This is an HTML version of an attachment to the Freedom of Information request 'EMCDDA documents about the Pompidou group report "Costs and Unintended Consequences of Drug Control Policies"'.

PUBLIC 
EXPENDITURE 
ON SUPPLY 
REDUCTION 
POLICIES
PREMS 062717
Authors: 
Anne Line Bretteville-Jensen 
Claudia Costa Storti 
Thomas Kattau 
Sania Mikulic 
Fatima Trigueiros 
Fivos Papamalis 
Laura Piscociu 
Sergey Tsarev

PUBLIC 
EXPENDITURE 
ON SUPPLY 
REDUCTION 
POLICIES
Authors: 
Anne Line Bretteville-Jensen 
Norwegian Institute of Public Health, Oslo
Claudia Costa Storti 
EMCDDA
Thomas Kattau 
Pompidou Group
Sania Mikulic 
Office for combating drugs abuse Government 
of the Republic of Croatia, Zagreb
Fatima Trigueiros 
General Directorate for Interventions  on Addictive 
Behaviours and Dependencies, Lisbon
Fivos Papamalis 
Advisor to the Greek National Drug Coordinator, Athens
Laura Piscociu 
Romanian National Anti-Drug Agency, Bucharest
Sergey Tsarev 
State Medical Institution 
“Chapaevsk Narkology Hospital”, Samara
Council of Europe

The opinions expressed in this work are the responsibility 
of the authors and do not necessarily reflect the official 
policy of the Council of Europe or the Pompidou Group. 
All requests concerning the reproduction or 
translation of all or part of this document should 
be addressed to the Directorate of Communication 
(F-67075 Strasbourg Cedex or xxxxxxxxxx@xxx.xxx). 
All other correspondence concerning this document 
should be addressed to the Pompidou Group.
Cover and layout: Documents and Publications 
Production Department (SPDP), Council of Europe
This publication has not been copy-
edited by the SPDP Editorial Unit to correct 
typographical and grammatical errors.
© Council of Europe, May 2017 
Printed at the Council of Europe

link to page 6 link to page 7 link to page 8 link to page 9 link to page 11 link to page 11 link to page 11 link to page 11 link to page 12 link to page 13 link to page 13 link to page 13 link to page 13 link to page 13 link to page 14 link to page 14 link to page 14 link to page 16 link to page 17 link to page 17 link to page 17 link to page 18 link to page 18 link to page 19 link to page 19 link to page 19 link to page 19 link to page 20 link to page 20 link to page 20 link to page 20 link to page 20 link to page 21 link to page 22 link to page 23 link to page 24 link to page 26 link to page 28 link to page 28 link to page 31 link to page 32 link to page 37 link to page 37 Contents
FOREWORD 5
EXECUTIVE SUMMARY 
6
ACKNOWLEDGEMENTS 7
INTRODUCTION 8
DEFINING CONCEPTS 
10
Public expenditure 
10
Drug-related public expenditure 
10
Public expenditure on supply reduction initiatives 
10
EMPIRICAL ESTIMATES OF DEMAND AND SUPPLY POLICY EXPENDITURE 
11
STEPS IN COST ESTIMATION AND ANALYSIS 
12
Defining the scope and objects 
12
Making an inventory of service providers 
12
Mapping financing entities 
12
Data collection 
12
Classifying and identifying data on drug-related spending 
13
Extracting expenditure data from sources: labelled and unlabelled expenditure 
13
Reporting the value of estimates 
15
EXAMPLES OF SECTORIAL MODELS 
16
Police 16
Customs 16
Court systems 
17
Prisons 17
EXAMPLES OF NATIONAL STUDIES 
18
Croatia 18
Belgium 18
Italy 18
France 19
Luxembourg 19
Russia 19
Portugal 19
Other national studies 

19
INTERNATIONAL DATABASES USED TO MODEL DRUG-RELATED PUBLIC EXPENDITURE 
20
CONCLUSIONS 21
RECOMMENDATIONS 22
GLOSSARY 
23
REFERENCES 
25
APPENDIX 1 – AVAILABLE DATABASES AND POTENTIAL INDICATORS FOR DRUG-RELATED PUBLIC 
EXPENDITURES 
27
APPENDIX 2 – THE INTERNATIONAL CLASSIFICATION OF THE FUNCTIONS OF GOVERNMENT (COFOG) 
30
APPENDIX 3 – SUMMARY TABLES: DATA FROM INTERNATIONAL DATABASES 
31
ACRONYMS 36
TABLE REFERENCES 

36
 ► Page 3


Foreword
confirmed political will to address the drugs  are available for European countries. To facilitate and 
problem in Europe lies not only in the develop-
promote future empirical expenditure studies, relevant 
ment of appropriate policies, but in the amount 
data sources and methodologies applied in empirical 
of public funds assigned to implement cost effective 
estimations are listed and discussed. 
policies. Currently, however, analysing what these 
This publication brings together the findings of wider 
funds are is still difficult. Information and data are 
study conducted by the Pompidou Group in cooper-
still sparse and national estimates tend to neither use 
ation with the EMCDDA seeking to identify the unin-
comparable definitions nor agreed methodologies. 
tended effects and associated costs of drug control 
Supply reduction is an approach used for addressing 
policies. The aim of this publication is threefold. First, 
the illicit drug phenomena. It comprises the whole 
increase international awareness about the impor-
system of laws, regulatory measures, courses of action 
tance of estimating public expenditure on supply 
and funding priorities intended to reduce the avail-
reduction initiatives. Second, stress the importance 
ability of illegal drugs. The EU Drugs Strategy (2013-
of harmonizing definitions and increasing availability, 
2020) sets the dissemination of the evaluation of 
comparability and reliability of data as well as methods 
interventions results as a priority. Providing sound 
for sound estimates. Third, contribute to developing 
methods to estimate drug-related public expenditure 
sound estimation practices to obtain accurate, com-
is an important step in this direction. Such estimates 
plete and reliable drug policy evaluations. 
aim to calculate the amount of resources spent on 
Promoting international cooperation and develop-
implementing targeted interventions and may reveal 
ing of effective working partnerships between drug 
the extent to which policy intentions are reflected in 
international organizations, policy makers, specialists 
relevant budgets.
in accountancy, law enforcement agents and those 
This publication is a first step towards a systematic 
in charge of economic modelling is the way forward. 
analysis. It examines a set of recent and representative 
While recognising the limitations imposed by currently 
attempts to estimate public expenditure on supply 
available data sets, this publication sheds light on 
reduction policies. Consequently, it proposes a com-
current practice and, in doing so, suggests areas of 
mon set of definitions aiming to establish a common 
focus for future desired methodological development. 
basis for understanding such complex subject matter 
In this way it hopes that the estimation of drug-related 
and to facilitate comparability in three main dimen-
public expenditure and policy evaluation will move 
sions: time, policy and countries. Although the study 
forward, in Europe. For continuous improvements to 
is mainly focused in supply reduction expenditures, 
take place, however, it is essential that partnerships are 
it reports data on the balance between spending on 
extended and maintained with the goal of developing 
demand and on supply reduction, when estimates 
good practices, standards and guidelines in this field.
Jan Malinowski 
Alexis Goosdeel 
Executive Secretary of the Pompidou Group
Director of the EMCDDA
 ► Page 5

Executive summary
Supply reduction has been normally used for  This report takes a first step towards a systematic 
addressing the illicit drug phenomena in Europe. 
analysis, by examining a set of representative attempts 
It uses the whole system of laws, regulatory mea-
to estimate public expenditure on supply reduction 
sures, courses of action and funding priorities used 
interventions. It proposes a common set of defini-
by governments and their representatives. 
tions, aiming to establish a common basis for under-
Evaluating drug policy is an integral part of a cost-ef-
standing this topic and facilitating comparability in 
ficient approach to tackle illicit drugs. Assessing and 
three main dimensions: time, policy and countries. 
estimating drug-related public expenditure is a first 
Although it is mainly confined to supply reduction 
step in evaluations exercises. Estimates aim to cal-
expenditures, in order to set the context, it describes 
culate the amount of resources spent, or needed, to 
the proportion that total drug-related expenditure 
implement these targeted interventions. Therefore, 
represents of national public spending and; presents 
estimates may reveal to what extent policy intentions 
the balance between demand and supply reduction 
are reflected in relevant budgets, if considered that 
spending for a number of European countries. With 
the size of the phenomena and resources available 
the aim of facilitating and promoting future empirical 
condition choices.
expenditure studies and of setting the ground for the 
development of good practices, relevant data sources 
Until now, estimates for the funds spent by govern-
and methodologies applied are listed and discussed 
ments in this field are sparse. They have been mostly 
and examples of sectorial models of public spending 
produced at national level and applied different defi-
are selectively provided. Finally, some conclusions and 
nitions, with no commonly agreed methodologies or 
recommendations are offered. 
comparable datasets. Uncertainty about the most 
appropriate economic models to use also exists. These 
factors have constituted effective barriers to rapid 
developments of policy evaluation and cost-effective 
analysis in the field.
► Page 6

Acknowledgements
The authors would like to thank the EMCDDA and the Pompidou Group Secretariat for the support ren-
dered in the preparation of this publication. In addition gratitude is expressed to the following experts 
who contributed to the content of this publication in the context of their work in the Pompidou Group’s 
expert group on effects and associated costs of drug control policies:
Peyman Altan
Turkish Public Health Institution, Ankara, Turkey
Pavel Bem 
Charles University, Prague, Czech Republic
Torbjørn Brekke
Senior Adviser, Ministry of Health and Care Services, Norway
Yossi Harel-Fisch 
Bar-Ilan University, Ramat Gan, Israel
Elena Hedoux
Pompidou Group, Council of Europe
Janusz Sieroslawski 
Institute of Psychiatry & Neurology, Warsaw 
 ► Page 7

Introduction
The aim of this publication is threefold. Firstly, to  the whole system of laws, regulatory measures, courses 
increase international awareness concerning the 
of action and funding priorities concerning illicit drugs 
importance of estimating public expenditure on 
put into effect by a government or its representatives 
supply reduction initiatives. Secondly, to raise public 
(law enforcement officers such as police and customs 
awareness of the need to agree upon harmonising 
officers, judges, prison guards, etc.). Reduced drug 
definitions and increasing the availability, compa-
availability and accessibility is achieved through a 
rability and reliability of data, as well as methods for 
disruption of illicit drug trafficking; dismantling of 
producing sound estimates. And thirdly, to contribute 
the criminal organisations that are involved in drug 
to developing national and international estimation 
production and trafficking; efficient use of the criminal 
practices with a view to obtaining accurate, complete, 
justice system; effective intelligence-led law enforce-
reliable and comparable drug policy evaluations.
ment and increased intelligence sharing; and a focus 
The target audience includes officials involved in the 
on large-scale, cross-border and organised drug-re-
evaluation of drug policy; entities wishing to evaluate 
lated crime (EMCDDA, 2016). 
drug policy priorities, develop drug policy strategies 
and action plans and analyse their economic, social 
As stated in documents such as the 2013-2020 
and political consequences; accounting authorities; 
European Union Drug Strategy (Council of the 
entities seeking funds to finance their service provi-
European Union, 2013) and the EU Action Plan on 
sion; and researchers.
Drugs 2013-2016 (Official Journal of the European 
Union, 2013), an evaluation of drug policy is an inte-
Most European countries have a national drug policy 
gral part of the approach to combating illicit drugs. 
presented in a drug strategy document (EMCDDA, 
Estimation of drug-related public expenditure can be 
2015). National drug strategies tend to reflect a bal-
seen as a first step in this direction. Public expenditure 
anced approach between drug demand and drug 
estimates aim to calculate the amount of resources 
supply reduction (EMCDDA, 2016). An optimal balance, 
spent, or needed, to implement targeted interventions 
however, may not imply that the two approaches 
in a particular policy field and may reveal to what 
receive an equal share of resources and attention. 
extent policy intentions are reflected in the relevant 
Instead, it will depend on country specific priorities 
budgets and are conditioned by the size and charac-
and aims for the different drug policy sectors, as well 
teristics of the drug phenomenon. 
as on the relative price of implementing each activity 
in a cost-effective manner. 
Accurate estimates of public spending on implement-
Supply reduction is often the main approach used 
ing drug policy initiatives will help policymakers to 
for addressing the illicit drug problem. Nonetheless, 
plan relevant interventions and make the required 
efforts aimed at reducing demand (mainly prevention 
funds available to the authorities in charge of policy 
and treatment measures) are also important and 
implementation. A thorough assessment of drug 
harm reduction initiatives have gained in significance 
policy expenditures will also contribute to improved 
over the years. In addition, the effects of supply and 
transparency and accountability of public institutions. 
demand reduction efforts are often interrelated. For 
Estimates may provide information on factors such 
instance, successfully reducing drug availability may 
as the relative importance of demand and supply 
influence the consumption of drugs and also have an 
expenditures and enable cross-country comparisons 
impact on problem drug use and adverse drug use 
of the level and composition of spending on the fight 
consequences. 
against illicit drugs (EMCDDA, 2008). Sound planning, 
improved knowledge of the resources allocated to this 
The overarching objective of supply reduction is a 
policy field, and cost-effective resource allocation are 
measurable reduction in the availability and accessibil-
particularly necessary in times of economic downturn 
ity to illicit drugs. Supply reduction initiatives comprise 
when fewer resources are available. 
► Page 8

A subsequent step would be to systematically compare 
useful and desirable. This report takes the first step 
public expenditure and other possible costs to the pol-
towards a systematic analysis by examining a num-
icy’s measured outputs or results. Depending on how 
ber of representative attempts to estimate public 
the results are defined and measured, a cost-benefit or 
expenditure on supply reduction policies. It proposes 
cost-effectiveness analysis can be conducted (see glos-
a common set of definitions to be used for public 
sary). In this case, resource inputs (the costs of labour, 
expenditure assessment and evaluation. In addition, 
capital and/or equipment) are linked to intermediate 
it aims to establish a common basis for understanding 
outcomes (e.g. number of drug dealers arrested); final 
this complex subject and to facilitate comparabil-
outputs (e.g. lives saved, life years gained, number of 
ity in three main areas: time, policy and countries 
drug users, reduction in drug-related harm, percent-
concerned. Although, the report mainly focuses on 
age reduction in crimes committed); or policy goals. 
supply reduction expenditures, in order to contex-
Irrespective of the chosen output measures, however, 
tualise them, it describes the proportion that total 
public expenditure will be a central cost factor, since 
drug-related expenditure represents of gross domestic 
governments constitute the main provider of drug 
product. It also shows how spending is balanced 
supply reduction services in Europe.
between demand and supply reduction initiatives 
A thorough economic evaluation can provide pol-
in a number of European countries. To facilitate and 
icymakers with the information required to make 
promote future empirical expenditure studies, the 
well-informed decisions. Although the data and a 
relevant data sources and methodologies applied in 
quantification of all the outcomes and cost elements 
making empirical estimates are listed and discussed. 
required for conducting the most comprehensive 
Examples of sectorial models of public spending and 
analyses are currently not available, a somewhat less 
examples of national supply reduction expenditure 
extensive analysis and an improved understanding 
studies are also provided. Finally, some conclusions 
of the individual elements involved are still possible, 
and recommendations are offered. 
Introduction ► Page 9

Defining concepts
Public expenditure
government documents or single budget lines; 
these are labelled expenditure. The required data 
The term «public expenditure» refers to the value of 
are instead embedded in budgets for larger sectors 
goods and services purchased by  general govern-
or programmes (unlabelled expenditure), which 
ments (at central, regional and local level) in order to 
means that modelling and calculations are needed. 
perform its functions. For instance, it refers to resources 
For instance, it is common that prisons do not have a 
spent on healthcare, justice, public order and safety, 
separate budget for drug-law offenders, because they 
education, social protection and so on (Eurostat, 2011), 
usually have one single budget for their entire activity. 
and its quantification is a costing exercise undertaken 
Therefore, the values of this embedded expenditure 
from the government’s perspective (EMCDDA, 2008). 
can only be estimated through modelling approaches 
The role of private expenditure in drug policy varies 
(EMCDDA, 2014). This requires skills, modelling tools 
across countries, timescales and policy areas. In many 
and techniques. 
countries, drug treatment is partly financed by the 
private sector (insurance companies, drug users or 
Despite the various factors which may challenge the 
their employers, relatives, etc.). In other drug policy 
robustness of estimation results (limited data availabil-
areas, such as supply reduction, private funding usu-
ity, layering of assumptions, changes in definitions or 
ally constitutes a negligible share of total spending 
regulations over time, etc.), the application of existing 
(European Commission, 2012). 
models can provide useful insights, as various coun-
tries’ experience shows (see the examples below). 
Drug-related public expenditure
Drug-related public expenditure is the sum spent by 
Public expenditure on supply 
governments on goods and services with the aim 
reduction initiatives
of tackling the illegal drug phenomenon. Although 
drug policy expenditure estimates are deemed useful, 
In this report, public expenditure on drug supply 
most countries do not produce separate drug-related 
reduction comprises the funds spent by general gov-
budgets as part of their ordinary budgeting exercise. 
ernment with the broad purpose of reducing the 
Relevant analyses and estimations can be complicated 
availability with the support of the police, law courts 
since several inter-ministerial and cross-governmental 
and prison services geared towards combating the 
sectors are involved in drug control programmes, 
illegal drug phenomenon, as defined by Eurostat 
including justice, policing and border control, prisons, 
(2011). In general, police services comprise, among 
social protection, education and health. Disentangling 
others, the regular and auxiliary policing of ports 
drug policy expenditure across government depart-
and borders, coast guards and customs, as well as 
ments and inter-sectorial policies remains a significant 
road traffic regulations and supervision. The services 
challenge. Changes in legislation and the structure of 
provided by law courts comprise the operation or 
public administration can further hamper compara-
support of civil and criminal law courts and judicial 
bility over time.
systems, the prosecution service, fine enforcement 
An additional challenge lies in the fact that drug-re-
and probation systems. Prison services comprise the 
lated programmes and activities can be found at many 
activities of prison administration and the operation or 
different levels of public administration. For instance, 
support of prisons and other places for the detention 
the funding for imprisoning drug-law offenders is usu-
or rehabilitation of offenders, such as prison farms, 
ally provided by central government, while prevention 
workhouses, reformatories, borstals, forensic wards, 
of street dealing or social reintegration programmes 
etc. (Eurostat, 2011). 
for former drug dealers are frequently financed by 
local authorities. This makes it necessary to compile 
In the case of public expenditure on drug supply 
data at different administrative levels, which can be 
reduction initiatives, the vast majority of resources 
a demanding task. 
will be spent on enforcement targeting producers and 
dealers, but may also include legal action targeting 
In addition, often only a small fraction of drug- related 
users for drug possession when required by national 
public expenditure can be traced back directly to 
judicial systems.
► Page 10


Empirical estimates of demand 
and supply policy expenditure
Over the last decade at least 16 European coun- Interestingly, however, the information available sug-
tries have provided comprehensive estimates 
gests that supply reduction activities accounted for 
of drug-related public expenditure (EMCDDA, 
the largest share of drug-related public expenditure 
2014b). Country estimates suggest that drug-related 
in most countries. Of the 16 countries which pro-
expenditure ranged from 0.01% to 0.5% of gross 
duced complete estimates in the last decade, only 
domestic product (GDP). Since the studies may not 
four countries spent less than 50% of their total drug 
have applied the same expenditure classifications 
budget on supply reduction, while five countries spent 
or the same estimation methods, caution is required 
70% or more. The other countries spent between 
when making cross-country comparisons (EMCDDA, 
50% and 70% of their drug-related expenditure on 
2014b). 
supply reduction. 
Figure 1 Breakdown of drug-related expenditure between demand and supply reduction. 
Source: EMCDDA, 2014b
Analysis has also shown that funds allocated to 
classified under health and social protection, this 
drug-related initiatives account for only a small pro-
may further suggest that European countries give 
portion of the overall public expenditure on the public 
higher political priority to supply reduction initiatives, 
order and safety sector. For instance, in 2008 (the only 
as part of public order and safety activities, than to 
year this exercise was systematically conducted in 
demand reduction initiatives as part of overall public 
European Union countries), supply reduction expendi-
health activities (EMCDDA, 2008). Annually, EMCDDA 
ture represented between 2% and 12% of total public 
reports the most recent estimates available for national 
expenditure in this sector. This proportion compares 
drug-related public expenditure in percentage of 
to the proportion of drug-related spending on the 
the gross domestic product (GDP), in the European 
health and social protection sectors. The proportion 
Union countries, Norway and Turkey. When available, 
of drug-related expenditure on these items accounted 
EMCDDA reports also the proportion of funds spent 
for less than 1% of total public spending on health 
on supply reduction initiatives (http://www.emcdda.
and social protection during that period. Since most 
europa.eu/countries). 
public spending on demand reduction initiatives is 
 ► Page 11

Steps in cost estimation 
and analysis
Clarifying definitions, improving estimation  services are not always obvious and easy to identify. 
methods, agreeing on best practices and find-
For instance, when drug treatment services are pro-
ing reliable, standardised data will enhance the 
vided within prisons, the entity in charge has public 
utility of public expenditure estimates, as analysis 
order and safety as its first function but health as its 
over time and across policy areas and countries can 
“real” goal. Therefore, analysts must consider whether 
be improved(Single, 2009). Better quality data and 
to include the costs of these activities as supply reduc-
further methodological developments are needed. To 
tion or demand reduction initiatives. Eurostat, along 
this end, we list below some recommended, general 
with most international organisations concerned with 
methodological steps in cost estimation and analysis. 
policy evaluation, includes the provision of services 
in the main function that the funds are used for, even 
Defining the scope and objects
where the provider is less obvious. In this case, public 
expenditure on drug treatment provided in prisons 
Globally speaking, a first step for a viable estimate is 
should be excluded from expenditure estimates for 
defining the scope and type of public expenditure 
supply reduction services and accounted for as drug-
considered. In addition, clear indications of the geo-
related health expenditure. Sometimes, provision will 
graphical area and which function of public service 
be the responsibility of private entities while financing 
provision the estimates cover are needed. 
is a government responsibility. 
It should be noted, however, that the same service 
Making an inventory 
may have multiple policy purposes and double count-
of service providers
ing should be avoided. For instance, in the case of 
social reintegration programmes in deprived neigh-
Secondly, it is necessary to identify the public entity 
bourhoods, financing may serve both the purpose 
or institutions responsible for the provision of drug-
of preventing drug crime (and should be added to 
related services – in the case of this report supply 
supply reduction expenditure) and the purpose of 
reduction measures and interventions. The govern-
preventing drug use (and should also be accounted 
ment authorities and public institutions and services 
for as health spending in demand reduction expendi-
responsible for the implementation of the drug policy 
ture). For public accounting purposes the same funds 
initiatives, on the different competency levels, have 
should not be counted twice. Therefore, researchers 
to be made an inventory. 
will have to include this spending only once, choosing 
to record it under either preventive health or crime 
prevention. Sometimes, making a decision is difficult 
Mapping financing entities
and the best way to deal with such situations is to 
guarantee that researchers document the different 
The third step is then to identify who finances these 
choices and assumptions they make.
service providers. The starting point for a public expen-
diture analysis is accordingly the different public 
authorities which fund the respective aspects of the 
Data collection
drug policy. Irrespective of the governmental struc-
ture, expenditure by all relevant national, regional 
The fourth step is to determine a strategy for collecting 
or local government institutions, directly or indi-
the required data on public expenditure. In order to 
rectly associated with drug policy, should always be 
obtain the relevant information, analysts will have to 
included.
examine policy documents and accounting data. It is 
also recommended that interviews be conducted with 
Matching stakeholders responsible for providing drug 
the major stakeholders in the field, as a way to obtain 
policy services with their financing entities can be chal-
better information about where financial data might 
lenging, as the entities in charge of providing public 
be available, and to search for international data sets. 
► Page 12

Classifying and identifying data 
public expenditure spent with the main aim of tackling 
on drug-related spending
the drug phenomenon. However, no systematic data 
collection has taken place. 
It is essential to classify public expenditure according 
to the purpose for which the expenditure is intended 
The research community has not formally adopted any 
(Reuter et al., 2004, and Eurostat, 2011), so the next 
of these classification systems. However, as Eurostat 
step to consider is how to group drug-related spending 
publishes data annually in accordance with the COFOG 
according to these sub-purposes. Taking into account 
classification, their system is frequently used. Eurostat 
the fact that drug-related expenditure on supply reduc-
publishes data on public spending with the purpose 
tion initiatives comprises funds spent with the aim of 
of guaranteeing public order and safety, which is split 
combating the illegal drug phenomenon through the 
into the above-mentioned classes. Researchers still 
police, law courts and prison service, the classification 
have to opt for criteria and models to disentangle 
commonly used in international comparisons is the 
drug-related spending within these overall expen-
Classification of the Functions of Government.1
diture classes.
 
► The Classification of the Functions of  In fact, supply reduction initiatives are often embedded 
Government (COFOG) provides a useful frame-
in policy projects that have broader objectives and bud-
work for classifying public spending according 
gets. Therefore, firstly, it is important to look beyond 
to its purpose. Under COFOG, most drug control 
expenditure that is exclusively used for drug policy 
policy expenditure is included in the “public 
and also include spending intended for broader policy 
order and safety” class of expenditure. The most 
domains that indirectly, but significantly, contribute to 
directly relevant subclasses are “police services”, 
drug policy or impact upon it. For instance, investing 
“law courts”, “prisons” and “R&D public order and 
in effective policing in certain problematic neighbour-
safety” (Eurostat, 2011). 
hoods, in order to prevent all types of crime, may also 
contribute to preventing drug dealing. Consequently, 
 
► Reuter (2006) relates public expenditure to the 
supply and demand sides of the market. He 
it is relevant to take into account overall budgets for 
counts public spending on supply reduction 
initiatives which may have direct synergies with drug 
under “enforcement programmes” and con-
policy objectives. Secondly, modelling techniques are 
siders that these are “programmes aimed at 
required in order to disentangle drug-related expendi-
traffickers and producers to shift up the supply 
tures from overall expenditures. For instance, specific 
curve for drugs; other things being equal, they 
estimates and well-defined methodologies are needed 
should raise the price of drugs and lower quantity. 
to disentangle expenditure on drug-related crime from 
Programmes aimed at users and retailers raise 
overall public spending on law courts (more details on 
the transaction costs of buying drugs”. In other 
methodologies are given below). 
words, enforcement programmes will make 
In the event that not all the required data are available 
drug producing, trafficking or dealing more 
in international data sets, national databases should 
expensive, because they either bring about an 
be mapped. Every country has different structures 
increase in the unitary costs of production or 
for drug control services, provision and financing. 
introduce greater risk into the business (Costa 
National data mapping can be achieved in different 
Storti and De Grauwe, 2009).
ways: information from registration systems, annual 
These two classification systems are substantially 
reports, interviews with key experts and/or contacts 
different. COFOG has been co-designed by the statis-
working in this field (De Ruyver et al., 2007). Detailed 
tical office of the European Union and the European 
mapping of available data can be demanding and 
Commission, with well-defined concepts and data 
makes intensive use of resources. However, it is a 
collection methodologies. Annual mandatory data 
fundamental step for any estimate of public spending 
collection has been implemented in every European 
on drugs control.
Union member state since early 2000. The system 
covers all functions provided and financed by govern-
Extracting expenditure data 
ments. Drug-related activities are among the overall 
from sources: labelled and 
tasks provided and financed by the public sector, 
unlabelled expenditure
but there are no specific methods specified or data 
collected on drug-related expenditure. Drug-related 
Some of the funds allocated by governments for 
expenditure is embedded in broader items, such as 
drug-related expenditure are identified as such in the 
public expenditure on public order and safety, security, 
budget (labelled expenditure). Often, however, the 
health, education or social protection. Conversely, 
majority of drug-related expenditure is not identified 
the Reuter’s classification was designed to organize 
(unlabelled expenditure) and must be estimated using 
modelling approaches. Total drug-related expenditure 
1.  National estimates sometimes use alternative definitions. 
is the sum of labelled and unlabelled drug-related 
See (Lievens et al., 2016) or (Kopp, 2006) for further details.
expenditures (EMCDDA, 2016).
Steps in cost estimation and analysis ► Page 13

Since labelled expenditures are clearly identified in 
For instance, the number of drug-law offenders in 
budgets, calculation methods are not required. Time 
prison may allow estimating the proportion that con-
series data are often available for labelled expenditure. 
victed prisoners for drug-law offences represent from 
The biggest challenge when data on labelled expen-
total prison population, and therefore to approach the 
diture are compiled is to ensure complete mapping 
proportion that drug-related expenditure on prisons 
of all entities in charge of providing these services, as 
represents from total prison spending; or the propor-
they can be spread across different government levels. 
tion that drug-related cases handled by the police, 
Depending on the national structures, expenditures 
by prosecutors or by drug-law courts on the total 
from all relevant national, regional or local government 
number of cases handled by these institutions may 
institutions that are directly or indirectly associated 
allow starting approaching their drug-related costs.
with drug policy should always be included.
To design attributable fractions, models use the sup-
For unlabelled expenditure, a modelling procedure 
port of data on crime, police, law courts or prisons 
is necessary to estimate these different expenditures 
activity. Appendix 3 presents information and data 
and the modelling is based on either a top-down or 
by groups of variables. These groups encompass total 
a bottom-up approach. Frequently, these estimates 
public expenditure, drug-related public expendi-
require the use of activity data to develop estimates 
ture, supply reduction public expenditure; drug law 
(for example, number of offences, offenders, criminal 
offences; crime reported by the police, drug-related 
cases, prisoners, etc.)
crime, conviction statistics and prison population. 
Within groups, variables directly relevant are listed. 
Modelling unlabelled expenditure
For each variable, data available are listed by source, 
country and time period. Finally, this annex reports 
The top-down modelling approach is mainly used when 
the number of observations available for each vari-
the data available are embedded in programmes with 
able. The relevant sources include data from the 
broader goals and the fraction attributable to drugs can 
Council of Europe, EMCDDA, EUROSTAT, Univeristeé 
be identified as the proportion of the overall budget. 
de Criminoligie et de Droit Penal de Lausanne and 
In order to identify this proportion, models lay down 
the UNODC. 
objective criteria and calculate attributable fractions. 
Despite that data available are still referent to a short 
Unlabelled drug-related expenditure = Overall expenditure × Attributable fraction
period of time and that data are still missing in many 
countries/years, gathering available information shall 
There is no general methodology to determine attrib-
allow developing better methods and more accurate 
utable fractions also known as repartition keys. In 
estimates in the future. 
practice, the appropriate repartition key is determined 
by the object of the estimate, data availability and the 
When international sources are not available, publicly 
modelling approaches available. Repartition keys are 
available national statistics and data from competent 
determined in different ways on the basis of informa-
public bodies should be used.
tion from activity data, extracted from registration 
Advantages of the top-down approach
systems, annual reports and/or contacts working in 
this field (De Ruyver et al., 2007). When determining 
 
► Availability of data: the availability of aggre-
attributable fractions, the data used should prefer-
gated budgetary data means that top-down 
ably be publicly available or, even better, be stored 
approaches can be easily applied.
within international databases. This can guarantee 
 
► Low cost: the availability of aggregate cost data 
the possibility of producing similar estimates in the 
means that the time and costs required to esti-
years that follow and in other countries.
mate a top-down unit cost can be reduced.
Appendix 3 summarizes the information and data 
 
► Versatilitythe methodology enables an analyst 
available in the most relevant international data-
to forecast how costs may change as a result 
bases that can be used to estimate unlabelled public 
of a reduction/an increase in service usage 
expenditure on supply reduction. It describes the 
(for instance, when there are less/more drug- 
activity data reported, the reporting countries and 
related crimes committed in a certain year than 
time periods. 
expected) and how these costs change over 
time.
This Appendix reports the data available concern-
ing the annual statistics on national public expen-
There are, however, some limitations associated with a 
diture on police, law courts and prisons reported by 
top-down approach. Firstly, it does not clearly identify 
Eurostat. These data include not only expenditure on 
the different factors that may drive the costs and there-
drug-related initiatives, but the total spent to tackle all 
fore often masks the underlying factors that determine 
types of crime. Therefore, to disentangle drug-related 
why unit costs vary within a single, yet heterogeneous, 
expenditure and built attributable fractions, activity 
services group. The criteria laid down for estimating 
data shall be required. 
attributable fractions do not always take into account 
Page 14 ► Public expenditure on supply reduction policies

all of the characteristics that may impact the total costs, 
 
► Versatility: the methodology enables an analyst 
which means that cost functions are often simplified. 
to forecast how costs may change as a result of 
These estimates are therefore often not very precise. 
a reduction in service usage or demand.
Nevertheless, they are frequently used and provide 
valuable proxy indicators for average costs. 
However, the main disadvantage associated with the 
bottom-up approach is that it requires detailed infor-
An alternative method of estimating drug-related 
mation concerning both the type of costs associated 
expenditure is to base estimates on the cost of provid-
with the provision of each service (full knowledge of 
ing one unit of public service, known as the bottom-
the production function of each public service) and 
up modelling approach. This modelling approach 
the unit cost of each of the production factors. 
starts by detailing how much it costs to provide one 
unit of service or intervention. For instance, how 
A combination of the two approaches may be pre-
much does it cost to keep one drug-law offender in 
ferred. The advantage of this dual method is that 
prison? Considering the different costs borne by the 
it makes cross-verification possible; the data gath-
government for managing a prison facility, such as 
ered on the basis of the top-down approach can be 
the real costs of state property, prison staff, electric-
double-checked and supplemented with the data 
ity, water and gas, machinery, etc., it is possible to 
retrieved from project actors in the field.
estimate how much each detainee costs per day. This 
sum can then be multiplied by the number of drug-
related detainees, taking into account the different 
Reporting the value of estimates
costs associated with each type of detainee, based 
on the different lengths of prison sentences, different 
The basic format used to report the value of estimates 
security levels, etc. To obtain the total expenditure on 
is monetary value in nominal terms. However, to per-
drug control policy, all the cost elements should be 
mit comparability over time, if reported in monetary 
identified and totalised. 
units estimates should be adjusted for inflation. 
The bottom-up approach is particularly appealing 
In addition, some authors report the value as a per-
when relevant unit costs are readily available. If, on 
centage of GDP. This way of presenting the results 
the other hand, every type and element of the drug 
considers the economic dimension of a country. It is 
policy has to be separately estimated, the approach 
likely that drug-related spending is higher in a country 
can be demanding and challenging. 
with 85 million inhabitants than in a country with 
Advantages of using a bottom-up approach
10 million inhabitants. The same holds for a higher 
 
► Transparency: detailed cost data allow potential 
income country (EMCDDA, 2008). For these reasons, 
errors to be investigated and their impact tested 
reporting the value of estimates as a percentage 
– this facilitates a quality assurance process.
of GDP is a valid choice, since it takes account of 
both the inflation problem and the size and level of 
 
► Simplicity: the calculation required to estimate 
a country’s income.
unit costs is easy to understand and direct, 
providing a simple way to quantify the admin-
Another frequently used approach is reporting the 
istrative and overhead costs associated with a 
value of spending per number of problem drug users. 
range of public services.
In this case, authors take into account the dimension 
 
► Detail: detailed cost data can highlight vari-
of the drug problem. Reporting all these complemen-
ations, enable analysts to explore the factors 
tary measurements of drug-related public spending 
underlying variations and determine whether, 
facilitates the validation of the data through cross-ver-
for example, some service users account for a 
ification and increases the economic significance and 
disproportionate share of the costs.
utility of the estimates.
Steps in cost estimation and analysis ► Page 15

Examples of sectorial models 
In addition to collecting labelled public expenditure  trafficking and dealing in illicit drugs; and driving 
data, several examples exist of models applied to 
under the influence of drugs and alcohol. The pro-
identify unlabelled expenditure on drug control in 
portion that this time represented of the total work-
the national contexts. Different authors have applied 
ing time for the police forces was then used as an 
different definitions, data sets and models to estimate 
«attributable fraction» for disentangling the amount of 
items of drug-related expenditure. In this section, 
money that was spent on drug-related police activities 
examples of the definitions, data and models are 
from the total spending on police activity. Within the 
provided. The section aims to present the models 
funds allocated for supply reduction, 14% was spent 
utilised to estimate unlabelled drug-related spending 
on drug-police activity, while law courts and prisons 
on various types of supply control initiatives. 
absorbed the remaining 21% and 65% respectively.
Moolenaar (2009) developed a model and provided an 
Police
example of how to estimate public spending on supply 
reduction initiatives in the Netherlands. The author 
Public spending on drug-related police services is 
applied a top-down model based on the average cost 
probably best identified using a top-down approach.2 
of police time spent on this work. Moolenaar calcu-
In order to disentangle this expenditure from total 
lated the average duration of each type of criminal 
public expenditure on public order and safety, as 
investigation firstly by type of criminal activity (assum-
published by Eurostat, attributable fractions has been 
ing that different criminal activities have different 
calculated with the help of activity data. Authors 
investigation costs – based on an assessment of the 
have used auxiliary data to create these fractions, for 
severity of the crime) and secondly by the number of 
instance data on drug-related offences in proportion 
cases registered for each criminal activity.
to the total number of offences. The following are 
concrete examples of variables available in national 
Customs
and international data sets, which have all been used 
separately to estimate attributable fractions:
With regard to customs services, the share of customs 
(1)  The number of drug-related crimes per 100 000 
officers who deal with drug control activities and/or 
population.
the proportion of their working time compared to 
the total number of custom officers has been used 
(2) The number of drug-related cases reported 
as an attributable fraction. As input data, the number 
by the police out of the total number of police 
of customs officers who are involved in drug control 
cases.
activities forms the basis for the calculation. These 
(3)  The time the police forces spend on combating 
estimates are then applied to the total expenses of the 
the drug phenomenon in proportion to their 
customs administration (minus any labelled expen-
total working time.
diture specifically targeted towards this activity). It 
should, however, be noted that most customs officers 
To estimate the share of costs attributable to spend-
do not exclusively devote their working time to drug 
ing on police action against illicit drugs, the ratio is 
control activities, so, ideally, the percentage, or the 
multiplied by the total expenditure of the law enforce-
average, of working time devoted to drug control 
ment agencies and reduced by any available data on 
should be estimated.
labelled expenditure for drug control. 
Kopp and Fenoglio (2002) estimated the drug-related 
A concrete example is provided by the estimates for 
expenditure of customs services based on the propor-
Italy. Genetti (2014) estimated drug-related public 
tion of customs officers allocated to combating illicit 
expenditure for police forces based on the amount 
drug trafficking within the total number of customs 
of time that staff spent on drug control in 2011: pos-
officers. This proportion constituted the attributable 
session of illicit drugs for personal use; production, 
fraction applied to the total customs budget. The 
authors concluded that, in 2000, drug-related spend-
2.  Although it is also possible to use a bottom-up approach, 
ing on customs services represented approximately 
since police activity is normally financed by the central 
government budget, a pragmatic approach frequently used 
10% of total drug-related spending in France. As these 
is to prepare estimates based on these aggregated budgets. 
authors pointed out, omitting costs such as those of 
In this case, estimates for public spending are relatively 
detection equipment or detection dogs may consti-
complete, considering all relevant costs. Additionally, this 
tute a relevant limitation, since the costs of detection 
method facilitates the international comparability of results, 
since comparable data are available for most European 
equipment may have a strong impact on relatively 
countries.
small budgets such as that for customs.
► Page 16

Lievens et al. (2016) estimated drug-related expendi-
supreme court) was estimated based on a bottom-up 
ture by customs based on the proportion that drug-
approach, which combined the number of cases and 
law violations represented in the total number of 
the average cost per case (Ramstedt, 2006). The data 
violations registered by the ordinary customs services, 
were obtained from a judicial system official. It should 
investigation services and motorised brigades. They 
be noted that the average case cost was not recorded 
used a top-down approach based on the number of 
by type of crime, instead the average for all types 
drug-law offences in proportion to the total number 
of crime was used as an indicator for drug crimes. 
of offences. In 2012, customs spending represented 
Moreover, for the court of appeal and supreme court, 
3.6% of the total drug-related public spending on 
only the total number of criminal cases was available 
supply reduction in Belgium.
and the fraction of drug cases was estimated based 
on the situation in the district courts (9%). Regarding 
Court systems
the range of the estimates it should be noted that 
the author included, as an upper limit for estimates, 
Spending on drug-related court services has been 
a specific percentage (30%) of the costs of tackling 
extracted from total national expenditure on law 
other crimes, as they may have been committed under 
courts based on the following activity data: 
the influence of drugs. 
(1) The proportion of drug-related offences with 
regard to the total number of offences.
Prisons
(2)  The proportion of drug-related convictions with 
regard to the total number of convictions;
Unlabelled costs of drug-law offenders in the prison 
(3) The proportion of people imprisoned for 
system can be estimated using the number of con-
drug-related offences with regard to the total 
victed prisoners for drug-related offences expressed 
number of prisoners.
as a proportion of the number of overall convictions. 
For example, to estimate expenditure related to drug-
Kopp and Fenoglio (2002) estimated the expenditure 
law offences in prisons, two elements have been taken 
that drug-related crime represented in the French 
into account: overall prison expenditure for a given 
judicial system. They adopted a bottom-up approach, 
fiscal year and the attributable fraction of prisoners 
taking estimates of the time spent by the various 
convicted of drug-law offences. 
types of French judges and other types of adminis-
trative staff on drug-law cases and then multiplying 
EMCDDA (2014) provides an example of how public 
these estimates by their average salaries. Based on 
expenditure on drug-law offenders in prisons can be 
this method, the authors concluded that law courts 
estimated. Based on data for public expenditure on 
represented about 24.4% of total drug-related public 
prisons provided by Eurostat and data on the number 
expenditure in France in 2000.
of offenders provided by the Council of Europe, the 
proportion of prisoners sentenced for a drug-law 
In Croatia, drug-related spending on the courts 
offence as their main offence was applied to the total 
covered drug-related cases prosecuted by both the 
public expenditure on prisons. A range of estimates 
State and the courts (Budak et al., 2013). A top-down 
was calculated, with low estimates taking into con-
approach was used based on estimates of the num-
sideration only prisoners sentenced for a drug-law 
ber of drug-related crimes as a proportion of the 
offence and high estimates also including pre-trial 
total number of crimes registered by the police. The 
prisoners. Between 2000 and 2010, this expenditure 
researchers recognised that these estimates were 
was estimated to range, on average, between 0.03% to 
crude, but they could not obtain a better proxy for 
0.05% of GDP in 22 European countries. On applying 
this particular component of the estimates. 
these percentages to the entire EU for the year 2010, 
In Sweden, expenditure on drug-related prosecutions 
the estimated expenditure was within the range of 
and court cases (district court, court of appeal and 
3.7 billion euros to 5.9 billion euros.
Examples of sectorial models  ► Page 17

Examples of national studies
Several models and data sources have been  2008 (Vander Laenen, De Ruyver, Caulkins & Lievens, 
applied in different national contexts to identify 
2012). It further developed upon two earlier stud-
labelled and unlabelled expenditure allocated to 
ies (De Ruyver et al. 2004, 2007) by carrying out a 
drug control initiatives. Due to national specificities, 
new and more refined estimation of public expendi-
neither their external validity nor the comparability 
ture to combat illegal drugs. The study combined a 
of the methods used have been tested. The extent 
top-down and a bottom-up approach for estimating 
and specificity of labelled drug-related expenditure 
public expenditure. The vast majority (98.45%) of 
vary substantially across countries, as do the data and 
the expenditures were identified as a result of the 
methods applied for estimating unlabelled expen-
top-down approach. Public expenditures identified 
diture. The national estimates presented below are 
through the bottom-up approach (1.55%) concerned 
therefore not directly comparable. They nonetheless 
organisations that depended on the government for 
provide examples of useful models and estimates and 
most of their funding.
illustrate some of the approaches applied. 
The total drug-related expenditure was broken down 
Croatia
by programme: law enforcement, treatment, pre-
vention, harm reduction and other. For 2008, public 
Budak et al. (2013) aimed to identify the central gov-
expenditure on law enforcement constituted 45% of 
ernment’s total drug-related public expenditure and 
the total expenditure. This was slightly less than the 
to develop a method of estimating and allocating 
spending on treatment (49%) and substantially more 
unlabelled expenditure by type of drug policy pro-
than that on prevention (4%), harm reduction (0.8%) 
gramme (prevention, treatment, social reintegration, 
and other (1.2%). When estimated in the same way in 
harm reduction and law enforcement). For labelled 
2004 and 2008, public expenditure on law enforce-
expenditure, governmental institutions were asked 
ment showed a substantial increase, both nominally 
to classify budget expenditure by public function 
(from 186 038 337 euros to 243 000 490 euros) and 
and by type of programme. Unlabelled expenditures 
in relation to the other programmes (it increased by 
were identified indirectly with a system of reparti-
6 percentage points).
tion keys, which were applied to the total state unit 
budget (minus labelled costs). The repartition keys 
were estimated using supply reduction activity data. 
Italy
Unlabelled public expenditures were estimated on 
the assumption that they make up the part of public 
For the purpose of estimating drug-related public 
expenditure remaining after labelled public expendi-
expenditure in Italy (Reitox Italian Focal Point, 2014), a 
tures for combating drug abuse have been deducted 
model was developed to analyse the flow of cost infor-
from the total expenditure of a public body. 
mation from various sources. The model consisted of 
For the period 2009-2012 the study suggested that 
four components: private or indirect costs (individual 
public expenditure on law enforcement constituted 
costs and costs due to loss of productive capacity) and 
about 73% of total drug-related public expenditure by 
public expenditure or direct costs (law enforcement 
central government, whereas prevention, treatment, 
costs, social and health costs). To determine the costs 
social reintegration and harm reduction represented 
of law enforcement, different sources of information 
12%, 13%, 0.3% and 2%, respectively. When comparing 
were used: data concerning traffic control and traffic 
unlabelled expenditure for the different programmes 
accidents; police data on people caught with drugs 
in a single year (2011), unlabelled expenditure on law 
for personal use; data on the number of convictions 
enforcement represented 82% of total unlabelled 
for drug trafficking; and data on crimes related to 
drug-related expenditure. On the other hand, law 
drug trafficking.
enforcement accounted for 4% of the total labelled 
expenditure. Overall, the estimates indicated that 
For 2011, the cost of drug-related law enforcement 
drug-related expenditure stood at 0.2 % of the GDP.
was estimated at 1 600 435 296.60 euros, or roughly 
40 euros per inhabitant aged 15-64 years. The largest 
Belgium
cost component was prisons and alternative measures 
(65%), whereas trials and legal expenses, law enforce-
The study Drugs in Figures III measured how much 
ment activities and administration represented 21.3%, 
the Belgian Government spent on drug policy in 
13% and 0.7%, respectively.
► Page 18

France
costs due to loss of productivity) and public spend-
ing, including direct spending on supply reduction 
In a French study the method relied on analysing 
services. These were disaggregated into spending 
activity records, wherever available in the agencies 
on law enforcement and on criminal justice, which 
concerned (Kopp, 2015). The total expenditure for drug-
included factors such as law enforcement agencies 
related activities was then aggregated. The top-down 
and the federal drug control service. 
approach applied in this case provided an indication of 
the proportion of expenditure for drug control related 
Public expenditure on supply reduction services was 
activities compared to the overall expenditure of all 
estimated using a top-down approach and various 
the institutions and agencies concerned. To obtain an 
sources of information: police data on persons caught 
estimate, a fraction was applied to the total staff and 
with drugs for personal use; data on the number of 
routine operating costs of the agency concerned. In the 
sentences for drug trafficking; and data on crimes 
year 2010, for example, 10% of police activities were 
related to drug trafficking. As there was no pub-
attributable to drug control activities, which involved 
lished information on the fraction attributable to 
60 police units. In this example, police expenditures 
drug-related crime in Russia, the fraction estimated 
attributable to drug-related activities were calculated 
in a study by the US Office of National Drug Control 
by multi plying the total expenditure of the police 
(22%) was employed with a view to estimating the 
services by this fraction of 10%. 
law enforcement and judicial system expenditures. 
A bottom-up approach was also adopted, based on the 
working time of staff performing support functions in 
Portugal
connection with drug-related activities or the equip-
There are few examples of attempts to estimate 
ment used, as recorded by the agencies concerned. 
the impact of changes in the legal system on drug-
For example, the time spent giving prevention talks 
related public expenditure and drug-related bud-
in schools and the time spent by the police forces on 
gets. Gonçalves et al. (2015) are an exception as they 
alcohol tests were included in the calculations.
conducted a comprehensive social cost analysis of 
the situation before and after decriminalisation in 
Luxembourg
Portugal. The authors found a significant reduction in 
the non-health related costs of drug policy between 
Since 1999, the social costs of drugs have been esti-
2000 and 2004, in particular in the legal system (direct) 
mated annually in Luxembourg. These estimates take 
costs. Although these observations highlight signifi-
account of the total costs to public and private agents 
cant changes, prudence is still called for in concluding 
of the consequences of drug use and trafficking. Public 
causal relationships with the new Portuguese National 
spending is analysed in five sectors: prevention, treat-
Strategy for the Fight against Drugs (NSFAD).
ment, harm reduction, law enforcement and research. 
In the law enforcement field, as in other fields, the 
analysts face the twofold challenge of accounting 
Other national studies
for drug-related spending, as financed by different 
general government levels, and of developing models 
There are other examples of public expenditure studies 
to extract unlabelled drug-related expenditure from 
additional to those mentioned above. For instance, 
broader budgets (Origer, 2002).
Mostardt et al. (2010) estimated public expenditure 
in 2006 for Germany using data from Eurostat and 
Law enforcement was estimated to account for 39% 
the COFOG system, concluding that supply reduction 
of total drug-related public expenditure in 1999; pre-
represented close to 65% of the total drug-related 
vention, treatment and harm reduction expenditure 
public spending; Rigter (2006) estimated that 75% of 
amounted to 59%, whereas research and other stood 
public expenditure was spent on law enforcement in 
at 2%. Overall, drug-related public expenditure rep-
the Netherlands; Ramstedt (2006) presented public 
resented 0.013% of GDP.
expenditure estimates for Sweden, whereas public 
spending on supply reduction represented between 
Russia
70 to 76% of the total; and Lievens et al. (2016) pub-
lished a social cost study, including estimates of public 
For Russia, public expenditures on law enforcement 
expenditure to deal with legal and illegal drugs in 
agencies and on the judicial system were estimated as 
Belgium. There are also US (ONDCP, 1989-2015) and 
part of a social study (Potapchik and Popovich, 2014). 
Australian (Moore, 2008) estimates. Despite substantial 
The comprehensive model encompassed private and 
differences, the studies may all be viewed as necessary 
indirect costs (the cost for the individual and the 
first steps in national drug policy evaluations. 
Examples of national studies ► Page 19

International databases 
used to model drug-related 
public expenditure
The only available international compilation  Appendix 1 provides a list of relevant data sources. 
of updated estimates of drug-related public 
In addition to the two data sources already men-
expenditure on supply reduction is published 
tioned, there is information on international reporting 
by the EMCDDA for the EU member states3, report-
concerning supply reduction factors such as: drug 
ing the available national estimates of total drug-re-
related crime (EMCDDA and the European Institute 
lated spending and spending separated into sup-
for Crime Prevention and Control); prison activity 
ply and demand reduction initiatives. The scope for 
and costs (the Council of Europe);crime and criminal 
cross-country comparisons is nonetheless limited 
justice systems (Eurostat and the European Institute 
because the estimates often do not use comparable 
for Crime Prevention and Control). Appendix 3 makes 
definitions, data sets or methodologies.
an extensive description of data published by inter-
Another database of particular relevance is Eurostat. 
national institutions. 
This is partly because it is based on a consistent cat-
egorisation system and on internationally agreed 
definitions, which are required features for interna-
tional comparison. The Classification of the Functions 
of Government (COFOG) is a detailed classification 
system for the functions or socioeconomic objectives 
that general government units aim to achieve through 
a range of outlays. Eurostat has published annual data 
according to the COFOG classification for European 
countries since the early 1990s. This data source has 
proved to be relevant and amenable to a wide variety 
of analytic applications. However, the data set does 
not comprise data concerning specific spending on 
drug-related public initiatives. In order to disentangle 
drug-related expenditure from the broad classes of 
public spending, modelling approaches are adopted 
according to the sector of intervention. 
3.  See http://www.emcdda.europa.eu/topics/
drug-related-public-expenditure.
► Page 20

Conclusions
 
► Every European country allocates significant 
 
► The total budget for supply reduction services 
public resources to the drug policy field. Public 
is the sum of labelled and unlabelled expendi-
expenditure studies can reveal how much public 
tures. Labelled expenditures are clearly iden-
authorities are spending on drug policy and for 
tified in public budgets, whereas a modelling 
what purposes such expenditure is incurred. 
procedure is required for estimating unlabelled 
ones. The modelling is based on either a top-
 
► Public expenditure estimates can be used as a 
tool for assessing whether policy intentions are 
down or a bottom-up approach. Using both 
actually reflected in action, and they constitute 
approaches as complementary is advantageous 
a necessary tool for implementing thorough 
but expensive. A list of advantages and lim-
policy evaluations. Public expenditure studies 
itations for both alternatives is provided, in 
should mirror all relevant activities and policy 
addition to empirical expenditure studies for 
approaches and may be particularly appropriate 
supply reduction activities in some European 
in times of austerity.
countries. 
 
► Estimates exist for 16 EU countries, out of the 
30 potential reporting countries (EMCDDA, 
 
► While recognising the limitations imposed by 
2014b). Estimates suggested that drug-related 
the data sets currently available, this report 
expenditure ranged from 0.01 % to 0.5 % of GDP. 
provides examples of current practice and, in so 
12 out of the 16 reporting countries allocate the 
doing, suggests areas of future focus for desired 
largest share of drug-related public expenditure 
methodological development. It is hoped that 
for supply reduction. 
the estimation of drug-related public expendi-
ture on supply reduction initiatives and policy 
 
► Data availability is one of the main limitations in 
evaluation will move forward in Europe. For 
this field. The use of international databases is 
continued improvements to take place, how-
recommended, whenever possible. These data 
ever, it is essential that a network of experts is 
sets employ broadly accepted concepts and 
developed and maintained. Partnerships should 
definitions providing better comparable data. 
be extended and maintained with the goal 
Sometimes, however, national data sets can con-
of developing good practices, standards and 
tain more detailed or reliable information. 
guidelines in this field.
 ► Page 21

Recommendations
1.  Improving estimation methods with further meth-
5.  A methodology using a set of repartition keys 
odological developments; agreeing on best prac-
according to COFOG categories can be a starting 
tices and; finding reliable standardised data will 
point in order to estimate unlabelled drug-related 
enhance the utility of public expenditure estimates, 
expenditures. General agreement among all par-
as that will permit analysis over time and across 
ticipating countries on definitions and methods 
policy areas and countries. 
will help to improve the comparability of results 
2.  Improved data quality and developing relevant 
between countries. 
data sources is needed for conducting more precise 
6.  Public expenditure studies involve analytical work, 
estimations of spending on drug control measures 
which requires adequate human and technical 
and to measure the impact of drug control policies. 
capacities in all relevant stakeholder fields. This is 
One option is to develop guidelines for data col-
important for obtaining the data quality needed 
lection and economic modelling of evaluations.
for aggregation and comparison. To achieve this, 
3.  It is essential to classify public expenditure based on 
a network of experts could be established and a 
the purpose for which the expenditure is intended. 
working group of experts developed.
It is therefore useful to use a consistent categorisa-
7.  Developing methods to estimate public expendi-
tion system, such as the international Classification 
ture on supply reduction requires effective work-
of the Functions of Government (COFOG).
ing partnerships between drug policymakers and 
4.  Cross-country comparisons are important, but 
specialists in the police, law courts and prisons. 
they are only possible with a common methodol-
Collaboration with public accountancy experts and 
ogy of public expenditure estimates. International 
those in charge of economic modelling is required 
data sets and modelling techniques need to be 
to guarantee meaningful estimates. 
expanded and improved in order to increase the 
capacity to carry evidence based on drug policy 
evaluations in the drug field.
► Page 22

Glossary 
Attributable fractions also known as repartition 
Cost-benefit analysis converts all types of outcomes 
keys are coefficients estimated to help those who 
to a monetary equivalent, in contrast to cost-effective-
estimate drug-related expenditure with the purpose 
ness analysis (Chalk et al., 2013 and Drummond et al., 
of reflecting the proportion of expenditure allocated 
1997). As a result, the euro value of the intervention’s 
to finance drug-related initiatives. Therefore, attrib-
benefits can be directly compared with the euro value 
utable fractions are designed to accurately isolate 
of the intervention’s costs. Two common methods 
drug-spending, when drug-related expenditure is 
for comparing benefits and costs include calculating 
embedded into a broader budgetary structure. There 
net benefits (costs are subtracted from benefits) and 
is no general methodology to determine repartition 
benefit-cost ratios (benefits are expressed as a percent-
keys. It depends on the case (on the basis of the activ-
age of programme costs). A related type of analysis is 
ity information and data available) (Vander Laenen 
the cost-offset analysis in which future costs or cost-
et al, 2011). 
savings are examined. Since cost-benefit analyses 
combine multiple outcomes into a single measure 
Cost analysis provides monetary estimates of the 
and allow direct comparison of costs to benefits, they 
costs of a particular intervention or set of interven-
often provide clearer guidance than cost-effectiveness 
tions, and also information on the amount of resources 
analyses on which treatment programmes should be 
(e.g. labour, facility, supplies) used in their provision. 
adopted – namely those programmes whose ben-
The latter information is often used to identify critical 
efits exceed their costs. Cost-effectiveness analyses 
cost components of the intervention and to assess 
can provide a ranking of competing alternatives but 
whether the costs are affected by changes in key 
not information on the extrinsic value of any single 
assumptions (Bray and Zarkin, 2006). In addition 
intervention independent of the alternatives (Bray 
to being the first step in a cost-effectiveness and 
and Zarkin, 2006).
cost-benefit analysis, cost studies can also be used 
to compare the relative costs of one intervention to 
General government comprises the central govern-
another or to monetise savings from implementing 
ment, state government (in some countries it applies 
a particular action (Chalk et al., 2013).
to the federal level of government), local and social 
Cost-effectiveness analysis involves estimating 
security funds (Eurostat, 2011). 
the ratio of the difference in costs between two 
alternatives (net costs) divided by the difference in 
Government expenditure is defined as a particular 
the outcomes (net effectiveness) (Gold et al., 1996). 
set of transactions, comprising the expending under-
Traditionally, this measure has been used in health 
taken by general government sector units (Eurostat, 
economics. However, this evaluation tool can be used 
2011). 
in any framework of policy intervention, given that the 
outcome measures are those relevant for each type 
Economic evaluation is a comparative analysis of 
of public policy analysed. It is, essentially, the incre-
alternative actions in terms of both their costs and 
mental price of obtaining a unit outcome effect (e.g. 
consequences (Drummond et al., 1997).
a 10% reduction in the number of drug-law offences 
in the past month) from a given police intervention 
Labelled drug-related expenditure is the ex-ante 
(e.g. introducing drug squads in problem neighbour-
planned public expenditure made by general gov-
hoods) when compared to an alternative (e.g. regular 
ernment in the budget that reflects the public and 
policing). Intervention costs are estimated in monetary 
voluntary commitment of a country in the field of 
units, such as the euro. The effect of the intervention 
drugs. In addition, it is any expenditure identified 
can be any policy-relevant outcome that is collected 
as drug-related in public accountancy documents 
for all interventions under consideration. 
(EMCDDA, 2008).
 ► Page 23

Public expenditure is the value of goods and services 
costs caused by the consequences of drug use, which 
purchased by the general government of a state in 
can affect anyone in society, including those who 
order to perform each of its functions. The functions 
do not necessarily use drugs or are involved in drug 
of governments are, among others, the provision of 
demand, supply or drug policy. For instance, external 
health care, justice, public order, education and social 
costs are the expenditure on drug-related nuisance, 
protection. Public expenditure studies are important 
expenditure on tackling offences committed under the 
because they provide information about the size and 
influence of drugs, losses of productivity or absentee-
the composition of the costs of public programmes 
ism associated with either drug trafficking or dealing 
and interventions (Eurostat, 2011). 
activities, among others (Single et al., 2003).
Total drug-related public expenditure is the sum of 
Social costs of illegal drugs comprise all costs carried 
the labelled and unlabelled drug-related expenditure 
by the different sectors of society as a consequence 
(EMCDDA, 2008). 
of the illicit drug phenomenon. Public expenditure 
is only one of the cost elements here. Social cost is 
Unlabelled drug-related expenditure is the non-
the sum of public expenditure, private expenditure 
planned or non-publicly announced ex-post public 
and external costs. Private expenditure constitutes, 
expenditure incurred by the general government in 
for instance, the money spent by private citizens to 
tackling drugs that is not identified as drug-related 
purchase illicit drugs. External costs comprise the 
in the budget (EMCDDA, 2008).
Page 24 ► Public expenditure on supply reduction policies

References 
Bray J. W. and Zarkin G. A. (2006), “Economic evaluation 
European Commission (2012), “The quality of public 
of alcoholism treatment”, Alcohol Research and Health
expenditures in the EU”, Occasional papers 125, ISBN 
No. 29(1), pp. 27-33.
978-92-79-22932-9.
Budak J., Jurlina Alibegović D., Slijepčević S. and 
European Monitoring Centre for Drugs and Drug 
Švaljek, S. (2013), Analiza javnih rashoda za praćenje 
Addiction, EMCDDA (2007) “The State of the Drug 
ostvarivanja ciljeva u području suzbijanja zlouporabe 
Problem in Europe”, 2007 Annual Report, pp. 12-13, 
droga u Republici Hrvatskoj [Analysis of public expen-
Office for Official Publications of the European 
diture for monitoring achievement of the objectives 
Communities, Luxembourg.
in the field of combating drug abuse in the Republic 
of Croatia], Ekonomski institut, Zagreb
European Monitoring Centre for Drugs and Drug 
i Ured za suzbijanje zlouporabe droga Vlade Republike 
Addiction, EMCDDA (2008), “Selected Issue: Towards 
Hrvatske, Zagreb
a Better Understanding of Drug-Related Public 
Expenditure in Europe”, EMCDDA Papers, Office for 
Carnevale Associates (2008), “FY02-09 Budget empha-
Official Publications of the European Communities, 
sizes least effective ingredients of drug policy”, 
Luxembourg.
Carnevale Associates LLC, policy brief, available at: 
http://www.carnevaleassociates.com/Federal_Drug_
European Monitoring Centre for Drugs and Drug 
Budget_FY02_09_Trend.pdf
Addiction, EMCDDA (2014), “Estimating public expen-
diture on drug-law offenders in prison in Europe”, 
Chalk M., Alanis-Hirsch K., Woodworth A., Kemp J. and 
EMCDDA Papers, Office for Official Publications of the 
McLellan T. (2013), “FDA approved medication for the 
European Communities, Luxembourg.
treatment of opiate dependence: literature reviews 
on effectiveness and cost-effectiveness”, Treatment 
European Monitoring Centre for Drugs and Drug 
research institute.
Addiction, EMCDDA (2014a), “Financing drug policy 
Costa Storti C. and De Grauwe P. (2009), “The cocaine 
in Europe in the wake of the economic recession”, 
and heroin markets in the era of globalisation and 
EMCDDA Papers, Office for Official Publications of the 
drug reduction policies”, International Journal of Drug 
European Communities, Luxembourg.
Policy, No. 20(6), pp. 488-496.
European Monitoring Centre for Drugs and Drug 
Council of the European Union (2012), “EU Drug 
Addiction, EMCDDA (2014b), “European Drug Report 
Strategy (2013-2020)”, JAI901, 17547/2, Brussels.
– Trends and developments”, EMCDDA Papers, Office 
for Official Publications of the European Communities, 
De Ruyver B., Casselman J. and Pelc I. (2004), Drug 
Luxembourg.
policy in figures. Study of the actors involved, cost price 
calculation and population reached, Academia Press, 
European Monitoring Centre for Drugs and Drug 
Ghent.
Addiction, EMCDDA (2015), “European Drug Report 
De Ruyver B., Van Malderen S. and Vander Laenen F. 
– Trends and developments”, EMCDDA Papers, Office 
(2007), Study into public expenditure with regard to 
for Official Publications of the European Communities, 
national drug policies. A feasible plan for the national 
Luxembourg.
focal points, Academia Press, Ghent.
European Monitoring Centre for Drugs and Drug 
De Ruyver B. et al. (2007), Drug Policy in Figures II: 
Addiction, EMCDDA (2016), Countries’ overviews, avail-
Follow-up Research into the Actors, Public Spending 
able at www.emcdda.europa.eu/countries.
and Reached Target Groups, Academia Press, Ghent.
European Monitoring Centre for Drugs and Drug 
Drummond M. F., O’Brien B., Stoddart G. L. and Torrance 
Addiction and Europol (2016), “EU Drug markets 
G. W. (1997), Methods for the economic evaluation of 
report: in-depth analysis”, EMCDDA-Europol joint 
health care programmes (2nd edn), Oxford University 
publication, Office for Official Publications of the 
Press, Oxford.
European Communities, Luxembourg.
 ► Page 25

Eurostat (2011), “Manual on sources and methods for 
Mostardt et al. (2010), “Schaetzung der Ausgaben 
the compilation of COFOG statistics – Classification 
des offentlichen Hand durch den Konsum illegaler 
of the Functions of Government (COFOG)”, Eurostat 
Drogen in Deutschland”, Das Gesundheitswesen, No. 
methodologies and working papers (2011 edn), Office 
73(12), pp. 886-894.
for Official Publications of the European Communities, 
Luxembourg.
Official Journal of the European Union (2013), “EU 
Action Plan on Drugs 2013-2016”, 2013/C 351/01.
Fazey C.S.J. (2003), International Journal of Drug Policy
No. 14, pp. 155-169.
Origer A. (2002), Le coût économique direct de la poli-
tique et des interventions publiques en matière d’usage 
French Monitoring Centre for Drugs and Drug 
illicite de drogues au Grand-Duché de Luxembourg. 
Addiction (2012), Recent trends in drug-related public 
Research series No. 4, Point focal OEDT Luxembourg 
expenditure and drug-related services in France, avail-
– CRP-Santé, Luxembourg.
able at http://en.ofdt.fr/BDD/publications/docs/
nr2012si2.pdf 
Potapchik E. and Popovich L. (2014), “Social cost of 
substance abuse in Russia”, Value in health regional 
Genetti B. (2014), “First component of costs (costs of 
issues, No. 4C, pp. 1-5.
enforcing the Law) – conceptual model, methodol-
ogy and results in Italy”, presented at the conference 
Ramstedt M. (2006), “What drug policies cost. 
“A national study on drug-related social costs”, 7-11 
Estimating drug policy expenditures in Sweden, 2002: 
April 2014, Zagreb.
work in progress”, Addiction, No. 101, pp. 330-338.
Gold M.R., Siegel J.E., Russel L.B. and Weinstein M.C. 
Reuter P., Ramstedt M. and Rigter H. (2004), Developing 
(eds) (1996), Cost-effectiveness in Health and Medicine
a Framework for Estimating Government Drug Policy 
Oxford University Press, New York.
Expenditures, EMCDDA, Lisbon.
Gonçalves R, Lourenço A, Silva S. N. (2015), “A social 
Reuter P. (2006), “What drug policies cost. Estimating 
cost perspective in the wake of the Portuguese strat-
government drug policy expenditures”, Addiction, No. 
egy for the fight against drugs”, International Journal 
101(3), pp. 315-322.
of Drug Policy, No. 26(2), pp. 199-209.
Rigter H. (2004), “Drug policy expenditures in the 
Kopp P. and Fenoglio P. (2002), “Calculating the social 
Netherlands, 2003”, in Reuter P., Ramstedt M. and 
cost of illicit drugs”, Pompidou Group, Council of 
Rigter H. (eds), Developing a Framework for Estimating 
Europe Publishing.
Government Drug Policy Expenditures, European 
Kopp, P. & Fenoglio, P. (2003), Public spending on drugs 
Monitoring Centre for Drugs and Drug Addiction, 
in the European Union during the 1990s, EMCDDA, 
pp. 37-73, Lisbon.
Lisbon.
Rigter, H. (2006), “What Drug Policies Cost. Drug Policy 
Kopp, P. (2006), Économie de la drogue, Éditions La 
Spending in the Netherlands in 2003”, Addiction, No. 
Découverte, Paris.
101pp. 323-329.
Kopp, P. (2015), Le côut social des drogues en France
Serpelloni, G. et al. (2013), “Italy’s electronic health 
OFDT.
record system for opioid agonist treatment”, Journal 
of Substance Abuse Treatment, No. 45(2), pp. 190-195.
Lievens D., Laenen F. V., Caulkins J. and De Ruyver 
B. (2012), “Drugs in Figures III - Study of public 
Single E. et al. (2003), International Guidelines for 
expenditures on drug control and drug problems”, 
Estimating the Economic Costs of Substances Abuse 
European criminal justice and policy: Governance of 
(2nd edn), World Health Organization, Geneva.
Security Research Paper Series, No. 7, Maklu Publishers, 
Single E., (2009), “Why we should still estimate the 
Alperdoon.
costs of substance abuse even if we needn’t pay undue 
Moolenaar D. E. G. (2009), “Modelling criminal justice 
attention to the bottom line”, Drug and Alcohol Review
system costs by offence; lessons from the Netherlands”, 
No. 28(2), pp. 117-121.
European Journal of Criminal Policy Research, No. 15, 
Vander Laenen F., Vandam L., De Ruyver B. and Lievens 
pp. 309-326.
D., (2011) Studies on public expenditure in Europe: 
Moore T. J. (2005), Monograph No. 01: What is Australia’s 
possibilities and limitations, Bulletin on Narcotics, Vol 
“drug budget”? The policy mix of illicit drug-related gov-
LX, 2008, UNODC, Vienna 
ernment spending in Australia, Fitzroy: Turning Point 
Alcohol and Drug Centre. 
Vander Laenen F. and Lievens D. (forthcoming), “A 
cross-national comparison of public expenditures on 
Moore, T. (2008), “The size and mix of government 
drug treatment, context is key”, in Drug-related treat-
spending on illicit drug policy in Australia”, Drug and 
ment expenditure: a methodological insight, EMCDDA, 
Alcohol Review, No. 27pp. 404-413.
Lisbon.
Page 26 ► Public expenditure on supply reduction policies

Appendix 1 – Available databases 
and potential indicators for drug-
related public expenditures 
Examples of international databases, which can be used for estimating drug-related 
public expenditures
Level of 
estimation
Examples of databases
Estimation data
International EMCDDA Statistical 
 
– The EMCDDA statistical bulletin covers a broad range of 
bulletin
areas including the most recent estimates of drug-related 
and 
crime in the form of drug seizures, types of offence, price, 
purity and use in prison, and country responses to the 
Public expenditure 
drug situation in Europe. http://www.emcdda.europa.eu/
database
data/stats2015
 
– The EMCDDA also publishes the most recent national data 
on drug-related public expenditures available in Europe. 
 
– http://www.emcdda.europa.eu/topics/
drug-related-public-expenditure
WHO Database
 
– Global Information System on Resources for the 
Prevention and Treatment of Substance Use Disorders 
(includes information about: prevalence and burden of 
Disease, monitoring and surveillance; policy; treatment 
system and services; pharmacological treatment; preven-
tion programmes for substance use and related harm; and 
human resources and civil society involvement).
Eurostat
Public expenditure according to the Classification of the 
Functions of Government (COFOG)
COFOG published data according to two levels of classification 
(United Nations, 2008). The first classifies expenditure into 10 
general functions, one of which is “Public order and safety”. The 
second classifies expenditure into 69 groups, in which there 
are three indicators of interest: police service, law courts and 
prisons. The definitions below are provided by the UNODC.
From the Public order and safety section:
Police services
 
– Administration of police affairs and services, including alien 
registration, issuing work and travel documents to immi-
grants, maintenance of arrest records and statistics related to 
police work, road traffic regulation and control, prevention 
of smuggling and control of offshore and ocean fishing.
 
– Operation of regular and auxiliary police forces, of port, 
border and coast guards, and of other special police forces 
maintained by public authorities; operation of police labora-
tories; operation or support of police training programmes.
 ► Page 27

Level of 
estimation
Examples of databases
Estimation data
Law Courts
 
– Administration, operation or support of civil and criminal 
law courts and the judicial system, including enforcement 
of fines and legal settlements imposed by the courts and 
operation of parole and probation systems.
 
– Legal representation and advice on behalf of the govern-
ment or on behalf of others provided by government, in 
cash or in services.
Prisons
 
– Administration, operation or support of prisons and other 
places for the detention or rehabilitation of criminals such 
as prison farms, workhouses, reformatories, asylums for the 
criminally insane, etc.
UN-CTS (Crime and 
Data produced by UNODC have multiple sources. Mem-
Criminal Justice 
ber States regularly submit to UNODC statistics on drugs 
Statistics)
(through the Annual Report Questionnaire) and crime and 
criminal justice (through the annual Surveys on Crime Trends 
and Operations of Criminal Justice Systems). Other data are 
collected through national surveys implemented by UNODC 
in co-operation with national governments or are compiled 
from scientific literature. UNODC attempts to maximise the 
comparability of the data and estimate regional and global 
statistics.
SPACE
SPACE unites two related projects: SPACE I provides data 
on penal institutions and the population held in custody, 
as well as on certain conditions of detention, while SPACE 
II collects information on persons serving non-custodial 
sanctions and alternative measures.
Data are collected every two years by means of two question-
naires sent to the equivalents of the ministries of justice, the 
penitentiary administrations and the probation authorities of 
each country in Europe. The collection and validation of these 
data then takes place at the University of Lausanne, where 
analyses and interpretations for both projects are formulated 
through a common methodology. This methodology aims 
to allow comparisons among states at European level, by 
proposing SPACE categories instead of each country’s own 
national categories, while still including questions regarding 
the particularities of their specific sanctions and measures. The 
SPACE project produces two annual reports: SPACE I – Prison 
Populations and SPACE II – Persons Serving Non-Custodial 
Sanctions and Measures, presenting the data collected and 
the key points of the results.
Page 28 ► Public expenditure on supply reduction policies

Level of 
estimation
Examples of databases
Estimation data
European Sourcebook 
The Sourcebook contains data from 41 European countries 
on Crime and Criminal 
regarding their criminal justice systems. The book is structured 
Justice Statistics
into six main chapters covering different stages of the judicial 
system: Police Statistics, Prosecution Statistics, Conviction 
Statistics, Prison Statistics, Probation Statistics and, for the 2014 
edition, a final chapter on National Victimization Surveys. The 
data provided are systematically accompanied by texts and 
notes relating to the specificity of each country and which 
discuss the different challenges attributed to the comparison 
of the data.
Social Expenditure 
The OECD Social Expenditure Database (SOCX) provides a 
Database
unique tool for monitoring trends in aggregate social expendi-
ture and analysing changes in its composition. The main social 
policy areas are as follows: old age, survivors, incapacity-related 
benefits, health, family, active labour market programmes, 
unemployment, housing, and other social policy areas.
ESPAD
Drug abuse prevalence among teenagers in European 
countries.
National
Database of national 
Expenditures of different groups, in which can be found some 
statistics
indicators of interest: police service, law courts, prisons, medi-
cal and social services. 
Annual report from 
Data on Social Services Department expenditures at regional 
Social Services 
level and the number of drug users receiving social benefits 
Department
in connection with drug use.
Appendix 1 – Available databases and potential indicators for drug-related public expenditures  ► Page 29

Appendix 2 – The international 
Classification of the Functions 
of Government (COFOG)
The COFOG classification has three structural  (non-investment) transfers. Eurostat has published 
levels. At the first level, government expenditure 
annual data according to the COFOG definitions for 
is broken down into 10 functions. These are then 
the European Union countries since the early 2000s.
divided into 69 groups (second level of COFOG), which 
The extensive structure of COFOG contrasts with the 
are themselves divided into classes at the third level – 
four-category division introduced by Reuter (2006), 
the most detailed classification level. COFOG permits 
based on the likely effects of services provided by 
an examination over time of trends in government 
drug policy programmes (namely prevention, treat-
outlays on particular functions (Eurostat, 2011).
ment, enforcement and harm reduction). Reuter’s 
programme division is the classification of the recip-
The detailed three-level structure of COFOG includes 
ients (NPOs) with drug-policy programmes. 
financial flows of public finance, which are going from 
An example of an overview of public expenditure 
state and local (regional and municipal) budgets to 
groups, broken down according to the main public 
non-profit organisations (NPOs) with drug-policy 
functions pursuant to the international classification 
programmes. COFOG is a functional classification 
of the functions of the government at the third level, 
system used by the System of National Accounts 
is shown in the table below.
1993. COFOG is a useful international classification 
system for spatial comparison (between countries) 
A pragmatic approach towards drug-related research 
and also for time comparison (over time). In principle, 
and public expenditure estimates would suggest 
its units of classification are individual transactions. 
adopting a classification such as COFOG, as proposed 
This means that each outlay (purchase or transfer) 
by Eurostat. The COFOG classification system guar-
should be assigned a COFOG code according to the 
antees annually available data for most European 
function that the transaction serves. This principle is 
countries, according to harmonised definitions and 
valid for both capital transfers (investment) and current 
standard data collection procedures. 
Public expenditures according to the classification of public functions
Public functions
Public functions at the third level of classification
01 General public services
014 Basic research
03 Public order and safety 
031 Police services
033 Law courts
034 Prisons
07 Health
071 Medical products, appliances and equipment
072 Outpatient services
073 Hospital services
074 Public health services
075 R&D health
09 Education
091 Pre-primary and primary education
092 Secondary education
094 Tertiary education
095 Education non-definable by level
096 Subsidiary services to education
10 Social protection
105 Unemployment
106 Housing 
107 Social exclusion 
► Page 30

Appendix 3 – Summary tables: 
data from international databases
Table 1 - Public expenditure
Data and 
Dataset
Type of 
DATABASE
YEARS
Number of 
Statistics
information
observations(*) 
Expenditure 
Law Courts
Eurostat, 
Europe (31) = 473/651
Public 
of the 
Police Services
European 
Europe (31) = 473/651
expenditure general 
Union 
1995-2015
government
Prisons
Europe (31) = 473/651
(EU)
Total drug-
related public 
expenditure
Drug-
Public 
Country Drug 
related 
expenditure 
Percentage 
Profiles,
Last year 
public 
on supply 
spent on supply 
available
EU (30) =20/30
expenditure reduction
reduction
EMCDDA, (EU)
Percentage spent 
on demand 
reduction
(*) The number of observations reports the number of data records, taking into account the territory; countries and years available. 
The ratio compares the number of effectively reported observations with the total number of records, if no data were missing. 
Example: Europe (44) = 28/368: in Table 4, the conviction statistics of the European Sourcebook of crime and criminal justice sta-
tistics reports 28 data records, for the community sanctions imposed to drug offences in 2010, compared to the 368 data records 
that would exist if no data were missing, in the region Europe (which accounts with 44 countries). 
 ► Page 31

Table 2 – Drug law offences
Data and 
Statistics
Dataset
Type of information
DATABASE
Years
Number of 
observations(*)
Number  Offences
EU (30) = 364/600
of 
1995-2014
offences
Offender
EU (30) = 262/600
Offences  Use
EU (30) = 230/300
by Types
2004-2013
Supply
EU (30) = 238/300
Total
2004-2013
EU (30) = 203/300
Cannabis
Use
EU (30) = 163/270
2005-2013
Supply
EU (30) = 160/270
Total
2004-2013
EU (30) = 186/300
Heroin
Use
EU (30) = 159/270
2005-2013
Supply
EU (30) = 160/270
Total
2004-2013
EU (30) = 185/300
Cocaine
Use
EU (30) = 159/270
2005-2013
Supply
EU (30) = 176/270
Drug law 
Drug Law 
Total
EMCCDA
EU (30) = 50/270
offences
Offences
(EU)
Crack
Use
2005-2013
EU (30) = 47/270
Offences 
Supply
EU (30) = 37/270
by drug
Total
EU (30) = 163/270
Amphetamine
Use
2005-2013
EU (30) = 74/270
Supply
EU (30) = 87/270
Total
EU (30) = 98/270
Methamphetamine Use
2005-2013
EU (30) = 74/270
Supply
EU (30) = 87/270
Total
EU (30) = 162/270
Ecstasy 
Use
2005-2013
EU (30) = 144/270
Supply
EU (30) = 153/270
Total
EU (30) = 127/270
LSD
Use
2005-2013
EU (30) = 108/270
Supply
EU (30) = 95/270
Page 32 ► Public expenditure on supply reduction policies

Table 3 – Prison population
Data and 
Statistics
Dataset
Type of information
DATABASE
YEARS
Number of 
observations(*)
Persons held in 
Availability of 
institutions for drug  institutions for drug 
users offenders 
users offenders, 
2014
CoE (47) =28/53
outside penal 
outside penal 
institutions
institutions
Population on 
1st January 
Total number of 
prisoners (including 
pre-trial detainees)
Total number of 
detainees held in 
remand institutions/
Situation of prison 
2009 & 
2009: CoE (47) = 343/424
sections (pre-trials)
population
2014
2014: COE (47) = 255/265
Total number of 
prisoners held in 
institutions serving 
a sentence
Total capacity of 
penal institutions
Surface area per 
prisoner (m^2) 
Total number 
Evolution of prison 
of prisoners
CoE (47) = 707/795
population
2000-2014
Prison 
Space I, 
Prison population
CoE (47) = 683/795
population
Council of 
Untried detainees 
Europe (CoE)
(no court decision)”
Detainees found 
guilty but no 
sentence yet
Sentenced prisoners 
(appealed or 
can do so)
Detainees with no 
final sentence, but 
serving a prison 
sentence in advance
Legal status of 
2009 &
2009: CoE (47) = 274/424
prison population
Sentenced prisoners 
2014
2014: CoE (47) = 315/477
(final sentence), 
of which:
 
– fine defaulters
 
– in revocation, 
suspension or 
annulment of the 
conditional release 
or probation
Other cases
Total number of 
prisoners (including 
pre-trial detainees)
Appendix 3 – Summary tables: data from international databases ► Page 33

Data and 
Statistics
Dataset
Type of information
DATABASE
YEARS
Number of 
observations(*)
Main offence of 
sentenced prisoners  Drug offences
2009 & 
(Final Sentence) 
2014
CoE (47) = 88/106
Lengths of 
sentences imposed 
Length of the 
2009
CoE (47) = 405/583
(final sentenced 
sentences by month, 
prisoners)
years or lifetime
2014
CoE (47) = 557/689
Prison population
Prison population 
Europe (44) = 387/414
Stock – Total
(including pre-trial 
European 
2003-2011
detainees): stock
Prison population
Sourcebook 
Europe (44) = 356/414
Pre-trial detainees
of crime 
and criminal 
Total criminal offences
2006 & 
justice 
2010
Europe (44) = 88/92
statistics,
Convicted prison 
Drug offences 
population by 
(of which %)
2010
Europe (44) = 38/46
Université de 
type of offence
Lausanne
Convicted prison 
population in 2010 
2006 & 
Drug offences: Total
2010
Europe (44) = 46/92
Sentenced persons 
Drug Offences
Europe (26) = 49/81
UNODC
held in prisons
2010-2012
Drug Trafficking
Europe (26) =36/81
Page 34 ► Public expenditure on supply reduction policies

Table 4 – Cases registered by the police, prosecutors and law courts
Data and 
Statistics
Dataset
Type of information
DATABASE
YEARS
Number of 
observations(*)
Crime 
Total
1993-2007 Europe (36) = 536/585
Recorded by 
Eurostat
Unlawful acts involving 
1993-2007 Europe (36) = 486/585
the Police
controlled drugs or precursors
2008-2014 Europe (39) = 275/287
Drug-Related 
Total
Europe (40) = 215/258
Crimes at 
the national 
Drug Possession
Europe (21) = 101/138
level, number 
UNODC
2003-2008
of police-
Police 
recorded 
Drug Trafficking
Europe (37) = 175/240
statistics
offences
Criminal Offences
Europe (42) = 347/387
Offences
Drug Offences
Europe (42) = 333/387
Police 
Statistics- 
Drug Trafficking
Europe (41) = 269/387
Offences/
2003-2011
Criminal Offenders
Europe (42) = 263/396
Offenders
Offenders
Drug Offenders
Europe (42) = 245/396
Drug Trafficking
Europe (42) = 190/396
Output cases: Total
Europe (42) = 218/396
Percentage brought before 
2003-2011
a court of the total output of 
Criminal cases 
criminal cases handled by the 
Europe (42) = 198/396
handled by the  prosecuting authorities
prosecuting 
authorities 
Output 
Drug Offences
European 
Europe (42) = 33/88
cases by 
Sourcebook 
offence 
2010
Drug Trafficking
of crime 
group
Europe (42) = 25/88
and criminal 
Convictions 
Criminal offences
justice 
Europe (42) = 293/369
Statistics- 
statistics
Drug offences
2003-2011 Europe (42) = 272/369
Conviction  Persons 
statistics
convicted
Drug trafficking
Europe (42) = 193/369
2006
Europe (41) = 203/473
Criminal offences
2010
Europe (41) = 176/602
Total persons 
receiving 
2006
Europe (41) = 175/473
sanctions/
Drug offences
2010
Europe (41) = 158/602
measures
2006
Europe (41) = 113/473
Drug trafficking
2010
Europe (41) = 104/602
Community 
Criminal offences
Europe (44) = 52/368
sanctions and 
measures 
2010
imposed
Drug offences
Europe (44) = 28/368
Appendix 3 – Summary tables: data from international databases ► Page 35

Acronyms
Council of Europe (47) = CoE(47): Albania, Andorra, 
Europe (31): Austria, Belgium, Bulgaria, Croatia, 
Armenia, Austria, Azerbaijan, Belgium, Bosnia-
Cyprus, Czech Republic, Denmark, Estonia, Finland, 
Herzegovina,  Bulgaria,  Croatia, Cyprus, Czech  France, Germany, Greece, Hungary, Ireland, Italy, Latvia, 
Republic, Denmark, Estonia, Finland, France, Georgia, 
Lithuania, Luxembourg, Malta, Netherlands, Norway, 
Germany, Greece, Hungary, Iceland, Ireland, Italy, 
Poland, Portugal, Romania, Slovakia, Slovenia, Spain, 
Latvia, Liechtenstein, Lithuania, Luxembourg,  Sweden, Turkey, United Kingdom
Malta, Republic of Moldova, Monaco, Montenegro, 
Europe (36): Europe (31) + Liechtenstein, Montenegro, 
Netherlands, Norway, Poland, Portugal, Romania, 
TRF-Macedonia, Serbia, Turkey
Russian Federation, San Marino, Serbia, Slovak 
Republic, Slovenia, Spain, Sweden, Switzerland, “The 
Europe (37): Europe (21) + Republic of Moldova, 
former Yugoslav Republic of Macedonia”, Turkey, 
Russian Federation, Slovakia, Ukraine, Denmark, 
Ukraine, United Kingdom 
Estonia, Ireland, Bosnia-Herzegovina, Italy, 
Montenegro, Portugal, Serbia, Spain, TRF- Macedonia, 
European Union (30)= EU(30): Austria, Belgium, 
Germany, Monaco
Bulgaria, Croatia, Cyprus, Czech Republic, Denmark, 
Europa (39): Europe (36) + Albania, Bosnia-
Estonia, Finland, France, Germany, Greece, Hungary, 
Herzegovina, Kosovo
Ireland, Italy, Latvia, Lithuania, Luxembourg, Malta, 
Netherlands, Norway, Poland, Portugal, Romania, 
Europa (40): Belarus, Bulgaria, Czech Republic, 
Slovakia, Slovenia, Spain, Sweden, Turkey, United 
Hungary, Poland, Moldova, Romania, Slovakia, 
Kingdom
Ukraine, Denmark, Estonia, Finland, Iceland, Ireland, 
Latvia, Lithuania, Norway, Sweden, United Kingdom, 
Europe (21): Belarus, Bulgaria, Czech Republic, 
Albania, Bosnia-Herzegovina, Croatia, Greece, Italy, 
Hungary, Poland, Romania, Finland, Iceland, Lithuania, 
Malta, Montenegro, Portugal, Serbia, Slovenia, Spain, 
Norway, Sweden, United Kingdom, Albania, Croatia, 
TRF- Macedonia, Austria, Belgium, France, Germany, 
Malta, Slovenia, Belgium, France, Liechtenstein, 
Liechtenstein, Luxembourg, Monaco, Netherlands, 
Luxembourg, Switzerland
Switzerland
Europa (41): Europa (42), except Luxembourg
Europe (26): Bulgaria, Czech Republic, Poland, Romania, 
Europe (42): Europa (31) + Albania, Armenia, Bosnia-
Russian, Denmark, Estonia, Finland, Ireland, Latvia, 
Herzegovina, Georgia, Kosovo, Moldova, Russia, Serbia, 
Lithuania, Norway, Sweden, United Kingdom, Andorra, 
TRF- Macedonia, Turkey, Ukraine
Greece, Italy, Malta, Portugal, Serbia, Slovenia, Austria, 
Liechtenstein, Monaco, Netherlands, Switzerland
Europe (44): Europe (42) + Azerbaijan, Montenegro
Table references 
Aebi, M. F., Delgrande, N. (2011). SPACE I- Council 
EMCDDA. (2017). http://www.emcdda.europa.eu/
of Europe Annual Penal Statistics. Survey 2009. 
topics/drug-related-public-expenditure
Strasbourg: Council of Europe
Eurostat. (2017). http://ec.europa.eu/eurostat/data/
Aebi, M. F., Tiago, M. M. & Burkhardt, C. (2015). SPACE 
database?node_code=gov_10a_exp
I- Council of Europe Annual Penal Statistics: Prison pop-
UNODC. (2017). http://www.unodc.org/unodc/en/
ulations. Survey 2014.Strasbourg: Council of Europe
data-and-analysis/statistics/crime.html
EMCDDA. (2017). http://www.emcdda.europa.eu/
UNODC. (2017). http://www.unodc.org/unodc/en/
data/stats2015
data-and-analysis/statistics/crime.html
► Page 36

Evaluating drug policy is an integral part of a cost-efficient 
approach to tackle illicit drugs. This report takes a first 
step towards a systematic analysis, by examining a set of 
representative attempts to estimate public expenditure 
on supply reduction interventions. It proposes a common 
set of definitions, aiming to establish a common basis for 
understanding this topic and facilitating comparability in three 
main dimensions: time, policy and countries. Although it is 
mainly confined to supply reduction expenditures, in order to 
set the context, it describes the proportion that total drug-
related expenditure represents of national public spending and; 
presents the balance between demand and supply reduction 
spending for a number of European countries. Finally, with the 
aim of facilitating and promoting future empirical expenditure 
studies and of setting the ground for the development of 
good practices, relevant data sources and methodologies 
applied are listed and discussed and examples of sectorial 
models of public spending are selectively provided.
PREMS 062717
ENG
The Council of Europe is the continent’s 
leading human rights organisation. 
It comprises 47 member states, 28 of which 
www.coe.int
are members of the European Union. 
All Council of Europe member states have signed up to 
the European Convention on Human Rights, a treaty designed 
to protect human rights, democracy and the rule of law.
The European Court of Human Rights oversees 
the implementation of the Convention in the member states.

Document Outline