Baumol's Law in Brazil: Why We Migrated Directly From Agriculture to Low Productivity Services

Between 1950 and 1980, Brazil was one of the fastest growing economies on the entire planet. Average productivity per Brazilian worker jumped decade after decade. Afterwards, starting in the 1980s and even following currency stabilization with the Real Plan in 1994, something disconcerting happened: Brazilian agribusiness soared in technological efficiency, several cutting edge companies modernized, but average aggregate productivity per worked hour in Brazil remained practically at a standstill for thirty years.

How is it possible that a country has satellite connected farms, one of the most digitized banking systems in the world (Pix, Open Finance), and an aerospace giant like Embraer, and yet exhibits stagnant average labor productivity in the historical series of FGV IBRE?

The explanation involves one of the most elegant concepts in economic science, formulated in 1966 by North American economist William Baumol (1922 to 2017): Baumol's Law (also called Baumol's Cost Disease), combined with the phenomenon of early structural change in Brazil.

1. Beethoven's quartet and the two sectors of the economy

In his classic study, William Baumol divided the economy into two fundamental types of activities:

1. The progressive sector (technologically scalable): includes manufacturing, mechanized agriculture, chip manufacturing, and software development. Here, the introduction of machines, fertilizers, or algorithms causes a single worker to produce ten or fifty times more per hour today than they did fifty years ago.
2. The stagnant sector (intensive in human presence time): Baumol used as an example a string quartet playing a 40 minute piece by Beethoven. In 1826, it took four musicians working for 40 minutes (2.66 man-hours) to deliver the product. In 2026, it still takes the exact same four musicians and the exact same 40 minutes. If you try to "increase productivity" by placing only two musicians or by playing the song at double speed in 20 minutes, you destroy the service.

Into the same category as the string quartet fall dozens of traditional urban services: cutting hair at a barbershop, serving a restaurant table, caring for an elderly person, cleaning a building, providing on-site security, or performing an artisanal consultation without technological support. Physical productivity per hour in these activities grows very little over the decades.

2. The Brazilian structural error: skipping the industrial stage

What does this have to do with Brazil's history? Everything.

Between 1950 and 1980, about half of the Brazilian population left the countryside (where they practiced manual, extremely low productivity, almost subsistence agriculture) and migrated to cities. Even if a worker leaving the hoe in the interior went to work in simple civil construction or a basic textile factory in São Paulo, their productivity multiplied by four or five times instantly. Economists call this the structural change bonus.

In the 1990s and 2000s, however, the scenario changed. Brazilian agribusiness mechanized intensively, began producing record harvests with increasingly less labor in the field, and maintained productivity growth worthy of South Korea (above 3% to 4% per year).

Yet, where did the tens of millions of workers freed up by the countryside and the young people entering the urban market go? In South Korea, Taiwan, and Poland, they were absorbed by a high productivity export industry and sophisticated business services. In Brazil, due to the Brazil Cost, tax chaos (which punished industry with cumulative taxes), and low educational levels (72% below the math minimum in PISA), we experienced what Harvard economist Dani Rodrik calls premature deindustrialization.

Our workforce migrated en masse directly to traditional and informal low productivity urban services: small street businesses, domestic services, manual deliveries, informal repairs, and repetitive administrative roles.

3. How to break Baumol's curse in services and companies

If about 70% of Brazilian GDP and employment are already in the services sector, Brazil will not get rich by trying to recreate 1970s bolt factories. The only way to unlock national per capita income is to radically increase productivity within the services sector itself.

And here we reach the great historical opportunity of the 2020s and 2030s: for the first time since Baumol wrote his article in 1966, the combination of mobile connectivity, digital platforms, and generative Artificial Intelligence is transforming previously "stagnant" services into scalable activities:

1. In education and training: an adaptive platform (as seen in Uruguay's Plan Ceibal or in AI-based tutors) allows a great teacher to monitor the personalized learning of hundreds of students, identifying exactly where each young person got stuck in mathematics.
2. In health and diagnosis: clinical decision support systems, interoperable medical records, and telemedicine do not replace the human care of a doctor, but they cut in half the time spent on bureaucracy, triage, and repeated exams.
3. In corporate services (legal, accounting, customer service, and backoffice): in Brazil, millions of office workers function as "bureaucratic string quartets," copying data from a PDF to a spreadsheet at the exact same speed as thirty years ago. Automating these workflows with software engineering and AI agents is the direct path to defeating Baumol's Law inside your company.

Continue deepening your knowledge: also read The case of Chilean salmon: why copying a validated method comes before inventing and 2041: the year the Brazilian population starts to shrink. Also explore the Brazil 2050 Agenda with an interactive simulator and Brasília Arcade.

---

References and verifiable readings

• Baumol, W. J. (1967). Macroeconomics of Unbalanced Growth: The Anatomy of Urban Crisis. The American Economic Review, 57(3), 415-426.
• FGV IBRE (2025). Regis Bonelli Productivity Observatory: Productivity by Economic Sector in Brazil.
• Rodrik, D. (2016). Premature Deindustrialization. Journal of Economic Growth, 21(1), 1-33.