The World’s Largest Open Access Agricultural & Applied Economics Digital Library

This document is discoverable and free to researchers across the

globe due to the work of AgEcon Search.

Help ensure our sustainability.

Give to AgEcon Search

AgEcon Search
http://ageconsearch.umn.edu
aesearch@umn.edu

Papers downloaded from AgEcon Search may be used for non-commercial purposes and personal study only.
No other use, including posting to another Internet site, is permitted without permission from the copyright
owner (not AgEcon Search), or as allowed under the provisions of Fair Use, U.S. Copyright Act, Title 17 U.S.C.

No endorsement of AgEcon Search or its fundraising activities by the author(s) of the following work or their
employer(s) is intended or implied.
 The Likely Impacts of the EU
Deforesta�on Regula�on
Christopher L. Gilbert
Preliminary dra�: 18 February 2024
This revision: 26 February 2024

Abstract

The EU Deforesta�on Regula�on (EUDR) will introduce stringent due diligence

requirements on the import of seven major tropical agricultural commodi�es into
the EU with the objec�ve of limi�ng deforesta�on in the producing countries.
The greatest impact is likely to be in cocoa and coffee where Europe is responsible
for a large share of world consump�on; and in palm oil which has driven
substan�al deforesta�on. The commodity supply chains are complex. In
par�cular, crop produced by smallholder farmers is aggregated prior to export.
Tracking the deforesta�on status of these aggregated packets is a major and
poten�ally costly undertaking. It is likely that this will involve some restructuring
of supply chains, favouring large farms over smallholdings and interna�onal
trading companies over na�onally-based exporters. These developments are
seen by some producing country governments as imperialis�c. EUDR-compliant
supplies will earn a premium and this will raise prices for European consumers.
Producers who are able to comply will benefit from the premium but will bear
the compliance cost. Overall there will be a net pecuniary loss. Deforesta�on
benefits will only emerge as new plan�ng takes place and will depend on whether
other consuming countries introduce similar legisla�on.

SAIS Europe, Johns Hopkins University, via Beniamino Andreata 3, 40126 Bologna, Italy.
Email: christopher.gilbert@jhu.edu

This paper has been prepared for the 2024 mee�ng of the Agricultural Economics Society,
Edinburgh, 18-20 March 2024. It draws on analysis previously discussed in a workshop
“Reshaping Eco-Social Paradigm towards Innova�on and Sustainability” at the University of
Shkodër, Albania, 7-8 November, 2023. I am grateful to Robin Dand for comments on the
ini�al dra�.
1. Introduc�on

The EU Deforesta�on Regula�on (EUDR) is a regula�on of the European Parliament enacted

on 31 May 2023 (European Union, 2023). Ar�cle 1 of the Regula�on sets out two objec�ves:

• minimiza�on of the EU’s contribu�on to deforesta�on and forest degrada�on

worldwide, thereby contribu�ng to a reduc�on in global deforesta�on; and
• reduc�on of the EU’s contribu�on to greenhouse gas emissions and global
biodiversity loss.

In sec�on 2, I outline the opera�on of the EUDR. Sec�on 3 provides a brief introduc�on to
the literature on deforesta�on. It reports evidence that deforesta�on has tended to decline
over the past two decades but s�ll remains important. Increases in pasture for the
produc�on of catle meat produc�on have been the single most important driver of
deforesta�on with palm oil produc�on as the second most important driver. The EU, and
Europe more generally, have seen net increases in forest cover over the past two decades,
but may be seen as responsible for deforesta�on through the import of commodi�es that
have benefited from deforesta�on. Sec�on 4 reports es�mates of foresta�on embodied in
European imports. Palm oil is the single most important cause of this embodied
deforesta�on, followed by cocoa, soybeans and coffee.

Sec�on 5 turns to the problems that will be faced in implemen�ng the EUDR. The first
major issue is the availability of detailed forest cover maps to allow assessment of the
deforesta�on status of each producer’s commodity sales. The EUDR requires that European
importers ensure EUDR-compliance. The implied due diligence process requires the importer
to be able to track imports back to the individual producers. The vast majority of imports of
EUDR commodi�es are indirect and involve supply chain intermediaries. Product is
aggregated at each step in the chain. Accurate tracking of the deforesta�on status of
aggregated commodity parcels will be a complicated and poten�ally costly undertaking.
There are fears that this may result in European importers choosing to source from
planta�ons rather than smallholders or other smallholders where this is possible.

The EUDR is explicitly extra-territorial. It is resented and even seen as imperialis�c by

some producer countries. Voluntary agreements that cer�fy sustainability are an alterna�ve.
The best known is the Roundtable on Sustainable Palm Oil (RSPO). I discuss this voluntary
approach in sec�on 6. There is at best mixed evidence that the RSPO limits deforesta�on,
and the premium obtained by cer�fied producers appears insufficiently large to make
cer�fica�on atrac�ve to smallholders. Its proponents hope that the EUDR will be more
successful but risk that deforesta�on may undermine respect for the sovereignty of
producing na�ons.

I look at the likely impact of the EUDR from two perspec�ves. First, in sec�on 7, I set
out a simple model which can be used to evaluate the welfare impacts in terms of the
(possible) gains to producers and the costs imposed on European consumers. The model

1
pivots on the premium that emerges between EUDR-compliant and non-compliant supplies.
There will be no impact in the short run on global commodity supplies and litle impact on
demand so prices outside Europe should be unaffected. European consumers are made
unambiguously worse off though payment of the premium. Those producers that supplied
the EU market prior to the EUDR and who con�nue to supply compliant product will gain
addi�onal revenue but bear addi�onal costs. Producers of non-compliant product who
would have previously supplied European markets are worse off. It is not possible to say in
advance whether producers as a group will benefit or lose from the Regula�on but any
benefit will be less than the loss to consumers. The EUDR premium is likely to be highest in
cocoa and palm oil but there is no current basis for es�ma�ng its likely magnitude.

Secondly in sec�on 8, I look at the impacts on deforesta�on. I suggest that, in the

absence of comparable controls in other major consuming countries, the main impact of the
Regula�on in the beef, natural rubber, and soybean markets will be diversion of non-
compliant supplies to non-European markets with litle or no impact on deforesta�on. In
cocoa, coffee and palm oil, any deforesta�on benefits will only emerge over �me as new
plan�ng takes place. Sec�on 9 concludes that the EUDR will be effec�ve in reducing the
extent of deforesta�on embodied in EU imports, albeit at significant cost, but is likely to
have only a small effect on deforesta�on itself unless the United States and China adopt
similar measures.

2. The Regula�on

The EUDR regulates imports into the EU of seven primary commodi�es, the produc�on of
which is viewed as responsible for deforesta�on in the producing countries. The seven
commodi�es are catle (beef), cocoa, coffee, palm oil, natural rubber, soya, and wood
(European Union, 2023, Ar�cle 2). Annex 1 of the Regula�on specifies that regula�ons
extend to products that have been “fed” or “made using” these seven commodi�es. Hence
leather, chocolate and cocoa confec�onary, roasted coffee, glycerol, tyres and other
products made from natural rubber, manufactured wooden objects, paper and printed
books all fall under the regula�on. Imports of any of the listed products into the EU must be
accompanied by a “due diligence statement” demonstra�ng that the merchandise is
deforesta�on-free and has been produced in accordance with relevant na�onal legisla�on
(Ar�cle 3). In rela�on to the Regula�on, “deforesta�on-free” is taken as meaning that the
land on which the commodi�es have been produced has not been subject to deforesta�on
since 31 December 2020 (Ar�cle 2, defini�on 13). Forwood et al. (2023) provide a good
short summary of the regula�on.

The Regula�on makes importers responsible for ensuring compliance with the EUDR
provisions through the submission of due diligence statements although SME importers face
more limited obliga�ons. Importers are required to carry out risk assessment exercises to
ensure that merchandise is indeed EUDR-compliant (Ar�cle 10). These must be based on the
collec�on of detailed informa�on. EU member states are required to carry out annual checks

2
on importer compliance (Ar�cle 16). It is envisaged that compliance checks will be calibrated
against the perceived riskiness of the expor�ng country, with exporters based on a three-�er
classifica�on into “low, standard or high risk” countries (paragraph 68). Importers can be
fined for non-compliance with the size of the fine being propor�onal to assessed
environmental damage but up to 4% of total EU turnover (Ar�cle 25). The EUDR comes into
force from 30 December 2024 and applies to all products produced a�er 29 June 2023.

3. Deforesta�on

The quan�fica�on of deforesta�on raises conceptual, measurement and informa�on issues.

At a na�onal or regional level, we can dis�nguish between gross and net forest loss and
between forest loss and forest degrada�on. Further, forest loss can be permanent or
temporary, as when an area of forest is logged or destroyed by wildfire. Pendrill et al. (2022,
page 2) define deforesta�on as a “persistent conversion of natural forest to any other land
use”. This is a measure of Tree Cover Loss (TCL). Cur�s et al. (2018) reserve the term
deforesta�on for “abrupt transi�on for land with trees to land without trees” (emphasis
added). Their defini�on focuses aten�on on land clearing as dis�nct from gradual evolu�on.
Either defini�on is possible but, if concern is with clima�c impact, the abruptness of any
transi�on is not relevant. Hansen et al. (2013) produced high resolu�on satellite-based maps
of world forest cover. These maps derive from satellite-based remote TCL sensing and form
the basis for the widely-used Global Forest Change (GFC) dataset. 1 GFC TCL data may not
dis�nguish between permanent and transitory losses and between deforesta�on and forest
degrada�on (Cur�s et al., 2018).

The FAO publishes a five yearly Global Forest Resources Assessment (GFRA) primarily
based on land use sta�s�cs at the na�onal level but complemented by remotely sensed
informa�on.2 The FAO defini�on of deforesta�on excludes conversion of natural forest to
forest planta�on which is included in the
GFC defini�on. That dis�nc�on is
important for biodiversity concerns but
less so in terms of clima�c change. FAO
(2020) is the most recent report based on
the GFRA data and provides a coarse but
comprehensive assessment of the state of
the world’s forests. The report counsels
that net forest loss differs from
deforesta�on because of reforesta�on
Figure 1: Net forest loss (world), 1990-2020.
and natural forest expansion. According
Source: FAO (2020) to the report (page xi), the rate of net
forest loss decreased substan�ally over

1
htps://data.globalforestwatch.org/documents/941f17325a494ed78c4817f9bb20f33a/explore
2
htps://www.fao.org/forest-resources-assessment/en/

3
the period 1990–2020. See Figure 1. Net deforesta�on is some countries has been offset by
net forest expansion in others. Over the period analyzed in the report, South America and
Africa have seen the most substan�al net deforesta�on while Asia and Europe have seen
significant net reforesta�on. There has only been modest net change in Oceania and North
and Central America. On an acreage basis, net forest loss has been greatest in Brazil, the
DRC, and Indonesia and net reforesta�on greatest in China, Australia and India (ibid, page
18).

The EUDR refers both to deforesta�on and forest degrada�on. It defines

“deforesta�on” as the conversion of forest to agricultural use, whether or not human-
induced; and “forest degrada�on” as changes in forest cover including conversion of natural
(“primary”) forest into planta�on, cul�vated forest or other wooded land (European Union,
2023, Ar�cle 2). The Regula�on makes clear that its concern is with the impacts of
deforesta�on on all of climate change, biodiversity loss and the welfare of indigenous
peoples. The objec�ve is that of promo�ng “deforesta�on-free supply chains” (paragraph
41). This characteriza�on is closer to the Pendrill et al. (2022) deforesta�on measure than to
that underlying the GFRS.

Hosonuma et al. (2012) provided the first systema�c assessment of the role of
agriculture as a driver of deforesta�on. The importance of the various factors varies both
geographically and with regard to the country’s level of development, and specifically the
extent of pre-exis�ng forest cover. Using GFRA data, they es�mate that, overall, agriculture
was responsible for 73% of deforesta�on over the two decades from 1990. 3 This es�mate
predates the availability of the GFC data. Pendrill et al. (2019) arrive at a lower es�mate
(45%-65%) using a hybrid approach combining GFC data with a land balance model based of
the GFRA data. Pendrill et al. (2022) comment that the quality of the GFRA land alloca�on
data, which are o�en imputed or es�mated, can be problema�c. Using GFC data over the
five year period 2011-15, the later authors report a 60%-83% range for the importance of
agriculture as a deforesta�on driver, commen�ng that the wide range reflects uncertainty as
to what propor�on of TCL is permanent as dis�nct from resul�ng from shi�ing agricultural
produc�on (prevalent in much of Africa).

In order to assess the likely impact of the EUDR, we need to move to a crop-based
analysis. Pendrill et al. (2019) and Goldman et al. (2020) both provide es�mates. Remote
sensing is not currently sufficiently developed to give accurate crop iden�fica�on. This
difficulty is accentuated where there is inter-cropping and where plan�ng is extensive (as is
o�en the case in cocoa), and where tree crops are grown under forest cover (again o�en the
case in cocoa but also some�mes in coffee). The Pendrill et al. (2019) es�mates provide
greater detail but derive from the GFRA data and rely on an assump�on that crop
composi�on remain unchanged as agricultural land area expands. The es�mates in Goldman
et al. (2020) are based on overlaying informa�on from a number of different datasets. This

3
This es�mate is similar to an earlier es�mate given by Geist and Lambin (2002).

4
gives them very detailed data on oil palm, natural rubber and wood fibre; and for pasture
and soybeans in parts of South America. They adopt a coarser approach for cocoa and
coffee, where there is a general lack of detailed and comprehensive informa�on on the
precise loca�on of produc�on, and also for other commodi�es in areas not covered by their
overlaying data. Their numbers are obtained by overlaying detailed remotely sensed
informa�on on forest cover (or its absence) with na�onal informa�on on crop yields.

Figure 2: Es�mated forest loss by commodity (Pendrill et al. (2022)

Figure 2, which derives from the informa�on in Pendrill et al. (2020, Table S6),
compares the Pendrill et al. (2019) and Goldman et al. (2020) es�mates. There are two
points to be noted. First, Pendrill et al. (2019) es�mate overall deforesta�on over their
analysis period (2005-13) as 5.5 million hectares per annum (Mha pa) while Goldman et al.
(2020), who analyze the longer period 2001-15, imply a figure of 8.2 Mha pa. As noted
above, Pendrill et al. (2020) regard the Pendrill et al. (2019) es�mate as “conserva�ve”. This,
together with the different analysis periods, may go some way to explain the higher
es�mates given by Goldman et al. (2020). Second, Goldman et al. (2020) do not give figures
for forest loss to maize, rice or cassava, which are included in an unatributed category (not
charted) along with wood products and subsistence agriculture.

• The two sets of es�mates are in agreement that increases in pasture for the
produc�on of catle meat produc�on has been the single most important driver of
deforesta�on. The Pendrill et al. (2019) es�mates see increases in pasture as

5
 responsible for over 40% of the total. It has been par�cularly important in La�n
America where it accounted for 60% of the net forest loss (72% in Brazil).
• The two studies also agree in seeing palm oil produc�on as the second most
important driver with losses concentrated in South-East Asia. Pendrill et al. (2019)
report that it accounted for 46% of deforesta�on in Indonesia and 69% in Malaysia.
World Resources Ins�tute (WRI) �me series data on deforesta�on show a clear link
to palm oil prices in the previous year. 4
• The produc�on of soybeans is seen as accoun�ng for a similar degree of net forest
loss, again in La�n America, in par�cular in Brazil, but also in Argen�na, Bolivia and
Paraguay.
• Forestry products (not charted in Figure 2) were also an important contributor to net
forest loss. WRI figures show 90% of the loss of primary forest in Indonesia as
atributable to this category, mainly to forest planta�ons.
• Tropical export crops (cocoa, coffee and natural rubber) were seen as rela�vely
unimportant drivers of net deforesta�on accoun�ng around 5% of the total loss.

In summary, of the seven commodi�es specified in the EUDR, four (catle meat, oil palm,
soya and wood products) are seen as having been important drivers of deforesta�on and
three (cocoa, coffee and natural rubber) as rela�vely unimportant drivers. Nevertheless,
these three commodi�es may have played in important role in deforesta�on in specific
countries.

4. Imported deforesta�on

This sec�on addresses the ques�on of the extent to which deforesta�on is embodied in EU
imports of the seven commodi�es specified in the EUDR. Pendrill et al. (2019) dis�nguish
between deforesta�on atributable to commodi�es des�ned for consump�on in the
producing country and that embodied in exported commodi�es. They state that catle meat
is primarily consumed locally with only 11% in their dataset exported. On the other hand,
cocoa, coffee, natural rubber, and palm oil are primarily export crops. Even though they are
responsible for a smaller overall share of deforesta�on, they may poten�ally account for a
high propor�on of deforesta�on embodied in European imports.

Table 1 shows EU shares of six of the seven EUDR commodi�es in world produc�on of
the commodity.5 (I omit �mber since most wood is imported in manufactured wood
products making it difficult to es�mate the commodity share). The EU shares include semi-
processed commodity products converted into commodity equivalents. 6 The table shows

4
htps://research.wri.org/gfr/forest-extent-indicators/deforesta�on-agriculture#how-much-forest-
has-been-replaced-by-specific-agricultural-commodi�es
5
Data sources: World produc�on – FAOSTAT. EU-27 imports – Eurostat. All figures are 2017-2021
averages.
6
Beef: boneless beef, 1.16; leather 0.65. Coffee: roast, 1.19. Cocoa: powder 0.59; paste, 1.25; buter,
0.67; chocolate, 1.20 (powder and buter are treated as joint products). Oil palm: palm kernel oil,

6
substan�al import shares for cocoa and coffee, a negligibly small share for beef, and
intermediate shares for palm oil, soybeans and natural rubber.

Table 1
European import share of world produc�on
Beef Cocoa Coffee Natural rubber Palm oil Soybeans
EU-27 0.7% 39.4% 40.6% 11.0% 18.5% 9.3%
CH+N+UK <0.1% 2.8% 3.8% 0.2% 0.8% 0.6%
EU+3 0.7% 42.2% 44.3% 11.1% 19.3% 10.0%
Source and defini�ons: see text. Column totals may not sum because of rounding.

Bradford (2020) argues that, in certain industrial contexts, there is a regulatory “race
to the top” since firms are forced to choose between compliance with the most exac�ng set
of regula�ons or alterna�vely retrea�ng from the highly regulated markets. Since the EU has
tended to be the most stringent regulator, she argues that the EU has become the global
regulator in those industries. This is the “Brussels effect”, named a�er the “California effect”
(Vogel, 1995) in the United States automobile market. Perkins and Neumayer (2012)
examined whether the California effects extends across na�onal borders. The mechanism is
that global companies will wish to sell products manufactured to the same specifica�on in
all the countries in which they operate.

Figure 3: EU import shares compared with shares of forest loss

This argument appears weak in rela�on to products manufactured from the EUDR
commodi�es, with the possible excep�on of rubber goods, but we should nevertheless

5.71; oilcake from palm, 0.30. Natural rubber: rubber manufactures, 0.15. Soybeans: soybean oil,
2.81; soy sauce, 0.50; soy oil cake (meal), 0.6 (soybean oil and meal are treated as joint products).

7
expect a Brussels effect for the broader European market where Switzerland and the UK are
the two most important extra-EU countries. Indeed, the UK government has announced
similar regula�ons but on a more restricted list of commodi�es UK Parliament, 2023). 7
Table 1 therefore also reports the share of imports into three EU neighbours (Norway,
Switzerland and the UK) , ne�ng out EU-27 exports to these countries, in recogni�on that
importers into these countries are likely to insist on EUDR-compliance.

Figure 3 charts this informa�on and adds the Goldman et al. (2020) forest loss
percentages previously charted in Figure 2. Cocoa and coffee, in which Europe imports a
large propor�on of world produc�on, are rela�vely unimportant in accoun�ng for
deforesta�on at the global level; the opposite is true of catle produc�on, which is a major
driver of deforesta�on. The correla�on of the Goldman et al. (2020) forest loss percentages
with the EU import shares is -0.65.

Figure 4: Es�mated European imported deforesta�on (000ha pa)

By interac�ng the Goldman et al. (2020) forest loss es�mates over 2001-15 and the
European import shares charted in Figure 3, I can obtain an es�mate of the approximate
extent of deforesta�on embodied in European imports. The results are charted in Figure 4.
These figures indicate that EU palm oil imports embody the largest quantum of
deforesta�on, es�mated at 82,000 ha p.a., followed by cocoa es�mated at 30,000 ha p.a.
and coffee and soybeans each at 24,000 ha p.a. These es�mates are based on the strong
assump�on that the extent of deforesta�on within each country is independent of the crop
grown. It is, for example, possible that in a cocoa-expor�ng country with high recorded
deforesta�on, that deforesta�on is due en�rely to forest loss for �mber produc�on.

7
Only preliminary details are available at the �me of wri�ng. Four commodi�es are provisionally
specified: catle products (excluding dairy), cocoa, palm oil, and soy. There is no men�on of coffee,
natural rubber or �mber.

8
 Figure 5: Es�mated European imported deforesta�on as a share
of world total deforesta�on
The rankings shown in Figure 4 change drama�cally if imported deforesta�on is
expressed as a share of the Goldman et al. (2000) es�mates of total deforesta�on driven by
each ERUDR commodity – see Figure 5. EU imports are es�mated as embodying 4.0% of
total world deforesta�on resul�ng from agriculture and 2.8% of total world deforesta�on.
23.5% if coffee and 22.8% in cocoa. Although coffee and cocoa are not major drivers of
deforesta�on, EU imports embody a high propor�on of the deforesta�on arising from these
ac�vi�es

5. Commodity supply chains

Commodity supply chains are complicated, and possibly more complicated than the
European Parliament realized when it enacted the EUDR. This complexity makes
implementa�on of the required due diligence controls difficult and possibly costly. It may
result in unintended collateral effects favouring one group of supply chain par�cipants over
another.

There are two points in the commodity supply chain issues at which EUDR-compliance
issues may arise. The “first purchaser” (a company expor�ng to Europe or a supply chain
intermediary) of the commodity will need to provide evidence that the land on which the
commodity was produced has not suffered from deforesta�on since the cut-off date at the
end of 2020. It is envisaged that this will require precise geophysical loca�on of the farm or
plot on which produc�on has taken place. Absence of deforesta�on can than be established
by comparing satellite images showing the current forest cover at that loca�on with that at
or before the 2020 cut-off date. Governments and a number of commercial organiza�ons
produce interac�ve maps which permit iden�fica�on of the first cover in the small polygonal
region in which the producing farm is located. Figure 6 reproduces a forest cover map for
9
 Cameroon in 2020.8 It will only be
possible to demonstrate EUDR-
compliance by reference to a map of
this sort. It is currently unclear what
propor�on of the EUDR commodi�es
are produced in unmapped areas.

The second compliance issue

relates to aggrega�on. The EU importer
will need to report due diligence that
shows the en�re quan�ty imported
sa�sfies this requirement. In a small
propor�on of instances, a European
impor�ng company purchasing at
origin will be the first purchaser. This
may be the case with some soybean,
palm oil and natural rubber purchases
from large farms or planta�ons.
However, in the majority of instances,
EU imports will be intermediated by
Figure 6: Forest cover map, Cameroon, 2020 traders, coopera�ves, export
companies, or all of these. zu
Ermgassen et al. (2022), who discuss the catle, cocoa, palm oil and soybeans, show that
“indirect sourcing” is prevalent in all four supply chains and is universal in the Ivorian cocoa
and Brazilian live catle chains. This will always be the case with smallholder produc�on
where intermedia�on will involved aggrega�on of the products of a large number of
farmers. For example, West African cocoa beans are invariably bulked and exported either in
containers or bulk carriers. A 40 foot (12.2 m) container carries around 25 tons of cocoa
beans which may have been produced on around 400 farms. A bulk carrier will take around
5,000 tons of beans may have been grown on around 80,000 farms. 9 EUDR-compliance
requires that EU importers need to maintain no-deforesta�on records of the en�rety of their
imports. Under pre-EUDR arrangements, European importers had no interest in knowing the
detail of where crop was produced, and in any case, traceability would have been lost
whenever aggrega�on took place.

The extent of aggrega�on differs both across the EUDR commodi�es and across
origins:

• Prevalence of smallholder produc�on will increase the need for aggrega�on.

8
Forêts communautaires and communales are shown in blue ; managed forests (forêts UFA) are in
brown). Source: Ministère de Forêts et de la Faune, Cameroon; and World Resources Ins�tute,
htps://cmr.forest-atlas.org/
9
Federa�on of Cocoa Commerce, private communica�on.

10
 • This need will be further increased where crops are grown extensively and are inter-
cropped, as is o�en the case with cocoa, and so individual farmers only produce
small quan��es
• … and where the commodity is homogeneous (cocoa and robusta coffee but not
arabica coffee).

The FAO defines smallholder producers as “farmers … who manage areas varying from
less than one hectare to 10 hectares”. Lowder et al. (2016) provide es�mates of the global
distribu�on of smallholder farms at the na�onal level. I am not aware of comparable
informa�on at the commodity level. Table 2 reports the results of interac�ng the Lowder et
al. es�mates with shares of commodity produc�on averaged over the years 2017-21. 10 The
es�mates should be treated as approximate – the es�mates assume that the farm
distribu�on within each country is independent of the crop produced and the Lowder et al.
data are not comprehensive and derive from surveys undertaken at different dates.
Nevertheless, the rankings are the same irrespec�ve of the farm size taken as iden�fying a
smallholding: smallholdings are seen as the dominant produc�on mode for palm oil and
natural rubber, followed by cocoa. 11 Smallholdings are less prevalent in beef and soybeans
produc�on. Coffee is intermediate.

Table 2
Es�mated farm size distribu�ons
0-1 ha 0-2 ha 0-5 ha 0-10 ha
Beef 31.2% 40.1% 52.7% 59.6%
Cocoa 49.6% 64.8% 82.3% 90.6%
Coffee 42.9% 55.6% 71.8% 79.3%
Natural rubber 63.6% 79.5% 93.9% 97.5%
Palm oil 64.9% 81.3% 95.1% 97.8%
Soybeans 17.6% 25.1% 39.0% 48.3%
Author’s calcula�ons weigh�ng data in Lowther et al. (2016), WEB Table 3, by country produc�on
shares (averaged over 2017-21).

Textbook descrip�ons of market interac�ons typically see concentra�on as being

located either with producers (monopolis�c market structures) or purchasers
(monopsonis�c structures). Many supply chains in the primary commodity sector are more
complicated since the concentra�on is at an intermediate point in the chain. The commodity

10
Source for farm size distribu�on (111 countries): htps://www.fao.org/family-
farming/detail/en/c/273864/. The es�mates derive from survey undertaken at differing dates. Source
for produc�on sta�s�cs (193 countries): FAOSTAT. Numbers for the 82 countries not included in the
Lowder et al. (2016) dataset are extrapolated from regional averages.
11
According to the Interna�onal Ins�tute for Sustainable Development (IISD), around 90% of cocoa
farmers produce on less than five hectares. htps://www.iisd.org/ssi/commodi�es/cocoa-
coverage/#:~:text=About%2090%25%20of%20growers%20are,lands%20less%20than%205%20hectares.

11
is produced by a large number of farmers (a very large number in the case of smallholder
produc�on) and is processed by a large number of users. A much smaller number of
commodity trading companies sit between these two groups.

In the remainder of this sec�on, I focus on cocoa, coffee, and palm oil. Over 98% of EU
imports of cocoa beans originate from West Africa where produc�on is almost en�rely by
smallholders and where produc�on is o�en extensive involving intercropping. Fold (2002)
characterizes the cocoa market as bipolar – the structure resembles an egg-�mer with a
broad base and top but very narrow in the middle. See also Carodenuto and Buluran (2021)
and Gilbert (2024). Cocoa market concentra�on at the processing stage is atributable to
economies of scale in both bulk transport and in processing itself. See also Fold (2001),
Gilbert (2009) and Araujo Bonjean and Brun (2016). See van Duijn (2013) on palm oil. Lyons-
White and Knight (2018, page 306) report market par�cipants as describing the palm oil
supply chain as being “hourglass-shaped, with many different stakeholders at the supply and
demand ends and a small number of trading companies in the middle”.

The hourglass structure is illustrated for palm oil in Figure 7, reproduced from Lyons-
White and Knight (2018, Figure 2). The large number of oil palm farmers, signified by small
palm trees, are at the top of the hourglass. Local traders convey the palm oil and kernels to a
smaller number of millers and processors who sell to a small number of trading companies,
many of which will be mul�na�onals. This involves aggrega�on of the products of large
numbers of farmers. The traders in turn sell to refiners and manufacturers who use the palm
oil in the produc�on of mul�ple “fast moving consumer goods” (FMCGs) which are then
retailed to very large numbers of final consumers. The diagram allows either for complete
specializa�on at each stage of the supply chain (solid red line) or for a degree of ver�cal
integra�on (broken line).

The EUDR imposes compliance obliga�ons at the import stage, i.e. on the commodity
processors, refiners and manufacturers in the lower bulb of the Figure 7 hourglass. These
obliga�ons require full traceability of the products they purchase back to the farmers at the
top of the Figure. Traceability needs to be maintained through the lower part of the bulb in
which o�en small quan��es of the commodity are used in the manufacture of the FMCGs.

Indonesia and Malaysia are the most important producers of palm oil. Averaging over
the five years 2017-2021, Indonesia accounted for 60.3% of world palm oil produc�on, 12
mostly grown intensively. Smallholders predominate in Sumatra, which is the most
important producing region, but larger scale industrial planta�ons have dominated the
expansion in Kalimantan and Papua over the two most recent decades (San�ka et al., 2021).
Zhunusova et al. (2022) report 42% of Indonesian palm oil and 28% of Malaysian palm oil
was produced on planta�ons. Lyons-White and Knight (2018, page 306) discuss the prac�cal
difficul�es of ascertaining deforesta�on commitments in the palm oil supply chain. Supply

12
Source: FAOSTAT.

12
chain intermediaries stated that fruit bunches “were o�en traded between mul�ple
smallholders … before arriving at a mill” making traceability “difficult or impossible”. They
state (page 310) that “complexity of the palm oil supply chain was the major barrier to the
implementa�on of no-deforesta�on commitments”.

Figure 7: The palm oil supply chain (from Lyons-White and Knight, 2019)
The coffee and natural rubber supply chains combine smallholder produc�on and
larger planta�ons. The website Coffee Intelligence states that “the EU has failed to recognise
a difference [in rela�on to EUDR compliance] between smallholders and larger estates”. 13 In
other respects, the coffee supply chain is similar to that of cocoa. Coffee Intelligence argues
that “The coffee supply chain is complex and involves many intermediaries. … Traceability
has long been an issue for the industry, and it’s another factor why this may be so difficult to
implement” (ibid). However, unlike cocoa, coffee is typically grown intensively and the prices
paid for arabica beans can vary in rela�on to perceived quality. Moreover, processing is less

13
htps://intelligence.coffee/2023/08/eu-deforesta�on-laws/

13
concentrated than in cocoa and independent exporters play a greater role. The natural
rubber chain has similari�es with palm oil with a significant propor�on of direct links
between growers and processors and manufacturers at origin, in par�cular in the
manufacture of condoms and surgical and other latex gloves. By contrast, the soybeans
supply chain is simpler with a large propor�on of beans produced on large farms and sold
directly for export – see Weerdenburg (2023) for discussion.

EUDR compliance will involve less aggrega�on and hence be simpler and less costly
when imports into the EU are sourced from larger producers and not smallholders. This has
led to the charge that the EUDR will favour planta�ons over smallholders. Reuters reported
that EU coffee importers are turning away from African beans, predominantly produced by
smallholders and quoted one trader as sta�ng “Roasters are moving to big rich Brazilian
farmers”. 14 Similarly, the East Asia Forum reported that smallholders “may be
dispropor�onately affected by the EUDR” leading the authors to argue that “In the long run,
this may further reduce their ability to par�cipate in the palm oil market, further increasing
the dominance of large corpora�ons”. 15 Zhunusova et al. (2022) make similar arguments.

This issue does not arise in cocoa where smallholders account for almost the en�rety
of produc�on. However, a different collateral impact of the EUDR may arise in Côte d’Ivoire,
the most important producing country, accoun�ng for 63% of West African and 39% of world
produc�on.16 Six mul�na�onal trading companies 17 dominate Ivorian cocoa exports,
benefi�ng from economies of scale in transporta�on and superior access to credit, at the
expense of na�onal exporters. The Ivorian Conseil du Café-Cacao (CCC), which regulates the
Ivorian sector, regards this concentra�on as against the Ivorian na�onal interest 18 and has
moved to guarantee a share of the export trade to the local exporters, some of which are
poli�cally powerful. The mul�na�onal exporters are beter placed to ensure EUDR-
compliance than na�onal exporters with the consequence that the EUDR may strengthen
the posi�on of the former group. The na�onal authori�es will not welcome such a
development.

14
M. Angel and D. Kurniawa�, “Coffee firms turning away from Africa as EU deforesta�on law looms”.
Reuters, 19 December 2023. htps://www.reuters.com/markets/commodi�es/coffee-firms-turning-away-
africa-eu-deforesta�on-law-looms-2023-12-19/
15
M.A. Iswara, D.S. Nurshadrina, and A. Suryaahadi, “European Union palming off deforesta�on
regula�on to smallholders in Indonesia”, East Asia Forum, 10 October 2023.
htps://eastasiaforum.org/2023/10/10/european-union-palming-off-deforesta�on-regula�on-to-smallholders-
in-indonesia/
16
2017-21 average. Source: FAOSTAT.
17
Barry Callebaut, Cargill, Ecom Trading, Olam, Sucden, and Touton. The first four of these companies
accounted for 75% of global cocoa trading and processing in 2016-17 (Star�z et al., 2022).
18
“ «Le cartel de ces six mul�na�onales cons�tue un danger pour notre secteur cacao», a déclaré à
Reuters un responsable du Conseil du café-cacao (CCC) sous le sceau de l’anonymat. «Nous devons
réduire leur influence et assurer que tous les acteurs ont une place pour travailler.» ”.
CommodAfrica, 29 April 2021. htps://www.commodafrica.com/29-04-2021-la-cote-divoire-atribuerait-
20-des-exporta�ons-de-cacao-aux-operateurs-na�onaux/

14
 At the same �me, there may also be posi�ve collateral impacts in terms of improved
supply management. Cocoa in Côte d’Ivoire provides a case in point. Côte d’Ivoire
experienced civil war over the periods 2002-07 and 2010-11. The ending of each of the two
phases of war was associated with a large increase in produc�on in the years that
followed. 19 These drama�c increases in supply resulted in nominal cocoa prices at the world
level falling by 10.2% in deflated terms comparing the five years 2015/16 to 2020/21 with
the period 2005/06 to 2010/11. 20 The phases of rapid cocoa expansion took place around
four years a�er the end of each war episode, consistent with the �me cocoa trees reach
produc�on, and was associated with large scale internal migra�on from the north of the
country, and possibly also beyond, by families seeking land in the a�ermath of the war
disrup�on. The absence of accurate crop maps and clearly defined property rights were
instrumental in the inability of the Ivorian government to control this new plan�ng, some of
which may have been illegal. The need to comply with EUDR due diligence procedures
should assist in improving domes�c supply management in commodity-producing countries.

6. Extra-territoriality

The EUDR is explicitly extra-territorial. Net deforesta�on in the EU is nega�ve (see sec�on 3)
and of the seven commodi�es specified in the Regula�on, only �mber and a small quan�ty
of soybeans are produced within the EU. Although the EUDR does not atempt to regulate
either the loca�on or manner of commodity produc�on in the producing countries, it has
this effect indirectly since, by prohibi�ng the import of non-compliant supplies into the EU, it
also restricts the export possibili�es for these commodi�es. Furthermore, the EUDR
prohibits import into the EU of commodi�es produced on deforested land even where
deforesta�on is legal in the producing country (European Union, 2023, paragraphs 33-34).

The percep�on of extra-territoriality was ar�culated by the Indonesian Minister for

Economic Affairs Airlangaa Hartarto, with specific reference to palm oil where Indonesia is
the leading exporter. He stated “A key feature of the EUDR regula�on forcing other people
and countries to submit to internal EU regula�ons … is the regulatory imperialism that the
Global South must now confront”. 21 Again discussing the Indonesian palm oil sector, Iswara
et al. (2023) suggest that the EUDR might be seen as “a veiled protec�onist measure”
encouraging European consumers to subs�tute domes�cally produced sunflower for palm
oil. Almeida et al. (2023, page 5), who discuss EU environmental policy more generally, see

19
Produc�on rose by 44.6% from crop year 2009/10 (1.24 Mt) to 2014/15 (1.80 Mt) and by a further
17.2% to 2019/20 (2.10 Mt). Source: ICCO, Quarterly Bulletin of Cocoa Statistics (various issues).
20
Defla�on by the US Producer Price Index. Averages of the ICCO Indicator Price. 2005/06 – 2010/11:
$2354/t ($2410/t at 2010 values); 2016/17 – 2020/21: $2453/t ($2165/t at 2010 values). Source for
ICCO Indicator Price: ICCO, Quarterly Bulletin of Cocoa Statistics (various issues). Source for US PPI:
IMF, International Financial Statistics.
21
“EUDR: Hartarto versus Timmerman’s �n ear”. Palm Oil Monitor, 19 June 2023.
htps://palmoilmonitor.org/2023/06/19/

15
the EU as a “moral intervener” with the aim of “influencing decisions of foreign
governments … [and] crea�ng mechanisms of so� regulatory power for climate response”.

Voluntary cer�fica�on provides an alterna�ve approach to environmental

sustainability. The Roundtable on Sustainable Palm Oil (RSPO), 22 founded in 2004, is the best
known and most studied of such schemes. The Round Table on Responsible Soy Associa�on
(RTRS), 23 founded in 2006, has similar objec�ves but has atracted less academic interest.
Evidence on the effec�veness of the RSPO remains mixed. Brandi et al. (2015) find evidence
that smallholders o�en lack both the informa�on and the degree of organiza�on required
for RSPO-cer�fica�on. Catau et al. (2016) show that fire loss is less likely on RSPO-cer�fied
planta�ons rela�ve to non-cer�fied planta�ons in periods of low fire risk but not in periods
of high fire risk. Carlson et al. (2018) show that RSPO-compliance reduces deforesta�on.
Using village-level data, San�ka et al. (2021) find mixed impacts from RSPO cer�fica�on on
farmer welfare in villages with a history of market integra�on but nega�ve impacts in
villages which have previously depended on subsistence agriculture. Ruysschaert and Salles
(2024, page 440) see the RSPO as a bargain in which growers “preserve areas of forest of
high biodiversity value, to the benefit of … palm oil processors [and] consumer goods
manufacturers”. They remark that farmer par�cipa�on requires that they receive a sufficient
premium over non-cer�fied palm oil. They find that cer�fica�on proves too costly for many
smallholders in rela�on to the realized premium.

The major difference between voluntary cer�fica�on schemes, like the RSPO, and the
EUDR, is that farmers opt into voluntary schemes while, under the EUDR, they opt into
supplying EU customers. A propor�on of pre-EUDR suppliers of palm oil to the EU will not
have been cer�fied so the EUDR’s coverage should be larger, or at least, not smaller. It was
open to the Indonesian government to obligate RSPO cer�fica�on for palm oil exports but
they elected not to do this. EUDR due diligence controls take that decision out of Indonesian
hands in rela�on to exports to Europe.

Overall, the evidence indicates that the RSPO has at best been only moderately
effec�ve in limi�ng deforesta�on and that consumers have been unwilling to pay sufficiently
high premium to make the scheme atrac�ve to smallholder producers. By requiring due
diligence, which is analogous to cer�fica�on, the EUDR aims for complete producer coverage
with regard to supplies imported into Europe and may therefore hope for a larger impact on
deforesta�on but at the expense of over-riding elements of the sovereignty of producer
country governments.

7. Gainers and losers

The EUDR will cause a combina�on of higher intermedia�on costs, subs�tu�on between
EUDR-compliant and non-compliant producers, likely higher prices paid by European

22
htps://rspo.org/
23
htps://responsiblesoy.org/

16
consumers, and possibly higher prices to some EUDR-compliant producers. In Figure 8, I set
out a simple model that allows analysis of these impacts.

I consider a specific EUDR commodity. The model makes the simplifying assump�on
that world produc�on, European consump�on and non-European consump�on are all
unaffected by the EUDR. This allows me to suppose that the world, i.e. extra-European, price
of the commodity is unaffected. I also assume that, prior to the EUDR, producers self-sorted
into supplying the European and non-European markets on the basis of rela�ve
intermedia�on costs. The model therefore focuses on the costs of EU-compliance and of
diversion of supplies to and from the European market.

€
aa cc
slope sa slope sc

π
C
D
A F
bb
δ E
B slope sb

M1 M2 M
Figure 8: Simple analy�cal model
European imports, M, are measured along the horizontal axis. The ver�cal axis
measures the price paid by European consumers net of the world price. The compliance cost
for pre-EUDR suppliers is the line cc which is drawn as upward-sloping since costs will vary
depending on the origin country, farm loca�on and other factors. The final sec�on of cc is
drawn as ver�cal reflec�ng the possibility that a quan�ty M-M2 might have been produced
on deforested land and therefore could not be made compliant at any cost.

The line aa represents the supply curve of compliant commodity from producers who
were not supplying European markets prior to the EUDR. This line is again seen as upward
sloping, but reading from right to le�. The line aa intersects the ver�cal axis at δ (> 0),
represen�ng the addi�onal costs that these producers bear in supplying the European
market. The lines aa and cc intersect at the quan�ty M1 and premium π which equates
marginal compliance costs to marginal diversion costs. The quan�ty M1 is supplied by the
“tradi�onal” (pre-EUDR) suppliers and the quan�ty M-M1 is diverted from the non-European
market. (So long as marginal compliance costs are lower than π, these suppliers have no
incen�ve to divert to the non-European market). The line bb represents the supply curve of

17
the non-compliant commodity diverted from the European market. For simplicity, I suppose
that these quan��es incur the same diversion costs δ. 24

We can use Figure 8 to iden�fy welfare gains and losses from the EUDR:

• European consumers pay the premium π on the en�re quan�ty M imported. This
implies a welfare loss of A+B+C+D+E+F.
• EUDR-compliant tradi�onal suppliers supply the quan�ty M1. They obtain the EUDR
premium π and so obtain increased revenue A+B but bear the compliance costs given
by triangle B. Their net gain is triangle A.
• Non-EUDR-compliant tradi�onal suppliers divert quan�ty M-M1 to the non-European
market. Their addi�onal costs are E+F.
• Non-tradi�onal suppliers who divert the quan�ty M-M1 to the European market
obtain addi�onal revenue C+D+E+F but incur addi�onal costs of D+E. Their net gain is
C+F.
• Summing, the net producer gain is A+C-E which is drawn as posi�ve in Figure 8 but
which would be nega�ve if area E is sufficiently large.
• The overall welfare loss is B+D+2E+F. (The diversion costs E are borne both by the
non-compliant tradi�onal suppliers and by the non-tradi�onal suppliers that replace
them and hence enter this expression twice).

Overall, therefore, consumers lose from the EUDR by paying the EUDR premium. Compliant
producers benefit from the premium but also pay higher intermedia�on costs. Non-
compliant producers lose and bear addi�onal costs in supplying non-European markets.

The premium and the size of the welfare costs and benefits will become clear once the
Regula�on becomes opera�ve in 2025. However, we can use the Figure 8 model to advance
some general considera�ons. I consider two latent factors:

• the European share σ of total import of the commodity, displayed in Table 1; and
• the smallholder share τ of smallholders in domes�c produc�on displayed in Table 2.

I hypothesize that marginal compliance costs, measured by the slope sa, will rise with the
quan�ty of the commodity for which due diligence is required, so dsc dσ > 0 , and also with
the propor�on of producers that are smallholders, so dsc dτ > 0 . Similarly, I hypothesize that
marginal diversion costs, measured by the slope sc, will rise with the extent of diversion
required, so dsa dσ > 0 but I would not expect any clear rela�onship with τ. The diversion
cost intercept δ will depend on commodity characteris�cs and produc�on loca�ons. It is
unclear why δ would be related to the consump�on or smallholder shares.

24
The diagram implicitly assumes that the quan��es M-M1 both diverted by tradi�onal suppliers
onto the non-European market and diverted by non-tradi�onal suppliers onto the European market
are intra-marginal since otherwise there would be an impact on the non-European price.

18
 Table 3
Possible premium levels
European share Smallholder share EUDR premium
σ τ π (score/10)
Beef very low low 3 low
Cocoa very high high 9 high
Coffee very high medium 8 high
Palm oil medium very high 8 high
Natural rubber low very high 7 medium
Soybeans low low 4 low
European shares (σ) from Table 1 (row 3): Under 3% = “very low”, 4%-15% low; 16-30% =
“medium”, 30-40% = “high”; >40% =”very high”.
Smallholder share (τ) from Table 2, (column 3): Over 90%, “very high”; 80-90%, “high”; 60-70%,
“medium”; under 60%, “low”.
Score: “very high” 5, “high”, 4; “medium”, 3; “low”, 2; “very low”, 1.

We can use these hypotheses to obtain qualita�ve predic�ons of the likely premium
levels arsing from the EUDR and hence the welfare costs and benefits. The calcula�ons and
results are displayed in Table 2 where I use a five-way classifica�on for each of the European
share, reported in Table 1, and for the smallholder shares es�mated in Table 2. This allows
me to obtain a score out of ten for the likely premium level by summing the scores in the
two columns. On this, very crude, basis, the premium is seen as likely to be rela�vely high in
the cocoa, coffee, and palm oil markets, but somewhat lower in the natural rubber market. I
expect low premiums in the beef and soybean markets.

The analysis is crude and relates only to rela�ve magnitudes. Ot says nothing about the
absolute size of any premium and hence nothing on the absolute size of the welfare costs
and benefits sketched in Figure 8. The discussion is premised on a lack of comparable
regula�ons in other major consuming countries, in par�cular the United States and China.
Such developments would raise diversion costs by elimina�ng alterna�ve sales loca�ons and
hence would raise premiums and welfare costs and benefits.

8. Impact on deforesta�on

It is, of course, inevitable that, by moving away from the compe��ve alloca�on, the EUDR
will impose net pecuniary costs. The substan�ve issue is whether these costs are offset by
the benefits from reduced deforesta�on. The Regula�on states as objec�ve minimiza�on of
the EU’s contribu�on to deforesta�on and forest degrada�on worldwide, thereby
contribu�ng to a reduc�on in global deforesta�on – see sec�on 1. The Regula�on will
ensure a reduc�on in European imported deforesta�on but, to the extent that non-
compliant supplies are diverted to other jurisdic�ons, it may have only a small effect on
global deforesta�on.

As stated in sec�on 7, a crucial considera�on in this regard is whether similar

regula�ons are introduced in other consuming countries. So long as there is large

19
unregulated market, diversion costs will be low and incen�ves to expand produc�on will
remain despite the EUDR. Diversion costs will be higher the smaller is the unregulated share
of the world market.

China is a major consumer of natural rubber, palm oil and soybeans but accounts for
only a small share of world cocoa and coffee consump�on.25 There is no reason to expect
market-based pressure on non-European commodity processors to line up with the EUDR (a
“Brussels effect”) since processed products are largely sold on regional and not world
markets – see sec�on 4. Extension of deforesta�on controls to the North American market
would make diversion very costly in cocoa and coffee but China will need to be involved for
this to result in the other five EUDR commodi�es. Absent such an extension, the EUDR is
likely to do more go calm the consciences of European voters with regard to deforesta�on
arising from natural rubber, palm oil and soybeans produc�on than to have a significant
effect on deforesta�on itself. To quote Iswara er al. (2023), the EUDR will have “palmed off”
responsibility for deforesta�on onto other consuming countries without significantly
reducing its overall level.

This is a harsh judgment and may not apply to cocoa and coffee where Europe has
large import shares – see Tables 1 and 2. In cocoa, where the expansion of cocoa produc�on
in Côte d’Ivoire is o�en seen as the driver of deforesta�on, the EU accounted for an average
of 54.3% of Ivorian cocoa exports over the five crop year 2017/18 to 2021/22. 26 Renier et al.
(2023) report a detailed analysis of deforesta�on in Côte d’Ivoire over the period 2000-15.
The authors report that, over that period, 77% of the ini�al 6.8 Mha of tropical moist forest
was deforested or degraded. 18% of this total (1.25 Mha, approximately 83,300 ha pa) was
atributable to cocoa with the remainder to palm oil, natural rubber, food crops, logging and
mining.27 This is an es�mate of the quan�ty M-M2 in Figure 8. Cocoa yields vary enormously
across farms depending in part on farm size, rainfall and the extent of inter-cropping. Taking
a yield of 485 kg/ha 28 and supposing a constant rate of deforesta�on, I obtain an order of
magnitude es�mate that approximately 121 million metric tons (Mt) of Ivorian cocoa may
currently (2023/24) be produced on land that has suffered deforesta�on or degrada�on over
the three years since the EUDR cut-off date of end 2020. Annual Ivorian produc�on averaged
2,094 Mt over the five crop years 2016/17 to 2020/21 so this gives an order of magnitude

25
Average (mainland) China imports as shares of world produc�on 2017-21: cocoa 1.6%, coffee
0.9%, palm oil 13.7%, natural rubber 17.5%,, soybeans 27.2%. (Bean equivalents in cocoa, coffee, and
soybeans; oil equivalent in palm oil). Source: FAOSTAT.
26
Bean equivalents. Source: ICCO, Quarterly Bulletin of Cocoa Statistics, various issues.
27
The authors imply (ibid, page 7) that this may be an under-es�mate since cocoa if o�en planted
under forest cover with deforesta�on or forest degrada�on only becoming clear a�er three to five
years once the new trees become produc�ve and the cover is cleared.
28
This figure is based on a study undertaken by Barry-Callebaut and Agri-Logic that reported a range
of 350-620 kg/ha. Green, E, “Côte d’Ivoire cacao produc�on: Barry Callebaut and Agri-Logic target
yield, farm size and price”, Food Ingredients 1st (5 May 2023).
htps://www.foodingredientsfirst.com/news/cote-divoire-cacao-produc�on-barry-callebaut-and-agri-logic-
target-yield-farm-size-and-price.html

20
es�mate of 6% of the Ivorian crop (121,250 tons) that would fail to meet the EUDR import
criteria. (There will be a further quan�ty, possibly larger, for which compliance costs will
exceed the value of selling in Europe). This tonnage would need to be sold into non-
European markets. It is possible that some of all of this quan�ty of cocoa would not have
been produced if the EUDR had been effec�ve in the past, but, once planted, trees are
seldom grubbed up. The relevant issue is therefore how the Regula�on will affect future
forest clearing.

An implica�on of the EUDR is that future increases in produc�on should come by

raising yields from trees on exis�ng cul�vated land and on replan�ng on the same land when
trees reach the end of their produc�ve life. Yields of tropical commodity crops vary
substan�ally across farms in the same country resul�ng in a large gap between poten�al
yields. Research in cocoa shows that yield differences have many drivers, but the quality of
farm management and fer�lizer applica�on are among the most important (Uribe et al.,
2001; Hoffman et al., 2020). Coffee and palm oil yields can also be increased by fer�lizer
applica�on (Cannell, 1973; Sanputawong et al., 2017). However, increasing fer�lizer
applica�on is not straigh�orward. Fer�lizers need to be purchased early in the crop year and
hence financed out of the previous year’s revenue or on credit. Development economists
have stressed that, absent government schemes to provide fer�lizer, poorer farmers are
likely to be either limited by lack of collateral or will face high interest rates (Binswanger and
Sillers, 1983). In these circumstances, some farmers may find it more atrac�ve to expand
plan�ng, perhaps into currently forested area.

In cocoa, replan�ng on the same land is not always an atrac�ve proposi�on. An old
literature in economic geography sees cocoa as a “fron�er crop” (Mikesell, 1960; Hill, 1963).
Ruf et al. (2015, page 102) explain that “cocoa famers, o�en migrants, took advantage of the
‘forest rent’ of more fer�le soil” when establishing new farms on cleared land, while ageing
farms, on which this rent has been extracted, are seldom replanted with cocoa. Knudsen and
Agergaard (2015, page 325) document how “The movement of the produc�on of cocoa in
Ghana from one region to another in search of fer�le land and con�nued high outputs and
revenues has resulted in successive openings of fron�er land for cocoa produc�on”.

Ruf and Varlet (2017) suggest that the poten�al for further deforesta�on in Côte
d’Ivoire may be small as most of the previously exis�ng forest has been lost. The implica�on
is that the EUDR is likely to have only a small impact on deforesta�on even over the longer
term. The more important EUDR impact may come from discouraging deforesta�on in
countries where produc�on is increasing, in par�cular Brazil and Colombia, but also the DRC,
Guinea, Liberia, and Madagascar which are currently only minor producers. If Ruf et al.
(2015) are correct in arguing that cocoa only grows well under tropical cover, there is a
danger that the EUDR, in conjunc�on with similar schemes in other consuming countries,
may severely limit future supplies of cocoa by making plan�ng of cocoa impossible in areas
to tropical forest. From this perspec�ve, the sustainability focus in cocoa should be on

21
managed plan�ng and reforesta�on at the end of the (approximately) forty years of
produc�ve tree life�me rather than on deforesta�on per se.29

9. Concluding comments

The EUDR imposes stringent due diligence requirements on importers of a group of seven
agricultural commodi�es into the European Union with the objec�ve of reducing
deforesta�on in the producing countries. These requirements will not be limited to imports
into the EU since the majority of importers in extra-EU European countries, in par�cular
Switzerland and the UK, will also choose to comply in order to be able to export processed
products into the EU.

EUDR due diligence will require importers to maintain and provide informa�on on the
deforesta�on status on the en�re quan�ty of the commodity imported. This will o�en
require informa�on on a very large number of farms, par�cularly where the commodity is
produced by smallholders. Supply chain intermediaries had no incen�ve to collect this
informa�on prior to the introduc�on of the EUDR. Furthermore, such informa�on, if
complied, would have been lost as the commodity moved through the supply chain. The
EUDR therefore requires a complex and costly revolu�on in supply chain informa�on
keeping. It may also result in significant and costly restructuring of tropical commodity
supply chains to the likely detriment of smallholders rela�ve to larger producers, and to
na�onally-based supply chain intermediaries rela�ve interna�onal trading companies.

Governments of the commodity-producing countries see the EUDR as unwarranted,

and even imperialis�c, interven�on in their economies. These governments might be
happier with a voluntary system of controls based on cer�fica�on. The Roundtable on
Sustainable Palm Oil provides a possible model but the evidence suggests that this scheme
has had only a limited impact on deforesta�on and that cer�fica�on costs leave many
smallholder farmers outside the scheme. There is therefore a poten�al conflict between
deforesta�on, supposing that the EUDR can deliver this, and respect for the sovereignty of
producing country governments.

EUDR compliance will impose costs which will be compensated by the emergence of a
premium for compliant supplies. European consumers will be obliged to pay this higher
premium and will hence suffer a welfare loss. It is difficult to predict the likely scale of this
premium prior to the Regula�on becoming effec�ve at the start of 2025. Nevertheless, it
seems likely that the premium will be highest in the cocoa and palm oil markets. Tradi�onal
suppliers to the European markets who are able to comply with the EUDR will benefit from
the premium but will bear compliance costs. Producers that are unable to comply will need
to divert supplies to the non-European market thereby incurring diversion costs. It is not
possible to state a priori whether or not producers, as a group, will benefit or lose but any

29
Managed to ensure con�nuity of par�al tree cover. Ruf and Varlet (2017) argue for cer�fica�on of
forest, and not cocoa, trees.

22
benefit will be lower than the costs to consumers. Given this net loss, European voters and
governments must be persuaded that this will be offset by the gains arising from limi�ng
deforesta�on.

The EUDR will reduce the extent to which deforesta�on is embodied in commodi�es
and products imported into the EU, and Europe more widely but, unless similar regula�ons
are adopted by other impor�ng countries, the main impact is likely to be the diversion of
non-compliant material to non-European markets. Only in cocoa and coffee, is the European
share of global sufficiently large for there to be the poten�al of a significant impact on
deforesta�on. Any such impact will materialize over �me as new plan�ng takes place. Coffee
land can be replanted as trees age and yields decline and this may be sufficient to meet
demand growth. Replan�ng cocoa land is less successful, and it is likely that new plan�ng is
more likely to take place away from the tradi�onal producing areas.

In summary, the EUDR should result in a substan�al reduc�on in the quantum of

deforesta�on embodied in European commodity imports (“imported deforesta�on”). It will
do this at substan�al cost to European consumers without commensurate benefits to
commodity producers, in par�cular smallholder farmers. Unless other consuming countries
(par�cularly China and the United States) enact similar regula�ons, it will have litle effect on
deforesta�on in the beef, natural rubber, palm oil and soybean sectors. There may be some
reduc�on in deforesta�on resul�ng from cocoa and coffee produc�on but this will only
happen as new plan�ng takes place.

23
References

Almeida, D.V., V. Kolinjivadi, T. Ferrando, B. Roy, H.Herrera, M. Vecchione Gonçalves, and G.

Van Hecken (2023). “The ‘greening’ of empire: The European Green Deal as the EU first
agenda”. Political Geography, 105, 102925.
Araujo Bonjean, C., and J-F. Brun (2016). “Concentra�on and price transmission in the cocoa-
chocolate chain”. Chapter 17 of M.P. Squicciarini and J. Swinnen (eds.), The Economics
of Chocolate. Oxford: Oxford University Press.
Binswanger, H.P., and D.A. Sillars (1983). “Risk aversion and credit constraints in farmers'
decision-making: A reinterpreta�on”. Journal of Development Studies, 20, 5-21.
Bradford, A. (2020). The Brussels Effect. Oxford: Oxford University Press.
Brandi, C., T. Cabani, C. Hosang, S. Schirmbeck, L. Westermann, and H. Wiese (2015).
“Sustainability standards for palm oil: Challenges for smallholder cer�fica�on under
the RSPO”. Journal of Environment and Development, 24, 292–314.
Cannell, M.G.R. (1973). “Effects of irriga�on, mulch and N-fer�lizers on yield components of
arabica coffee in Kenya”. Experimental Agriculture, 9, 225-232.
Carlson, K.M., R. Heilmayra, H.K. Gibbs, P. Noojipady, D.N. Burns, D.C. Morton, N.F. Walker,
G.D. Paoli, and C. Kremen (2018). “Effect of oil palm sustainability cer�fica�on on
deforesta�on and fire in Indonesia”. Proceedings of the National Academy of Sciences,
115, 121-126.
Carodenuto, S., and M.M. Buluran (2021). “The effect of supply chain posi�on on zero-
deforesta�on commitments: evidence from the cocoa industry”. Journal of
Environmental Policy and Planning, 23, 1-16.
Catau, M.E., M.E. Marlier, and R. DeFries (2016). “Effec�veness of Roundtable on
Sustainable Palm Oil (RSPO) for reducing fires on oil palm concessions in Indonesia
from 2012 to 2015”. Environmental Research Letters, 11, 105007.
Cuny, C.J. (1993). “The role of liquidity in futures market innova�ons”. Review of Financial
Studies, 6, 57-78.
Cur�s, P.G., C.M. Slay, N.L. Harris, A. Tyukavina, and M.C. Hansen (2018). “Classifying drivers
of global forest loss” Science, 361 (6407), 1108–1111.
htps://science.sciencemag.org/content/361/6407/1108
European Union (2023). “Regula�on (EU) 2023/1115 of the European Parliament and of the
Council of 31 May 2023 on the making available on the Union market and the export
from the Union of certain commodi�es and products associated with deforesta�on
and forest degrada�on and repealing Regula�on (EU) No 9”. Official Journal of the
European Union, L.150/206.
FAO (2020). Global Forest Resources Assessment 2020. Main Report. Rome: FAO.
Fold, N. (2001). “Restructuring of the European chocolate industry and its impact on cocoa
produc�on in West Africa”. Journal of Economic Geography, 1, 405–420.
Fold, N. (2002). “Lead firms and compe��on in ‘bi-polar’ commodity chains: grinders and
branders in the global cocoa-chocolate industry”. Journal of Agrarian Change, 2, 228-
247.

24
Forwood, G., C. Connellan, J. Killick and S. Nordin (2023). “10 key things to know about the
new EU Deforesta�on Regula�on”. London: White & Case.
htps://www.whitecase.com/insight-alert/10-key-things-know-about-new-eu-deforesta�on-regula�on
Geist, H. and E. Lambin (2002). “Proximate causes and underlying driving forces of tropical
deforesta�on”. BioScience, 52, 143–150.
Gilbert, C.L. (2009). “Cocoa market liberaliza�on in retrospect”. Review of Business and
Economics, 54, 294-312.
Gilbert, C.L. (2024). “Cocoa: origin differen�als and the Living Income Differen�al”. European
Journal of Development Research. htps://doi.org/10.1057/s41287-023-00612-x
Goldman, E.D., M. Weisse, N. Harris, and M. Schneider (2020). “Es�ma�ng the role of seven
commodi�es in agriculture-linked deforesta�on: Oil palm, soy, catle, wood fiber,
cocoa, coffee, and rubber”. Washington DC: World Resources Ins�tute.
htps://www.wri.org/research/es�ma�ng-role-seven-commodi�e-sagriculture-linked-deforesta�on-oil-
palm-soy-catle
Hansen, M.C., P. V. Potapov, R. Moore, M. Hancher, S. A. Turubanova, A. Tyukavina, D. Thau,
S.V. Stehman, S. J. Goetz, T. R. Loveland, A. Kommareddy, A. Egorov, L. Chini,
C.O. Jus�ce, J.R.G. Townshend (2013). “High-resolu�on global maps of 21st-century
forest cover change”. Science, 342, 6160, 850-853.
Hill, P. (1963). The migrant cocoa-farmers of southern Ghana: a study in rural capitalism.
Cambridge: Cambridge University Press.
Hoffmann, M.P., J. Cock, M. Samson , N. Janetski , K. Janetski, R.P. Röter , M. Fisher, and T.
Oberthür (2020). “Fer�lizer management in smallholder cocoa farms of Indonesia
under variable climate and market prices”. Agricultural Systems, 178, 102759.
Hosonuma, N., M. Herold, V. De Sy, R.S. De Fries, M. Brockhaus, L. Verchot, A. Angelsen, and
E. Romijn (2012). “An assessment of deforesta�on and forest degrada�on drivers in
developing countries”. Environmental Research Letters, 7, 044009.
Hull, J.C. (2017). Fundamentals of Futures and Options Markets (8th edi�on). Boston:
Pearson.
Iswara, M.A., D.S. Nurshadrina, and A. Suryahadi (2023). “European Union palming off
deforesta�on regula�on to smallholders in Indonesia”. East Asia Forum, 10 October
2023. htps://doi.org/10.59425/eabc.1696975216
Knudsen, M.H., and J. Agergaard (2015). “Ghana’s cocoa fron�er in transi�on: The role of
migra�on and livelihood diversifica�on”. Geografiska Annaler. B, Human Geography,
97, 325-342.
Lowder, S.K., J. Skoet, and T. Raney (2016). “The number, size, and distribu�on of farms,
smallholder farms, and family farms worldwide”. World Development, 87, 16-29.
Lyons-White, J., and A.T. Knight (2018). “Palm oil supply chain complexity impedes
implementa�on of corporate no-deforesta�on commitments”. Global Environmental
Change, 50, 303-313.
Mikesell, M. W. (1960). “Compara�ve studies in fron�er history”, Annals of the Association of
American Geographers, 50, 62-74.

25
Pendrill, F., U.M. Persson, J. Godar and T. Kastner (2019). “Deforesta�on displaced: trade in
forest-risk commodi�es and the prospects for a global forest transi�on”, Environmental
Research Letters, 14, 055003.
Pendrill, F., T.A. Gardner, P. Meyfroidt, U.M. Persson, J. Adams, T. Azevedo, M.G. Bastos Lima,
M. Baumann, P.G. Cur�s, V. De Sy, R. Garret, J. Godar, E.D. Goldman, M.C. Hansen, R.
Heilmayr, M. Herold, T. Kuemmerle, M.J. Lathuillière, V. Ribeiro, A. Tyukavina, M.J.
Weisse, and C. West (2022). “Disentangling the numbers behind agriculture-driven
tropical deforesta�on”. Science, 377, 1168.
Perkins, R., and E. Neumayer (2011). “Does the ‘California effect’ operate across borders?
Trading- and inves�ng-up in automobile emission standards”. Journal of European
Public Policy, 19, 217-237.
Renier, C. , M. Vandromme, P. Meyfroidt, V. Ribeiro, N. Kalischek, and E.K.H.J zu Ermgassen
(2023). “Transparency, traceability and deforesta�on in the Ivorian cocoa supply
chain”, Environmental Research Letters, 18, 024030.
Ruf, F., G. Schroth, and K. Diffangui (2015). “Climate change, cocoa migra�ons and
deforesta�on in West Africa: What does the past tell us about the future?”
Sustainability Science, 10, 101-111.
Ruf, F., and F. Varlet (2017). “The myth of zero-deforesta�on cocoa in Côte d’Ivoire”. ETFRN
News, 58, 86-92. htps://agritrop.cirad.fr/586983/
Ruysschaert, D., and D. Salles (2014). “Towards global voluntary standards: Ques�oning the
effec�veness in ataining conserva�on goals. The case of the Roundtable on
Sustainable Palm Oil (RSPO)”. Ecological Economics, 107, 438-446.
Sanputawong, S., K. Chansathean, N. Peakchantuk, and C. Chuiruy (2017). “Study of proper
fer�lizer management on growth and yield of oil palm (Eleais guineensis Jacq.)”.
International Journal of Agricultural Technology, 13, 2631-2639
San�ka, T., K.A. Wilson, E.A. Law, F.A.V. St John, K.M. Carlson, H. Gibbs, C.L. Morgans, M.
Ancrenaz, E. Meijaard, and M.J. Struebig (2021). “Impact of palm oil sustainability
cer�fica�on on village well-being and poverty in Indonesia”. Nature Sustainability, 4,
109-119.
Staritz, C., B. Tröster, J. Grumiller and F. Maile (2022). “Price-se�ng power in global value
chains: the cases of price stabiliza�on in the cocoa sectors in Côte d’Ivoire and Ghana”.
European Journal of Development Research, 35, 840-868.
UK Parliament (2023). “Introduc�on of forest risk commodi�es regula�ons”. Statement UIN
HCWS117 (12 December 2023). htps://ques�ons-statements.parliament.uk/writen-
statements/detail/2023-12-12/hcws117
Uribe, A., H. Méndez, and J. Man�lla (2001). “Effect of balanced fer�liza�on on cocoa yield”.
Better Crops International, 15(2), 1-5. htps://cadenacacaoca.info/CDOC-
Deployment/documentos/Effect_of_Balanced_Fer�liza�on_on_cocoa_yield.pdf
van Duijn, G. (2013). “Traceability of the palm oil supply chain”. Lipid Technology, 25, 15-18.
Vogel, D. (1995). Trading Up: Consumer and Environmental Regulation in a Global Economy.
Cambridge (Mass.): Harvard University Press.

26
Weerdenburg, F. (2023). “From bean to green: EU Deforesta�on Regula�on and the soy
supply chain”. M.Sc. thesis, Wageningen University. htps://edepot.wur.nl/642972
Zhunusova, E., V. Ahimbisibwe, L.T.H Sen, A. Sadeghi, A., T. Toledo-Aceves, G. Kabwe, and S.
Günter (2022). “Poten�al impacts of the proposed EU regula�on on deforesta�on-free
supply chains on smallholders, indigenous peoples, and local communi�es in producer
countries outside the EU”. Forest Policy Economics, 143, 102817.
zu Ermgassen, E.K.H.J, M.G. Bastos Lima, H. Bellfield, A. Dontenville, T. Gardner, J. Godar, R.
Heilmayr, R. Indenbaum, T.N.P. dos Reis, V. Ribeiro, I. Abu, Z. Szantoi, and P. Meyfroidt
(2022). “Addressing indirect sourcing in zero deforesta�on commodity supply chains”.
Science Advances, 8, eabn3132.

27