Costa Rica: Selected Issues and Analytical Notes

Selected Issues and Analytical Notes


Selected Issues and Analytical Notes

Selected Fiscal Sector Issues

A. Distributional Impact of Tax Reforms1

This note provides an assessment of the distributional impact of recent income tax and VAT reform proposals. 2 The main conclusions of the analysis are that proposed income tax and VAT reforms would contribute to reduce inequality and extreme poverty, though this would be conditional on smooth implementation of targeted transfers proposed as part of the VAT reform. The reduction in inequality would be greater under the main proposal with a higher marginal tax rate on higher income brackets, higher VAT rates, and larger share of population receiving proposed transfers.

1. Costa Rica has relatively high inequality of market income. Costa Rica is an upper middle income country with low poverty levels by regional standards. However, inequality of market income, before taxes and transfers, is still high relative to OECD countries and to other more advance countries in Latin America. In contrast with other countries in the region, inequality has increased since 2008, with rising public sector salaries—particularly of qualified workers in public agencies outside central government—making the largest contribution to this trend in recent years. The extent to which inequality is reduced through current tax and transfer policies is also limited in Costa Rica, reflecting both the country’s relatively low tax pressure and relatively small cash transfers (OECD, 2016).3


Gini ratios

Citation: IMF Staff Country Reports 2017, 157; 10.5089/9781484304563.002.A002

Source: OECD (2016)

2. Costa Rica has relatively low taxation, with greater reliance on indirect taxes. Like other countries in Central America, Costa Rica has a low ratio of tax revenues, at about 13 percent of GDP, compared to an average of about 19 percent in other upper-middle income countries. Also consistent with trends in the region,4 the tax structure is biased toward indirect taxation, with the sales tax and other indirect taxes contributing over 60 percent of total tax revenue, compared to less than 50 percent on average in OECD countries. This share, however, has been declining steadily over the last decade—from a peak share of almost 75 percent in 2006. This is the result of both a decline in indirect taxation—driven by lower tariffs from gradual trade liberalization under CAFTA, as well as a decline in sales tax collections relative to GDP, possibly related to the greater contributions to growth from the exempted services sector—and an increase in direct taxation, mainly reflecting increased income tax collections efforts in recent years.


CAPDR: Tax Revenues, 2016

(Percent of GDP)

Citation: IMF Staff Country Reports 2017, 157; 10.5089/9781484304563.002.A002

Source: WEO and Fund staff calculations.
Sources: National authorities, OECD, and Fund staff calculations.

3. Costa Rica’s tax system generally favors the middle class. Greater reliance on indirect taxation in Latin America has negative implications for the progressivity of tax systems. In the case of Costa Rica, tax progression is U-shaped, with the lowest tax burden on the mid-section of the income distribution (Cubero and Vladkova Hollar, 2010). The strong concentration of income tax burden at the highest deciles of the income distribution is likely a reflection of the high tax-free threshold, at about two times the average wage in the private sector (OECD, 2016). As for the VAT, in addition to the typical concentration of the tax burden—measured as taxes paid over pre-tax income—at the lowest deciles of the income distribution, the large basket of possibly poorly targeted basic goods and services could explain the comparatively low burden of VAT on the middle class. Overall, direct taxes and transfers have some limited impact on reducing inequality, which is not substantially reversed by indirect taxes. This reflects both the relative progressivity of fiscal provisions and their relatively small size by international standards.

Sources: INEC; and authors’ estimates.Note: Gross market income includes salaries, income from investments, contributory pensions, and private transfers. Net market income deducts direct tax payments and social contributions from gross market income. Disposable income adds public transfers—mostly cash transfers but also some in-kind transfers, such as school food and transport programs—to net market income. Consumable income deducts indirect taxation from disposable income.

4. Tax reform proposals are aimed at increasing revenue collections without increasing inequality.

  • The key elements of the main tax reform proposal are: i) conversion of the current sales tax into a VAT, widening its base to include services, and gradually increasing the tax rate to 15 percent;5 while the exempted basic consumption basket would be drastically reduced, a new mechanism of targeted transfers would be introduced to compensate taxpayers in the lowest four deciles of the income distribution; and ii) introduction of new marginal tax rates of 20 and 25 percent on higher income brackets, elimination of exemptions on the 15 percent tax on income from investments, and introduction of a new capital gains tax at the same rate.6

  • The government has also presented an alternative tax proposal viewed as a second-best revenue mobilization option that it considers could have greater probability of Congressional approval ahead of the 2018 elections. The main differences in this alternative plan relative to the main proposal are: i) the absence of VAT rate increases, restriction of transfers to the first three deciles of the income distribution, and reduced VAT rates on private education and healthcare services mostly exempted in the original proposal; and ii) limiting the increase in marginal income tax rates to a new 20 percent rate, and reducing the tax rate on income from investments and on capital gains to 12 percent.

5. The impact of proposed reforms can be assessed with simulations based on household survey data. The simulations presented in this note are based on data from the household survey carried out by the Instituto Nacional de Estadística y Censos (INEC) in 2013. The analysis is static, reflecting the estimated impact of amendments to income tax and VAT provisions without any assumptions about possible dynamic adjustments in tax payers’ behavior in response to the reforms. The simulations also assume full compliance of new tax obligations and perfect targeting of transfers aimed at mitigating the impact of VAT payments on lower income households.7 While income and direct taxation are simulated at the individual level given the schedular nature of income tax in Costa Rica, expenditures and related indirect taxation are simulated at the household level. The distributional impact of tax reforms is assessed by looking at changes in the income distribution caused by the per capita impact of reforms. The key measure is the difference between estimated Gini coefficients for relevant income definitions before and after the reforms.8 The impact on poverty is assessed based on national poverty lines by region estimated by INEC, with national averages of $7 and $3 of daily per capita income for poverty and extreme poverty respectively.

6. The proposed reforms of Personal Income Taxes (PIT) would decrease the inequality of income distribution. To evaluate the distributional impact of the PIT reform, we compare Gini coefficients under the baseline Net Market Income (NIM) with those under the two reform proposals. The Gini coefficient decreases compared to the baseline under both reform scenarios—pointing to a more equal distribution of income after the reforms—but to a bigger extent under the main proposal. Intuitively, this result is due to the introduction of the new marginal rates on higher income brackets under the main proposal, and the limitation to only one additional bracket under the alternative. Overall, the reform scenarios increase the average PIT rate from 2.2 percent under the current system to 4.3 and 4.1 percent respectively under the main and alternative scenarios, by reducing rates at the bottom decile, and increasing them in the rest of the income distribution, with the main increases concentrated in the 7th and 8th income deciles.

7. However, poverty could marginally increase. The incidence of poverty measured on NMI following the reform would marginally increase from 28.8 percent to 29.3 percent under both scenarios, and extreme poverty would also rise from 8.6 to 8.8. Such increases are largely due to the generalization of the tax on investment income, but are also partially attributable to our estimation method, as we use actual tax collections under the baseline and theoretical payments under the reform scenarios. The latter may overestimate payments by not taking into account tax evasion or avoidance that can be widespread at the very bottom of the income distribution. Theoretical tax payments are also attributed to individuals by source of income without controlling for the composition of the households, thus not incorporating the small deduction for each dependent—a factor which is particularly relevant at the bottom of the income distribution where the household size and number of dependents tend to be higher. Hence the poverty increases under the reform scenarios are likely to be overstated.


PIT Rates by Income Deciles

(Percent of gross market income)

Citation: IMF Staff Country Reports 2017, 157; 10.5089/9781484304563.002.A002

Sources: INEC;and authors’ estimates.
Sources:INEC; and authors’ estimatesNote: Net market income deducts direct tax payments and social contributions from gross market income.

8. Absent the envisioned mitigating measures, the VAT reform would worsen the progressivity of indirect taxation. From a distributional perspective, indirect taxation is generally regressive. The conversion of the current sale tax into a VAT would worsen the distributional impact of indirect taxation by widening its base to include services under both proposals, and gradually increasing the tax rate to 15 percent under the main plan. The average VAT rate as a share of gross market income would increase from 3.6 percent in the baseline, to 6.4 and 5.9 percent in the main and alternative reform scenarios respectively. Although the total amount collected would be largely paid by high income households (over 60 percent of the total amount collected contributed by the top three deciles) due to their absolute higher consumption, the relative rate increase would be particularly high at the bottom of the income distribution, where consumption accounts for most of income, with the first income decile paying 16 and 14.2 percent of gross income under the main and alternative reform scenario, compared to 7.6 in the baseline.

9. Income inequality and poverty would increase accordingly, if compensation is not provided to poor households. The cumulative distributional impact of the income and VAT reforms can be measured by comparing Gini coefficients for consumable income (CI, which adds public transfers and deducts indirect taxation from NMI) net of the proposed VAT refunds.9 These coefficients would marginally increase under both reforms scenarios compared to the baseline (from 0.5118 to 0.5134 and 0.5124 under the main and alternative proposals respectively), more so under the main reform plan due to the higher rate absent in the alternative one. Poverty and extreme poverty measured on CI net of VAT refunds would also worsen as the purchasing power of households in the bottom deciles is eroded by the tax, with the share of generally (extreme) poor people increasing by 3.2 (1.5) and 2.6 (1.3) percent of the total population under the main and alternative plan respectively.


VAT Rates by Income Deciles

(Percent of grossmarket income)

Citation: IMF Staff Country Reports 2017, 157; 10.5089/9781484304563.002.A002

Sources:INEC;and authors’estimates.

10. However, the VAT refunds to lower incomes largely offset the VAT impact on income distribution. Both reform scenarios envisage a system of public transfers to compensate poor households for VAT payments. Such transfers would be distributed to the bottom 40 and 30 percent of the population under the main and alternative reform plans respectively. According to both proposals, all beneficiaries would receive the average amount spent by individuals in the second and third income deciles, resulting in over- and under- compensation in the first and fourth deciles respectively. Our estimates assume perfect targeting of such transfers—without errors of inclusions and exclusion—and do not discount the administrative cost of their distribution, hence they show the maximum potential redistributive effect of the refunds.

11. As a result, disposable income inequality would substantially fall and so would poverty. The Gini for disposable income (DI, which adds government transfers to NMI) decrease substantially under the reform scenarios compared to the baseline, and so do poverty and extreme poverty. In particular, poverty measured on DI would fall from 24.2 percent in the baseline to 20.6 percent according to the main proposal, and extreme poverty would almost halve from 4.7 to 2.4 percent of population. This suggest that the cumulative effect of the PIT reform and VAT refunds would be highly progressive and poverty reducing under both proposals, although to a lesser extent under the alternative one.10

Sources:INEC; and authors’ estimatesNote: Disposable income adds public transfers, both cash andin-kind, to net market income.

12. Overall the proposed reforms would improve the progressivity of the Costa Rican tax system and reduce income inequality, beside increasing tax revenues. The overall effect of the reform can be assessed by looking at the distribution of CI including the effect of VAT refunds, which were netted out to analyze the impact of the VAT reform alone. When looking at the overall CI, the Gini coefficient would fall from 0.51 in the baseline, to 0.49 and 0.5 under the main and alternative proposals respectively, suggesting that the progressivity of the VAT refund more than offsets the regressivity of the “core” VAT reform. This means that both proposals would increase the progressivity of the system and reduce income inequality while at the same time providing much needed tax revenues to the government.

13. Although extreme poverty is reduced under both reform scenarios, overall poverty decreases under the main proposal, but not under the alternative one. The tax reform would decrease extreme poverty from 6.2 to 4.8 percent of the population, under both plans. Overall poverty measured on CI, however, would fall from 28.2 in the baseline to 27.3 under the main proposal, but slightly increase to 28.9 under the alternative proposal. Hence, the alternative proposal would result in lower government revenues, a smaller decrease in income inequality compared to the main reform plan, and an absolute increase in overall poverty.

Sources:INEC; and authors’ estimatesNote: Consumable income deducts indirect taxationfrom disposable in come.

B. Efficiency of Public Expenditure11

This note provides an assessment of the efficiency of public expenditure in Costa Rica.12 The main conclusion regarding economic classification of expenditure is that containment of salary increases in the public sector could support needed consolidation efforts and rebalancing of expenditure toward capital investment over the medium-term. From a functional perspective, efficiency of expenditure in the education and healthcare sectors are critical given their large share in the budget. Increased focus on raising the efficiency, rather than the level, of expenditure on education would be critical to achieve higher educational outcomes and further enhance the economy’s growth potential. While efficiency of expenditure on healthcare appears high, continued gains will also be important given long-term pressures from aging trends. The authorities should take advantage of the upcoming Public Expenditure Review by the World Bank to make advances in these areas.

14. Public spending is low by international standards, but has increased sharply in recent years, especially the wage bill. Costa Rica has a low level of public spending as a share of GDP. Capital spending in particular is one of the lowest in Latin America, highlighting the need to allow some room for increases in public investment in any fiscal consolidation plan aimed at ensuring fiscal sustainability.13 The economic classification of expenditure shows that the wage bill is high by international standards, both in percent of GDP and as a share of total expenditure—almost 40 percent of the total, compared to less than 30 percent in emerging markets and latin America in particular, and to just over 20 percent in OECD countries. This was aggravated by the counter cyclical response of expenditure policy to the onset of the global financial crisis. In addition to being significantly more pronounced than in other countries in LAC and other regions, the stimulus was focused to a much larger extent on expanding the wage bill to the detriment of social benefits.14

Sources: INEC;and authors’ estimates.Note: Dashed lines are the average of LAC. Social benefits are transfers in cash or in kind to protect the entire population or specific segments of it against certain social risks. Examples of social benefits are the provision of medical services, unemployment compensation, and social security pensions (Garcia-Escribano and Yue Liu, 2017).

15. International comparisons suggest scope to contain wages in the public sector while still attracting highly qualifed employees needed to deliver high quality public services. While the share of the public sector in total employment is in line with other emerging market countries, the share of the wage bill in total public expenditure is comparatively high. The large wage premium paid in the public sector relative to employees in the private sector with similar skills suggests significant scope to contain wage growth in the public sector without negatively affecting policy outcomes due to the loss of skilled labor to the private sector. Rationalization of public compensation bonus schemes that have contributed to total salary increases well above inflation could be particularly important. Increases in salaries in other decentralized entities of the public sector that finance part of their current expenditures with transfers from the central government (CG) have also contributed to the deterioration in the fiscal situation since the global financial crisis. Going forward, savings in the public-sector wage bill would be important to support needed consolidation efforts and rebalancing of expenditure toward capital investment over the medium.

Sources: EAT, OECD, national authorities, and Fund staff calculations.Note: Wage premia refer to the wage differential between the non-financial public sector and the private sector not explained by differences in the skill mix (IMF 2016). Other transfers are mainly destined to finance current expenditures of decentralized public institutions.

16. Efficiency of expenditure in the education and healthcare sectors are critical given their large share in the budget. From a functional perspective, Costa Rica’s budget is dominated by expenditures in the education and healthcare sectors. Expenditure on these two sectors represents over sixty percent of the total budget, more than double the share in emerging markets and the OECD on average. Within Latin American, Costa Rica has the largest expenditures on both education and healthcare as a share of GDP.15 Attaining efficiency gains in these sectors could yield potentially large benefits in terms of expenditure rationalization over the long-term. This note focuses on benchmarks of spending levels and composition against comparators, as well as a high-level assessment of spending relative to outcomes to get a sense of spending efficiency. The authorities should take advantage of the detailed recommendations from the upcoming Public Expenditure Review of the World Bank to increase efficiency in these key sectors.

Sources: EAT, and Fund staff calculations.

17. Education outcomes do not reflect the high level of public spending on education. Notwithstanding the country’s high expenditure on education, only behind Denmark and Sweden among advanced economies, education outcomes are not significantly better than in other emerging markets. In secondary school, teacher-student ratios are only slightly higher than the average in emerging markets, pointing to relatively elevated salaries of teachers by international standards. Regarding policy outcomes, while school enrollment ratios in primary education are in line with OECD countries, they are not significantly higher than in other emerging markets in secondary school. OECD standardized test (PISA) scores for secondary school students are also broadly consistent with those of other emerging markets that spend much less on education, suggesting significant scope for improving efficiency. Scores are well below those of advanced economies consistently for PISA tests on reading, math, and science. In terms of policy recommendations, the authorities should move away from an exclusive emphasis on increasing spending toward establishing better educational outcomes as the main policy target. Specific recommendations made by the OECD in the context of the evaluation for accession included improving evaluation mechanisms and enhancing accountability for teachers, as well as improving their professional development and harmonizing their qualifications. Reinforcing the vocational technical track would also contribute to reduce drop-outs at secondary level and tackle high youth unemployment (OECD, 2016).

Figure 1.
Figure 1.

Costa Rica: Education Spending and Outcomes

Citation: IMF Staff Country Reports 2017, 157; 10.5089/9781484304563.002.A002

Sources: EAT, OECD, and Fund staff calculations.1/ Dashlines are the average of LAC.

18. The scope for early gains in efficiency appear much more limited in the healthcare sector. Public spending on healthcare as a share of GDP is high relative to the country’s level of development, significantly higher than the average for emerging markets, and also higher than in richer countries in the region like Chile, Uruguay, and Mexico. Efficiency of public spending on healthcare also appears to be elevated, with near universal coverage and health outcomes close to OECD levels despite significantly lower total health expenditure per capita. These favorable outcomes are not explained by significant additional private sources of health spending, as Costa Rica has one of the lowest shares of out-of-pocket and other private health expenditures in the region. Notwithstanding this favorable assessment, with Costa Rica being one of the countries in the region where population aging is more advanced, health expenditures are projected to double over the next 50 years if universal access and current service levels are to be maintained (Section C). This highlights the need for measures to gradually increase efficiency. OECD recommendations include updating information systems to better monitor performance indicators, forward-looking allocation of resources that considers changing demographic patterns and disease trends, and introducing diagnosis-related funding schemes that provide stronger incentives to control spending than fee-for-service schemes that can result in service oversupply (OECD, 2016).

Figure 2.
Figure 2.

Costa Rica: Health Spending and Deficiency

Citation: IMF Staff Country Reports 2017, 157; 10.5089/9781484304563.002.A002

Sources: EAT, WEO, World Health Organization, and Fund staff calculations.1/ Trendline based on worldwide distribution of countries.2/ Dashlines are the average of LAC.3/ Includes expenditures covered by private plans and the non-profit sector.

19. Expenditure on social protection is relatively low, reflecting the country’s low poverty levels and universal coverage of basic services by the public sector. Transfers directly aimed at alleviating poverty and addressing income inequalities are comparatively low in Costa Rica, at ¾ percent of GDP, about half the average in emerging markets. This is likely a reflection of Costa Rica’s relatively low poverty levels—22 percent of the population as of 2015—as well as its social model aspiring to universal coverage in the education and healthcare sectors, which reduces the need for large social assistance spending. The OECD finds that benefits in the form of healthcare and education services provided by the state reduce inequality, as measured by the Gini coefficient, by almost one-fourth, a larger reduction than in other Latin America countries covered in the analysis (OECD, 2016). Efficiency of expenditure on social protection spending, as measured by the extent of targeting of lower income households achieved, is close to the average in other emerging and OECD countries.

Sources: EAT, and Fund staff calculations.1/ Includes unconditional cash transfers, cash transfers, social pension, food and in kind transfers, school feeding, public works, fee waivers and other social assistance.2/ Dashlines are the average of LAC.3/ Coverage is equat to: (number of individuals in the quintile who live in a household where at least one member receives the transfer)/(number of individuals in that quintile). Benefit incidence is equal to: (sum of all transfers received by all individuals in the quintile)/(sum of all transfers received by all individuals in the population).

20. The absence of subsidies to energy avoids a typical drag on expenditure efficiency in some other countries in the region. Movements in international oil prices are automatically passed through to domestic petroleum product prices, avoiding the use of government subsidies to cushion the impact on consumers that produce fiscal costs, economic inefficiencies, and environmental damages in some other countries in the region. Implicit post-tax subsidies from foregone consumption tax—as petroleum products are taxed at lower rates than other consumption products—and estimated externalities on the environment are also fairly low, in line with the OECD average and consistent with the country’s environment-friendly policies (AN V). At the same time, efficiency could be enhanced by replacing cross-subsidization of lower-income households (through exemptions from VAT for low residential consumption customers) with more transparent targeted social transfers.


Energy Subsidies by Component, in percent of GDP, 2014 1/

Citation: IMF Staff Country Reports 2017, 157; 10.5089/9781484304563.002.A002

Source: IMF FAD Expenditure Assessment Tool (EAT), IMF Energy Subsidy Estimates.1/ Dashlines are the median for countries in each region.

C. Long-term Fiscal Gaps from Population Aging16

This note estimates the fiscal costs of population aging in Costa Rica and provides recommended reforms to make these costs manageable. The main finding is that demographic trends make Costa Rica one of the region’s most vulnerable countries to long-term fiscal pressures from population aging. Under current policies, expenditure on pensions and healthcare would reach 16 and 25 percent of GDP respectively by 2100, generating one of the largest long-term fiscal gaps in the region. As indicated in the recent independent actuarial report, significant parametric reforms will be needed in the PAYG pension system to meet the objectives of maximizing coverage, while maintaining socially acceptable benefits and preventing at the same time the emergence of large long-term fiscal gaps. Stricter budget controls and improved incentives systems are also required to contain healthcare spending growth while preserving access to high-quality healthcare.

21. Costa Rica is one of the countries of Latin America with more advanced population aging. Costa Rica’s fertility and life expectancy are broadly in line with the averages for high income countries (HIC), and significantly below and above, respectively, Latin American averages. While the share of the old in total population is currently similar to the average in Latin America, it is expected to rise sharply over the long-term, surpassing the average for HICs from 2060, according to the United Nations’ long-term population projections. This will directly affect fiscal sustainability by putting pressure on public pension and healthcare systems.

22. The main purpose of this note is to quantify the long-term fiscal gaps from population aging by projecting public expenditure on pensions and healthcare. These projections can be produced through a stylized exercise that requires limited parametrization, based on countryspecific system characteristics. The exercise draws on demographic projections from the United Nations and methodologies developed by the IMF (Clemens et al, 2015) to derive spending projections. Notwithstanding the large uncertainties surrounding these long-term projections, they provide a useful basis for the necessary discussion on the fiscal implications of population ageing based on a consistent analytical framework across countries.

  • Pension expenditures are projected based on the following identity:
    PEGDP=population65+population1564Old-age dependency ratiopensionerspopulation65+Elderly coverage ratioaveragepensionaveragewageReplacement ratepopulation1564workersInverse employment ratiolaborincomeGDPLabor share of GDP
    where PE/GDP denotes the ratio of pension spending to GDP, population 65+ is the population aged 65 years or older, and population 15 – 64 is the population between ages of 15 and 64. The employment ratio (labor force participation) and the labor income share of GDP are typically assumed to be constant over time. Hence, future pension spending growth is essentially determined by parametric changes in the system (that is, changes in the elderly coverage ratio or the replacement rate) and demographic changes captured by the old-age dependency ratio.
  • Health expenditures are computed using the following formula:
    HEtGDPt=(ΣiPopulationi,t.HEi,tPopulationi,tHEref,tPopulationref,tAge spending profile).HE0PopulationGDP0Generosity of health system .(1+excesscostgrowth)t
    where HEt denote public health expenditures to GDP at time t, the subscript i corresponds to the specific age-group i, and the ref subscript indicates the age-group used as reference to compute the age-spending profile (assumed to be the 40–44-year group), xcesscost_growth represents the excess growth in health spending in real per capita terms over real GDP per capita not due to demographic changes, for instance due to costly medical innovation. This is assumed to be constant over time and equal to 1 percent based on historical trends in advanced economies, which in many cases already had universal coverage like Costa Rica (IMF, 2012). Population aging affects the first term of the product as the population in older age groups, for which health spending per capita is higher relatively to the health spending per capita in the reference group, increases compared to the population in younger groups.
  • Long-term fiscal gaps can in most countries be measured by the present discounted value (PDV) of projected government pension spending increases, assuming constant revenue ratio to GDP. In the case of Costa Rica, given already approved increases in the contribution rate for the main pay-as-you-go (PAYG) system, from current 8.5 percent rate to 10.5 percent by 2035, the PDV of projected pension system deficits provided a more accurate measure of long-term fiscal gaps.17

23. Age-related public spending is already high by regional standards reflecting legacy costs from closed pension systems in the public sector and universal coverage in healthcare. Despite having an old age dependency ratio broadly in line with the regional average, Costa Rica already ranks high in regional comparisons of expenditure on pensions and especially healthcare, the latter reflecting the universal coverage and high expenditure per capita identified in Section B. Regarding pensions, those paid out of the main PAYG pension system for private sector and active public sector workers administered by the Caja Costarricense de Seguridad Social (CCSS) are about 2¾ percent of GDP,18 in line with current median pension expenditure levels in the region. However, total pension payments are almost double, nearing the 75th percentile in the region at 5¼ percent of GDP, including pensions paid out of the CG budget under special regimes for civil servants and teachers that have been closed to new participants since 1992. Costa Rica also has a mandatory individually funded pillar that was introduced in 2000.

Sources: Country authorities and Fund staff calculations.1/ The upper end of the box represents the 75th percentile of regional spending, the middle line is the median, and the bottom end of the box the 25th percentile of spending. The ends of the whiskers represent the 5th and 95th percentiles respectively of regional pension spending in 2015.2/ Pension spending in Costa Rica includes pensions paid out from the CCSS and pensions paid out from the CG budget for special regime systems for civil servants and teachers closed to new participants.

24. The analysis in the remainder of this note focuses on long-term fiscal gaps from the main PAYG pension system and from the healthcare system that will be the main sources of long-term pension spending pressures.

  • Given an average age of participants of 62 years, expenditures on the special regimes covered by the budget are expected to gradually decline over the long-term and are therefore less relevant for the main exercise in this note, focused on potential contingent liabilities from increases in ageing pressures over the long-term, and related parametric reforms needed to contain these risks.

  • In principle contingent fiscal liabilities could eventually arise in countries with low coverage and defined contributions pillars generating replacement rates below socially acceptable levels. In these countries, advised reforms would go in the direction of increasing coverage and benefits, with contribution rates also adjusted as needed to support these objectives. But this is not the case in Costa Rica, where coverage is already relatively high and combined projected replacement rates of the PAYG and the individually funded pillars are almost 80 percent of earnings, significantly above the regional and OECD averages.19

  • Therefore, our analysis of long-term gaps from pensions in the case of Costa Rica focuses on the main PAYG pension system administered by the CCSS where large fiscal gaps would emerge under current policies and projected demographic trends.20 Despite low contribution rates of 8½ percent of earnings, compared to average of almost 15 and 20 in LAC and the OECD respectively, the PAYG system is currently still in surplus, with accumulated reserves of about 6½ percent of GDP. The current level of pension expenditures in line with the median regional level reflects the combination of a relatively low replacement rate compared to other regional PAYG systems—though still high by regional standards—and a desirable relatively high passive coverage among the current retirement-age population. This is projected to increase further over the long-term, following the significant welcome increase in active coverage during the last decade—from less than half of the economically active population contributing in the early 2000s to over 60 percent in recent years.21

Figure 3.
Figure 3.

Contribution Rates, Replacement Rates, and Passive Coverage of Pension Systems in Latin America

Citation: IMF Staff Country Reports 2017, 157; 10.5089/9781484304563.002.A002

Sources: IDB, OECD, World Bank Pension database, and Fund staff calculations.1/ Gross replacement rates are theoretical estimates based on a specific set of assumptions, including 100 percent contribution density and equal earnings in percent of economy-wide earnings throughout a worker’s career; while actual replacement rates are lower—reflecting rising earnings of individuals over time due to career advancement—this set of assumptions allows for consistent cross-comparison of generosity of statutory characteristics under different pension systems (see OECD/IDB/The World Bank, 2014). The OECD average gross replacement rates refer in both charts to combined defined benefit and defined contribution systems.

Key Pension System Parameters in LAC

article image
Sources: OECD, Pensions at a Glance 2013 (OECD average contribution rates); OECD, Pensions at a Glance 2015 (OECD average pensionable age); OECD/IDB/The World Bank (2014), Pensions at a Glance: Latin America and the Caribbean (type of system and gross replacement rates); United States Social Security Administration (2016), Social Security Programs Throughout the World: The Americas, 2015 (pensionable age, vesting period, contribution rates).

DC systems may include a DB pillar in the process of being phased out. For countries with fragmented systems, the largest single component is taken as a benchmark.

Numbers in parentheses are for women, where different than men.

Comparisons are based on a specific set of assumptions. See IDB for detailed information.

Only includes old-age, disability, and survivors. Values for Argentina are net of location-based tax credit for employers.

25. Costa Rica has one of the largest projected long-term fiscal gaps in the region, though the comparison also reflects low coverage and socially unsustainable benefits in some other countries in the region. Under policies prevalent at end-2016, pension expenditures of the PAYG system would increase from around 2¾ percent of GDP in 2015 to over 13 percent of GDP by 2065 and almost 16 percent of GDP by 2100, driven by a sharp projected increase in the old age dependency ratio. Even with approved gradual increases in the contribution rate until 2035, reserves would run out between 2025 and 2030, when the system would start having cash deficits. The long-term fiscal gap measured by the PDV of projected government pension spending increases until the year 2100 is almost 400 percent of GDP. Considering projected increases in contributions from approved reforms, the PDV of pension system deficits would reach more than 350 percent of GDP.22 While these estimated gap highlights the need for significant parametric reforms to ensure the long-term fiscal sustainability of the system—while also ensuring socially acceptable coverage and replacement rates—the unfavorable comparison with other countries in the region also reflects the limited coverage of PAYG systems and socially unacceptable replacement rates of defined contribution systems in some of these countries.

Note: Countries highlighted in red are countries with either DC systems or mixed DB/DC systems. More developed economies include Australia, Canada, European countries, Japan, New Zealand, and the United States, following the UN grouping (Clements et al, 2015).Sources: Caja Costarricense de Seguro Social, and Fund staff estimates and projections.

26. The contribution of public healthcare costs to long-term fiscal gaps is even larger. Costa Rica has the highest estimated long-term fiscal gap in the region from projected increases in healthcare spending. This reflects the impact of more severe aging trends as well as the higher starting spending level—consistent with desirable universal access and high quality aspects of the health system by regional standards—than in most other countries in the region. Total healthcare spending amounted to 8 percent of GDP in 2015 and is projected to increase to 25 percent of GDP by 2100, with a PDV of spending increases equal to almost 500 percent of GDP.

Note: More developed economies include Australia, Canada, European countries, Japan, New Zealand, and the United States, following the UN grouping (Clements et al, 2015).Sources: Country authorities, and Fund staff estimates and projections.

27. Significant parametric reforms will be needed to eliminate the long-term fiscal gap arising from the PAYG pension system. Consistent with the approach taken in the cross-country study cited above, two objectives are defined for parametric reforms: the first one aims at maintaining the current balance of the pension deficit constant until 2030; the second limits the deterioration in the system balance to 3 percent of GDP by 2065. Multiple sets of parametric reforms would be consistent with these objectives, although the size of any adjustment is logically larger the longer it is delayed. For illustration purposes, the following simulations indicate the adjustments that would be needed to achieve the reform objectives if only one of the key system parameters were to be amended:

  • A gradual increase in contribution rates to 14.5 percent of wages by 2030 would be sufficient to reach the first objective set above. While the second objective of limiting the increase in the deficit by 2065 could also be accomplished with an immediate increase to a not much higher contribution rate of 16.4 percent, more gradual reform scenarios could require increases in contribution rates of up to 31.5 percent by 2065. Under almost all these scenarios without reforms in benefits and retirement age, however, system deficits would continue to rise after 2065 as demographic pressures would eventually offset the buffers built with higher contribution rates.

  • Focus on reducing spending through cuts in replacement rates seems likely in the case of Costa Rica, where the introduction of a mandatory individually funded pillar in the early 2000s would lift the combined theoretical replacement rate of the two systems to 80 percent once the second pillar matures. Cuts in PAYG system benefits of about 25 and 35 percent would be sufficient by themselves to achieve the objectives of preventing an increase of the PAYG pension deficit by 2030 and limiting the deterioration to 3 percent of GDP by 2065, respectively.

  • Gradual increases in the retirement age to 69 by 2030 and to 75 by 2065 would also be sufficient by themselves to meet the reform objectives set in this exercise. While the retirement age is not currently low by international standards and reforms are therefore likely to be less focused on the retirement age during the initial phase, some increases are likely in the longer-term in line with global trends as population ageing continues to advance.

  • As an additional theoretical exercise, pension adequacy—defined as replacement rate around the OECD average—could be maintained even under the scenario where sustainability of the PAYG system is achieved solely through a cut of 35 percent in benefits if this were accompanied by a moderate increase in the DC contribution rate from 4.25 to 6.5 percent.23 Actual future replacement rates under this scenario would logically be subject to the uncertainty of future market returns, and moving too far in this direction could increase the risk of not delivering socially acceptable levels of pensions. Alternatively, the same replacement rate around the OECD average could be achieved with a smaller 25 percent cut in benefits under the PAYG system, which would require an increase in the PAYG system contribution rate of about 3.5 percent to at the same time achieve the objective of limiting the increase in the PAYG system deficit to 3 percent of GDP by 2065.


Costa Rica: PAYG Contribution Rates and Pension System Balance Under Reform

Citation: IMF Staff Country Reports 2017, 157; 10.5089/9781484304563.002.A002


Costa Rica: Combined PAYG and DC Pension System Replacement Rates

Citation: IMF Staff Country Reports 2017, 157; 10.5089/9781484304563.002.A002


Costa Rica: Combined PAYG and DC Pension System Reforms

Citation: IMF Staff Country Reports 2017, 157; 10.5089/9781484304563.002.A002

28. International comparisons suggest a greater initial focus of parametric reforms on raising contribution rates and cuts in benefits. Given the comparatively high combined replacement rates of the PAYG and funded pension pillars, and the low contribution rates of the PAYG system in Costa Rica, a reform to ensure sustainability of the pension system is more likely to entail a combination of increases in contribution rates and cuts in benefits than substantial early increases in the retirement age, which is already relatively high in the region, consistent with the already mature stage of the demographic transition in the country. However, over the long-term as life expectancy continues to rise, increases in the retirement age are also likely both in Costa Rica and globally. The range of reforms delineated above represent extreme scenarios focused on reforms in single parameters, but are broadly in line with the refined recommendations made in an independent report commissioned by the pensions supervisor. The report included a menu of multi-parametric reform options with increases in the retirement age up to 70 years, cuts in the replacement rate of about one fifth, and gradual increases in contribution rates up to 25.9 percent by 2055 (Universidad de Costa Rica, 2016).

29. Standard recommendations to contain healthcare costs in countries with more extensive healthcare coverage put the emphasis on enhancing budget controls. While the upcoming World Bank Review of Public Expenditure can provide more specific proposals to enhance efficiency in the Cosa Rican healthcare system, standard recommendations in countries with already high coverage focus on measures needed to contain costs and improve spending efficiency while preserving access to high-quality healthcare (Clements, Coady, and Gupta 2012). Typical measures include:

  • Budget caps with central oversight: these should be based on reasonable and objective expenditure projections, as opposed to simply reimbursing all spending.

  • Public management and coordination of healthcare services to control costs and screen out unnecessary services: For example, gatekeeping, through which a primary care physician manages a patient’s healthcare services and coordinates referrals to specialists, is widely considered crucial to constraining the growth of costly and often unnecessary hospital treatment.

  • Local and state government involvement in key health resource decisions: this can help tailor services to local conditions, while also increasing incentives to avoid overruns when coupled with increased financial responsibility of local governments.

  • Use of market mechanisms in the healthcare system: these include increasing patient choice of insurers, allowing greater competition between insurers and providers, and relying on more private services. Moving away from simple reimbursement of provider costs toward contracting systems with built-in incentives for providers to minimize waste and improve services also enhances efficiency.

  • Increasing the share of costs borne by patients: either higher copayments or expanded private insurance—accompanied by targeted measures to ensure access to health services by the poor and chronically ill—has also been successful in containing the growth of public health spending.

Annex I. Costa Rica Pension System Structure

article image
Sources: OECD/IDB/World Bank (2014) and SSA (2016).


  • Clements, B., D. Coady, S. Gupta, 2012, “The Economics of Public Health Care Reform in Advanced and Emerging Economies”, International Monetary Fund, Washington.

    • Search Google Scholar
    • Export Citation
  • Clements, B., K. Dybczak, V. Gaspar, S. Gupta, and M. Soto, 2015, The Fiscal Consequences of Shrinking Populations. IMF Staff Discussion Note, October 2015.

    • Search Google Scholar
    • Export Citation
  • Cubero, R., Vladkova Hollar, I., 2010, “Equity and Fiscal Policy: The Income Distribution Effects of Taxation and Social Spending in Central America,IMF Working Paper No. 10/112 (Washington: International Monetary Fund).

    • Search Google Scholar
    • Export Citation
  • Garcia-Escribano, M, Yue Liu, C., 2017, Expenditure Assessment Tool (EAT), Technical Notes and Manuals (Washington: International Monetary Fund).

    • Search Google Scholar
    • Export Citation
  • OECD, 2013, Public Spending on Health and Long-term Care: A New Set of Projections, OECD Economic Policy Papers No. 6.

  • OECD/IDB/The World Bank, 2014, Pensions at a Glance: Latin America and the Caribbean, OECD Publishing,

    • Search Google Scholar
    • Export Citation
  • OECD, 2016, OECD Economic Surveys: Costa Rica 2016: Economic Assessment, OECD Publishing, Paris.

    • Search Google Scholar
    • Export Citation
  • SSA, 2016, Social Security Programs Throughout the World: The Americas, 2015. SSA Publication No.13-11802, March 2016.

  • UN, 2015, World Population Prospects: The 2015 Revision. Key Findings and Advance Tables. United Nations.

  • Universidad de Costa Rica, Facultad de Ciencias, Escuela de Matemáticas, “Estudio Actuarial del seguro de invalidez, vejez y muerte administrado por la Caja Costarricense de Seguro Social,Diciembre, 2016.

    • Search Google Scholar
    • Export Citation

Prepared by Valentina Flamini, Luis Ángel Oviedo, Jaume Puig, and Juan Diego Trejos.


This study is part of a collaborative effort between the IMF and the Commitment to Equity (CEQ) Institute. The CEQ project is an initiative of the Center for Inter-American Policy and Research (CIPR) and the Department of Economics, Tulane University, the Center for Global Development and the Inter-American Dialogue. For more details visit”


Transfers in kind, including universal access to public education and healthcare services, make a much more significant contribution to reducing inequality in Costa Rica, but they are outside of the scope of the analysis in this note focused on the distributional impact of proposed income tax and VAT reforms combined with cash transfers.


Cross-country comparisons of tax structures in Latin America are available in Cubero and Vladkova Hollar (2010).


VAT simulations presented in this note look at the distributional impact once the transition to the 15 percent tax rate is completed.


The distributional impact of the capital gains tax is not modelled in the simulations below due to lack of survey data on capital gains.


The assumption about tax compliance implies a possible bias toward assessment of less progressivity/more regressivity of tax reforms if informality and tax evasion are greater among taxpayers at the lower end of the income distribution. The assumption about perfect targeting of transfers could be subject to biases in both directions depending on extent of errors of inclusions and exclusion in the transfers system.


The relevant income definitions are net market income for income tax reforms and post-fiscal income for VAT reforms.


Technically, the CI also includes the effect of government transfers other than VAT refunds, and other direct and indirect taxation, in both the baseline and reforms scenarios. However, as these components are not changed under the reforms, they do not bias the comparison.


Although VAT refunds are conceived to compensate households from the erosion of purchasing power caused by the VAT, from an income classification perspective, they are transferred to households before the latter spend the corresponding amount of disposable income in consumption. Hence, DI incorporates the redistributive effect of the PIT reform and VAT refunds, while CI net of VAT refunds show the effect of PIT and VAT reforms.


Prepared by Jaume Puig.


The analysis is based mainly on cross-country comparisons of expenditure levels and measures of efficiency relative to key policy outcomes available from the Expenditure Assessment Tool (EAT) developed by the IMF’s Fiscal Affairs Department. The tool focuses on expenditure at the general government level as reported in the IMF’s WEO. Expenditures of the social security administrator, which are typically included in general government but are not part of the data reported in the WEO for Costa Rica, have also been included in the analysis.


Efficiency aspects of public capital expenditure are covered in AN IV on addressing infrastructure bottlenecks to boost competitiveness.


The also comparatively high increase in other current expenditures since 2008 was mostly driven by current transfers to decentralized institutions, which are in a significant part also destined to pay salaries.


Long-term fiscal sustainability of healthcare spending is covered as part of the analysis of the long-term fiscal implications of aging in Section C of this note.


Prepared by Jaume Puig, based on a study of long-term fiscal gaps in Latin America and the Caribbean by a team led by Lorenzo Figliouli.


The exercise presented in this note does not include the impact of the recent decision to accelerate the pace of increases in the PAYG contribution rate, with an increase to 9.5 percent starting from June 2017, compared to previously approved increases of 0.5 percentage points every five years starting in 2020. The impact of this measure is limited relative to the size of the long-term fiscal gaps estimated below.


This includes expenditure on non-contributory pensions of about ¼ percent of GDP, financed by transfers from the CG and public enterprises.


Projected gross replacement rates are theoretical estimates based on a specific set of assumptions, including 100 percent contribution density. Actual replacement rates in all countries are lower, around 65 percent in Costa Rica reflecting actual contribution density closer to 75 percent.


There are also separate special regimes for magistrates and teachers that joined the public sector after 1992. While these systems are currently in surplus, the special regime for magistrates is also deemed to be actuarially imbalanced. A recent reform proposal by the superintendence of pensions envisaged annual CG transfers of about 0.1 percent of GDP to cover pensions of current retirees over the next three decades (retirees would also pay an extraordinary contribution rate); under the same proposal, the system for current active contributors would undergo parametric reforms to reach actuarial balance, including an increase in the retirement age and an increase in the reference period to determine pensions.


While increases in labor formality and active coverage of pensions reduce fiscal gaps in the short term, they lead to higher long-term deficits over the long-term. While revenues increase relative to expenditures as long as new contributors remain in the labor force, as they retire, passive coverage increases proportionately, and the additional pension benefit increases offset previous gains in contributions. In the case of Costa Rica with an actuarially imbalanced contributory system and limited coverage and generosity of non-contributory pensions–which the elderly would have received if they remained in the informal sector–the net effect of higher coverage by contributory pensions on the long-term fiscal gaps is negative.


Key assumptions, in line with the methodological approach described above include: 1) underlying demographic trends based on UN projections; 2) 45-year transition period assumed for system to mature–with gradual increase in coverage of pensioner age population up to current level of contributions’ coverage among working age population; the latter is assumed to remain constant since it is already one of the highest in the region; 3) replacement rate assumed to remain constant (i.e. initial pension grows with average wages); 4) as mentioned above, projected increases in contribution rates only include the increase to 10.5 percent by 2035 approved in 2005–projections do not change materially with the planned 1 percent increase in worker contributions from June 2017 announced by the CCSS in early 2017 following the publication of an actuarial report by the Universidad de Costa Rica commissioned by the CCSS.


The simulation is for combined DB/DC replacement rates of male pensioners, assuming 75 percent contribution density.

Costa Rica: Selected Issues and Analytical Notes
Author: International Monetary Fund. Western Hemisphere Dept.