r+3 6586 0 PSP Discussion Paper Series Ukraine: Reforming the Pension System Cheikh T. Kane January 1996 Poverty and Social Policy Department Human Capital Development and Operations Policy The World Bank This Booklet of Abstracts contains short summaries of recent PSP Discussion Papers; copies of specific papers may be requested from Patricia G. Sanchez via All-in-One. The views expressed in the papers are those of the authors and do not necessarily represent the official policy of the Bank. Rather, the papers reflect work in progress. Tlhey are intended to make lessons emerging from the current work program available to operational staff quickly and easily, as well as to stimulate discussion and comment. They also serve as the building blocks for subsequent policy and best practice papers. ABSTRACT This paper provides a detailed discussion of all the programs comprising the ULkrainian pension system. It highlights the redundancies of the system as well as some of its regressive features. Efficiency issues are also addressed through a cross-country comparison of marginal effective taxes on labor. On average these taxes are higher in Ukraine than in OECD countries. A simulation model and a cross-country comparison of implicit pension debt are used to highlight the pension system' structural fiscal imbalance. In order to maintain a viable pay-as-you-go system a comprehensive reform that includes an increase in the pension age to 65 would be necessary. The report also shows that provided the promises of the public pension system are scaled back Ukraine could meet the challenge of stabilization, reduce labor taxation, and introduce a multipillar pension system with a fiily-finded component. i I I Glossary Adverse Selection People with higher risks are more likely to seek insurance coverage than people with average or below average risk exposure. Benefit Ratio Ratio of average pension to average wage. Cost Rate Ratio of value of benefits to the covered wage bill. Defined Contribution In these plans the worker's annual contribution is defined in advance but the benefit --which depends on the investment return-- is not. The worker bears the investment risk, as the pension varies according to contributions and investment earnings on contributions. Defined Benefit Plans In these plans, the pension forrnula is defined in advance usually based on years of service and earnings (e.g., the worker is guaranteed an agreed slice of his/her final salary). The rest of society bears the risk that the economy does not do well or people live longer. Moral Hazard Individual causing a deliberate loss in order to collect the proceeds of an insurance policy. Morale Hazard Attitude of indifference to loss created by the purchase of an insurance. Pay-as-you-go System Benefits received by current retirees are financed by contributions paid by current active workers. Population Dependency Ratio of population over 60 to that between 20 and 59. Replacement Rate The proportion of the average wage given as pension. Saving Rate of wealth accumulation. Savings Stock of financial assets. System Dependency Ratio of beneficiaries to contributors. ii I I I OUTLINE OF THE REPORT ABSTRACT ............................................. i GLOSSARY ............................................. ii OUTLINE OF REPORT .................................................... iii CHAPTER I.......................................... THE FUNCTIONING OF THE UKRAINAN PENSION SYSTEM .......................................... 1 I. INTRODUCTION .............................................. I I. THE OLD-AGE PENSION ......................................................... . .......... .... 2 m. DISABiL1TY PENSION .... 3 IV. SURVIVOR PENSION ... 5 V. SERVICE PENSION .... 6 VI. SOCIAL PENSIONS ..............7 ............... 7 VII. YEARS OF SERVICE VS YEARS OF CONTRIBUTION ................................................ 8 VII. PENSION CALCULATION ................................................ 9 A. The Old-Age Pension ................................................ 9 B. The Disability Pension . : 13 C. Survivor Pension ....................... 14 VIIII. SUMMARY ........................ 15 CHAPTER II ............................. 16 FISCAL SUSTAINABILITY OF THE PENSION SYSTEM .................................................... 16 I. INTRODUCTION ................................................ 16 II. THE MACROECONOMIC CONTEXT ................................................. 16 A. Pension Fund and Consolidated Fiscal Deficit ................................................ 16 B. Inflation and Pension Fund Deficits .................... ............................ 17 C. The Stabilization Program ................................................ 19 III. TRANSITIONAL IssuEs IN THE LABOR MARKET .......................................... 20 A. Labor Taxation ................................................. 20 B. Recent labor market developments and future trends: .................................... 22 C. Lessons from Other Transition Economies ................................................ 22 iii I I I IV. BASELINE FINANCIAL PROJECTIONS OF THE PENSION FUND ....................................... 24 A. Presenting the Baseline Scenarios: ........................................................ 24 B. Projection Results ........................................................ 25 V. TE PENSION FUND'S IMPLICiT DEBT ........................................................ 30 CHAPTER IH .....................3....2......................................... 32 TRANSITION AND POST-TRANSITION REFORM OPTIONS .32 I. INTRODUCTION .32 II. REFORM OBJECTIVES .32 A. Meeting the Challenge of Transition .32 B. Post-Transition Objectives .33 III. PENSION LEVEL AND COVERAGE: WHAT IS AFFORDABLE . 33 A. Sharing the Burden between current and future pensioners .33 B. Affordable Coverage. 3 5 IV. OPTIONS FOR KEEPING THE PAY-As-You-Go SYSTEM:. 3 7 A. Increasing the Contribution Rate is Not Desirable .37 B. Making the Pay-As-You-Go Fiscally Sustainable .37 V. INTRODUCING A DEFINED-CONTRIBUrION PENSION SCHEME AFTER THE TRANsmoN. 41 iv I CHAPTER I THE FUNCTIONING OF THE UIKRAINIAN PENSION SYSTEM I. Introduction In December 1991, the Ukraine became independent of the Union of Soviet Socialist Republics (USSR) and took sole responsibility of its Social Security Program. The Soviet system itself was established in 1921 and was significantly modified in 1956 and 1964. A new pension law was adopted in the USSR in 1990 with separate legislation and provisions for the Military, and each republic modified their own version of the pension law in 1990 as well. There are mainly three subsystems in the overall social protection: (i) the pension fund; (ii) the social insurance fund; and (iii) the unemployment fund. This report deals with the pension fund alone, which covers the following five programs: (i) Old-Age; (ii) Disability; (iii) survivor (or loss of breadwinner); (iv) length of service; and (v) Social pension. The relative importance of these five programs with respect to total pensioners (i.e., stock) or new pensions granted during the year (i.e., flow) is summarized in Table 1.1. Table 1.1: Number of Pensioners by Major Programs (1000s) (January 1, 1994) All Pensioners New Pensioners (Stock) (Flow) Total 13,596 (100%) 983 (100%) 1. Old-Age 10,865 (80%) 688 (70%) -standard 8,805 (66%) 539 (55%) -preferential 2,060 (14%) 149 (15%) 2. Disability 1,355 (10%) 153 (16%) 3. Survivor 837 (6%) 78 (8%) 4. Service 39 (0%) 13 (1%) 5. Social 501 (4%) 51 (5%) Source: Pension Fund I I IL The Old-Age Pension This program is by far the most important one accounting for about 80 percent of total pensioners and 70 percent of newly granted pensions (January, 1993). This program can be subdivided into two categories: the standard Old-Age pension, and the Preferential Old-Age pension. About 80 percent of Old-Age pensioners fall under the standard sub-category. As shown in Table 1.2, eligibility conditions for the Standard Old- Age pension are determined by age and total years of service (TYS). Men can draw a pension starting at age 60 and after at least 25 years of service. For women, eligibility starts at age 55 after at least 20 years of service. Eligibility conditions for the Preferential Old-Age pension varies across sectors based on age, total years of service, and years of service in the sector subject to a preferential regime (YSS). For instance, men working in activities belonging to the so called "List I", are eligible to a pension at age 50 after 20 years of service, of which a minimum of 10 must be effectively spent in the listed area. In general, the age requirement for the preferential regime varies between 50-60 for men and between 45-55 for women. Workers in mining and metallurgy, who belong to the preferential regime, are entitled to an Old-Age pension based on years of service alone. It is important to note that according to the pension law, some of the Old- Age pensions subject to a preferential retirement condition have to be paid by the firm directly up to the point the worker reaches the standard retirement age. After reaching the standard Old-Age pension age, the pensions start being paid by the pension fund. This measure of having the firm bear the direct cost of the preferential retirement provision is supposed to be implemented gradually, starting with 50 percent of this cost in 1991 and reaching 100 percent in 1996. As we shall see, this provision is the key difference between the preferential retirement provisions of the Old-Age program and the so called Service Pension. It should be mentioned that the years of service requirements for the Old- Age program are not limited to period of effective work, but include in some cases time spent by mothers taking care of children as well as time spent in school. In other words, these "years of service" do not necessarily imply years of contribution. Assuming the earliest entrance into the labor market occurs at age 20, the years of service under the standard Old-Age pension would then vary between 40 and 25. This did not mater as much in a central planning economy because of guaranteed job provisions and limited scope for informal employment. In market economies, vesting periods have important implications for formal labor market participation (see Section VII for more discussion). 2 I I Table 1.2: Eligibility Rules for the Old-Age Program MEN WOMEN YSS Ag T l____________Age TYS e YS YSS I Standard 60 25 55 20 II Preferential 2.1 List I 50 20 10 45 15 7.5 2.2 List II 55 25 12.5 50 20 10 2.3 Agriculture 55 25 20 50 20 15 2.4 Textile - - --- 50 . 20 ? 2.5 Drivers 55 25 20 50 20 15 2.6 Mining & ANY 25-20 ? ANY 25-20 ? Metallurgy I_I_I I_I Note: TYS= total years of service required YSS = years of service required in the sector. The value of the pension is set at 55 percent of the average salary for workers retiring after the minimum vesting period of the standard Old-Age program: 25 years for men and 20 from women. The pension, however, is subject to both a minimum and a maximum pension. For this reason, the effective replacement rate (value of pension relative to the average salary to be replaced) will vary with the statutory replacement rate and across income groups as well (see Section VEII). The statutory replacement rate is increased by 1 percentage point for every year of service above the minimum requirement. The statutory replacement rate is subject to a ceiling of 70 percent, except for workers under the so called list I, for whom the ceiling statutory replacement rate is raised to 85 percent. m. Disability Pension The Disability program is the second most important one behind Old-Age, accounting for 10 percent of total pensioners and 16 percent of newly granted pensions (January 1994). Disability pensions are granted for total or partial disability resulting either from on-the-job or general disease or injury. However, disabilities from childhood are not treated. Students are generally treated as being employed. Employment-related related disabilities are granted regardless of the length of employment. Disabilities stemming from general disease require a number of years of employment that varies with the age of onset of disability as appearing in Table 1.3. The years of employment for sickness-related disabilities vary between 1 and 15. When the 3 I worker has been employed for an insufficient period to be granted a full pension, the disability benefit is granted proportionally to available employment duration but no lower than the social pension, which is a multiple of the Old-Age minimum pension (see Section VI). Table 1.3: Years of Service Requirements for Disability Pensions Age at Onset Years of Employment < 23 1 23 - 25 2 26 -30 3 31 -35 5 36 -40 7 41 - 45 9 46 -50 1 1 51 -55 13 56 - 60 14 61+ 15 The disability types are classified into three categories by decreasing order of severity (Group I, II, IE). The amount of a disability pension is determined as a percentage of earnings and vary with the degree of disability: 70 percent for Group I, 60 percent for group II, and 40 percent for group HI. The minimum pension is set at the level of the social pensions for the corresponding category of disability. When disability has occurred during a war or as a result of unjustified political repression, the minimum pension is 3 times the minimum Old-Age pension for group I and II disabled, and 1.5 the minimum Old-Age pension for group m disabled. Maximum pensions are 3 times the Old-Age pension, except for those disability workers from list I, for whom a ceiling of 4 minimum pensions apply. If a disabled person has accumulated enough years of service to qualify for an Old-Age pension, the disability pension received is equal to that of an Old- Age. Unemployed disability pensioners are entitled to a pension supplement when they have disabled family members as dependents, provided the dependent is not already receiving a social or labor pension. The pension supplement is equal to I social pension per disabled dependent. Single disability pensioners of Group I and II who need constant care are also entitled to a pension supplement of 1 social pension. Disability pensioners who reach the normal pension age are also entitled to a pension supplement of 1 social pension. The pension law explicitly allows for the possibility of cumulating two pension supplements when both criteria are met: (i) unemployed disabled with a disabled dependent, and (ii) victim of disability m at normal age of retirement. It should be noted that because some disability supplements are tied to the normal retirement age, changing this age would reduce the future cost of both the Old-Age and the Disability programs. 4 I Disabilities are granted for a duration that is determined by bodies of socio- medical commission. Men disabled over age 60 and women disabled over age 55 are granted pensions for life. Once the duration is determined, reexamination is made only upon application. Otherwise, pensions are automatically terminated at the end of the month of recovery. IV. Survivor Pension This is the third most important program and covers 6 percent of total pensioners and 8 percent of newly granted pensions. Survivors of a deceased worker are granted pensions if they are (i) disabled dependents, (ii) children under 18, (iii) aged spouse or parent (60 for men and 55 for women), (iv) non dependent family member who lost his source of income, (iv) family member caretakers of children under 8. Each survivor is entitled to a pension. Inmates and students are entitled to a survivor pension, but not after they reach 23 years of age. Families of deceased inmates and students are also entitled to a pension under the same conditions as the disability pension for this same group. All family members who were fully supported by the deceased or received constant assistance, which was the main source of living, are considered dependents. Members of a family who qualify as dependent but who received a pension have the option to switch from their pension to a new survivor pension. For disabled dependents, the disability must have occurred prior to reaching 18 years of age. If death is work related, the full survivor pension is granted regardless of years of service of deceased. If death is not work related, then a full survivor pension is subject to a years of service provision, which is similar to that of disability benefits that are not work related (see Table 1.3). The basic full-rate pension of a survivor is 30 percent of the worker's earnings, but no lower than a social pension. As shown below, the social pension varies across categories of pensioners and is a function of the minimum Old-Age pension (see Section VI). The full rate survivor pension applies to each disabled member of a family. This implies that the survivor pension can exceed the average wage of the deceased if the number of dependents exceeds 3. Pension is given on a prorata basis when eligibility is tied to years of service, which are not fully met. Children with a disabled single mother or having neither father nor mother have a minimum pension of 2 social pensions. Full rate pensions are granted if the worker was employed for a specified duration that varies with age at death. The duration is the same as required for a full rate disability that is work- related (see Table 1.3). Pensions are granted on a prorata basis for workers with insufficient duration of employment to earn a full-rate benefit. Orphans who are taken care of by the state receive a full survivor pension. Other children under state support receive 50 percent of the survivor pension. The value 5 of survivor pensions increases or decreases as the total number of family member changes, keeping the total pension unchanged. V. Service Pension This program is the smallest within the Ukrainian pension system, covering 0.3 percent of total pensioners and 1 percent of newly granted pensions. It is intended for workers with occupations that are considered detrimental to abilities even prior to the normal retirement age. As a result, these categories of workers are granted pensions under more liberal eligibility conditions. The program covers workers in aviation, railroad, truck drivers at mines, longshoremen, seamen, lumbermen, artist, long-term care workers, and sportsmen. In general, these workers are entitled to retire at a younger age or with a shorter duration of employment. In addition, these workers may include in the years of service periods of military service and education. The key difference between the Service Pension and the Old-Age preferential retirement has to do with who is responsible for the payment of pensions. For the Service Pension, benefits are filly paid by the pension fund. For the Old-Age preferential retirement, the company is ultimately responsible for the payment of pensions prior to the normal retirement age. The specific conditions for eligibility by activities are summarized in Table 1.4. A number of activities are entitled to a service pension based on years of service alone. For women in such category the minimum years of service is typically 17 or 20 years. For men, the corresponding minimum is 20 years. The remaining activities, which are entitled to service pension, are subject to a minimum pension age: 50 or 55 for men, and 45-50 for women. It should be noted that the pension is calculated the same way as the Old-Age pension. 6 Table 1.4: Basic Requirements for the Service Pension Activities Men Women Age TYS YSS Age TYS YSS I Aviation -Aircraft Crew ANY 20-25 20-25 ANY 15-20 15-20 -Air Traffic Disp 50 25 12 45 17.5 10 -Traffic Control ANY 20 20 ANY 17 17 -Engineers 55 25 20 50 20 15 -Stewards 55 25 15 45? 20 10 II Other Sectors -Subway Railroad 55 25 12.5 50 20 10 -Expeditor 55 25 12.5 50 20 10 -Timber Cutting 55 25 12.5 50 20 10 -Machinists 55 25 20 50 20 15 -Fishing Industry 55 25 12.5 50 20 10 -Specified Vessels ANY 25 25 ANY 20 20 -Education Health ANY --- ANY _ -Sportsmen ANY 20 20 ANY 20 20 -Artists ANY 20-30 20-30 ANY 20-30 20-30 Note: TYS total years of service required YSS = of total, numbers of years of service required in the corresponding sector appearing in the first column. VL Social Pensions This program accounts for 4 percent of total pensioners and 5 percent of newly granted pensions. Social Pensions are provided to unemployed citizens that are not eligible for labor pensions (e.g., Old-Age, Disability, Survivor) and meet specific conditions. Eligible social pensioners fall into one of the following categories: (i) disabled of categories I and II, including children disabled from childhood, and disabled of category m, (ii) men over 60 or women over 55, (iii) children of a deceased parent or disabled children. The amount of a social pension varies as a percentage of the minimum Old- Age pension by category. The highest pension is given to disabled of category I (200 percent of the minimum Old-Age pension). The amount of Social Pension granted to persons that have reached the normal pension age depends on whether the reasons for not fulfilling the years of service required for an Old-Age pension are deemed "valid" or not. 7 Without a "valid" reason, the social pension is set at 30 percent of the minimum Old-Age pension. This minimum is raised to 50 percent when the reasons are deemed "valid".' One can interpret the social pension given to men and women who reach the normal pension age as a first pillar of flat pension that applies to workers not covered by the regular retirement programs. This pension is implicitly guaranteeing a universal coverage with a flat rate of 30-50 percent of the minimum pension. The problem of course is that as unemployment grows the social pension cost might increase significantly. Although so far unemployment has hit harder younger workers, older workers are likely to feel the blow in the following stages of the labor market adjustment. VIIL Years of Service vs Years of Contribution The eligibility rules of the different programs described above are based on the years of service. These years of service, however, will most likely differ from the actual number of years of contribution, which is more relevant from an actuarial perspective. One reason for this is that the pension law explicitly allows for including in the calculation of years of service non-contributory activities. A selected list of these activities is presented below on the basis of their relevance for equity or financial consideration: (i) Training on higher and secondary education, special educational establishments in vocational schools, postgraduate courses, clinical studies. The 1989 census indicates that about 93 percent of the working age population had secondary school or university level of education. For this large majority, 7-14 years of education can be counted as years of service and therefore be deducted from years of contribution. To the extent that more educated people also have steeper income profiles, this provision will tend to push towards regressiveness; (ii) The period spent by an unemployed mother taking care of her children under 3, is included as years of service. Under the Ukrainian law mothers are allowed to have an unpaid maternity leave of three years after giving birth. This period is counted as years of service. In other words, given an average fertility of 2, mothers can have up to 6 years of service that are non-contributory. Again given the preferential retirement of women this will tend to increase the return on their pension even more compared to that of men. This might bring progressiveness into the system because ' The pension law, however, does not specify what is considered a "valid" reason, leaving room for arbitrariness. 8 women usually earn less than their male counterparts, and poor families tend to have larger families; (iii) The time spent taking care of an invalid of Category I is counted as years of service. A generous eligibility rule for disability could therefore affect the vesting of other retirement programs. The time spent taking care of a disabled child under 16 of age, or a pensioner who according to the doctor is in need of constant care is considered for pension eligibility; (iv) Period of residence of wives of military personnel in localities where there was no possibility of employment in their fields. The years that are considered as service for pension eligibility cannot exceed 10 years; (v) For underground workers, who are already entitled to a preferential Old-Age pension, periods of temporary disability are counted as years of service; (vi) For world war II veterans, the time spent in military service is doubled for computing the years of service. The time spent in nazi concentration camps, ghetto and other forced confinement within war period is increased threefold. Work in Leningrad within the period of siege during World War II is included into service length threefold. Time spent as resident in Leningrad during this period is increased twofold; (vii) Work in hospitals for lepers and plague-carriers, establishments for aids and treatment of people infected by aids in pathological anatomic and reanimation departments of medical establishments is included twofold in the service length. VIII. Pension Calculation A. The Old-Age Pension An individual's Old-Age pension is calculated in five steps. First, an average salary is calculated. Second, a "replaceable salary" is calculated by using conversion factors to scale down the average wage. Third, a replacement rate that depends on years of service is applied to the replaceable salary to determine an unconstrained pension. Fourth, the unconstrained pension calculated in this manner is subject to a floor and a ceiling. Fifth, a supplemental pension benefit is available if the pension plus other earnings fall below a given threshold. 9 Step 1: average salarv. The average salary for purposes of the pension calculation is either the average of the two preceding years or the average of any 60 consecutive months. The pension law does not indicate whether a reflation factor is applied to old salary months before the average is taken. It should be noted that the average is taken for the two years preceding pension calculation rather than retirement. People drawing Old-Age pensions without retiring can pull their average salaries up even after. The first average wage adjustment is made at least three years after the original pension was granted. Subsequent adjustments are allowed in intervals of two years. The number of years of service is also adjusted at the time the average wage is recalculated.2 Step 2: replaceable salary. If the average salary determined in Step 1 is less than four minimum wages, then the replaceable salary is equal to the average salary. If the average salary is greater than four minimum wages, then the replaceable salary is a smaller and smaller percentage of the average salary, as shown in Table 1.5 below. As the table shows, the replaceable salary is capped at 6.9 minimum wages. Table 1.5: Average Wage and Replaceable Salary Average Wage Incremental Average Conversion Replaceable Salary (number of Conversion Factor Coefficient (number of minimum wages) (percent) (percent) minimum wages) up to 4 100.00 100.00 equal to AW 5 85.00 97.00 4.85 6 70.00 92.50 5.55 7 55.00 87.14 6.10 8 40.00 81.25 6.50 9 25.00 75.00 6.75 10 15.00 69.00 6.90 > 10 0 6.9/ AW 6.90 Step 3: statutory replacement rate. The statutory (unconstrained) replacement rate is an increasing function of years of service. The minimum length of service required of men and women is 25 and 20 years, respectively. For these lengths of service, the statutory replacement rate is 55 percent. The rate increases one percentage point for each year of service over the minimum, up to 85 percent for men with 55 years of service and women with 50 years of service. 2 This adjustment does not apply, however, for disability pensions, which are based on years of service completed at the time disability occurs. 10 Step 4: application of floor and ceiling. The minimum pension until October, 1994, was twice the legislated minimum wage, which was in fact the operative minimum wage. Starting in November 1994, the minimum pension relative to the minimum wage was lowered from 2 to 1.5.3 The maximum pension is triple the minimum pension, except for people on "List I", for whom the minimum pension is four times the minimum pension. List I is shown in an attachment. Table 1.6 summarizes the interaction of the replaceable salary (Step 2 above), the statutory replacement rate (Step 3 above) and the floor and ceiling. The pension level by wage bracket, expressed in minimum wages, is shown upper panel of the table. A worker with a pre-retirement average salary of two minimum wages would receive a pension equal to 1.5 the minimum wage regardless of the statutory replacement rate. This occurs because the minimum pension relative to the minimum wage is set at that level. Given the progressive conversion coefficients shown above, the actual maximum pension becomes 4.5 minimum wages. It is important to note that an increase in the maximum pension relative to the minimum, say from 3 to 4, would not have any effect on the pension structure. If the overall average salary were used to compute the benefit, this would raise the maximum pension from 6 to 8 minimum wages. However, the conversion coefficients presented earlier limit the replaceable salary itself to 6.9 minimum wages. The lower panel of Table 1.6 shows the effective replacement rate, which is defined as the ratio of the pension level to the average wage. The effective replacement rate varies between 150 percent for a worker earning the minimum wage to 25 percent for a worker earning 15 minimum wages and retiring with the minimum years of service. It should be noted that over time the effective replacement rate will also be affected by the degree of indexation of benefits. Subindexation will result in lowering the effective replacement rate. In other words, the pension levels shown should be interpreted as the one that apply for the initial benefit, the level of the pension over its duration will depend on the degree of indexation, which does not follow any well-specified rule. 3 In November, 1994 the minimum pension was raised from KBVl20,000 to KBV360,000, while the minimum wage was raised from KBV60,000 to KBV240,000. 11 Step 5: supplementary pension benefits. In June 1993, a means-tested social assistance scheme was put in place for pensioners and low-income persons.4 Under this scheme, a Social Assistance Intervention Line (SAIL) was set at KBVI97,000 as of December 1993. If the per-capita income of a pensioner's household is lower than the SAIl, a top up is provided to fill the gap. The financing of this scheme falls under the responsibility of the Pension Fund and local authority budgets. This scheme unduly burdens the pension fund with a social assistance activity. Table 1.6: Old-Age Pension Level and Effective Replacement Rates (pension level in minimum wages replacement rates in %) Wage SRR=55 SRR=60 SRR=65 SRR=70 SRR=75 SRR=85 Bracket I lMW 1.5 1.5MW 1.5MW 1.5MW 1.5MW 1.5MW 2MW 1.5MW 1.5MW 1.5MW 1.5MW 1.5MW 1.5MW 4MW 2.2MW 2.4MW 2.6MW 2.8MW 3.OMW 3.4MW 6MW 3.0MW 3.3MW 3.6MW 3.9MW 4.2MW 4.5MW 8MW 3.6MWf 3.9MW 4.2MW 4.5MW 4.55MW 4.5MW 0OMW 3.8MW 4. 1MW 4.5MW 4.5MW 4.5MW 4.5MW 12MW 3.8MW 4.1MW 4.5MW 4.5MW 4.5MW 4.5MW 15MW 3.8MW 4.1MW 4.5MW 4.5MW 4.5MW 4.5MW llMW 150.0 150.0 150.0 150.0 150.0 150.0 2MW 75.0 75.0 75.0 75.0 75.0 85.0 4MW 55.0 60.0 65.0 70.0 75.0 85.0 6MW 50.9 55.5 60.1 64.7 69.4 75.0 8MW 44.7 48.7 52.8 56.2 56.2 56.2 lOMW 37.9 41.4 44.8 45.0 45.0 45.0 12MW 31.6 34.5 37.4 37.5 37.5 37.5 15MW 25.3 27.6 29. 30.0 30.0 30.0 Note: Using a minimum wage of KB240,000 and minimum pension of KB360,000. 4 This was established by a Decree of the Prime Minister (No .394) and a Joint Executive Order of the Ministries of Social Protection, Labor and Finance. 12 B. The Disability Pension The procedure for determining the replaceable salary is exactly the same as for the Old-Age pension. In other words there is ceiling of 6.9 minimum wages on the replaceable salary that arises because of the coefficients used to convert the average salary into the replaceable salary. The pension level is also subject to a minimum. This minimum, however, varies across types of disabilities and across sectors from 100 to 300 percent of the Old-Age minimum pension. It should be noted that the pension law fixes the minimum disability pension as a function of the social pension, which in turn is set as a function of the minimum Old-Age pension. Pensions are also subject to a maximum pension, which is set a 3 minimum Old-Age pensions but raised to 4 for full time underground workers. The pension levels expressed in minimum wages, and the effective replacement rates are shown in Table 1.7. Table 1.7: Disability Pension and Effective Replacement Rates (pension in minimum wages, replacement rates in %) Wage Regular . Milita Duties Undergr und GI GIU GI GI GIl Gi GI GIl GMl 1MW 3MW 1.5 1.5 4.5 4.5 2.2 3.0 1.5 1.5 2MW 3 1.5 1.5 4.5 4.5 2.2 3.0 1.5 1.5 4MW 3 2.4 1.6 4.5 4.5 2.2 3.0 2.4 1.6 6MW 3.9 3.3 2.2 4.5 4.5 2.2 3.9 3.3 2.2 8MW 4.5 3.9 2.6 4.5 4.5 2.6 4.5 3.9 2.6 O1MW 4.5 4.1 2.8 4.5 4.5 2.8 4.8 4.1 2.8 12MfW 4.5 4.1 2.8 4.5 4.5 2.8 4.8 4.1 2.8 15MW 4.5 4.1 2.8 4.5 4.5 2.8 4.8 4.1 2.8 1MW 300 150 150 450 450 225 300 150 150 2MVV 150 75 75 225 225 112 150 75 75 4MW 75 60 40 112 112 75 75 60 40 6MW 65 55 37 75 75 56 65 55 37 8MW 56 49 32 56 56 32 57 49 32 1OMW 45 41 28 45 45 28 48 41 28 12MW 47 34 23 37 37 23 40 34 23 15MW 30 28 18 30 30 18 32 28 18 13 Examining the pension levels shown above help capture some of the redundancies of the benefit structure. For instance, although the maximum pension for underground workers is set at four minimum wages compared with three for other workers, this does not affect the actual level of the benefit. The reason is that given that the highest statutory replacement rate for disability is 70 percent (group I), and the existence of a ceiling on the replaceable salary of 6.9 minimum wages, it follows that even without imposing any additional constraint on the maximum pension, the highest pension possible would have been 4.8 minimum wages (70%*06.9). It follows that whether the maximum pension is set at 4.5 minimum wages, (3 minimum Old-Age pension), or 6 minimum wages (4 minimum Old-Age pension) is irrelevant. C. Survivor Pension The full rate survivor pension is 30 percent of the replaceable salary of the deceased worker for each survivor. In other words, the pensions generated by a survivor benefit can exceed 100 percent of the replaceable salary in case there are more than three survivors. The survivor benefit is also subject to a minimum pension. This minimum is set at a social pension, which varies with the degree of disability of the survivor. A dependent that is subject to a disability type I receives 200 percent of the minimum Old-Age pension, while other disabled dependents receive the minimum Old-Age pension. As shown in Table 1.8, the highest survivor pension per-dependent is currently at about 3 minimum wages (2 minimum Old-Age pensions) and applies to disabled dependent of type I and children under 18 with a disabled single mother. For aged dependents the survivor pension varies between .75-2 minimum wages. The pensions shown below are the fill rates, which apply in case death is work related or worker met the years of service requirements for death outside work (see Table 1.3). Table 1.8: Survivor Pension and Effective Replacement rates (pension in minimum wages and replacement rates in %) Wage Aged Dependents -Disabled I -Other Disabled Bracket -Children with Disabled -Children <18 l_______________ _ M other l. 1MW .75MW 3.0MW 1.5MW 2MW .75MW 3.0MW 1.5MW 4MW 1.2MW 3.0MW 1.5MW IOMW 2.1MW 3.0MW 2.1MW 15MW 2.lMW 3.0MW 2. 1MW 1MW 75 300 150 2MW 37 150 75 4MW 30 75 37 lOMW 21 30 21 15MW 14 20 . 14 Note: The pension level is for each dependent. 14 VIIM Summary The Ukrainian pension system works as a pay-as-you-go with very complicated eligibility criteria and benefit formulae. In addition to the usual retirement programs (e.g., Old-Age, Disability, and Survivor), the so called Social Pension works as a universal flat pension. The cost of this program is likely to increase rapidly as unemployment grows. The unnecessary complexity of the system arises mainly from the existence of minimum and maximum pensions, which make many other parameters (e.g., coefficient applied to the salary to obtain the replaceable salary) redundant. Furthermore, because indexation does not follow a well-specified rule, even the relevant parameters entering the benefit formulae affect more the initial pension level than its value over its duration. Adding to the complexity of the system is the use of means-tested supplementary benefits, which burdens the pension fund with social assistance programs. A hidden cost to the system comes from the fact that the years of service that determine eligibility conditions include many non-contributory years, such as student years. In addition to putting a financial burden on the system, counting student years for the vesting period makes the system more regressive. * ~~~~15 CHAPTER II FISCAL SUSTAINABILfTY OF THE PENSION SYSTEM L Introduction Ukraine is confronted with two major economic challenges. The first one is to correct its macroeconomic imbalances. In 1993, monthly inflation reached 39 percent and the consolidated budget deficit --including extrabudgetary accounts- had a deficit of 11.8 percent of GDP. Double digit inflation and deficit relative to GDP persisted during most of 1994. The second challenge is that of moving from a command to a market economy. It is estimated that about 78 percent of the labor force is employed by state-owned enterprises (1990). The policies that bring about these changes must be taken into account when assessing the pension fund's fiscal outlook. ]EL The Macroeconomic Context A. Pension Fund and Consolidated Fiscal Deficit Ukraine's fiscal accounts are summarized in Table 2.1. One can see that the pension fund is an important component in these accounts averaging 25 percent of total revenues and 16 percent of total expenditures (1991-93). Prior to 1993, the pension fund, unlike the rest of the public sector, was running a surplus. This is to say that the pension system cannot be blamed for the original fiscal imbalances that triggered inflation. Nevertheless, in 1993 the pension fund operated in a deficit of 1 percent of GDP, hence contributing for roughly one-tenth of the monetary financing of the fiscal gap.' A further deterioration in the pension fund's fiscal accounts could undermine or prolong the fiscal adjustment that is at the cornerstone of the stabilization plan. 'Following the 1993 liquidity crisis, the pension fund was brought under the umbrella of the Ministry of Finance. 16 Table 2.1: The Pension Fund in the Consolidated Fiscal Accounts (in percentage of GDP) _1991 1992 1993 I. Pension Fund 1.1 Revenues 10.4 12.4 8.0 1.2 Expenditures 9.6 8.8 9.0 1.3 Balance +0.8 +3.6 -1.0 II. Budget and Other Accounts 2.1 Revenues 27.7 32.4 33.8 2.2 Expenditures 42.2 64.7 44.6 2.3 Balance -14.5 -32.3 -10.8 m Consolidated Balance -13.7 -28.7 -11.8 Memo: (monthly inflation) (5.5 - (25.6) (38.6) Source: World Bank RSMX data. Note: Other Accounts include extrabudgetaiy accounts such as the employment fund. B. Inflation and Pension Fund Deficits There are at least two ways of looking at the interaction between inflation and the pension fund's fiscal accounts. The first one, which we alluded to earlier, is related to the pension fimd's contribution to the overall fiscal gap that is financed through monetary creation. There is, however, another linkage between inflation and the pension accounts. This linkage relates to the way inflation affects real benefits and contributions. As far as contributions are concerned, this comes from the existence of collection lags, which in high inflation lead to an erosion of revenues collected (i.e., Tanzi-Olivera effect). According to the Pension Fund officials, Ukraine has reduced the official collection lag by about a week in 1993.2 In Table 2.2, we present estimates of losses arising from collection lags for different levels of inflation. With an inflation rate of 40 percent, roughly the 1993 average, and a collection of one week, the loss would be about 7.5 percent. Between 1991 and 1993, the average contribution fell by 38 percent, of which 5 percent is attributed to the Tanzi-Olivera effect, and the rest to the combined effect of a falling real wage and compliance. As Ukraine moves to a single digit monthly inflation rate, the Tanzi-Olivera effect will become insignificant. It is also important to look at the impact of inflation on real benefits. The reason is that in many countries inflation has been used to erode the real values of pensions. Brazil and 2 Before the change inter-regional transfers could take up to a week and intra-regional transfers up to 3 days. Under the new rules, transfers to the pension fund are made within a day of settlement regardless of the origin. 17 Venezuela are two cases in hand.3 Table 2.3 presents estimates of the reduction in real pensions under different price indexing rules and inflation rates. For simplification, all rules assume that the original real pension value is restored after a certain time. Pensions are still eroded during the interval separating the periodic adjustments. If the real value of the pension were restored every quarter while average inflation were running at 40 percent, the reduction in real outlays would be 26 percent. In the context of the stabilization plan, it is important to realize that a drastic fall in inflation while keeping the pension promises unchanged would lead to a corresponding increase in real expenditures. In Ukraine, indexation of real pensions is done in an ad-hoc way. In 1993, for instance, benefits were multiplied by a factor of 2.9 in September and 3 in December. In general, pension indexation lag inflation. Between 1991 and 1993, the average pension in real terms fell by 36 percent.4 Table 2.2: Revenue Erosion and Inflation: The Tanzi-Olivera Effect (in percentage of before erosion revenues) Monthly Lag= 3 days Lag=5 days Lag= 7days lag=15 days Inflation 40 3.3 5.4 7.5 15.5 25 2.2 3.6 5.1 10.6 20 1.8 3.0 4.1 8.7 15 1.4 2.3 3.2 6.7 10 0.9 1.6 2.2 4.6 5 0.5 0.8 1.1 2.4 1 0.0 0.2 0.2 0.5 Source: Staff Estimates 3 In Brazil, for instance, the indexation mechanism used before the Cardoso Plan (June 1994) entailed eroding pensions by about 20 percent a year. 4 It is misleading to use the Pension Fund's fiscal position in one month, which is the way the original data are presented (January 1), to assess the degree of erosion of real pensions over the year. The figure presented here is based on a yearly average rather than end-of-period. 18 Table 2.3: Erosion of Real Pensions and Indexation Rules (in percentage of before erosion pension) Monthly Bimonthly Quarterly Semestrial Inflation 40 14 26 49 25 10 19 39 20 8 16 33 15 7 12- 27 10 5 9 20 5 2 5 11 1 0 1 2 Source: Staff Estimates C. The Stabilization Program Under the IMF supported stabilization plan, the deficit of the consolidated state budget, including directed credits and the Pension Fund, will be limited to 10 percent of the expected GDP in 1994, and the ratio of deficit to GDP will be reduced by half in 1995. This would be consistent with the objective of reducing the monthly inflation rate to less than 5 percent by end 1995. The Government also intends to limit the growth of the wage bill paid by state enterprises to no more than 80 percent of the programmed rate of inflation in the fourth quarter of 1994. The stabilization program does not envisage --at least in its first stage-- the use of an exchange rate anchor. Such anchor is typical of adjustment programs in Latin American countries such as Argettina and more recently Brazil. The adjustment process in Latin America seems to suggests that the convergence to World inflation rates is often a long process.5 It took Argentina three years and a fixed exchange rate to reduce inflation to international levels. Mexico without an exchange rate anchor has not yet achieved convergence with the US rate after about 7 years of fiscal and monetary austerity. It should be noted that for Ukraine the use of a fixed exchange rate is not envisaged by the stabilization plan because of the low level of reserves.6 5 Kiguel and Leviatian (1992), based on recent Latin American experience point out that the "process of restoring price stability has been longer and more costly than in classical cases" (i.e., Poland, Austria, and Germany in the 1920s). 6 In 1993 Ukrainian foreign reserves were estimated at $193 million, only 18 percent of the monetary base (using the parallel exchange rate) and 10 months of imports. In Brazil the level of reserves at the time the exchange rate peg was introduced (May 1994) was $41 billion or 18 months of imports. In Argentina, reserves at the time the fixed exchange rate was introduced 19 IIL Transitional Issues in the Labor Market A. Labor Taxation The Financing of the Ukrainian pension system is based on a payroll tax of 33.56 percent, of which 1 percent is paid by workers and the rest by employers. By way of comparison, the average payroll tax for pensions is 16.3 percent in OECD countries and 25.5 percent in Eastern Europe and Former Soviet Union. In addition, Ukrainian employers contribute 3 percent of wages to the Employment fund, 4.44 percent to the Social Insurance Fund, and 12 percent to the Chernobyl Fund. The total contributions on wages, ignoring the personal income tax, is thus 53 percent. There is empirical evidence showing that in the OECD countries, high payroll tax rates and associated transfers resulted in lower competitiveness.7 This is to say that for Ukraine the high penalty on labor has implications not only for labor allocation but also for competitiveness. One measure of this penalty is the marginal effective on labor (METL), which is defined as --for an additional unit of labor-- the difference between the gross labor cost to the employer and the consumption available to the worker. As such, the METL takes into account both direct and indirect taxes. The METL in Ukraine is shown below for different tax brackets and after taking into account the family benefits paid by firms. The progressiveness of the METL in Ukraine is a direct result of the progressive tax schedule for the personal income tax appearing in the last column of Table 2.4. This progressiveness is partially offset by flat benefits such as family benefits, which therefore push the gross labor cost higher for low income workers.8 (Mars 1991) was US$5.4 billion or 1.5 times higher than central bank liabilities. The level of reserves relative to base money is an indication of the currency board's vulnerability to a run. 7 See Alesino and Perotti: "The Welfare State and Competitiveness", NBER No 4810, July 1994. 8 Although the family benefit itself is progressive because family size is often positively correlated with poverty, from a tax standpoint it creates an additional disincentive for employing poor workers. 20 Table 2.4: Ukraine: Marginal Effective Taxes on Labor (in percentage) Wage Bracket METL Personal (minimum wage) Income Tax lMW 49.7 0.0 5MW 49.5 8.0 10MW 49.5 9.0 15MW 52.0 13.3 20MW 53.0 15.0 50MW 71.7 44.0 Source: Staff Estimates Table 2.5: A Cross-Country Comparison of Marginal Effective Taxes on Labor (1987-93) Country MEFL Ukraine* 50 LAC Peru 65 Venezuela 44 Uruguay 48 Costa Rica 49 Mexico 58 Brazil 60 Argentina 63 OECD 46 Austria 42 France 52 Germany 41 Luxembourg 42 Sweden 63 U.K 43 USA 35 Source: Mission Estimates for Ukraine, World Bank (1994) report No.1 1943-PE for Latin America, and OECD (1986) for the rest. Note: *: the Ukrainian rate applies to the wage bracket from I to 10 minimum wages. The OECD average is based on a larger sample of countries than the one shown in the table. Although Mexico has recently joined the OECD, it was not a member when these estimates were done. 21 A cross country comparison of METL is shown in Table 2.5. The Ukrainian rate is higher than the average prevailing in OECD countries. France and Sweden are the only two OECD countries in the sample with a higher rate than Ukraine. In Latin America, Argentina has a tax rate 13 percentage points higher than that of Ukraine. It should be noted, however, that both Peru and Argentina have privatized their pension system, so that the payroll tax becomes closer to a deferred compensation than to a tax. Moreover, Argentina recently reduced the payroll tax rate itself.9 B. Recent labor market developments and future trends:"' Countries in transition such as Ukraine have seen and will continue to see drastic changes in the labor market with far reaching implications for their pension systems. Among these is an initial decline in employment and output. Between 1990 and 1993, real GDP in Ukraine declined by a cumulative 32 percent, while employment declined by 1.5 million or 5 percent of the labor force (1993). So far, this fall in employment has disproportionately affected the working-age population. For instance, the number of pensioners in employment increased by 100,000 (1990-93). A probable corollary to the employment decline is a growing size of the informal sector. Corroborating this is the finding by the Ukrainian National Employment Service that 40 percent of youths in metropolitan areas have been operating in the informal sector. Another important development is the growing importance of unpaid leave and short-term working. During the first quarter of 1994, short-term working and unpaid leave accounted to 6 percent and 15 percent, respectively, of workers. By the fourth quarter of 1994, short-term working had increased to 16 percent, and unpaid leave to 23.6 percent. Overall, while the registered unemployment rate is as low as 0.3 percent, hidden unemployment is estimated at 25- 40 percent. It is important to quantify the impact of the expected increase in open unemployment on the pension fund's fiscal posture. C. Lessons from Other Transition Economies The decline in employment and output observed in Central and Eastern Europe has been drastic. In Bulgaria, Hungary and Poland employment fell by more than 25 percent between 1989 and 1991, and open unemployment rates reached 14 percent in 1993. In the republics of the former Soviet Union the labor shakeout is still to come; unemployment rates are still below 2 percent. Slovenia provides an example of what major labor market changes to expect in transitional economies. Orazem and Vodopivec (1994) document these changes by focusing on 9 As of January 1, 1994, Argentina reduced its employee payroll contribution to social security from 16 to 11 percent. 10 This section draws from a 1994 ILO report "The Ukrainian Challenge: Reforming Labor Market and Social Policy." 22 the following labor market characteristics: returns to education; returns to job experience; wage inequality. Their findings based on earning profiles for Slovenia can be summarized as follows: (i) average returns to years of education rose dramatically during the transition"1; (ii) returns to the most experienced rose relative to those with least experience. For workers reaching the pensionable age, relative wage increased very rapidly12; (iii) the variance of wage income increased, and gap between rich and the poorest has increased. By 1991, wages earned by the tenth percentile were 56 percent of the wage earned by those at the tenth percentile in 1987. Most of these changes are likely to take place in all economies in transition and have far reaching implications for the pension system. For instance, the significant rise in the return to education and experience is likely to have regressive impact on the pension system because of the short time period used to compute the average salary. The cost of a minimum wage provision could very well increase as a result of the growing inequality in wage income --even if the minimum pension relative to the average wage is kept unchanged. It is important to keep these changes in mind when reforning the pension system of a transition economy. For Ukraine, the speed and magnitude of labor restructuring depends on many factors, including the success of the privatization program, and the degree of labor mobility. Such mobility will be necessary to enable the labor shift away from agriculture towards services. Today in Ukraine about 21 percent of the labor force is employed in agriculture (1993), compared with 9 percent for upper middle-income countries. 13 The ILO estimates that about 4 million Ukrainian workers will have to be shifted away from the agricultural sector. The extent to which the wage dispersion observed in Slovenia will take place in Ukraine will depend on many factors, including the actual removal of subsidies to loss making state owned enterprises, and labor regulations. "Higher gains for workers with 4 years of University training lower gains for workers with vocational training. 12 This implies that the social cost of a low retirement age, as measured by the opportunity cost (i.e., the value of the marginal product of labor at retirement age) is likely to increase during the transition. 13 Labor share in agriculture is 6.6 percent in France, 2.3 percent in the U.K, and 2.9 percent in the U.S. 23 IV. Baseline Financial Projections of the Pension Fund A. Presenting the Baseline Scenarios: Transition Period: It is useful to distinguish between the transition and post- transition periods. The transition period is the one in which stabilization takes place, as it is a -prerequisite for all structural reforms. We assume that by the end of the transition, the monthly inflation would have reach a single digit level. The expected rise in unemployment is also an integral part of the adjustment that take place during the early stages of the transition. The privatization program is also expected to be undertaken during the first period. 14 The different ways of characterizing the transition hinge mainly on the speed at which inflation is brought under control, the magnitude of initial labor shedding and economic contraction. We assume that the transition period will continue up to the year 2000. Post-Transition: The post-transition period goes from the year 2001 up to the last year of the projections (2030). During this period, the average unemployment rate is assumed to be at its natural rate. Basically, the different assumptions about the post transition period will depend on the long run GDP growth and natural rate of unemployment. We consider two baseline scenarios that are summarized in Table 2.6. Baseline Scenario I: It is designed to capture a less optimistic outcome that one could associate with a delayed or partial reform program. During the transition inflation declines but at a lower pace than envisaged by the stabilization program. The targeted monthly rate of 5 percent is reached two years later than anticipated by the stabilization program (1997). During the transition, the observed real GDP contraction in the past two years continues until 1997 and the upturn starts in 1999 with a modest 1 percent. The unemployment rate during the transitional period rises at a peak of 25 percent in 1997 before declining 18 percent by the year 2000. The post-transition era is characterized by a modest real GDP growth of 3 percent, the long run unemployment rate is set at 10 percent, which is the estimated natural rate in Western Europe (see OECD jobs study)."5 Baseline Scenario II: This optimistic scenario assumes that during the transition period, the monthly inflation rate falls to 5 percent in 1995 and to 1 percent by the year 2000. The unemployment rate during the transition peaks at 15 percent in 1995 and falls to 10 percent by 1997. Real GDP growth starts in 1996 reaching 5 percent by the year 2000. As far as the post-transition period is concerned, this scenario envisages a long run unemployment rate that is similar to the United States'natural rate (6 percent) and 4 percentage points below the long run 14 The World Bank Rehabilitation loan envisages the privatization of 90 percent of all small enterprises by end 1995. 15 The natural rate of unemployment is defined as the threshold rate that cannot be lowered by an upturn in economic activity. 24 rate under the pessimistic scenario. The long term GDP growth is assumed to be 6 percent, or twice higher than the one associated with the pessimistic scenario. There are a number of assumptions and model specifications that are common to both scenarios. First, we assume a Cobb-Douglas production function with perfect competition and profit maximization behavior. As the result, real wage growth rate is equal to the difference between output growth and labor growth. Second, the minimum pension relative to the average wage is kept at 20 percent. One could do some sensitivity analysis to quantify the impact of increasing the minimum pension relative to the average wage to its 1990-91 level (31 percent). Pension benefits are assumed to be fully indexed to wage inflation, as stipulated by Article 97 of the pension law. 16 Table 2.6: Key Macroeconomic Assumptions of Baseline Scenarios (in percentage) SCENARIOS TRANSMON PERIOD )_|_POST 1994 1995 1996 1997 1998 1999 2000 2001-30 CASE I Inflation 20 10 8 5 4 2 2 0.5 Unemployment 14 18 20 25 23 20 18 10 GDP growth -21 -5 -4 -3 0 1 2 3 CASE II Iflation 20 5 4 3 2 1 1 0.5 Unemployment 14 15 13 12 11 10 9 6 GDP growth -21 0 1 2 3 4 5 6 Note: Inflation is a monthly Rate, unemployment and GDP growth rates are annual. B. Projection Results Baseline Scenario I: The results are presented in Table 2.7. The first important observation is that the deficit rises during the transition period from 2.6 percent of GDP in 1996 to 4.0 percent of GDP by the year 2000. It is helpful to look at the components of the cost rate to understand the causes of the growing deficit. As shown in the upper panel of the table, the system dependency rate (i.e., ratio of pensioners to contributors) rises from 72 percent in 1995 and to 89 percent in 1997. This is the direct result of the assumed increase in the unemployment rate, which reaches its peak rate of 25 percent in 1997. The population dependency rate (i.e., population over 16 The Article states that indexation should be based the "monetary income of the population" but cannot be less than 2 percentage points below the covered wage inflation. 25 60 to that in the 20-59 age bracket) itself does not change very much during the transition period. The average benefit relative to the average wage (i.e., benefit ratio) also rises during the transition from 42 percent in 1995 to 47 percent by the year 2000. This is attributed mainly to the fact that pensions are fully indexed to wages, and the minimum pension is set at 20 percent of the average wage. In 1992-93 the minimum pension relative to the average wage was reduced to 15 percent, compared with 31 percent in 1990-91. Another important observation is that the cost of pensions attributed to workers who retired before 1994 still accounts for-over 50 percent of the total cost until the end of the transition (see lower panel of the table). This is to say that policies that affect the number of new pensioners (e.g., increasing the pension age) or the average value of their pension (e.g., decreasing the replacement rate) are unlikely to prevent deficits from emerging during the transition period. A central question to the transition period is the extent to which the projected pension fund deficits could undermine the overall fiscal adjustment. In this respect, it is important to note that the rising unemployment, that partially explains the pension fund imbalance during the transition, will also lead to a corresponding fiscal stress in the unemployment fund, as registered unemployed increase from its current level of .03 percent of the labor force. In 1993, the unemployment fund spent .01 percent of GDP in unemployment compensations, or 38 percent of its revenues, to support 84 thousand registered unemployed. If all the estimated unemployed in 1997 registered, the cost would jump to .60 percent of estimated GDP. Assuming that revenues from the unemployment fund relative to GDP stay constant, and indexing the average unemployment compensation to the average wage, the unemployment fund deficit would reach .58 percent of GDP in 1997. The combined deficit of the unemployment fund and the pension fund would reach 4.4 percent of GDP in 1997. Recall that under the stabilization plan the target deficit of the consolidated public sector is 5 percent of GDP in 1995. Clearly, the bleak fiscal prospects of the pension and unemployment funds during the transition period are a serious challenge to reaching the stabilization plan's fiscal deficit targets. 26 TABLE 2.7: LONG TERM FINANCIAL PROJECTIONS OF THE UKRAINIAN PENSION SYSTEM (BASELINE SCENARIO I) 1995 1996 1997 1998 1999 2000 2005 2010 2015 2020 2025 2030 I COST RATE ANALYSIS 1.1 Dependency Ratio 0.717 0.786 0.886 0.903 0.874 0.856 0.759 0.726 0.696 0.676 0.661 0.636 1.2 Benefit Ratio 0.416 0.430 0.442 0.453 0.460 0.467 0.494 0.515 0.534 0.548 0.558 0.560 1.3 Cost Rate 0.298 0.338 0.392 0.410 0.402 0.339 0.375 0.374 0.372 0.371 0.369 0.356 II Expenditures % of GDP 8.09 8.97 10.00 10.51 10.37 10.36 9.72 9.69 9.63 9.61 9.56 9.24 2.1 Benefits 7.92 8.78 9.78 10.28 10.14 10.13 9.51 9.48 9.42 9.40 9.35 9.03 -Old Age 6.69 7.46 8.35 8.82 8.67 8.66 8.08 8.01 7.93 7.90 7.86 7.50 -Disability 0.81 0.86 0.93 0.95 0.94 0.93 0.82 0.74 0.66 0.59 0.51 0.49 -Survivor 0.28 0.30 0.32 0.33 0.33 0.33 0.38 0.47 0.55 0.61 0.65 0.70 -Service Pension 0.03 0.04 0.04 0.05 0.05 0.05 0.06 0.06 0.06 0.06 0.05 0.05 -Social Pension 0.10 0.12 0.14 0.14 0.15 0.16 0.18 0.20 0.23 0.25 0.27 0.29 2.2 Administrative Cost 0.18 0.20 0.22 0.23 0.23 0.23 0.21 0.21 0.21 0.21 0.21 0.20 III Contributions 6.48 6.37 6.19 6.24 6.31 6.34 6.37 6.37 6.37 6.37 6.37 6.37 3.1 Employer 6.29 6.18 6.00 6.05 6.13 6.16 6.18 6.18 6.18 6.18 6.18 6.18 3.2Employee 0.19 0.19 0.18 0.19 0.19 0.19 0.19 0.19 0.19 0.19 0.19 0.19 IV Surplus/Deficit -1.61 -2.60 -3.81 -4.27 -4.05 -4.01 -3.36 -3.32 -3.26 -3.24 -3.19 -2.86 Memo Items Population Dependency 0.40 0.40 0.40 0.40 0.40 0.42 0.40 0.43 0.46 0.50 0.50 0.50 Cost for pre 1994 retirees 6.03 5.85 5.82 5.48 5.07 4.71 3.10 2.13 1.29 0.68 0.26 0.05 27 Looking at the post-transition period, one notes that the pension fund deficits persist until the year 2030, reaching a peak of 3.2 percent of GDP during the year 2020. During the post-transition year, the system dependency rate declines since unemployment falls from 25 percent in 1997 to 10 percent in 2001-30. Although the population dependency rate rises somewhat from 42 percent in the year 2000 to 50 percent in 2020-30, this is more than offset by the decline of the unemployment rate towards its natural rate. By contrast, the average pension relative to the average wage continues to rise during the post-transition era from 47 percent in the year 2000 to 56 percent in the year 2030. If the average pension of workers who retired prior to 1994 were equal to the average pension of new pensioners who retired in 1994 and after, then the benefit ratio would remain constant. The reason is that pensions are indexed to wages. However, with a minimum pension relative to the average wage at 20 percent, and full wage indexation, new pensioners gradually push up the overall average pension for the stock of retirees offsetting the observed decline in the system dependency rate. This highlights the need to push down the long run system dependency rate further by increasing the pension age today. Baseline Scenario II: The results are presented in Table 2.8, and do not change the fundamental message conveyed by the first scenario. The pension fund is consistently in deficit during as well as after the transition. Of course, the deficits are lower than the ones associated with the first scenario. For instance, a fiscal deficit of 3.8 percent of GDP would emerge in 1988 under the preceding scenario, in this case the corresponding deficit would be 3 percent of GDP. The system dependency rate still rises during the transition, though by less than before as the unemployment rate is now lower. For every single year, the average wage relative to the average pension is the same as the one under the preceding scenario since the indexing rules, the minimum pension, and the replacement rate are unaltered. During the post-transition period, the deficit, compared to its end of transition level (3.13 percent of GDP), remains more or less at the same level before declining modestly to to 2.6 percent of GDP in the year 2030. In short, despite more favorable macroeconomic assumptions, the unsustainability of the pension system still prevails. This is corroborated by the estimates presented below regarding the pension system's implicit debt. -28 TABLE 2.8: LONG TERM FINANCIAL PROJECTIONS OF THE UKRAINIAN PENSION SYSTEM (BASELINE SCENARIO II) 1995 1996 1997 1998 1999 2000 2005 2010 2015 2020 2025 2030 I COST RATE ANALYSIS 1.1 Dependency Ratio 0.691 0.721 0.748 0.774 0.770 0.764 0.723 0.691 0.663 0.644 0.630 0.606 1.2 Benefit Ratio 0.416 0.430 0.442 0.453 0.460 0.467 0.494 0.515 0.534 0.548 0.558 0.560 1.3 Cost Rate 0.287 0.310 0.331 0.351 0.354 0.357 0.357 0.356 0.354 0.353 0.351 0.339 II Expenditures % of GDP 8.04 8.71 9.30 9.86 9.94 10.03 10.07 10.04 9.97 9.95 9.90 9.56 2.1 Benefits 7.86 8.52 9.09 9.64 9.72 9.81 9.85 9.82 9.76 9.73 9.69 9.35 -Old Age 6.64 7.24 7.76 8.27 8.31 8.38 8.36 8.29 8.21 8.18 8.15 7.77 -Disability 0.80 0.84 0.87 0.89 0.90 0.90 0.85 0.77 0.68 0.61 0.53 0.51 -Survivor 0.28 0.29 0.30 0.31 0.32 0.32 0.39 0.48 0.57 0.63 0.68 0.73 -Service Pension 0.03 0.04 0.04 0.05 0.05 0.05 0.06 0.06 0.06 0.06 0.05 0.05 -Social Pension 0.10 0.11 0.13 0.14 0.15 0.15 0.19 0.21 0.24 0.26 0.28 0.30 2.2 Administrative Cost 0.18 0.19 0.20 0.22 0.22 0.22 0.22 0.22 0.22 0.22 0.22 0.21 III Contributions 6.77 6.83 6.85 6.87 6.89 6.90 6.92 6.93 6.93 6.93 6.93 6.93 3.1 Employer 6.57 6.62 6.64 6.67 6.69 6.70 6.72 6.72 6.72 6.72 6.72 6.72 3.2 Employee 0.20 0.20 0.20 0.20 0.21 0.21 0.21 0.21 0.21 0.21 0.21 0.21 IV Surplus/Deficit -1.26 -1.88 -2.45 -2.99 -3.05 -3.13 -3.15 -3.11 -3.05 -3.03 -2.97 -2.63 Memo Items Population Dependency 0.40 0.40 0.40 0.40 0.40 0.42 0.40 0.43 0.46 0.50 0.50 0.50 Cost for pre 1994 retirees 5.99 5.68 5.41 5.14 4.86 4.56 3.21 2.21 1.33 0.70 0.27 0.06 29 V. The Pension Fund's Implicit Debt Another way of examining the pension fund's fiscal stress is to estimate the value of its implicit debt. This debt has two components. The first component is the present value of pension payments up to the point when the last pensioner dies. This component is therefore based only on the current stock of pensioners. The second component captures the promises made to current workers based on their work history (i.e., years of contribution and wages). We present below a cross-country comparison of the implicit debt for both components using a discount rate of 4 percent (see Table 2.9). For Ukraine the total implicit debt is 217 percent of GDP, of which 138 percent is attributed to workers. Women account for more than half of the pension debt. The higher debt for women is mainly due to their higher life expectancy, which affects both the initial stock and the average duration of pensions. In most countries, the implicit debt of the pension system exceeds the explicit debt. Ukraine is no exception. The external debt of Ukraine is about 21.4 percent of GDP, by far lower than the pension implicit debt. 17 One can see that the liabilities of the pension system are higher in Ukraine than in Hungary. Hungary has a cost rate of about 34 percent (1993), of which 66 percent is due to the system dependency rate and 38 percent to the benefit ratio. The benefit cost rate is higher in UJkraine (49 percent) than in Hungary because of its higher benefit ratio (55 percent). It is interesting to note that Ukraine has the second highest implicit debt of the sample after Italy (242 percent). Clearly, Ukraine will be facing serious fiscal crisis if no corrective measures are taken. It should be noted that a measure such as a reduction of the retirement age will not affect the value of the implicit debt attributed to pensioners but only that of contributors. 17 Using the figure of $US 4301.6 Million for the stock of debt outstanding and disbursed for 1993. GDP at domestic prices is converted in dollars using the average official nominal exchange rate (7629.0) rather than the PPP exchange rate. 30 Table 2.9: Implicit Debt of the Pension Fund (percent of 1993 GDP) Pensioners Contributors Total* Ukraine 79 138 217 Men 29 64 93 Women 50 74 124 Hungary 102 70 172 United States 42 70 89 Japan 51 112 144 Germany 55 102 157 France 77 139 216 Italy 94 140 242 United Kingdom 58 98 156 Canada 42 79 121 Source: Mission Estimates for Ukraine, Palacios for Hungary and OECD for the rest. Notes: #; For OECD countries the figures are expressed as a percentage of 1990 GDP. *: Total liabilities exclude (from the sum of pensioners and contributors liabilities) pension fund's assets. These assets exist only for the US (89 percent of GDP) and Japan (18 percent of GDP). The figures for Ukraine are based on US mortality Tables. For both Ukraine and Hungary linear accrual rates are used for contributors (i.e., each worker gets a pension proportional to its years of service). For all countries, the discount rate used is 4 percent. 31 CHAPTER m TRANSMON AND POST-TRANSMTON REFORM OPTIONS L Introduction Reforming a pension system is generally a difficult process because of vested interests, expectations, and the sharp contrast between the short-term political horizon and the inherent long-term dynamics of pension schemes. This process is made more difficult for economies in transition, such as Ukraine, because of the pressing need to achieve a fiscal adjustment and insulate the more vulnerable groups from the transition's social cost. Even after successfully meeting these short term objectives sooner or later the need to reform the pension system in a more profound way will arise. The purpose of this Chapter is to examine the financial implications of different pension reform options for Ukraine. The rest of the Chapter is organized as follows. Section II discusses the short and long-term reform objectives. Section III illustrates the trade-off between pension level and coverage, hence highlighting the need to increase the statutory pension age. Section IV lays down the changes that would be required in order to keep the pay-as-you-go system sustainable in the long run while meeting the constraints of the transition period. Finally, Section V explores the implications of adopting a multipillar pension system. I. Reform Objectives A. Meeting the Challenge of Transition Meeting the Overall Fiscal Targets: As emphasized in the previous Chapter, fiscal adjustment is at the cornerstone of the Ukrainian stabilization program. Because pension expenditures are an important component of the overall fiscal accounts, they must be kept at a level that is consistent with the fiscal targets. Therefore the question that must be addressed is as follows: given the magnitude of the needed fiscal adjustment and expected labor shedding, what part of the adjustment has to he done within the pension fund accounts and which part in the rest of the public sector? The Benchmark used in this report is that the pension fund must remain at least in balance during the next five years to meet the stabilization objectives. In the absence of any corrective measure, however, the baseline scenarios (see Chapter II) indicate that fiscal balance would not be achieved during or after the transition. It should be pointed out that ensuring fiscal balance can be achieved by different measures to which the Government might attach different priorities. Protecting the most vulnerable groups: It is important both for the political viability of the reforms as well from an equity standpoint. In spite of popular perceptions, it is not likely that pensioners as a group are worse off than the rest, but certain limited categories of pensioners might need extra protection (e.g., pensioners living alone). This issue, however, is not 32 directly addressed in this report. This is best achieved by a direct transfer that does not affect the basic eligibility rules of the pension system. Establishing the foundations for longer term reforms: Some measures with only long- term effects, such as increasing the pension age, would have to be implemented now. As mentioned in Chapter I, the current pension system should be streamlined by eliminating the redundancies of the benefit formulae and separating the social insurance from the social assistance programs. There is also a need to reduce labor taxation. Ukraine seems to be moving in that direction by reducing the payroll contribution channeled to the unemployment fund. A further reduction in labor taxation is likely to be required in order to compete in world markets. The exact timing of the reduction will have to be determined in the broader context of tax, social assistance, and social insurance reform. B. Post-Transition Objectives Adopt a Better Pension System: Pension systems are supposed to fulfill at least three objectives: redistribution, insurance, and saving. There is growing acceptance that a single pillar pay-as-you-go system is not a satisfactory way of achieving these three objectives. Moreover, pay-as-you -go schemes are more often than not associated with significant distortions in labor markets due to the combination of high labor taxes and loose linkages between benefits and contributions. Nevertheless, in most countries, these systems are likely to prevail for a long time even after the introduction of a defined-contribution scheme (e.g., Peru, Colombia, Argentina). To some extent this reflects the political economy of reforming pension systems. It should be pointed out, however, that measures designed to bring the promises made by a pay-as- you system to a realistic level are an indispensable ingredient of any type of pension reform. In general, Ukraine could follow one of the following three options: (i) make the current public pension viable in the long term; (ii) introduce a defined contribution scheme as a complement to a reformed public pension pillar; (iii) substitute the public pension with a capitalization scheme. The following sections are devoted to the first two options. m. Pension Level and Coverage: What is Affordable A. Sharing the Burden between current and future pensioners The first question we ask is whether it is fiscally viable to achieve a balanced pension budget during the transition without scaling back the level of benefits of pensioners who have already started drawing a pension (as opposed to those that will start drawing a pension for the first time in the following years).' One should recognize, however, that the distinction 1 This is not a mere mathematical exercise, some countries facing fiscal stress in their pension system specifically commit themselves not to change the eligibility or pension level of existing retirees. The party that won the latest presidential election in Uruguay is a case in hand. 33 between new and old pensioners poses serious equity problems (vertical equity, as well as horizontal ). From a legal standpoint this merits attention because depending on the degree of contractual obligation towards pensioners the Government might be denied the legal or political possibility of affecting the pensions granted to current retirees. There are also practical difficulties because changes in the pension level of new beneficiaries can be affected by the replacement rate, minimum pension, and pension age, while for those already drawing pension this has to be done in an ad-hoc way or simply by reducing the degree of inflation adjustment. Often Government rely on ad-hoc inflation adjustments, which permit significant erosion in the level of benefits to avoid political fall-out. However, this is s short-sighted strategy that usually fails to fool pensioners for very long and obscures the real trade-off in social spending. We therefore use the model projections presented in the earlier Chapter to derive the average pension that would be consistent with a balanced budget, while meeting all current and future pension obligations. The results are presented in Table 3.1 under two different assumptions: that only new benefits are altered (upper panel), and all benefits can be reduced (lower panel). It should be pointed out that in practical terms not altering old pensions implies fully indexing them to wage inflation. Table 3.1: Balanced Budget vs Projected Average Pensions (as a percentage of the average wage) BASELINE I 1995-2000 2001-2030 I Only New Pensions Reduced 1.1 Projected 53.8 56.6 1.2 Balanced Budget *11.1 35.4 II. All Pensions Adjusted .2.1 Projected 44.5 53.5 2.2 Balanced Budg 29.8 36.4 Note: Staff Estimates Under the current rules, and with full wage indexation, the pension granted to a typical worker drawing a pension after 1994 would be 54-57 percent of the average wage. For a balanced pension budget to be attained --while keeping the pension level of those already drawing pension unchanged -- the new pensions granted would need to be reduced to roughly 11 percent of the average wage during the transition and to 35 percent of the average wage in the long run. A number of conclusions can be drawn from this exercise. First, the average pension for future retirees that is consistent with a balanced budget during the transition is too low, even far below the 1993 minimum pension (17 percent of the average wage), which applied to only 2.6 percent of new pensioners. In other words, letting only new pensioners bear the cost of transition would put an excessive burden on them. From a policy standpoint one can therefore conclude that a measure such as reducing the replacement rate for new pensions is not sufficient to achieve fiscal balance. During the transition, partial indexation of all benefits is virtually unavoidable. Second, as shown in the lower panel of the Table 3.1, even if both current and future pensioners bear the 34 cost, the average pension consistent with a balanced budget would need to be reduced to 30 percent of the average wage during the transition and 36 percent later. This suggests that unless Ukraine is willing to grant pensions of these low magnitudes some changes that reduce quantity (i.e., the total number of pensioners) are unavoidable. This leads us to the issue of affordable coverage. B. Affordable Coverage If one assumes that the pension budget must be balanced each year, then a trade- off between the average level of pensions and the maximum number of beneficiaries emerges. This trade-off is summarized in Table 3.2, where the total number of beneficiaries is computed as a ratio of the population over a given pension age. The lower panel uses the current statutory retirement age for both men and women. For all pension levels exceeding 50 percent of the average wage, the coverage would fall below 100 percent, corroborating previous findings. One can see that as the average pension is increased the coverage satisfying the budget constraint falls, reaching 62 percent with an average pension of 65 percent of the average wage. Again Ukraine cannot fulfill the promise of granting a meaningful pension to all women over 55 and men over 60 while meeting the budget constraint of the pension fund. The lower panel of Table 3.2 provides the same information as the upper one, but this time the coverage is assessed relative to the population over 65. Implicitly, the exercise entails examining whether the resources available would permit granting a given pension to the entire population over 65. An average pension of 50 percent of the average wage can indeed be granted to the entire population over 65, since the coverage exceeds 100 percent (114 percent in 2005-30).2 However, as the average pension is increased, the affordable coverage would decrease, reaching 88 percent with an average pension of 65 percent of the average wage. To sum up. a p2ension exceeding 55 p2ercent of the average wage is not compatible with a balanced budget. even if eligibility were restricted to the population over 65: The point of these simulations is to stress the basic dilemma facing the pension system: it cannot promise generous benefits to all potential claimants. To avoid future fiscal imbalances, either benefits have to be reduced, or the number of pensioners cut back, or both. 2 When trying to make a policy recommendation out of this it should kept in mind that a significant fraction of the population below 65 is already drawing a pension. 35 Table 3.2: The Trade-off Between Pension level and Coverage under a Balanced Budget (Maximum pensioners as a percentage of eligible population) Pension = 50% of Pension= 55% of Pension =60% of Pension = 65% of Average age Average W e Average age Average age 1995 to 2005 to 1995 to 2005 1995 2005 1995 2005 2000 2030 2000 2030 2000 2030 2000 2030 I CASE I 1.1 Men over 60 103.8 111.9 94.3 101.8 86.5 93.3 79.8 86.1 1.2 Women over 55 69.4 74.3 63.1 67.6 57.9 61.9 53.4 57.2 1.3 Both 80.3 80.4 73.0 73.1 66.9 67.0 61.7 61.9 II. CASE II 2.1 Men over 65 159.4 153.4 144.9 139.5 132.8 127.9 122.6 118.0 2.2 Women over 65 128.1 102.2 116.4 92.9 106.7 85.2 98.5 78.6 2.3 Both 128.1 114.5 116.4 104.0 106.7 95.4 98.5 88.0 Source: Staff Estimates Note: Case I captures the coverage associated with current statutory retirement ages. 36 IV. Options for Keeping the Pay-As-You-Go System A. Increasing the Contribution Rate is Not Desirable It was emphasized in the preceding Chapter that the tax on labor was excessively high in Ukraine. But because an increase in the contribution rate is still often mentioned in Ukraine as a solution to the pension problem, it is worth showing the extent to which even an extreme increase in the contribution rate to 50 percent would not solve the long term problems of the system. This holds even under the most optimistic scenario that higher taxes do not affect compliance. In the more plausible case of a resulting increase in evasion, increasing the contribution rate could make things worst by lowering actual revenues collected. Table 3.3 shows the impact of increasing the contribution rate to 50 percent under four different assumption about evasion responsiveness to such increase. We define the evasion elasticity as the percentage increase in evasion for each percentage increase in the contribution rate. Under the unlikely assumption that the increase in the contribution rate does not affect evasion (i.e., elasticity of 0), one can see that the pension fund after a brief surplus in 1995 would revert to a deficit as early as 1997, and this imbalance would persist throughout the simulation period. When the evasion elasticity is raised to .5, the previously observed deficits are even bigger. For instance, during the transition the average deficit would increase from .29 percent of GDP (0 elasticity ) to 1.1 percent of GDP (.5 elasticity). It should be noted that the elasticity of 1.42 corresponds to the estimated one for Brazil. With an elasticity similar to that of Brazil, the fiscal gap would average 2.5 percent of GDP during the transition and 2.3 percent in the long term. For both efficiency and financial considerations, an increase in the contribution rate should be avoided. B. Making the Pay-As-You-Go Fiscally Sustainable As mentioned earlier, one option for Ukraine would be to adopt a comprehensive reform package that would restore the long run fiscal balance of the current system. We therefore identify such comprehensive reform package of five components, which is assumed to be implemented in 1996, and simulate its effects. The following changes are considered, after taking into account the main findings regarding the trade-off between coverage and pension level: (i) increase the statutory retirement age of both men and women to 65 (each year, the retirement age would be increased by six months, reaching the target in the year 2005 for men and 2015 for women. In addition to containing the future cost of the old-age program, this would also push a downward pressure on disability benefits. The reason is that all disability benefits at or after the normal retirement age are considered permanent); (ii) reduce the old-age statutory replacement rate from 65 to 50 percent; (iii) reduce the maximum from 3 to 2 times the minimum pension; (iv) indexing pensions to the minimum of wage and price inflation, with partial indexation during the transition. During the transition degree of price indexation is increased gradually from 80 37 percent in 1996 to 100 percent in the year 2000 (by 5 percentage points each year); and (v) reduce the contribution rate to 25 percent after the transition (2000-2030). The results are shown in Table 3.4 using the macroeconomic assumptions of the first baseline scenario. Except for a brief period following the reduction of the contribution rate (2000-2005), the pension system shows consistently a modest surplus, averaging 0.2 percent of GDP during the transition. By way of comparison, under the no-reform scenario, the deficit during the transition would average 3.4 percent of GDP. Engineering such an improvement while reducing significantly payroll taxes would be a remarkable achievement. As the result of the reform package, the cost rate (e.g., ratio of pension expenditures to the wage bill) falls to 16-17 percent in the long run. Again, without any reform the long run cost rate would be around 37 percent. Another important beneficial effect of putting in place a reform that lowers the long term cost rate is that it leaves room for adding a capitalization scheme. This issue is explored in the next Section. 38 Table 3.3: Increasing the Contribution Rate to From 33.56 to 50 Percent (BASED ON MACROECONOMIC SCENARIO I) 1995 1996 1997 1998 1999 2000 2005 2010 2015 2020 2025 2030 I COST RATE ANALYSIS 1.1 Dependency Ratio 0.717 0.786 0.886 0.903 0.874 0.856 0.759 0.726 0.696 0.676 0.661 0.636 1.2 Benefit Ratio 0.416 0.430 0.442 0.453 0.460 0.467 0.494 0.515 0.534 0.548 0.558 0.560 1.3 Cost Rate 0.298 0.338 0.392 0.410 0.402 0.399 0.375 0.374 0.372 0.371 0.369 0.356 II Expenditures%ofGDP 8.09 8.97 10.00 10.51 10.37 10.36 9.72 9.69 9.63 9.61 9.56 9.24 III. A Contribution: Baseline 6.48 6.37 6.19 6.24 6.31 6.34 6.37 6.37 6.37 6.37 6.37 6.37 III.B. Contribution:C = 50%, E = 0 9.65 9.49 9.22 9.30 9.41 9.45 9.49 9.49 9.49 9.49 9.50 9.49 III.C Contributions:C=50%, E= 0.5 8.84 8.69 8.46 8.53 8.63 8.68 8.71 8.72 8.72 8.71 8.72 8.72 III.D Contributions:C = 50%, E = 1.42 7.34 7.23 7.05 7.11 7.21 7.25 7.28 7.29 7.29 7.29 7.29 7.29 III.E Contributions:C=50%, E= 2 6.40 6.31 6.16 6.22 6.32 6.35 6.38 6.39 6.39 6.39 6.39 6.39 IV. A Deficit/Surplus Baseline -1.61 -2.60 -3.81 -4.27 -4.05 -4.01 -3.36 -3.32 -3.26 -3.24 -3.19 -2.86 IV.B Surplus/Deficit:C=50%, E = 0 1.56 0.52 -0.78 -1.21 -0.96 -0.91 -0.24 -0.20 -0.14 -0.21 -0.07 0.26 IV. C. Surplus/Deficit:C=50%, E 0.5 0.75 -0.28 -1.55 -1.98 -1.73 -1.68 -1.01 -0.98 -0.92 -0.90 -0.84 -0.52 IV. D Surplus/Deficit:C=50%, E= 1.42 -0.75 -1.74 -2.95 -3.40 -3.15 -3.11 -2.44 -2.40 -2.34 -2.32 -2.27 -1.95 IV. Surplus/Deficit C=50%, E=2 -1.69 -2.66 -3.84 -4.29 -4.05 -4.01 -3.34 -3.30 -3.24 -3.22 -3.17 -2.85 Source: Staff Estimates 39 TABLE 3.4: MAKING THE PAY-AS-YOU-GO FISCALLY SUSTAINABLE BY IMPLEMENTING A COMPREHENSIVE REFORM PACKAGE (BASED ON MACROECONOMIC SCENARIO 1) 1995 1996 1997 1998 1999 2000 2005 2010 2015 2020 2025 2030 I COST RATE ANALYSIS 1.1 Dependency Ratio 0.717 0.780 0.872 0.882 0.846 0.822 0.676 0.600 0.540 0.528 0.527 0.513 1.2 Benefit Ratio 0.287 0.269 0.260 0.267 0.277 0.287 0.305 0.301 0.304 0.313 0.319 0.320 1.3 Cost Rate 0.206 0.210 0.227 0.235 0.234 0.236 0.206 0.181 0.164 0.165 0.168 0.164 II Expenditures(%ofGDP) 5.59 5.60 5.84 6.11 6.11 6.19 5.42 4.75 4.32 4.35 4.43 4.32 2.1 Benefits 5.47 5.48 5.71 5.97 5.98 6.05 5.30 4.65 4.22 4.26 4.33 4.22 -OldAge 4.61 4.66 4.88 5.12 5.11 5.16 4.44 3.81 3.41 3.47 3.55 3.44 -Disability 0.56 0.53 0.53 0.54 0.55 0.55 0.49 0.44 0.40 0.37 0.34 0.34 -Survivor 0.19 0.19 0.19 0.20 0.21 0.22 0.25 0.27 0.28 0.28 0.29 0.29 -Service Pension 0.02 0.02 0.03 0.03 0.03 0.02 0.01 0.01 0.00 0.00 0.00 0.00 -Social Pension 0.07 0.08 0.08 0.09 0.09 0.10 0.11 0.12 0.13 0.14 0.14 0.15 2.2 Administrative Cost 0.12 0.12 0.13 0.13 0.13 0.14 0.12 0.10 0.09 0.10 0.10 0.09 III Contributions (% of GDP) 6.48 6.40 6.24 6.31 6.39 4.79 4.81 4.82 4.82 4.82 4.82 4.82 3.1 Employer 6.29 6.21 6.05 6.12 6.20 4.60 4.62 4.62 4.63 4.63 4.63 4.63 3.2 Employee 0.19 0.19 0.19 0.19 0.19 0.19 0.19 0.19 0.19 0.19 0.19 0.19 IV Surplus/Deficit 0.89 0.80 0.40 0.20 0.28 -1.40 -0.61 0.06 0.50 0.47 0.39 0.50 Memo Items Population Dependency Ratio 0.40 0.40 0.40 0.40 0.40 0.42 0.40 0.43 0.46 0.50 0.50 0.50 Cost for pre 1994 retirees 3.78 3.23 2.92 2.67 2.48 2.31 1.44 0.87 0.46 0.21 0.07 0.01 NPV = + 4.36 % of GDP Source: Staff Estimates Note: NPV = Present Value of Pension Liabilities Net of Contributions as percentage of 1994 GDP (using a 4 percent discount). 40 V. Introducing a Defined-Contribution Pension Scheme After the Transition After making the necessary adjustments to its public pension system Ukraine could very well consider introducing a defined-contribution scheme as a complement. Essentially, this would entail adopting a multipillar pension system. This report does not deal with the reforms in the financial sector that would be required to have a well-functioning capitalization scheme. Instead, the focus will be mainly on the main characteristics of this scheme and the timing of its implementation. One question that arises is whether a capitalization scheme should be privately or publicly managed. International experience indicates that when such schemes are managed primarily by the state the returns tend to be lower and even negative. For this reason, when introducing such scheme, Ukraine should rely primarily on a competitive and privately managed system. Another important issue has to do with the compulsory or optional characteristic of this scheme. The vast majority of recent reformers in Latin America adopted an optional capitalization scheme.3 In most cases, however, this was more the result of a political compromise than an initial objective of the reform. A rmandatory defined-contribution scheme has the advantage of allowing workers of all income groL.p. to benefit from the potential high investment returns relative to wage growth. As such it also guarantees that the pay-as-you-go system would be reduced to a reasonable scale. With a mandatory multipillar pension scheme, the pension offered by the public scheme could be either earnings-related or flat. Argentina, for instance adopted a flat pension of 27-32 percent of the average wage, depending on the number of years of service.4 In general, a flat pension leads to greater redistribution towards the poorest low-income groups, and making it a function of years of service removes the disincentive to participate after the vesting period. Table 3.5 shows the impact of adopting a mandatory multipillar pension system. The mandatory savings pillar is assumed to be introduced in the year 2000. The contribution to the private pension fund is gradually raised from 2 percent of payroll in the year 2000 to 10 percent by the year 2005 (2 percentage points increase each year). During the same time the 3 The savings pillar is optional in Peru, Argentina, and Colombia. Under the recently adopted reform in Uruguay, the savings pillar is optional for all workers except the very few with a monthly earning above 5000 pesos ($US800). 4 The lower bound applies to workers with 30 years of service and the upper bound to workers with 45 years of service. 41 contribution to the public pillar is reduced to 25 percent in the year 2000 and 17 percent by the year 2004. As the result, the total contribution rate to both pillars would be 27 percent, of which 10 percent going to the savings pillar would become a deferred compensation rather than a payroll tax. Except for the replacement rate, all the reforms to the public pillar outlined in the previous section (e.g., increasing the pension age to 65) are assumed to hold. The replacement rate, which was reduced to 50 percent under the previous reform package is assumed to be further reduced to 30 percent. Again this could be either a flat rate or the average of an earnings-related scheme. It should be kept in mind that with a contribution rate of 10 percent, the savings pillar could provide for an annuity exceeding 70 percent of final salary. Early withdrawals would have to be permitted for workers who would have accumulated enough savings for an annuity above a certain threshold. This would mitigate some the efficiency costs associated with excess mandatory savings. The reduction in the public pillar's replacement rate to 30 percent would be introduced gradually to avoid penalizing the first wave of retirees after the reform who would have contributed very little to the capitalization scheme. This would be done in order to keep the same accrual rate (ie., ratio of replacement rate to years of service) for all workers regardless of whether they retire immediately after the introduction of the fully-funded scheme. The results show that the pension system would have a modest surplus during the first five years. In the year 2000, however, a deficit of 1.4 percent of GDP would emerge reaching a peak of 2 percent of GDP in the year 2005. Subsequently, this deficit would fall as the decline in the replacement rate starts taking effect, turning into a surplus of 0.5 percent of GDP in the year 2030. On a net present value basis, and using a 4 percent discount rate, pension liabilities would exceed contributions by only 8.5 percent of 1994 GDP (1995-2030). In most OECD countries facing future bankruptcies in their pay-as-you-go systems, pension liabilities exceed contributions by about 160-250 percent of GDP.5 The contributions as well as the value of the accumulated pension fund are also shown in Table 3.5. Annual contributions would be around 1.9 percent of GDP and the accumulated fund would reach 157 percent of GDP by the year 2020. One of the ways of financing the fiscal gap emerging after the adoption of a savings pillar would be by issuing bonds to the newly created private pension funds. Given the size of pension savings, these bonds could be easily absorbed by the private pension funds. The main message of this exercise is that, provided the promises of the public pension system are scaled back, Ukraine could meet the challenge of stabilization, reduce labor taxation, and introduce a multipillar pension system with a fully-funded component. 5 Among OECD countries only the U.S have a double digit net present value of liabilities relative to GDP (43 percent). 42 TABLE 3.5: INTRODUCING A MULTIPILLAR PENSION SYSTEM AFIER THE TRANSImTON (BASED ON MACROECONOMIC SCENARIO I) 1995 1996 1997 1998 1999 2000 2005 2010 2015 2020 2025 2030 I COST RATE ANALYSIS 1.lDependencyRatio 0.717 0.780 0.872 0.882 0.846 0.822 0.676 0.600 0.540 0.528 0.527 0.513 1.2 Benefit Ratio 0.287 0.269 0.260 0.267 0.277 0.287 0.296 0.272 0.246 0.225 0.212 0.203 1.3 CostRate 0.206 0.210 0.227 0.235 0.234 0.236 0.200 0.163 0.133 0.119 0.112 0.104 11 Expenditures %ofGDP 5.59 5.60 5.84 6.11 6.11 6.19 5.26 4.30 3.49 3.12 2.95 2.75 2.1 Benefits 5.47 5.48 5.71 5.97 5.98 6.05 5.14 4.20 3.42 3.05 2.88 2.69 -OldAge 4.61 4.66 4.88 5.12 5.11 5.16 4.32 3.47 2.79 2.50 2.37 2.18 -Disability 0.56 0.53 0.53 0.54 0.55 0.55 0.46 0.37 0.29 0.24 0.21 0.20 -Survivor 0.19 0.19 0.19 0.20 0.21 0.22 0.24 0.24 0.23 0.22 0.21 0.20 -Service Pension 0.02 0.02 0.03 0.03 0.03 0.02 0.01 0.01 0.00 0.00 0.00 0.00 -Social Pension 0.07 0.08 0.08 0.09 0.09 0.10 0.11 0.11 0.10 0.10 0.10 0.10 2.2AdministrativeCost 0.12 0.12 0.13 0.13 0.13 0.14 0.11 0.09 0.08 0.07 0.06 0.06 JR Contributions to Public Pension 6.48 6.40 6.24 6.31 6.39 4.79 3.27 3.28 3.28 3.28 3.28 3.28 IV Surplus/Deficit 0.89 0.80 0.40 0.20 0.28 -1.40 -1.99 -1.02 -0.22 0.16 0.33 0.53 V. Savings Pillar 5.1 Contributions(%ofGDP) 0.0 0.0 0.0 0.0 0.0 0.38 1.92 1.93 85173 1.93 1.93 1.93 5.2 Accumtulated Fund (% of 0.0 0.0 0.0 0.0 0.0 0.63 14.34 41.55 157.43 273.79 462.63 GDP) Memo Itemns Population Dependency 0.40 0.40 0.40 0.40 0.40 0.42 0.40 0.43 0.46 0.50 0.50 0.50 Cost forpre 1994 retirees 3.78 3.23 2.92 2.67 2.48 2.31 1.44 0.87 0.46 0.21 0.07 0.01 NPV = -8.51 % of GDP Source: Staff Estimates Note: NPV = Present Value of Pension Liabilities Net of Contributions as peretage of 1994 GDP (using a 4 percent discount). 43