The Number, Size, and Distribution of Farms, Smallholder Farms, and Family Farms Worldwide

Abstract

Numerous sources provide evidence of trends and patterns in average farm size and farmland distribution worldwide, but they often lack documentation, are in some cases out of date, and do not provide comprehensive global and comparative regional estimates. This article uses agricultural census data (provided at the country level in Web Appendix) to show that there are more than 570 million farms worldwide, most of which are small and family-operated. It shows that small farms (less than 2 ha) operate about 12% and family farms about 75% of the world’s agricultural land. It shows that average farm size decreased in most low- and lower-middle-income countries for which data are available from 1960 to 2000, whereas average farm sizes increased from 1960 to 2000 in some upper-middle-income countries and in nearly all high-income countries for which we have information.

Full text

The Number, Size, and Distribution of Farms, Smallholder Farms,
and Family Farms Worldwide
q
SARAH K. LOWDER, JAKOB SKOET and TERRI RANEY
*
Food and Agriculture Organization of the United Nations, Rome, Italy
Summary. —Numerous sources provide evidence of trends and patterns in average farm size and farmland distribution worldwide, but
they often lack documentation, are in some cases out of date, and do not provide comprehensive global and comparative regional
estimates. This article uses agricultural census data (provided at the country level in Web Appendix) to show that there are more than
570 million farms worldwide, most of which are small and family-operated. It shows that small farms (less than 2 ha) operate about 12%
and family farms about 75% of the world’s agricultural land. It shows that average farm size decreased in most low- and lower-middle-
income countries for which data are available from 1960 to 2000, whereas average farm sizes increased from 1960 to 2000 in some upper-
middle-income countries and in nearly all high-income countries for which we have information.
Such estimates help inform agricultural development strategies, although the estimates are limited by the data available. Continued ef-
forts to enhance the collection and dissemination of up-to date, comprehensive, and more standardized agricultural census data, includ-
ing at the farm and national level, are essential to having a more representative picture of the number of farms, small farms, and family
farms as well as changes in farm size and farmland distribution worldwide.
Ó2016 Food and Agriculture Organization of the United Nations. Published by ELSEVIER Ltd. This is an open access article under the
CC BY-NC-ND license (http://creativecommons.org/licenses/by-nc-nd/4.0/).
Key words — family farm, small farm, farm size, smallholder, farmland distribution
1. INTRODUCTION
Agricultural economists and other development specialists
generally agree that investing in agriculture is an effective
strategy for reducing poverty, inequality and hunger, espe-
cially in countries where the sector employs a large share of
the population (FAO, 2012; World Bank, 2007). There is con-
siderable debate regarding what type or scale of agriculture
should be promoted in order to most effectively achieve these
goals (Larson, Otsuka, Matsumoto, & Kilic, 2014). Many
advocates emphasize the importance of ‘‘smallholder farming”
or ‘‘family farming”, with claims often made that smallholders
or family farms are responsible for a large share of the world’s
food production (e.g., Fairtrade International, 2013)—or that
a large share of the food consumed in Africa and Asia is pro-
duced by smallholders in those regions (HLPE, 2013; IFAD &
UNEP, 2013). The terms smallholder and family farm are
often used interchangeably or in combination without clear
definitions. Lack of clarity regarding terminology as well as
the basic composition and diversity of the agricultural sector
is a serious barrier to effective policy dialog.
A 2010 special issue of World Development, devoted to
small farms,
1
classified the agricultural development literature
on farm size according to one or more of the following topics:
scale and efficiency; small farms, and poverty; and changing
agrarian structure and the future of small farms (Wiggins,
Kirsten, & Lambi, 2010). A number of sources, including arti-
cles from the special issue of World Development, state that
there are about 500 million farms smaller than 2 ha world-
wide, but mostly without providing documentation for the
claim. Other work provides evidence of trends and patterns
in average or median farm size and/or farmland distribution
(see Table 1), but it is often lacking documentation, is in some
cases out of date, and does not provide comprehensive global
and comparative regional estimates. This article makes use of
agricultural census data from 167 countries and territories to
provide comprehensive and well-documented estimates for
the number of farms worldwide. It also considers data for
more than 100 countries to assess the evolution of average
farm size as well as distribution of farms, small farms, and
family farms at global and regional levels and by country
income groups. All country-level data are provided in Web
Appendix. These estimates can inform agricultural develop-
ment strategies and policy decisions on a wide range of issues.
This article proceeds as follows. Section 2reviews the avail-
able literature on the number of farms, average farm size, and
farmland distribution. Section 3describes the data used for the
analysis. Section 4presents estimates of the total number of
farms in the world, as well as by region and income group clas-
sification. Section 5presents trends in average farm size by
region and income grouping. Section 6examines the distribu-
tion of farms and farmland at the global and regional level as
well as for country income groups. Section 7presents an esti-
mate of the number of family farms worldwide as well as the
qThe authors are economists in the Agricultural and Development
Economics Division (ESA) of FAO and are members of the team that
produces FAO’s major annual flagship publication, The State of Food and
Agriculture.
*This paper was prepared as background research for The State of Food
and Agriculture 2014, available at http://www.fao.org/publications/sofa/
en. The authors are grateful for the provision of agricultural census
reports by the Statistics Division of FAO (ESS) and helpful research
assistance from Saumya Singh and Raffaele Bertini. Feedback was
received from numerous colleagues in FAO; in particular, the authors
wish to thank Jairo Castano, Giorgi Kvinikadze, Adriana Neciu, Andre
´
Croppenstedt, Cheng Fang, Panagiotis Karfakis, Michelle Kendrick,
George Rapsomanikis, Kostas Stamoulis, and Dominique Van Der
Mensbrugghe. Any errors and omissions are those of the authors. The
views expressed in this publication are those of the author(s) and do not
necessarily reflect the views or policies of FAO.
World Development Vol. xx, pp. xxx–xxx, 2016
0305-750X/Ó2016 Food and Agriculture Organization of the United Nations.
Published by ELSEVIER Ltd. This is an open access article under the
CC BY-NC-ND license (http://creativecommons.org/licenses/by-nc-nd/4.0/).
http://dx.doi.org/10.1016/j.worlddev.2015.10.041
www.elsevier.com/locate/worlddev
1
Please cite this article in press as: Lowder, S. K. et al. The Number, Size, and Distribution of Farms, Smallholder Farms, and Family
Farms Worldwide
q
,World Development (2016), http://dx.doi.org/10.1016/j.worlddev.2015.10.041

Table 1. Published literature on average or median farm size and farmland distribution worldwide, 2010–14
Author, year of
publication
Data used Geographical Coverage Sample size Time period Findings:
Eastwood et al. (2010) Agricultural censuses Global – 1930–90 rounds The log of average farm size increased from 1950 to 1990 in North
America and Europe. It decreased from 1950 to 1990 in Asia and from
1970 to 1990 in Africa. There was no clear long-run trend for South
America
Hazell et al. (2010) Agricultural censuses Africa, Asia and Latin
America
16 countries 1970–2000 rounds Median farm size decreased in most countries considered
Deininger & Byerlee
(2012)
Literature review Selected land abundant
countries in Latin America,
Eastern Europe and Central
Asia, Southeast Asia and
Sub-Saharan Africa
– 1970–2000s Farmland distribution: In land abundant countries of Latin America
and Eastern Europe and Central Asia there has been an increase in
large scale farming. In South East Asia the palm oil industry has seen
an increase in the number of larger plantations or large firms
contracting with outgrowers, but rubber production has shifted from
being cultivated mainly by large plantations to production by
smallholders. In countries of Sub-Saharan Africa efforts to move
toward large-scale agriculture in the 1970s and ’80s largely failed and
small farming operations have persisted, with an increase in large-scale
investments in more recent years.
FAO (2013) Agricultural censuses Global 114 countries 1930–2000 rounds Average farm size: At the global level (106 country sample, with the
number of countries varying from year to year) the pooled average of
mean farm size and the median of mean farm sizes decreased from 1930
to 2000. This reflects a decrease in average farm size for most regions
except Europe. For the African region a decrease in average farm size
was registered from 1960 to 1980 followed by a slight increase from
1980 onward. Countries where an increase rather than decrease was
observed included Australia, New Zealand, the United States, Canada,
Argentina, and Uruguay
HLPE (2013) Agricultural censuses Global 81 countries 1930–2000 rounds Average farm size has decreased for the African region as a whole and it
has also decreased in China. Farmland distribution: For a sample of 81
countries, using data from the 2000 round, the HLPE report found that
73% of farms are smaller than 1 ha and 85% are smaller than 2 ha. On
average for 14 African countries 80% of holdings are smaller than 2 ha
in size; they operate about 25% of the agricultural land. In the
European Union 50% of farms are smaller than 2 ha in size and operate
only about 2% of the agricultural land
Masters et al. (2013) Rural and urban
population data
Africa and Asia – 1950–2050 Since 1950 average farm sizes have been decreasing for Africa and Asia
as a whole, but in recent years they have begun increasing for Asia as a
whole, while they will continue to decrease in Africa for quite some
time
Adamopoulos &
Restuccia (2014)
Agricultural censuses Global 63 countries 1990 round Average farm size is larger in countries with higher average per capita
GDP. Farmland distribution: In countries with high average incomes
farms larger than 20 Ha operate 70% of land, while in the poorest
countries 70% of land is operated by farms smaller than 5 ha
Jayne et al. (2014) Agricultural
censuses/surveys
Africa 12 countries 1980–2010 round Africa is typically characterized as land abundant (Deininger et al.,
2011), but the majority of the region’s uncultivated arable land is
concentrated in a few countries. In all land-constrained countries for
which data are available, average farm size has decreased. Most but not
all land abundant African countries have shown an increase in average
farm size
Note: ‘‘–”indicates data not available.
2 WORLD DEVELOPMENT
Please cite this article in press as: Lowder, S. K. et al. The Number, Size, and Distribution of Farms, Smallholder Farms, and Family
Farms Worldwide
q
,World Development (2016), http://dx.doi.org/10.1016/j.worlddev.2015.10.041

share of farmland operated by family farms. In conclusion,
Section 8considers the implications of the findings.
2. LITERATURE REVIEW
Nagayets (2005) used agricultural census data from FAO to
estimate that there are about 525 million farms of all sizes in
the world. Several other sources maintain that worldwide
there are about 500 million farms smaller than 2 ha (see for
example Hazell, Poulton, Wiggins, and Dorward (2010),
Wiggins et al. (2010), IFAD (2011) and HLPE (2013)), many
of them refer to Nagayets (2005). Unfortunately, Nagayet’s
paper is unpublished, and neither it nor any of the other
sources provide the information needed to reproduce the esti-
mates.
Table 1 summarizes the results of a literature review on
average or median farm size and farmland distribution. We
first consider the results on average or median farm size.
Eastwood, Lipton, and Newell (2010) considered unweighted
averages of the log of mean farm size by region from the
1950 to 1990 rounds of the agricultural census; they found that
the average farm size increased and the number of small farms
decreased in Europe and North America, while average farm
size decreased and the number of farms increased in Asia from
1950 to 1990 and in Africa from 1970 to 1990. They found no
clear long-run trend for South America. Hazell et al. (2010)
report on median farm size for a selection of 16 low- and
middle-income countries in Africa, Asia, and Latin America
and find that, for most countries considered, median farm size
decreased from 1970 to 2000.
A 2013 report by FAO shows that at the global level (106
country sample, with the number of countries varying from
year to year) the pooled average of mean farm size and the
median of mean farm sizes decreased from 1930 to 2000. This
reflects a decrease in average farm size for most regions except
Europe. Another report shows that average farm size has
decreased for the African region as a whole as well as in China
(HLPE, 2013).
Masters et al. (2013) argue that there is more of a distinction
to be made among regions, based on data and projections on
population in urban and rural areas from 1950 through 2050.
They show farm sizes decreasing in Africa and Asia from 1950
to 2010. From 2010 onward, they project that average farm
size will continue to decrease for Africa as a whole and to
increase for Asia as a whole. Adamopoulos and Restuccia
(2014) use a sample of data from agricultural censuses for 63
countries to show that average farm size is larger in countries
with higher average per capita income.
Jayne, Chamberlin, and Headey (2014) note that though
Africa is typically characterized as land abundant (e.g.,
Deininger, Byerlee, Lindsay, Norton, Selod, & Stickler,
2011), the majority of the region’s uncultivated arable land
is concentrated in a few countries. Using data on average farm
size over time from agricultural censuses and surveys for 12
African countries, the authors also show that in all of the
land-constrained African countries, average farm size
decreased from the 1980s to 2000s and that in some, but not
all of the land-abundant African countries, average farm size
increased from the 1980s to 2000s.
Several articles have considered farmland distribution and
concentration (Table 1). Deininger and Byerlee (2012) con-
ducted a review of literature on farmland concentration over
the period 1970–2000 for selected land-abundant countries in
Latin America, Eastern Europe and Central Asia, Southeast
Asia, and Sub-Saharan Africa. They conclude that in land-
abundant countries of Latin America and in Eastern Europe
and Central Asia there has been an increase in large-scale
farming. They also find that in South East Asia, the palm
oil industry has seen an increase in the number of larger plan-
tations or large firms contracting with out-growers, while rub-
ber production has shifted from being cultivated mainly by
large plantations to production by smallholders. In countries
of Sub-Saharan Africa, on the other hand, efforts to move
toward large-scale agriculture in the 1970s and ‘80s largely
failed and small farming operations have persisted, but with
an increase in large-scale investments in more recent years.
For a sample of 81 countries, using data from the 2000
round of the agricultural censuses, a report by the HLPE
found that 73% of farms are smaller than 1 ha and 85% are
smaller than 2 ha. It also considered farmland distribution
in 14 African countries, showing that 80% of holdings are
smaller than 2 ha in size and operate about 25% of the agricul-
tural land. The report also found that in the European Union
(EU) 50% of farms are smaller than 2 ha in size but operate
only about 2.4% of the agricultural land (HLPE, 2013).
Finally, Adamopoulos and Restuccia (2014) look at farm-
land distribution by country income level. They find that in
the richest countries farms larger than 20 ha operate 70% of
land, while in the poorest countries 70% of land is operated
by farms smaller than 5 ha.
In summary, the literature on farm size generally finds that
average farm size has decreased in the developing and
increased in the developed world and that farmland distribu-
tion is more unequal in high-income countries than in develop-
ing regions. However, the literature suffers from some
limitations. None of it is replicable, because none of the
sources provide full documentation of the methodology as
well as the underlying data. Furthermore, since some of these
sources were published, more recent information has been
made available for some countries. Finally, many of the
sources are not comprehensive in their country coverage. This
article seeks to address these shortcomings, by providing well-
documented, up-to-date, and comprehensive estimates of the
number of farms, farm size, and farmland distribution world-
wide. This is largely possible due to the country-level estimates
made available in international comparison tables in a recent
FAO Publication (2013).
3. DATA USED
This article relies on data from numerous agricultural cen-
suses, which are representative of all farms or farm households
in a country. The Food and Agriculture Organization of the
United Nations (FAO) has promoted the Programme for the
World Census of Agriculture (WCA) since 1950 by providing
governments with guidance on standard methodology and
contents for their agricultural census. The WCA was first con-
ducted in 1930; both the 1930 and 1940 rounds were sponsored
by the International Institute of Agriculture (IIA). For the first
six rounds (from 1930 to 1980), countries conducted the census
in the same year, but for the 1990 census this constraint was
relaxed and countries conducted the census during the period
1987–93 (FAO, 2005). The 2000 and 2010 rounds span the
periods 1996–2005 and 2006–15, respectively.
FAO recommends that the census considers farms of all
types throughout a country. Whereas agricultural censuses
are often nationally representative of all farms, some other
sources of information often used by the agricultural and
development economics profession are not. Household income
surveys, such as the Living Standard Measurement Study
THE NUMBER, SIZE, AND DISTRIBUTION OF FARMS, SMALLHOLDER FARMS, AND FAMILY FARMS WORLDWIDE
Q
3
Please cite this article in press as: Lowder, S. K. et al. The Number, Size, and Distribution of Farms, Smallholder Farms, and Family
Farms Worldwide
q
,World Development (2016), http://dx.doi.org/10.1016/j.worlddev.2015.10.041

(LSMS) surveys (World Bank, 2015), are often used for study-
ing agricultural activities in developing countries. The LSMS
and some other household income and expenditure survey
data are made widely available and provide a rich source of
information at the household level on sources of income and
expenditures as well as agriculture for many countries. How-
ever, a limitation of household surveys is that they are repre-
sentative of farm households but not representative of all
the farms in a country. Household surveys generally do not
include farms that are not family-owned (which are for the
most part large farms) and thus underestimate the contribu-
tion of large farms. The possible implications are illustrated
by the example of Guatemala, shown in Table 2. The agricul-
tural census data reveal that, in Guatemala, a small number of
very large farms (from 45.2 to more than 9,000 ha) represent
the minority (2%) of holdings, but the majority (57%) of farm-
land. Non-household farms and thus, most of the country’s
large farms, are not reported by the household survey data
indeed, the largest farm reported by the household survey
encompasses 98.8 ha. Clearly the agricultural census is key
to our gaining a more comprehensive picture of the agricul-
tural sector in Guatemala.
The FAO suggests that the census be conducted by using
complete enumeration and/or sampling methods. It encour-
ages countries to include a core module on number of holdings
and household size and status as well as 12 supplementary
themes. The supplementary themes are: agricultural land; irri-
gation, and water management; crops; livestock; agricultural
practices; agricultural services; demographic, and social char-
acteristics; farm labor; household food security; aquaculture;
forestry, and management of the holding.
Most of the data used in this article were collected under the
agricultural censuses’ core module (number of holdings,
household size, and status), as well as the agricultural land
and farm labor themes. The FAO’s theoretical definition of
an agricultural holding is:
‘‘an economic unit of agricultural production under single management
comprising all livestock kept and all land used wholly or partly for
agricultural production purposes, without regard to title, legal form,
or size. Single management may be exercised by an individual or
household, jointly by two or more individuals or households, by a clan
or tribe, or by a juridical person such as a corporation, cooperative or
government agency’’
[FAO, 2005]
The agricultural holder is the person who makes strategic
decisions regarding use of the farm resources and who bears
all risks associated with the farm. The agricultural holder
may undertake all management responsibilities or delegate
day-to-day work management responsibilities to a hired
manager. The difference between the hired manager and the
agricultural holder (the manager of the holding) is that the for-
mer is a hired employee who implements the decisions of the
agricultural holder while the latter makes all strategic deci-
sions, takes all economic risks, and has control over all pro-
duction resulting from the agricultural holding (FAO, 2005).
Agricultural holdings and agricultural area reported by the
census include crop and livestock production only; holdings
engaged in forestry or fisheries are only included if they are
also engaged in crop and livestock production. Communal
lands are generally not included in the agricultural census.
As with any data, a number of problems arise when making
international comparisons and global estimates using the agri-
cultural census data. First, information is not available for all
countries and all decades for each indicator. Second, although
the WCA provides a large degree of standardization among
countries, there is naturally variation among countries in terms
of how indicators are defined or reported. Third, there is varia-
tion among agricultural holdings within countries; that is, in
most countries a wide range of diverse entities are considered
agricultural holdings, making aggregation of them problematic.
In order to address the first problem and ensure the broadest
country coverage for each indicator, we combine estimates
from numerous census rounds. We use the most recent obser-
vation found for each indicator-country combination. Thus,
by considering the 1960 round through the 2010 round, we
have an estimate of the total number of farms which includes
observations for 167 countries and territories. For the number
of farms by size, we reach a sample of 111 countries and ter-
ritories by combining international comparison tables from
the 1990 and 2000 census rounds. Country coverage is the least
comprehensive for the number of family farms, for which we
have a sample of 52 countries taken from an international
comparison table from the 2000 round agricultural census.
Our approach of combining data from several rounds of the
census is by no means without fault, since, as the available
data show, the farm sector is dynamic and the number of
farms in a country changes over time.
Second, some definitional and methodological variation exists
across countries despite the high degree of standardization
achieved through the WCA. Despite the recommendation that
the agricultural census cover farms of all types, some agricultural
censuses survey household farms rather than all farms; see for
example, in the 2010 round, Ethiopia, Malawi, and Rwanda
(Government of Ethiopia, 2007; Government of Malawi, 2010;
Government of Rwanda, 2010). Furthermore, most countries
establish a minimum threshold farm size only above which farms
Table 2. Farmland distribution in Guatemala, agricultural census versus household survey
Farm size class (ha) Agricultural census, 2003 Household survey, 2006
Holdings Area Household farms Operated area
(percentage) (percentage)
<0.7 45 3 50 13
0.7–1.4 22 5 24 19
1.4–3.5 19 8 20 33
3.5–7.1 6 6 5 18
7.1–22.6 5 13 1 13
22.6–45.2 1 9 0 3
>45.2 2 57 0 0
Notes: For the household survey operated area equals land owned and used for crop production plus land share cropped or rented in minus land share
cropped or rented out.
Sources: Government of Guatemala (2004) for the agricultural census and Government of Guatemala (2006) for the household survey.
4 WORLD DEVELOPMENT
Please cite this article in press as: Lowder, S. K. et al. The Number, Size, and Distribution of Farms, Smallholder Farms, and Family
Farms Worldwide
q
,World Development (2016), http://dx.doi.org/10.1016/j.worlddev.2015.10.041

are included in the census; this minimum threshold varies from
one country to the next. In China and India, very small farms
are included in the census; China reported farms as small as
0.07 ha whereas India did not apply a minimum threshold
(Government of India, 2012; Government of the People’s
Republic of China, 2009). Had those two countries used larger
minimum thresholds, similar to those used by other countries
(Bangladesh, for example, excluded farms smaller than 0.2 ha
(Government of Bangladesh, 2010), fewer total holdings would
have been recorded.
There is also variation across countries in what land size
classes they use in reporting their agricultural census results
and in the smallest land size class reported by FAO in its inter-
national comparison tables. The smallest land size class
reported by many countries where average farm sizes are large
is upto 2 ha; in many cases, farms smaller than one hectare are
included in that group. For example, in France there are
111,740 farms smaller than 2 ha in size operating a total of
212,000 ha of agricultural land. Such farms are reported by
FAO’s international comparison tables in the cohort 1–2 ha,
although some of them are in fact smaller than 1 ha (FAO,
2013). A similar limitation applies to the farms in the larger
farm size classes, particularly for countries where average farm
sizes are relatively small. In the Philippines, for example, the
largest farm size class reported is 20–50 ha, where there are
a total of 20,666 farms operating 1,072,845 ha of land. The
average farm size for this group is thus larger than 50 ha,
and we can therefore infer that some of the Filipino farms
reported as being between 20 and 50 ha in size actually belong
to a larger farm size class (FAO, 2013). For some other coun-
tries, similar issues arise; for details refer to FAO (2001, 2013).
Our third challenge is the diversity of agricultural holdings
within countries. Perhaps the most extreme example is that
of Russia. For its 2006 agricultural census, the Russian Feder-
ation reports several different types of agricultural holdings
(Government of the Russian Federation, 2006). The vast
majority of holdings (about 98%) are owned by private house-
holds and other individual operators; they represent only about
2% of the farmland in the country. Farms in the other cate-
gories of agricultural holdings (small and large agricultural
enterprises, institutional farms, private farms, individual entre-
preneurs, and non-profit citizen associations) represent only
about 2% of the holdings in Russia but about 98% of the farm-
land. The average farm size for the farms in the private-
household and other individual-operator category is about
0.4 ha whereas it is far larger for the other categories. The Rus-
sian census recognizes this heterogeneity, and, because the var-
ious entities are so different, it does not provide a total number
of agricultural units for the country; rather, it reports separate
subtotals for the different types of agricultural holdings. For
our total number of farms, we have chosen to consider and
aggregate all agricultural units enumerated by each country,
including all the different categories in the Russian census.
For all these reasons, it is important to stress that, when we
report a total number of farms or family farms, we are report-
ing a very rough estimate of the total number of very diverse
entities; they are diverse in terms of scale, livelihood strategy,
and socio-economic status. Throughout the world, farms and
family farms include large-scale industrial ventures as well as
small and medium-sized operations. The holders may be spe-
cialized in agriculture, with farm income and production pro-
viding a substantial share of their income and/or food
consumption. Alternatively, the holder may have diverse
sources of income or may be specialized in off-farm income
sources. The holder may or may not have adequate income
to provide for his family. Given this diversity, we therefore
strongly caution the reader that when we refer to a total num-
ber of farms in the world we are describing entities which vary
greatly from one country to another as well as within each
country. Despite its limitations, we have nevertheless chosen
to use the agricultural holding as reported by the census as a
proxy for the total number of farms.
4. THE NUMBER OF FARMS IN THE WORLD
We produce an estimate of the total number of farms in the
world for the widest possible country coverage by using infor-
mation from six different WCA rounds dating back to 1960.
We recorded the most recent estimate of the number of agricul-
tural holdings for each country or territory for which an agricul-
tural census has been carried out and for which a report was
available. Because we are interested in the number of farms
worldwide, we consider estimates of farms in all geographical
entities, whether recognized as sovereign states or not. For ease
of communication, from this point forward the article refers to
number of ‘‘countries”used in various calculations without
drawing the distinction among actual geographic areas that
are widely recognized as sovereign states and those that might
be considered as territories of sovereign states.
Estimates are available for 167 countries which represent
96% of the world’s population, 97% of the population active
in agriculture, and 90% of agricultural land worldwide. The
resultant estimate of the total number of farms in the 167
countries comes to nearly 570 million. For country-level esti-
mates and sources, see Web Appendix Table 1.
For numerous reasons, we may expect that the actual number
of farms in the world is larger than this 570 million estimate.
First, no estimates were available for about 40 smaller coun-
tries; including estimates from those countries would increase
the total but probably only slightly. Second, although the vast
majority of agricultural holdings were reported from fairly
recent agricultural census rounds (1990, 2000 or 2010), data
for some low- and middle-income countries are from older agri-
cultural census rounds. For example, the last census for three of
the countries (Brunei Darussalam, Nigeria, and Zimbabwe) was
conducted in 1960. The number of farms in many low- and
middle-income countries has increased since 1960 (see Section 5
on average farm size), and it is likely to have done so in many of
the countries for which we used data from older census rounds.
For these reasons, our estimate of 570 million farms worldwide
probably represents a lower-bound estimate.
Of the estimated 570 million farms in the world, 74% are
located in Asia (Figure 1). China alone represents 35% and
India 24% of the 570 million farms. Nine percent of the farms
are found in Sub-Saharan Africa and 7% in Europe and Cen-
tral Asia. Farms in Latin America and the Caribbean repre-
sent 4% of farm holdings worldwide. Only 3% of the world’s
farms are located in the Middle East and North Africa.
The majority of farms are found in lower- or upper-middle-
income countries (representing respectively 36% and 47% of
the 570 million farms worldwide), this largely results from
the classification of India as belonging to the former and
China the latter group (Figure 1). Thirteen percent of farms
are in low-income countries. Farms in high-income countries
represent 4% of the world’s farms.
5. AVERAGE FARM SIZE
Average farm size is generally thought to have increased in
the developed world, while it has decreased in the developing
THE NUMBER, SIZE, AND DISTRIBUTION OF FARMS, SMALLHOLDER FARMS, AND FAMILY FARMS WORLDWIDE
Q
5
Please cite this article in press as: Lowder, S. K. et al. The Number, Size, and Distribution of Farms, Smallholder Farms, and Family
Farms Worldwide
q
,World Development (2016), http://dx.doi.org/10.1016/j.worlddev.2015.10.041

world; our findings suggest the story is not so clear cut and
that data are inadequate to have a globally representative pic-
ture of the evolution of average farm size over time. All
country-level data used for estimates in this section are avail-
able in Web Appendix Table 2.
International comparison tables from the 2000 census round
allows us to consider average farm sizes for a sample of 107
countries for which estimates of average farm size are avail-
able for at least two of the census periods from 1960 to
2000. We first consider the trends at the country level, record-
ing an increase (decrease) for any country where average farm
size has consistently increased (decreased) from one period to
the next allowing for only one exception over the period. The
level of significance of the increase or decrease is not consid-
ered. We see clear patterns according to income group
(Table 3). In most low- and lower-middle-income countries,
farm sizes have decreased. A slightly larger share of upper-
middle-income countries has exhibited increasing average
farm size, and the majority of high-income countries has seen
a clear increase. Patterns according to regional groupings of
low- and middle-income countries show that an increase in
average farm size has occurred in some countries in Latin
America and the Caribbean, but in few countries elsewhere.
For another indication of what trends are evident in high-
income countries as opposed to low- and middle-income coun-
tries, we examine the evolution of average farm size from the
1960 census to the 2000 census round for a limited sample of
72 countries. We consider only those countries for which we
had an average farm size from at least three census rounds.
We imputed average farm size (using linear interpolation) in
all cases where the missing value was preceded and followed
by an observed average farm size. When the missing observa-
tion(s) was (were) at the beginning or end of the series, we used
the adjacent value(s). To calculate the weighted average farm
size by income group, we used a fixed weight over time, which
is the number of agricultural holdings reported in the coun-
try’s most recent agricultural census.
Our sample is not globally representative, but it does allow
us to consider representative trends for high-income countries,
Latin America and the Caribbean, and South Asia. It includes
30 high-income countries which represent 96% of holdings and
98% of agricultural area in the category of high-income coun-
tries other than Australia. The 18 countries from Latin Amer-
ica and the Caribbean included in the sample represent 79%
and 87% of that region’s agricultural holdings and agricultural
area. The five countries from South Asia represent 98% of the
region’s holdings and 86% of its agricultural area. Of the
remaining 19 low- and middle-income countries in our sample,
five are from East Asia and the Pacific, five from the Middle
East and North Africa, and nine from Sub-Saharan Africa.
A notable exclusion from the sample is that of China. Also,
we have no information on average farm size over time for
countries in Europe and Central Asia, due in part to changes
in boundaries in that region.
We find that average farm size has increased for the 30 high-
income countries as a whole. For Latin America and the Car-
ibbean as a whole average farm size decreased from about
80 ha per farm in 1960 to 50 ha in 1990, after which point it
increased to 54 ha in 2000. The trend in South Asia is a clear
decrease in average farm size from about 2.6 ha per farm in
9% East
Asia and
the
Pacific,
excluding
China (14)
35% China
7%
Europe
and
Central
Asia (14)
4% Lan
America
and the
Caribbean
(26)
3%
Middle
East and
North
Africa (12)
6% South
Asia,
excluding
India (6)
24% India
9% Sub-
Saharan
Africa (41)
4% High-
income
countries
(46)
570 million farms in 161 countries*, by
region or country group
13%
Low-
income
countries
(30)
36%
Lower-
middle-
income
countries
(38)
47%
Upper-
middle-
income
countries
(47)
4%
High-
income
countries
(46)
570 million farms in 161 countries*,
by income group
Figure 1. Share of farms worldwide, by country group, most recent observation. Sources: Country-level data are available in Web Appendix Table 1. Notes:
Number of countries included are shown in parentheses. Country regional and income groupings are the same as those used by the World Bank (2012).
*
Only
161 of the 167 countries with observations are classified by the World Bank groupings.
6 WORLD DEVELOPMENT
Please cite this article in press as: Lowder, S. K. et al. The Number, Size, and Distribution of Farms, Smallholder Farms, and Family
Farms Worldwide
q
,World Development (2016), http://dx.doi.org/10.1016/j.worlddev.2015.10.041

1960 to 1.3 ha in 2000. For the remaining 19 low- and middle-
income countries as a whole, average farm sizes decreased over
the period. We do not have a sufficient sample to have a rep-
resentative picture of what has happened to average farm sizes
for Sub-Saharan Africa, China, the rest of East Asia and the
Pacific, the Middle East and North Africa, or Europe and
Central Asia.
The average farm size worldwide is determined by countries
with the largest number of farms and those with the largest
share of the world’s agricultural land. Among countries for
which we have an estimate of the number of farms, the five
countries with the largest share of the world’s agricultural
holdings are, in decreasing order of magnitude: China, India,
Indonesia, Russian Federation, and Bangladesh. In 2010,
those with the largest share of the world’s agricultural area
(measured as arable land plus permanent crops and permanent
pastures) were, in decreasing order of magnitude: China, Aus-
tralia, the United States of America, Brazil, and the Russian
0.0
0.5
1.0
1.5
2.0
2.5
3.0
0
10
20
30
40
50
60
70
80
90
1960 1970 1980 1990 2000
High-income-countries excluding Australia (30), primary axis
Lan America and the Caribbean (18), primary axis
South Asia (5), secondary axis
Other low- & middle-income countries (19), secondary axis
Hectares Hectares
Figure 2. Average farm size, 1960–2000. Sources: Authors’ calculations using (FAO, 2013) for average farm size together with the most recent observation
for the number of farms. See Web Appendix Table 2. Notes: Total country coverage is indicated in parentheses.
Table 3. Number of countries exhibiting a decrease or increase in the average size of agricultural holdings, 1960–2000
Decrease Increase Neither clear increase nor decrease
High-income countries 7 26 4
Low- and middle-income countries, by income group
Low-income countries 12 2 1
Lower-middle-income countries 24 2 0
Upper-middle-income countries 19 5 1
Low- and middle-income countries, by regional grouping
East Asia and the Pacific 8 2 0
Latin America and the Caribbean 18 7 2
Middle East and North Africa 10 0 0
South Asia 5 0 0
Sub-Saharan Africa 15 3 1
Sources: Authors’ compilation using FAO (2013). See also Web Appendix Table 2.
Table 4. Top five countries by number of agricultural holdings and agricultural area
Agricultural holdings (in millions) Agricultural area (in millions ha)
Top five countries Most recent
estimate
Share of the world’s
agricultural holdings (%)
Top five countries 2010 Share of the world’s
agricultural area (2010) (%)
China 201 35 China 525 11
India 138 24 Australia 456 9
Indonesia 25 4 United States of America 414 8
Russian Federation 23 4 Brazil 261 5
Bangladesh 15 3 Russian Federation 217 4
World total 570 World total 4,889
Source: Authors’ compilation using FAO, 2014a.
THE NUMBER, SIZE, AND DISTRIBUTION OF FARMS, SMALLHOLDER FARMS, AND FAMILY FARMS WORLDWIDE
Q
7
Please cite this article in press as: Lowder, S. K. et al. The Number, Size, and Distribution of Farms, Smallholder Farms, and Family
Farms Worldwide
q
,World Development (2016), http://dx.doi.org/10.1016/j.worlddev.2015.10.041

Federation (Table 4). Of these countries, in our Figure 2 data
were missing or not included for Australia, China, and the
Russian Federation.
We deliberately excluded Australia from the 72-country
sample in Figure 2 because its farms are so large (averaging
in the thousands of hectares) compared to farms in other parts
of the world that it appears to be an outlier. Considering Aus-
tralia on its own (Figure 3), we see that average farm size
clearly increased from 1960 to 2000, although it declined dur-
ing 1990–2000; this lends greater evidence to the tendency for
farm size to have increased in high-income countries as
depicted in Figure 2.
Given the sheer number of agricultural holdings as well as
the amount of agricultural land in China, what happens in
terms of average farm size in that country will be a major
determinant of the trend in average farm size observed for
the world as well as for low- and middle-income countries as
a whole. Information from the 2000 and 2010 rounds of agri-
cultural censuses indicate a decrease in average farm size in
China in the most recent period (Figure 3). However, some
experts argue that farm consolidation is likely to have begun
or to begin, given current changes in land policy, which facil-
itate private ownership (as opposed to merely use) of agricul-
tural land by individuals (Nie & Fang, 2013). Should
individuals be able to own (rather than simply use) their land,
sales of land will be facilitated and consolidation should
become much easier.
The Russian Federation has likely exhibited a decrease in
average farm size since liberalization and land reform (Sedik
& Lerman, 2008), this lends further evidence to the tendency
of farm size to have decreased in low- and middle-income
countries from 1960 to 2000. Regardless, the distribution of
agricultural land in Russia remains bimodal, with the vast
majority of farmers operating small areas, while the majority
of land farmed is concentrated among a few large-scale corpo-
rate farms (Lerman & Sedik, 2013). This would suggest that
the measure of average farm size at the national level is less rel-
evant for the Russian Federation than it may be for some
other countries. To have an idea of the scale of most farms
in Russia the median farm size would be a more informative
statistic than average farm size. However, few agricultural cen-
suses offer estimates of median farm size.
6. FARMS AND FARMLAND DISTRIBUTION
The High-Level Panel of Experts of the Committee on Food
Security (HLPE) surveyed definitions of smallholders, present-
ing a wide array of criteria used by countries in their national
definitions (HLPE, 2013). The most commonly used dimen-
sion for measuring farm size is land, but some countries also
consider the number of livestock held or the assets used, while
others use gross sales as a criterion. The report recognizes that
small is a relative concept, depending on agro-ecological as
well as socio-economic considerations. What is small in most
countries in Latin America and the Caribbean is considered
large in most countries in Sub-Saharan Africa or in Asia.
Although simplistic, a 1 or 2 ha threshold is frequently used
to designate farms as small. For a sample of 81 countries,
using data from the 2000 round, the HLPE report found that
73% of farms are smaller than 1 ha and 85% are smaller than
2 ha (2013). Here we consider a more comprehensive sample,
with nearly identical results.
Combining data from the international comparison tables
for the 1990 and 2000 rounds of the WCA (FAO, 2001,
2013), we are able to consider farms by size for a sample of
111 countries and territories with a total of nearly 460 million
farms (Figure 4 and Web Appendix Table 4). Seventy-two per-
cent of the farms are smaller than one hectare in size; 12% are
1–2 ha in size; and 10% are between 2 and 5 ha. Only 6% of
the world’s farms are larger than 5 ha. Assuming that these
0
1,000
2,000
3,000
4,000
1960 1970 1980 1990 2000
Australia
Hectares
0.0
0.1
0.2
0.3
0.4
0.5
0.6
0.7
2000 2010
China
Hectares
Figure 3. Average farm size over time in Australia and China. Sources: Authors’ compilation using FAO (2013) and Government of the People’s Republic of
China (2009).
72%
less than
1 Ha
12%
1 - 2 Ha
10%
2 - 5 Ha
3%
5 - 10 Ha
1%
10 - 20
Ha 2%
Over 20
Ha
460 million farms in 111 countries
Figure 4. Share of farms worldwide, by land size class. Sources: Authors’
compilation using data from FAO (2001, 2013). See Web Appendix Table 3.
Notes: The figures are based on the most recent available estimate for each
country from the World Agricultural Census 1990 and 2000 rounds.
8 WORLD DEVELOPMENT
Please cite this article in press as: Lowder, S. K. et al. The Number, Size, and Distribution of Farms, Smallholder Farms, and Family
Farms Worldwide
q
,World Development (2016), http://dx.doi.org/10.1016/j.worlddev.2015.10.041

averages are representative of farm sizes worldwide, we esti-
mate that worldwide there are more than 410 million farms
less than 1 ha in size and more than 475 million farms that
are less than 2 ha in size (this is the result of more than 570
million farms multiplied by 72% and 84%, respectively). This
would support claims that worldwide there are about 500 mil-
lion small farms (those with less than 2 ha) (see, for example,
Hazell et al., 2010; HLPE, 2013; IFAD, 2011; Wiggins et al.,
2010).
Estimates of the number of farms by farm size class are
fairly widespread, since many countries provide that informa-
tion in their agricultural census. It is more difficult to estimate
how farmland is distributed among farms of different sizes,
since fewer countries report information on agricultural area
by land size class. To date, no comprehensive estimates of
farmland distribution have been made for the world as a
whole, but some work does give an indication of farmland dis-
tribution for selected countries or country groups (see Sec-
tion 2).
Here we present the most comprehensive estimate possible
of the distribution of farms and farmland by land size class;
the data underlying the estimates in this section are presented
in Web Appendix Table 3. We have data on both the number
of farms and the agricultural area by land size class for 92
countries and territories. This sample of 240 million farms rep-
resents 42% of the world’s farms, as well as about 56% of the
world’s population, 43% of the population active in agricul-
ture, and 38% of agricultural land worldwide (FAO, 2014a).
We cannot claim that this sample is representative of the
world’s farms. It does not include China, thus omitting about
35% of the world’s farms; the Russian Federation and Aus-
tralia are likewise not included. It under-samples countries
in Sub-Saharan Africa and the low-income category, with only
nine and eight countries in each group, respectively.
We first consider the pattern of agricultural land distribu-
tion observed based on a country’s income level. For nearly
all income levels, on average, a large share of farms (between
40% and 80%) are smaller than 2 ha; they control anywhere
from a few percent to about 40% of farmland (Figure 5).
The farmland share represented by the larger cohorts would
seem to increase with each income category. For example,
farms greater than 5 ha in size cover 27% of the farmland in
low-income countries, 43% in the lower-middle-income coun-
tries, 96% in the upper-middle-income countries, and 97% in
the high-income countries. In short, it would appear that the
share of farmland controlled by larger farms is higher in coun-
tries with larger average incomes.
We now examine regional patterns among the 58 low- and
middle-income countries for which agricultural censuses
report information on both number of farms and farmland
by land size class (Figure 6). In all regions except Latin Amer-
ica and the Caribbean, the majority of farms are in the small-
est cohort (smaller than 1 ha). The distribution of farms and
farmland according to farm size seen for the nine countries
located in Sub-Saharan Africa is similar to that of the three
South Asian countries, where about 80% of farms are smaller
than 2 ha and operate close to 40% of the farmland; in those
regions, few farms reach a size larger than 50 ha and the few
that do comprise only a small share of total farmland. The dis-
tribution observed for the nine countries in East Asia and the
Pacific (excluding China) is also quite similar, with about 75%
of farms being smaller than 2 ha and representing almost 30%
of farmland, and farms rarely reaching sizes exceeding 100 ha.
In Latin America and the Caribbean, the pattern is radically
different. There are relatively fewer very small farms and the
bulk of the land is operated by a small number of very large
farms. Farms in Latin America and the Caribbean are gener-
ally larger and a few very large farms exceeding 1000 ha in size
Figure 5. Distribution of farms and farmland area by land size classes and income group. Sources: Authors’ compilation using FAO (2001, 2013). See Web
Appendix Table 3. Notes: Country groupings are the same as those used by the World Bank (2012). Number of countries shown in parentheses.
THE NUMBER, SIZE, AND DISTRIBUTION OF FARMS, SMALLHOLDER FARMS, AND FAMILY FARMS WORLDWIDE
Q
9
Please cite this article in press as: Lowder, S. K. et al. The Number, Size, and Distribution of Farms, Smallholder Farms, and Family
Farms Worldwide
q
,World Development (2016), http://dx.doi.org/10.1016/j.worlddev.2015.10.041

represent nearly half of the total farmland. In the Middle East
and North Africa, 70% of all farms are below 2 ha, but more
than 50% of the land is farmed by holdings above 10 ha in
size.
We can improve the representativeness of the sample by
including an additional 14 countries for which we have infor-
mation on the number of farms by farm size class as well as the
total farmland operated nationally, but not by land size class.
These countries include China as well as Canada, Guinea Bis-
sau, Hungary, Lesotho, Libya, Malawi, Mozambique, New
Zealand, Slovakia, Slovenia, Spain, St. Kitts & Nevis, and
Viet Nam. We estimate the farmland distribution in each of
these countries by assuming that the average farmland size
in each cohort in the 14 countries is the midpoint of the
cohort. That is, in each of the countries the farms in the 0–
1 ha cohort average 0.5 ha; those in the 1–2 ha cohort average
1.5 ha, etc. The hectares of farmland in the largest farm size
class reported by each country are calculated as the difference
between the total number of hectares of agricultural land in
the country and the sum of the estimated hectares of land in
the smaller land size classes (using the midpoint assumption).
Adding the 14 countries, including China, to the 92 country
sample, we have a 106 country sample covering a total of 450
million farms. These countries represent 80% of the world’s
Figure 6. Average distribution of farms and farmland area by land size classes and region. Sources: Authors’ compilation using FAO (2001, 2013). See Web
Appendix Table 3. Notes: Country groupings are the same as those used by the World Bank (2012) with the following additions: Cook Islands were classified
as East Asia and the Pacific; French Guiana, Guadeloupe, and Martinique as Latin America and the Caribbean; and Re
´union as sub-Saharan Africa. Number
of countries shown in parentheses.
10 WORLD DEVELOPMENT
Please cite this article in press as: Lowder, S. K. et al. The Number, Size, and Distribution of Farms, Smallholder Farms, and Family
Farms Worldwide
q
,World Development (2016), http://dx.doi.org/10.1016/j.worlddev.2015.10.041

farms, about 85% of the world’s population active in agricul-
ture, 80% of the world’s population, and 60% of agricultural
land worldwide (FAO, 2014a). We find that 84% of the farms
are smaller than 2 ha and operate about 12% of the farmland
(Figure 7). Otherwise stated, only 16% of the world’s farms are
larger than 2 ha, but they represent 88% of the world’s farm-
land.
For two reasons this is a fairly conservative estimate of the
degree to which farmland distribution is unequal at the global
level. First, by using the midpoint assumption for the 14 coun-
tries that do not report information on agricultural land oper-
ated by land size class, our estimate probably overstates the
share of farmland held by small farms worldwide. Secondly,
as data were not available for Australia or the Russian Feder-
ation, our estimate does not include either of those two coun-
tries. Both Australia and the Russian Federation have vast
amounts of farmland, the majority of which is found on large
farms; were either of those two countries included in our esti-
mates they would show the distribution of agricultural land
worldwide as being even more unequal than it appears in Fig-
ure 7 (FAO, 2013; Lerman & Sedik, 2013).
In conclusion, in countries at lower levels of income smaller
farms operate a far greater share of farmland than do smaller
farms in the higher income countries. This is perhaps best
illustrated by considering the share of farms and farmland rep-
resented by farms that are smaller than 2 ha. Globally, about
84% of farms are smaller than 2 ha, and they operate about
12% of farmland. In low- and lower-middle-income countries,
as well as in countries of East Asia and the Pacific (excluding
China), South Asia and Sub-Saharan Africa, about 70–80% of
farms are smaller than 2 ha and operate about 30–40% of land
in those countries. In upper-middle-income countries (exclud-
ing China), high-income countries, Latin America and the
Caribbean, Middle East and North Africa a large share of
farms are likewise smaller than 2 ha in size, but they operate
less than 10% of farmland.
Claims are often made that small farms are responsible for
large shares of the world’s food production, often without a
clear definition of a small farm. What can meaningfully be
considered a small farm clearly depends on the country and
the context. However, as we have seen, globally farms smaller
than 2 ha operate only 12% of the world’s agricultural land.
We do not know the relative land productivity of farms smal-
ler than 2 ha relative to those larger than 2 ha, although a
large body of literature on land productivity by farm size
has shown a phenomenon referred to as the inverse land
productivity relationship, i.e. that in a number of countries
smaller farms have higher crop yields than do larger ones
(Barrett, Bellemare, & Hou, 2010; FAO, 2014b; Larson
et al., 2014). We likewise do not know what share of farmland
on holdings smaller than 2 ha is used for food production as
opposed to agricultural production or what share of land
included in our estimates of farmland area is land that lies fal-
low. Nevertheless, it seems implausible that farms smaller than
2 ha in size are able to produce the majority of the world’s
food using a mere 12% of the world’s farm land, and, in any
case, claims to the contrary are unsubstantiated. For some
regions and country groups, however, these farms manage a
larger share of the region’s land and are thus likely responsible
for a larger share of the region’s food or agricultural produc-
tion than at the global level. In East Asia and the Pacific,
excluding China; South Asia; Sub-Saharan Africa, and in
low- and lower-middle-income countries farms smaller than
2 ha operate 30–40% of agricultural land. This would suggest
that their share of regional or group-wide agricultural produc-
tion is sizeable, but it does not provide proof that they are
responsible for a majority of agricultural production in those
regions or country groups.
7. THE NUMBER AND DISTRIBUTION OF FAMILY
FARMS
Family farms have figured prominently in the discourse
about agricultural development, but definitions are often
unclear. The term family farm may even be used interchange-
ably with smallholder farm. In an attempt to lend some clarity,
we provide an estimate of the total number of family farms
worldwide as well as the share of agricultural land they oper-
ate, which has implications for their contribution to total food
and agricultural production.
There is no universally agreed definition of family farms,
although various stakeholders have established definitions
either for purely analytical purposes or for the implementation
of government programs. Despite wide variation among defi-
nitions of family farms, there are some commonalities. A sur-
vey of 36 definitions of family farms found that the most
common aspect of such definitions is the use of family labor
and that many of the definitions also specify that the farm is
managed by the family (Garner & de la O Campos, 2014).
Some definitions limit the size of the farm explicitly by estab-
lishing a maximum land area for the farm, beyond which the
farm is no longer considered a family farm. Finally, a defini-
tion may require that the share of household income from
non-farm activities not exceed a certain level.
Information from national agricultural censuses can provide
insights into the prevalence of some of the characteristics that
commonly enter into the various definitions of family farms.
International comparison tables are available describing such
characteristics for some countries included in the 2000 round
(see Web Appendix Table 4 and FAO, 2013). The sample sizes
are small since many countries chose not to report such infor-
mation in their census. We first consider whether the farm is
owned or operated by a family and next whether the labor is
supplied by the family or by hired workers.
Some censuses report on the legal status of the holder of the
farms (see Section 3for the definition of agricultural holder),
but the censuses generally do not report on ownership of the
farm. In most of the 52 countries for which we have informa-
tion (FAO, 2013), more than 90% of farms (and often close to
100%) are held by a single individual, a group of individuals or
a household, either with or without a formal contract; only a
0%
10%
20%
30%
40%
50%
60%
70%
80%
90%
less
than 2
2 - 5 5 - 10 10 - 20 20 - 50 50 -
100
100 -
200
200 -
500
500 -
1000
over
1000
Share of agricultural holdings Share of farmland area
Figure 7. Distribution of farms and farmland area by land size classes, 106
country sample. Sources: Authors’ compilation using FAO (2001, 2013).
See Web Appendix Table 3.
THE NUMBER, SIZE, AND DISTRIBUTION OF FARMS, SMALLHOLDER FARMS, AND FAMILY FARMS WORLDWIDE
Q
11
Please cite this article in press as: Lowder, S. K. et al. The Number, Size, and Distribution of Farms, Smallholder Farms, and Family
Farms Worldwide
q
,World Development (2016), http://dx.doi.org/10.1016/j.worlddev.2015.10.041

very small share of farms are held by a corporation, coopera-
tive, governmental institution, religious institution, or an
unknown arrangement.
Several definitions of family farms also require the family to
supply the majority of the labor used on the farm. Data on the
use of family and hired labor are quite limited in the censuses.
Only 55 countries report data on the number of permanent
hired workers; for nearly all of these, the average is very small,
far less than 1 hired worker per farm (FAO, 2013). Only 30
countries report data on both family and permanent hired
labor; for these countries the total number of family members
engaged in agriculture exceeded the total number of perma-
nent hired workers by a ratio of 20–1. In many contexts, sea-
sonal hired workers provide an important source of labor for
farms, but data on seasonal hired workers are available only
from very few agricultural censuses. Therefore, in the absence
of more information, we conclude that family labor exceeds
hired labor on the vast majority of farms (FAO, 2013).
We use this information, together with our estimate of the
total number of farms worldwide, to provide a rough estimate
of the total number of family farms worldwide. We assume
that at least 90% of the world’s more than 570 million farms
are held by an individual, small group of individuals, or house-
hold, as was the case for our 52 country sample. This leads us
to estimate that there are more than 500 million family farms
worldwide.
Due to data limitations, the estimate of more than 500 mil-
lion family farms worldwide should be considered an approx-
imation. Current agricultural censuses are not available for
many countries where the number of farms has probably
increased in recent years, so the total number of farms likely
exceeds the 570 million estimated here. Furthermore, in almost
all countries for which data are available, 90% represents a
conservative estimate of the share of family farms in the total.
On the other hand, agricultural censuses do not provide data
on seasonal workers, who are an important source of labor for
farms in many contexts. Accurate data on the use of seasonal
labor might lead to lower estimates of the share of family
farms in several countries, depending on the threshold used
for the share of non-family labor in the family farming defini-
tion.
Information on how farmland is distributed among family
and non-family farms is limited. However, for a subset of 28
countries for which data are available the unweighted average
share of total agricultural land operated by farms which we
have classified as family farms is 73%. Calculating a weighted
average (using the amount of agricultural land in each of the
28 countries as weight), we find that 77% of the land is oper-
ated by family farms.
The sample is small, but if it is representative of the world as
a whole, it would suggest that more than 90% of the world’s
farms are family farms and that they operate about 75% of
the world’s land. This would imply that family farms are likely
to be responsible for the majority of the world’s agricultural
and food production. However, family farms, as defined in
this article, are a diverse group which includes farms of all
sizes. In designing policies for agricultural development, it is
necessary to bear this diversity in mind and distinguish among
different types of family farms.
8. CONCLUSIONS AND IMPLICATIONS
In summary, the main findings of this article, which repre-
sent very broad estimates, are:
There are more than 570 million farms in the world; more
than 475 million farms are smaller than 2 ha, and more than
500 million are family farms.
From 1960 to 2000, average farm size decreased in most
low- and lower-middle-income countries and in South Asia
as a whole. It decreased from 1960 to 1990 in Latin America
and the Caribbean as a whole, before increasing from 1990
to 2000. Average farm sizes increased from 1960 to 2000 in
some upper-middle-income countries and in nearly all high-
income countries for which we have information. Data on
average farm size are lacking in particular for China as well
as many other countries in East Asia and the Pacific and for
countries in Sub-Saharan Africa. Expectations of land consol-
idation in China, together with a slight increase in average
farm sizes for a small sample of low- and middle-income coun-
tries from 1990 to 2000, raises the question: has average farm
size for the world as a whole already begun to increase or will
it increase in the near future? For many low- and lower-
middle-income countries, however, average farm sizes are
likely to continue to diminish for some time still.
Globally, about 84% of farms are smaller than 2 ha, and
they operate about 12% of farmland. In countries at lower
levels of income, smaller farms operate a far greater share of
farmland than do smaller farms in the higher-income coun-
tries. In low- and lower-middle-income countries, as well as
in countries of East Asia and the Pacific (excluding China),
South Asia, and Sub-Saharan Africa, about 70–80% of farms
are smaller than 2 ha and operate about 30–40% of the land.
Our findings illustrate that the terms family farm and small
farm must not be used interchangeably. According to the most
commonly used definitions, more than 90% of the world’s
farms can be considered family farms, while 84% of all farms
are small farms (less than 2 ha) (Figure 8). Indeed, while there
must necessarily be a considerable degree of overlap between
the two categories, they are not the same. While family farms
operate the majority of the world’s agricultural land (about
75%), small farms (below 2 ha) operate only about 12% of
the world’s land.
These estimates of farm distribution provide insights into the
share of the world’s food or agricultural production that is pro-
duced by either of the two groups. With family farms operating
75% of the world’s agricultural land, and with a consistent high
share across almost all countries, it follows that they are likely
responsible for the majority of the world’s food and agricultural
production. However, it is implausible that with only 12% of the
world’s land, small farms, defined as those operating less than
2 ha, are able to produce a large share of the world’s food,
0
10
20
30
40
50
60
70
80
90
100
Small farms (< 2 ha) Family farms
Share of holdings Share of agricultural area
Percentage
Figure 8. Share of holdings and agricultural area, by farm type. Sources:
Authors’ compilation using data in Web Appendix Tables 3 and 4.
12 WORLD DEVELOPMENT
Please cite this article in press as: Lowder, S. K. et al. The Number, Size, and Distribution of Farms, Smallholder Farms, and Family
Farms Worldwide
q
,World Development (2016), http://dx.doi.org/10.1016/j.worlddev.2015.10.041

and claims that they do remain unsubstantiated. For some
regions and country groups, these farms control a larger share
of the land and are thus likely responsible for a large share,
but not necessarily a majority, of food or agricultural produc-
tion in those regions or country groups.
These estimates attempt to provide a comprehensive
description of the number and distribution of farms and
farmland worldwide, but there are important caveats, both
in terms of the analysis and the data used. Future work
might improve upon these results by conducting a compre-
hensive statistical analysis of the data we have used, which
are made available in Web Appendix. In terms of data, agri-
cultural census reports have achieved a high degree of stan-
dardization as a result of the World Programme for the
Census of Agriculture. Nevertheless, as with any interna-
tional dataset, variations among surveys in terms of defini-
tions, methodologies, geographical coverage, and the
inclusion (or not) of non-household farms mean that the
estimates are often times not comparable across countries
or over time. Continuing to improve documentation of the
coverage and methodology used in agricultural census
reports is important to gain a better understanding of the
degree to which such variations are a problem.
Furthermore, agricultural census data are rarely made avail-
able to researchers at the farm level. As a result, the agricul-
tural and development economics profession relies mostly on
household surveys, which, despite being a rich source of
detailed information, do not include larger commercial or gov-
ernment owned, non-household farms. A comprehensive
understanding of the agricultural landscape requires access
to data that cover all entities in the sector and not only farm
households. It would seem appropriate to redouble efforts to
make farm-level agricultural census data more easily accessible
to researchers as well as to encourage all countries to conduct
comprehensive agricultural censuses which either sample or
enumerate all types of farms throughout a country.
Were all agricultural censuses to include information on
farmland distribution it would help us improve our under-
standing of farmland distribution worldwide. Also useful for
analytical purposes would be the systematic inclusion of data
on use of hired labor—both permanent and seasonal.
Continued efforts to enhance the collection of up-to date,
comprehensive and more standardized agricultural census
data are essential to having a more representative picture of
the number of farms, small farms, and family farms as well
as changes in farm size and farmland distribution worldwide.
NOTE
1. Small is a relative concept and highly context specific, with national
definitions of small farms varying widely from country to country and
wide variation in approaches to measuring the number of smallholders.
The most widely used approach for global or regional estimates is that of
using the size of an agricultural holding. An overview of this issue is
beyond the scope of this paper. For more, see, for example, FAO, 2015
and HLPE, 2013.
REFERENCES
Adamopoulos, T., & Restuccia, D. (2014). The size distribution of farms
and international productivity differences. The American Economic
Review, 104(6), 1667–1697.
Barrett, C., Bellemare, M., & Hou, J. (2010). Reconsidering conventional
explanations of the inverse productivity-size relationship. World
Development, 38(1).
Deininger, K., & Byerlee, D. (2012). The rise of large farms in land
abundant countries: Do they have a future? World Development, 40(4),
701–714.
Deininger, K., Byerlee, D., Lindsay, J., Norton, A., Selod, H., & Stickler,
M. (2011). Rising global interest in farmland: Can it yield sustainable
and equitable benefits? Washington, DC: World Bank.
Eastwood, R., Lipton, M., & Newell, A. (2010). Farm size. In P. L.
Pingali, & R. E. Evenson (Eds.), Handbook of agricultural economics.
Elsevier: North Holland.
Fairtrade International. (2013). Powering up smallholder farmers to make
food fair: A five point agenda.
FAO (2001). Supplement to the report on the 1990 world census of
agriculture. FAO statistical development series 9a. Rome: FAO.
FAO (2005). A system of integrated agricultural censuses and surveys.
Volume 1: World programme for the census of agriculture 2010: FAO
statistical development series 11. Rome: FAO.
FAO (2012). The State of Food and Agriculture 2012: Investing in
agriculture for a better future. Rome: FAO.
FAO (2013). 2000 world census of agriculture methodological review (1996–
2005): FAO statistical development series 14. Rome: FAO.
FAO. (2014a). FAOSTAT. Retrieved from <http://faostat.fao.org/site/
291/default.aspx>.
FAO (2014b). The State of Food and Agriculture 2014: Innovation in family
farming. Rome: FAO.
FAO. (2015). A data portrait of smallholder farmers: An introduction to a
dataset on small-scale agriculture. Retrieved from <http://www.fao.
org/economic/esa/esa-activities/esa-smallholders/dataportrait/en/>.
Garner, E. & de la O Campos, A. (2014). Identifying the ‘‘family farm”:
An informal discussion of the concepts and definitions. ESA Working
Paper 14-10. Rome: FAO.
Government of Bangladesh (2010). Census of Agriculture 2008. Structure
of Agricultural holdings and livestock population (Vol. Volume 1).
Dhaka: Bangladesh Bureau of Statistics.
Government of Ethiopia (2007). Agricultural Sample Survey 2006/2007.
Volume I report on area and production of crops. Private peasant
holdings, meher season. Addis Ababa: Central Statistical Agency.
Government of Guatemala (2004). IV Censo Nacional Agropecuarios.
Caracterı
´sticas generales de las fincas censales y de productoras y
productores agropecuarios (Resultados Definitivos). Tomo I. Guatema-
la City: Instituto Nacional de Estadı
´stica.
Government of Guatemala (2006). Encuesta Nacional de Condiciones de
Vida. Guatemala City: National Institute of Statistics.
Government of India (2012). Agriculture Census 2010–11. Phase 1. All
India report on number and area of operational holdings (provisional).
New Delhi: Ministry of Agriculture.
Government of Malawi (2010). National Census of Agriculture and
Livestock 2006/07. Main report. Lilongwe: National Statistics Office.
Government of Rwanda (2010). National Agricultural Survey 2008. Kigali:
National Institute of Statistics of Rwanda.
Government of the People’s Republic of China (2009). Abstract of the
Second National Agricultural Census. Beijing: National Bureau of
Statistics of China.
Government of the Russian Federation (2006). All-Russia Census of
Agriculture: Russian Federation Summary and Country-Level Data.
Moscow: Statistics of Russia Information and Publishing Center.
Hazell, P., Poulton, C., Wiggins, S., & Dorward, A. (2010). The future of
small farms: Trajectories and policy priorities. World Development, 38
(10), 1349–1361.
HLPE (2013). Investing in smallholder agriculture for food security. A
report by The High Level Panel of Experts on Food Security and
Nutrition (Vol. 6). Rome: FAO.
THE NUMBER, SIZE, AND DISTRIBUTION OF FARMS, SMALLHOLDER FARMS, AND FAMILY FARMS WORLDWIDE
Q
13
Please cite this article in press as: Lowder, S. K. et al. The Number, Size, and Distribution of Farms, Smallholder Farms, and Family
Farms Worldwide
q
,World Development (2016), http://dx.doi.org/10.1016/j.worlddev.2015.10.041

IFAD (2011). Proceedings. In IFAD conference on new directions for
smallholder agriculture, 24–25 January 2011. Rome: IFAD.
IFAD & UNEP. (2013). Smallholders, food security and the environment.
Rome.
Jayne, T. S., Chamberlin, J., & Headey, D. D. (2014). Land pressures, the
evolution of farming systems, and development strategies in Africa: A
synthesis. Food Policy, 48, 1–17.
Larson, D. F., Otsuka, K., Matsumoto, T., & Kilic, T. (2014). Should
African rural development strategies depend on smallholder farms? An
exploration of the inverse-productivity hypothesis. Agricultural Eco-
nomics, 45(3), 355–367.
Lerman, Z., & Sedik, D. (2013). Chapter 22: Russian agriculture and
transition. In M. Alexeev, & S. Weber (Eds.), The Oxford handbook of
the Russian economy (pp. 514–543). New York, NY: Oxford University
Press.
Masters, W. A., Djurfeldt, A. A., De Haan, C., Hazell, P., Jayne, T.,
Jirstro
¨m, M., et al. (2013). Urbanization and farm size in Asia and
Africa: Implications for food security and agricultural research. Global
Food Security, 2(3), 156–165. http://dx.doi.org/10.1016/j.
gfs.2013.07.002.
Nagayets, O. (2005). Small farms: Current status and key trends. In The
future of small farms: Proceedings of a research workshop, Wye, UK,
June 26-29, 2005. Washington, DC: International Food Policy
Research Institute.
Nie, F. & Fang, C. (2013). Family farming in China: Structural changes,
government policies and market development for growth inclusive of
smallholders. Rome: Presentation made at FAO Friday, 13 December
2013.
Sedik, D., & Lerman, Z. (2008). Rural development and food security.
Development and Transition, 2–5.
Wiggins, S., Kirsten, J., & Lambi, L. (2010). The future of small farms.
World Development, 38(10), 1341–1349.
World Bank (2007). World Development Report 2008: Agriculture for
development. Washington, DC: World Bank.
World Bank (2012). World Development Indicators. Washington, DC.
World Bank. (2015) Living Standards Measurement Survey. Available at
<http://econ.worldbank.org/WBSITE/EXTERNAL/EXTDEC/EX-
TRESEARCH/EXTLSMS/0,menuPK:3359053~pagePK:64168427~
piPK:64168435~theSitePK:3358997,00.html>.
APPENDIX A. SUPPLEMENTARY DATA
Supplementary data associated with this article can be
found, in the online version, at http://dx.doi.org/10.1016/
j.worlddev.2015.10.041.
ScienceDirect
Available online at www.sciencedirect.com
14 WORLD DEVELOPMENT
Please cite this article in press as: Lowder, S. K. et al. The Number, Size, and Distribution of Farms, Smallholder Farms, and Family
Farms Worldwide
q
,World Development (2016), http://dx.doi.org/10.1016/j.worlddev.2015.10.041