social expenditures. In 2000-2006, the proportion of pension expenditures was in the range of 36-41%. In 2006, it amounted to 38% (see Figure 13). It must be noted also that pensions constitute more than 2/3 of the total social insurance and social security expenditures: 72.5% in 2005 and 69.6% in 2006.

4.1.3 Comparison of the dynamics of pension expenditures in 2000-2006 with replacement rates reveals the following important trend: Volumes of pension expenditures increased by 2.6 times, while their ratios to the GDP (3.4%) and total social expenditures (39%) remained unchanged. The replacement rate, however, declined by 22% in the same period: amounting to 16.9% in 2006 compared to 20.6% in the previous year.

Figure 13. Composition of social expenditures 2000-2006, percentage of total expenditures

4.1.4 The above-mentioned analysis of social and pension expenditures reveals that:

4.2 Financial sustainability and pension adequacy advocacy

4.2.1 Financial sustainability of the pension system is a crucial issue in any country. It became even more important in 1980s, when demographers all around the world recorded serious changes in demographic trends and warned about the steady trend of population aging in the world. In those years, governments became concerned about the financial sustainability of their pension systems. Academics and experts faced the challenge of assessing financial sustainability based on forecasts of demographic trends. The formula currently used for calculating the financial sustainability of pension systems is the following:

AP x NP = AW x NC x CT Where:

AP – average pension NP –number of pensioners (insurance) AW – average wage, from which contributions are paid NC – number of actual contributors CR – current contribution rate.

4.2.2 The average insurance pension for 2005 was calculated by the mentioned formula:

AP = {AW x NC x CR}. NP = {AMD52.1 thousand x 435.5 thousand people x 25.8%}: 476.5 thousand people = AMD12,285

Comparing the calculated result (AMD12,285) with the actual average insurance pension (AMD 11,863) presented in Box 6 of section 2.2, we can state that we have reached the “limit” of the financial sustainability of the pension system

4.2.3 Financial sustainability of the pension system, naturally, depends on the behavior of its components (see the formula above). From this point of view, it is very important to analyze the dynamics of those components and factors impacting them. This analysis will be the basis for future forecasts.

4.2.3.1 Number of pensioners (insurance) (NP) largely depends on the perspectives of the demographic situation the country, which are described in detail in section 3.1. Forecasts based on demographic trends in the country reveal that the gap between the numbers of retirement age (63+) and working age (15-63) populations will become wider. Experts call this the “pipeline effect” and forecast a dangerous decline in “pipeline conductivity” by 3.3 times (see Figure 14). In calculations of the financial sustainability of the pension system the purely demographic indicator “retirement age (63+) population” is replaced by the “number of pensioners” which currently has a higher value. In 2005, the difference between the mentioned indicators was more than 165 thousand people. The higher number of actual pensioners is a result of the following factors:

· Until 2002, 55+ retired women and 60+ retired man were included in the actual number of pensioners,

· Inclusion of pensioners who retired with privileged conditions,

· Inclusion of those who received pensions due to disability or loss of breadwinner.

In effect, the “pension load” of the system was larger and consequently the “pipeline conductivity” was lower.

Figure 14. The “pipeline affect” of Armenia’s demographic composition, 2005-2080

4.2.3.2 The next important element of financial sustainability of the pension system is the number of actual contributors (NC), which is a derivative of the “working-age (15-63) population” demographic indicator. Analysis of the specific manifestations of the labor market and its trends presented in section 3.2 reveals that the differences between the mentioned two indicators are more tangible. In 2005, according to official statistics, labor resources amounted to 2072.4 thousand people, while according to the state social insurance fund information, social contributions were made by 435.5 thousand people, or 4.8 times less (see Annex 3, Table 5). This difference is a result of the following factors:

· In the total number of the working resources, 1175.8thousand people or 56.7% were employed. The figure includes employment in the informal sector and agricultural employment.

· Only 44.2% or 519.3 thousand of the employed were working in the formal sector of the economy

· Only 83.9% or 435.5 thousand people employed in the formal economy made social contributions.

As a result, the pension load is redistributed over a lower number of working-age population.