Human Becoming
The Engineer Who Lives in Both Countries
She drives forty minutes from Escazú to her office in a free-trade zone on the western edge of San José. The office belongs to a Fortune 100 company. The building has floor-to-ceiling glass, a cafeteria that serves sushi on Thursdays, and a gym that opens at six in the morning. She earns $38,000 a year — base, before the annual bonus and the stock options that vest over four years. She is twenty-nine. She studied computer engineering at the University of Costa Rica, graduated near the top of her class, speaks fluent English, and was recruited through a campus pipeline that funnels the top fifteen percent of STEM graduates directly into multinational hiring processes.
Her mother works at a soda — a small family restaurant — in Desamparados, twelve kilometers east. Her mother earns roughly 350,000 colones a month, which is about $640. This is not unusual. This is the national median income. Her mother has worked at the soda for nineteen years. The soda does not offer health insurance beyond the mandatory Caja contribution. There is no bonus. There are no stock options. There is no gym.
The daughter earns approximately five times what the mother earns. They live in the same metropolitan area. They shop at some of the same stores, though the daughter now shops more frequently at Auto Mercado, where imported products cost what imported products cost. They vote in the same elections. They breathe the same air, which in the Central Valley during dry season carries particulate matter from diesel buses that the multinational employees do not ride.
The daughter knows she is lucky. She uses the word suerte, which means luck, though she also knows it was not purely luck — it was four years of engineering school, two years of English classes her family paid for with difficulty, and a recruitment system that identified her as convertible labor. She fits the profile. Her mother does not fit the profile. Her mother was never going to fit the profile. The profile was not designed with her mother in mind.
This is not a parable. This is a data point. And it repeats, with variations, 195,000 times across Costa Rica's multinational sector.
Structural Read
The Architecture of the Divide
In 2025, CINDE — Costa Rica's Investment Promotion Agency — reported 67 new foreign direct investment projects, continuing a trajectory that has made the country the most successful FDI destination per capita in Latin America.[1] The multinational sector now employs approximately 195,000 people directly, concentrated in free-trade zones across the Greater Metropolitan Area. The number has grown every year for the past decade. The companies are real. The jobs are real. The money is real.
Intel, which first established operations in Costa Rica in 1997 and briefly scaled down its assembly operations in 2014, has reversed course dramatically. The company expanded its Costa Rica workforce to 2,400 employees in 2025 and committed $150 million to expanding its design and testing operations in the country.[2] Intel's Costa Rica site now handles advanced chip design work, not assembly — a shift that repositions the country in the semiconductor value chain from manufacturing floor to engineering center. With the CHIPS Act reshaping global semiconductor geography, Costa Rica is projected to reach $6.2 billion in semiconductor-related exports by 2027, according to CINDE's investment pipeline data.[3]
Amazon Web Services trains approximately 2,000 Costa Rican professionals per year through its re/Start and Skills to Jobs programs, channeling graduates into cloud architecture roles that pay between $25,000 and $45,000 annually.[4] IBM has certified 850 professionals in cybersecurity through its SkillsBuild partnership with local universities, creating a talent pipeline that feeds directly into its regional security operations center in Heredia.[5] Microsoft, Accenture, Procter & Gamble, HP, and Amazon all operate significant shared-services or technology centers. Twenty-nine of the Fortune 100 now have operations in Costa Rica.[1]
The tech workforce specifically — defined as professionals in software development, data science, cybersecurity, cloud infrastructure, and related fields — reached 74,000 in 2025, a 13% year-over-year increase. Salaries in this segment range from $20,000 for junior developers to $45,000 for senior engineers and architects. The national median income is approximately $15,000.[6]
The numbers describe a success story. Costa Rica bet on education, political stability, and strategic geographic positioning. It won the bet. The multinationals came. They kept coming. They are still coming.
But numbers describe conditions. They do not describe what happens inside the condition. What happens inside is the signal.
The Two-Speed Economy
The OECD, which admitted Costa Rica as a member in 2021, published its 2025 Economic Survey of the country with a finding that deserves quoting directly: the domestic economy, which accounts for approximately 85% of national production, is stagnating — characterized by low productivity growth, limited innovation, and weak integration with the multinational sector.[7]
This is not a peripheral observation. This is the central structural diagnosis. Costa Rica does not have one economy with a successful sector. Costa Rica has two economies operating in parallel, with minimal linkage between them.
The multinational economy operates in free-trade zones that offer tax exemptions, streamlined regulation, and infrastructure that the domestic economy does not receive. Companies in free-trade zones pay no income tax for the first eight years and 50% of the statutory rate for the following four. They import equipment duty-free. They operate under labor regulations that are technically identical to the domestic economy but practically different — because the multinationals offer salaries, benefits, and career development structures that make their workforce legally voluntary but economically captive. You do not leave a $35,000 job with stock options to start a soda.
The domestic economy operates under a different set of constraints. Small and medium enterprises, which account for the vast majority of non-multinational employment, face a tax burden that the free-trade zone companies do not share. They compete for labor against entities that can offer three to five times the prevailing wage. They lack access to the training pipelines that AWS, IBM, and Intel have built for their own hiring needs. The result is predictable: the best-educated, most English-proficient, most technically capable segment of the workforce is systematically extracted from the domestic economy and absorbed into the multinational sector.
The Gini coefficient — the standard measure of income inequality — stands at 45.8 for Costa Rica, higher than the OECD average of 31.5 and among the highest in Latin America.[7] This number has not improved meaningfully in a decade. The FDI success has not compressed the income distribution. It has stretched it.
The Geography of Partition
The multinational operations are concentrated in a corridor that runs from Escazú through western San José to Heredia and Alajuela. This is where the free-trade zones are. This is where the airport is. This is where the roads work. This is where property prices have risen 15-20% in the past three years, driven by multinational employee purchasing power and a growing expatriate population drawn by Costa Rica's digital nomad visa program.[8]
Twelve kilometers east of this corridor, in Desamparados and Alajuelita and La Unión, the infrastructure degrades visibly. The roads narrow. The bus system — which remains the primary transportation for the 85% of the workforce that does not work in free-trade zones — runs on schedules that have not been updated to reflect actual commute patterns. A domestic worker living in the eastern districts of San José faces a ninety-minute bus commute to reach the free-trade zones in the west, assuming the bus arrives. The multinational employees drive. Or they take the company shuttle.
The geographic partition maps onto the economic partition. The western corridor has fiber-optic internet, reliable electricity, water pressure that does not drop during dry season, and streets that are paved and maintained. The eastern and southern districts have intermittent service across all four dimensions. The free-trade zones are, functionally, infrastructure islands — zones where the Costa Rica that investors see actually exists, surrounded by the Costa Rica that investors do not visit.
This is not new. What is new is the acceleration. The nearshoring wave triggered by COVID-19 supply chain disruptions, the U.S.-China trade realignment, and the CHIPS Act has compressed a decade of gradual investment growth into three years of rapid expansion. Between 2022 and 2025, Costa Rica attracted more FDI projects than in the entire previous decade combined. Each project deepens the divide — not because the investment is harmful, but because the investment is structurally isolated from the domestic economy it nominally inhabits.
The Education Bottleneck
Costa Rica's public education system produced the workforce that attracted the multinationals in the first place. The country's literacy rate is 98%. Its public universities — the University of Costa Rica, the Instituto Tecnológico de Costa Rica, the Universidad Nacional — are among the best in Central America. The investment in human capital over decades is what made the CINDE pitch credible: come to Costa Rica, we have the people.
But the pipeline has a filter, and the filter is English.
Multinational hiring in Costa Rica requires functional English proficiency. This is non-negotiable for virtually every position above entry-level manufacturing. The public education system teaches English, but the quality varies dramatically by school and region. Students in the western San José metropolitan area — where the best-funded schools are, where the parents who work at multinationals live — receive English instruction that produces functional proficiency by graduation. Students in rural areas and lower-income urban districts receive English instruction that does not.
The result is a sorting mechanism that operates before the labor market begins. The children of multinational employees attend schools (or private institutions) that prepare them for multinational employment. The children of domestic economy workers attend schools that prepare them for the domestic economy. The system reproduces itself. The bridge that was supposed to connect the two economies — public education — has become the mechanism that separates them.
CINDE and the government have launched programs to address this. The Empleo para Vos initiative targets workers outside the free-trade zone corridor. INA, the national training institute, has expanded its English and technical training programs. AWS and IBM's certification partnerships are designed to create alternative entry points. These programs are real. They produce real graduates. But the scale is insufficient. AWS trains 2,000 per year. IBM certifies 850. The domestic economy employs over two million people. The gap between the pipeline and the need is not a rounding error. It is a structural mismatch.
Pattern Confirmation
The Template That Repeats
Costa Rica's two-speed economy is not unique. It is the clearest example of a pattern that is emerging across every successful nearshoring destination in Latin America and Southeast Asia.
Vietnam's semiconductor and electronics manufacturing zones around Ho Chi Minh City exhibit the same partition: multinational wages running three to four times the domestic median, infrastructure concentrated in export processing zones, a workforce filter based on language proficiency (English or Japanese), and a domestic economy that accounts for the vast majority of employment but receives a minority of the investment benefit.[9] Mexico's nearshoring corridor from Monterrey through Guadalajara shows identical dynamics: Tesla, BMW, and hundreds of auto-parts suppliers operating in a high-wage, high-infrastructure envelope while the surrounding domestic economy stagnates at lower productivity levels.
The pattern has five structural components, and Costa Rica manifests all five:
First, the tax asymmetry. Free-trade zone companies receive fiscal incentives that domestic companies do not. This is rational as attraction policy. It is irrational as development policy. The domestic economy subsidizes the infrastructure and education that produces the multinational workforce, but does not share in the fiscal revenue that the multinational operations generate. Costa Rica's free-trade zone tax exemptions cost the treasury an estimated $800 million per year in foregone revenue.[7] That money does not build roads in Desamparados.
Second, the labor market siphon. Multinational wages are set by global benchmarks, not local market conditions. When Intel pays a design engineer $40,000 in a country where the median income is $15,000, it does not raise wages across the economy. It creates a parallel wage structure that the domestic economy cannot match. The domestic sector loses its best talent not to competition but to a compensation framework that operates in a different currency of possibility.
Third, the supplier linkage failure. The OECD's 2025 survey found that local supplier integration with multinational operations remains "limited." Most multinationals in Costa Rica source inputs from global supply chains, not from domestic producers. The free-trade zone architecture incentivizes this: duty-free imports for zone companies means there is no cost advantage to sourcing locally. The multinationals exist in the geography but not in the economy.[7]
Fourth, the infrastructure divergence. Public investment follows the multinational footprint. Roads to the airport are maintained. Roads to the free-trade zones are maintained. Roads to the neighborhoods where the 85% live are maintained when the budget allows, which is inconsistently. The multinationals did not cause this divergence. But their presence determines where the divergence concentrates.
Fifth, the educational sorting. The skills required by the multinational sector — English, STEM, digital literacy — are produced by the education system at a rate sufficient to fill multinational hiring needs but insufficient to transform the domestic economy. The education system does not fail. It succeeds — for the fifteen percent. The other eighty-five percent receive an education calibrated to an economy that is not growing.
This five-factor model is not a Costa Rican problem. It is a structural feature of FDI-led development when the receiving economy does not build integration mechanisms alongside attraction mechanisms. Costa Rica built the best attraction mechanism in Latin America. The integration mechanism remains a policy aspiration.
The Semiconductor Question
The stakes are about to increase. Costa Rica's positioning in the global semiconductor supply chain — driven by Intel's expansion, the CHIPS Act reallocation, and CINDE's strategic focus on chip design and advanced testing — is projecting the country toward $6.2 billion in semiconductor-related exports by 2027.[3] If this projection holds, semiconductors will surpass medical devices as Costa Rica's largest export category.
Semiconductor operations are more capital-intensive, more technically demanding, and more narrowly staffed than shared-services centers or medical device assembly. The workforce requirements are steeper: advanced degrees in electrical engineering, materials science, or physics. The English requirement is absolute. The salary premiums are larger. The geographic concentration is tighter — semiconductor operations cluster around existing Intel facilities and the infrastructure that supports them.
If Costa Rica becomes a significant node in the global semiconductor supply chain — and the trajectory suggests it will — the two-speed dynamic does not moderate. It intensifies. A semiconductor engineer in Heredia will earn not three times the national median but five or six times. The infrastructure investment required to support chip fabrication or advanced testing will flow to the facilities, not to the neighborhoods. The educational filter will narrow further: from "speaks English and has a technical degree" to "speaks English, has an advanced degree in a hard science, and can pass Intel's security clearance process."
The semiconductor opportunity is real. The semiconductor opportunity also accelerates every structural partition that already exists.
Alternative Explanations
It is possible that the two-speed economy is a transitional phase rather than a structural condition — that the multinational sector will eventually generate sufficient spillover effects to raise domestic productivity. This is the classic FDI development thesis: foreign investment brings technology transfer, management practices, and supplier development that gradually upgrades the domestic economy. This thesis has empirical support in cases like South Korea, Taiwan, and Ireland, where FDI-led development eventually produced broad-based economic transformation.
The counterargument is timing. South Korea's transformation took forty years and involved massive state-directed industrial policy, including forced technology transfer requirements that Costa Rica's free-trade agreements prohibit. Ireland's transformation required EU membership, structural funds, and a population of five million — roughly Costa Rica's size, but with access to a 450-million-person single market. Costa Rica has neither the industrial policy tools nor the market access that made those transformations possible. The question is not whether spillover can occur. It is whether spillover can occur under the specific structural conditions that Costa Rica's free-trade zone model creates — and the OECD's 2025 data suggests it is not occurring at meaningful scale.
A second counterargument holds that inequality is acceptable as long as absolute living standards are rising. Under this reading, the mother earning $640 per month is better off than she would be without the multinational sector, because the FDI inflows support government revenue (through indirect taxes and social contributions even if direct corporate taxes are exempted), maintain macroeconomic stability, and generate construction and service employment around the free-trade zones. This argument has partial merit: Costa Rica's poverty rate, while stubbornly high at approximately 23%, would likely be higher without the multinational sector. But "better than the counterfactual" is a low bar for a country that markets itself as Latin America's most stable and equitable democracy.
What is not known: The exact degree of local supplier integration. CINDE reports that local procurement by multinationals has increased, but does not publish granular data on the percentage of multinational inputs sourced domestically versus imported. Without this data, the supplier linkage failure cannot be precisely quantified, only structurally inferred from the OECD's assessment.
What is not confirmed: Whether the new semiconductor investments will create broader workforce pipelines than existing tech operations. Intel's public commitments include community training programs and university partnerships, but the scale and effectiveness of these programs relative to the hiring need is not independently verified.
What would change the signal: If Costa Rica implements mandatory local-content requirements for free-trade zone companies — requiring a minimum percentage of domestic supplier procurement — the integration mechanism would shift from voluntary to structural. This would be politically difficult given existing free-trade agreements. If the Gini coefficient begins declining while FDI continues growing, the spillover thesis gains support. If the Gini remains flat or increases, the structural partition thesis is confirmed.
Monitoring indicators: Track CINDE FDI project count quarterly. Track OECD productivity data for the domestic versus multinational sector annually. Monitor Gini coefficient trends from INEC (national statistics institute). Track semiconductor export volume against the $6.2 billion projection. Monitor Intel workforce count and local training program enrollment. Track property price differentials between the western corridor and eastern/southern districts of the Greater Metropolitan Area.
[1] CINDE Costa Rica, "2025 Annual Investment Report: FDI Projects and Employment," 2025. cinde.org — Tier A
[2] Intel Corporation, "Intel Expands Costa Rica Operations with $150M Investment," press release, 2025. intel.com — Tier A
[3] CINDE Costa Rica, "Semiconductor Sector Strategy: Export Projections 2025-2027," investment pipeline briefing, 2025. cinde.org — Tier B
[4] Amazon Web Services, "AWS re/Start and Skills to Jobs: Costa Rica Program Impact," 2025. aws.amazon.com — Tier B
[5] IBM, "SkillsBuild Costa Rica: Cybersecurity Certification Program Results," 2025. skillsbuild.org — Tier B
[6] Costa Rica tech salary data aggregated from Glassdoor, LinkedIn Salary Insights, and CINDE sector reports, 2025. — Tier C
[7] OECD, "OECD Economic Surveys: Costa Rica 2025," Paris, 2025. oecd.org — Tier A
[8] Property price data aggregated from Colliers International Costa Rica, Newmark Central America, and local real estate market reports, 2024-2025. — Tier C
[9] World Bank, "Vietnam Development Report: Economic Transformation and Dual Economy Dynamics," 2024. worldbank.org — Tier B
[10] INEC Costa Rica (Instituto Nacional de Estadística y Censos), "Encuesta Nacional de Hogares 2025: Indicadores de Desigualdad." inec.cr — Tier D