WHAT THE ARTICLE "REESTIMATING BRAZIL'S GDP GROWTH FROM 1900 TO 1980" BY BACHA, TOMBOLO, VERSIANI, TEACHES US ABOUT OUR OWN HISTORY

The article "Reestimating Brazil’s GDP Growth from 1900 to 1980" by Edmar Bacha, Guilherme Tombolo, and Flávio Versiani, accepted for publication in the Revista Brasileira de Economia, undertakes a comprehensive review of Brazil’s economic growth rates over the 20th century, specifically from 1900 to 1980.

During the era when the national accounts were compiled by the Getúlio Vargas Foundation (FGV), from 1947 to 1980, a substantial segment of the services sector—encompassing government, financial, and rental services—was excluded from the national accounts. These sectors represented approximately 30% of the GDP during this period. The growth rate of the excluded GDP—averaging 3.1% per annum, based on documentary evidence from the time—was lower than that of the included components, such as agriculture, industry, trade, and transport. This discrepancy led to an overestimation of economic growth during this era. Similar issues were identified for the earlier period from 1900 to 1946. The paper’s summary indicates that the corrections to the GDP series lead to notable reductions in annual growth rates: from 7.4% to 6.2% for 1947-1980 and from 4.4% to 4.0% for 1900-1947. For the entire period from 1900 to 1980, the suggested adjustments reduce the annual GDP growth rate from 5.7% to 4.9%.

From 1980 onwards, official figures from the Brazilian Institute of Geography and Statistics (IBGE) have included all economic sectors, ensuring accurate measurement of the economy's level in 1980. Consequently, the overestimation of growth between 1900 and 1980 implied an underestimated economic baseline in 1900.

Clear evidence exists that conventional statistics underestimated the level of the Brazilian economy in 1900. According to the latest Maddison database update, Brazil's per capita income in 1900 was 11% of that of the United States. In the same database, Italy's per capita income was 40% and Spain's was 33% of the American level. The significant migration from Italy and Spain to a country considered three to four times poorer appears paradoxical. In 1900, as per the Maddison database, Mexico's income was 23% of the US level, Bolivia’s was 18%, and the rest of Latin America, excluding Brazil, averaged 26%. Brazilian per capita income was slightly lower than India’s, which was 12% of the American per capita GDP (compared to Brazil’s 11%). All data are in 2011 US dollars, adjusted for systematic cost of living differences.

The authors suggested downward adjustments in the average annual growth rates for various periods: by 0.2 percentage points (pp) for 1900-1919; 0.6 pp for 1920-1947; 1 pp for 1948-1966; and 1.5 pp for 1967-1980. Revisiting the Brazilian economy's trajectory up to 1900 and adjusting the rates accordingly, it is inferred that Brazil's per capita income in 1900 was 80% higher than previously estimated. Applying this adjustment to the Maddison database’s series for Brazil, it is concluded that Brazil's income in 1900 was 20% of the US level, not much below the Latin American average.

The principal consequence of the authors' revision is a significant reconfiguration of Brazil's performance during the national developmentalism era (1930-1980), when interventionist policies were adopted to stimulate growth. Current data suggest that Brazil's performance was commendable. According to the Maddison database, only Japan and Romania outpaced Brazil in per capita GDP growth during the five developmentalist decades. Brazil’s per capita GDP grew at 3.7% annually, compared to Romania's 4.3% and Japan's 3.8%. All other 53 countries performed worse.

With some latitude, the column reconstructs the Maddison series for Brazil in the 20th century. The same correction factors calculated from standard data (Ipeadata) were applied to the Brazilian data from the Maddison database. Differences may arise due to deflator variations between series. Under this working hypothesis, corrected data show per capita GDP growth between 1930 and 1980 at 2.7% per annum. This places Brazil 17th in growth rate among 56 countries in the Maddison series for this period. The average growth rate among these countries was 2.2% per year, positioning Brazil’s performance 0.5 pp above the average—a good outcome, but not spectacular.

The following figure illustrates Brazilian per capita GDP from 1900 to 2022, presented on a logarithmic scale for clarity. Values are in 2010 reais. The black line represents the Ipeadata series, and the gray line shows the corrections proposed by Bacha, Tombolo, and Versiani. Growth rates for the periods differ slightly from Maddison's rates. The average growth from 1930 to 1980 was 3.7% annually, revised to 2.7% with the proposed correction. For the Ipeadata series, growth from 1930 to 1980 was 3.9%, revised to 2.9% with the proposed correction.

Visually, the new series retains structural breaks identified in 1918 and 1980. For 1900-1918, the rate of 0.6% per annum was revised to 0.4%. From 1918 to 1980, the rate was 3.9%, revised to 3.0%. The recent period was not revised, with growth at 0.8%.

Finally, the finding that Brazil's economy in 1900 was almost double what was previously thought suggests that 19th-century performance was better than assumed. The same authors have another study reviewing Brazil’s experience from 1820 to 1900.

For the 19th century, the adjustment pertains to GDP inflation, known as the implicit GDP deflator. The utilized statistics overestimate 19th-century inflation. The authors also propose a new series for Brazilian exports in the first half of the 19th century.

According to the Maddison database, Brazilian per capita growth from 1820 to 1900 was 0.03% annually. The proposed revision increases this to 0.7% per annum, aligning with the Latin American average, excluding Brazil. These revised figures are more coherent.

Much remains to be done. The 19th-century text by the authors is undergoing peer review, and the adjustments to the new series for the 20th century in the Maddison database are preliminary. Nonetheless, this reassessment of Brazil’s growth experience in the 19th and 20th centuries is highly valuable.