1
Economic Commission for Europe
Conference of European Statisticians
Group of Experts on Measuring Poverty and Inequality
Geneva, Switzerland, 28-29 November 2023
Agenda item: Inflation and its impact on poverty and inequality
The Cost of Food for those with Food Insecurity:
The Brazilian Case1
Note by IBGE, Brazil
Leonardo dos Santos Oliveira, IBGE, Luciana Alves dos Santos,
IBGE, Ana Luiza Neves de H. Barbosa, IPEA
Abstract
Since the beginning of the 2000s, the Brazilian Household Food Insecurity
measurement Scale (EBIA) has been the official measure of household food
insecurity (FI) in Brazil. But it was only in 2017-2018 that EBIA was included in the
National Household Budget Survey (Pesquisa de Orçamentos Familiares – POF)
which collects data on household’s expenditure on goods and services (food
expenditure, in particular).
The main objective of this paper is to identify the food costs of the vulnerable
populations at risk of food insecurity (FI) in Brazil. The methodology is based on the
construction of corresponding spatial price indexes obtained from POF conducted in
1 IBGE and IPEA is exempt from any responsibility related to the opinions, information, data and concepts stated in
this article that are of exclusive responsibility of the authors.
2 The authors would like to thank CNPQ - National Council for Scientific and Technological Development for the
financial support and granting of scholarships in the project "Prices, consumption and demand of food products by
processing level in Brazil: evolution and projections", which gave rise to this article
Working paper 3
Distr.: General
19 November 2023
English
2
2017-2018, which collected data from different Brazilian geographical areas. It is
worth noting that Brazil also does not evaluate official spatial price indexes which
specifies differences in the cost of living among different Brazilian regions.
Following the EBIA, we identify the vulnerable population as one that is at risk of
mild, moderate, and severe food insecurity.
Our study points to relevant disparities in price indexes between the regions of the
country for this already population vulnerable to food insecurity that represents 60%
of the Brazilian population. Among more than 40 food products selected for product
price analysis, chicken was the product that recorded the highest average monthly
household expenditure, with the Metropolitan Region of São Paulo being the
geographic context that presented the greatest positive variation in relation to Brazil,
0,5% above. The average household expenses with products classified as fresh or
minimally processed represented 55.5% of the total expenses for Brazil. Next was
the expenses for ultra-processed foods, 26.3%. The share of processed foods and
processed culinary ingredients was 13.4% and 4.8%, respectively.
To the best of our knowledge, this is the first time that is possible to investigate,
simultaneously, data based on food expenditure and on food insecurity in the same
survey. This study also offers a food regional price index for both, the whole
population and the vulnerable one. Finally, these indexes can be used in future
studies to provide information for public policies on poverty.
Keywords: Food Insecurity, Poverty, Social Vulnerability, spatial price index
JEL: I32, D63, C43, C01, C50, I38, D12
ANPEC ÁREA 12 – Economia Social e Demografia Econômica
1. INTRODUCTION
Identifying which population subgroups are eligible to a public policy is always one of the
main issues and challenges when a project of social-economical format is elaborated. Such a
challenge includes the study of profile of each group. When working with the population suffering
from social vulnerability and the risk of food insecurity or even hunger, this dilemma becomes more
challenging. The usual poverty statistics, based on relatively low lines, for example, does not seem
suitable for the identification of families facing this vulnerability state because they focus more on
extreme poverty and hunger than on vulnerability itself. Moreover, vulnerable families can transit
among the poverty states over time, being poor in one moment and not poor in another moment;
they can also coexist with the risks without becoming poor. They can have an income above the
poverty line but live with legitimate concerns and doubts related to the maintenance of income
and the capacity to buy appropriate food. Information like the food basket of this vulnerable part
of the population, what is the minimum income to pay for this basket are questions that would help
identify this target audience more efficiently.
It is based on these questions that this article suggests the calculation of a regional price index
built from the food basket acquired by families vulnerable to food insecurity. To identify these
families, the results of the Brazilian Scale of Food Insecurity – EBIA will be used.
EBIA identifies the families suffering of severe food insecurity (disruption of eating patterns
among the residents in the household), moderate (quantitative reduction of food among adults and
children) or mild (uncertainty related to the access to food in the short term). Previously, EBIA was
studied using surveys not related to the mapping of the family budget3. In 2017-2018, it was
incorporated into the Brazilian Household Budget Survey (POF) enabling for the first time
the identification of the food basket of families in moderate or severe food insecurity.
Another aspect that is considered regarding the food basket of the vulnerable population is
the nutritional quality of the products selected. The analysis was made with the use of the
NOVA classification that divides food according to the extent and the purpose of the industrial
processing they were exposed to before they are acquired by the individuals (MONTEIRO et al.,
2010, 2018,
• 3 Oliveira, Leonardo (2017) La medición de la inseguridad alimentaria y los indicadores no monetarios en el Sistema de Encuestas de Hogares IBGE, Brasil," Seminarios y
Conferencias 44098 Chapter: XVII Publisher: Naciones Unidas Comisión Económica para América Latina y el Caribe (CEPAL) - LC/TS.2017/149
2019a). The NOVA classification is composed of four groups: i) Natural or minimally processed
foods, ii) processed culinary ingredients, iii) processed foods and iv) ultra-processed foods.
Since the vulnerable population is already identified, as well as the food basket that represents
this group, and the items were classified according to the level of processing, the next step is to
measure the price variations of these products regionally. The regional price variations of similar
goods are even more significant in developing countries once the integrated distribution system
implies higher transportation costs. Given the Brazilian territorial dimensions, these discrepancies
become even more significant. As a result, the price differences among the areas can be higher both
in relative and absolute terms. Nowadays, IBGE (the Brazilian official institute of statistics and
geography) does not calculate spatial price indexes that indicate (even in approximate rates) the
differences in the cost of living in different geographical contexts or in a way that the values of income
and consumption are reviewed according to the variations of regional prices. Although the release of
IBGE (2008)4 had made use of spatial price indexes of foods to map inequality and poverty in the
Brazilian municipalities.
Therefore, this article aims to calculate spatial deflators of foods according to the NOVA
processing level. The building of the price indexes that will be presented adopt five methodological
steps: (1) Selection of a subgroup of the population; (2) Classification of products by processing level;
(3) Selection of foods available in all geographical contexts; (4) Definition of implicit prices and
average amounts of each context; (5) Definition and calculation of price indexes.
In addition to this introduction, this article has five more sections. The second one explains
the EBIA methodology and then how it is possible to identify the vulnerable population from the
classification of food insecurity. The third section presents the classification of foods according to
the processing level using NOVA. The fourth section explains the method used for the definition of
the food basket by geographical area and the calculation of implicit price by product and average
amount by region. A descriptive analysis of expenses amounts and prices of products that compose
the basket is also made. The fifth section presents the results of spatial deflators by processing level.
In conclusion, the final considerations are made.
2. FOOD INSECURITY AND THE VULNERABLE POPULATION
One of the main issues faced in the study of family vulnerability is the definition of the target
audience. That is also impacted by monetary and non-monetary factors that influence the budget or
reveal other information about the living conditions of the population. Collecting the family income
is not a simple task and it involves the identification of different monetary and non-monetary
components. Monetary factors are not always capable of representing the true status of the pattern of
family life since non-monetary acquisitions of goods and services are relevant components in
consumption and income. Thus, non-monetary income has an important participation in the
composition of the family budget, especially of the population with lower income, and in Brazil this
percentage reaches around 9%. POF collects the information of non-monetary acquisitions in a
systematic way and counts on a broad survey of monetary components of income. Additionally, other
non-monetary factors that are not part of the income are covered by POF. This section explains how
it is possible to use measures of Food and Nutrition Security (SAN) to determine the target population.
2.1. The Brazilian Scale of Food Insecurity
The limitation of the family income or other monetary indicators used to identify families at
risk of Food Insecurity (IA) led to the development of a direct scale to measure IA and Hunger by the
United States Department of Agriculture – USDA (BICKEL et al., 2000). This assessment tool of
SAN at household level is suitable for the elaboration of a diagnosis of the condition of food security
and the indication of populations at higher insecurity risk, also helping to observe the impact of public
policies on the circumstances where the population has access to adequate food.
* Research analyst of the Brazilian Institute of Geography and Statistics (IBGE) and Scholarship Holder of CNPQ
4 “MAPA de pobreza e desigualdade: municípios brasileiros 2003. Rio de Janeiro: IBGE, 2008”. 1 DVD. Available at: https://biblioteca.ibge.gov. br/index.php/biblioteca-
catalogo?view=detalhes&id=241385. Accessed in: Nov., 2021.
The Brazilian Scale of Food Insecurity - EBIA is a psychometric scale of the family access to
food, based on the design of a quantitative measure scale of 14 questions that covers both the
perception of the concern with a future food insufficiency and the problems related to the number of
available calories, as well as the quality of the family diet (IBGE, 2006). An advantage of the use of
psychometric scales is that they measure the phenomenon directly from the IA experience lived and
noticed by the affected people. As a result, they capture not only the difficulty in having access to
foods but also the psychosocial dimension of IA considering the households as unit of analysis.
Besides, they can be adapted – with the use of qualitative methodologies – to different local
sociocultural contexts and their application and analysis are relatively simple (PÉREZ-ESCAMILLA;
SEGALL-CORRÊA, 2008).
The direct measure scales of IA, such as EBIA, provide essential information for the
management of policies and social programs because they allow both the identification and
quantification of social groups at risk of IA in relation to their determinants and consequences.
Considering the perception of the experience of a household in the last 90 days, EBIA indicates one
of the following levels of IA experienced by the families (IBGE, 2020):
Frame 1: Description of the levels of food security and insecurity
Food security situation Description
Food safety
The family/household has regular and permanent access to quality food, in
sufficient quantity, without compromising access to other essential needs
Mild food insecurity
Concern or uncertainty about access to food in the future; inadequate food quality
resulting from strategies that aim not to compromise food quantity
Moderate food insecurity
Quantitative reduction of food among adults and/or disruption in eating patterns
resulting from lack of food among adults
Severe food insecurity
Quantitative reduction of food also among children, that is, disruption in eating
patterns resulting from lack of food among all residents, including children. In this
situation, hunger becomes an experience at home
Search: Pesquisa de Orçamentos Familiares – POF / IBGE, 2017-2018
The first time EBIA was applied in Brazilian household surveys was in 2004 in the National
Household Sample Survey – PNAD. Later, it was studied again in PNADs in 2009 and 2013 and in
the National Survey of Demographics and Health of the Child and the Woman – PNDS, both
elaborated by the Brazilian Institute of Geography and Statistics – IBGE. The results obtained in these
surveys confirm IA is directly related to socioeconomic factors as well as of factors that compose the
household unit such as, for example, the presence of residents under the age of 18, the number of
residents, the gender or race of the reference person in the family, and the household income.
In 2017, EBIA started to be collected through the Household Expenditure Survey – POF, edition
2017-2018, also elaborated by IBGE. It was noticed that when the application of EBIA is transferred
to a survey that captures food acquisition and analyzes the living conditions of families the
possibilities of analysis are amplified.
The analysis of EBIA is based “on the sum” of affirmative answers of 14 aspects of the
questionnaire5, classified according to the cut-off points demonstrated in Table 1.
Table 1: Cut-off points for households, with and without residents under the age of 18, according to the status of
food security
Food security situation
Cut off points for households
With people under 18 No people under 18
Food safety 0 0
Mild food insecurity 1 - 5 1 - 3
Moderate food insecurity 6 - 9 4 - 5
Severe food insecurity 10 - 14 6 - 8
Search: Pesquisa de Orçamentos Familiares – POF / IBGE, 2017-2018
5 IBGE, Pesquisa de Orçamentos Familiares 2017-2018 Análise da Segurança alimentar no Brasil, p.24., 2020.
2.2. The selection of the population from the level of Food Insecurity
After the clarification on the methodology and the importance of EBIA, this section will show
how to identify the target audience for the building of a spatial prices deflator that identifies the
regional discrepancies among the populations more vulnerable to food insecurity.
According to IBGE (2020), the proportion of the Brazilian population suffering from severe,
moderate or mild IA is of 41%, severe or moderate IA is of 13.9% and severe IA is of 5%. However,
the distribution of this population along hundredths of income varies significantly, as well as the level
of IA in which it is inserted, as shown in Figure 1. In this chart is calculated the proportion of people
in IA for each hundredth of income. It is clear that as the percentile group of income grow all the
groups of people in IA tend to be zero, which reinforces that although the income is not capable of
identifying the population in IA with accuracy, it is responsible for keeping people away from this
situation. Only in percentile group 60, with per capita income close to R$1200, the probability of a
person to be in Severe or Moderate IA is under 10%. For a status of Severe IA this percentile group
is far below, indicating that almost 15% of the population is at risk of IA above 10%.
Figure 1: Proportion of people in food insecurity by percentile group of per capita income – Brazil –2017-2018
Search: Pesquisa de Orçamentos Familiares – POF / IBGE, 2017-2018
By observing Figure 2 it is possible to notice how people vulnerable to IA are concentrated
along the income distribution. Thus, 90% of the people in Severe IA are among the 60% with lower
income, while for people with Severe or Moderate IA this percentage is of almost 90%.
Figure 2: Concentration curves by type of food insecurity – Brazil –2017-2018
Search: Pesquisa de Orçamentos Familiares – POF / IBGE, 2017-2018
In order to predict the probability of IA, a logit model was estimated for each depending on
variable, that is, the levels of IA: Severe Insecurity, Severe or Moderate Insecurity and Severe,
Moderate or Mild Insecurity. The independent variables (factors) used were: twentieths of income,
number of residents in the family, gender of the reference person in the family, color or race of the
reference person in the family, composition of the family (family structures formed only by adults,
adults and children, only elderly people and remaining possibilities), location of the household (urban
area / rural area and the Federation Units (UF: the 26 Brazilian states and the Federal District). Figure
3 shows the result of the probability of Severe or Moderate IA estimated for each person. There is a
big variability in the predictions but a clear trend of reduction for this probability according to the
increase of percentile group of income.
As a result of the variability of predictions observed in Figure 3, Figure 4 shows the propensity
of people with probability of 20% or more and the proportion of people with probability equal to or
greater than 10% to present a better idea of how this variability moves along the percentile group of
income. Clearly, the drop is sharper in the beginning of the distribution, in the line that represents the
people with probability of 20% or more. However, for people with chances equal to or greater than
10% this drop is less marked. The black line represents the average of all the estimated probabilities
in each percentile group, that is, the average risk of the is around 10% or less around percentile group
60.
Figure 3: Estimated probability of Food Insecurity with the Logit model – Brazil –2017-2018
Search: Pesquisa de Orçamentos Familiares – POF / IBGE, 2017-2018
Figure 4: Probability of people in Food Insecurity according to predictions of the Logit model, by hundredths of
per capita income.
Search: Pesquisa de Orçamentos Familiares – POF / IBGE, 2017-2018
Figure 5 shows how the predictions for the risk of IA is concentrated in the population according
to the per capita income. It is noticed that 60% of the population with lower per capita income
accumulate more than 90% of the cases with risk of Severe or Moderate IA of 20% or more. For the
group of people with chances of 10% or more of Severe or Moderate IA the concentration is of nearly
90% in percentile 60.
Considering the results of charts 3, 4 and 5, it is possible to take the operational concept that
60% of the population with lower income is the population vulnerable to the risk of IA, which meets
the objectives of this article to identify the cost of a representative food basket to these people, and
they will be designated onward as population vulnerable to IA.
Figure 5: Concentration curves of predictions of Food Insecurity – Brazil – 2017-2018
Search: Pesquisa de Orçamentos Familiares – POF / IBGE, 2017-2018
3. NOVA CLASSIFICATION
NOVA is a classification of foods based on the extent and purpose of industrial processing
developed by Monteiro et al (2010) that results in four groups: 1) Natural or minimally processed
foods, 2) processed culinary ingredients, 3) processed foods and 4) ultra-processed foods. This
classification is internationally recognized and has been extensively used in epidemiological studies
on food consumption, quality of diet and health conditions of individuals (MONTEIRO et al., 2019b),
as well as basis for food guidelines of several countries, including Brazil (BRAZIL, 2014).
Natural foods, Group 1, are those obtained directly from plants or animals (such as leaves and
fruits or eggs and milk) and acquired for consumption without any change after leaving nature. The
acquisition of Natural foods is limited to some varieties, such as fruits, vegetables, roots, tubers, and
eggs. Minimally processed foods are Natural foods that were subjected to processes like removal of
inedible or unwanted parts, drying, dehydration etc. Most of these processes aim to increase the
durability of Natural foods, allowing extended storage.
Group 2, the Processed Culinary Ingredients are substances extracted directly from foods of
Group 1 or from nature and they are usually consumed as items of culinary preparations. The
processes involved in the extraction of these substances include pressing, grinding, milling,
pulverization, drying and refinery. The purpose of processing is the manufacturing of products used
to season and cook Natural or minimally processed foods and, in general, for culinary preparations
based on these foods.
The third group that refers to Processed Foods is characterized by products manufactured with
the addition of salt or sugar and possibly oil, fat, vinegar, or other substance of Group 2 to a food of
Group 1, and most of the products have two or three ingredients at most. The processes involved in
the manufacturing of these products can include different methods of cooking and, in the case of
cheese and bread, non-alcoholic fermentation. The purpose of the processing underlying the
manufacturing of processed foods is to increase the durability of Natural or minimally processed
foods or to modify their flavor being, therefore, similar to the purpose of the processing employed in
the manufacturing of foods from Group 1.
The Ultra-processed Foods compose Group 4, which includes products manufactured with
several ingredients and involving, in addition to substances of Group 2 (such as salt, sugar, oil and
fat), substances also extracted directly from foods of Group 1 but not usually used in culinary
preparations (such as casein, whey, soya protein and other foods isolate and hydrolyzed proteins),
substances synthesized from food constituents (such as hydrogenated or inter-esterified oil, modified
starch and other substances not naturally present in foods) and additives used with cosmetic function
to modify the organoleptic characteristics of the products (colour, smell, taste or texture). Several
industrial techniques are used in the manufacturing of ultra-processed products, including extrusion,
molding and pre-processing by frying.
4. BUILDING METHODOLOGY OF SPATIAL PRICE INDEXES FOR BRAZIL
The building of price indexes that will be presented follow five methodological steps: (1)
Selection of a subgroup of the population; (2) Classification of products by level of processing; (3)
Selection of foods available in all geographical contexts; (4) Definition of implicit prices and average
amounts of each context; (5) Definition and calculation of price indexes.
The first step, the selection of a subgroup of the population, was made in section 2, when the
target audience of our study was defined, the population vulnerable to food insecurity. The second
step refers to the classification of foods by level of processing that was made in the previous section.
Thus, this section will present the way of selecting the food basket acquired by the target population
(step 3) and the definition of implicit prices and average amounts of each context (step 4). Having the
food basket defined, a brief analysis of the expense values and the average amounts of products in
Brazil and in the geographical contexts is also made.
Proceeding with the steps to build the spatial deflator, the third step is the identification of the
food products that are common to all geographical contexts. The geographical contexts are formed
by the stratification of the Great Regions in metropolitan areas, urban areas (except the metropolitan
regions) and rural areas. The decision to use the geographical contexts was made due to the existence
of differences in regional prices, according to the research made in previous articles (IBGE, 2008;
Oliveira et al 2016, 2017) and because of the possibility of using the same stratification in the other
editions of POF (2008-2009 and 2002-2003).
The twenty geographical contexts used are mutually excluding and they were created for the
following areas: Metropolitan Urban Regions - MUA (Belém, Fortaleza, Recife, Salvador, Belo
Horizonte, Rio de Janeiro, São Paulo, Curitiba and Porto Alegre); and the Federal District; Non-
metropolitan Urban Area and Rural Area of each of the five Brazilian Great regions.
In order to build a representative price deflator for all the Brazilian vulnerable population, there
was a selection of the foods items acquired for household consumption registered by POF 2017-2018
that were common to all geographical contexts. The basket obtained is composed of 191 food items
that are listed in Appendix 1. Having the definition of the basket and the information related to the
expense value and the amount of each item, one reaches step 4 and it is possible to calculate the
implicit price (Pij) by product that is obtained by the ratio of the total expenditure with the product
divided by the total amount of the item acquired in the respective geographical context, as shown in
Equation (1):
𝑃𝑖𝑗 =
∑ 𝑉𝑛𝑖𝑗𝑛
∑ 𝑞𝑛𝑖𝑗𝑛
=
𝑇𝑜𝑡𝑎𝑙 𝑒𝑥𝑝𝑒𝑛𝑑𝑖𝑡𝑢𝑟𝑒 𝑜𝑛 𝑡ℎ𝑒 𝑝𝑟𝑜𝑑𝑢𝑐𝑡 𝑖 𝑖𝑛 𝑐𝑜𝑛𝑡𝑒𝑥𝑡 𝑗
𝑇𝑜𝑡𝑎𝑙 𝑎𝑚𝑜𝑢𝑛𝑡 𝑜𝑓 𝑝𝑟𝑜𝑑𝑢𝑐𝑡 𝑖 𝑖𝑛 𝑐𝑜𝑛𝑡𝑒𝑥𝑡 𝑗
(1)
Another measure that can be obtained from the information of items that compose the basket is
the average amount acquired by families (Qij), which is the result of the ratio of total expenditure
divided by the product acquired in the corresponding geographical context and the total of families
(UCs) in each geographical context with food expenditure registered for the household, as shown in
Equation 2:
𝑄𝑖𝑗 =
∑ 𝑞𝑛𝑖𝑗𝑛
∑ 𝑈𝐶𝑛𝑗𝑛
=
𝑇𝑜𝑡𝑎𝑙 𝑒𝑥𝑝𝑒𝑛𝑑𝑖𝑡𝑢𝑟𝑒 𝑜𝑛 𝑝𝑟𝑜𝑑𝑢𝑐𝑡 𝑖 𝑖𝑛 𝑐𝑜𝑛𝑡𝑒𝑥𝑡 𝑗
𝑇𝑜𝑡𝑎𝑙 𝑓𝑎𝑚𝑖𝑙𝑖𝑒𝑠 𝑖𝑛 𝑐𝑜𝑛𝑡𝑒𝑥𝑡 𝑗
(2)
Similarly, to Equations (1) and (2) the implicit prices and the average amounts are calculated
for Brazil.
4.1. ANALYSIS OF THE FOOD BASKET OF THE POPULATION VULNERABLE (CAPV) TO
FOOD INSECURITY
Using the food basked elaborated with the products acquired by the vulnerable population in
all the geographical contexts, it is possible to calculate the expenditure, the average amount and the
implicit prices of each product in Brazil and in each geographical area.
The basket of food item that is common to all vulnerable families in Brazil is composed of 191
types of foods acquired for consumption of the family in the household, combined in 68 groups, and
the average monthly expenditure is of R$ 324.38 and the monthly average amount is of 62.57 kg (the
amount of all the products was standardized in the unit of measure kilogram (kg), according to the
POF publication). Table 2 presents a list of 20 products that registered the largest monthly average
expenditure in the basket of food item acquired by the vulnerable population in Brazil, as well as the
expenditure value, the monthly average amount, and the corresponding implicit price.
Table 2: Average and quantity monthly family expenditure and implicit price of the 20 items with the highest
average monthly family expenses that make up the food basket of the population vulnerable to food insecurity -
Brazil - 2017-2018
Ranking Selected products
Average monthly
family expense
Average monthly
family amount
Implicit
price
Processing level
1º Chicken meat 32.06 4.040 7.93
Natural ou minimally processed
food
2º Bread roll 24.90 3.441 7.24 Processed foods
3º Second Category Beef 18.89 1.433 13.18
Natural ou minimally processed
food
4º Rice 18.54 7.268 2.55
Natural ou minimally processed
food
5º First Class Beef 17.57 0.955 18.41
Natural ou minimally processed
food
6º Other Beef Meat 13.38 0.859 15.59
Natural ou minimally processed
food
7º Coffee 12.73 0.800 15.92
Natural ou minimally processed
food
8º Milk 12.65 5.199 2.43
Natural ou minimally processed
food
9º Chicken's egg 8.65 1.007 8.59
Natural ou minimally processed
food
10º Sugar 8.14 3.917 2.08 Culinary preparations based
11º Soda 7.88 2.860 2.76 Ultra-processed foods
12º Sausage 7.87 0.654 12.04 Ultra-processed foods
13º oils 7.31 1.668 4.38 Culinary preparations based
14º Sweet cookie 7.16 0.617 11.61 Ultra-processed foods
15º Powdered milk 7.15 0.38 18.68
Natural ou minimally processed
food
16º crackers and snacks 7.03 0.622 11.31 Ultra-processed foods
17º
Fermented Alcoholic
Beverages
5.70 0.896 6.37 Processed foods
18º Tomato 5.26 1.15 4.58
Natural ou minimally processed
food
19º
Other Tropical Climate
Fruits
5.21 2.02 2.58
Natural ou minimally processed
food
20º Bean 5.16 1.42 3.62
Natural ou minimally processed
food
Search: Pesquisa de Orçamentos Familiares – POF / IBGE, 2017-2018
Considering the 20 food products acquired with the largest monthly average expenditure in the
household, five of them are related to a type of meat: chicken, beef of the second and first categories,
other beef cuts and sausage. Chicken (that includes the other cuts and the bowels) was the food
product that registered the largest monthly average expenditure by families, with the value of R$32.06
and implicit price of R$7.93. Beef of the second category was the third food product with the largest
expenditure, while beef of the first category was only the fifth. It is worth noticing the difference in
the amount acquired of these products. If on the one hand chicken had an average acquisition of 4 kg
per month, the amount acquired of meet of the second category drops to 1.4kg and the amount of beef
of the first category is less than 1 kg.
The traditional products of the Brazilian breakfast, French bread, coffee, and milk are also
among the products with largest expenditures, respectively: R$24.90, R$12.73 and R$12.65. In
average, the Brazilian families acquire around 3.5kg of French bread per month and 5.2 kg of milk.
It is curious to notice that the famous rice and beans is no longer so present in the food basket
of the country. Rice is still a significant product in the household budget, being the fourth product
with largest expenditure (R$18.54) and average amount of 7.3 kg. On the other hand, bean is on the
twentieth position in terms of expenditure (R$ 5.16) and average acquisition of only 1.4 kg per month.
Products of low nutrient content such as soft drinks, biscuits and fermented alcoholic drinks (beer,
for example) consume more of the family budget than beans.
An analysis of the food basket of Table 2 showing the level of food processing demonstrates
that most products acquired with larger expenditure are still the minimally processed or Natural, that
is, the healthiest ones, such as chicken, rice, beefs, coffee, and milk. However, the ultra-processed
foods have a strong presence in the composition of the Brazilian food basket being the second most
important. The acquisition of sweet and salted biscuits demonstrates that not always the option to
acquire ultra-processed products is because they have lower value. These products have an implicit
price of R$ 11.61 and R$ 11.31, respectively, a value that is higher than those of other products like
milk, chicken, chicken eggs, fruits and with little difference in comparison with beef of the second
category, for example.
In order to dimension the regional differences of price, Table 3 shows the price index of the
product for the 10 products with the largest average expenditures for Brazil and according to the
geographical contexts. The price index of the product is the ratio of product price i of geographical
context j divided by the product price i in Brazil, and as a result it is possible to see how the
geographical area is above or below the average Brazilian price.
Table 3: Product price index of the 10 items with the highest average family expenses, according to the products
selected from the food basket of the population vulnerable to food insecurity, by geographic context, 2017-2018
Geographical Context
Chicken
meat
Bread
roll
Second
Category
Beef
Rice
First
Class
Beef
Other
Beef
Meat
Coffee Milk
Chicken's
egg
Sugar
Brazil 1.00 1.00 1.00 1.00 1.00 1.00 1.00 1.00 1.00 1.00
Metropolitan urban area of
Belém 1.10 1.05 1.03 1.14 0.86 1.00 0.95 1.54 1.06 1.12
Urban North excluding
metropolitan urban areas 0.96 0.90 0.92 0.99 0.91 0.95 0.96 1.11 1.17 1.13
Rural North 0.98 0.95 0.87 1.05 0.85 0.83 1.07 1.00 1.21 1.16
Metropolitan urban area of
Fortaleza 1.05 0.89 1.01 1.55 0.99 1.16 1.03 1.28 1.21 1.12
Metropolitan urban area of
Recife 1.02 0.85 1.00 1.15 0.95 0.93 1.16 1.14 0.97 0.94
Metropolitan urban area of
Salvador 1.00 0.79 0.98 1.14 0.99 1.03 1.03 1.20 0.88 1.07
Urban Northeast excluding
metropolitan urban areas 1.04 0.82 1.02 1.05 1.00 1.05 0.95 1.17 1.01 1.06
Rural Northeast 1.05 0.84 0.99 1.03 0.96 0.95 0.96 1.13 1.05 1.08
Metropolitan urban area of Belo
Horizonte 0.91 1.40 1.09 0.99 1.08 0.99 1.01 0.90 1.10 0.91
Metropolitan urban area of Rio
de Janeiro 1.00 0.86 0.94 0.96 1.00 0.95 1.10 1.11 0.80 1.07
Metropolitan urban area of São
Paulo 1.07 1.22 1.16 0.98 1.21 1.18 0.88 0.96 0.93 1.07
Urban Southeast excluding
metropolitan urban areas 1.01 1.25 1.07 0.97 1.11 1.04 1.05 0.97 0.93 0.89
Rural Southeast 0.98 1.16 1.04 0.95 1.00 1.07 1.05 0.94 1.13 0.88
Metropolitan urban area of
Curitiba 0.80 1.10 0.88 1.00 1.03 0.95 1.00 0.93 0.78 1.09
Metropolitan urban area of
Porto Alegre 0.87 1.11 1.08 1.17 0.75 1.08 0.88 0.85 0.85 1.05
Urban South excluding
metropolitan urban areas 0.84 1.15 0.97 0.90 0.99 1.04 1.06 0.94 0.85 0.86
Rural South 0.83 1.14 0.92 0.82 0.91 0.77 1.06 0.89 0.87 0.97
Distrito Federal 0.87 1.20 0.92 0.87 1.04 1.03 1.07 0.95 1.05 0.82
Urban Midwest excluding
Distrito Federal 0.92 1.20 0.98 0.86 0.92 0.96 1.08 1.10 1.09 0.87
Rural Midwest excluding
Distrito Federal 1.03 1.26 0.86 0.90 1.05 0.88 1.11 1.00 1.19 0.82
Source: Pesquisa de Orçamentos Familiares – POF / IBGE. 2017-2018
The first aspect to notice is that the price variation of the selected products is in most cases
more distinguishable among the geographical contexts of the North and Northeast regions in relation
to the Southeast, South and Midwest regions. When the product tends to have a price that is lower
than the average price in Brazil in the North and Northeast areas, the movement is reverse in most
contexts located in the remaining regions and vice-versa. The price of the bread roll that in the
geographical contexts of North and Northeast tend to be lower than the national average, have higher
prices registered in the rest of the country, with emphasis on the MUA of Belo Horizonte that
presented a value 40% higher in comparison with the value in Brazil. From this group, the MUA of
Rio de Janeiro was the only one where the bread roll had a price below the average.
Milk was another product that registered this trend of differentiation of the “North-South” areas
in the prices, while in the contexts of the North and Northeast regions the variation in relation to
Brazil was positive, having in the MUA of Belém a difference of 54%, while in the areas located in
the South, Southeast and Midwest the price of the milk presented was lower than the average of the
country. The MUA of Curitiba had prominence with the lowest value registered, -15%.
Another factor to consider is that despite the major converging prices between the “North-
South” areas, among the geographical contexts that are in the North-Northeast or in the South,
Southeast and Midwest, there is divergence in the level of price variation and even of the upward or
downward trend. In the case of MUA of the Northeast (Fortaleza, Recife and Salvador), the price of
the chicken egg. Recife and Salvador presented a negative difference in relation to Brazil of 3% and
12%, respectively. However, the price in MUA of Fortaleza was 21% above, and the price of sugar
was higher than the price of Brazil in MUA of Fortaleza (12%) and MUA of Salvador (7%) and lower
in MUA of Recife (6%). The same analysis can be made among MUAs of the Southeast Region (Belo
Horizonte, Rio de Janeiro and São Paulo) that are geographically closer. In Belo Horizonte the
difference in the price of milk was of 10% in relation to the price in Brazil, and in Rio de Janeiro and
São Paulo the variation was negative of 20% and 7%, respectively.
Relevant price divergences are also found in Rural Areas. The price of the bread roll in the
Rural North was 5% lower than the price in Brazil but in the Rural Northeast it was -16%. Only the
Rural Northeast registered the price of coffee below the national average, 4%. All the remaining Rural
Areas had the product value higher than the price in Brazil, having the Midwest Region the biggest
difference, 11%.
Although most product prices follow the regional trend explained above, some items had a
distinct behavior in certain geographical contexts such as, for example, rice. This product in a major
part of the North and Northeast Regions had a variation higher than the price in Brazil with values
from 3% to 15%, but in MUA of Fortaleza this addition was of 55%, being the higher price registered
for the product in Brazil. Similar cases that also worth mentioning are milk in MUA of Belém (+54%).
the bread roll in MUA of Belo Horizonte (+40%), meet of the second category in MUA of São Paulo
(+16%) and in Rural Midwest (-14%) and beef of the first category in MUA of Porto Alegre (-25%).
Knowing a little about the areas where the farming activities of each product are prominent or
even the local food culture makes it a little easier to understand the behavior of certain prices, such
as in the case of meet in rural Midwest. However, it is not a rule that is observed in all the products
not even those that do not depend of an economy of scale to be produced or of a region with specific
soil, climate, etc. As a result, the creation of a spatial price index in Brazil proves to be essential for
the incorporation of inequalities in the food basket of the vulnerable Brazilian population.
5. AGGREGATED REGIONAL PRICE INDEX
Resuming the steps to build the price index, after the definition of the population vulnerable to
food insecurity as target audience (step 1), the classification of foods by processing level (step 2), the
selection of the food basket representing the target audience (step 3) and the calculation of the implicit
prices and the average amounts of each product in all the geographical contexts (step 4), the fifth and
last step will be finally demonstrated: the definition and calculation of the price indexes.
In order to study the behavior of prices, Brazil is taken as basis (B). Thus, PB and QB denote the
implicit price and the average amount of Brazil. Based on the information of the implicit prices in
each context and in Brazil are presented the price ratios (Pij/PiB) for each product that compose the
food basket created in the previous section (See Appendix 1). In addition, to the price ratios are
calculated the price indexes of Laspeyres (L), Paasche (I) and Fisher (F) according to equations (3),
(4) and (5), respectively. The indexes were built as per the processing level and for foods in general.
According to OECD6, the Laspeyres index is a price index defined as a fixed weight, or fixed
basket, which uses the basket of goods and services of the basis period. The basis period works as the
reference period of the weight and the reference period of the price. The Paasche index is a price
index defined as a fixed weight, or fixed basket, that used the basket of goods and services of the
current period. The current period serves as the reference period of the weight and the basis period as
the reference period of the price. The Fischer index is the result of the geometric average of the two
indexes. Laspeyres and Paasche.
Paasche Index (Ij)
𝐼𝑗 =
∑ 𝑃𝑖𝑗.𝑄𝑖𝑗𝑖
∑ 𝑃𝑖𝐵.𝑄𝑖𝑗𝑖
=
𝐶𝑜𝑠𝑡 𝑜𝑓 𝑐𝑜𝑛𝑠𝑢𝑚𝑝𝑡𝑖𝑜𝑛 𝑖𝑛 𝑡ℎ𝑒 𝑔𝑒𝑜𝑔𝑟𝑎𝑝ℎ𝑖𝑐 𝑐𝑜𝑛𝑡𝑒𝑥𝑡 𝑗
𝐶𝑜𝑠𝑡 𝑜𝑓 𝑐𝑜𝑛𝑠𝑢𝑚𝑝𝑡𝑖𝑜𝑛 𝑜𝑓 𝑔𝑒𝑜𝑔𝑟𝑎𝑝ℎ𝑖𝑐 𝑐𝑜𝑛𝑡𝑒𝑥𝑡 𝑗 𝑎𝑡 𝑏𝑎𝑠𝑒 𝑝𝑟𝑖𝑐𝑒𝑠 (𝐵𝑟𝑎𝑧𝑖𝑙)
(3)
Laspeyres Index (Lj)
𝐿𝑗 =
∑ 𝑃𝑖𝑗.�̅�𝑖𝐵𝑖
∑ 𝑃𝑖𝐵.�̅�𝑖𝐵𝑖
=
𝐶𝑜𝑠𝑡 𝑜𝑓 𝑐𝑜𝑛𝑠𝑢𝑚𝑝𝑡𝑖𝑜𝑛 𝑜𝑓 𝑡ℎ𝑒 𝑏𝑎𝑠𝑒 𝑈𝐹 (𝑆ã𝑜 𝑃𝑎𝑢𝑙𝑜) 𝑎𝑡 𝑡ℎ𝑒 𝑝𝑟𝑖𝑐𝑒𝑠 𝑜𝑓 𝑈𝐹 𝑗
𝐶𝑜𝑠𝑡 𝑜𝑓 𝑐𝑜𝑛𝑠𝑢𝑚𝑝𝑡𝑖𝑜𝑛 𝑜𝑓 𝑡ℎ𝑒 𝑏𝑎𝑠𝑒 𝑈𝐹 (𝑆ã𝑜 𝑃𝑎𝑢𝑙𝑜)
(4)
Fisher Index (Fj)
𝐹𝑗 = √𝐿𝑗 . 𝐼𝑗 = 𝑔𝑒𝑜𝑚𝑒𝑡𝑟𝑖𝑐 𝑚𝑒𝑎𝑛 𝑜𝑓 𝑡ℎ𝑒 𝑡𝑤𝑜 𝑖𝑛𝑑𝑖𝑐𝑒𝑠 (5)
Table 4 presents the result of the aggregated regional price index calculated from the three, as
per the geographical contexts. With this data, it is possible to observe the differences of spatial prices
in each index model (analysis by column) and analyze the differences of values in the contexts among
the models (analysis by line). Brazil is used as reference basis and the values of the three indexes
equals 1.
The values of the MUAs of Belém (+9.0%) and São Paulo (+7.5%) and the Federal District
(+8.1) were the areas that presented the biggest positive difference in relation to Brazil according to
the Laspeyres index, while the Urban North (-7.9%), the Urban South (- 6.9%) and the MUA of
Recife (-5.9%) were the areas that registered the biggest negative differences. That is, as per this
index the Urban North is the geographical context with the cheapest food basket in the country.
Considering the results calculated with the Paasche index, the Federal District (+9.8%) and the
MUAs São Paulo (+7.7%) and Salvador (+5.8%) were the areas that registered the biggest index
values above the basis Brazil. On the other hand, the MUA Recife (-8.7%), the Urban Southeast and
the Rural South with indexes around -6% were the geographical contexts that had a prominent
position with bigger negative differences in relation to Brazil. The Fischer index, since it is calculated
6 OECD: https://stats.oecd.org/
from the values of the other two indexes cited. has a similar result in terms of regions with variations
that are superior or inferior to the national parameter.
It is possible to emphasize the results of the Federal District and of the MUA of São Paulo that
regardless of the index model used are always the areas with the biggest distance (more expensive)
compared to the basis of reference. In contrast, with MUA of Recife occurs the opposite, and in the
three indexes presented it is the region that is always among the ones with the smallest values in
relation to Brazil.
Table 4: Price indexes for Brazil and geographical contexts - Brazil - 2017-2018
Geographical Context
LASPEYERES
INDEX
PAASCHE
INDEX
FISHER
INDEX
PLS
Brazil 1.000 1.000 1.000 0.000
Metropolitan urban area of Belém 1.090 1.043 1.066 0.044
Urban North excluding metropolitan urban areas 1.011 0.989 1.000 0.022
Rural North 0.981 0.969 0.975 0.012
Metropolitan urban area of Fortaleza 1.013 1.003 1.008 0.011
Metropolitan urban area of Recife 0.954 0.916 0.935 0.041
Metropolitan urban area of Salvador 0.992 0.973 0.983 0.019
Urban Northeast excluding metropolitan urban areas 0.977 0.964 0.970 0.013
Rural Northeast 0.944 0.952 0.948 0.009
Metropolitan urban area of Belo Horizonte 1.008 0.992 1.000 0.016
Metropolitan urban area of Rio de Janeiro 1.016 1.009 1.013 0.007
Metropolitan urban area of São Paulo 1.091 1.086 1.089 0.005
Urban Southeast excluding metropolitan urban areas 1.033 1.026 1.029 0.006
Rural Southeast 0.984 0.980 0.982 0.004
Metropolitan urban area of Curitiba 1.011 1.001 1.006 0.010
Metropolitan urban area of Porto Alegre 1.054 1.020 1.037 0.033
Urban South excluding metropolitan urban areas 0.985 0.974 0.980 0.012
Rural South 0.937 0.915 0.926 0.024
Distrito Federal 1.018 1.013 1.015 0.004
Urban Midwest excluding Distrito Federal 1.016 1.004 1.010 0.012
Rural Midwest excluding Distrito Federal 0.976 0.977 0.976 0.001
Search: Pesquisa de Orçamentos Familiares – POF / IBGE. 2017-2018
Still in Table 4 are found the measure values of the Dispersion Paasche – Laspeyres (Paasche–
Laspeyres Spread - PLS) designed by Hill (1999) that indicate the heterogeneity of prices among
these indexes and can be analyzed from the measure expressed by Equation 6:
𝑃𝐿𝑆𝑠𝑡 = |ln (
𝑃𝑠𝑡
𝐿
𝑃𝑠𝑡
𝑃)| (6)
Where PLSst is the absolute value of the log for the Laspeyres price index divided by the
Paasche index for period t.
In Figure 6 are presented the PLS values for all the geographical contexts. making the visual
comparison among the areas easier. The MUAs of three different Major Regions were the
geographical contexts that presented the biggest dispersions: in the North Region, the MUA of Belém
with 0.044, in the Northeast Region the MUA of Recife with 0.041 and in the South Region the MUA
of Porto Alegre with 0.033. On the other hand, the MUA of São Paulo was among the metropolitan
areas the one with the smaller distance among the values of the two indexes.
Comparing the urban areas, the Urban North was the one that registered the biggest dispersion
with 0.022, followed by the Northeast with 0.013. The urban context with the smallest value was the
Southeast, 0.006. While in the Rural Areas, the South (0.024) and the North (0.012) were the regions
with the biggest dispersions, and again the Southeast was the region with the smallest dispersion,
0.004.
Figure 6: Paasche–Laspeyres Spread (PLS) by geographical contexts
Search: Pesquisa de Orçamentos Familiares – POF / IBGE. 2017-2018
After the results of the aggregated price indexes by geographical context and following the idea
of evaluating the cost of the Brazilian food basket of the vulnerable population according to its
composition, the values of the spatial deflators were calculated for the three models (Laspeyres,
Paasche and Fischer) as per the level of processing of the products, for the geographical contexts. As
before, Brazil is used as reference area, and the values of the three indexes equal 1. These values can
be seen in Appendix 2. However, to demonstrate the comparison of indexes by processing level and
geographical context. Figures 7, 8 and 9 will be used, representing the Paasche index for the
processing levels: Natural or minimally processed, processed and ultra-processed for all the areas.
The Paasche index was chosen for the demonstration. since according to Deaton and Zaidi
(2002)7 the use of the Paasche index is suggested as spatial deflator for two reasons. Oliveira et al
(2016)8 also suggest the use of the Paasche index to evaluate regional differences by POF. First of
all, this index emphasizes the consumption habits of each geographical context. Secondly, the money
metrics utility function is obtained, or approximated, when the expense is divided by the Paasche
index, according to the explanation in the sequence below.
Considering the utility of the Xa basket observed in area a (taking the price vector of the area
Pa as given) in the monetary metrics (Maa) expressed as:
𝑀𝑎𝑎 = 𝐸(𝑃𝑎 , 𝑈(𝑋𝑎)) = 𝐸(𝑃𝑎, 𝑈𝑎) = 𝑚𝑖𝑛𝑋(𝑃𝑎
′𝑋) 𝑎𝑠 𝑈(𝑋) ≥ 𝑈𝑎 = 𝑈(𝑋𝑎), (6)
where Pa and X are vectors and a represents an area (or location). As result, the expenditure in area a
is: Maa=Pa’Xa. where Xa is the argument that minimizes the expression above.
According to Shepard´s Lemma9 :
𝜕𝐸 𝜕𝑃𝑎⁄ = 𝑋𝑎 , with Xa being the demand for prices Pa. (7)
The monetary metrics utility Mba=E(Pb .Ua) takes the price vector Pb and Ua as references. Mba
is the expenditure that people in area a should incur to obtain the utility Ua taking the price vector Pb
as given. The first order approximation of Mba results in Equation (8):
𝑀𝑏𝑎 = 𝐸(𝑃𝑏 , 𝑈𝑎) ≈ 𝐸(𝑃𝑎, 𝑈𝑎) + (𝑃𝑏 − 𝑃𝑎)′. (𝜕𝐸 𝜕𝑃)⁄ |𝑃=𝑃𝑎
𝑀𝑏𝑎 = 𝐸(𝑃𝑏 , 𝑈𝑎) ≈ 𝑃𝑎
′. 𝑋𝑎 + (𝑃𝑏 − 𝑃𝑎)′ . 𝑋𝑎 = 𝑃𝑏
′𝑋𝑎 (8)
7 Deaton, A. and S. Zaidi, “Guidelines for Constructing Consumption Aggregates for Welfare Analysis”, Living Standards Measurement Survey Working Paper 135,
Washington DC, The World Bank, 2002.
8 Previously mentioned in footnote number 3.
9 Varian, Hal (1992) Microeconomic Analysis 2nd ed. USA: W. W. Norton & Company Inc.
The monetary metrics utility is approximately Pb’Xa. This approximation occurs without the
need to suppose specific utilities functions and stronger restrictions (homogeneity, quadratic
functions or translog).
Finally, if the expenditure observed in area a (Maa=Pa’.Xa) is divided by the Paashe index that
takes Pb as reference (Iba=Pa’.Xa/Pb’.Xa). the result is (approximately) the money metrics utility
function (Mba≈Pb’.Xa). as reported in equation 8.
According to the Paasche index for natural or minimally processed foods (Figure 7), three
MUAs of different regions presented the biggest differences in relation to base 1, represented by
Brazil. The MUA of Fortaleza is the geographical context with the biggest difference, 12.3%,
followed by the MUAs of Belém (4.4%) and São Paulo (3.9%). That demonstrates the relevance of
studying the behavior of prices in the different realities of Brazil, since it is not possible to infer that
a type of food is more expensive or cheap in a certain area considering the general cost of living of
this location. Considering the contexts with results below of the value in Brazil, the three contexts
with the lowest values were in the South Region. having the Rural South -11.9%, MUA of Porto
Alegre -7.2% and the Urban South registering -6.9%.
Figure 7: Paasche price index of Natural or minimally processed products by geographical contexts
Search: Pesquisa de Orçamentos Familiares – POF / IBGE. 2017-2018
Among the processed foods (Figure 8), the Paasche index with the highest value was in the
MUA Belo Horizonte, 22.7%. The Urban Southeast and MUA of São Paulo have also registered high
differences in relation to Brazil, 16.6% and 14.6%, respectively. The MUA of Salvador (-17.4%) and
Urban Northeast (-14.9%) had the biggest negative variations, as well as the MUAs of Recife, Rio de
Janeiro and Fortaleza around 9%.
The MUA of Porto Alegre (Figure 9) was the geographical context with the highest price of
ultra-processed foods in relation to the reference Brazil, 5.0%. In second place, is the Rural Midwest
with a variation 3.5% above the national reference. The MUA of Recife and the Rural Northeast were
the areas with the biggest negative variations, respectively, 8.2% and 4.2%.
Figure 8: Paasche price index of processed foods by geographical contexts
Search: Pesquisa de Orçamentos Familiares – POF / IBGE. 2017-2018
Figure 9: Paasche price index of ultra-processed foods by geographical contexts
Search: Pesquisa de Orçamentos Familiares – POF / IBGE. 2017-2018
6. FINAL CONSIDERATIONS
The inclusion of EBIA in POF enabled an operational definition of vulnerable subgroup of the
population. Up to 60% of the total population has a non-negligible chance of experiencing moderate
or severe food insecurity. The operational definition allows the monitoring of the target audience of
public programs that aim to combat hunger and promote safe food security, identifying consumption
habits and regional differences in the cost of living, especially in urban and rural areas. Such
information and monitoring will only be possible in the future with the maintenance of this
information in POF. Thus, it is recommended to update EBIA on a regular basis in POF.
The methodology developed can be applied to studies that search for the analysis of poverty
and inequality from the perspective of the vulnerable population in IA. Therefore, the next steps
involve the use of deflators in the calculation of poverty measures and monetary inequality revealing
the regional disparities and the differences of among significant subgroups of the population, such as
the Rural Northeast and the Urban Southeast. This information is relevant because Brazil still does
not have an official calculation to measure these regional price changes. This article seeks to
contribute to fill this gap in a country of continental extent.
A second development is the estimation of demand systems that make it possible to assess the
impacts of public policies on the subgroup identified as vulnerable, such as tax increases and
reductions.
Another contribution that can be applied in the future would be a time analysis with the
elaboration of a historical series of price indexes that are based on the food basket defined for the
vulnerable population. This temporal price index can serve as a basis for readjustments, for example,
for benefits and public programs aimed at combating hunger and food insecurity in Brazil.
7. REFERENCES
AIZCORBE, Ana M. A Practical Guide to Price Index and Hedonic Techniques. Oxford, Oxford
University Press, 2014.
BLACKORBY, C. and D. Donaldson, “Welfare Ratios and Distributionally Sensitive Cost-Benefit
Analysis,” Journal of Public Economics, 34, 265–90, 1987.
DEATON, A. and S. ZAIDI, “Guidelines for Constructing Consumption Aggregates for Welfare
Analysis,” Living Standards Measurement Survey Working Paper 135, Washington DC, The World
Bank, 2002.
HILL, R. J. (1999a), “Comparing Price Levels Across Countries using Minimum Spanning Trees,”
Review of Economics and Statistics 81(1), 135-142.
IBGE. PESQUISA DE ORÇAMENTOS FAMILIARES 2017-2018: primeiros resultados. Rio de
Janeiro: IBGE, 2019. 64 p. Disponível em:
https://www.ibge.gov.br/estatisticas/sociais/populacao/24786-pesquisa-de-orcamentos-familiares-
2.html?edicao=25578&t=publicacoes. Acesso em: out. 2021.
_____. PESQUISA NACIONAL POR DOMICÍLIOS 2004: Segurança Alimentar. Rio de Janeiro:
IBGE, 2004. 140 p. Disponível em: https://biblioteca.ibge.gov.br/visualizacao/livros/liv29775.pdf.
Acesso em: fev. 2023.
_____. Mapa de pobreza e desigualdade: municípios brasileiros 2003. Rio de Janeiro: IBGE, 2008”.
1 DVD. Disponível em: https://biblioteca.ibge.gov. br/index.php/biblioteca-
catalogo?view=detalhes&id=241385. Acesso em: nov. 2021.
_____. TRATAMENTOS das informações. In: PESQUISA de orçamentos familiares. Rio de
Janeiro: IBGE, 1991. v. 2. 198 p. (Série relatórios metodológicos, v. 10. Disponível em:
https://biblioteca.ibge.gov.br/visualizacao/livros/liv81836_v2.pdf. Acesso em: out. 2021.
ILO, Consumer Price Index Manual: Theory and practice, 2004 (revisions as at 15/08/2010)
http://www.ilo.org/public/english/bureau/stat/guides/cpi/.
OLIVEIRA, L. S. de et al. Construction of a consumption aggregate based on information from
POF 2008-2009 and its use in the measurement of welfare, poverty, inequality and vulnerability of
families. The Review of Income and Wealth, Malden: Blackwell, v. 62, n. 1, p. 179-210, Aug.
2016. Disponível em:
https://www.researchgate.net/publication/315943671_Construction_of_a_Consumption_Aggregate
_Based_on_Information_from_the_Brazilian_Consumer_Expenditure_and_its_use_in_the_Measur
ement_of_Welfare_Poverty_Inequality_and_Vulnerability_of_Families_Construc. Acesso em: out.
2021.
_________, L. S. de et al. The expansion of consumption and the welfare dynamics of the Brazilian
families: a decomposition analysis of poverty and inequality. In: SOCIETY FOR THE STUDY OF
ECONOMIC INEQUALITY - ECINEQ MEETING, 17., 2017, New York. Papers [...]. New York:
City University of New York, The Graduate Center, 2017. p. 1-38. Disponível em:
https://www.researchgate.net/publication/332211428_The_expansion_of_consumption_and_the_we
lfare_dynamics_of_the_Brazilian_families_a_decomposition_analysis_of_poverty_and_inequality.
Acesso em: out. 2021.
PÉREZ-ESCAMILLA, R.; SEGALL-CORRÊA, A. M. Food insecurity measurement and
indicators. Revista de Nutrição, Campinas: Pontifícia Universidade Católica - PUC-Campinas, v.
21, p. 15-26, jul./ago. 2008. Suplemento. Disponível em: https://www.scielo.br/scielo.
php?script=sci_arttext&pid=S1415-52732008000700003&lang=es. Acesso em: ago. 2020.
RAO, D. S. Prasada, Welfare Comparisons with Heterogeneous Prices, Consumption, and
Preferences, in Adler, M.; Fleurbaey, M. (ed.) The Oxford Handbook of Well-being and Public
Policy. New York: Oxford University Press, 2016.
RAVALLION, M. “Poverty Lines in Theory and Practice”. LSMS Working Paper – Number 133.
Washington DC, TheWorld Bank, 1998.
SEGALL-CORREA et al. Refinement of the Brazilian Household Food Insecurity Measurement
Scale: Recommendation for a 14-item EBIA, Rev. Nutr., Campinas, 27(2):241-251, mar./abr., 2014
VARIAN, Hal (1992) Microeconomic Analysis 2nd ed. USA: W. W. Norton & Company Inc
WOOLDRIDGE, Jeffrey M., Introductory Econometrics: A Modern Approach, Seventh Edition.
Cengage Learning inc 2020.
__________________ Econometric Analysis of Cross Section and Panel Data Second
Edition. London:MIT Press, 2010.
APPENDIX 1 – ITEMS OF FOOD BASKET OF THE POPULATION VULNERABLE
POF'S
CODE
DESCRIPTION ITEMS POF'S CODE DESCRIPTION ITEMS
64006 Massa De Pastel 81029 Peito Bovino
65026 Pipoca Para Viagem 71011 Banana (Nao-Especificada)
85012 Queijo tipo De Minas 68011 Vinagre De Álcool
79020 Creme De Arroz 70033 Peixe Sardinha Em Conserva
65001 Aveia Em Flocos 77030 Suco De Fruta Ou Vegetal em caixa
65024 Macarrão Sem Ovos 82047 Carne Não-Especificada
65032 Couve 71089 Milho Verde Com Ervilha Em Conserva
67005 Repolho 65004 Tangerina
67009 Chuchu 68022 Lagarto Comum
67041 Tomate 71007 Pao De Forma De Padaria (Salgado)
67051 Cebola 80004 Feijão Preto
67057 Abacaxi 63015 Manga
68026 Maçã 68032 Melancia
68030 Maracujá 68034 Salsa
68033 Chá De Dentro 67010 Orégano (Tempero Industrializado)
71004 Costela Bovina 70022 Ervilha Em Conserva
71013 Carne Moída De Primeira 77002 Pão De Forma Industrializado
71014 Fígado Bovino 80005 Toucinho De Porco Defumado
71025 Milho Verde Em Conserva 81010 Batata Frita Para Viagem
77004 Peito De Galinha Ou Frango 85015 Açúcar Indeterminado
78004 Muçarela 69066 Flocos De Milho
79018 Mortadela 65009 Massa De Lasanha
81026 Fubá De Milho 65029 Abóbora Moganga
65006 Macarrão instântaneo 67033 Bolo industrializado
65048 Salsicha 80025 Peixe Não-Especificado
81021 Óleo De Milho 76009 Biscoito Não-Especificado
84004 Macarrão Com Ovos 80052 Leite De Coco
65033 Tempero industrializado 70038 Vinho De Uva E Outros
70118 Refrigerante De Laranja 83024 Limão Nao Especificado
82002 Água Mineral 68093 Ovo De Páscoa
82010 Cerveja 69058 Pão Não-Especificado
83001 Carne Assada Ou Bife Preparado Para Viagem 80015 Tempero Não-Especificado
85011 Batata Doce 70084 Queijo Não-Especificado
64004 Alface 79030 Aguardente De Cana
67001 Banana D'água 83003 Óleo De Girassol
68001 Laranja Pêra 84008 Pá (carne bovina de segunda)
68014 Mamão 71009 Salame
68031 Patinho 81027 Sopa Desidratada
71005 Frango Congelado 77014 Goiaba
78002 Parte De Galinha Ou Frango Não-Especificada 68042 Brigadeiro
78003 Torrada 69036 Capa De Filé
80019 Batata Não-Especificada 71012 Molho De Soja
64008 Linguiça 70036 Confeitos de bolos e doces
81022 Fécula De Mandioca 69022 Queijo prato
65015 Banana Prata 79017 Salsicha em conserva
68002 Doce De Frutas Em Pasta De Qualquer Sabor 77028 Cesta Básica
69012 Pão Integral 90005 CESTA BASICA
80014 Presunto De Qualquer Tipo
Search: Pesquisa de Orçamentos Familiares – POF / IBGE. 2017-2018
APPENDIX 2 - Laspeyres. Paasche and Fisher price index. according to food processing levels.
by geographic context - Brazil - POF 2017 – 2018
Geographical Context Processing level Laspeyres Index Paasche Index Fischer Index
Brazil
Natural ou minimally processed food 1.000 1.000 1.000
Culinary preparations based 1.000 1.000 1.000
Processed foods 1.000 1.000 1.000
Ultra-processed foods 1.000 1.000 1.000
Metropolitan urban area of
Belém
Natural ou minimally processed food 1.085 1.044 1.064
Culinary preparations based 1.140 1.123 1.131
Processed foods 1.035 1.023 1.029
Ultra-processed foods 0.992 0.981 0.986
Urban North excluding
metropolitan urban areas
Natural ou minimally processed food 1.002 0.974 0.988
Culinary preparations based 1.025 1.007 1.016
Processed foods 0.941 0.923 0.932
Ultra-processed foods 1.008 0.996 1.002
Rural North
Natural ou minimally processed food 1.001 0.979 0.990
Culinary preparations based 1.060 1.063 1.062
Processed foods 0.966 0.962 0.964
Ultra-processed foods 1.029 0.991 1.010
Metropolitan urban area of
Fortaleza
Natural ou minimally processed food 1.124 1.123 1.124
Culinary preparations based 1.112 1.086 1.099
Processed foods 0.942 0.904 0.923
Ultra-processed foods 1.021 1.004 1.012
Metropolitan urban area of
Recife
Natural ou minimally processed food 1.025 1.008 1.016
Culinary preparations based 1.052 1.035 1.043
Processed foods 0.926 0.901 0.913
Ultra-processed foods 0.966 0.918 0.942
Metropolitan urban area of
Salvador
Natural ou minimally processed food 1.015 0.996 1.005
Culinary preparations based 1.115 1.124 1.120
Processed foods 0.834 0.826 0.830
Ultra-processed foods 0.956 0.959 0.957
Urban Northeast excluding
metropolitan urban areas
Natural ou minimally processed food 1.030 1.021 1.025
Culinary preparations based 1.033 1.021 1.027
Processed foods 0.877 0.851 0.863
Ultra-processed foods 0.975 0.969 0.972
Rural Northeast
Natural ou minimally processed food 1.017 1.004 1.010
Culinary preparations based 1.011 1.016 1.014
Processed foods 0.891 0.866 0.878
Ultra-processed foods 0.973 0.958 0.965
Metropolitan urban area of
Belo Horizonte
Natural ou minimally processed food 1.004 0.977 0.990
Culinary preparations based 1.062 1.049 1.055
Processed foods 1.252 1.227 1.240
Ultra-processed foods 0.992 0.971 0.981
Metropolitan urban area of
Rio de Janeiro
Natural ou minimally processed food 1.009 1.000 1.005
Culinary preparations based 1.133 1.119 1.126
Processed foods 0.904 0.904 0.904
Ultra-processed foods 1.029 1.010 1.019
Natural ou minimally processed food 1.068 1.039 1.053
Metropolitan urban area of
São Paulo
Culinary preparations based 1.036 1.000 1.018
Processed foods 1.145 1.146 1.145
Ultra-processed foods 1.038 1.021 1.029
Urban Southeast excluding
metropolitan urban areas
Natural ou minimally processed food 1.028 1.019 1.023
Culinary preparations based 1.018 0.995 1.006
Processed foods 1.180 1.166 1.173
Ultra-processed foods 1.030 1.026 1.028
Rural Southeast
Natural ou minimally processed food 1.010 1.002 1.006
Culinary preparations based 0.899 0.898 0.898
Processed foods 1.121 1.109 1.115
Ultra-processed foods 1.009 1.008 1.008
Metropolitan urban area of
Curitiba
Natural ou minimally processed food 0.970 0.939 0.954
Culinary preparations based 1.072 1.051 1.062
Processed foods 1.049 1.015 1.032
Ultra-processed foods 1.050 1.025 1.037
Metropolitan urban area of
Porto Alegre
Natural ou minimally processed food 0.983 0.928 0.955
Culinary preparations based 1.050 0.971 1.010
Processed foods 1.049 1.021 1.035
Ultra-processed foods 1.149 1.050 1.098
Urban South excluding
metropolitan urban areas
Natural ou minimally processed food 0.958 0.931 0.944
Culinary preparations based 0.944 0.905 0.924
Processed foods 1.095 1.046 1.070
Ultra-processed foods 1.031 1.019 1.025
Rural South
Natural ou minimally processed food 0.905 0.881 0.893
Culinary preparations based 0.967 0.960 0.963
Processed foods 1.087 1.040 1.063
Ultra-processed foods 1.018 1.006 1.012
Distrito Federal
Natural ou minimally processed food 0.986 0.964 0.975
Culinary preparations based 0.999 0.987 0.993
Processed foods 1.147 1.120 1.133
Ultra-processed foods 1.006 0.999 1.002
Urban Midwest excluding
Distrito Federal
Natural ou minimally processed food 1.001 0.982 0.992
Culinary preparations based 0.915 0.897 0.906
Processed foods 1.149 1.125 1.137
Ultra-processed foods 1.033 1.025 1.029
Rural Midwest excluding
Distrito Federal
Natural ou minimally processed food 1.030 1.016 1.023
Culinary preparations based 0.894 0.873 0.883
Processed foods 1.166 1.099 1.132
Ultra-processed foods 1.030 1.035 1.032
Search: Pesquisa de Orçamentos Familiares – POF / IBGE. 2017-2018
Forest products conversion factors provides ratios of raw material input to the output of wood-based forest products for 37 countries of the world. Analysts, policymakers, forest practitioners and forest-based manufacturers often have a need for this information for understanding the drivers of efficiency, feasibility and economics of the sector.