José-Francisco Pacheco
Alberto González Pandiella
Expanding Social Protection and Addressing Informality in Latin America
This chapter builds on an original background report prepared by José-Francisco Pacheco for the OECD. The current version has been revised and edited by Alberto González Pandiella (OECD), incorporating additional updates.
Abstract
Costa Rica’s social protection system has historically ensured broad coverage and good quality, but it now faces growing pressures from demographic change, widespread informality, and persistently high poverty and inequality. With around 40% of workers in informal employment, many remain excluded from protection, while high payroll contributions discourage formal job creation. To expand coverage and reduce informality, reforms are needed to secure at least a basic level of social protection for all. A gradual strategy could begin by extending coverage to the elderly poor and poor children. The chapter proposes shifting financing away from payroll contributions toward general taxation, a reform that would both facilitate formal employment and strengthen social protection.
Costa Rica has a long tradition of social protection initiatives aimed at protecting its population from different risks. While informal initiatives were already implemented since the mid-19th century, the country’s first formal social protection pillar emerged during the 1940s, with the launch of the Caja Costarricense de Seguro Social (CCSS) and the labour code reforms. Since then, the contributory social insurance system has expanded to other sectors. Non-contributory social assistance initiatives aimed at poverty reduction emerged during the 70s. Such initiatives included the creation of the Fondo de Desarrollo Social y Asignaciones Familiares (FODESAF), a fund financed by payroll contributions and public budget transfers. Over time, a network of social assistance programs, primarily funded by FODESAF, expanded to cover children, the elderly and the poor. Social protection financing as it stands today relies heavily on social contributions. Overall, mandatory social contributions amount to about 37% of wages, with around two-thirds of these resources directed to the Costa Rican Social Security Fund (CCSS). Employers bear most of the cost, covering roughly 72% of total contributions.
Despite being characterised by consolidated institutional and financing models, the social protection system in Costa Rica suffers from fragmentation in service delivery and falls short in several aspects. In particular, labour informality levels have grown in the last decades, and now affect around 40% of total workers. Although multiple factors may explain labour informality (Pacheco, 2020), as analyzed in the OECD Economic Surveys of Costa Rica, there is increasing acknowledgement that social protection financing, and in particular, high mandatory contributions, are an important factor (Flores-Estrada Pimentel, 2021).
Various options for reducing informality in Costa Rica have been discussed in the past decade, but few concrete actions have materialised. This chapter examines the option of reducing reliance on social contributions and increasing the use of general taxation to finance social protection in Costa Rica, and how this shift could lower informality and strengthen social protection.
8.2.1. Poverty and inequality
Since 1994, household poverty levels in Costa Rica have remained stable at around 22% (Figure 8.1). Extreme poverty rates have also remained stable, affecting roughly 6 out of every 100 households.
% of households
Source: INEC historical records.
During the pandemic (2020), most social indicators worsened, with household poverty rates growing to 26.2% (30% of the population). These rates started to decline again in 2021, and by 2024 poverty incidence was below pre-pandemic levels and similar to 2007 levels.Different sectors of the population experience different rates of poverty, with children, unemployed persons, and people without formal education experiencing the highest rates (Figure 8.2).
% of population, 2024
Source: ENAHO 2024.
Inequality trended up over the 1990s and 2010s (Figure 8.3). It started falling in 2021, although it remains higher than in the average OECD economy.
Gini coefficient, from 0 ('perfect equality") to 1 ("perfect inequality")
Source: INEC historical records.
8.2.2. Labour Market Structure
Unemployment increased substantially in Costa Rica during the pandemic (Figure 8.4). The Government adopted a series of sanitary measures to prevent the spread of the virus including driving restrictions, business closures and the shutdown of beaches, cultural and religious sites and frontiers. Face-to-face attendance to the school year was replaced by virtual lessons. During the first quarter of full pandemic control measures (2nd quarter of 2020), unemployment doubled from 12% to 24% and remained over 20% during the following three quarters. The trend reversed in 2021 and unemployment has been on a declining path, partly due to a fall in participation among some segments of the population. By 2023, unemployment achieved a one-digit rate (7.3%), being this the lowest rate since 2016. It has stayed broadly constant since then.
% of labour force
Source: INEC historical records.
Like poverty, different populations experience different rates of unemployment (Figure 8.5). Women, young people, and residents of the Central Pacific and Brunca regions experienced the largest pre-pandemic unemployment rates (IV-2016 to I-2020). The youth unemployment rate affected 25% of citizens aged 15 to 24 years, female unemployment averaged 13.3%, and Central Pacific and Brunca regions observed rates between 13.5% and 13.9% of their labour force, respectively. During the pandemic period (II-2020 to IV-2021), youth unemployment reached 48% during the second quarter of 2020, and unemployment among women reached an unprecedented 24.6%.
The pandemic also increased unemployment rates among people aged 25 or above from 7.3% (pre-pandemic) to 14.9% (pandemic), partially closing the gap between youth unemployment and adult unemployment. The gap between women and men widened during the pandemic, from an unemployment rate of 1.6 to 1.65 times greater among women than among men.
Finally, the post-pandemic period (I-2022 to I-2024) shows reduced unemployment rates in all groups in comparison to pandemic years. Perhaps the most important feature of this period is the level of recovery that each group experienced in relation to pre-pandemic years. In this regard, unemployment in the Central Region, Huetar Norte Region and people with secondary education is higher than before COVID-19. On the contrary, rates are now lower in the rest of regions, women and those unemployed persons who used to work in construction, lodging and food and commercial activities. Youth unemployment remains at the same rate as in pre-pandemic years, although the rate itself remains high.
Average unemployment rate
Source: INEC database.
Higher informality rates is also a growing matter of concern given that informality is associated with lower levels of social protection, productivity, and resilience to macroeconomic shocks. Informality has been hovering around 42% over the last 15 years although in that period the rate experienced an inverse U-shaped curve. In this way, informality moved from 37.6% of total occupation in 2010-2011 to 45.1% in 2018-2019, reaching a maximum 47.1% in the first quarter of 2020. Since 2021, informality dropped to 41% of total workers. Based on the information of Figure 8.6, informality tend to be higher among less-educated workers over 25 years, living in rural areas. Rates by sex show no significant differences.
Informal workers, who already had low incomes, experienced a sharp decrease in income during the pandemic. In pre-pandemic years, 42.1% of informal workers earned less than the minimum wage (US$ 2.2 per hour) and just 11.7% earned two times the minimum wage or more. In the second quarter of 2020, 48.7% earned less than the minimum wage and just 8.3% earned two times the minimum wage or more.
Informality rate, %
Source: INEC database.
This section provides an overview of the social protection system in Costa Rica including contributory social insurance schemes, key non-contributory social assistance programs, and the main labour market-related initiatives.
8.3.1. Overview of Existing Systems
The Costa Rican social protection sector includes the three typical components that characterise modern systems: social insurance (contributory), social assistance (non-contributory), and labour market policies. Social insurance programs are contributory programs that typically cover formal workers. Social assistance programs are non-contributory programs that typically cover informal workers. Labour market policies are the regulations that define the labour market structure.
The primary social insurance institution is the Caja Costarricense del Seguro Social (CCSS), which was first created in 1941. The CCSS covers healthcare services, social security provisions and the contributory pension regimen. The Caja grants medical care, sickness, old-age (contributory pensions), maternity, invalidity, and employment injury benefits that are provided to direct affiliates and, in some cases, cover dependents (as in the case of healthcare) and widows (survivors' benefits).
The social insurance scheme is financed through social contributions from employers, workers and the State. The State contributes in multiple ways. First, as an employer. Second, there is considerable legislation that has created special health programmes (vaccination, in vitro fertilization, vulnerable groups insurance coverage, smoking cessation) that are financed through taxation or debt. Third, the State also contributes with 1.91% to the contributory pension fund and with 0.25% to the health insurance scheme. Each year, the Ministry of Finance allocates a determined amount of funds in the National Budget that are later transferred to the CCSS.
The CCSS has set a minimum contributory income of reference under which no salaried worker can contribute. In 2023, the corresponding floors were defined at US$564 for pensions and US$603 for healthcare. In 2022, the Board of the CCSS reduced the minimum contributory income for independent workers and defined it at US$283.
The CCSS also administers the non-contributory pension regime (NCR). The NCR, created by Law 5662 of 1974, is a monetary transfer “with the objective of protecting all those people who are in need of immediate economic protection and do not qualify in any of the existing contributory or non-contributory regimes” (Abarca and Fernández, 2017: 5). The subsidy is mainly addressed to protect elderly, orphans, widows and people with cerebral palsy with a subsidy of US$136 per month.
The social assistance component is financed with the Fondo de Desarrollo Social y Asignaciones Familiares, or FODESAF, which allocates resources to 21 institutions in charge of more than 35 social assistance schemes, including the non-contributory scheme previously mentioned. This large number of institutions illustrates the high level of fragmentation that characterise social assistance in Costa Rica, with multiple schemes following similar objectives in terms of protecting people from poverty and deprivation. Established in 1975, the FODESAF is funded with two sources of funding. First, there is a payroll contribution paid by employers that is equivalent to 5% of the total wages1 they disburse. In addition, the existing legislation requires the Ministry of Finance to allocate and transfer to FODESAF the equivalent of 593,000 base salaries, an amount that represented 0.7% of GDP in 20212.
FODESAF allocates around 1.7% of GDP (Pacheco and García, 2022) with five key interventions accounting for two-thirds of that expenditure (Table 8.1). Individually speaking, the main programs are the non-contributory pensions (0.35% of GDP), the Fund for Housing (0.27%), the IMAS Program of Welfare and Family Promotion (0.16%), Avancemos (0.16%) and the School Feeding Program (0.23%). The rest of the funding is addressed to smaller initiatives administered by the Equity Programs of the Ministry of Public Education to avoid/reduce school dropout, the Instituto Mixto de Ayuda Social (IMAS), the Instituto Nacional de la Mujer (INAMU), the Patronato Nacional de la Infancia (PANI), the Ministry of Health and the Consejo Nacional de la Persona con Discapacidad (CONAPDIS), among others.
Source: Pacheco and García (2022)
As it was mentioned before, the non-contributory pension scheme is funded with taxes that are allocated to FODESAF and then to the CCSS. The subsidy is granted to poor elderly people 65 years old or older and other vulnerable groups (widows, orphans, people with cerebral palsy) who receive US$136 per month. By the end of 2020, the scheme benefitted 129,265 persons (FODESAF, 2021). Beneficiaries of this transfer have, on average, 76.5 years with 43.4% living in rural areas and half belonging to zones outside the Central Region. The pension represents, on average, 20% of the income of beneficiary households (Pacheco and García, 2022).
The Subsidy Fund for Housing (FOSUVI, in Spanish) is the second largest programme in terms of spending accounting for around 0.3% of GDP. The Housing Fund averages around US$12,375 and since 1991 it has awarded 404,423 benefits. In total, 18.1% of the Costa Rican households have received the Bond according to information in 2021’s Household Survey. It is important to mention that the subsidy is not limited to poor households but it may be granted to families with gross income up to US$2,785 per month.
Avancemos and Crecemos are two of most important conditional cash transfer programs financed by FODESAF for young populations. These two initiatives, both administered by IMAS, targets persons below 25 years old in Costa Rica and are linked to school attendance and school year completion.
In the case of Avancemos, the programme grants a monetary transfer of ¢30,000 (US$45) per month with the objective of contributing to “the inclusion, permanence, assistance and reintegration into the education system” of secondary school students (aged 12 years or older) belonging to poor or vulnerable families. Crecemos, on the other hand, provides a ¢18,000 (US$27) monthly subsidy to poor children in Early Childhood or primary school levels (11 years old or younger) with the objective of promoting permanence and assistance to the education system (IMAS, 2021).
The last component includes labour market policies. Both the Ministry of Labour and Social protection and the National Vocational Institute (INA) rank as the top two institutions with relevant functions and responsibilities in this category. According to the Ministry of Labour, its key responsibilities in terms of labour market policies and regulations include:
Manage the process of public policy in socio-labour matters, through the management of specific policies and general guidelines to exercise the stewardship and steering responsibilities
Design and implement strategies to promote labour equity, aimed at employers and population groups with special needs.
Supervise and guarantee the adequate application and compliance of labour legislation, including that related to occupational health
Set, review and advise on minimum wages in the private sector.
Direct the design and implementation of actions for the progressive eradication of child labour and protection of adolescent workers, the elimination of the worst forms of child labour, special protection for the worker and his family.
Contribute to the generation of employment by supporting micro-businesses, by financing productive initiatives for entrepreneurs.
The INA, on the other hand, is the institution in charge of training and certifying individuals and companies in those technical areas that promote human talent, according to market conditions and the competitiveness requirements of the business environment (INA, 2021). In this regard, the Training and Vocational Training Services is the most important budgetary programme currently administered by the Institute.
The mission of this program is to promote and develop high-quality Vocational Training and Training Services through their design and execution, aimed at people over 15 years of age and the productive sector, in search of attention to the needs of civil society and the labour market.
While the Ministry receives most of its funding for labour-related policies from general taxation (through the National Budget), the INA is financed with a payroll tax.
Table 8.2 provides an overview of existing social protection programs during the life cycle.
Source: Own elaboration based on ILO and institutional web pages
In terms of coverage, the Costa Rican social protection system performs relatively well, especially when it comes to social security. As can be seen in Table 8.3 below, health insurance ranks at the top of the list. Statutory coverage (Bonnet and Tessier, 2014) grants access to 100% of the population in case of emergency services. Contributory health coverage currently covers about 89% of the population while contributory pension regimes (CCSS and other schemes) cover almost two thirds of the labour force.
While average coverage rates are high, there are significant gaps in coverage for vulnerable groups, most notably the elderly poor, children, persons with disability, and the unemployed. Just 71% of poor elderly persons (over 64 years old) receive either a contributory or a non-contributory pension. Only 29% (127,830 beneficiaries) of poor children between 0 and 17 years old receive a monetary benefit from IMAS; this rate declines to 17.7% of the whole group below 18 years. In short, there are about 313,000 poor children and adolescents without Avancemos transfers. This low rate of coverage is in part due to the fact that the cash transfer program targets only school-aged children, and not children in early childhood.
Regarding other groups, one in three poor persons with disability receives any type of monetary transfers, including pensions. Unemployed persons experience the lowest level of social protection coverage, with only 2% receiving any benefit. Although there is no formal unemployment insurance scheme in the country, salaried people who lose their jobs usually receive some compensation as part of the Labour Capitalisation Fund (LCF) mechanism. The FCL is a savings mechanism, financed with contributions by the employer on behalf of the worker. The final amount is made up of 1.5% of the employee's salary and the respective investment returns. The beneficiary is entitled to request this money in case of resignation, dismissal or retirement.
Source: Estimations based on ENAHO 2024
Beyond their coverage, the extent of financial protection provided by benefits (i.e. the size of the transfers) is another important dimension. Non-contributory pensions provide a high level of financial protection, with an average (US$125 per month) amounting to 76.7% of the average poverty line. However, as there are different poverty lines for urban and rural residents, the level of protection is higher among rural beneficiaries (93.2%) than among urban beneficiaries (71.3%). In the case of contributory pensions, the average amounts were calculated at US$484.6 among urban residents and US$309.3 among rural pensioners (data from Household Survey 2021). Respectively, they account for 3.4 and 1.7 times their corresponding poverty lines.
Childhood-oriented benefits provide lower levels of protection. In the case of Crecemos, aimed at protecting children below 12 years, the average benefit amounts US$30 per month representing between 16.5% (urban) and 21.4% (rural) of the poverty lines. Avancemos, on the other hand, average US$50 per month with sufficiency representing 26.7% and 34.7% of the regional poverty lines. However, it is important to consider that infant-based lines should have a reduced value as several services consumed in the household, like some utilities, experience Economies of Scale. Finally, in the case of healthcare benefits, financial protection in the country is high. According to World Bank Development Indicators, out-of-pocket expenditures is estimated at 22.3% of current health expenditures, lower than the average in Latin America region, where this indicator amount 28.3%. Looking forward the health sector face the challenges of buttressing the financial sustainability of the CCSS, as highlighted in recent actuarial assessment (Pacheco and Itriago, 2022), delivering adequate services, and boosting efficiency and cost control (the country ranks among the bottom four OECD countries in terms of medical doctor productivity, measured as medical visits per doctor).
While benefit coverage and amount are relatively high, there remain substantial targeting problems among social assistance programs. Figure 8.7 shows the distribution of benefits from three cash transfer programs (Avancemos, Crecemos, and non-contributory pensions) between poor and non-poor households, as determined by pre-benefit household income. One-third of non-contributory pensions (RNC) were allocated to non-poor households, and over 40% of the beneficiaries of the Avancemos and Crecemos programs were from non-poor households. In total, resources allocated to non-poor households would amount to 0.59% of GDP out of which 0.16% comes from the non-contributory pensions, 0.26% from Crecemos and 0.17% from Avancemos. There is, consequently, room to improve programmes targeting. A 15% reduction in mistargeting would liberate roughly 5,729 non-contributory pensions, 4,197 scholarships of Avancemos and 7,060 transfers coming from Crecemos.
Source: Author elaboration using ENAHO-INEC 2021.
8.3.2. The financing of Social Protection
Social protection programs in Costa Rica can be categorised into three types based on their financing mechanisms. The first type of program is funded through contributions, such as the health insurance and contributory pension programs managed by CCSS or the FODESAF Fund. The second has legislation linking their financing to an existing source, mainly general taxes. The PANI, for instance, receives 7% of the income tax according to Act 7648 of 1996. The third category has no defined financing, and it is usually funded from the general budget.
The financing of social protection programs in Costa Rica relies heavily on high payroll contributions, which, as discussed in chapter 1 of this book, are barriers to formalisation. Table 8.4 presents contribution rates for each system. In total, contributions account for 37% of wages. Of this burden, employers pay 72% (26.5% of wages) while workers pay the remaining 28% (10.5% of wages). Of total contributions, 64.8% are allocated to health insurance and contributory pension financing (CCSS). FODESAF is the second largest recipient at 13.5% of contributions, while the National Vocational Institute absorbs 4%. Institutions under the umbrella of the Law on Worker´s Protection account for an additional 5.75%. Part of the contributions are also used to finance a state-owned bank.
Source: CCSS
When compared with OECD countries, payroll charges paid by employers in Costa Rica, at 21% of labour costs, are higher than in the average country (Figure 8.8).
Employer contributions, % of labour costs, 2024
Source: OECD Taxing Wages database.
The previous analysis showed that the level of social protection varies considerably by population. In the analysis that follows, we will focus our recommendations on three groups, children, unemployed and the elderly, for different reasons. Children below 15 years currently experience the highest poverty rate in the country (about 40%) while the proportion of elderly people (65 or older) is experiencing the highest rate of growth, compared to all other age groups, so in the course of the next 15-20 years they would represent roughly 20% of the population in Costa Rica. Not keeping the path in terms of group coverage may imply a considerable reduction of elderly people coverage by 2040. Finally, just a very small fraction of unemployed persons receives any type of social protection monetary support.
8.4.1. Universal coverage for children
To provide universal coverage for children, benefits provided by cash transfer programs and option would be to increase the coverage of Avancemos. The expansion would target children between ages 0 and 14, and would provide a monthly benefit of ¢24,000 (US$36) that represents the average of the Avancemos subsidy. If the target group included all children, this would imply an additional fiscal cost equivalent to 0.45% of GDP (Figure 8.9). If only poor children were targeted, then cost would be equivalent to 0.12% of GDP.
% of GDP
Source: Author elaboration based on ENAHO 2024 and FODESAF.
The benefit amount could also be raised to different benchmarks, with varying cost implications. If the transfer amount was increased to the level of the current non-contributory pension (US$124 per month), then the cost would be 1.5% of GDP (Figure 8.10). Several reasons could be outlined for the use of this benchmark. The first is that, being a program already established, it allows us to know the social investment in case of standardizing the benefits between groups. Second, the amount granted to the non-contributory program is the highest of all subsidies granted in mass coverage programs so that it could represent a "ceiling" on how much to allocate to beneficiaries.
Matching the benefit amount with the poverty line (US $127 for urban areas and US $133 in rural areas) would imply a cost of 2.1% of GDP (1.6% of GDP for urban children and 0.5% of GDP for rural children). Matching the benefit amount with the food poverty line (US $78 in urban areas and US $65 in rural areas) would imply a cost of 1.1% of GDP (0.8% of GDP for urban children and 0.3% of GDP for rural children).
% of GDP
Source: Author elaboration based on INEC and FODESAF databases.
Finally, if the former scenarios are estimated for poor children only, then achieving full coverage may cost between 0.32% (benefit equals food poverty line) and 0.65% of GDP (benefit equals poverty line). Additional estimates for poor urban children yield 0.23%-0.48% of GDP while poor rural children may demand between 0.10% and 0.17% of GDP. A benefit equivalent to the non-contributory pension implies 0.45% of GDP in additional funds (Table 8.5).
Source: Author elaboration based on INEC and FODESAF database.
8.4.2. Closing the coverage gap among elderly
Elderly people (aged 65 or older) in Costa Rica experience high levels of social protection coverage in comparison with other countries in the region, with most receiving benefits from contributory pension or non-contributory social transfer systems. The non-contributory transfer that elderly people currently receive, US$124 per month, amounts to 72% of the urban poverty line (93% of the rural poverty line), or 1.6 times the urban food poverty line (1.9 times the rural food poverty line). This is the amount of reference that these estimations will utilise to calculate the costs of increasing coverage.
Still, according to the Household Survey, in 2021 31.8% of persons aged 65 or over (a total of 211,019 people) did not receive either a contributory, a non-contributory pension or a foreign pension. This is the case of people who: 1) never contribute to the CCSS; 2) contribute but did not reach the minimum number of years to qualify for a minimum contributory pension; 3) do not qualify for a non-contributory pension or 4) never applies to non-contributory benefits. In other words, they are fully unprotected and depend on secondary sources or support to maintain a basic consumption. The proposal considered in this section aims to close this coverage gap by expanding the existing non-contributory pension system. This may imply creating new funding that would be transferred to FODESAF and then to the CCSS as administrator of the scheme.
Closing the gap for elderly people using a transfer that would be equal to the non-contributory pension would cost US$ 314.6 million, an amount equivalent to 0.5% of GDP (Figure 8.11). If the exercise uses the poverty line as reference, the additional fiscal cost of 0.70% of GDP out of which 0.57% is linked to urban elderly people while the remaining 0.13% refers to the additional cost for rural old age persons.
% of GDP
Source: Author elaboration based on INEC and FODESAF databases.
If the expansion targets only poor elderly persons without any type of protection, then the target population drops to 50,855 people (36.4% of poor elderly). In this case, the demand for new resources to achieve universal coverage is estimated at 0.13% of GDP in the case of matching the current benefit amount, or 0.17% of GDP in the case of matching the poverty line.
8.4.3. Expanding coverage to unemployed
Unemployment social protection support provides monetary benefits over a determined period of time to unemployed people who are capable of working (Social Protection and Human Rights web page). For the purposes of this exercise, the costing assumes that all unemployed persons receive a transfer during 4 months equivalent to the non-contributory pension. Then, a second exercise assumes that only poor unemployed (43.1% of total unemployed or 112,160 persons) would receive the benefit.
In the first case, the cost of covering all unemployed persons is estimated in US$129.0 million or 0.11% of GDP (Figure 8.12). About 78% of this cost would go to unemployed persons in urban regions. In the case of the scenario in which only poor unemployed receives the transfer, the cost would be US$55.6 million or 0.04% of GDP.
% of GDP
Source: Author elaboration based on INEC.
Alternative scenarios can be constructed based on changes in key of parameters. If the decision is to grant the subsidy during 5 months instead of 4 months, then the incremental funding needs would be US$32.3 million (all unemployed) and US$13.9 million (poor unemployed) in relation to the base scenario. That would represent a 0.05% of GDP increment up to 0.25% in the first case and of 0.02% of GDP (to reach 0.11%) in the second situation.
The previous section described Costa Rica’s social protection system as an extensive, but highly fragmented system, with low social expenditure levels compared to OECD averages, high social insurance coverage, and limited coverage with monetary transfers (non-contributory, social assistance). In addition, social protection financing relies heavily on payroll contributions, which disincentivises formalisation. Indeed, recent debates have pointed to high contribution rates as one of the main causes of the country’s high unemployment and informality rates (Robles, 2022). More empirical research is needed, but initial results show a connection between informality and number of hours worked, which is partly explained by the existence of the Minimum Contributory Rate (Mora-Guerrero, 2020).
Against this background, this section outlines options for shifting financing coming from payroll contributions to other financing sources, as a means to facilitate higher formalisation. While here we focus on restructuring social protection financing, it’s important to note to reduce informality policy actions are also needed in other areas, as outlined in Chapter 1 of this book and in OECD Economic Surveys for Costa Rica (OECD, 2023). This includes simplifying formalisation procedures for firms and workers, enhancing workers’ skills through higher-quality education and training opportunities, and strengthening institutional capacity to ensure broader enforcement of existing regulations.
8.5.1. Options for reducing contributions to incentivise formalisation
This section identifies a series of policy options to reduce the cost of formalisation by shifting tax burden from payroll taxes (or contributions) to general taxes. This section uses the value added tax (VAT) to assess the necessary increase in general taxes that would be needed to make up for the lost revenue of reducing payroll taxes. Importantly, VAT is just one potential policy tool for replacing lost resources as it generates an important volume of revenues. The VAT was introduced in Costa Rica in 2018, by transforming an existing sales tax into a fully-fledged value-added tax. The current VAT rate in Costa Rica is 13%, against a 20% average in the OECD. An increase in the VAT rate is used here as an illustrative option and is not the only possibility. As explored in recent OECD Economic Surveys, there is room to combine increase in the VAT rate with increases in direct taxes (see also Table 8.10).
8.5.2. Reducing employers’ contributions to the CCSS
Total payroll contributions to the CCSS amount to 24% of gross wages (14.5% from the employer and 9.5% from the worker). One option for lowering the barrier to formalisation is to reduce employer contributions. This reduction may apply to the healthcare insurance scheme, to the pension scheme, or to both, and to all employers, or just to employers in specific sectors.
Employer contributions amounted to ¢1.32 billion (US$2,129 million) to health insurance in 2021 and ¢ 495,309.4 million (US$798.9 million) to pensions in 20203. A one percentage point cut in employer contribution rates each to the healthcare and pension systems would result in a reduction in healthcare revenues equivalent to 0.4% of GDP and a reduction in pension revenues equivalent to 0.3% of GDP. Taken together, the total effect of a one-point cut in the contributory rates to the CCSS of the two schemes would be 0.7% of GDP (Figure 8.13).
To counteract the proposed reduction in employer contributions, the VAT would need to be raised by 1 percentage points for health and 0.8 percentage points for pension systems, or a total increase of 1.8%.
Source: Author elaboration based on CGR and CCSS databases.
The main limitation of the former estimate is that it only considers the negative effects in CCSS revenues of reducing the contributory rates. A dynamic model would be required to calculate how affiliation may react after the rate is cut, but this product is beyond the purpose of this document. Instead, some related analysis may illustrate those effects. For instance, León-Murillo (2021) estimated the effects of reducing the contributory rate paid by the employers on labour formality. To proceed, the author simulated a 25%, 50%, 75% and 100% reduction in the corresponding rate. The results can be observed in Table 8.6. For instance, if the rate is set at 19.88% (25% reduction), so the level of labour formalisation would increase from 56% to 61.4%. Subsequent declines motivate increasing gains in formalisation. In the extreme case that no employer rate exists, formalisation would achieve a high 83.3% rate.
Two related outcomes were relevant in the discussion. The first one refers to the estimated decline in social contributions collected by the CCSS. As per the simulation, contributions would fall from 7.5% of GDP to 6.9%, 6.0%, 5.0% and 3.5%, depending on the size of the reduction. Second, although lower rates may have a positive effect on employment, the author considers that additional collection due to new workers may not be sufficient to compensate the reduction in contributions. As a result, the net effect in CCSS revenues is an estimated loss of 0.63% of GDP under the smallest reduction but it may grow up to 3.92% loss if the contribution rate were 0%.
Source: Leon-Murillo (2021)
8.5.3. Reducing contributions for those earning less than 3 minimum wages
A second option is to reduce the social security contributions of workers earning below three times the minimum wage. The minimum wage in Costa Rica was set at US$712.7 per month at the end of 2024, so this would cover workers with wages below US$2,138, which accounts for about 90% of salaried workers (Figure 8.14).
Source: Author elaboration using INEC and Ministry of Labour databases.
This policy option establishes a 2-point reduction (one per scheme) in the social contributions that both employers and salaried workers pay for health insurance and pensions. Empirical evidence estimated using the Household Survey indicates that the CCSS revenues may fall 0.13% per year if the proposed 2-point cut in the corresponding rates is applied. In other words, the joint revenue collection may go from 6.23% to 6.10% of GDP.
By social security scheme and contributor, the results indicate that 38.5% of the reduction (0.11% of GDP) is linked to employer healthcare contributions while an additional 22.9% (0.06%) is due to reduced workers’ contributions to health, too. In other words, the contraction of healthcare-related contributions explains 61.5% of the total fall. Pensions, as expected, would experience a 0.11% of GDP revenue shrinkage.
To fiscally compensate this measure, an increment in VAT of 0.33 points would be sufficient to cover the initial loss. If, instead of VAT, the decision is to increase income tax (that generates, on average, 0.33% of GDP per point), then the increment should be 1 point.
An alternative scenario, in the same line as the former, assumes no contributions applied to all workers earning below the minimum wage. According to the Household Survey, about 674,993 salaried workers received less than that, something equivalent to 39% of all wage-earners. However, if only those salaried persons working 40 hours per week or over are considered, then the number receiving less than the minimum wage drops to 459,277 or 31.7% of all workers with working weeks of 40 hours or more (26.6% of all wage-earners). So this measure may significatively affect CCSS revenues as the proportion of workers fully exempted from contributions is high.
Being the median wage of ¢400,000 (US$776.5 approximately), the level of foregone contributions would amount US$1.38 billion in a year. This result is equivalent to 1.45% of GDP. However, as pension contributions cannot be fully exempted, if only health insurance is 100% subsidized, then the foregone income would be 0.89% of GDP.
8.5.4. Eliminating, totally or partially, the Minimum Contributive Base
A third option for reducing contributions is to eliminate totally, or partially, the minimum contributive base (MCB). Figure 8.15, below, shows the distribution of wages for salaried and independent workers income (in logs) as reported in the 2024 Household Survey. Self-employed workers are the most affected group as around 64% of self-employed workers earn less than the minimum base, as compared to 39% of wage-earners.
Source: Author estimations based on INEC and Ministry of Labour.
The elimination of the minimum base would mean that affiliated workers and employers would contribute according to their actual earnings. So, for instance, all affiliates who today earn less than the MCB and who today pay based on that reference point, would start contributing based on their actual income.
Table 8.7, below, presents calculations for the total income loss that could be expected if the MCB were eliminated for the health system. Income loss is calculated by estimating collected contributions from salaries and independent earnings with and without the MCB, using the 2024 Household Survey.
The full elimination of the MCB would result in a total income loss of US$ 864.9 million per year, equivalent to 0.9% of GDP, with reduced revenues from salaried workers amounting to US$500 million, or 0.53% of GDP, and with lower contributions from self-employed amounting to US$364.9 million, or 0.38% of GDP). Although there are much more salaried workers than independent ones (the ratio is 3.5:1), the gap between average wage and the MCB and the one between independent earnings and the MCB is much wider in the latter. Consequently, although there are much more wage earners, the gap compensates this unbalance on the independent side and for that reason the loss between groups closes. For instance, independent earnings are about two-thirds the median value of wages. According to the ENAHO 2024, the median value of self-employed labour-related income is US$485 per month versus median wages of US$776.5 per month. Those two elements seem to compensate the reduced participation of self-employed among occupied persons and equalise the expected losses due to the elimination of CBM.
Source: Author elaboration based on ENAHO 2024- INEC
Alternatively, the MCB could be reduced to 50% of the minimum wage (US$356 per month). In this case, the associated loss for the health insurance scheme would be 19.3% of total revenue. This would imply a net loss of US$705.8 million, or 0.74% of GDP (Table 8.8). Losses related to independent workers would account for 35.8% of the total losses.
Source: Author elaboration based on ENAHO 2024- INEC
Initial estimations show that full elimination of the CBM can be compensated by a 2.33-point increment in the VAT. The second scenario, in which the CBM is set at 50% of the minimum wage, would demand a 1.89 VAT point adjustment. This last result runs in line with the recently approved reform that reduced the CBM to US$282 for independent workers.
8.5.5. Temporarily reducing payroll contributions of young persons
A fourth option is to temporarily reduce the contributions of young people, who experience the highest unemployment rates of any age group. By IV-2021, youth unemployment (15 to 24 years) was estimated at 16.5% of the young labour force (aged 15 to 35 years) or 26.5% of 15-24 year-olds.
Robles (2022) suggests a temporary reduction of social contributions for young people. The scenario considered in this section assumes a 5-point decline in the rate, from 31.25% to 26.25% (i.e. excluding only contributions to the entities under the Law on Worker´s Protection). The results of the simulation show that revenues coming from young workers’ contributions may fall by 16%. This represents a loss of 0.09% of GDP if the reduction applies to workers aged 15 to 24, and 0.44% of GDP if it applies to workers aged 15 to 35. About 75% of this reduction may finally affect the finances of the CCSS. To balance revenues, VAT would have to increase by between 0.24 and 1.11 percentage points.
Increasing the VAT may not be necessary if other effects occur. In the long term, for instance, the loss in revenues may be compensated by increased formal employment of young people as it is expected that lower contributory rates would motivate employers to hire more young people. A 128,245 person increase (19%) in youth formal employment (ages 15 to 35), would be enough to fully offset the reduction in contributions. If the policy applies only to 15-24 year olds, then just 37,855 workers would need to be added to the formal workforce.
8.5.6. Eliminating specific payroll contributions
A fifth policy option is the elimination of a select few outdated and oversized contributions. Workers contribute 0.25% of wages to the Popular Bank, and 1.5% to the National Vocational Institute (INA). The contribution to the Popular Bank, created in 1969, is outdated, as the bank is already a mature institution integrated into the financial market, and the funds are no longer required for its operation. In 2019 and 2021, the Bank received US$46.2 million (6% of its total revenues) each year from employer contributions. Removing this 0.25% payroll contribution would reduce employer’s contributory rates by less than 1% of the current burden.
In the case of the INA, contributions are oversized, as the Institute reports persistent financial surpluses. Between 2018 and 2021 (excluding pandemic year 2020), the Institute observed annual revenues of US$208 million per year, of which social contributions accounted for 96.2%. Expenditures, on the other hand, amounted to just US$167 million per year, resulting in surpluses to the order of US$24.5 million per year, representing about 20.6% of the received payroll contributions. Cutting contributions to the INA by anywhere between 0.20% to 0.40% of wages4 would reduce these surpluses and have minimal impact on INA operations. This policy option would also require no change to VAT taxes.
8.5.7. Looking for fiscal space for social protection: alternatives
In the above discussion, VAT was suggested as a policy tool for offsetting revenues lost through by reducing social contributions. Official data show that VAT generated 4.6% and 5.1% of GDP in 2020 and 2021, respectively, with 2020 revenues heavily impacted by the COVID-19 pandemic. A comparison with other OECD countries, suggests that Costa Rica has room to increase VAT revenues. However, such a policy may not be politically feasible in the immediate future. A fiscal reform in December 2020 transformed the former Sales Tax to the Value Added Tax. Although the rate remained the same at 13%, one of the lowest in Latin America, the political space to negotiate an incremental rate at this point of time may be difficult to assure.
Alternative options for generating revenues exist. Such options should be explored not only as an alternative to VAT increases to make up for lost revenue, but also as a means to fund the higher expenditures on social protection systems that are needed to increase coverage of children and the elderly. Based on the prior analysis, increased coverage of children and elderly persons would require additional funding of between 0.3%-1% of GDP, depending on the strategy pursued (Table 8.9).
Source: Author elaboration based on INEC and FODESAF databases.
Other financing options include improved expenditure allocation, improving the revenue-enhancing capacity of existing taxes and broadening tax bases by removing tax exemptions or tax evasion. Table 8.10, below, presents a list of options available in the literature. Most options would be sufficient on their own for funding increased coverage of poor children and poor elderly persons. Combining several options would be needed to fully fund universal coverages for all children and elderly persons.
Source: Authors calculations.
Other measures that could increase resources available include improve the targeting of social benefits, as based on 2021 Household Survey, around 30% and 40$ of total benefits go to middle or high-income households. The SINIRUBE, a social registry of beneficiaries of social programmes, could be the central building block to select beneficiaries of social benefits and improve the targeting.
These options are, like VAT increases, not free of limitations. The elimination of existing tax exemptions, for example, would face political challenges, and reducing tax evasion may take several years, as it requires increased institutional capacity.
The current macroeconomic situation implies also some constraints, as the country is undergoing a, a gradual process of fiscal consolidation, which was needed to reduce the public-debt-to GDP ratio from its very high level (OECD, 2022). Hence, the fiscal rule, created under Law No. 9635 on “Strengthening of the public finance”, currently in place establishes limits to future increases in public spending, depending on the level of the level of public debt. One exception is the non-contributory pensions program, currently exempted of the application of the rule.
Moreover, if lower contribution rates motivate higher levels of formalisation, from other perspective, there is a debate of whether the compensation via VAT increment should be a temporary or a permanent measure. If the expected outcome of increasing formality finally materialises, initial losses due to the reduction of contributions may be compensated by new affiliates that were informal workers before. The length of the temporary transfer via increasing VAT may also be a subject of debate.
Over the years, the Costa Rican social protection system has provided a good level of coverage and quality of the services, particularly when it comes to social security. However, the society in which this system operates is changing fast. The population pyramid is shifting, with the proportion of elderly increasing. Informality remains high, affecting around 40% of all workers. Inequality remains high and poverty has remained little unchanged for the past 25 years. In this context, a review and reform of the social protection system is needed with a view to increase its coverage. Social protection coverage and labour informality are interlinked, as high social contributions is a key barrier to formal job creation in Costa Rica. Reforming social protection systems to ensure some basic social protection coverage for all, regardless of whether they work in the formal or informal sector, while simultaneously reducing the cost of formal employment, has significant potential to reduce labour informality and decrease poverty and inequality, all of which are long-standing challenges in Costa Rica.
Achieving universal coverage, even for specific population groups, remains a significant challenge that will require time and sustained effort. Given current constraints, an initial strategy could focus on expanding coverage for the elderly poor and poor children. Our analysis indicates that there is fiscal space to mobilise additional revenues to support this expansion. Crucially, these additional resources should not only increase overall funding but also enable a reduction in payroll contributions. A shift in the financing structure of social protection, towards greater reliance on general taxation and less on payroll contributions, would help promote formal job creation. As detailed and quantified in this chapter, multiple policy options are available to both expand social coverage and raise the necessary revenues through general taxation.
References
Abarca, Alejandro and Fernández, Sergio (2017) Efecto de las pensiones del régimen no contributivo sobre el mercado laboral de las personas y los ingresos por trabajo de los hogares. San José: Departamento de Evaluación, Control y Seguimiento (DECS), Unidad de Evaluación (UEval) de FODESAF, Ministerio de Trabajo y Seguridad Social
Flórez-Estrada Pimentel María. La única solución para muchas personas es la informalidad. Entrevista al economista Pablo Sauma. Semanario Universidad, 17 agosto, 2021. Accesed in the following link: https://semanariouniversidad.com/pais/la-unica-solucion-para-muchas-personas-es-la-informalidad/
INEC (2020) Tasa de desempleo abierto se situó en 12,4 %. In: https://inec.cr/noticias/tasa-desempleo-abierto-se-situo-124
León-Murillo, Jorge (2021) Simulación de los efectos de una reducción en las cargas patronales sobre el empleo formal y la recaudación. San José: Banco Central de Costa Rica.
Ministerio de Trabajo y Seguridad Social (2017) Estrategia Nacional para la Transición a la Economía Formal Implementación de la Recomendación 204 de la OIT en Costa Rica. San José, Costa Rica: Gobierno de la República.
Mora Guerrero, José Carlos (2020) La Informalidad del Empleo en Costa Rica: Trabajo y Justicia Social. Caracterización y recomendaciones. San José: Fundación Friedrich Ebert.
OECD (2017) Tackling Wasteful Spending on Health. OECD Publishing, Paris.
OECD (2016) OECD Economic Surveys: Costa Rica 2016. OECD Publishing, Paris.
OECD (2020) OECD Economic Surveys: Costa Rica 2020. OECD Publishing, Paris.
OECD (2023) OECD Economic Surveys: Costa Rica 2023. OECD Publishing, Paris.
OECD (2021) Government at a Glance 2021. Country Fact Sheet, Costa Rica. OECD Publishing, Paris.
Organización Internacional del Trabajo (ND) El piso de protección social en Costa Rica: Universalizando el derecho a la protección social. San José, Costa Rica.
Pacheco, José and Elizondo, Hazel (2019) Indicador para medir la calidad del empleo en Costa Rica. Ponencia presentada como investigación base para el Informe Estado de la Nación en Desarrollo Humano Sostenible 2019 (no. 25), Capítulo: Equidad e Integración Social. San José, Costa Rica.
Pacheco, José and Itriago, Deborah (2022) Seguimiento a los indicadores de cobertura, calidad y financiamiento de los servicios de salud. San José: Programa Estado de la Nación
Pacheco, Jose and García, Heiddys (2022) Consideraciones sobre Política Social Selectiva. San José: Programa Estado de la Nación
Robles, Edgar (2022) ¿Cómo generar empleo? Access in the following link: https://www.academiaca.or.cr/opinion/como-generar-empleo/
Yip, Winnie and Hafez, Reem (2015) Improving Health System Efficiency: Reforms for improving the efficiency of health systems: lessons from 10 country cases. Geneva, WHO: Health Systems Governance & Financing.
← 1. This includes private employers except those ones with payroll-related expenses below one base salary and agricultural companies with salary expenditures below two base salaries.
← 2. The base salary refers to the salary of a clerk category 1, used by the Judiciary to set fines and penalties. In 2023, this parameter was defined at US$853 per month.
← 3. Latest available figures per scheme.
← 4. These figures were the result of dividing the annual surplus by the income generated per contributory point.