FAO Statistics Working Paper Series

Issue 24/41

ESTIMATING THE FOOD VALUE CHAIN

DECOMPOSITION BY INDUSTRIES AND
PRIMARY FACTORS
FAO Statistics Working Paper Series / 24-41

ESTIMATING THE FOOD VALUE CHAIN DECOMPOSITION BY

INDUSTRIES AND PRIMARY FACTORS

Silvia Cerilli, Michele Vollaro, Veronica Boero

Food and Agriculture Organization of the United Nations, Rome

Dianna Tran, Henry Williams, Miguel I. Gomez

Charles H. Dyson School of Applied Economics & Management, Cornell University, Ithaca,
United States of America

and

Jing Yi
Department of Agricultural and Applied Economics, University of Wisconsin-Madison,
United States of America

Food and Agriculture Organization of the United Nations

Rome, 2024
Required citation:
Cerilli, S., Vollaro, M., Boero, V., Tran, D., Williams, H., Yi, J. and Gomez, M.I. 2024. Estimating the food
value chain decomposition by industries and primary factors. FAO Statistics Working Paper Series,
No. 24-41. Rome, FAO. https://doi.org/10.4060/cd0532en

The designations employed and the presentation of material in this information product do not imply the expression of any opinion
whatsoever on the part of the Food and Agriculture Organization of the United Nations (FAO) concerning the legal or development
status of any country, territory, city or area or of its authorities, or concerning the delimitation of its frontiers or boundaries. The mention
of specific companies or products of manufacturers, whether or not these have been patented, does not imply that these have been
endorsed or recommended by FAO in preference to others of a similar nature that are not mentioned.

The views expressed in this information product are those of the author(s) and do not necessarily reflect the views or policies of FAO.

ISBN 978-92-5-138740-5

© FAO, 2024

Some rights reserved. This work is made available under the Creative Commons Attribution-NonCommercial-ShareAlike 3.0 IGO
licence (CC BY-NC-SA 3.0 IGO; https://creativecommons.org/licenses/by-nc-sa/3.0/igo/legalcode).

Under the terms of this licence, this work may be copied, redistributed and adapted for non-commercial purposes, provided that the
work is appropriately cited. In any use of this work, there should be no suggestion that FAO endorses any specific organization,
products or services. The use of the FAO logo is not permitted. If the work is adapted, then it must be licensed under the same or
equivalent Creative Commons licence. If a translation of this work is created, it must include the following disclaimer along with the
required citation: “This translation was not created by the Food and Agriculture Organization of the United Nations (FAO). FAO is not
responsible for the content or accuracy of this translation. The original [Language] edition shall be the authoritative edition.”

Disputes arising under the licence that cannot be settled amicably will be resolved by mediation and arbitration as described in Article
8 of the licence except as otherwise provided herein. The applicable mediation rules will be the mediation rules of the World Intellectual
Property Organization http://www.wipo.int/amc/en/mediation/rules and any arbitration will be conducted in accordance with the
Arbitration Rules of the United Nations Commission on International Trade Law (UNCITRAL).

Third-party materials. Users wishing to reuse material from this work that is attributed to a third party, such as tables, figures or
images, are responsible for determining whether permission is needed for that reuse and for obtaining permission from the copyright
holder. The risk of claims resulting from infringement of any third-party-owned component in the work rests solely with the user.

Sales, rights and licensing. FAO information products are available on the FAO website (www.fao.org/publications) and can be
purchased through [email protected]. Requests for commercial use should be submitted via: www.fao.org/contact-
us/licence-request. Queries regarding rights and licensing should be submitted to: [email protected]
Abstract
This statistical working paper presents an estimation methodology for decomposing food expenditure
across the industries and the primary factors of the food value chain (FVC). The approach outlined is based
on the Global Food Dollar methodology developed by the Economic Research Service of the United States
Department of Agriculture (USDA-ERS) and Cornell University (Yi et al., 2021). FAO has enriched the
analytical scope of the methodology by adding the industry and primary factors decomposition. Country
coverage has also been increased by adapting the methodology to different data types and sources.

The methodology, source datasets and computational results have been published on FAOSTAT. The FVC
domain currently covers 65 middle- and high-income countries and territories with annual data from 2005
to 2015, and several countries and years will be added in the upcoming data release.

The paper describes methods for estimating three food value measures: (i) Food At Home (FAH); (ii) Food
and Tobacco At Home (FTAH); and (iii) Food and Accommodations Away From Home (FAAFH). The paper
also describes two decompositions, by food industries (agriculture, processing, retail, etc.) and by primary
factors (labour, taxes, imports and gross operating surplus). Through collaboration with Cornell University,
the Global Food Dollar methodology was integrated from SAS and STATA into a single programming
language (STATA), to limit software licenses needed to apply the methodology and to further facilitate
upscaling in terms of data source applicability by widening its country and data coverage. Further
improvements will focus on the migration of the computation technique towards reliable open-source
programming software, such as Python or R, to minimize the technical costs, to improve the efficiency of
the computation in each stage of the process, and to be able to directly interact with other open-source
analytical software.

The decomposition relies on the Leontief input-output reduction method (Leontief, 1967), which is
applied to the food value chain industries and related primary factors accruing to the value added,
(Canning, 2011), thereby providing elements to assess the impact analysis by sector.

The methodology proposed allows the analysis of the entire FVC, from producers to consumers, in its farm
and post-farm gate component. The methodology uses the international statistical standards of the
System of National Accounts (SNA), which provides a comprehensive, consistent and flexible set of
macroeconomic accounts for policymaking, analysis and research purposes. Since it is applied in several
countries, it facilitates the comparison of macroeconomic variables across countries and at the global
level.

Finally, the data and methodology presented in this paper allow to expand a simple “upstream” or
producer-behaviours-oriented, or “downstream” (consumer-oriented) analysis of the food value by
analysing all the food industries involved (such as agricultural producers, processing, transport and
retailers) and therefore providing a more complete FVC analysis. This includes the ability to properly
quantify post-farm gate components, which have become increasingly important over time in the analysis
of agrifood systems. Further, it may inform national, regional, and global food policy, including additional
measures in the World Food Summit framework, and it may be relevant for assessing sustainable
consumption and production patterns under Sustainable Development Goal (SDG) 12 (Responsible
Consumption and Production) as well as progress towards SDGs 1 (No Poverty) and 2 (Zero Hunger).

iii
Contents
Abstract ........................................................................................................................................................ iii
Acknowledgements...................................................................................................................................... vi
Abbreviations .............................................................................................................................................. vii
1 Introduction .......................................................................................................................................... 1
1.1 The food value chain domain ........................................................................................................ 1
1.2 Methodological background: “food dollar” and “Global Food Dollar” ......................................... 2
1.2.1 The food dollar series............................................................................................................ 4
1.2.2 The Global Food Dollar.......................................................................................................... 5
1.3 The joint FAO–Cornell upscaling exercise and the industry and primary factors decomposition 5
2 Methodology ......................................................................................................................................... 6
2.1 National accounts data used to derive FVC domain information ................................................. 6
2.2 Food value measures and related data needs .............................................................................. 7
2.3 The data system ............................................................................................................................ 9
2.3.1 The Z matrix .......................................................................................................................... 9
2.3.2 The Y matrix ........................................................................................................................ 10
2.3.3 The V matrix ........................................................................................................................ 10
2.4 Leontief application for the derivation of food value measures ................................................ 10
2.4.1 Dealing with imports trade and transportation margins .................................................... 11
2.4.2 Leontief model and the direct and indirect requirement matrices .................................... 12
2.4.3 Computation of the three food value chain measures ....................................................... 13
2.5 Industry and primary factors decomposition ............................................................................. 13
3 Results ................................................................................................................................................. 15
3.1 Strengths and limitations of the Food Value Chain domain ....................................................... 15
3.2 Main findings............................................................................................................................... 15
4 Potential applications ......................................................................................................................... 19
5 Way forward ....................................................................................................................................... 20
Annex I Industries mapping to ISIC code ................................................................................................ 21
A. OECD database data mapping with ISIC ..................................................................................... 21
B. Asian Development Bank database data mapping with ISIC ...................................................... 22
References .................................................................................................................................................. 23
Acknowledgements
The authors gratefully acknowledge the methodological input received from José Rosero Moncayo, Piero
Conforti (Food and Agriculture Organization of the United Nations [FAO]) and Patrick Canning (USDA-ERS),
and the technical input received from Nick Grandstaff and Tianzi Liu (Cornell University). The authors also
thank Rob Vol (International Food Policy Research Institute) and Francesco Tubiello (FAO) for their useful
insights during the peer review process.

vi
Abbreviations
FAAFH Food and Accommodation Away From Home

FAH Food At Home

FAO Food and Agriculture Organization of the United Nations

FTAH Food and Tobacco At Home

FVC food value chain

IOTs input-output tables

ISIC International Standard Industrial Classification of All Economic Activities

NSOs National Statistical Offices

OECD Organisation for Economic Co-operation and Development

SDG Sustainable Development Goal

SEEA System of Environmental-Economic Accounting

SNA System of National Accounts

SUTs supply and use tables

USDA-ERS United States Department of Agriculture - Economic Research Service

vii
1 Introduction
1.1 The food value chain domain
The FAO Strategic Framework 2022-2031 seeks to support the 2030 Agenda, including SDG 1 (No Poverty),
SDG 2 (Zero Hunger), and SDG 12 (Responsible Consumption and Production) through the transformation
to more efficient, inclusive, resilient and sustainable agrifood systems for a “better production, better
nutrition, a better environment, and a better life, leaving no one behind” (FAO, 2021).

The 2021 United Nations Food Systems Summit explicitly recognized the importance of adopting a food
systems approach inclusive of FVC analysis, when considering policies to address farm, food safety and
environmental issues. Both international fora (Oddone and Padilla, 2014; Henriksen et al., 2010) and
academia (Barrett et al., 2022) show increasing attention to the FVC analysis in the food system
framework.

The analysis of the food value chain and the data of the FVC domain need to be framed in terms of
definition, scope and limitation with respect to the analysis of food systems. The terms “food system” and
“food value chain” need to be clearly distinguished: the Food Systems Countdown Initiative (FSCI) defines
food systems as comprising “all the people, institutions, places, and activities that play a part in growing,
processing, transporting, selling, marketing, and, ultimately, eating food” (FSCI, n.d.). It is a generic
framework that may be analysed from the food security, food safety, environmental or socioeconomic
perspective, among others. The FVC concept is instead based on a specific perspective of the food system
analysis, which comprises “all activities required to bring farm products to consumers, including
agricultural production, processing, storage, marketing, distribution, and consumption” (Gómez et al.,
2011). It considers linkages between participating actors (e.g. farmers, manufacturers, retailers and
consumers) and examines the flow of foods and their value from farmers to distributors and retailers.

In this context, FAO has built a FAOSTAT statistical domain on food value chains to disseminate data on
the distribution of value added along the FVC from final consumers back to the farmers, going through
the intermediate sectors and related primary factors. Data published in the domain, as further described
in Section 1.2, enrich and upscale the Global Food Dollar methodology developed by USDA-ERS and
Cornell University. Published data show the integrated distribution of value added of both the farm and
post-farm gate food-related activities for 65 countries and territories between 2005 and 2015.

The FVC data allow analysing food supply from an economic perspective: in the case of the FVC domain,
as further explained in the next sections, the economic principles applied are those of the input-output
modelling and the Leontief matrix reduction methods applied to the supply chain industries (Leontief,
1967). Figure 1 below shows the main farm and post-farm gate economic activities, classified by the ISIC,
considered in the FVC analysis as presented in the FAOSTAT domain.

1
Figure 1. Food value chain main economic activities components by ISIC category

Farm gate

Manufacture
Accomodation
Agriculture, of food,
Wholesale and Transportation and food
forestry and beverages and
retail trade and storage services
fishing tobacco
(ISIC G) (ISIC H) activities
(ISIC A) products
(ISIC I)
(ISIC C)

Post-farm gate

Source: Author’s own elaboration.

The FAOSTAT FVC domain provides data on the shares of food value across industries and primary factors
belonging to the food value chains identified through different food expenditure measures. It is limited in
terms of countries and territories covered (65) and years available (from 2005 to 2015). Work is ongoing
to expand both the country and time coverage.

A detailed description of the estimation methodology is provided in Section 2, followed by the

presentation in Section 3 of the type of results that can be obtained with the methodology. Strengths and
limitations of the current dataset and planned improvements are discussed in Sections 4 and 5,
respectively.

1.2 Methodological background: “food dollar” and “Global Food Dollar”

The FVC domain estimates the distribution across the farm and post-farm gate value chain of domestic
food expenditures.

The FVC identifies three food value measures, which reflect the domestic portion of personal consumption
expenditures, at purchaser prices, related to three different food systems: (i) Food At Home (FAH), (ii)
Food and Tobacco At Home (FTAH), and (iii) Food and Accommodation Away From Home (FAAFH).

FAH refers to the domestic personal consumption expenditures for food consumed at home. The FTAH
measure is similar in its target, but refers to a broader set of expenditure categories, inclusive of tobacco,
as food and tobacco expenditures are not always separable in the original data. FAAFH, instead, refers to
domestic personal consumption expenditures for food consumed away from home (e.g. in restaurants)
and includes expenditures for accommodation in all the cases where the two types of expenditures were
not separable in the original data. The Food at Home and the Food Away from Home measures are
therefore distinct and complementary.

2
In the FVC methodology applied, the value added of the food value chain industries associated with the
three food value expenditures described above is then decomposed by industry groups (agriculture,
forestry and fishing; manufacture of food, beverages and tobacco; transportation and storage; wholesale
and retail trade, as shown in Figure 1) and primary factors contributing to the industry value added
(imports, labour, gross operating surplus, taxes) using a methodology based on Leontief’s matrix reduction
method (Leontief, 1967). The FVC domain is the result of an input-output modelling based on supply and
use tables and official SNA data and information. Currently the domain covers 65 countries and territories
for the period 2005–2015: 65 for FAAFH, 37 for FTAH, and 10 for FAH. As a result, 28 countries are covered
only by FAAFH, 27 countries are covered by both FAAFH and FTAH and 10 countries are covered by all
three measures, as shown in Figure 2.

The FVC complies with the main international statistical standards and classifications on economic
statistics, such as the ISIC, the System of Environmental-Economic Accounting (SEEA) and the SNA, and
can therefore be scaled up at the global level when adequate information is made available by countries.

Figure 2. Data coverage of the food value chain domain by measure

Source: FAO. 2022. FAOSTAT: Industry and primary factors decomposition. In: FAO. Rome. [Cited April 2024].
https://www.fao.org/faostat/en/#data/GFDI based on UN Geospatial. 2020. Map geodata [shapefiles]. New York,
USA, UN.

The FVC domain methodology is built on the USDA-ERS food dollar series (Canning, 2011), published on a
yearly basis since 1947 with the aim of “achieving a more efficient and orderly marketing, as well as
reducing the price spread between the producer and the consumer” as stated in the 1946 Agricultural
Marketing Act, and the more recent Cornell University and USDA-ERS research paper on the post-farm
gate food value chain (Yi, Meemken, Mazariegos-Anastassiou et al., 2021).

3
1.2.1 The food dollar series
The food dollar series measures annual expenditures on domestically produced food by individuals living
in the United States of America. In particular, the industry group series identifies the value added from
distinct food supply chain industry groups to one food dollar (that is, the marginal contribution of each
industry group to the final food product), as shown in Figure 3.1

Figure 3. Industry group series in the food dollar analysis

Source: Canning, P. 2011. A Revised and Expanded Food Dollar Series: A Better Understanding of our Food Costs.
ERR-114. Washington, DC, USDA-ERS. https://www.ers.usda.gov/webdocs/publications/44825/7759_err114.pdf

The USDA-ERS analysis on the food dollar industries is complemented with the marketing bill series, which
“identifies the distribution of the food dollar between farm and marketing shares”, and the primary factor
series, which “focuses on the distribution of the food dollar in terms of U.S. worker salaries and benefits,
rents to food industry property owners, taxes, and imports” (Canning, 2011). The food dollar series
supplies a detailed analysis on the national supply chain trough a methodology that is based on input-
output tables.

1
Note that Figure 3 refers to the food dollar industry decomposition, while Figure 1 refers to the food value chain FAOSTAT domain. The food
dollar analysis is applied only to the United States of America, as the granularity of the source data allows for a more detailed industry
decomposition than in the food value chain domain, which is applied instead to 65 countries and territories.

4
1.2.2 The Global Food Dollar
Yi, Meemken, Mazariegos-Anastassiou et al. (2021) apply Leontief’s inverse multiplier, the same
methodological principle as in the food dollar series, to a larger set of countries2 in their paper, which
represents the second critical point of reference in the FVC methodological background. They develop a
standardized method and data series to estimate the distribution of consumer food expenditures
between activities on farms and in the post-farm gate value chain. The focus of this analysis is the
distinction between national domestic expenditure of food accruing to the agricultural sector (“farm”) or
to all the other food industries (“post-farm gate”). It identifies the agricultural shares in two food value
chains, the “Food at Home”, which refers to the agricultural share on domestic personal consumption
expenditures for food consumed at home, at purchaser prices, and the “Food and Accommodations Away
From Home” which instead refers to the “farm” share on domestic personal consumption expenditures
for food consumed away from home.

1.3 The joint FAO–Cornell upscaling exercise and the industry and primary factors
decomposition
Since May 2021, FAO has collaborated with USDA and Cornell University to further upscale and enrich the
Global Food Dollar methodology with a special focus on middle- and low-income countries, and on the
industry and primary factors decomposition. FAO has led a stocktaking exercise and has collaborated with
international and regional organizations such as the Asian Development Bank (ADB), the United Nations
Statistics Division (UNSD), the Economic Commission for Latin America and the Caribbean (ECLAC), the
Economic and Social Commission for Asia and the Pacific (ESCAP) and the United Nations Economic
Commission for Africa (UNECA), as well as National Statistical Offices (NSOs) and relevant Ministries, to
coordinate the data collection of supply and use tables (SUTs), and input-output tables (IOTs).

In collaboration with Cornell University, FAO has translated in a unique technical and programming
language (STATA) the Global Food Dollar methodology and has applied the resulting scripts to the larger
datasets. While the previous USDA-ERS–Cornell activity was focusing on the farm and post-farm
component of the FVC, the FAO–Cornell collaboration has allowed a more precise and accurate industry
and primary factors analysis of the three food value measures: (i) Food At Home (FAH), (ii) Food and
Tobacco At Home (FTAH), and (iii) Food and Accommodation Away From Home (FAAFH), further described
in Section 2.4.3, for 65 countries and territories from 2005 to 2015.

Based on a literature review,3 a previous data collection exercise (in May–August 2021) and the related
data assessment, FAO has developed a conceptual framework for deriving IOTs from SUTs that has been
implemented by Cornell University. Converting SUTs to IOTs is a major source of information and data for
Africa and in general countries that are not members of the Organisation for Economic Co-operation and
Development (OECD).

2
These countries are those covered by the OECD input-output tables: Argentina; Australia; Belgium; Brazil; Brunei Darussalam; Bulgaria;
Cambodia; Canada; Chile; China, mainland; China, Hong Kong SAR; China, Taiwan Province of; Colombia; Costa Rica; Croatia; Cyprus; Czechia;
Denmark; Estonia; Finland; France; Germany; Greece; Hungary; Iceland; India; Indonesia; Ireland; Israel; Italy; Japan; Kazakhstan; Latvia; Lithuania;
Luxembourg; Malaysia; Malta; Mexico; Morocco; Netherlands (Kingdom of the); New Zealand; Norway; Peru; Philippines; Poland; Portugal;
Republic of Korea; Romania; Russian Federation; Saudi Arabia; Singapore; Slovakia; Slovenia; South Africa; Spain; Sweden; Switzerland; Thailand;
Tunisia; Türkiye; United Kingdom of Great Britain and Northern Ireland; United States of America; and Viet Nam.
3
European Commission (2008); UNSD (2018); Miller and Blair (2022).

5
 2 Methodology
2.1 National accounts data used to derive FVC domain information
The Global Food Dollar analysis is based on nationally produced IOTs and SUTs. Applying national datasets
built through the SNA allows the replication of the Global Food Dollar analysis, when adequate
information is made available. NSOs regularly produce SUTs and IOTs, which are essential tools for
computing the national gross domestic product (GDP). Putting together SUTs and IOTs allows an accurate
analysis of the production and use of goods and services (products), and of the income generated through
this process.

Figure 4 shows an example of a supply table, which is a product by industry account, with products in rows
and industries and imports in columns. A supply table records the supply of goods and services by product
and by type of supplier, distinguishing supply by domestic industries and imports from those of other
countries. The last row shows the total output by industry, total imports and total supply. The last column
reports the total supply by product, consisting of domestic and imported products.

Figure 4. Section of a supply table

Industries Industries
Other Service Imports Total supply
Products Agriculture
industries activities
Agricultural products 270 30 50 20 370
Other industrial products 10 430 100 50 590
Services 20 40 550 30 640
Total 300 500 700 100 1 600
Source: Author’s own elaboration.

Complementary to the supply table, the use table (example shown in Figure 5) shows the use of goods
and services by product and by type of use, i.e. products for intermediate consumption by industries and
services, and products for final consumption by households. Gross fixed capital formation (GFCF) and
exports are also recorded in this accounting framework as products use. The use table records (by row)
the value added by industry. The SUTs are linked by the key accounting identity that total use equals total
supply (1 600 units in the examples).

Figure 5. Section of a use table

Industries Industries Final uses
Total
Other Service Final supply
Products Agriculture GFCF Exports
industries activities consumption
Agricultural products 34 59 143 81 21 32 370
Other industrial products 106 119 77 123 103 62 590
Services 70 112 75 291 61 31 640
Total 210 290 295 495 185 125 1 600
Value added 90 210 405 705
Output 300 500 700 1 500
Source: Author’s own elaboration.

6
 A simplified example of an industry-by-industry IOT is shown in Figure 6. It is a square matrix that includes
information related to the intra-industry relationships, final demand expenditures and value added
produced by industries.4

Figure 6. Simplified input-output table

Industries Industries Final uses
Other Service Final Uses
Industries Agriculture GFCF Exports
industries activities consumption
Agriculture 26.78 43.7 111.59 73.44 19.25 25.24 300.00
Industry 78.17 90.47 72.33 121.43 85.28 52.32 500.00
Service activities 75.05 115.83 96.06 295.13 75.47 42.45 700.00
Imports 30.00 40.00 15.00 5.00 5.00 5.00 100.00
Value added 90.00 210.00 405.00 0.00 0.00 0.00 705.00
Output 300.00 500.00 700.00 495.00 185.00 125.00
Source: Author’s own elaboration.

SUTs and industry-by-industry IOTs are all needed to derive the complete Global Food Dollar
methodology. However, it is possible to derive the FAAH and the related industry and primary factors
decomposition when only IOTs are available.

2.2 Food value measures and related data needs

The FAOSTAT Food Value Chain domain computes three different “food value measures”: (i) FAH, (ii) FTAH
and (iii) FAAFH. These food value measures refer to the domestic expenditures on three different food
baskets, belonging to two separate groups:

a. Consumption at home

i. FAH: refers to domestic personal consumption expenditures for food consumed at home, at
purchaser prices.

ii. FTAH: is similar to FAH, but includes tobacco, which is not an edible commodity. Moreover,
tobacco has, compared with most of other items included in the basket, a high retail price.

b. Consumption away from home

iii. FAAFH: reflects domestic personal consumption expenditures for food consumed away
from home. It includes both purchases in retail food stores and purchases in food service
establishments (for example, restaurants and hotels).

It follows that the expenditure quota for food at home is not included in the food away from home
measure.

4
IOTs describe the sale and purchase relationships between producers and consumers within an economy; they can be either industry by industry,
as in the example, or commodity by commodity. A more detailed description of the different types of IOTs may be found in UNSD (2018). On
deriving IOTs from SUTs, see European Commission (2008).

7
The ERS-food dollar methodology is based on a type I input-output (IO) multiplier model, or Leontief input-
output model.5 In this economic model, the gross industrial output (x) and the final consumption
expenditure (y) are linked through a multiplier matrix (L), such that x = L · y.

L is the Leontief (or total requirements) matrix, containing the whole inter-industry transactions
coefficients. The application of the type I IO multiplier model allows for the identification of the desired
supply chains related to FAH, FTAH and FAAFH.

In order to identify the final expenditures of the desired food value chain (y(FAH; FTAH; FAAFH)), the food-share
vector, denoted as s, containing the share measures of each element in the final demand vector (y) is
needed: (s(FAH; FTAH; FAAFH))″y = y(FAH; FTAH; FAAFH).6 Assuming that the supply and use of agricultural
industries are fully depicted in the first column and row of the Leontief matrix L, it follows that the gross
agricultural share of the desired food value chain is measured as in equation (1) below:

gross farm share(FAH; FTAH; FAAFH) = x(FAH; FTAH; FAAFH)/y(FAH; FTAH; FAAFH) (1)
For example, if food expenditure refers to the food at home and is USD 100 billion (denominator: y(FAH)),
and the farm gross output (numerator: x(FAH)) to satisfy the food at home demand is USD 1 billion, then
the gross farm share of FAH is 1 percent. Therefore, the FAH expenditure (USD 100 billion) multiplied by
1 percent results in the monetary values received by farmers and linked to FAH spending. The three food
value measures (FAH, FTAH, and FAAFH) are computed with the same methodology7 but substantially
differ for the reference food expenditure y, the denominator of simplified equation (1).

As the reference expenditures (y) are different, they are also computed using different variables. The
yFAAFH are derived from IOTs, while the yFAH and the yFTAH refer to national consumption expenditure on
food consumed at home, inclusive of the trade margin between the producer (e.g. the smallholder selling
commodities at farm gate prices) and the consumer (who buys at retail prices). In providing a more precise
measure of the food nationally produced and consumed in a country, FAH and FTAH require larger and
more granular data that is available in the SUTs in basic and purchaser’s prices.

Food and Accommodation Away From Home should also be distinguished from Food Away From Home
(FAFH).8 The real complementary measure of Food At Home is indeed Food Away From Home. Ideally, our
analysis should be concentrated only on FAH and FAFH. However, data granularity very often does not
allow to separate the tobacco component from the at home measures, or accommodation from food
expenditure in the away from home measures. Indeed, in most IOTs, accommodation and food services
are reported together, covering both ISIC divisions 55 and 56.

5
One strength of Leontief input-output economics is that output multipliers measuring the combined effects of the direct and indirect
repercussions of a change in final demand can be readily calculated. In particular, the type I multiplier, used in the methodology, shows industry-
to-industry transactions, and explains direct and indirect effects of exogenous changes on outputs of sectors in the economy. Additional details
on the ERS methodology and its application over countries can be found in Canning (2011) and Yi, Meemken, Mazariegos-Anastassiou, et al.
(2021). For additional information on input-output multipliers, see Miller and Blair (2022) and Steenge and van den Berg (2007).
6
In this paper, the symbol ″ means diagonalized vector.
7
On top of the gross farm share, the methodology further nets out farm to farm direct and indirect sales and nets out imports. For further details
on the methodology, see Yi, Meemken, Mazariegos-Anastassiou et al. (2021).
8
Yi, Meemken, Mazariegos-Anastassiou et al. (2021).

8
2.3 The data system
Information collected from national accounts data as described in Section 2.1 is organized by country and
by year, according to a specific data system. It is constituted by an IOT, which can be described as a
portioned matrix comprising the following matrices: Z, Y and V (as shown in Figure 7). These three
matrices represent the building blocks of the total gross output and the basis for the decomposition
procedures.

Figure 7. Food value chain data system

Z matrix Y matrix
Intermediate use (industry by industry) Final demand (industry by sector)
Household
personal Private and Government
consumption public consumption
Industry 1 Industry 2 Industry 3 Industry 4 ... Industry n expenditure investment expenditure Gross export
Industry 1 Industry 1
Industry 2 Industry 2
Industry 3 Industry 3
Industry : Industry :
Industry n Industry n

V matrix
Value added (factor by industry)
Industry 1 Industry 2 Industry 3 Industry 4 ... Industry n
Labour
Gross operating surplus
Output taxes
International suppliers

Source: Author’s own elaboration

In such a data system, the sum by row of Z and Y is equal to the total use of industry output, while the
sum by columns of Z and V produces the total industrial supply. By construction, total supply equals total
use and both equal to the total gross output.

The Y matrix shows four expenditure sectors, namely: (i) Household personal consumption expenditure;
(ii) Private and public investment; (iii) Government consumption expenditure and (iv) Gross export that are
consumers of the food produced by industries. The V matrix shows the four factors that contribute to the
value added of each industrial sector: (i) Labour, (ii) Gross operating surplus, (iii) Output taxes and (iv)
International suppliers. The Z matrix is an industry-by-industry matrix, where the number of industrial
sectors may vary according to the country data availability.

For example, a country taken from the OECD database has 36 industries of reference, while in the case of
Nepal, the Z matrix reports 25 industries. A complete list of industries used in the FVC domain is provided
in Annex I, while a more complete description of the Z matrix is included in Section 2.3.1.

2.3.1 The Z matrix

The Z matrix shows the intra-industry relationship. This matrix has industries in the rows and industries in
the columns and reports how much from each industry is produced and used by other industries. From
this matrix the direct coefficients are derived, which express the amount of output from one sector used
in a different sector to produce one unit of output.

9
2.3.2 The Y matrix
The Y matrix is the final demand matrix. While the Z matrix refers to the intermediate consumption, the
Y matrix defines the final consumption. The Y matrix has industries in the rows and four sectors in the
columns, namely: (i) Personal consumption expenditure, (ii) Private and public investment (GFCF),9 (iii)
Governmental expenditure of goods and services, and (iv) Exports of goods and services, that collectively
represent the final consumption expenditure. From the Y matrix, processed as described in Section 2.4,
the food expenditures are derived, which, together with the Leontief multiplier L allow for the estimation
of the food industry gross income X, according to equation (2):

𝐿Y = X (2)
Where:

𝐿 = Leontief multiplier; Y= Final consumption; and X = Industrial gross output

2.3.3 The V matrix

The V matrix records value added payments to primary factors, by industries. It has the primary factors
accruing to the industries value added in the rows, and the industries relevant for our analysis in the
columns. The four primary factors contributing to the industrial value added are:

1. Gross operating surplus: defined in the OECD Glossary of Statistical Terms as “the surplus generated
by operating activities after the labour factor input has been recompensed” (OECD, 2008). When an
industry has purchased all of its inputs and paid its labour force, the remainder from their sales
proceeds is the gross operating surplus, which is typically distributed to capital.
2. Labour: refers to payments to labour (salaries, wages accruing to the labour production factor).
3. Output taxes: covers payment of fees, taxes or sales taxes, which in the accounting framework is the
same as setting aside some operating surplus to pay indirect taxes. This does not include taxes on
labour income or capital income.
4. International suppliers: imports.

2.4 Leontief application for the derivation of food value measures

Three main types of multipliers are used within input-outputs models:

• Type I focuses on industry-to-industry transactions but does not consider the induced effects (for
example, it may consider that most part of the analysed national country is self-sufficient).
• Type II includes in the intra-industry transaction model the effect of household expenditures. This
model is particularly useful in analysing small regions in the short term, as public and private
investments are not amortized.
• Type III includes type II and amortized public and private investments input-output multipliers. It
is generally used for long-term and regional scenarios.

Type I is the most suitable approach when dealing with official data and statistics, and when no local or
regional forecast is needed, as in the FVC domain. In the type I input-output multiplier model, only inter-
industry transactions, recorded in the Z matrix described in Section 2.3.1 are considered endogenous.

9
GFCF refers to the acquisition of produced assets (including purchases of second-hand assets), including the production of such assets by
producers for their own use, minus disposals.

10
Final demand (recorded in the Y matrix described in Section 2.3.2) is considered an exogenous injection,
or input, into the system. Finally, the V matrix, representing the industry value added, is counted as a
leakage of the system and is determined following Walras’s law, according to which everything that is
produced in an economy, in a perfect competitive market, is also bought. Figure 8 shows the relationship
between the Y, Z and V matrices.

Figure 8. Type I input-output multiplier model

Y matrix Z matrix V matrix

• Final • Industrial • Value

demand production added

Source: Author’s own elaboration

In line with these assumptions, to compile the multiplier model used in the Global Food Dollar analysis,
the Y and V matrices (exogenous accounts) are reduced to vectors by summing each of their columns into
one row and the endogenous inter-industry transaction matrix Z is normalized column-wise by its import-
inclusive gross industry output X, which is the transpose of the sum of the column of Z and V matrices
defined in equation 3 below:

X = [𝑐𝑜𝑙𝑠𝑢𝑚(𝑍) + 𝑐𝑜𝑙𝑠𝑢𝑚(𝑉 )]′ (3)

2.4.1 Dealing with imports trade and transportation margins

Imports are dealt in the Global Food Dollar methodology as a specific vector m that defines the monetary
amount of imported goods by industry type with the industries in rows and the values in the column.
Then, the Global Food Dollar analysis associates the industrial imports with the purchase from final
consumers and creates the YM matrix (equation 4):

𝑌𝑀 = [𝑌 𝑚] (4)
As an additional step, from the final consumption related to the Food At Home (YFAH) and the Food and
Tobacco At Home (YFTAH) matrices we extrapolate the following variables:10

1) Trade, which represents the retail and wholesale trade margin; and

2) Tran, the transportation margin.

The trade and transportation parameters are fundamental to assess the trade margins, which are
reflected in the difference between basic and retail prices and are a key element of the food value chain
distribution.

Finally, the value free on board (VFOB_mil), which is the farm value of the purchase prior to mark-up, the
personal consumption expenditure (PCE), the government expenditure (GOV) and the exports (EXP)

10
The applied STATA command for this step of the data processing is “mkmatz“. This command allows to convert and extrapolate variables to
matrix and vice versa.

11
elements defining the Y matrix allow for the application of the Leontief principle expressed in
equation (2).

2.4.2 Leontief model and the direct and indirect requirement matrices
The fundamental purpose of the Leontief model is to analyse the interdependences of industries in an
economy. In fact, both the direct requirement matrix A, which is the coefficient matrix of industrial inputs
needed by industry type for one unit of additional product, and the total requirement matrix L, which
allows to define the total industrial output needed to satisfy a unit of final demand, are both derived from
the industry-by-industry matrix Z of the data system.

To illustrate the coefficient matrices concept, we assume two sectors (𝑋1 , 𝑋2 ) in the economy, where:
𝑥1 = 𝑜𝑢𝑡𝑝𝑢𝑡 𝑜𝑓 𝑠𝑒𝑐𝑡𝑜𝑟 1 ; and 𝑥2 = 𝑜𝑢𝑡𝑝𝑢𝑡 𝑜𝑓 𝑠𝑒𝑐𝑡𝑜𝑟 2.

The total output of a two-sectors economy can be expressed as:

𝑋1 = 𝑥11 + 𝑥12 + 𝑌1 (5)

𝑋2 = 𝑥21 + 𝑥22 + 𝑌2
Where 𝑋1 is the total amount produced by sector 1, 𝑥11 is the output of sector 1 used as input for the
same sector, 𝑥12 is the output of sector 1 used as input of sector 2 and 𝑌1 is the output of sector 1 used as
final demand. Similarly, 𝑋2 is the total amount produced by the second sector, 𝑥21 is the output of sector
2 used as input for sector 1, 𝑥22 is the output of sector 2 used as input for sector 2 and 𝑌2 is the output of
sector 2 used as final demand. This is a simplification of the granularity of information dealt with in the
original Leontief model (with over 500 industrial sectors) and in the FVC data source (over 36 sectors in
the OECD database and more than 100 in the original dataset of Chile for 2018). This simplified model,
however, allows to define:
𝑥𝑖𝑗 𝑦𝑖𝑒𝑙𝑑𝑠 (6)
= 𝑎𝑖𝑗 → 𝑥𝑖𝑗 = 𝑎𝑖𝑗 𝑋𝑗
𝑋𝑗

Where the input coefficient 𝑎𝑖𝑗 indicates the amount of output of sector i used in sector j to produce one
single unit of output.

Therefore, equation (6) can be rewritten as:

𝑋1 = 𝑎11 𝑋1 + 𝑎12 𝑋2 + 𝑌1 (7)

𝑋2 = 𝑎21 𝑋1 + 𝑎22 𝑋2 + 𝑌2
Using simple matrix algebra, we can arrange equation (8) as follows:
𝑋1 𝑎11 𝑎12 𝑋1 𝑌1 (8)
= [𝑎 𝑎22 ] 𝑋2 + 𝑌2
𝑋2 21
𝑎11 𝑎12 11
Where 𝑎
21 𝑎22 is the coefficient matrix generally named A. In addition, equation (9) can be rewritten
as:
(𝐼 − 𝐴 ) 𝑋 = 𝑌 (9)

11
The direct requirement matrix is also named A in this paper.

12
 Or

𝑋 = (𝐼 − 𝐴)−1 𝑌 (10)
Where 𝑋 is the output in monetary terms, 𝐼 is the identity matrix; 𝑌 is the total final demand and
(𝐼 − 𝐴)−1 is the Leontief multiplier, generally named 𝐿.

As shown in equations (2) and (10) the Leontief multiplier is extremely important as it allows, knowing the
value of final expenditure 𝑌, to derive the output production value and vice versa.

The Leontief multiplier, or total requirement matrix, is a coefficient matrix, explaining for each industry
the amount of gross output needed to satisfy a unit of additional final demand. The columns of the matrix
show the input requirements, both direct and indirect, for all industries, generated for one unit of product
(spillover effect).

In the FVC domain, 𝐴 is computed as equal to the 𝑍 matrix multiplied by the inverse diagonal of the gross
output 𝑋:

𝐴 = 𝑍 (𝑋′′)−1 (11)
Where Z is the intra-industry transaction matrix described in Section 2.3.1. The resulting matrix of
coefficients A tells how many dollars of inputs are needed to produce one dollar of monetary output.

In addition, based on equation (10), 𝐿 is equal to:

𝐿 = (𝐼 − 𝐴)−1 (12)

2.4.3 Computation of the three food value chain measures

Matrices A, L, Z, V and Y (including the YFAH and YFTAH expenditures measures) are recomputed and
updated to accurately reflect imports and exports treatment in line with the SNA.

When data availability allows, it is important to identify expenditure only for national food consumed at
home (FAH) or, as a proxy, for food and tobacco at home (FTAH). However, this is rarely possible as already
discussed in section 2.2, and this is why the current FVC domain also features the FAAFH food value
measure. Despite the above-mentioned specificities related to at home or away from home food systems,
the three food value measures are computed using equation (2).

2.5 Industry and primary factors decomposition

The relationship between the technical coefficients matrix 𝐴, the total output X and the final demand 𝑌,
for all sectors in the economy can be expressed as in equation (13) derived from equations (2) and (12):
(𝐼 − 𝐴 )𝑋 = 𝑌 (13)
To apply the decomposition model to the food value chain (Canning, 2016) we decompose the original
matrices in portioned matrices and we distinguish the industries of interest, which are the supply chain
industries, and we include them into group 1, while the non-supply chain industries are combined into
group 2. The partitioned version of equation (13) is shown in equation (14) below:
(𝐼 − 𝐴11 ) − 𝐴12 𝑥1 (14)
[ 1 ][ ]
−𝐴21 (𝐼2 − 𝐴22 ) 𝑥2

13
With all industries combined into two groups, 𝐴11 and 𝐴22 are square matrices whose technical
coefficients govern internal flows. They represent the exchange of goods and services from industries
within groups 1 and 2. Matrices 𝐴12 and 𝐴21 need not be square and describe the direct requirements of
industries of group 1 for outputs of group 2 and vice versa. By multiplying both sides of equation by
𝐿 = (𝐼 − 𝐴)−1 , we obtain equation (15):

𝑥1 (𝐼 − 𝐴11 ) − 𝐴12 −1 𝑦1 (15)

[𝑥 ] = [ 1 ] [𝑦 ]
2 −𝐴21 (𝐼2 − 𝐴22 ) 2

Applying the Leontief definition of 𝐵 as equal to (𝐼−𝐴) in equation (15) leads to equation (16):
𝑥1 𝐵 − 𝐵12 𝑦1 (16)
[𝑥 ] = [ 11 ][ ]
2 𝐵21 −𝐵22 𝑦2
Based on Leontief (1967) and Canning (2016), this method enables to describe the entire economy in
terms of the supply chain industries of interest without distorting any of the basic structural relationships.
The value added can then be decomposed by supply chain industries and by primary factors.

In national accounts in general and in the Global Food Dollar methodology in particular, imports that are
used for immediate re-export and therefore do not contribute to the national production system
(transhipment), do not need to be accounted. In line with these principles, the Global Food Dollar
methodology creates the ad hoc “export assembly” industry, which aggregates all exports of all industries.

The industry and primary factors subdomains identify five supply chain industries, described in Table 1
below according to their ISIC section and division.

Agrifood value chain industries classification according to ISIC Rev 4

Agri-food value chain industry ISIC section ISIC division
Accommodation and food service activities I 55, 56
Agriculture, Forestry and Fishing A 1-3
Manufacture of food, beverages and tobacco products C 10, 11, 12
Transportation and storage H 49, 50, 51, 51, 53
Wholesale and retail trade G 46, 47
Source: FAO. 2022. Industry and primary factors decomposition. In: FAOSTAT. Rome. [Cited April 2024].
https://www.fao.org/faostat/en/#data/GFDI

A similar approach is applied to further analyse by industry the factors accruing to the value added
production: operating surplus, labour, taxes and imports.

14
3 Results
3.1 Strengths and limitations of the Food Value Chain domain
An important strength of the FVC domain is its consistency with the main international statistical
standards, and namely the SNA, ISIC and SEEA. These standards are in fact applied in all United Nations
Members and therefore the FVC methodology can be scaled up at the global level when adequate
information is made available.

Another important analytical strength, jointly developed by FAO and Cornell University, is the detailed
analysis, by sectors and by primary factors (see Sections 2.4 and 2.5) that allows to shed light on each
single phase of the farm and post-farm food value chain, as well as on the primary factors contributing to
each food industry value added.

However, this level of analytical detail is also the main limitation of the methodology, as it requires a high
level of data granularity. As a result, the analysis covers 65 countries and territories from 2005 to 2015 for
FAAFH, 37 for FAAFH and FTAH and 10 for FAAFH, FTAH and FAH (see Section 1.2).

3.2 Main findings12

For 65 countries and territories that represent more than 90 percent of the global gross income,
6.7 percent of consumer expenditure on food consumed away from home went to the agricultural sector
over the 2005–2015 period. The share declined by about 1 percentage point between 2005 and 2015, and
the declining trend is expected to continue. The value of the entire food value chain (from agriculture to
retail trade) went up by around 35 percent between 2005–2007 and 2013–2015, from USD 1.4 trillion to
USD 1.9 trillion. This contrasts with the slow decline of the average share of agriculture, suggesting a lower
“farm” quota in the overall food supply chain, as shown in Figure 9. Farm shares dropped slightly in 2008–
2010) and rebounded in 2009–2011 and 2010–2012. The high food prices seen in 2007–2009 during the
food prices crisis may have prompted small producers in low-income countries to increase agricultural
production (FAO, 2009).

12
This section draws from FAO (2022).

15
Figure 9. Value of the food value chain and average share of agriculture using the Food and
Accommodation Away From Home measure
2.0 8

1.5 6

Percentage
USD trillion

1.0 4

0.5 2

0.0 0

Total value Share of agriculture

Source: FAO. 2022. Industry and primary factors decomposition. In: FAOSTAT. Rome. [Cited April 2024].
https://www.fao.org/faostat/en/#data/GFDI

Figure 10 shows that the agricultural share in FAAFH varies significantly between countries, ranging from
0.6 percent in Singapore to 28 percent in India, which is consistent with the correlation between the
agricultural share in FAAFH and the share of agricultural value added in GDP at the country level (FAO,
2022). The countries with the smallest spread in the FAAFH agricultural share (Singapore, the United
Kingdom if Great Britain and Northern Ireland and Austria) were also among the ones with the smallest
FAAFH agricultural share. India had both the largest FAAFH agricultural share and the largest spread.

As seen on Figure 11, wholesale and retail trade represented 46–48 percent of the total food industries
for FAH and FTAH. Manufacturing of food and agriculture had similar shares of 23–25 percent and 20–
22 percent, respectively. The breakdown by industry for FAAFH is quite different since accommodation
and food service activities accounted for two-thirds of the total, thus reducing the share of the other
industries significantly (12 percent for wholesale and retail trade, and 11 percent for manufacturing of
food).

16
Figure 10. Average share of agriculture using the Food and Accommodation Away From Home measure
(2005–2015), top and bottom countries

Percentage
0 5 10 15 20 25 30 35 40 45

Singapore

China, Hong Kong SAR

United Kingdom of Great Britain and Northern Ireland

Austria

United States of America

China, mainland

Malaysia

Indonesia

Viet Nam

India

Note: Nepal is not shown as data are available for only one year.

Source: Based on FAO. 2022. Industry and primary factors decomposition. In: FAOSTAT. Rome. [Cited April 2024].
https://www.fao.org/faostat/en/#data/GFDI

Figure 11. Food value measures by industry (2005–2015 average)

100
90
80
46% 48%
70
67%
Percentage

60
50 7%
9%
40
25%
30 23%
12%
20 3%
10 22% 20% 11%
7%
0
Food At Home Food and Tobacco At Home Food and Accommodation Away
From Home
Agriculture, forestry and fishing Manufacture of food, beverages and tobacco products
Transportation and storage Wholesale and retail trade
Accommodation and food service activities

Source: FAO. 2022. Industry and primary factors decomposition. In: FAOSTAT. Rome. [Cited April 2024].
https://www.fao.org/faostat/en/#data/GFDI

17
For the 65 countries and territories currently available in the FVC domain, the order of the primary factors
in the decomposition is the same in the three food value chain measures, even though the values vary
slightly. The larger factor is the gross operating surplus (36–42 percent of the total), followed by labour
(31–38 percent), imports (16–30 percent) and taxes (3–6 percent), as shown in Figure 12.

Figure 12. Food value measure by primary factors excluding accommodation and food service activities
(2005–2015 average)
100 3%
4% 6%
90

80 36%
42% 40%
70
Percentage

60

50
31%
40
34% 38%
30

20
30%
10 19% 16%
0
Food At Home Food and Tobacco At Home Food and Accommodation Away
From Home

Import Labour Operating surplus Taxes

Source: FAO. 2022. Industry and primary factors decomposition. In: FAOSTAT. Rome. [Cited April 2024].
https://www.fao.org/faostat/en/#data/GFDI

18
4 Potential applications
The FVC data can be employed in different analytical frameworks, such as economic, social and
environmental analyses, allowing for comparability across countries and geographical regions because of
their consistency with the main international statistical standards such as the SNA and the SEEA.

The FVC data and information can also provide useful insight for the social accounting matrix (SAM), which
focuses on transactions and transfers between different production activities, factors of production and
institutions (households, corporate sector and government) within a country, when the social and
economic analysis is food-oriented.13

The industry decomposition of the food value provides a static picture of the relative economic weight of
each sector with respect to the FVC that they belong to. Cross-country comparisons of the shares of the
food industries contribute to highlight the potential underlying economic and policy drivers leading to
differences in the relative weights of the sectors belonging to the same FVC. They can inform international
trade food policies, shedding light on the comparative advantages of national industrial sectors. For
example, FVC data may help evaluating whether if it is more economical for a country to process raw food
commodities locally or to export them; by describing the role of wholesale and retail trade in the FVC, FVC
data help better quantify the impact of a short versus a long food value chain.

The changes of economic dimensions over time inform about the evolution of the FVC and how they relate
to changes in sectoral or general economic policy measures. The same type of analysis can be applied to
the decomposed values of the primary factors employed by each sector in the FVC: they disentangle the
relationships between the economic and social dimensions, especially when the decomposed economic
value of primary factors can be directly associated with the related social component such as workers
(labour), companies (business) and governments (economic and social policies).

The longitudinal nature of FVC data offers the opportunity to build analytical models for programme
evaluation to quantify the performance of specific objectives or outcomes in all analytical frameworks.

The decomposition methodology can be applied to any dimension of the FVC provided data are available
in an input-output or supply-use format and could be further developed in terms of environmental-
extended SUTs and IOTs (UNSD, 2018). In this perspective, the availability of production factors or waste
data along the FVC, such as water and greenhouse gas emissions, would allow for the decomposition of
the relative value per sector and per primary factor, paving the way for deeper analytical applications to
other dimensions of the FVC, including the impact on the environment and natural resources.14

13
For a detailed description of the social accounting matrices and the importance of the food component in them, see Breisinger, Thomas and
Thurlow (2009). Additionally, a number of national SAMs are available in the International Food Policy Research Institute (IFPRI) Dataverse.
14
See Canning, Rehkamp and Yi (2022) or Camanzi, Alikadic, Compagnoni and Merloni (2017).

19
5 Way forward
The availability of IOTs and SUTs data usable in the FVC methodology is currently limited to most OECD
countries and some non-OECD countries. It is expected to upscale the FVC domain to more countries, in
collaboration with UN regional agencies as well as NSOs and Ministries of agriculture. While data
availability is currently the main limitation, it is possible in some cases to further adapt the methodology
to the data available in the countries. To this end, a methodology was recently developed that allows IOTs
to be derived directly from SUTs in basic and purchaser prices, since countries produce SUTs (especially in
Africa and Asia) more often than IOTs.

The conceptual framework for SUTs versus IOTs conversion has been elaborated by FAO and implemented
by Cornell University. In testing the conceptual model proposed by FAO, Cornell University has introduced
methodological advancements and in particular has developed an additional STATA module to derive the
trade margins, using the SUTs in basic and purchaser’s prices.

The upscaling exercise will focus on a continuous improvement of analytical tools, including open-source
languages such as Python and R and reducing when possible private software licences.

Following the upscaling procedure, the FVC data will be updated annually, expanding the country and
time coverage.

20
 Annex I Industries mapping to ISIC code
A. OECD database data mapping with ISIC
STATA label OECD ISIC
Z01T03 D01T03: Agriculture, forestry and fishing 01, 02, 03
Z05T06 D05T06: Mining and extraction of energy producing 04, 05, 06
products
Z07T08 D07T08: Mining and quarrying of non-energy 07, 08
producing products
Z09 D09: Mining support service activities 09
Z10T12 D10T12: Food products, beverages and tobacco 10, 11, 12
Z13T15 D13T15: Textiles, wearing apparel, leather and 13, 14, 15
related products
Z16 D16: Wood and products of wood and cork 16
Z17T18 D17T18: Paper products and printing 17, 18
Z19 D19: Coke and refined petroleum products 19, 19
Z20T21 D20T21: Chemicals and pharmaceutical products 20, 21
Z22 D22: Rubber and plastic products 22
Z23 D23: Other non-metallic mineral products 23, 23
Z24 D24: Basic metals 24
Z25 D25: Fabricated metal products 25
Z26 D26: Computer, electronic and optical products 26
Z27 D27: Electrical equipment 27
Z28 D28: Machinery and equipment, nec 28
Z29 D29: Motor vehicles, trailers and semi-trailers 29
Z30 D30: Other transport equipment 30
Z31T33 D31T33: Other manufacturing; repair and 31, 33
installation of machinery and equipment
Z35T39 D35T39: Electricity, gas, water supply, sewerage, 35, 36, 37, 38, 39
waste and remediation services
Z41T43 D41T43: Construction 41, 42, 43
Z45T47 D45T47: Wholesale and retail trade; repair of motor 45, 46, 47
vehicles
Z49T53 D49T53: Transportation and storage 49, 50, 51, 52, 53
Z55T56 D55T56: Accommodation and food services 55, 56
Z58T60 D58T60: Publishing, audiovisual and broadcasting 58, 59, 60
activities
Z61 D61: Telecommunications 61
Z62T63 D62T63: IT and other information services 66, 63
Z64T66 D64T66: Financial and insurance activities 64, 65, 66
Z68 D68: Real estate activities 68
Z69T82 D69T82: Other business sector services 69, 70, 71, 72, 73, 74, 75, 76, 77, 78, 79, 80, 81, 82
Z84 D84: Public admin. and defense; compulsory social 84
security
Z85 D85: Education 85
Z86T88 D86T88: Human health and social work 86, 87, 88
Z90T96 D90T96: Arts, entertainment, recreation and other 90, 91, 92, 93, 94, 95, 96
service activities
Z97T98 D97T98: Private households with employed persons 97, 98, 99

21
 B. Asian Development Bank database data mapping with ISIC
STATA label ADB IOT label (name) SUT label: ISIC codes
c1 c1: Agriculture, hunting, forestry, and fishing 01, 02, 03
c2 c2: Mining and quarrying 05, 06, 07, 08, 09
c3 c3: Food, beverages, and tobacco 10, 11, 12
c4Tc5 c4: Textiles and textile products 13, 14, 15
c4Tc5 c5: Leather, leather products, and footwear
c6 c6: Wood and products of wood and cork 16
c7 c7: Pulp, paper, paper products, printing, and 17, 18
publishing
c8 c8: Coke, refined petroleum, and nuclear fuel 19
c9 c9: Chemicals and chemical products 20, 21
c10 c10: Rubber and plastics 22
c11 c11: Other nonmetallic minerals 23
c12 c12: Basic metals and fabricated metals 24, 25
c13 c13: Machinery, nec 28
c14 c14: Electrical and optical equipment 26, 27
c15 c15: Transport equipment 29, 30
c16 c16: Manufacturing, nec; recycling 31, 32, 33
c17 c17: Electricity, gas, and water supply 35, 36, 37, 38, 39
c18 c18: Construction 41, 42, 43
c19Tc21 c19: Sale, maintenance, and repair of motor vehicles 45
and motorcycles; retail sale of fuel
c19Tc21 c20: Wholesale trade and commission trade, except 46
of motor vehicles and motorcycles
c19Tc21 c21: Retail trade, except of motor vehicles and 47
motorcycles; repair of household goods
c22 c22: Hotels and restaurants 55, 56
c23Tc26 c23: Transport and storage
c27 c27: Post and telecommunications 61
c28 c28: Financial intermediation 64, 65, 66
c29c31Tc35 c29: Real estate activities
c29c31Tc35 c31: Public administration and defense; compulsory
social security
c29c31Tc35 c32: Education 58, 59, 60, 62, 63, 68, 84, 85, 86, 87, 88, 90, 91, 92,
c29c31Tc35 c33: Health and social work 93, 94, 95, 96, 97, 98
c29c31Tc35 c34: Other community, social, and personal services
c29c31Tc35 c35: Private households with employed persons
c30 c30: Renting of M&Eq and other business activities 69, 70, 71, 72, 73, 74, 75, 77, 78, 79, 80, 81, 82

22
References
Barrett, C., Reardon, T., Swinnen, J. & Zilberman, D. 2022. Agri-food Value Chain Revolutions in Low-and
Middle-Income Countries. Journal of Economic Literature, 60(4): 1316–1377.
http://dx.doi.org/10.1257/jel.20201539

Breisinger, C., Thomas, M. & Thurlow, J. 2009. Social accounting matrices and multiplier analysis: An
introduction with exercises. Food Security in Practice technical guide 5. Washington, DC, International
Food Policy Research Institute. http://dx.doi.org/10.2499/9780896297838fsp5

Camanzi, L., Alikadic, A., Compagnoni, L. & Merloni, E. 2017. The impact of greenhouse gas emissions in
the EU food chain: A quantitative and economic assessment using an environmentally extended input-
output approach. Journal of Cleaner Production, 157: 168–176
https://doi.org/10.1016/j.jclepro.2017.04.118

Canning, P. 2011. A Revised and Expanded Food Dollar Series: A Better Understanding of our Food Costs.
ERR-114. Washington, DC, USDA-ERS.
https://www.ers.usda.gov/webdocs/publications/44825/7759_err114.pdf

Canning, P., Rehkamp, S., & Yi, J. 2022. Environmental input-output (EIO) models for food systems
research: Application and extensions. In: C. Peters & D. Thilmany, eds. Food Systems Modelling. Academic
Press. https://doi.org/10.1016/B978-0-12-822112-9.00014-X

European Commission. 2008. Eurostat Manual of Supply, Use and Input-Output Tables. Luxembourg,
Office for Official Publications of the European Communities.
https://ec.europa.eu/eurostat/web/products-manuals-and-guidelines/-/ks-ra-07-013

FAO. 2009. High food prices and the food crisis; experiences and lessons learned. Rome.
https://www.fao.org/3/i0753e/i0753e.pdf

FAO. 2021. Strategic Framework 2022-31. Rome. https://www.fao.org/3/cb7099en/cb7099en.pdf

FAO. 2022. Agriculture and the food value chain 2005–2015. FAOSTAT Analytical Brief Series No 51. Rome.
http://www.fao.org/3/cc2733en/cc2733en.pdf

FSCI. n.d. About Food Systems. In: Food Systems Dashboard. [Cited 29 February 2024].
https://www.foodsystemsdashboard.org/information/about-food-systems

Gómez, M.I., Barrett, C.B., Buck, L.E., De Groote, H., Ferris, S., Gao, H.O., McCullough, E., Miller, D.D.,
Outhred, H., Pell, A.N., Reardon, T., Retnanestri, M., Ruben, R., Struebi, P., Swinnen, J., Touesnard, M.A.,
Weinberger, K., Keatinge, J.D.H., Milstein, M.B. & Yang, R.Y. 2011. Research Principles for Developing
Country Food Value Chains. Science, 332(6034): 1154–1155. https://doi.org/10.1126/science.1202543

Henriksen, L.F. et al. 2010. Agro-Food Value Chain Interventions in Asia: A review and analysis of case
studies. Vienna, UNIDO.
https://www.unido.org/sites/default/files/2011-01/WorkingPaper_VC_AsiaFinal_0.pdf

Leontief, W. 1967. An Alternative to Aggregation in Input-Output Analysis and National Accounts. The
Review of Economics and Statistics, 49(3): 412–419

23
Miller, R.E. & Blair, P.D. 2022. Input–Output Analysis, Foundations and Extensions, third edition.
Cambridge, Cambridge University Press.

Oddone, N. & Padilla, R. 2014. Upgrading value chains through professional and supporting services:
lessons from three agro-industry chains in El Salvador and Guatemala. Mexico, United Nations.
https://hdl.handle.net/11362/36898

OECD. 2008. OECD Glossary of Statistical Terms. Paris, OECD Publishing.

https://doi.org/10.1787/9789264055087-en

Steenge, A.E. & van den Berg, R. 2007. Transcribing the Tableau Économique: Input-Output Analysis à La
Quesnay. Journal of the History of Economic Thought 29(3): 331–58.
https://doi.org/10.1080/10427710701514745

UNSD. 2018. Handbook on Supply and Use Tables and Input-Output Tables with Extensions and Applications.
New York, United Nations. https://unstats.un.org/unsd/nationalaccount/docs/SUT_IOT_HB_Final_Cover.pdf

Yi, J., Meemken, EM., Mazariegos-Anastassiou, V. et al. 2021. Post-farmgate food value chains make up
most of consumer food expenditures globally. Nat Food 2, 417–425. https://doi.org/10.1038/s43016-021-
00279-9

24
Contact:

Statistics Division – Economic and Social Development

[email protected]
www.fao.org/food-agriculture-statistics/resources/publications/working-papers/en/

Food and Agriculture Organization of the United Nations

Rome, Italy