Chapter

Chapter3. Assessing the Reform of the Caja Fiscal

Author(s):
Jeffrey Franks, Randa Sab, Valerie Mercer-Blackman, and Roberto Benelli
Published Date:
September 2005
Share
  • ShareShare
Show Summary Details

A. Background

Paraguay's existing pension system is fragmented into eight independent public institutions. Few workers—less than 10 percent of the active labor force—participate in a pension system; 95 percent of pension coverage is provided by two institutions, the Instituto de Previsión Social, which covers private sector employees, and the caja fiscal, which covers public employees (public administration employees, teachers, university professors, police officers, army officers, and judicial employees).58 These institutions operate as defined benefit pay–as–you–go plans, that is, the pension benefits are financed in principle by the contributions paid by the active workers who participate in the pension plan. In addition to the contributing plans for public employees, the caja fiscal pays noncontributory pension benefits to Chaco War veterans and their survivors.

In recent years, the caja fiscal has generated large deficits, causing serious concerns about its long–term fiscal sustainability. In 2002, the deficit amounted to 1.9 percent of GDP, about 60 percent of which was due to Chaco War pensions. The deficit was somewhat smaller in 2003, 1.5 percent of GDP, because the year–end bonus was not paid as a result of a new law passed by Congress. Retirement ages as low as 40 and high replacement rates, applied on the last wage, contributed to boosting these deficits. Lack of transparency, especially regarding the legitimacy of Chaco War pensions, exacerbated the generosity of the system. Although systematic evidence is not available, anecdotal evidence suggests widespread fraud, possibly on the order of 30 to 40 percent of claims.

The purpose of this chapter is to assess the long–term fiscal implications of the reform of the caja fiscal recently approved by Congress. As part of its reform program, Paraguay's government committed to undertaking a wide–ranging overhaul of the caja fiscal. In particular, the government committed to reform key parameters, such as the contribution rate, the minimum retirement age, and the replacement rates, to ensure the long–term viability of the caja fiscal. Furthermore, as part of its more general anticorruption and transparency campaign, the government also committed to conducting an in–depth review of the registry of the beneficiaries, with the aim of purging it of illegitimate claims. Although we recognize the utmost importance of improving transparency, this chapter focuses on the long–term quantitative impact of the parametric reform of the caja fiscal. As is common in studies that make projections far into the future—our projections extend to 2050—we recommend some caution in interpreting the findings.

Our main finding is that the reform brings about major savings but leaves large unfunded liabilities that will require further reform in the future. The reform addresses explosive dynamics—prior to the reform, the net present value of deficits over 2003–50 reached almost 90 percent of GDP in 2003. However, the reform leaves large unfunded liabilities, in the form of a large net present value of deficits, of about half the prereform scenario, mostly arising from the caja fiscal's main three contributing plans (for teachers, the army, and the police). As a result, ensuring the financial equilibrium of the caja fiscal over the long run will require further reform in the future.

This chapter is organized as follows. Section B describes the main features of the reform approved by Congress relative to the existing regime. Section C discusses our assumptions and methodology for the actuarial calculations. Section D presents our findings. Finally, Section E draws the main policy implications of our study.

B. Reform of the Caja Fiscal

Both the contributing plans and the Chaco War pensions contributed to the large deficit of the caja fiscal in 2003 (Table 3.1). Specifically:

Table 3.1.Caja Fiscal in 2003
ContributionsDisbursementsOperational Balance1
MillionsPercentMillionsPercentDepen-MillionsPercent
ActiveofofPensionofofdencyofof
EmployeesguaraniesGDPRecipients guaraniesGDPratio2guaraniesGDP
Caja Fiscal144,537352,7820.9455,167906,5332.43-553,750-1.48
Contributing plans3144,537352,7820.9432,132659,9771.770.22-307,195-0.82
Public administration32,50268,4980.187,785113,7500.300.24-45,252-0.12
Army13,73830,5650.086,813181,2590.490.50-150,694-0.40
Police15,52435,7180.103,55696,8680.260.23-61,151-0.16
University professors10,96518,1560.0552214,0950.040.054,0610.01
Judicial employees8,54630,1220.0844111,2560.030.0518,8660.05
Teachers63,262169,7230.4513,015242,7490.650.21-73,025-0.20
Chaco War veterans23,035246,5550.66-246,555-0.66
Veterans7,62199,0730.27-99,073-0.27
Survivors15,414147,4820.40-147,482-0.40
Sources: Ministry of Finance; and Fund staff estimates.

Difference between contributions and disbursements.

Ratio between number of pension recipients (including beneficiaries of survivor pensions) and active employees.

Disbursements include survivor pensions.

Sources: Ministry of Finance; and Fund staff estimates.

Difference between contributions and disbursements.

Ratio between number of pension recipients (including beneficiaries of survivor pensions) and active employees.

Disbursements include survivor pensions.

  • The caja fiscal as a whole generated an operational deficit—defined as the difference between annual contributions and annual disbursements—of 1.5 percent of GDP. The six contributing plans and the Chaco War pensions accounted for 55 and 45 percent of this deficit, respectively.59

  • The dependency ratio—defined as the ratio of pension recipients to active employees—for the contributing plans as a whole stood at 0.22, implying that there were, on average, about 4.5 active workers per pension recipient. However, this ratio varied dramatically across plans, ranging from 0.50 for army officers to 0.05 for judicial employees and university professors.

  • The plans for army officers and teachers accounted for one–half and one–fourth of the overall deficit of the contributing plans, respectively. As regards army officers, contributions were especially low, covering only about half of the disbursements.

  • The plans for university professors and judicial employees were the only ones in surplus. However, these are the smallest plans, both in number of active employees and amount of contributions.

  • As regards the Chaco War pensions, disbursements to survivors accounted for almost two–thirds of total disbursements, with disbursements to veterans accounting for the rest.

Congress approved a new law reforming the caja fiscal on December 22, 2003. The new law,60 with administrative regulations issued on January 30, 2004, aims at improving the actuarial balance of the system and at reducing heterogeneity across old–age pension plans. However, Congress retained a few exceptions, most notably for teachers but also for the police and army plans. As regards the six contributing plans of the caja fiscal, the new law modifies the existing regime by (Table 3.2):

  • Increasing the contribution rate from 14 to 16 percent for all pension plans.61

  • Increasing the retirement age to 62, making retirement mandatory at this age, except for police and army officers—for whom the old rules continue to apply—and teachers. Early retirement is allowed at 50 with at least 20 years of service.

  • Broadening in the base wage used to compute the pension benefit by defining it as the average wage in the last five years of service rather than the last wage.62

  • Defining the replacement rate as 20 percent of the base wage plus 2.7 percent for each year of service (previously 93 percent of the base wage).63

  • Eliminating the 13–month bonus (aguinaldo).

  • Limiting pension indexation to consumer price index (CPI) inflation rather than to public wage increases.64

Table 3.2.Comparison Between the Preexisting Law and the New Law Regulating the Caja Fiscal
Old LawNew LawComments
I. General provisionsContribution rate14 percent.16 percent.New rate has to remain effective at least until financial balance is achieved.
Contribution baseBase salary.All taxable remunerations, including base salary, overtime, representation charges, bonuses.The remuneration base excludes family subsidy and health care contributions.
Ordinary retirement age (jubilation obligatoria)Between 40 and 50 years depending on occupation, gender, and years of service.62 years.Retirement is compulsory at age 62. Employees with less than 10 years of service receive 90 percent of their inflation–adjusted contributions. The new rules do not affect police and army officers.
Early retirement age(jubilation)Retirement age can be moved forward through various mechanisms.50 years of age.Early retirement is possible with at least 20 years of service.
Replacement rate93 percent (lower for early retirement). Between 50 and 100 percent depending on years of service, for police and army officers.20 percent plus 2.7 percent per each year of service.The replacement rate reaches its maximum at 100 percent after 40 years of service. In case of early retirement, the replacement rate is multiplied by the the ratio of the age at retirement and 62. The new rules do not affect police and army officers.
Base wage for computing pension benefit (Remuneration base)Last wage.Average wage in last five years of service.Past wages are not adjusted for inflation in the computation of the base wage.
Pension adjustment for inflationBased on wage increase of working cohort.Based on average public wage increase but limited to CPI inflation.The adjustment to Chaco War pensions is determined in the budget law.
Thirteenth–month bonus (aguinaldo)Customarily paid but not mandated by law.Prohibited.The prohibition applies to all beneficiaries, including Chaco War pensions.
II. Special provisions for teachers Ordinary retirement45 or 40 years of age for males and females, respectively, with at least 25 years of service. Female teachers can count up to five children as years of service.At least 28 years of service, or at least 25 years of service at a lower replacement rate. Female teachers can count up to three children as years of service.
Early retirement20 years of service (for causes preventing ordinary retirement).Between 15 and 24 years of service (only for physical or mental incapacity).
Replacement rate93 percent (lower for early retirement).87 percent; 83 percent for at least 25 years of service.
Base wage for computing pension benefitLast wage.Average wage in last five years of service (or 10 years in case of an increase in hours or shifts worked).Past wages are not adjusted for inflation.
Sources: Ministry of Finance; and Fund staff analyses.
Sources: Ministry of Finance; and Fund staff analyses.

The new law retains special provisions for the teachers’ pension plan, most important by allowing ordinary retirement after 25 years of service at a replacement rate of 83 percent. These exceptions were negotiated with teachers’ unions following strong protests from teachers. Furthermore, the new law introduced a transitory regime that allowed teachers with at least 20 years of service at the time of approval of the new law to opt for the old regime until December 31, 2004 (about 6,000 teachers were eligible).

The new provisions for teachers are nonetheless an improvement over the existing situation. First, the new minimum retirement age for teachers is still higher than the age at which many teachers now retire. Second, the new definition of base wage is particularly important for teachers, given the common current practice of teachers working more hours and shifts just prior to retirement. Finally, the law essentially eliminates early retirement for teachers, which remains only as a disability pension.

The new law also retains important exceptions for army and police officers. The new minimum retirement age and the associated replacement rate do not apply to them; they are allowed to retire using the graduated replacement rates under the old law. Importantly, however, the base wage will be computed using the new law, that is, the replacement rate will be applied to the average wage in the last five years of service rather than to the last wage.

Finally, the law aims at generating savings from the Chaco War pensions. In particular, the law reduces survivor pensions to 75 percent of the original pension (from 100 percent), tightens somewhat the eligibility criteria, and abolishes aguinaldo. Furthermore, by separating the administration of Chaco War pensions from the other contributing plans and by requesting that the former be fully financed in the annual budget law, the reform will make explicit the cost to society of Chaco War pensions.

Besides legislative action, the government has started reorganizing the public pension administration and reviewing the rolls of beneficiaries. With assistance from the International Development Bank (IDB), the government recently hired an external consultant, PricewaterhouseCoopers, to help reorganize the pension administration and computerize its databases. On a separate but related front, the government is also completing a census of public employees that will eventually provide a detailed database of contributors to the pension system. In a second stage, the government plans to extend this census to the pension recipients, further deepening the cleanup of the registry of beneficiaries. These efforts to purge the registry of beneficiaries of fraudulent claims—which are believed to be widespread, especially in the administration of Chaco War pensions—should generate large savings.

C. Assumptions and Methodology

To make projections about future contributions and pension disbursements, we need a variety of assumptions on demographic, macroeconomic, and sector–specific variables. This section describes our assumptions, together with the methodology that we use to calculate our actuarial projections. Tables 3.3 and 3.4 summarize the assumptions for the contributing plans and the Chaco War pensions, respectively.

Table 3.3.Assumptions for Actuarial Computations on Contributing Plans
Public administrationArmy
Before reformAfter reformBefore reformAfter reform
Hiring policyNew hires maintain constant shares of active male and female public employees in total population.Same number of new hires as before reform.New hires maintain constant shares of active male and female officers in total population.Same number of new hires as before reform.
Retirement ageShare of each cohort that is retired as in 2003.Starting from 2004, no retirement before age 62; every employee retires at age 62.Share of each cohort that is retired as in 2003.Same as before reform.1
WageAverage wage of active employees.Same as before reform.Average wage of active officers.Same as before reform.
Last wageEstimated from average pension of retired employees using replacement rate of 93 percent.Same as before reform.Average pension of retired officers.Same as before reform.
Wage growthNominal wage grows at inflation rate plus productivity growth rate.Same as before reform.Nominal wage grows at inflation rate plus productivity growth rate.Same as before reform.
Contribution rate14 percent (on 13 monthly wages).16 percent (on 13 monthly wages).14 percent (on 13 monthly wages).16 percent (on 13 monthly wages).
Replacement rate93 percent.93 percent.100 percent.100 percent.
AguinaldoPaid.Not paid.Paid.Not paid.
Base wageLast wage.Average of last wage in current and previous four years.Last wage.Average of last wage in current and previous four years.
Pension indexationGrowth rate of current nominal wage.CPI inflation rate.Growth rate of CPI inflation rate. current nominal wage.
Number of survivor pensionsConstant share of survivor pensions to old–age pensions.Same as before reform.Constant share of survivor pensions to old–age pensions.Same as before reform.
Amount of survivor pensionsAverage survivor pension in 2003 adjusted by wage inflation (with aguinaldo).Average survivor pension in 2003 adjusted by wage inflation (without aguinaldo).Average survivor pension in 2003 adjusted by wage inflation (with aguinaldo).Average survivor pension in 2003 adjusted by wage inflation (without aguinaldo).
Table 3.3Assumptions for Actuarial Computations on Contributing Plans
Public administrationArmy
Before reformAfter reformBefore reformAfter reform
Hiring policyNew hires maintain constant shares of active male and female officers in total population.Same number of new hires as before reform.New hires maintain constant shares of active male and female teachers in population aged 5–19.Same number of new hires as before reform.
Retirement ageShare of each cohort that is retired as in 2003.Same as before reform.Share of each cohort that is retired as in 2003.Starting from 2004, no retirement before age 50; same as in 2003 for other cohorts.
WageAverage wage of active officers.Same as before reform.Average wage of active teachers.Same as before reform.
Last wageAverage pension of retired officers.Same as before reform.Estimated from average pension of retired officers using replacement rate of 93 percent.Same as before reform.Congress approved a new law reforming the caja fisca
Wage growthNominal wage grows at inflation rate plus productivity growth rate.Same as before reform.Nominal wage Same as before reform. grows at inflation rate plus productivity growth rate.
Contribution rate14 percent (on 13 monthly wages).16 percent (on 13 monthly wages).14 percent (on 13 monthly wages).16 percent (on 13 monthly wages).
Replacement rate100 percent.100 percent.93 percent.87 percent.
AguinaldoPaid.Not paid.Paid.Not paid.
Base wageLast wage.Average of last wage in current and previous four years.Last wage.Average of last wage in current and previous four years.
Pension indexationGrowth rate of current nominal wage.CPI inflation rate.Growth rate of CPI inflation rate. current nominal wage.
Number of survivor pensionsConstant share of survivor pensions to old–age pensions.Same as before reform.Constant share of survivor pensions to old–age pensions.Same as before reform.
Amount of survivor pensionsAverage survivor pension in 2003 adjusted by wage inflation (with aguinaldo).Average survivor pension in 2003 adjusted by wage inflation (without aguinaldo).Average survivor pension in 2003 adjusted by wage inflation (with aguinaldo).Average survivor pension in 2003 adjusted by wage inflation (without aguinaldo).
Table 3.3Assumptions for Actuarial Computations on Contributing Plans
Public administrationArmy
Before reformAfter reformBefore reformAfter reform
Hiring policyNew hires maintain constant share of active male and female professors in total population.Same number of new hires as before reform.New hires maintain constant share of active male and female judges in total population.Same number of new hires as before reform.
Retirement ageShare of each cohort that is retired as in 2003.Starting from 2004, in retirement before age 62; same as in after 2003. Every professor retires after age 62. retires after age 75.no Share of each cohort that is retired as 2003Starting from 2004, no retirement before age 62; every employee age 62.
WageAverage wage of active professors.Same as before reform.Average wage of active judges.Same as before reform.
Last wageEstimated from average pension of retired employees using replacement rate of 93 percent.Same as before reform.Estimated from average pension of retired employees using replacement rate of 93 percent.Same as before reform.
Wage growthNominal wage grows at inflation rate plus productivity growth rate.Same as before reform.Nominal wage grows at inflation rate plus productivity growth rate.Same as before reform.
Contribution rate14 percent (on 13 monthly wages).16 percent (on 13 monthly wages).14 percent (on 13 monthly wages).16 percent (on 13 monthly wages).
Replacement rate93 percent.93 percent.93 percent.93 percent.
AguinaldoPaid.Not paid.Paid.Not paid.
Base wageLast wage.Average of last wage in current and previous four years.Last wage.Average of last wage in current and previous four years.
Pension indexationGrowth rate of current nominal wage.CPI inflation rate.Growth rate of current nominal wage.CPI inflation rate.
Number of survivor pensionsConstant share of survivor pensions to old–age pensions.Same as before reform.Constant share of survivor pensions to old–age pensions.Same as before reform.
Amount of survivor pensionsAverage survivor pension in 2003 adjusted by wage inflation (with aguinaldo).Average survivor pension in 2003 adjusted by wage inflation (without aguinaldo).Average survivor pension in 2003 adjusted by wage inflation (with aguinaldo).Average survivor pension in 2003 adjusted by wage inflation (without aguinaldo).
Sources: Ministry of finance; and Fund staff analyses.
Sources: Ministry of finance; and Fund staff analyses.
Table 3.4.Assumptions for Actuarial Computations on Chaco War Pensions
Before ReformAfter Reform
Creation of new survivor pensionsIn case of death of a veteran, a new pension is paid to a female survivor aged 50–54.Same as before reform.
Elimination of veteran pensionsIn case of death only.Same as before reform.
Elimination of survivor pensionsIn case of death only (also for pensions of single daughters).Same as before reform.
Replacement rate for survivor pensions100 percent of veteran pension.75 percent of veteran pension.
AguinaldoPaid.Not paid.
Pension indexationGrowth rate of current nominal wage.CPI inflation rate.
Sources: Ministry of finance; and Fund staff analyses.
Sources: Ministry of finance; and Fund staff analyses.

Demographics

Population projections for Paraguay are obtained from the World Development Indicators (WDI). The WDI provide projections for the future male and female population disaggregated by cohort; each cohort spans five years of age up to the cohort 70–74, except for the oldest cohort, which includes the population older than 75. Since projections are available only at a frequency of five years, we interpolate them to obtain projections for the intermediate years not covered by WDI.

These population projections allow us to compute the time evolution of the death rates by cohort that we employ in our study (Table 3.5). For example, because the cohort aged 25–29 in year 2025 turns into the cohort aged 30–34 in year 2030, we can compute the death rate for this cohort over the period 2025–30. As expected, the death rates decline over time across all the cohorts. However, the decline is more pronounced for the older cohorts, especially for the female population, as Paraguay's population ages.

Table 3.5.Death Rates by Cohort1
200520102020203020402050
Males
20-24-0.1-0.2-0.1-0.1-0.1-0.1
25-29-0.2-0.1-0.1-0.1-0.1-0.1
30-34-0.2-0.2-0.1-0.1-0.1-0.1
35-39-0.2-0.2-0.1-0.2-0.1-0.1
40-44-0.2-0.2-0.3-0.2-0.2-0.1
45-49-0.4-0.5-0.4-0.3-0.3-0.3
50-54-0.8-0.7-0.7-0.5-0.5-0.4
55-59-1.1-1.0-1.0-0.9-0.9-0.7
60-64-2.0-2.0-1.6-1.4-1.3-1.2
65-69-3.0-2.6-2.4-2.2-1.9-1.7
70-74-4.2-4.1-3.7-3.4-3.0-2.6
75+-18.3-15.9-16.1-16.0-13.3-12.9
Females
20-24-0.1-0.1-0.1-0.1-0.1-0.1
25-29-0.1-0.1-0.10.0-0.10.0
30-34-0.1-0.1-0.1-0.1-0.1-0.1
35-39-0.1-0.1-0.1-0.10.0-0.1
40-44-0.1-0.2-0.2-0.1-0.1-0.1
45-49-0.3-0.4-0.2-0.2-0.2-0.1
50-54-0.5-0.4-0.3-0.4-0.3-0.2
55-59-0.5-0.5-0.5-0.5-0.4-0.4
60-64-1.0-1.1-0.9-0.7-0.6-0.5
65-69-2.0-1.7-1.3-1.2-1.0-0.8
70-74-3.2-2.7-2.6-2.1-1.7-1.5
75+-15.9-15.2-12.9-12.7-10.4-9.7
Sources: World Development Indicators; and Fund staff estimates.

The minus sign denotes a positive death rate.

Sources: World Development Indicators; and Fund staff estimates.

The minus sign denotes a positive death rate.

The Ministry of Finance provided us with data on active public employees as of October 2003. These data included, for each category of public employees covered by the caja fiscal, information on monthly wages and contributions by age and gender. Unfortunately, the information on age turned out to be very inaccurate since, for a large number of employees, age was either missing or implausible. As a result, we had to discard those individuals for whom age was either missing or implausible to estimate the age distribution of the active employees in 2003. Subsequently, we allocated the individuals with missing or inaccurate age information to the various cohorts using the estimated age distribution.65 For each category of public employees, we computed annual wages and contributions in 2003 by multiplying the monthly contributions by 13 (since contributions are paid on the aguinaldo as well).

The Ministry of Finance also provided us with data on pension recipients as of October 2003. The data consisted of all the individual records of pension recipients in the caja fiscal. A complete record includes gender, type of pension (i.e., ordinary or early retirement), age at retirement, current age, and monthly pension but, unfortunately, no information on the length of service at the time of retirement. As for the age information on active employees, we found many gaps in the individual records. As for the active employees, therefore, we first had to compute the age distribution of current pension recipients on the subsample of recipients for whom age information was available and then use the resulting distribution to allocate the individuals with missing or inaccurate age information to the various cohorts.66 For each category of pension recipients, we computed annual pension disbursements in 2003 by multiplying the monthly disbursements by 13 (to account for the payment of aguinaldo).

Retirement decisions

Future pension disbursements depend on future retirement decisions, which in turn depend on the incentives and requirements contained in the pension law.

Modeling retirement decisions is therefore critical for the evolution of pension disbursements. To model these decisions, we made the following assumptions:

  • The individual attitudes toward retirement can be described entirely by the share of each cohort of the total population of active and retired employees who, at each point in time, are retired. For example, if, in 2003, a share of 20 percent of all the active and retired teachers aged 50–54 are retired, this means that individual preferences toward work and retirement are such that 20 percent of the teachers of this age choose to retire, given their years of service, the minimum service requirements, the replacement rates, and the other provisions in the pension law. Table 3.6 shows the share of pension recipients by cohort across the various pension plans in 2003 before the reform.

  • The individual attitudes toward retirement remain constant over time. That is, holding the pension law and future wages unchanged, the same share of each cohort of the total population of active and retired employees is retired during each year.

  • Given the previous two assumptions, we computed the number of employees who retire every year to keep the shares of pension recipients within each cohort constant over time.

  • The new law affects the shares of employees who retire during each year if it raises the minimum retirement age or introduces mandatory retirement. Accordingly, we adjust the share of employees who retire within each cohort if the pension reform affects the retirement age. For example, if retirement becomes mandatory at 62, as is the case for public administration employees, then 60 percent of the cohort aged 60–64 (corresponding to the individuals who are 62 or older in this cohort) and 100 percent of the older cohorts retire after the reform; for all the younger cohorts, this share is zero. On the other hand, if a cohort is not affected by the increase in the retirement age, as is the case for army and police officers, then the same share of employees retire before and after the reform—that is, we ignore the effect that other changes introduced by the reform (such as in the replacement rates) may have on retirement decisions. Table 3.6 summarizes the effect of the reform on the shares of retired employees for all the contributing plans.

  • Data limitations prevent us from considering the years of service as a determinant of retirement.

Table 3.6.Assumptions on Retirement Age by Cohort and Pension Plan(Percent of retirees per cohort)
Public AdministrationArmyPolice
Before reformAfter reformBefore reformAfter reformBefore reformAfter reform
Males
30-340.00.00.60.60.00.0
35-390.30.01.31.30.70.7
40-441.00.04.84.82.82.8
45-494.00.011.911.96.56.5
50-5410.50.034.534.524.524.5
55-5924.70.067.667.649.749.7
60-6442.460.080.280.273.673.6
65-6960.6100.084.284.296.096.0
70-7470.6100.087.087.097.897.8
75+81.3100.092.192.197.697.6
Females
30-340.00.00.00.00.40.4
35-390.20.00.40.41.21.2
40-441.30.00.00.03.23.2
45-495.90.00.80.84.84.8
50-5416.00.05.55.511.911.9
55-5934.30.026.226.223.523.5
60-6452.960.047.147.148.848.8
65-6970.3100.048.648.680.780.7
70-7478.9100.064.964.9100.0100.0
75+89.4100.083.083.0100.0100.0
TeachersUniversity ProfessorsJudicial Employees
Before reformAfter reformBefore reformAfter reformBefore reformAfter reform
Males
30-340.00.00.00.00.00.0
35-390.00.00.00.00.00.0
40-440.10.00.00.00.00.0
45-492.90.01.20.00.50.0
50-5420.020.03.30.010.90.0
55-5933.433.46.90.020.90.0
60-6447.047.018.611.247.660.0
65-6947.047.029.129.154.4100.0
70-7452.852.834.034.067.9100.0
75+100.0100.060.2100.096.3100.0
Females
30-340.00.00.00.00.00.0
35-390.00.00.00.00.00.0
40-441.50.00.00.00.00.0
45-4915.60.03.50.01.70.0
50-5446.646.68.40.04.40.0
55-5970.470.410.50.025.60.0
60-6484.184.113.17.831.960.0
65-6989.289.216.416.419.4100.0
70-7490.490.430.830.859.0100.0
75+100.0100.055.8100.0100.0100.0
Sources: Ministry of Finance; and Fund staff estimates.
Sources: Ministry of Finance; and Fund staff estimates.

Contributions

Future contributions depend on the future evolution of the population of active employees. Determining the evolution of the active population is particularly difficult in the present context, since it depends not only on the future demographic developments at the country level but also on the government's future policies regarding the size of the public sector. In this study, we sought to keep a neutral stance by making the following assumptions:

  • As regards teachers, in the prereform scenario, every year the government hires new male and female 24–year–old teachers to keep the ratios of active male and female teachers to the population aged 5–19 constant at the 2003 level.

  • As regards the other plans, in the prereform scenario, every year the government hires new male and female 24–year–old employees to keep the ratios of active male and female employees to the whole population constant at the 2003 level.67

  • The same number of new employees is hired every year under the postreform scenario as in the prereform scenario. This assumption implies that although the total population of active and retired employees is equal in the pre–and postreform scenarios, its composition in terms of active and retired employees is different (if retirement decisions are different).

  • Finally, the population of active employees declines only as a result of death or retirement of active employees.

We computed the contributions to each pension plan during a given year by multiplying the number of active employees, the average annual wage, and the contribution rate. The contribution rate is 14 percent before the reform and 16 percent after the reform. The evolution of future wages is described below.

Old–age pensions

Future wages matter for future contributions and for the future pensions of newly retired employees. In particular, an employee's wage history at the time of retirement determines the base wage, which is used in turn to compute the initial pension. To compute the evolution of average wages and wage histories, we made the following simplifying assumptions:

  • For each category of workers, we computed a measure of the last (nominal) wage of a newly retired employee in 2003 as the average of the pensions paid in 2003, adjusted for the relevant replacement rate. Two features of the prereform pension plans—that the base wage coincides with the last wage and that old pensions are indexed to the current wage growth—imply that this yields an accurate yet straightforward measure of the last wage of new pension–recipient employees in 2003.

  • We assumed that average and last wages grow at an annual rate equal to the sum of CPI inflation and productivity growth. The latter is defined as the difference between real GDP growth and population growth. This assumption implies that we keep the slope of the wage structure constant over the horizon of our study.

The base wage is a key variable for the evolution of pension disbursements. Together with the replacement rates, it determines the initial pension for new pension recipients. In the prereform scenario, we compute the base wage as the last wage; in the postreform scenario, we compute it as the average of the last nominal wages in the current and the previous four years.68 It is worth noting that the new law does not require that past wages be adjusted for inflation when computing the base wage, although old pensions are indexed to CPI inflation.

The new law modifies the replacement rates. In our study, we accommodated for these changes as follows:

  • As regards teachers, we used a replacement rate of 87 percent in the postreform scenario, down from 93 percent in the prereform scenario—we thus ignored the possibility that teachers can retire early at the replacement rate of 83 percent.

  • As regards public administration employees, university professors, and judicial employees, the new law makes the replacement rate dependent on the years of service at the time of retirement. Since we lacked information on years of service, we used the same replacement rate of 93 percent in the pre–and postreform scenarios.

  • As regards army and police officers, the new law does not change the replacement rates (which vary as a function of years of service). Since the replacement rate can rise to 100 percent (after 30 years of service), we assumed conservatively the highest replacement rate of 100 percent in the pre–and postreform scenarios.

We computed the annual pension disbursements for each pension plan as the sum of new pensions and the pensions of the pensioners surviving from the previous year. This required the following steps:

  • We computed the total amount of new pensions by multiplying the number of new pension recipients, the monthly amount of a new pension, and the number of months a pension is paid during the year.

  • We computed the total amount of old pensions by multiplying the number of existing pensioners who survive (using the cohort–specific death rates in Table 3.5), the adjusted monthly pension from the previous year, and the number of months a pension is paid.

  • We adjusted the nominal value of pensions by the growth rate of wages for active employees in the prereform scenario and by CPI inflation in the postreform scenario.

Survivor and Chaco War pensions

Data limitations prevented us from modeling the evolution of survivor pensions in detail. Thus, we made the following simplifying assumptions:

  • The ratio of survivor pensions to old–age pensions remains constant over time at its value in 2003.

  • Nominal survivor pensions grow at the same rate as old–age pensions.

There are two categories of Chaco War pensions, those paid to the war veterans themselves and those paid to their survivors. As for the data on active employees and old–age pension recipients, the data exhibited many gaps that forced us to estimate the age distributions of the various categories on subsamples of the population of pensioners. To model the evolution over time of these two categories of pensions, we made the following assumptions:

  • We assumed that the number of veteran pensions declines only as a result of death. Therefore, in our calculations we ignored the effect that the ongoing efforts to purge the registry of beneficiaries may have on the number of pensions.

  • We assumed that, when a veteran dies, a new survivor pension is created; in particular, the survivor is female, aged 50–54.

  • We assumed that the number of survivor pensions declines only as a result of death. As for the veteran pensions, we ignored the effect of purging the registry of beneficiaries.

  • We assumed a replacement rate for new survivor pensions of 100 in the prereform scenario and 75 percent in the postreform scenario.

  • We assumed that pensions are indexed to the growth of nominal wages for public administration employees in the prereform scenario and to CPI inflation in the postadministration scenario.

Thirteen–month bonus (aguinaldo)

Aguinaldo to recipients of old–age, survivor, and Chaco War pensions is paid in the prereform scenario but not in the postreform scenario. This implies that the monthly pension is paid 13 times in the prereform scenario and 12 times in the postreform scenario.

Macroeconomic assumptions

Finally, we needed a few macroeconomic assumptions. These are summarized in Table 3.7 for selected years. In the long run, real GDP is assumed to grow at an annual rate of 3.4 percent, and inflation is assumed to be stable at 6 percent. The real interest rate is assumed constant at 5 percent; the assumption on the real interest rate is critical for determining the net present value of deficits of a pension plan but does not affect the size of the operational deficits. Finally, we computed productivity growth as the difference between GDP growth and population growth.

Table 3.7.Paraguay: Macroeconomic Assumptions
2004200520102020203020402050
GDP at constant prices (percent change)2.73.23.43.43.43.43.4
CPI inflation8.06.06.06.06.06.06.0
Real interest rate (percent)5.05.05.05.05.05.05.0
Productivity growth10.30.81.42.02.22.42.7
Source: Central Bank of Paraguay; and Fund staff estimates.

Difference between GDP growth and population growth

Source: Central Bank of Paraguay; and Fund staff estimates.

Difference between GDP growth and population growth

D. Findings

Prereform scenario

In the prereform scenario, the balance between active workers and pension recipients worsens dramatically over 2003–50 (Figure 3.1). For the contributing plans as a whole, the dependency ratio more than doubles over this period, rising from 0.22 to 0.55. The increase in this ratio affects all the contributing plans, but it is particularly steep for the largest plan (for teachers), where the ratio quadruples to 0.8 between 2003 and 2050. With regard to Chaco War pensions, although the number of veteran pensions quickly dwindles to zero as old veterans die, the number of survivor pensions falls more gradually, an implication of the relatively young age of current survivors and of our conservative assumption that a female survivor aged 50–54 replaces each dying veteran.

Figure 3.1.Demographic Evolutions of Pension Plans in the Prereform Scenario

Sources: Ministry of Finance; and Fund staff estimates.

Worsening dependency ratios generate rapidly growing operational deficits (Table 3.8. In the prereform scenario, the caja fiscal is on an explosive path, with the net present value of deficits in 2003–50 approaching 90 percent of initial GDP,69 as the operational deficit in percent of GDP grows from 1.5 percent to 4.2 percent in 2050. The contributing plans account for almost all of these cumulated deficits (about 78 percent of initial GDP), with the Chaco War pensions accounting for the rest (just below 7 percent of initial GDP).

Table 3.8.Actuarial Balances in the Prereform Scenario(Percent of GDP)
20032004200520102020203020402050NPV1
Caja Fiscal-1.5-1.5-1.4-1.6-2.3-3-3.7-4.2-86.5
Contributions0.91.01.01.01.01.01.11.134.8
Disbursements2.42.42.42.63.34.04.75.4121.4
Contributing plans2-0.8-0.9-0.9-1.2-2-2.9-3.6-4.2-77.8
Contributions0.91.01.01.01.01.01.11.134.8
Disbursements1.81.81.82.23.03.94.65.3112.7
Public administration-0.1-0.1-0.1-0.1-0.2-0.3-0.3-0.4-8.1
Contributions0.20.20.20.20.20.20.30.37.5
Disbursements0.30.30.30.30.40.50.60.715.6
Army-0.4-0.4-0.4-0.4-0.5-0.6-0.7-0.9-18.1
Contributions0.10.10.10.10.10.10.10.13.4
Disbursements0.50.50.50.50.60.70.81.021.5
Police-0.2-0.2-0.2-0.2-0.4-0.6-0.7-0.9-15.5
Contributions0.10.10.10.10.10.10.10.13.9
Disbursements0.30.30.30.30.50.70.91.119.4
University professors0.00.00.00.00.0-0.1-0.1-0.1-1.3
Contributions0.00.00.00.10.10.10.10.12.0
Disbursements0.00.00.00.10.10.10.10.23.3
Judicial employees0.10.10.10.10.00.0-0.1-0.1-0.2
Contributions0.10.10.10.10.10.10.10.13.3
Disbursements0.00.00.00.00.10.10.20.23.5
Teachers-0.2-0.2-0.3-0.5-1-1.4-1.6-1.7-34.7
Contributions0.50.50.50.50.40.40.40.414.7
Disbursements0.70.70.70.91.41.82.02.249.4
Chaco War pensions-0.7-0.6-0.6-0.4-0.2-0.2-0.1-0.1-8.7
Veterans0.30.20.20.10.00.00.00.01.4
Survivors0.40.40.40.30.20.20.10.17.3
Sources: Ministry of Finance; and Fund staff estimates.

NPV represents net present values of balances, contributions, and disbursements between 2003 and 2050 at an annual real interest rate of 5 percent.

Disbursements for contributing plans include old–age and survivors pensions.

Sources: Ministry of Finance; and Fund staff estimates.

NPV represents net present values of balances, contributions, and disbursements between 2003 and 2050 at an annual real interest rate of 5 percent.

Disbursements for contributing plans include old–age and survivors pensions.

Three plans—for teachers, the army, and the police—account for more than three-fourths of the net present value of deficits of the contributing plans. In the prereform scenario, the plan for teachers, in particular, generates a net present value of deficits of almost 35 percent of initial GDP, as its operational deficit in percent of GDP grows from 0.2 percent in 2003 to 1.7 percent in 2050, boosted by a threefold increase in disbursements not matched by an increase in contributions—contributions as a percent of GDP remain more or less flat over the period under consideration.70 The pension plans for the army and the police also experience steep increases in their operational deficits—their joint deficit climbs to almost 1 percent of GDP in 2050—generating net present values of deficits of about 18 and 15 percent of initial GDP, respectively.

Box 3.1.Computing Net Present Values of Future Deficits

In this study, we use net present value calculations extensively. This box presents the analytical underpinnings for such calculations, following Chand and Jaeger (1996).

Let bt,dt,rt, andgtdenote the debt position accumulated at the beginning of year tin percent of GDP in year t, the operational deficit of a given pension plan in year t (defined as the difference between disbursements and contributions) in percent of GDP in year t, the real interest rate in year t, and the growth rate of real GDP in year t. Then, the debt position of the pension plan evolves as

bt+1 = Rt Bt + dt,

where the capitalization factor Rt is defined as Rt(1+rt)/(1+gt).Rt ≡ (1+rt)/(1+gt) This equation implies that the debt position in percent of GDP at the beginning of year t+N is equal to

From this equation, the net present value of the deficits incurred in year t through year t+N—1 (in percent of GDP in the initial year t) is given by

If the initial debt position is zero,bt = 0, then bt+N coincides with the net present values of future deficits using (approximately) the interest rates rt - gt, rt+1 - gt+1, etc., that is, the difference between the real interest rates and the rates of growth of GDP; this is the notion of net present value that we use in the main text. If the initial debt position were different from zero, then it would have to be added to the net present value of future deficits.

Despite the old age of veterans, Chaco War pensions generate deficits with a net present value of almost 9 percent of initial GDP. Offsetting the rapid decline in the number of veteran pensions, survivor pensions continue to generate sizable operational deficits well into the future. This behavior reflects the slow decline in the number of survivor pensions that was pointed out earlier (Figure 3.1).

Postreform scenario

The reform generates major savings but does not achieve financial balance over time (Table 3.9). The reform tackles the prereform explosive dynamics by cutting the net present value of deficits by more than half. However, it leaves large unfunded liabilities, in the form of a net present value of deficits of about 45 percent of initial GDP for the caja fiscal (38 percent of GDP for the contributing plans as a whole). The pension plans for teachers and army officers continue to be the plans generating the highest net present value of deficits between 2003 and 2050 (about 13 percent of initial GDP each).

Table 3.9.Actuarial Balances in the Postreform Scenario(Percent of GDP)
20032004200520102020203020402050NPV1
Caja Fiscal-1.5-1.0-1.0-0.9-1.2-1.5-1.8-1.9-45.0
Contributions0.91.11.11.11.21.21.21.339.9
Disbursements2.42.12.12.02.32.73.03.284.9
Contributing plans2-0.8-0.5-0.5-0.6-1.0-1.4-1.7-1.9-38.2
Contributions0.91.11.11.11.21.21.21.339.9
Disbursements1.81.61.61.72.22.62.93.278.1
Public administration-0.1-0.1-0.1-0.1-0.1-0.1-0.1-0.2-3.2
Contributions0.20.20.20.20.20.30.30.38.5
Disbursements0.30.30.30.30.30.40.40.511.7
Army-0.4-0.3-0.3-0.3-0.3-0.4-0.4-0.5-12.7
Contributions0.10.10.10.10.10.10.10.13.8
Disbursements0.50.40.40.40.40.50.60.716.5
Police-0.2-0.1-0.1-0.2-0.2-0.3-0.4-0.5-9.7
Contributions0.10.10.10.10.10.10.20.24.5
Disbursements0.30.20.20.30.40.50.60.714.1
University professors0.00.00.00.00.00.00.00.0-0.1
Contributions0.00.10.10.10.10.10.10.12.3
Disbursements0.00.00.00.00.10.10.10.12.4
Judicial employees0.10.10.10.10.10.00.0-0.10.7
Contributions0.10.10.10.10.10.10.10.13.7
Disbursements0.00.00.00.00.10.10.20.23.0
Teachers-0.20.00.0-0.1-0.4-0.6-0.6-0.6-13.2
Contributions0.50.50.50.50.50.50.50.517.1
Disbursements0.70.60.60.70.91.11.11.130.4
Chaco War pensions-0.7-0.6-0.5-0.3-0.2-0.10.00.0-6.7
Veterans0.30.20.20.10.00.00.00.01.2
Survivors0.40.40.30.30.20.10.00.05.5
Sources: Ministry of Finance; and Fund staff estimates.

NPV represents net present values of balances, contributions, and disbursements between 2003 and 2050 at an annual real interest rate of 5 percent.

Disbursements for contributing plans include old-age and survivors pensions.

Sources: Ministry of Finance; and Fund staff estimates.

NPV represents net present values of balances, contributions, and disbursements between 2003 and 2050 at an annual real interest rate of 5 percent.

Disbursements for contributing plans include old-age and survivors pensions.

Although the reform generates savings across the board, the financial situation of some plans continues to cause concern. With regard to the impact on the components of the caja fiscal, the following facts deserve consideration:

  • The reform generates savings worth, in net present value terms, about 20 percent of initial GDP for the teachers’ plan. However, this plan, which accounts for almost half of the prereform net present value of deficits, continues to generate a net present value of deficits of more than 13 percent of initial GDP.

  • The reform has a relatively small impact on the plans for police and army officers. The reform reduces the combined net present value of their deficits by about one–third, from more than 30 percent of initial GDP in the prereform scenario to about 20 percent of initial GDP in the postreform scenario.

  • The reform brings the plans for judicial employees and university professors into financial balance. However, these plans are too small to generate a large impact on the overall deficit of the caja fiscal.

  • The reform almost achieves long–term equilibrium in the plan for public administration employees, bringing the net present value of deficits to 3 percent of initial GDP.

  • As regards Chaco War pensions, the reform generates savings of 0.1 percent of GDP in 2004 and of about 2 percent of initial GDP in net present value terms. These savings accrue almost entirely from lowering survivor pensions.

The relative importance of the sources of savings changes over time (Table 3.10). We can summarize our main findings about the sources of savings from reforming the contributing plans as follows:71

Table 3.10.Estimated Savings from Reforming the Contributing Plans 1
20032004200520102020203020402050NPV 2
Overall balancePercent of GDP
Before reform-0.8-0.9-0.9-1.2-2.0-2.9-3.6-4.2-77.8
After reform-0.8-0.5-0.5-0.6-1.0-1.4-1.7-1.9-38.2
Savings0.40.40.61.01.41.82.339.6
Source of savings:Percent of Savings
Raising retirement age 317.718.015.45.60.1-2.4-3.12.4
Lowering pension benefits6.512.935.961.874.781.384.769.0
Abolishing aguinaldo30.927.819.814.912.711.09.613.5
Raising contributions37.133.522.814.110.28.47.212.2
More employees contribute2.82.82.41.00.2-0.2-0.30.5
Survivors’ pensions5.25.03.82.62.12.01.92.4
Sources: Ministry of Finance; and Fund staff estimates.

Actuarial computations on the pension plans for teachers, public administration employees, army and police officers, university professors, and judges.

NPV represents the net present value of operational balances between 2003 and 2050 at an annual real interest rate of 5 percent.

Includes the effect of mandatory retirement (whenever applicable).

Sources: Ministry of Finance; and Fund staff estimates.

Actuarial computations on the pension plans for teachers, public administration employees, army and police officers, university professors, and judges.

NPV represents the net present value of operational balances between 2003 and 2050 at an annual real interest rate of 5 percent.

Includes the effect of mandatory retirement (whenever applicable).

  • In the short run, two measures, abolishing aguinaldo and raising the contribution rate, account for about two–thirds of the savings.72 This is because these measures came into effect immediately with the approval of the reform and do not require time for their full effect to play out.

  • In the long run, the bulk of savings—two–thirds of the net present value of savings between 2003 and 2050—arise from lowering the pension benefits. Lower pension benefits, in turn, are due to the joint effects of changing the replacement rates, the computation of the base wage, and the pension indexation mechanism. The effect of lowering pension benefits builds up slowly over time with the increase in the share of pensions regulated by the new regime, as current pension recipients die and are replaced by workers retiring under the new regime.

  • Raising the retirement age has a positive effect on savings in the earlier years following the reform—it explains about one–fifth of total savings in 2004—but this effect declines over time and eventually turns negative, implying that raising the retirement age accounts for only 3 percent of savings in net present value terms.

Box 3.2.Identifying the Sources of Savings of the Reform

In our calculations of the savings generated by the reform, we identified, for each pension plan, six distinct sources of savings (Table 3.10 reports the sum of the savings over the six underlying contributing plans). This box explains how we calculated each source of savings.

As regards the disbursement side, let n, p, and m denote, for a given year, the number of pensions, the average monthly pension, and the number of months, respectively, in the prereform scenario; we use the same letters with a prime to denote the same variables in the postreform scenario during the same year. Total savings on the disbursement side can be decomposed as follows:

Total savings from disbursement side ≡npm−n'p'm' = (n−n')pm + n'(p−p')m + n'p'(m−m').

The first term on the right–hand side represents the savings from raising the minimum retirement age at the old pension and without abolishing aguinaldo, since tighter age requirements tend to reduce the number of pension recipients in the postreform scenario relative to the prereform scenario, that is, n—n'>0. The second term represents the savings from lowering the pension benefits; because several measures introduced by the reform reduce the amount of pensions paid to the new and old pension recipients, p—p'>0. Finally, the third term represents the savings from abolishing aguinaldo, that is, m—m'=1.

As regards the contributionside, we used a similar approach to identify two sources of savings. Let a, cand mdenote the number of active employees, the average monthly contribution by employee, and the number of months contributions are paid in the prereform scenario in a given year; as before, a prime denotes the same variables in the postreform scenario during the same year. Total savings from the contribution side can be decomposed as follows:

Total savings from contribution side ≡ a'c'm'—acm = (a'—a)cm + a'(c'—c)m.

The first term on the right–hand side represents the savings from raising the number of employees who contribute in the postreform scenario (owing to raising the minimum retirement age). The second term represents the savings from raising the average contribution (owing to raising the contribution rate from 14 to 16 percent). Finally, there is no effect from changing the number of months on the contribution side, as employees contribute the same number of months in the pre–and postreform scenario.

Finally, we included in Table 3.10 the savings from survivors' pensions. Under our simplifying assumptions, these result from three factors: the effect of the reform on the number of old–age pensions (since we kept the ratio of survivors’ pensions to old–age pensions constant over time); the adjustment over time of survivors' pensions (based on wage increases in the prereform scenario and on CPI inflation in the postreform scenario); and the abolition of aguinaldo.

Mandatory retirement explains why raising the retirement age does not generate important savings in the long run. The reason for this is mandatory retirement, which the reform introduced as age 62 for public administration employees and judicial employees and as age 75 for university professors. Mandatory retirement raisesover time the number of pension recipients in the postreform scenario relative to the prereform scenario because, when the population ages, forcing older workers to retire offsets the effect of raising the minimum retirement age. This offsetting occurs for public administration employees, university professors, and judicial employees (Figure 3.2), for whom the dependency ratios in the postreform scenario eventually exceed the dependency ratios in the prereform scenario. For the contributing plans as a whole, the implication is that, although the reform initially lowers the dependency ratio, eventually it has no effect.

Figure 3.2Dependency Rations Before and After the Reform

Sources: Ministry of Finance; and Fund staff estimates.

E. Assessment and Policy Implications

Our main conclusion is that, facing explosive dynamics, the reform generates major savings in net present value terms but leaves large unfunded liabilities. While bringing the caja fiscal into financial equilibrium may have been politically unrealistic, a net present value of deficits in the order of 45 percent of initial GDP in the postreform scenario will not be sustainable in the long run. As a result, further wide–ranging reforms will be required in the future to achieve long–term financial viability.

Decisive action will be required in the future to further tighten the minimum retirement age requirements. After the reform, the plans for teachers, the army, and police officers—the largest contributing plans—continue to generate very large net present values of deficits. These are also the plans for which the reform in minimum retirement ages was most timid, implying that for these plans the projected increase in dependency ratios is the highest among the contributing plans—the ratios for these plans stand above 0.6 in 2050. Future reform will have to address this shortcoming to counteract this rapid increase in dependency ratios and its effect on the deficits.

Mandatory requirement will also have to be reconsidered in the future. Apparently, the motivation for mandatory requirement was the desire to counteract a common practice of many older, low–productivity public employees to remain on the registry of active employees. However, mandatory requirement will interact with population aging in a perverse way, boosting dependency ratios and in turn operational deficits.

Chaco War pensions will require funding well into the future, albeit of declining size. By separating the administration of Chaco War pensions from the other contributing plans and by requesting that the former be fully funded in the annual budget law, the reform makes explicit the cost to society of Chaco War pensions. In this sense, earmarking specific tax revenues to the funding of Chaco War pensions would be highly desirable, since it would underscore the commitment of the Paraguayan society to its war heroes and their survivors.

While not considered in this chapter, efforts to purge the beneficiary rolls of false claims may generate important savings. Preliminary results achieved both in the context of contributing and noncontributing plans are particularly encouraging in terms of the savings that can be reaped through administrative improvements and the census of beneficiaries.

References

    ChandSheetal K. and AlbertJaeger1996Aging Population and Public Pension Schemes IMF Occasional Paper No. 147 (Washington: International Monetary Fund).

    Oficina Internacional del Trabajo2003Paraguay: Evaluación Actuarial del Régimen de Jubilaciones y Pensiones Administrado por el Instituto de Previsión Social (IPS) (Santiago).

    World Bank2003Paraguay: Policy Options for the New Administration. Creating Conditions for Sustainable Growth (Washington).

Other, smaller institutions cover railroad workers, elected Congress officials, employees of the publicly owned electricity company, banks, and municipalities. See World Bank (2003) for a more detailed description of the pension systems in Paraguay and Oficina Internacional del Trabajo (2003) for an actuarial study of the pension system for private sector employees.

The six contributing plans of the caja fiscal cover teachers, public administration employees, university professors, police officers, army officers, and judicial employees.

Law 2345/2003.

Article 1 also says that this rate will be maintained until the system achieves financial sustainability.

Furthermore, the law does not require that past wages be adjusted for inflation for the purpose of computing the base wage.

This schedule applies only after 10 years of service. In case of early retirement, the replacement rate is further reduced by multiplying the ratio of the years of service to 62.

If wages increase by less than the inflation rate, pension adjustment is limited to wage increases.

In other words, we implicitly assumed that the inaccuracy in age information was random.

It is worth noting that our implicit assumption that information inaccuracy is random may be particularly delicate in this context, as missing information may conceal fraud and thus be nonrandom.

To better match the data, new male hires for the army are 19 years old.

As a conservative assumption, we do not allow for the lengthening to 10 years of the period used to compute the base wage for teachers. The new law allows for this lengthening if a teacher works considerably more hours or shifts at the end of his or her career.

Box 3.1 explains the details behind the computation of net present values in percent of initial GDP.

Contributions from active teachers decline starting in 2020 as the number of active teachers declines. This decline, in turn, is an implication of our assumptions that the number of active teachers in proportion to the population aged 5–19 remains constant over time. As the latter is expected to undergo a temporary decline starting in 2020, contributions from active teachers decline with the decline in the population of active teachers.

Box 3.2 explains our method of identifying the sources of savings.

In 2004, these measures accounted for even more than implied by the findings in Table 3.10, since these do not take into account the transitory regime for teachers. Since this transitory regime reduces the savings from the teachers’ plan, the relative importance of abolishing aguinaldo and raising the contribution rate is even greater.

    Other Resources Citing This Publication