Brazil Data Book

Last update: September 2025

This page contains information about some of the data available in the FEWS NET Data Explorer (FDE) for Brazil. This is not a comprehensive guide.

For information about using the filters and fields for specific domains in the FDE, see Choose a Data Domain.

Summary table

Country food security context

Brazil has faced a decline in food security since 2015, with high food inflation, a weakening labor market, and the dismantling of social protection programs contributing to rising food insecurity and hunger, particularly affecting vulnerable regions and populations. Despite a significant decrease in severe food insecurity in 2023, overall progress remains challenged by factors like more frequent and severe extreme weather events, deforestation, and policies favoring large agribusiness over small-scale farmers.

Brazil is a global agricultural powerhouse, ranking among the world's largest exporters of various commodities like soybeans, coffee, sugar, and beef. Earnings from this sector are a fundamental pillar of the Brazilian economy, contributing substantially to the country's GDP, employment, and foreign exchange earnings. The country’s dynamic large-scale corporate agricultural sector benefits from a policy environment which favors commercial farms and exports, but the smaller family farming sector remains crucial for employment. The entire sector is increasingly vulnerable to extreme weather events like droughts and heatwaves, deforestation, and a decline in soil fertility, which pose both short- and long-term risks to its continuing viability. 

Constant high food inflation has reduced household purchasing power, making food less accessible for many. Vulnerable population groups and the North and Northeast regions of the country generally experience higher rates of food insecurity than others. Economic crises and a politically unstable environment have led to increased unemployment, declining real wages and a dismantling of key social protection programs, like the Food Acquisition Program (PAA).

Statistical reporting units

The first three levels of the actual administrative unit hierarchical structure of the country include the following entities, in this order:

Admin 1: States

Admin 2: Municipios

Admin 3: Districts

Municipio-level statistical reporting units

Brazil, and specifically the Brazilian Institute of Geography and Statistics (Instituto Brasileiro de Geografia e Estatística, IBGE), uses a combination of administrative units (states, municipios) and IGBE-defined statistical regions (meso-region, micro-region) as a national 4-level hierarchical structure for municipio-level statistical reporting as follows:

• State

• Meso-Region

• Micro-Region

• Municipio

FEWS NET has adopted this hybrid statistical-reporting structure to organize Brazil’s crop statistics. 

Crop data

Explore our crop data.

View our documentation on using the Crop Domain.

The Brazilian crop dataset is the largest in the FEWS NET Crop domain, and the largest dataset of any included in the FEWS NET Data Warehouse. For the 1974-2023 period, there are more than 3.5 million crop results, or about 14 million datapoints.

Crop estimate data sources

As of 1938¹, Brazilian municipio-level crop statistics have been collected, managed and disseminated by the Brazilian Institute of Geography and Statistics (Instituto Brasileiro de Geografia e Estatística, IBGE). Municipality crop statistics can be found as a well-documented long-duration time-series referred to as the Produção Agrícola Municipal (PAM) dataset at IGBE’s Sistema IBGE de Recuperação Automática – SIDRA.

These data are organized and downloadable in the structure of annual tables (tabela) for which a user may specify desired features of a subset of crop data. Table 1612 in the PAM data is a record of all temporary crops (planted and harvested within an annual cycle).  

Since 2003, a small number of crops have also been broken out into other tables to provide seasonal data. FEWS NET uses Table 839 for seasonal maize (milho) results, Table 1000 for seasonal common beans (feijão), Table 1001 for seasonal potatoes (batata-doce) and Table 1002 for seasonal groundnut (amendoim) data.

Crop reporting units

As described above, FEWS NET uses the IBGE municipio-level statistical reporting unit structure to organize its crop statistics into a Crop Region “R” type set of boundaries. The structure includes the following hierarchical entities:

Crop Region 1: 26 states (unidad federativa, UF) and one federal district (distrito federal) serve as the first level in Brazil’s crop statistics reporting. The federative units of Brazil are subnational entities with some autonomy (self-government, self-regulation, and self-collection) and their own government and constitution, which together form the Federative Republic of Brazil. The states are generally based on historical, conventional borders which have developed over time.

Crop Region 2: Meso-geographic regions (Meso-Região Geográfica), of which there are 133 as of 2024, serve as the second level in Brazil’s crop reporting. Although the Municipality (see Crop Region 4, below) is the country’s true second-level administrative unit, FEWS NET uses an IBGE convention in which this geographic region is identified as the second level unit in its crop statistics, because it helpfully groups municipalities found in large geographically diverse “States” into relatively homogenous sub-state areas. The Meso-geographic Region has no administrative purpose.

Crop Region 3: Micro-geographic regions (Micro-Região Geográfica), of which there are 554 as of 2024, serve as the third level in the country’s crop statistics reporting hierarchy. Within each Meso-geographic Region, Micro-geographic Regions further divide the area into relatively homogenous sub-Meso-geographic Region areas. The Micro-geographic Region has no administrative purpose.

Crop Region 4: Municipality (Municipio). The country is divided into more than 5,570 municipalities which serve as the fourth, and currently final, reporting level for the country’s crop statistics. There are no municipalities in the Federal District, which functions both as a state and a municipality. Municipalities can be subdivided, only for administrative purposes, into districts and sub-districts. The IBGE maintains an annual list of municipalities over time and a list of municipio name changes between 1939-present.

A future Crop Region 5?  Consistent with the crop region hierarchy described above, FEWS NET crop data users may find a need in the future for a Crop Region 5 to capture District-level crop data from the administrative unit immediately below the municipio.  Although there is no regular reporting mechanism for district-level crop statistics, many districts do enumerate their crop production and those results are presented in the standard national and annual municipio-level reporting as municipio crop statistics for municipios that have yet to be officially created. As such, their production data must either be added to their parent municipio or treated as an un-official municipio.  FEWS NET treats them as a municipio, identifiable by their FEWS NET ID (FNID, see the example below) which incorporates an identifier for the year of their subsequent creation as a municipio.

The FEWS NET Crop data Zone field

Given the large size of the country, and depending upon user needs, there is an additional frequently-used method Brazil uses to divide the country into a small number of relatively homogenous zones. It has created the Grande Região, (Grand region), a concept that divides the country into a total of 5 regions (Norte/North, Nordeste/NorthEast, Centro-Oeste/Center-West², Sudeste/SouthEast and Sul/South). These large regions have been defined upon agro-ecological and other features which lend a relatively homogenous internal character to each individually. Grand regions have no administrative purpose. FEWS NET crop statistic data carry the Brazilian Grande Região notation to identify the relative location of each municipio in this framework. 

Crop data reporting unit evolution

The IBGE-sourced PAM crop statistics data in the FEWS Data Warehouse cover a 50-year period (1974 to 2023), and relate to the municipios, and in some cases the districts, which existed in Brazil in each of those years. The number of reporting units increased from roughly 4,000 in 1974 to 5,571 in 2023. 

In order to compare Brazil’s area planted, production, and yield over time at a sub-national level (generally, municipio), with agro-climatic and other socio-economic data each year, it is important to associate the crop statistic for each unit with the unit’s location and surface area each year. To do this, FEWS NET compiles a record of the changes over time in the shape and size of all reporting units and records this information in a relationship table. The relationship table contains the name of each municipio-level reporting unit, its place in the national hierarchy of reporting units, and any changes in the unit’s shape or location that occur due, usually, to splitting (cutting one unit into two, or more), or merging (combining two reporting units into one). Additionally, it is important to record any name changes that are applied over time.

The Brazil 1974-2023 relationship table contains approximately 134,000 rows. A row includes the name, year, and FEWS NET identification code (FNID) of a reporting unit in a previous year (from-unit), and the name, year, and FNID of that same, or altered, reporting unit in the following year (to-unit), along with other associated identifying information. 

The FEWS NET FNID has a standard structure, as shown in the example below.  Learn more about FNIDs.

Because there are numerous municipios with the same name, they are identified in FEWS NET crop statistics with both their official name and, in parentheses following the name, the two-character alpha code (there are also numeric UF codes) representing the state (unidad federativa or UF) they are located in.

It is important to know that the IBGE also permanently identifies each municipio with its own 7-digit numerical code which consists of the State (UF) code, and its own 5-digit municipio code, in which the code of Chapada da Natividade (TO) is 1705102. The FEWS NET Relationship Table also records the IBGE code for each municipio.

For each year in which there are changes of shape in any municipio, a new annual set of FNIDs is created for all, reflected in the year code portion of the FNID. Each reporting unit in the Relationship Table will therefore contain in as many rows as there were annual FNID sets in the 1974-2023 period. 

For years in which there are no changes to any municipios, a new annual set is not created. There were no changes in the number and/or shape of any municipios in Brazil in 1984, 2002, 2004 to 2012, and 2014 to 2022, which means the maximum number of annual boundary sets for any municipio would be 30 (50 minus 20) for the 1974-2023 period. Notably, every one of the 30 annual boundary sets for Brazil was the result of a split of a pre-existing municipio – there were no merges or redistributions in this period. Learn more about FEWS NET Geographic Unit Relationships.

While there are many cases in Brazilian administrative history of municipios being created in a certain year and then being reversed back to a district status in subsequent years, there is no case in the crop statistics period where crop production moves back and forth between the new and the old municipio. Although there are multiple moments and events which occur to mark the creation of a new municipio in Brazil, FEWS NET uses the commonly reported date of a municipio’s emancipation or emancipação, as reported primarily by the IBGE, as the date by which the municipio came into being. Municipio emancipation dates are given for all 5,571 municipios. Those emancipation dates for the era before 1974 have not been confirmed in the same way that post-1974 dates have been, and there are likely errors and anomalies in them which will not affect the crop statistics. 

Post-1974 emancipation dates are also useful in determining whether a municipio was submitting crop statistics before it became a municipio (i.e. because it was a district at the time), a situation which is actually quite common in the 1974 to 2023 period. If that is the case, then those statistics might lengthen the record of a municipio’s yield and production over time, but it may also mean that there were other districts at the time whose statistics would have to be added together to get an accurate picture of the putative municipio’s total production.

During the 1974-2023 period, the IBGE recorded 122 municipio name changes (see Annex A), consisting of either a minor change in spelling, or an entire change of name. Note that there are a small number of municipios (marked with an asterisk: *) for which IGBE crop statistics still use the old municipio name, requiring an edit to match the FDW’s standard municipio names and codes.

New Brazilian states that were added/changed in the 1974-1988 period include:

• Amapá, 1988, previously a state territory

• Roraima, 1988, previously a state territory  

• Tocantins, 1988, previously a part of Goiás

• Rondônia, 1982, previously a state territory

• Mato Grosso do Sul 1977, previously a part of Mato Grosso.

In the Brazil Relationship Table, the crop statistics for municipios which previously belonged to Goiás and were then moved in 1988 to Tocantins are coded as Tocantins municipios over the entire 1974-2023 time period, and do not duplicate any Goiás results during the 1974-1987 period. The same procedure has been applied to the 1974-1987 Mato Gross do Sul municipio crop statistics. In the other three cases, the reported crop statistics simply show up in the reporting (with the conversion of an unincorporated territory to a federal state) and were never associated with any other municipio.   

Year and season definition

The timing of planting and harvests varies widely in this very large country between states, crops, micro-climate zones, and production systems. 

Year definition

The annual crop statistic reporting period is identified by IBGE as January 1 through December 31. Not infrequently, crops will be referred to with a two-year attribution (e.g. 2017/2018), because they may be planted in the last 2-3 months of the first year and are mostly harvested in the second of the two years. Because almost all Brazilian crops are harvested between January and September, the single calendar-year attributed reporting (e.g. 2018) does not introduce ambiguity about annual crop production figures. 

Note, however, that many descriptions of the annual crop cycle in Brazil describe it informally as generally beginning in late September and ending in early September of the following year.

Seasons

Meteorological seasons are opposite to seasons in the Northern Hemisphere: 

• Spring (Primavera): September 21-December 20

• Summer (Verão): December 21-March 20

• Autumn (Outono): March 21-June 20

• Winter (Inverno): June 21-September 20

However, because the timing of planting and harvests varies widely, defining crop season names and dates for Brazilian crop statistics is an exercise for which FEWS NET’s normal practices are stretched to their limit. 

Prior to 2003, IBGE PAM crop statistics follow the calendar year, describing the product of harvests that occur in a 12-month period between January 1 and December 31. Beginning with 2003, the IBGE began reporting seasonal crop results for:

• maize (milho)

• sweet potatoes (batata-doce)

• beans (feijão)

• groundnuts (amendoim)

These seasons are defined using the term safra, which is best translated as harvest/colheita or crop, and indicate the sequence of the specified harvest if there are multiple in a year. Thus, there is a first crop/harvest (primeira safra) and there may be a second (segunda). Since 2019, the IBGE has also recognized a third harvest (terceira) in the North and Northeast.  

First, second, and third refer only to a harvest’s sequence among multiple annual harvests for a crop and therefore are not true season descriptions. Indeed, the timing of each of these harvests varies greatly from one state to another, one crop to another, and even one production system to another.  

The IBGE does not provide a finite description of the time period relative to the first, second³, and third crop harvests, but other authoritative sources offer state-by-state planting and harvesting dates for the maize safras. Although not stated, it appears that the other crops for which IBGE provides seasonal results may also consist of multiple, sequential harvests with generally similar timing to the maize crops in each area.

Other crops (wheat, cotton, etc.) are harvested only once per year. While the variation in the planting and harvesting dates is still notable, it is easier to provide a rough guide of the timing of their seasonal growth that is generally applicable for the whole country, even if IBGE identifies their production only as occurring within the calendar year.

The challenge for documenting Brazilian crop production in the FDW is finding a happy medium in how much seasonal timing information to attach to each crop in each state. Too much may suggest, in effect, a false precision and complicate user understanding of seasonal patterns in Brazil. Too little may mislead users who want to broadly compare crop phenological stages and outcomes with earth observation measurements.

Primary crops

Municipality crop statistics include 31 crops considered temporary (planted and harvested in the same 12 months) and 33 considered permanent (planted crops remain alive for more than 12 months and are harvested more than one time over that same period). FEWS NET generally collects only temporary crop statistics for Brazil. Other metadata is available on the IBGE metadata site.  

Pineapples, sugarcane, castor beans, and cassava are considered temporary cultures of a long duration. They commonly exhibit a vegetative cycle of more than 12 months. Their quantity of production reflects only what was harvested in the calendar year and their area planted reflects the portion of the total area planted which was harvested in the calendar year.

Crop names, in Brazilian Portuguese and English, and FEWS NET crop codes are in the following table:

Crop statistics context

Brazil’s Municipality-level crop statistics are available from 1974 to the present, though there may be a one- to two-year lag in publishing the most recent results. Municipios which do not report at least one crop are not reported in the PAM at all.

Notable changes in crop reporting for Brazil include:

• Area planted began to be reported in 1988.

• Before 2001, the quantity produced and the yield of all fruits was by number of fruits. Pineapple (abacaxi), and coconuts (coco-da-baía) are still reported by the number of fruits, not metric tons.

• Triticale and sunflower seed data were first included in the PAM dataset in 2005.

Methodology

There is extensive documentation of IBGE crop estimation methodologies on the IBGE Biblioteca site and in the FDW document management system in the Brazil crop cabinet document.

Annex A: Municipio Name Changes (1974-2023)

¹ Previously managed by the Ministry of Agriculture, Livestock and Food Supply ( https://www.gov.br/agricultura/pt-br/internacional/english)

² Often referred to as Center-South in the US. 

³ Also widely-known as the “Safrinha”, or little harvest, which has become the largest of the three in recent years.