Money laundering: Who, When and How?

Report 2023/29

Sunniva Heckel

Money laundering: Who, When and How?

Category	Reports
Sub-Categories	Statistics and Empirical AnalysisTaxes and public economics
Year	2023
Report Number	29
Author(s)	Sunniva Heckel
Download	file_download (1.6 MB)
Read in browser	PDF

Rapport 202	3/29 | Masteropp	gave 	 	
 	
 	
 	
Money 	laundering: Who, When and How?	 	
An empirical analysis of the associations between the probability of committing 
money laundering and the sociodemographic and labour force characteristic	s 	
of individuals.	 	
Sunniva A	lice He	ckel

Money laundering: Who, When and How?	 	
 
Vista Analyse	 | Rapport 2023/29 	
 
Dokumentdetaljer	 	
Tittel	 	Money laundering: Who, When and How?	 	
Rapportnummer 	 	Rapport 2023/	29 	
Forfattere	 	Matilde Avdem Frankmo 	
ISBN	 	978	-82	-8126	-640	-7 	
Prosjektleder	 	Orvika Rosnes	 	
Dato for ferdigstilling	 	11. mai 2023 	
Tilgjengelighet	 	Offentlig 	 	
Nøkkelord	 	Statistikk og empirisk analyse, økonomisk kriminalitet, skatter og offentlig økonomi, panel-	
data 	
 
 
 
 
 
 
 
 
 
 
 
 
 
Om Vista Analyse	 	
Vista Analyse AS er et samfunnsfaglig 	analyseselskap med hovedvekt på økonomisk utredning, evaluering, rådgivning 	
og forskning. Vi utfører oppdrag med høy faglig kvalitet, uavhengighet og integritet. Våre sentrale temaområder er 
klima, energi, samferdsel, næringsutvikling, byutvikling og velfe	rd. Vista Analyse er vinner av Evalueringsprisen 2018.	 	
Våre medarbeidere har meget høy akademisk kompetanse og bred erfaring innenfor konsulentvirksomhet. Ved be-
hov benytter vi et velutviklet nettverk med selskaper og ressurspersoner nasjonalt og internasj	onalt. Selskapet er i 	
sin helhet eiet av medarbeiderne.

Money laundering:	 	
Who, When and How?	 	
An empirical analysis 	of the association	s between the probability of committing money laundering 	
and	 the	 sociodemographic and labour force characteristics of individuals.	 	
Sunniva Alice Heckel	 	
Master’s in Economics	 	
30 credits	 	
 
Department of Economics	 	
Faculty	 of Social Sciences	 	
May, 2023	 	
Master thesis

© Sunniva Alice Heckel	 	
2023	 	
Money laundering: Who, When a	nd How?	 	
http://www.duo.uio.no/	 	
Printed	: Graphic 	Center	, Universi	ty of	 Oslo

A	CKNO	WLEDGE	MENT	 	
I am very grateful for all the help I have received from my supervisor Manudeep Bhuller who 
has provided a lot of great input and ideas for this thesis. He has in addition helped me with 
decomposing this large task into several smaller and speci	fic tasks to be able to achieve the aim. 	 	
I am also very thankful to Statistics Norway for letting me use their data.	 The data analysis for 	
this thesis was carried out on a secure server area at Statistics Norway for the research project 
“Criminality, Vict	imization and Social Interactions (CIVICS)”. The data privacy impact 	
assessment (DPIA) is electronically archived under file no. 2018/1150.	 The thesis has also been 	
written in collaboration with Vista Analyse.	 	
I would also like to thank my family and frien	ds for their support during this journey. A special 	
thank is dedicated to my fathe	r who 	taught me all the necessary tools to succeed. 	 	
Any mistakes observed in the thesis are my responsibility. Data wrangling and estimations are 
performed in Stata. Do not 	hesitate to reach me at 	sunniva@hotmail.fr	 for any questions 	
regarding the codes.

A	BS	TRACT	 	
This thesis aims at investigating the associations between sociodemographic and 	labour market	 	
characteristics 	of individuals and their probability to be charged with money laundering or 	
receiving proceeds from crime. 	The 	results are estimated with a 	logit binary choice model 	and 	
are 	based on a	 panel dataset of 	reported 	charges in Norway from 2000 to 2019 from Stat	istics 	
Norway	 and the Norwegian Police Directorate	. 	
The estimations indicate that	 the chances of being charged with any offence	, including money 	
laundering	 or receiving proceeds from crime	, increase if a person is a	 male	 living in a city 	with 	
low	 after	-tax	 income and education	. A distinction is that the probability of being charged with 	
money laundering 	is highest if a	n individual 	is 	not an	 immigrant and 	relat	ive	ly older compared 	
to 	those charged with any offence for which	 the probability is highe	st 	for relatively younger 	
immigrants.	 	
Regarding the 	labour market	 characteristics, compared to individuals who are 	outside of the 	
labour force	, the chances of being charged with any offence, including money laundering, are 	
lower when being employed in the state, municipal 	or private sector 	in any	 industry	. However, 	
when 	being unemployed, the probability of being charged with money laundering 	decre	ases 	
while it 	increases	 for a	n offence in general, compared to those outside the labour force.	 	
When being self	-employed in the	 construction, manufacturing, hotel and restaurant	, and 	
financial	 industries, the probability of being charged with any offence 	increases, while 	the 	
chances of being charged with money laundering	 decrease for the three first	-mentioned 	
industries and 	does not change regarding the financial industry.	 Furthermore, the prob	ability of 	
being charged with any offence, including money laundering, decreases when self	-employed in 	
services or other industries.	 These findings 	are	 however	 based on 	reported criminality and might 	
therefore be biased by non	-reported statistics.

TABLE OF CONTENT	 	
1. 	Introduction	 ................................	................................	................................	................................	.....	 1 	
2. 	Institutional background	 ................................	................................	................................	..................	 2 	
2.1 Definition, implication and process of money laundering and receiving proceeds from crime	 ....	 2 	
2.2 The legislative context	 ................................	................................	................................	...................	 5 	
2.2.1 The Penal Code	 ................................	................................	................................	......................	 5 	
2.2.1 Anti	-Money Laundering Act	 ................................	................................	................................	.. 7 	
2.3 Related literature	 ................................	................................	................................	...........................	 9 	
3. 	Data and methodology	 ................................	................................	................................	...................	 11	 	
3.1 Data Sources	 ................................	................................	................................	................................	 11	 	
3.2 Methodology	 ................................	................................	................................	...............................	 15	 	
3.2.1 Logit Model for Money Laundering and Other Offences	 ................................	....................	 16	 	
3.2.2 Decomposing the Probability of Model Laundering By Co	-offences	 ................................	.. 18	 	
3.3 Descriptive statistics	 ................................	................................	................................	....................	 19	 	
3.3.1 Criminality	 ................................	................................	................................	............................	 19	 	
3.3.2 Gender	 ................................	................................	................................	................................	.. 23	 	
3.3.3 Age	 ................................	................................	................................	................................	.......	 25	 	
3.3.4 Immigration status	 ................................	................................	................................	................	 27	 	
3.3.5 Living in a city	 ................................	................................	................................	.....................	 29	 	
3.3.6 Education level	 ................................	................................	................................	.....................	 31	 	
3.3.7 Labour market characteristics	 ................................	................................	...............................	 33	 	
4. 	Results	 ................................	................................	................................	................................	...........	 46	 	
4.1. Main groups of offences	 ................................	................................	................................	.............	 46	 	
4.2. Money laundering and receiving proceeds from crime	 ................................	..............................	 53	 	
4.3 Implications for policymakers and countermeasures	 ................................	................................	.. 63	 	
5. 	Conclusion	 ................................	................................	................................	................................	.....	 65	 	
References	 ................................	................................	................................	................................	.............	 67	 	
Appendix 1: Decomposition of the outcome variable	 ................................	................................	...........	 70	 	
Appendix 2: Logit estimation of average marginal effects of the outcome alternatives when charged with 
money laundering or receiving proceeds from crime	 ................................	................................	............	 71

LIST OF FIGURES	 	
Figure 1: Number of offenders* in Norway from 2000 to 2019, by main categories of offences	 .........................	 20 	
Figure 2: Share of individuals charged with money laundering or receiving proceeds from crime* from 2000 to 	
2019, by year	 ................................	................................	................................	................................	................	 21 	
Figure 3: Combinations of co	-offences of individuals charged with money laundering or receiving proceeds from 	
crime* from 2000 to 2019, in proportion	 ................................	................................	................................	.....	 22 	
Figure 4: Gender ratio of offenders and the remaining population* from 2000 to 2019, by main categories of 	
offences and year	 ................................	................................	................................	................................	..........	 24 	
Figure 5: Gender ratio of individuals charged with money laundering or receiving proceeds from crime and the 	
remaining population* from 2000 to 20019, by year	 ................................	................................	...................	 25 	
Figure 6: Proportion of age of offenders and the remaining population* from 2000 to 2019, by main categories of 	
offences and year	 ................................	................................	................................	................................	..........	 26 	
Figure 7:	 Proportion of age of individuals charged with money laundering or receiving proceeds from crime and 	
the remaining population* from 2000 to 2019, by year	 ................................	................................	...............	 27 	
Figure 8: Immigrant ratio of offenders and the remaining population* from 2000 to 2019, by main categories of 	
offences and year	 ................................	................................	................................	................................	..........	 28 	
Figure 9: Immigration ratio of 	individuals charged with money laundering or receiving proceeds from crime and 	
the remaining population* from 2000 to 2019, by year	 ................................	................................	...............	 29 	
Figure 10: Proportion of of	fenders and the remaining population living in a city* from 2000 to 2019, by main 	
categories of offences and year	 ................................	................................	................................	....................	 30 	
Figure 11: Proportion of individuals charged 	with money laundering or receiving proceeds from crime and the 	
remaining population living in a city* from 2000 to 2019, by year	 ................................	.............................	 31 	
Figure 12: Proportion of educat	ion level of offenders and the remaining population* from 2000 to 2019, by main 	
categories of offences and year	 ................................	................................	................................	....................	 32 	
Figure 13: Proportion of education level of ind	ividuals charged with money laundering or receiving proceeds 	
from crime and the remaining population* from 2000 to 2019, by year	 ................................	......................	 33 	
Figure 14: Proportion of 	labour market status of active offenders and the remaining population* from 2000 to 	
2019, by main categories of offences and year	 ................................	................................	.............................	 34 	
Figure 15: Proportion of labo	ur market status of active individuals charged with money laundering or receiving 	
proceeds from crime and the remaining population* from 2000 to 2019, by main categories of offences and 
year	 ................................	................................	................................	................................	...............................	 35 	
Figure 16: Proportion of after	-tax income of offenders and the remaining population* from 2000 to 2019, by 	
main categories of offences and year	 ................................	................................	................................	...........	 36 	
Figure 17: Proportion of after	-tax income of individuals charged with money laundering or receiving proceeds 	
from crime and the remaining population* from 2000 to 2019, by year	 ................................	......................	 37 	
Figure 18: Proportion of offenders and the remaining population working in the private, municipal and state 	
sectors* from 2000 to 2019, by main categories of offences and year	 ................................	.........................	 38 	
Figure 19: Proportion of individuals charged with money laundering or receiving proceeds from crime and the 	
remaining population working in the private, municipal and state sector* from 2000 to 2010, by year	 ......	 39 	
Figure 20: Proportion of industries of employed offenders and remaining population* from 2000 to 2019, by 	
main categories of offences	 ................................	................................	................................	..........................	 41

Figure 21: Proportion of industries of employed individuals charged with money laundering or receiving 	
proceeds from crime and remaining population* from 2000 	to 2019, by year	 ................................	.............	 42 	
Figure 22: Proportion of industries of self	-employed offenders and remaining population* from 2000 to 2019, by 	
main categories of offences and 	year	 ................................	................................	................................	...........	 44 	
Figure 23: Proportion of industries of self	-employed individuals charged with money laundering and receiving 	
proceeds from crime* from 2000 to 2019, by mai	n categories of offences and year	 ................................	... 45 	
 
 
 
 
LIST OF TABLES	 	
Table 1: Logit estimation of average 	marginal effects of being charged with main categories of offences	 ........	 51 	
Table 2: Logit estimation	 of average marginal effects of being charged with money laundering or receiving 	
proceeds from crime	 ................................	................................	................................	................................	.....	 56

1 	
 	
1.	 INTRODUCTIO	N	 	
Money laundering is the fuel that drives organized crime and the illegal economy. 	Developing 	
knowledge o	f the profile of money launderers	 is of great relevance for both policymakers and 	
financial actors working agai	nst money laundering. The 	Norwegian 	government has 	in recent	 	
years increased 	its	 focus on preventing money laundering by	 developing the	 Anti	-Money 	
Laundering Act	 further and increasing financial actors' responsibility to prevent money 	
laundering	. (Endringer i 	regler på hvitvaskingsområdet	, 2021)	 The process	 of 	generating	 	
money issued from criminal activity	 and securing it 	has huge consequences	 on 	society, the 	
economy,	 and 	the political system	. It	 suppresses economic growth by weakening crime control 	
and distort	ing	 market mechanisms,	 besides that money laundering also	 threatens 	democratic 	
transition and consolidation	 (Jojarth, 2013)	. 	
To	 develop even more appropriate law	s and make the financial actors' work more efficient, it 	
is 	crucial	 to 	have information on the pr	ofile 	of 	a money 	launderer	. In particular, 	developing	 	
knowledge o	f the associated 	crim	inal activit	ies	 that 	trigger	 the process	 of money laundering	 is 	
important	 for 	lawmakers	, as it 	indicates	 the 	type	s of crime 	they	 should focus on 	when	 the aim 	
is to reduce money laundering.	 Assessing 	the level of 	risk 	associated with 	certain individuals 	
or firms might also help 	the Financial Intelligence Unit	 (FIU) 	to priorit	ise 	which reports on 	
suspicious activities received 	from	 financial actors they should 	analyse first.	 	
Th	e thesis aims to	 investigate	 the 	socio	demographic and 	labour 	market 	characteris	tics that are 	
associated with 	a person’s probability to	 launder money	 and receive proceeds from crime	 in 	
Norway	. The 	socio	demographic factors include individuals’ gender, age, immigration status, 	
place of residence and education level, 	while	 the labour 	market 	characteristics comprise 	
information on 	whether one is	 unemployed, 	employed,	 or self	-employed	, whether one is 	
working in the private	, municipal or state sector, and in which industry one is working in.	 The 	
analysis is 	relying on da	ta from the Norwegian Police Directorate on	 suspected offenders f	rom 	
2000 to 2019	, which is linked to sociodemographic data from Statistics Norway.	 	
In order to have a broader understanding of the extent of criminality in Norway and the 
characteristics of offenders in general, 	a first analysis will study how 	individuals’	 characteristics 	
are associated with the probability of being suspected of main categories of offences, 	such as 	
drug	, economic, violent and other offences.	 For this purpose, 	a logit binary choice model	 will 	
estim	ate the probability of being a suspected criminal, by the type of crime. 	Next, the same

2 	
 
analysis	 will be performed on more specifically money laundering and receiving 	profits from 	
offences as the variable of interest.	 The probability of being charged with 	money 	laundering 	
will further be	 decomposed 	depending on possible	 co	-offences	 committed by the charged 	
individuals, including no other co	-offences committed as one of the categories.	 	
The 	first section 	introduces the	 institutional background of the thesis	, followed 	by 	a section 	
presenting the data and the methodology. Thereafter, th	e results of 	the study will be presented, 	
to finally end with some concluding remarks.	 	
 	
2.	 INSTITUTIONAL BACKGROUND	 	
2.1 	Definition	, implication	 and 	process	 of money laundering and receiving proceeds from 	
crime	 	
Understanding the mechanism of 	money laundering is crucial 	when analysing the 	
characteristics of money launderers. 	Th	is practice	 can	 be defined as	 the process by which 	
individuals try to	 transform money 	arising	 from illegal sources into	 legitimate 	funds	. Criminals	 	
create profits from	 any	 illegal activity, such as	 drug trafficking, 	undeclared work or 	fraud	, and 	
launder these profits to make 	those 	appear as 	originating	 from legal sources. 	The aim is to 	
secure the proceeds originating from an offence. 	Once the money is laundered, criminal	s can 	
invest these illici	t funds into legitimate 	projects	. (Singh & Best, 2019	, 1	) One can launder 	
money 	on behalf o	f other	s or for 	oneself	 to secure own earnings originating from an offence	. 	
(Iversen, 2021)	 	
The	 other main offence of interest occurs when a person	 receives part	s of the p	roceeds of a 	
criminal act.	 As for money laundering, the profits	 can be obtained for the person himself or 	
other parts	 and	 they	 can be objects or 	money. 	(Riksadvokaten, 2018)	 Some examples include 	
obtaining profits generated	 from 	stolen goods,	 drug sales,	 fraud,	 corruption, 	tax evasion, 	
embezzlement, etc.	 This act occurs often 	before	 the process of	 money laundering.	 These two 	
offences are 	strongly 	linked to each othe	r. 	The main difference 	is that	 money laundering aim	s 	
to conserv	e and protect	 the proceeds, while the other offence	’s purpose is to 	acquir	e the 	
proceeds, but in both cases	, the profits are	 initially	 generated by a criminal act.	 	
The choice of analysing	 these two offences together is 	due to their strong connexion	. 	
Individuals are often 	charged with	 money laundering 	and receiving profits from crime at the 	
same time. A 	recent	 exampl	e is the case of a 	31	-year	-old	 man	 in Norway	 who has been judged

3 	
 
for money lau	ndering and receiving profits from 	an offence	 representing a 	value of 	7,3 million 	
NOK	. The funds originated in support schemes introduced during the COVID	-19 pandemic.	 	
(Dømt Til Fengsel for Heleri Og Hvitvasking	, 2022)	 Another e	vidence of the strong 	similarities 	
between the two offences is that in the older version of the 	Norwegian Penal Code	 1902	, money 	
laundering and receiving proceeds from crime were	 from 1993 to 2006	 treated together under 	
one common section	 of the code.	 (Straffeloven, 1902)	 The reason for not including any	 other	 	
offence	s that might be related to money launderin	g, such as fraud, tax evasion or	 corruption, is 	
that 	the variable 	of interest 	would	 in this case	 be 	too 	broad 	since the offender profile might vary 	
among these different offences. Moreover, 	it would be difficult to set a limit on	 which offence 	
should be included or not, and to 	truly	 know which offences are 	associated with money 	
laundering. Therefore, including	 the offence of receiving proceeds	 from crime	 seems to be a 	
reasonable choice for the 	aim of 	the	 thesis, 	considering	 the generality of this offence	. 	
Concerning	 the benefits of these offences, l	aundering money	 or receiving	 funds from crime	 is 	
valuable	 for those who	 commit an offence	 because it allows them to	 receive 	a reward f	or this 	
offence.	 If an offender cannot receive economic compensation f	or committing an offence when 	
this is the aim, there 	are	 no incentives for th	e person to commit a crime	. The process of money 	
laundering is essential for the existence of profit	-motivated offences as it 	allows criminals to 	
receive economic rewards from the offence	 and to conserve th	ese illicit	 funds	. When the 	
compensation is a high amount of money, it is even	 more important to be able to protect the 	
funds to keep them	 without being caught	 because the gain	s of succeeding 	are 	very high	. (Heleri	, 	
2020)	 	
However, 	society is paying the price for the occurrence of these offences.	 Money laundering is 	
affecting the economy negatively 	as it moderates economic growth by weakening the control 	
of crime	. Laundering illicit funds is 	an activity that 	represent	s an	 important part of the economy, 	
but as it is illegal it is not part of the ec	onomic statistics	. The 	International 	M	onetar	y Fund 	has 	
estimat	ed	 that the t	otal level of money laundering 	represents between 2 and 5 % of the world’s 	
annual gross domestic produc	t. (Singh & Best, 2019)	 Many criminals remain undetected 	
because they launder money. 	Since some resources are 	used to generate illegal profits instead 	
of being allocated to legal activities, society is 	losing from it as it implies that the state is 	losing 	
some revenues which migh	t have enabled provid	ing 	more goods, services and aid to the 	
population	. Other consequences are that 	some jobs 	will be	 lost because companies move to 	
safer place	s and some investments are stopped.

4 	
 
Economic growth is also suppressed 	because money laundering	 distort	s market mechanisms.	 	
Instead of investing in activities yielding 	the 	highest profits	 which makes the market effective	, 	
criminals invest in business	es in which the 	probability of being caught is lowest.	 It also affects 	
the competition within a sector	 as it allows	 one 	to for instance	 evade 	tax and 	offer	 a lower price 	
than the competitor who	 is operating legally and	 paying	 taxes.	 Money laundering 	has	 also some 	
political impacts on society	. An example is that 	a high level of crime might threaten the 	
legitimacy of the government since this one is failing in providing enough security among the 
population. 	 (Jojarth, 2013)	 	
Regarding the methods used to launder illicit earning	s, there are many of them 	but a	ccording to 	
Singh et al. (2019),	 these methods have in common that they are compos	ed of three steps.	 	
During the first stage, criminals are 	depositing illegal funds into the financial system	. This can 	
be done in	 several 	ways, but one approach is to divide a large amount of cash by depositing 	a 	
smaller amount of cash directly in a bank account. Another way	 is to	 use cash to 	buy 	products 	
that banks are offering such as credit card	s, checks 	or money orders, 	and 	deposit the money	 in 	
other accounts.	 The second stage is to layer the money	 and distanc	e the money from its sources 	
by making a certain 	number	 of transactions	 across different accounts or by selling and buying 	
different bank products	. The last stage is to invest the money into	 legitimate projects. 	This is 	
often done by investing in for instan	ce luxury 	goods, real estate 	or any kind of business. 	(Singh 	
& Best, 2019)	 	
Criminals’	 level of 	organization	 might also affect which method is used to launder money. 	
Organized criminals might for example not apply the same methods as low	-level criminals. 	
Some money laundering techniques are	 moreover	 easier 	to carry out than other	s. One might for 	
instance 	deposit the proceeds 	in the bank account of 	friends or family to thereafter withdraw 	
the money in cash. The proceeds might also be used to buy expensive 	consumer goods	, the 	
money can be 	transferr	ed	 abroad or	 smuggl	ed	 abroad	 in cash.	 Other methods are more 	
sophisticated.	 Proceeds might be incorporated into a business 	by for instance generating false 	
sales or by	 using the proceeds to cover	 the 	costs of the business 	without 	recording these costs 	
in the 	accounts of the firm. 	Th	ese funds 	will thereafter 	be withdrawn from the business and will 	
appear as being legitimate in the 	economy.	 (Iversen, 2021)	 	
The most 	complicated money laundering methods are often carried out by professionals. 	The 	
money is often transferred through different 	institutions, persons or businesses, which makes it 	
difficult to 	map the	 path of the transfer of money and identify the origin of 	the proceeds.	 This 	
can for instance be done by hiding the ownership of the proceeds by using tax haven companies.

5 	
 
Selling and buying cryptocurrency	 is also a new method which allows 	the transfer of money to 	
not be trac	eable, as 	traditional bank 	transfers would be. It makes it more difficult to know the 	
origin of the funds.	 (Iversen, 2021)	 The	 different 	methods are also changing with time an	d 	
depending on technical advances. 	The d	evelopment of cryptocurrency 	and 	the decreasing use 	
of cash in Norway 	might	 also	 influence	 the way criminals are laundering money	, as well as the 	
legislative context might do.	 	
 
2.2 	The legislative context	 	
Money laundering 	and 	receiving proceeds from crime are offences that are regulated by the 	
Norwegian Penal Code. 	Money laundering is moreover 	regulated 	by 	the Anti	-Money 	
Laundering Act	 introduced in	 2003	. These laws are necessary 	to avoid suffering from the	 	
earlier	-mentioned economic and political consequences o	f this criminality.	 	
2.2.1	 The Penal Code	 	
The Penal Code 1902	 of Norway	 had initially 	separate provisions for 	receiving proceeds from 	
crime	 (§ 317	) and for money laundering	 which 	was	 at that time called “subsequent 	assistance”	 	
(“etterfølgende bistand” in Norwegian	, § 320)	 (Straffeloven, 1902)	. In 1993, these two offences 	
were 	unified	 into one common provision §	317	, until	 it was sep	arated again	 when the new Penal 	
Code 2005 w	as introduced	 in 2006	, in which provisions 	§§ 332 	– 226 concern receiving 	
proceeds from crime and §§	 337 	– 341 	are related to money laundering	. (The Penal Code, 2005)	 	
According to the Penal Code 2005	 § 337 a	, one 	shall	 be punished	 for money laundering if 	this 	
person	 assists 	“in safeguarding the proceeds of a criminal act for another person by, for 	
example, collecting, storing, concealing, transporting, sending, transferring, converting, 
disposing of, pawning or 	investing them	” (The Penal Code, 2005)	. This statement 	is composed 	
of some examples of how one can 	help 	mak	e sure 	that 	the profits from a criminal act	 will 	remain 	
in the hands of the criminal	, but	 the enforcement applies to 	any 	other methods tha	t are not 	
mentioned	. The offence that generated proceeds is considered as being the primary offence.	 	
Moreover, 	the enforcement applies 	regardless of whether 	the act of 	money laundering occurred 	
before, after or at the same time as the	 primary offence	. 	
The Penal Code 2005	 § 337	 b 	specifies the 	case	 of “self	-washing”	 which was introduced in 	
2006	 and claims that one sh	all	 be punished i	n the case of “	converting or transferring assets or 	
by other means conceals or obscures where the proceeds of a criminal act he/she has personally 
committed are located or originate from, who controls them, their movements or rights

6 	
 
associated with them	” (The Penal Code, 2005)	. The	 main difference	 from the first part of	 § 337	 	
is th	at in this situation, the person	 laundering money	 has committed the 	offence which 	led	 to 	
profits.	 However, 	the enforcement does not require that	 the offender	 end	s up with self	-	
laundered 	earnings	. Moreover,	 the offender of the primary offence shall be punished 	for self	-	
washing if 	this one	 launders 	his 	profits	, but also if someone else launders this money for him. 	
(Riksadvokaten, 	2018)	 	
Furthermore, a	ccording to the Penal Code 	2005 	§332	, one shall be punished if 	this	 person 	
receives a part of the proceeds of a criminal act.	 It applies if the proceeds are received 	for	 the 	
person himself 	and	 if this person 	procured	 the proceeds for 	another subject	. However, if 	this 	
other 	subject	 is the offender of the primary offence	, the offence is instead punished as money 	
laundering.	 An example of this situation is that person 	A sells drugs, but person 	B acquire	s the 	
profits from 	the sale	 which 	are thus transferred to person B.	 In this case, person B will be 	
sanctioned for money laundering.	 	
A requirement i	n §332 is	 that the proceeds must 	have come into the offender’s possession or 	
transferred to 	a bank 	account controlled by the offender	, but the offender does not need to	 own	, 	
use or pledge the profits in the current version of the law.	 Moreover,	 the provision 	applies 	if 	
obtaining proceeds occurs at the same time as the primary offence, and	 if the proceed	s are partly 	
obtained before the primary offence	 by for instance 	being paid in advance	. 	
Common for 	provisions 	§§ 332 and 227 	is that	 profits are not exclusively economic, as the 	
penal code also includes objects, claims or services in their 	definition of proceeds	. Moreover, 	
one shall be punished for money laundering 	or receiving proceeds from 	a crime 	even though 	
there is no one to blame 	for the primary offence.	 	
The penalty for individuals is the same for bo	th offences and varies with the gra	vity of the 	
offence.	 The penalty for the violation of §§ 332 and 337 is up to 2 years	 of imprisonment	. 	
However, if these offences are related to severe robbery, 	severe 	human trafficking or 	severe 	
drug offence	, the penalty can be up to 15 years of imprisonment. 	If receiving proceeds from 	
crime	 and money laundering	 are considered as being severe offences, th	e penalty can be up to	 	
6 years of imprisonment	, cf. 	§§ 33	3 and 33	8. Whether these offences 	are	 considered 	severe 	
depends on	 which offence the proceeds 	originate 	from	, whether the value of the proceeds is 	
high and whether the offender has laundered money or received proceeds from crime 	regularly	. 	
If the profits originate from narcotic drugs offence, one should put weight on the type of drugs 
and the quantity of substance related to the profits.

7 	
 
Minor	 offences in 	§§ 334 and 339 	are	 penalised w	ith imprisonment of up to 2 years.	 Offences 	
are consider	ed minor when the	 offence from which the profits originate is less severe, the value 	
of the profits is lower, and the circumstances of the act suggest a lower degree of guilt	. 	
Moreover, negligent offences are according to §§ 335 and 340 	punished with	 a fine or 	
imprisonment of a period which does not exceed two years.	 Finally, the case in which t	wo 	
persons or more agree on committing money laundering or receiving proceeds from crime is 
penalised with	 imprisonment of up to 2 years.	 	
In some cases, it 	is cha	llenging to assess whether the offence is money laundering or receiving 	
proceeds from a crime. 	According to the guidelines of the Penal Code, 	one shall be punished 	
for receiving proceeds from an offence and not for money laundering, if there 	is no clear 	proof	 	
that the offender is 	protecting	 the proceeds for someone else. 	(Riksadvokaten, 2018)	 	
Individuals and enterprises can 	both 	be 	penalized	 for money laundering	 and receiving proceeds 	
from crime	, cf. § 27	 claiming that 	“when a 	penal provision is violated by a person who has 	
acted on behalf of an enterprise, the enterprise is liable to punishment	. This applies even if no 	
single person meets the culpability or the accountability requirement	”, cf. § 20	 (The Penal Code, 	
2005)	. I	t implies	 that a	 firm 	can be reprimanded 	when someone 	launders money or obtains 	
proceeds from a crime	 on behalf of the 	firm	. An enterprise is defined as	 being a “company, co	-	
operative society, association or other organisation, sole proprietorship, foundation, estate or 
public body”	 in § 27	 (The Penal Code, 2005)	. 	
The penalty 	for	 enterprises 	takes the form of a fine, but enterprises can also	 lose the	ir right to 	
operate	 and might also be subject to 	confiscation	. The Court can	, however,	 evaluate whether an 	
enterprise will be penalized based 	on different factors including	, for example	, the 	severity	 of 	
the offence,	 the preventi	ve effect of the penalty	 or the 	financial capacity of the enterprise	, cf. § 	
28 a, b and f. 	In the case of money laundering,	 the Court might take into consideration the firm’s 	
compliance 	system or preventive measures.	 	
 
2.2.1	 Anti	-Money Laundering Act	 	
In the process of money laundering, bank systems and other legitimate business activities are 
often 	misused,	 and this is why these institutions are regulated. 	The 	purpose	 of the Anti	-Money 	
Laundering Act is	 to “prevent and detect money laundering and terrorist financing”	 according 	
to 	the current 	version of the Act 	§ 1 a	 (The Anti	-Money Laundering Act, 2018)	. Terrorist 	
financing 	includes 	“financing of terrorist acts, terrorist organisations or individual terro	rists	”

8 	
 
(Veileder til hvitvaskingsloven	, 2022)	. The Act	 shall also “protect th	e financial and economic 	
system, as well as society as a whole, by preventing and detecting the use or attempted use of 
obliged entities for purposes of money laundering or terrorist 	financing”	, cf	. § 1 	b (The Anti	-	
Money Laund	ering Act, 2018)	. After the introduction of the Act, a new version has been 	
developed in 2009 and 2018	. 	
All the entities that are subject to the	 Act 	are required to prevent and detect 	activity that is	 	
suspicious and 	related to profits originat	ing	 from an offence and	 activity	 related to terrorist 	
financing, according to	 the Financial Supervisory Authority’s guidance on the Act.	 If 	such 	
activities were to be detected, they must be reported to 	the Norwegian Financial Intelligence 	
Unit (FIU). 	The 	Act	 was founded based	 on the 	recommendations of the 	Financial Action Task 	
Force (FATF) 	which is an intergovernmental organisation w	hich work	s with preventing money 	
laundering 	and the E	uropean 	Union’s Fourth Money Laundering Directive.	 	
The Act applies to finan	cial i	nstitutions that might 	be used during the process of money 	
laundering	, including	 “a. banks;	 b. credit institutions;	 c. financing institutions;	 d. the central 	
bank of Norway;	 e. e-money institutions;	 f. undertakings engaged in foreign exchange 	
activities;	 g. payment service undertakings and others entitled to provide payment services;	 h. 	
investment firms;	 i. management companies for securities funds;	 j. insurance undertakings;	 k. 	
undertakings engaged in insurance mediation that is not 	reinsurance broking;	 l. central 	
securities depositories, m.	 undertakings engaged in deposit activities;	 n. managers of alternative 	
investment funds;	 o. loan mediation undertakings	”, cf.	 § 4 (The Anti	-Money Laundering Act, 	
2018)	. 	
The Act 	defines	 how 	reporting firms shall work 	when trying to detect money laundering.	 Some 	
of the main elements	 are the following. Reporting entities shall 	establish 	needed measures 	based 	
on an assessment of the risk of money laundering and terrorist financing 	related to	 their	 clients	 	
or projec	ts, § 6	. They must implement internal control to ensure compliance with the 	
requirements of the Act, cf. § 35.	 Moreover,	 concerned 	entities 	are obliged to	 implement 	
customer due diligence measures 	to know their customer and how 	their products are used	 when 	
a new customer relationship is established	, in case of any suspi	cion of money laundering o	r 	
terrorist financing	, cf. § 10.	 If 	the reporting institutions discover 	any element indicating that 	the 	
activity	 is related to money laundering or terror financing, the entity must	 investigate it	, cf. § 	
25	, and 	it must be r	eported to FIU	 if this suspicion 	still exists 	and cannot be	 disproved	 after 	
further investigations, cf. § 26	.

9 	
 
The reporting 	corporations	 shall be sanctioned	 with a 	fine	 in case of	 violation of 	§§ 9, 12, 13, 	
17, 18, 24, 25, 26, 28, 30 or 42 of the Act	, cf. 	§ 51. If a company is penalized based on this 	
provision,	 members of the board, directors of the management or any person in a similar 	
position in this firm can also be 	penalized and must pay a fine, 	if they 	made any actions that are 	
related to this breach in an intentional way or with high negligenc	e, cf. 	§ 51	 a. Individuals that 	
are obligated to report 	can also be penalized on the same basis as above, cf.	 § 51 b. The penalty 	
does also apply to board members, managers, employees and other individuals who	 act on 	
behalf of the corporation	, who violate	 § 26 related to the obligation of reporting	 and	 informing	 	
the FUI and § 28 related to 	the prohibition of	 revealing 	investigations, report	s and inquiries, cf. 	
§ 51	 c. In case of aggravating 	circumstances,	 all the	 persons	 mentioned above	 can	 be	 sent to 	
prison for a maximum of one year	, cf.	 51 a, b and c	. 	
 
2.3 	Related literature	 	
Before analysing what has been done empirically, it is important to have an overview o	f existing 	
theories 	to understand	 the reason wh	y individuals commit 	crimes	 and launder money.	 	
According to Gil	mour (2015), th	e theory of r	ational choice 	is 	the most appropriate theory to 	
explain why individual	s are laundering money. It states	 that offenders commit 	a crime in cases 	
where the benefit	s of it	 are relatively	 higher than 	the	 cost.	 According to this theory focusing on 	
the individual	-level, 	offenders have 	preferences towards	 safe	 crime	s that are easier to commit	 	
and that	 give 	any form 	of recompense, while considering	 how much	 risk and effort 	this act 	
implies.	 Offenders	 are thus	 making rational decisions and	 are	 subject to cost	-benefit analysis	 	
when committing a crime.	 (Gilmour, 2016	, p. 3	) 	
On	 the opposite	, traditional criminology	 is focusing on the social and working environment as 	
being 	the 	reason 	for	 committing crimes	 instead of the individual characteristics 	of offenders. 	
According to the routine activity theory, a person will commit a crime	 if three conditions a	re 	
fulfilled at the same time and 	in 	the same place. 	Firstly, an	 offender must be willing and able to 	
commit a crime. 	Secondly	, there must be 	a “suitable target”, implying	 that	 it is easy for the 	
offender to damage 	the victim	 which might be a person, an obje	ct or a system	. In the case of 	
money laundering,	 this target might be the bank system or the use of cash which makes it 	
possible to launder money.	 Thirdly	, there must be no “suitable guardian”	 which refers to a 	
person preventing another one to commit a crime actively or just by its presence at the right 
time and place.	 (Cohen & Felson, 1979	, p. 589	) In the 	thesis	’s context, a suitable guardian 	might

10	 	
 
be the existence of 	an effective and 	well	-functioning 	legal 	and 	financial	 system	 including agents 	
such as the 	police or laws	, who 	prevent	 individuals to launder money.	 	
On a more general basis, 	two other theories	 that are	 opposed to each other	 have been developed 	
regarding the 	profile	 of white	-collar crimina	ls. 	White	-collar crime refers to a non	-violent 	
offence	 with the aim of	 obtain	ing	 money, property or any advantages. 	(Gottschalk, 2018)	 On 	
the one hand, 	Sutherland affirms that 	the sociodemographic profile of 	white	-collar 	criminals is 	
homogenous	 and specific to them. 	A criminal of this sort is	 a typical “entrepreneur with high 	
education and high or regular income”. On the other hand, the revi	sionists state that the profile 	
of whi	te-collar offenders is rather heterogenous, implying that any offender could commit this 	
sort of offence.	 (Alalehto & Larsson, 2012)	 	
Regarding the empirical research in this field, there are	 many	 analyses made	 on the profile of	 	
criminals 	in general	. The 	findings vary depending on which category of crime i	s concerned. 	A 	
study found	 for instance that 	“sex	, age, employment, marital status, education and occupation 	
increased the 	probability	 of being charged with robbery, burglary, mo	tor vehicle theft, or violent 	
crime”. The study is based on 	reported offences	 in the state o	f Hawaii	. (Hawaii Criminal Justice 	
Information Data Center, 198	2) Moreover, 	there is 	a common agreement in the empirical 	
literature that there is a 	strong 	relationship between 	ag	e and criminal activity.	 Crime tends to 	
increase and peak in th	e early twenties to start declining with age after th	e peak. 	(Hansen, 2003)	 	
It is interesting to compare 	the socio	-demographic characteristics 	of 	conventi	on	al crime 	
offenders 	and	 white	-collar offenders	. Conventional crime 	includes 	murder, rape, robbery, 	
burglary, 	aggravated assault, and auto theft	. A r	esearch 	based on data from the Federal courts 	
in the United States of America 	found that	 conventional crime offenders are often relatively 	
very young compared to white	-collar 	crime offenders	 for who the probability of	 committing a 	
crime is the highest when 	they are middle	-aged or older.	 (Benson & Kerley, 2001)	 Moreover, 	
white	-collar offenders are more likely to be employed and 	to have a higher education level	 	
compared to conventional 	crime	 offenders.	 (Weisburd et al., 2002)	 However, other studies are 	
challenging these findings by stating that 	white	-collar crime offenders are rather belonging to 	
the middle or lower class 	and are entrepreneurs or employees. 	(Daly & Tonry, 1997)	 	
The e	mpirical literature on the	 profile	 of	 money launderers is 	however more 	limited	. Most of 	
these	 papers 	are	 based on the	 reports	 on suspicious financial transactions sent from 	financial 	
institutions that are obligated to report	 to the FIU	, instead of being based on the	 Police’s	 data 	
on conviction. 	The	 paper from	 Ningtyas, 	Ismail	 and	 Wahyudi	 (2022)	 found	 for example	 that

11	 	
 
“the social demographic factors affecting the money laundering risk committed by individuals 
are 	gender	, age, marital status, occupation, and the history of financial service usage, in which 	
occupation has the highest effect	 “ in Indonesia	. (Ningtyas et al., 2022)	 Another research	 based 	
on the data of the Financial Action Task Force of Latin 	America	 states 	that money launderers 	
are mainly foreign males	. (Buele et al., 2019)	 	
The limitation of the empirical research is that there is no 	data 	on criminality 	representing 	a 	
correct picture of reality.	 Research is mostly based on reported criminality or 	surveys, 	but data 	
on true criminality	 are missing	. This is important to have in 	mind when reading th	is thesis as it 	
is based on reported criminality.	 Moreover, s	ome profiles of offenders might be better at not 	
being caught	, which can bias the results	. However, 	making research on the subject can 	still 	give 	
some relevant indication	s and	 information on it.	 	
 	
3.	 D	ATA AND METHODOLOGY	 	
3.1 Data 	Sources	 	
The data from the Norwegian Police	 Directorate	 includes 	information	 on	 individuals who have 	
been	 charged with	 any offence from	 2000 to 2019	 in Norway	. These data are further linked to 	
administrative population	-level 	sociodemographic data from Statistics Norway for the same 	
period	. Th	is data is	 composed of	 the gender, age	, labour market status	, after	-tax income	 and	 	
municipality number	 of the	 domicile	 of individuals. It also includes	 the number of years one 	
has 	spent	 on education	 and 	a variable	 which 	categoris	es individuals according to their	 	
immigration background. Moreover, 	for employed and self	-employed workers, 	it includes	 the 	
coding units	 for industrial classification 	follow	ing 	the European Community	’s standard called 	
NACE	 for	 which each code corresponds to a specific industry.	 A variable 	co	rresponding to 	the	 	
standard for institutional sector grouping	 is also included	 concerning employed and self	-	
employed workers.	 	
In the 	datasets	 of	 criminality and sociodemographic characteristics	, each individual has 	an 	
anonymized personal number which is common for both datasets.	 The two datasets have been 	
merged based on this personal number	, implying that each observation includes information on	 	
its criminal background and sociodemographic characteristics for each year.	 The 	linked dataset	 	
is restricted to Norwegian residents who are between 15 and 75 year	s old each year and is 	
composed of	 72	 995 982 observations	 in total.

12	 	
 
The dataset	 on sociodemographic 	characteristics	 contains	 first of all the	 gender of each 	
observation	. A dummy variable has been generated taking the value of 1 if the observation is a 	
male and 	0 if the person is a female. Secondly, the dataset 	is composed of 	the	 age 	of the 	
observations 	and is restricted to individuals who are between 15 and 75 years old.	 The reason 	
for	 selecting this	 proportion of the population	 is that	 one cannot be 	penalised by the	 police or 	
the court	 if this person	 undertakes a criminal act and 	is under	 the age of	 15	, cf	. Penal Code 2005 	
§ 20. 	Moreover, most of	 the	 individuals older than 75 are not 	working, which might 	bias the 	
analyses	 when 	studying the working sector or the income of individuals	. In 	addition, a very 	
small proportion of this	 population 	is 	charged with	 any offence.	 	
Some changes have been made to the independent variables to fit the model better and increase 
its predictive power.	 The age variable has been scaled to 10 to facilitate 	the 	interpretation of the 	
estimated regression coefficients.	 Moreover, a squared version of th	is variable	 has been added 	
to the regression to capture the non	-linear relationship between these variables a	nd the 	
dependent variable.	 	
Furthermore, 	the dataset is composed of a dummy variable which takes the value of 1 if the 	
person is considered as being an immigrant, and if it is not the case.	 Immigration is defined 	
as including	 all inhabitants who have emigrated from another country to Norway or 	those who 	
are born in Norw	ay	 but	 have immigrant parents.	 This categorization	 does not include persons 	
born in Norway 	having	 one parent born abroad	 or persons 	born	 abroad having one	 or tw	o 	
Norwegian paren	ts.	 	
The identification number of municipalities 	of the domicile of observations 	is also part of the 	
dataset.	 A 	binary	 variable	 indicating 	whether a person lives in a city or not 	has been created 	
based on 	the size of their 	respective municipality.	 A municipality 	has been defined as a city if 	
it compri	ses more	 than 5	 000 inhabitants	. The variable is based on the number of inhabitants 	
and the municipal composition in 2019.	 Changes 	in the	 municipal 	structure	 in Norway 	in 2017 	
and 2018 have been taken into account	 when generating the dummy variable.	 	
The level of education is also included in the dataset	 as a variable 	indicat	ing the number of 	
years one has studied,	 starting from 	primary school. 	This 	variable takes the value of if a person 	
has no education level. Expect for 0, t	he lowest level of education observed is of 	6 year	s and 	
the 	highest is of 23 years.	 3 999	 816 observations are missing data on education, and this is 	
mainly due to missing education data for a subset of immigrants. To keep these observations to 
not create any bias in the analysis results and control for it, a new indicator has been created in

13	 	
 
the dataset which takes the value of if the data on education is missing and of 1 if not. The 
missing data is replaced by “	-1” in the original education variable.	 	
Next, 	the dataset 	contains the 	after	-tax income	 of individuals	, which corresponds to 	the tot	al 	
income after deducting 	the applicable taxes and negative transfers,	 such as pension 	
contribution	s and 	child support paid under the public scheme. 	The 	after	-tax income	 has been 	
adjusted for inflation 	based on the 	Consumer 	Price 	Index of 20	19.	 This 	variable has been scaled 	
to 100	 000	 to facilitate 	the 	interpretation of the	 estimated regression coefficients.	 This method 	
has been chosen instead of taking the logarithm of income because 	this	 variable includes zeros 	
and negative values. 	A squared version of the variable of 	after	-tax income	 has	 also	 been added 	
to the regression to capture the non	-linear relationsh	ip between these variables and the 	
dependent variable.	 Missing values on 	after	-tax income	, representing 674 233 observations, are 	
set equal to 0. Since the proportion of missing data on 	after	-tax income	 is not very high relative 	
to missing data on for inst	ance education, it is not considered necessary to control for that.	 	
The dataset also contains 	Statistics Norway’s c	ategorization of labour market status. 	This code 	
list is composed of “0” if a person is 	out of the labour force, “1” if employed in a 	firm, “2” if 	
self	-employed	”, “3” if unemployed and “4” 	if participatin	g in active labo	ur market programs	. 	
Individuals out of the labour 	force represent on average 31 % of the sampl	e, while employed 	
observations represent	 63.2 % and self	-employed individual	s represent	 4 % of the sample. 	
Unemployed persons represent 1.4 % 	and individuals participating in active 	labour market 	
programs 	represent	 0.1 % of the sample.	 	
In the dataset, 	three different dummy variables have been create	d according to 	observations’	 	
labour market status	 based on the above list	. One	 variable 	corresponds to	 individuals who are 	
employed	 in a firm	, the 	second	 applies to	 observation	s who	 are self	-employed	 and t	he third 	
concerns 	person	s who are 	unemployed	 or participating in active labour market programs	. The 	
basis group corresponds to 	the	 remaining	 group of persons who are out of the labour market	. 	
The three last	-mentioned variables take the value	 of 1 if a	n individual belongs t	o the 	mentioned 	
labour mark	et 	groups	 and it takes the value	 of if an observation belongs to the basis group	 	
which has been generated to	 avoid the dummy variable trap	. 	
Moreover, the dataset is describing which sector	 individuals are working in if they are active.	 	
Sectors are 	initially 	defined 	based on	 the	 sectoral grouping of Statistics Norway. Observations 	
from 2000 to 2012 	are based on 	version 	1987 of this grouping and th	ose from 2013 to 2019 are 	
based on 	version	 2012.	 The changes in versions over time have been taken into account when

14	 	
 
generating variables.	 In the dataset, 	three different	 dummy	 variables have been defined 	
depending on whether 	an	 active person works in the private, 	municipal,	 or 	state sector	. 	
Next,	 the dataset contains the NACE codes of the firms	 in which individuals are working	, which 	
has led to the formation of 6 	main	 industries. 	The first one includes 	businesses related	 to 	
electricity, gas, steam 	and 	water supply,	 sewerage and waste management, 	and	 construction	 	
activities.	 The second industry 	corresponds to accommodation and food service activities.	 The 	
third one	 belongs to activities	 related to	 producing new products from natural resources, also 	
named 	manufacturing	, whole	sale and retail trade,	 in addition to	 the 	repair of motor vehicles, 	
transport and storage.	 	
The fourth industry is related	 to real estate, financial and insurance activities.	 The fifth	 one is 	
associated with	 all sorts of 	other 	services	 that 	do not belong to any already mentioned 	
categories,	 such as scientific and technical services which require 	higher education,	 health and 	
social work activities,	 education,	 public administration, defence, 	and 	social security schemes 	
under the responsibility of the government	. This category also includes	 business services,	 	
cultura	l and	 enter	tainment activities	, information and 	com	munication	, paid work in private 	
households	 and other services	. Finally, the last category is compos	ed of the remaining industries 	
such as 	agriculture, forestry, fisheries, mining, quarrying of oil and gas, and international 	
organisations.	 	
Some changes hav	e been made to the labour market variables for the purpose of the model 	
estimations. The unemployment variable 	is included 	per se 	in the regression, but in 	the case of 	
a person being employed, 	two	 dummy variable	s have been created depending on if working i	n 	
the state 	or 	municipal sector. According to our data, money laundering and proceeds from crime 	
are mostly committed by individuals working in the private sector. For that reason, being 
employed in the private sector has been linked to the different 	industrial categorie	s, generating 	
a dummy variable for each industrial grouping. Such dummy variables have also been generated 
for self	-employed persons in the private sector. This is only performed for 	individuals	 working 	
in the private sector as most self	-employed individua	ls belong to the private sector based on the 	
information in the dataset.	 	
Next, the 	data on criminality include the registration date of the offence, the start and the end 	
date of the offence, which category of criminality the offence belongs to, and a more	 specific 	
description of the nature of the offence.	 Based on this information, dummy variables have been

15	 	
 
generated taking the value of	 1 if an observation has been charged with 	a given	 offence in 	
Norway at least once during a year	. 	
The criminality data is 	based 	on whether one has been charged 	with	 an offence, not if this 	
person has truly been judged and condemned for it.	 Although a	 charge 	does not necessarily 	
imply that a person will be judged for the offence 	this one	 is 	charged with	, it is still relevant to 	
use data on 	charges	 as it 	indicates	 the extent of criminality. Moreover, some cases are suspended 	
due to insufficient resources	 and	 ev	idence, or because 	a case is of minor gravity relative to other 	
cases. 	There is a lack of data on criminality as there is a lot of unreported criminality, which is 	
why it might be beneficial to 	use data on 	charges	 instead of judgement as it 	enables collecting	 	
more observations	 and cover	ing	 a larger extent of criminality.	 	
The different 	categories 	of 	offence	s of Statistics Norway	 have been 	regrouped into 	4 main 	
categories.	 One category includes 	drug	 offences	, referring to a person	 who	 “unl	aw	fully	 	
produces, imports, exports, acquires, stores sends or supplies a substance” that is considered a 
narcotic drug, 	according to	 the 	Pe	nal Code	 2005 § 231	 (The Penal Code, 2005)	. The next 	
grouping is including all economic	, profit	-motivated 	and working environment 	offences	. To 	
name some examples, this categor	y includes offences related to 	money laundering,	 fraud, theft, 	
corruption, tax fraud, 	and 	data breach of financial infidelity.	 Another category is composed of 	
all 	violence	-related 	offences	, such	 as murder, rape, 	physical assault, 	torture, human trafficking	 	
or purchase of sexual 	services. The last category includes all other sort	s of offences which do 	
not 	belong to	 any of the three first categories of offences. It comprises 	environmental	 crime, 	
traff	ic offences, criminal damage and other 	offences.	 	
In addition to these main categories, a dummy variable has been generated indicating whether 
one has been charged with any offence. Another mo	re specific 	dummy variable 	indicates	 	
whether one has been charged with money laundering or proceeds from crime.	 These last 	
charges 	correspond to breaches of the 	Anti	-Money Laundering Act and	 the 	relevant sections of 	
the Penal Code mentioned in section 2.2.1.	 All the mentioned variables h	ave 	led to	 the 	
formation of a	 panel dataset	 used in the model estimations.	 	
 
3.2 Methodology	 	
This section will first specify the	 logi	t binary choice models 	estimating	 the association between 	
sociodemographic and labour market characteristics of individuals 	and their probability to be 	
charged with the main categories of offence, including money laundering and receiving

16	 	
 
proceeds from crime.	 The second section will descri	be how 	the probability of 	being charged 	
with money 	laundering	 is decomposed 	depending on possible	 co	-offences	 committed by the 	
charged individuals, including no other co	-offences committed as 	one of the	 categor	ies.	 	
3.2.1 Logit Model for 	Money	 Laundering and Other Offences	 	
Log	it binary choice models are estimated 	to measure the associations 	between individuals’ 	
characteristics and their chances of committing a crime. 	This method has been	 chosen because 	
this paper aims to estimate the probabilit	y of an event which has a binary outcome and is limited 	
between and 1, while estimations based on a linear probability model could sometimes result 
in predicted probabilities that lie outside this interval	. 	
It is important to first get an overall overvie	w 	of the associations betwee	n individuals’ 	
characteristics and the probability of being charged with the main categories of crime to be able	 	
to compare in better way	 the relationship between these factors and money laundering.	 A first 	
logit binary choice	 mode	l will	 therefore	 be 	estimated in which the dependent variables are 	
binary 	and depend	 on whether	 one is charged with any offence, economic offences, drug 	
offences,	 violent offences	 or other offences. In the	 second 	model	, the dependent variable is 	
related to whether one 	is charged with	 money laundering or receiving proceeds from 	a crime	 or 	
not	. The dependent variables take the value of 1 or depending on whether an individual has 	
chosen to 	commit	 the mentioned offences or not.	 	
The logit	 binary choice	 model is buil	t on an underlying latent model.	 When a person	 is choosing 	
whether to commit an 	offence or not, this person	 is evaluating the relative utility of each of the 	
outcomes,	 and it is assumed that individuals are maximizing their utility when mak	ing the	ir 	
decision. 	More precisely, the difference between the utilit	y of 	an individual 	������ of 	co	mmitting a	 	
crime 	������ or not	 is expressed 	in the	 variable 	�������������∗. This utility	 difference is not 	observed,	 and 	the 	
variable is	 therefore	 called latent	. The difference in utility between these two outcomes depends 	
on 	individuals’	 characteristics	 which are observed in the dataset,	 expressed in the vector	 �������′, 	
and on unobserved characteristics, expressed in 	������������������. This relationship is 	illustrated	 in equation 	
(1)	. 	
�������������∗=	 �������′∗������������+	������������������ 	 (1)	 	
The vector of observed characteristics is composed of individuals’	 age, squared age, immigrant 	
status, city dweller status, education 	years	, 	after	-tax 	income, squared income, 	and 	
unemployment status	. Binary variables indicating whether one is 	employed in the municipal 	
sector or whether one is employed in the municipal sec	tor are also included.	 For those working

17	 	
 
in the private sector, 	binary 	variables are also included depending on their industrial belonging 	
and whether they are employed or self	-employed.	 	
The model assumes that an individual	 will commit a crime if the util	ity difference is above a 	
certain threshold	 corresponding to zero.	 ������������� is observed 	in the data 	and takes the value of 1 if 	
an individual 	has been	 charged with a	 crime	 and	 takes the value of if not.	 The assumption	 	
implies that an offence will be executed	, implying	 that	 �������������	=	1, only	 if 	�������������∗>	0. If this condition 	
is not fulfilled	 for an individual, this one will not commit a crime,	 meaning that	 �������������	=	0. 	
Equation (2) express	es the consequences of this assumption 	and 	the 	relationship between 	the 	
observed outcome in the data an	d the unobserved utility of committing a crime	: 	
�������������	=	1 →	 �������������∗>	0 →	 �������′∗������������+	������������������	>	0 →	 −������������������	≤	�������′∗������������ 	(2)	 	
The 	variable	 ������������������ is 	regarded as being a stochastic variable	, leading to the following expression:	 	
Pr	(�������������	=	1)=	Pr	(�������������∗>	0)=	Pr	(�������′∗������������+	������������������	>	0)=	Pr	(−������������������	≤	�������′∗������������) (3) 	
It is further assumed that 	������������������ is logistically distributed	, leading to the following 	function	 	
expressing the probability 	������������������ of an individual 	������ to be charged with a	 crime 	������: 	
������������������	=	������������������′∗������������	
1+	������������������′∗������������ (4) 	
������������������ is 	moreover	 assumed to be in	dependently and identically distributed across individuals	 ������. 	
This implies that unobserved factors ar	e not correlated 	with observed factors	 across individuals	. 	
The variance of these	 unobserved	 characteristics is the same for everyone.	 As a consequence of 	
these assumptions	, the parameter 	������������ can be estimated with 	the 	help of the method of maximum 	
likelihood.	 This parameter 	estimates the effect of individuals’ characteristics on the probability 	
of committing 	a crime 	������. (Verbeek, 2017)	 	
To facilitate the interpretation of the results, the average marginal effect	s of any individual 	
characteristics	 on the probability of being charged with an offence	 will be estimated. This is 	
estimated by c	omputing 	the	 average	 change in probability of being charged with an offence 	
when the value of an independent variable change	s by one unit	, holding all the other variables 	
constant.	 These	 average marginal	 effects are also estimated in the next analysis.

18	 	
 
3.2.2 Decomposing the Probability of Model Laundering By Co	-offences	 	
In this second analysis, the aim	 is 	to find whether 	committing 	money laundering and receiving 	
proceeds from crime 	is associated with other categories of offences such as 	drug	, profit	-	
motivated, violent and other crimes	. The 	probability of being charged with	 money laundering	 	
and receiving proceeds from crime has theref	ore been decomposed by co	-offences	, including 	
no other co	-offences as one of the categories	. In other words,	 16 new models have been 	
generated in which	 each of	 the dependent variable	s describes 	a possible combination of 	
offences that have been committed in addition to be	ing	 charged with money laundering	. An 	
example of such a combination is that one person is 	charged with	 money laundering or receiving 	
proceeds from a crime and has also committed an	other profit	-motivated offence but no traffic, 	
violen	t or other offences.	 	
Appendix	 1 illustrates all the possible 	combinations of outcomes. 	The 	16 combinations of 	
offences are mutually exclusive, implying that when one observation satisfies the 	condition of 	
belonging to one alternative of offence, it cannot belong to any other group.	 A person who has 	
been charged with money laundering must	 belong to one of the offence alternatives and 	this 	
person can only belong to one of the alternatives 	for	 a year.	 Given that 	�������������	 represents 	the 	
observed 	binary outcome of an individual	 ������ committing	 the offence 	������	 of laundering or 	
receiving proceeds from 	a crime	 or not	, this variable must be equal to	 the sum o	f all the binary 	
variables 	indicating 	each	 alternative 	������ of 	combination	s of additional offences	, including no 	
other co	-offences as one of the categories	. This equality 	is expressed in the next equation	. 	
 �������������	=	∑	������������� (5) 	
In order to estimate the probability of choosing one 	alternative 	of the possible combinations, I 	
estimate separate 	logic binary choice 	models 	following the	 same assumptions as presented	 in 	
the last sub	-section. 	Specifically, t	he probability 	������������������ of an	 individual 	������ being charged with 	
alternative 	������ of combinations 	of crimes 	is now expressed	 as 	follows.	 	
 ������������������	=	������������������′∗������������	
1+	������������������′∗������������ (6) 	
The models include some weaknesses as the dataset is very large in terms of observations and 
variables	, besides that 	the outcomes are quite rare compared to the total number of observation	s. 	
This is especially the case for the analysis of individuals who have been charged with money 
laundering or receiving proceeds from crime at least once during a year since they represent

19	 	
 
approximately 0.1 % of the total number of observations during the 	whole period. 	It takes a 	
considerable amount of time for 	the logit 	estimations to	 converge to a maximum due to the size 	
of the dataset. The 	estimations	 ha	ve	 therefore had to be stopped before it converges when it 	
seemed that it was very close to the conver	gence point. 	The data will 	however 	first be descri	bed 	
before presenting the 	estimation results.	 	
 
3.3 	Descriptive statistics	 	
In the following, 	some 	descriptive statistics 	will be provided on the extent of criminality	, and 	
the 	sociodemographic and labour market characteristics of offenders	. The extent of 	the main 	
offences	 in addition to the	 charges of money laundering and receiving proceeds from crime will 	
first 	be 	presented, 	followed by a 	description	 of the share of each combination of 	additional 	
offences 	for individuals charged with money launderin	g. Next, the composition of gender, age, 	
immigration status, 	place of domicile	, education years 	and labour market status 	among charged 	
and non	-charged indi	vidual	s for each year in the sample	 will be considered	, to further show the 	
composition of these characteristics across individuals charged 	with	 specific types of crime, 	
including money laundering.	 	
3.3.1 Criminality	 	
Before analysing	 the estimation results, it is valuable to have a picture of the extent of each 	
offence. Figure 1 illustrates the evolution in the number of 	offenders	 of economic, violent, drug, 	
traffic and other offences from 2000 to 2019. 	An offender is 	considered as s	uch if a person has 	
been charged with at least one offence during a 	year.	 The total amount is lower than it would 	
be if 	the number of offences 	instead of offenders 	per year w	as analysed.	 	
Individuals	 charged with 	traffic	 offences represent the largest group of offenders	, as 	this group	 	
is composed of	 on average 	41 305 	offenders per year from 2000 to 2019	. The 	second largest 	
group of offenders is composed of 	other	 offenders	, with an average number of	 37 674	 offenders	 	
per year	 during the same period. In the 	regressio	n analyses that will follow, these two groups 	
will consist of one common group, but 	they have been described separately in the graph to show 	
the importance of traffic offences and to improve the r	eadability of the graph.	 	
The next largest group	 is composed of economic offenders, representing 	on average 3	5 783 	
offenders per year from 2000 to 2019.	 Economi	c offenders 	have	 exper	ienced an almost 	
continuous	 decline in their number during the whole 	period.	 Although less 	pronounced	, traffic 	
and other offences 	have also	 followed the same declining tendency over time.

20	 	
 
The 	second smallest	 group is composed of violent offenders representing an averag	e of 	27 674	 	
offenders per year from 2000 to 2019.	 The smallest group of offenders includes drug offences 	
with an average of 	20	 096 	offenders per year during the same period.	 Since the year 2000,	 the 	
number of violent	 offences has increased until 2014 and the number of dru	g 	offenders	 has 	
increased until 	2017. This increase is followed by a 	slight decrease in until 2019. 	 	
A remark to have in mind when analysing this data is that there might be some missing reported 	
cases in the last years of the studied period as it takes some time until one offence is registered 
in the police system. The number of cases might therefore be h	igher than it is in the last years 	
of the studied period, especially for the years 2018 and 2019.	 	
Figure 	1: Number of offen	ders	* in Norway from 2000 to 2019, by main categor	ies	 of offences	 
 	
Regarding the offences of money 	laundering and receiving proceeds from crime, Figure 2	 	
illustrates the evolution of the share of 	individuals charged with these offences	 compared to the 	
rest of the observations who have not been charged 	with these offences 	from 2000 to 2019. 	
Overall, the 	proportion	 of these offences has decreased on average during the whole period. It 	
first increased from representing 0.15 % of the observations in 2000 to 0.17 % in 2002. In the 
aftermath of this year, the share of these offenders has constantly decreased t	o end up 	
representing only 0.05 % of the observations in 2019. However, it is 	notable that 	there might

21	 	
 
be some missing reported cases in the last years of the studied period as it takes some time until 
one offence is registered in the police system	, implyi	ng that the observed number of cases might 	
be lower than the real	 in the last years.	 	
Figure 	2: Share of individuals charged with money laundering or receiving proceeds from 	
crime	* from 2000 to 2019	, by year	 	
 	
Next, F	igure 	3 illustrates the share of each combination of offences among those who have been 	
charged with	 money laundering and receiving proceeds from crime from 2000 to 2019. The 	
combination of offences which represents the largest proportion of all the combinations 
includes offenders who have been 	charged with	 drug	, economic and other 	co	-offences	, 	
excluding violent co	-offences. 	This group represents 21	.6 % of the total amount of offenders.	 	
The second largest grouping of offenders is composed of offenders who have onl	y been charged 	
with money laundering or receiving proceeds from crime, and no other offences, representing 
16	.7 % of those who have been convicted of this offence.	 The third largest combination 	of 	
additional offences is composed of those who have committed	 all four main cate	gories of co	-	
offences, representing 12.1 % of money launderers. 	This might indicate that a large share of 	
money launderers 	is often convicted of another offence, especially drug, economic and other

22	 	
 
offences. Those who have been 	charged with	 another economic offence than money laundering 	
or receiving proceeds of crime represent 8 % of the studied population.	 	
An interesting observ	ation is that	 being charged with drug offences in addition to money 	
laundering	 or receiving proceeds from crime	, regardless of	 being charged with any other 	
offences or not, seems to be very usual. Indeed, in 58 % of money laundering cases, at least one 
oth	er drug offence has on average been committed by the same person during a year.	 	
Moreover,	 the 8 combinations of outcomes which represents the lowest share of money 	
launderers, varying from 0.8 to 4.9 % of the total population studied, have also committed a	t 	
least one violent offence during a year. This might indicate that the relationship between 
committing a violent offence and money laundering or receiving proceeds from crime is 
relatively weaker compared to other offences, assuming that this relationship	 is existing.	 	
Figure 	3: Combinations of co	-offences of individuals charged with money laundering or 	
receiving proceeds from crime	* from 2000 to 2019, in proportion

23	 	
 
3.3.	2 Gender	 	
Starting with the gender of individuals, Figure 	4 illustrates the development of the proportion 	
of gender between 2000 and 2019 among individuals who have not 	been 	charged with	 any 	
offence in Panel A and people who have been 	charged with	 at least one offence during a year 	
belonging to any category of offence in Panel B. 	Next, 	Panel C, Panel D, Panel E and Panel F 	
describe the 	yearly gender ratio	 among individuals who have been 	charged with	 at least one 	
offence during a year belonging to t	he category of pro	fit-motivated offence, 	drug	 offence, 	
violent offence or other types of offences	, respectively.	 	
One can observe a clear tendency for men being on average 	charged with	 more offences than 	
women. This observation applies to all the categories	 of crime although there are some small 	
variations 	across	 categories. Violent offences are for instance executed by a larger proportion 	
of men on average than economic offences. 8	8.9% of violent crime offenders in 200	0 were 	
men, although, in the same year,	 only 78	.1% of economic offenders were men. 	 	
Remarkably, this tendency amongst violent offences has been declining with time as women 
comprise a larger proportion of 	violent	 offenders at the end of the period than in the beginning.	 	
This tendency is also observed amongst profit	-motivated offences, as 	women’s invol	ve	ment in 	
economic offences increased from representing 	21.9 % in 2000 to 2	7.6 % in 2019 of all 	
economic offenders. 	Drug crime	s are	 on the contrary experiencing a 	slight 	decrease in women	's 	
participation. Regarding other types of offences, the	 gender 	ratio	 is quite stable over 	the 	years 	
and similar to the gender disparity amongst all the offences.

24	 	
 
Figure 	4: Gender ratio of offenders and the remaining population	* from 	2000 to 2019, by main 	
categori	es of offences and 	year	 	
 	
As illustrated in Figure 	5, a larger share of men has also been 	charged with 	money laundering 	
or receiving proceeds from crime, compared to those who have not been 	charged with 	it. On 	
average from 2000 to 2019, 84.	4 % of these offenders were men, wh	ich is higher than the 	
average ratio of offenders of any crime, but lower than the share of men violent offenders. The 
gender ratio has been quite stable over time.

25	 	
 
Figure 	5: Gender ratio of individuals charged with money 	laundering or receiving proceeds 	
from crime and the remaining population	* from 2000 to 20019, by year	 	
 	
3.3.	3 Age	 	
When compared to the rest of the population who has been 	charged with	 any crime, what is 	
common for the different categories of crime regarding their age, is that there is a larger 
proportion of young individuals who tends to be charged with any offence relative to older 
persons. This is especially true for those 	charged wit	h drug offences, where on average during 	
the whole period, 3	8.7 	% of 	drug	 offenders 	were	 between 15 and 24 years old, whereas this 	
group of age represents only 16	.4 	% of the rest of the population who has not been 	charged with	 	
any offence. Individuals betw	een 25 and 34 years represent also a larger proportion 	of drug	 	
offenders compared to the other types of offenders and the rest of the population. Individuals 
who are over 55 years represent a smaller proportion of all sorts of offenders compared to the 
res	t of the population.	 	
The age repartition amongst economic, 	violent,	 and other offenders is overall quite similar, 	
although economic offenders tend to be a little bit older on average compared to violent 
offenders. 2	7.7 	% of economic offenders 	were	 45 years or older, whereas only 2	2.9 	% of violent 	
offenders belong to the same age group. Another similarity among economic, 	violent,	 and other

26	 	
 
offences, is that there has been, over time, a slight increase in older offenders’ participation in 
offences compared to younger, although this incre	ase does not se	em to be very significant. The 	
proportion of person	s who 	are	 between 35 and 44 is the group which has 	the	 most similar 	
proportion among all the categories of offenders and the rest of the population.	 	
Figure 	6: Propor	tion of age of offenders and the remaining population* from 2000 to 2019, by 	
main categories of offences and year	 	
 	
As illustrated i	n Figur	e 7, individuals 	charged with 	money laundering or receiving proceeds 	
from crime are also on average younger than 	the rest of the population. However, it may seem 	
that offenders between 25 and 44 years represent a larger share compared to the main categories 
of offences. There is a smaller proportion of them who are 45 years and older compared to those 
who have be	en	 charged with 	economic crime, but their share of offenders between 15 and 24 	
years is smaller than 	drug	 offenders. Over 	time,	 it seems that money launderers follow 	
approximately the same trend as economic offenders, as the share of offenders older than 24 
ye	ars increased until 2015 	and	 start	ed	 afterwards 	to decrease.

27	 	
 
Figure 	7: Propo	rtion of age of individuals charged with money laundering or receiving proceeds 	
from crime and the remaining population* from 2000 to 	2019, by year	 	
 	
3.3.	4 Immigration status	 	
In Figure 	8, one can observe that individuals with a foreign background tend to represent a 	
larger proportion of offenders compared to the rest of the population. This is especially the case 
for 	violent	 offence	s. Moreover, the propo	rtion of immigrants involved in offences has increased 	
over time, but this is partly explained by the fact that the proportion of immigrants has increased 
in the rest of the population 	as well	. However, the share of immigrants has increased more 	
among offe	nders than those who have not committed any crime. This is particularly true for 	
violent and economic offenders 	for	 which 32	 % of offenders were immigrants in both categories 	
of offences in 2019, whereas only 1	7.9 	% of the rest of the population were compo	sed of 	
immigrants in the same year. The evolution of immigrant participation in drug offences and 
other offences is quite similar. The share of immigrants has also increased in these two 
categories of offences, but the difference between this evolution and	 the one that the rest of the 	
population experienced is not as pronounced as it is for economic and violent offences.

28	 	
 
Figure 	8: Immigrant ratio of offenders and the remaining population* from 2000 to 2019, by 	
main 	categories of offences and year	 	
 	
As well as for all other offences, Figur	e 	9 illustrates that individuals 	charged with	 money 	
laundering or receiving proceeds from an offence represent on average a larger proportion of 
immigrants compared to those who have	 not been 	charged with 	it. This observation 	applies 	each 	
year	 of the period	, even if those who have not been charged with these offences have also been 	
experiencing a growing immigration ratio over 	time	. On average during the whole period, 	20	 	
% of money la	underers were immigrant	s.

29	 	
 
Figure 	9: Immigration ratio of individuals charged with money laundering or receiving proceeds 	
from crime and the remaining population	* 	from 2000 to 2019, by year	
 	
3.3.	5 Living in a city	 	
Next, there are only some small variations in the proportion of individuals living in a city among 
offenders and non	-offenders as illustrated in Figure 	10	. There is a larger share of offenders 	
living in a city among those who are 	charged with	 drug	 crimes than those who are not 	charged 	
with	 any crime since around 9	2.6 	% of 	drug	 offenders lived in a city on average from 2000 to 	
2019 and only around 88	.5 	% of i	nhabitants who were not 	charged with	 any crime lived in a 	
city during the same period. This tendency is also noticeable among profit	-motivated offenders, 	
as 9	1.7 	% of them were living in a city during the studied period.	 	
This observation does not apply to 	offenders of violen	t and other types of crime, as their share 	
of offenders living in a city is very close to the proportion of non	-charged	 individuals living in 	
a city. There are neither been any big changes in the proportion of individuals living in a city 
within each category of offence year after year. It is also noticeable that these proportions would 
vary	 depending on the definition of 	a city. 	Defining a city as comprising more than 5000 	
inhabitants is a quite broad definition which explains the large proportion of individuals living 
in a city. 	If a city were defined as including more inhabitants, the share of individuals living in 	
a cit	y would also be smaller	.

30	 	
 
Figure 	10	: Proportion of offenders and the remaining population living in	 a city* from 2000 to 	
2019, by main categories of offences and year	 	
 	
A larg	e proportion of 	individuals	 charged with 	money laundering 	or receiving profits from an 	
offence	 were	 on average living in a city. On average from 2000 to 2019,	 94.7	 % of them lived 	
in a city of more than 5	 000 inhabitants. 	This proportion of offenders living in a city is largest 	
than 	the share of 	the main categories of offences and it has moreover	 remained quite stable over 	
time	.

31	 	
 
Figure 	11	: Proportion of individuals charged with money laundering or receiving proceeds from 	
crime and the remaining population living in a 	city* from 2000 to 2019, by year	 	
 	
3.3.	6 Education level	 	
Figure 	12	 describes the evolution of the share of education of the different groups of persons 	
from 2000 to 2019. What is common for all the categories of offences, is that offenders in 
general have a lower education level than individuals who have not been charged	 with a crime. 	
This is particularly the case for drug offenders, as on average during the whole period, 66	.5% 	
of them	 had secondary school as their highest level of education, whereas only 26	.6% of those 	
who were not charged 	with	 any crime reached the same	 level of education. As expected, on the 	
higher side of the education level, a smaller proportion of offenders are having a higher 
education level compared to those who have not been charged with a crime. This is particularly 
the case for the bachelor leve	l and higher, but it also applies to the high school level.	 	
There has also been an increasing tendency for a higher education level among all types of 
offenders, especially for those studying at the bachelor	's or higher	 level	. However, this seems 	
to be explained by the fact that the population, in general, 	is having a higher education level, as 	
this increase is also observed among those who have not been charged with any crime. This

32	 	
 
trend is less marked among drug and violent 	offenders and more among economic and other 	
offenders.	 	
An important remark is that 	the age distribution is quite different among charged and non	-	
charged individuals. Since offenders tend to be younger than non	-offenders,	 it might be the case 	
that	 some 	offe	nder	s who have a low level of education 	will	 end up studying more later.	 Another	 	
graph could ha	ve	 compare	d otherwise similar non	-charged and charged individuals in terms of 	
age, gender and immigration. However, this is 	what will be done in the regression analysis later, 	
when controlling for all the variables at the same time.	 	
Figure 	12	: Proportion of 	education level of offenders and the remaining population* from 2000 	
to 2019, by main categories of offences and year	 	
 	
Individuals charged with money laundering or receiving proceeds from crime, as described in 
Figur	e 1	3, have also a lower education level on average compared to non	-offenders. The 	
proportion of each education level is most similar to the proportion of 	individuals 	charged with 	
drug	 offences, as	 70.9	 % of money launderers or 	individuals	 who have received pro	fits from an 	
offence have stopped studying already after secondary school. However, those of them who 
have reached a bachelor, master or 	higher	 degree, are representing a smaller share compared to 	
the other main categories of offence.

33	 	
 
Figure 	13	: Proportion of education level of individuals charged with money laundering or 	
receiving proceeds from crime and the remai	ning population* from 2000 to 2019, by year	 	
 	
3.3.	7 Labour 	market 	characteristics	 	
Regarding the labour 	market 	status, as illustrated in Figure 	14, the different categories of 	
offences have in common that offenders have on average a larger unemployment rate than the 
rest of the population who has not been convicted of any crime. Whereas 	the average 	
unemployment rate of observations who have not been 	charged with	 an offence has on average 	
been 2	.1 	% from 2000 to 2019, this rate has been up to 19	.4 % on average among drug offenders. 	
Again, offenders of this type of criminality 	distinguish themselves the most from the rest of the 	
population compared to the other categories of offenders. The unemployment rate has also been 
quite high for the other types of offences, as it has been on average 12	.4 % for economic 	
offences, 9	.7 % for	 violent offences and 7	.1 % for other sorts of offences.	 	
Regarding the share of self	-employment, offenders in general tend to have a larger share of self	-	
employees than those who have not committed any offence.	 Moreover, profit	-motivated 	
offenders distingu	ish themselves from 	any other 	offenders as 15	.6% of economic offenders 	
have on average during the whole period been holding a sole proprietorship and only 6% of the 
non	-charged	 are self	-employed. This share of self	-employment is also higher for the other

34	 	
 
categories of offence	s compared to non	-offenders, although it is less different 	for drug and other 	
offenders.	 	
As a result of this, offenders had a lower employment rate than non	-offenders. Among 	
offenders, violent and other offenders have the highest employ	ment rate whereas economic and 	
drug offenders had the lowest employment rate. With some fluctuations from year to year, the 
employment rate has slightly increased on average from 2000 to 2019 among the four categories 
of offence.	 	
Figure 	14	: Proportion of labour market status 	of active offenders and the remaining population* 	
from 2000 to 2019, by main categories of offences and year	 	
 	
Regarding individuals 	charged with money laundering and receiving proceeds from offences	 in 	
Figure 1	5, the unemployment and self	-employment rates are also relatively 	higher,	 and the 	
employment rate lower compared to those who have not been charged with these offences. The	 	
proportion of unemployed, self	-employed and employed 	money launderers	 rates are 	11.5, 10.1 	
and 78.4 %, respectively.	 The employment distribution is quite similar to the shares of 	
economic and drug offenders.

35	 	
 
Figure 	15	: Proportion of labour market status of active individuals charged with money 	
laund	ering or receiving proceeds from crime and the remaining population* from 2000 to 2019, 	
by main categories of offences and year	 	
 	
Figure 	16 demonstrates the proportion of	 after	-tax	 income among offenders and the rest of the 	
population. A common feature of the different categories of offence is that offenders earn on 
average less than individuals who are not charged with any crime. However, the extent of it 
varies among the different	 offences. Drug offences are the category of offence in which 	
offenders earn on average 	the least compared to	 the other types of crime. 86	.3 % of 	drug	 	
offenders had a	n after	-tax income	 of less than 300	 000 NOK on average between 2000 and 	
2019, whereas during the same period, only 50	.6 % of individuals who were not convicted of 	
any crime	 earned such an income. A very small proportion of drug offenders, around 1	.6 % of 	
them, had on averag	e a higher income than 500	 000 NOK, versus 12	.8 % for non	-offenders in 	
the same period.	 	
Although this tendency is also observed among violen	t, economic and other offen	der	s, the 	
difference in 	after	-tax income	 from the rest of the population is less pronounc	ed for these 	
categories of offence. This is especially the case for individuals 	charged with	 other offences. A

36	 	
 
common feature for all the offenders and the rest of the population is that their 	after	-tax income	 	
has on average increased over time, even thoug	h the income is adjusted for inflation.	 	
Figure 	16	: Proportion of after	-tax income of offenders and the remaining population* from 	
2000 to 2019, by main categories of offences and year	 	
 	
As well as for all the offences, persons charged with money laundering or receiving proceeds 
from crime are also earning on average a lower 	after	-tax income	 than non	-offenders	 as shown 	
in Figure 1	7. 85.7	 % o	f them 	had 	an	 after	-tax income	 of less than 300	 000 NOK on average 	
between 2000 and 2019, 	which is quite close to the share of drug offenders earning this range 	
of income. In general,	 individuals	 charged with	 money laundering and receiving proceeds from 	
crime have a pretty similar repartition and evolution of their 	after	-tax income	 compared to 	drug	 	
offenders.

37	 	
 
Figure 	17	: Proportion of after	-tax income of individuals charged with money l	aundering or 	
receiving proceeds from crime and the remaining population* from 2000 to 2019, by year	 	
 	
As described in Figure 	18, among employed and self	-employed individuals, 	the proportion 	of 	
individuals working 	in the private sector is on average higher 	for	 charged 	than	 non	-charged 	
individual	s. Around 80	.1 % of offenders were on average 	working	 in the private sector between 	
2000 and 2019, whereas only 63	.8 % of those who have not been 	charged with	 any offence 	
were	 working	 in the private sector during the same period. The proportion of persons 	working	 	
in the private sector is more or less equal 	across	 the different categories of offence	s, although 	
it is a bit higher among drug offenders as 85	.5% of them were 	working	 in the private sector 	
from 2000 to 2019.	 	
As a result of this, the proportion of individuals working in the state and municipal sector is 
lower among offenders 	compared to	 the rest of the population. The proportion of persons 	
working in the state sector is quite similar among all the types of offences, but the 	share	 of 	
individuals working in the municipal sector varies slightly more among offences	 as it is	 a little 	
bit lowe	r among economic and 	drug	 offences 	compared to 	violent and other offences. 	The 	
distribution of each sector has 	remained quite stable over time.

38	 	
 
Figure 	18	: Proportion of offenders and the remaining popu	lation	 working in the private,	 	
municipal and state sector	s* from 2000 to 2019, by main categories of offences and year	 	
 	
Similar 	to all offences, money launderers and persons 	charged with	 receiving profits from an 	
offence	 who were employed and self	-employed were 	working	 relatively	 more in the private 	
sector and less in the municipal and state sector compared to those who have not been 	charged 	
with	 these offences	. As	 described in Figure 1	9, t	he share of 	those 	charged with	 the last	-	
mentioned offences working in the private sector represents 	87.9	 % and is closest to the 	
proportion of drug offenders working in this sector.

39	 	
 
Figure 	19	: Proportion of individuals charged with money laundering or receiving pr	oceeds from 	
crime and the remaining population working	 in the private, municipal and state sector* from 	
2000 to 2010, by year	 	
 	
Next, Figure 	20	 illustrates the share of 	private sector 	employed persons working in the different 	
industrial categories. A larger proportion of offenders are 	employed	 in the sector of electricity, 	
gas, steam and water supply, sewerage and waste management and construction activities 
among offenders 	compared to	 the rest of the population. 18	.3 % of all the employed offenders 	
from 2000 to 2019 were on average working in this industry, whereas only 11	.1 % of the rest 	
of the population were employed	 in this industry in the same period.	 	
Regarding the hotel and restau	rant 	industry	, a slightly larger proportion of drug, economic and 	
violent offenders were working in this branch than people 	charged with	 other offences and 	non	-	
charged individuals 	in the period between 2000 and 2019.	 Moreover, there is a larger proportion 	
of employees from the private sector working in the 	industry of manufacturing	, retail trade, 	
repair of motor vehicles, 	transport and storage, among	 violent	 and other	 offenders compared to 	
non	-offenders	. For economic and 	drug	 offenders, the tendency is the opposite as a marginally 	
smaller proportion of them 	were employed	 in the	 manufacturing	 industry relative to non	-	
offenders, although this difference is tiny.

40	 	
 
Next	, a smaller share of offenders 	were	 employed in the finance and	 real estate	 industry	 	
compared to 	non	-offenders	. Around 4	.9 % of non	-offenders have been working in 	this industry	 	
whereas only 2% of all the offenders were on average 	employed	 in this 	industry	 between 2000 	
and 2019. There are no 	large	 variation	s across	 the different types of offenders.	 	
Furthermore, offenders tend to be less employed in the 	industry	 of services	 than non	-offenders. 	
The extent of this tendency varies across the different types of	 offences. Persons 	charged with	 	
other offences are the ones who are less employed in the service 	industry	 as 18	.7% of them 	
were on average employed in this 	industry	 from 2000 to 2019, whereas 28	.2% of active non	-	
offenders were employed in this 	industry	. Af	ter this group of offenders, violent offenders were 	
relatively less employed in the service 	industry	, followed by 	drug	 offenders and economic 	
offenders.	 	
Finally, the proportion of employees 	in the private sector working in 	the 	category	 regrouping 	
all the other industries such as agriculture, forestry, fisheries, mining, quarrying of oil and gas, 
and international organisations, is relatively equal across offenders and the rest of the 
population. Over 	time	, the proportion of employees fr	om the private sector working in the sector 	
of services and construction has been increasing, whereas the proportion of those working in 
the 	manufacturing industry	 has decreased 	among the whole population.

41	 	
 
Figure 	20	: Proportion o	f industries of employed offenders and remaining population* from 	
2000 to 2019, by main categories of offences	 	
 	
On average from 2000 to 2019, a larger proportion of individuals 	charged with	 money 	
laundering or receiving proceeds from an offence being employed in the private sector have 
been working in the 	industry	 of construction, hotel and restaurant, and a smaller share of them 	
have been working in the four other 	industries	 compared	 to tho	se who have not been 	charged 	
with	 these offences. 	The 	relative size of each sector is similar 	to the one of 	economic and 	drug	 	
offences.	 The proportions are	 illustrated in Figure 	21	.

42	 	
 
Figure 	21	: Proportion of industries of 	employed individuals charged with money laundering or 	
receiving proceeds from crime and remaining population* from 2000 to 2019, by year	 	
 	
Among self	-employees in the private sector as well, a larger proportion of offenders compared 	
to non	-offenders 	were	 working in the construction industry	, as described in Figure 	22	. On 	
average, 21	.4 % of all offenders were self	-employed in this 	industry	 fro	m 2000 to 2019, whereas 	
only 13	.5 % of non	-offenders were self	-employed in this 	industry	 during the same period. This 	
tendency is especially true for economic offences, as 26 % of 	them	 were self	-employed in the 	
construction industry from 2000 to 2019. 	 	
The share of self	-employees in the hotel and restaurant industry is quite similar among off	enders 	
and non	-offenders, although it is a little bit higher among offenders. There are no substantial 	
variations 	across	 the different categories of offence. This industry is also representing a smaller 	
share of all the industries among self	-employees comp	ared to employees. 	 	
Moreover, the share of self	-employees from the private sector working in the 	manufacturing 	
industry	 has also been higher among all the offenders compared to non	-offenders in the period 	
from 2000 to 2019. The industry of finance and real	 estate	, however, represent	s a very little 	
share of the different industries among self	-employed persons in the private sector, 	
independently from whether one is 	charged with	 an offence or not.

43	 	
 
Regarding the service	s industry, a larger share of self	-employ	ed offenders from the private 	
sector 	is working in this 	industry	 compared to 	non	-offenders.	 Only 20	.6 % of non	-offenders 	
were 	self	-employed	 in th	is industry from 2000 to 2019, whereas approximately 30	.3 % of 	
offenders were 	self	-employed	 in this industry in the same period. There are no 	large	 variations 	
across	 offences. 	 	
Finally, the proportion of self	-employed persons from the private sector working in 	the 	
industrial category regrouping all the other industries is on average smaller for 	offenders than 	
non	-offenders. 	The e	conomic offence	 is 	the group of offence which distinguished itself the 	
most, as the share of self	-employees of this group of offenders was 	on average 	12	.5% from 	
2000 to 2019, whereas the proportion of self	-employees of no	n-offenders working in this 	
industry	 was on average about 26	.1 %.	 	
Another interesting observation is that the share of self	-employed offenders increased suddenly 	
in 2006 whereas	, at the same time	, there were relatively 	fewer	 offenders 	self	-employed	 in the 	
manufacturing	 industry	. This sudden increase was especially drastic among 	drug	 offen	ces	 and 	
a bit among violent and other offences, but it was not very pronounced among economic 
offences. This development did not occur among non	-offenders as the share of s	elf	-employment 	
in 	manufacturing	 and other 	industries	 has increased constantly over time, and the proportion of 	
self	-employment in the service has been reduced with time.

44	 	
 
Figure 	22	: Proportion of industries of self	-employed offender	s and remaining population* from 	
2000 to 2019, by main categories of offences and year	 	
 	
The share of self	-employment across different 	industries	 is varying much more over	 time	 	
among money launder	ers	 than for 	offenders of 	the main categories of crime. This might be 	
explained by the fact that the number of self	-employed 	charged with	 these offences is relatively 	
lower than for the main categories of offences. However, on average during the studied period, 
a larger proportion 	of self	-employed individuals 	charged with	 money laundering or receiving 	
proceeds from crime 	were self	-employed in the industry o	f construction, 	manufacturing	, hotel 	
and restaurant relative to 	those 	non	-charged	. Consequently, a smaller share of them 	were	 	
wo	rking in the services and other 	industries	. The finance and real estate	 sector	, however, 	
represent	s a very little share of all the sectors, among 	the whole population	.

45	 	
 
Figure 	23	: Proportion of industries of self	-employed individuals 	charged with money 	
laundering and receiving proceeds from crime	* from 2000 to 2019, by main categories of 	
offences and year	 	
 	
Overall, compared to persons who have not committed any crime, offenders 	represent	 a larger 	
proportion of 	individuals who are young	, men, immigrant	s and have a lower level of education 	
and 	after	-tax income	. Offenders tend to work relatively more in the private sector and less in 	
the state and municipal sector	s compared to non	-offenders	. Moreover, a larger proportion of 	
offenders are unemployed or self	-employed and a lower share of them is employe	d. 	
There is a higher proportion of 	self	-employ	ed individuals in t	he industries of construction, 	
restaurant, 	manufact	uring,	 and finance among offenders, including money launderers, 	
compared to non	-offenders.	 The employment	 rate is higher in the indus	tries of construction, 	
hotel and restaurant and lower 	in the industries of finance and services among offenders, 	
including 	money launderers, compared to non	-offenders.	 	
Regarding 	manufacturing 	and other 	industries	, the proportion of individuals employed in these 	
industri	es is higher	 for offenders in general compared to non	-offenders but is lower for money 	
launderers relative to those who have not been 	charged with	 this offence.	 Finally, 	we note that

46	 	
 
it seems that drug	-related offenders often distinguish themselves the most from non	-offenders 	
compared to other offenders in terms of their age, education level, 	after	-tax income	, 	
unemployment rate, and whether they live in a city and work in the private sector	. These 	
descriptive data 	provide a starting point for	 the 	estimation results.	 	
 	
4.	 R	ESULTS	 	
4.1. 	Main groups of offences	 	
The average marginal effects	 of socio	demographic and 	labour market	 characteristics 	on the 	
probabilities of being charged with different offences are 	estimated with the help of binary 	
choice logit models and the results are provided in Table 1. 	The size of each group of offences 	
might be relevant to take into account when comparing t	hese groups to each other. 	Individuals 	
charged with any offence 	represent	 3.3 % of all the observations, whereas other offenders 	
represent	 2.03	 %, 	economic offenders 0.98 %, violent offenders 0.	76 % and drug offenders 0.55 	
% of the sample.	 Furthermore, t	he log	-likelihood 	values 	of each model 	are	 quite 	low	. This is 	
main	ly 	due to the 	complexity of the models in terms of the number of observations and 	
variables.	 The higher the dependent mean is the lower the log	-likelihood of the respective model 	
is.	 	
First of all, being a man increases the probability of 	committing a crime of any sort by 4	.23 	
percentage points, while holding all other variables constant. This 	estimate	 as well as the 	one 	
related to 	each 	category of offence	s are positive and 	significant at a 99 % confidence level, 	
implying that the gender of	 a person is associated with 	a change in the	 chance	s of committing 	
a crime. Among 	the different offences	, the coefficient is highest for offenders of other types of 	
criminality	 and lowest for drug offenders, as being a male increases the chances of committ	ing 	
the last	-mentioned offences by 2.	76	 and 0.	85	, respectively. However,	 when taking into account 	
the 	dependent mean	, the association between gender and 	criminality	 is 	strongest concerning 	
violent offences and weakest regarding 	economic	 offences.	 This corresponds to the 	data 	
described in section 3.3.2.	 	
All the 	estimate	s of 	average marginal effects of age are positive and those of age squared are 	
negative, 	regardless of the category of	 offences. These coefficients are all significant at a 9	9 % 	
confidence level. The signs of the 	two 	age coefficients imply that the relationship	 between age 	
and the probability of being 	charged with	 an offence is concave when holding all the other

47	 	
 
variables constant.	 We 	also 	note that as age variables are scaled by 10, the coefficient estimates 	
show the effects associated with an increase in age b	y 10 years.	 Based on the estimates	 in 	
Column (1)	, the average marginal effect of an 	additional year of age	 for those aged a is 	0.0024	-	
0.000816	(a/10	). This	 implies that 	going from age 15 to 16 increases the probability of 	
committing any offence	 by 0.11 	percentage points while going from age 50 to 51 decreases this 	
probability by 0.17 percentage points, holding everything else constant	. The 	probability starts 	
to decrease when going from age 30 to 31.	 	
Among the different categories of offences, 	going 	from 	age 	15 to 16	 increases	 the probability 	
of being charged with 	economic and drug offences the most, as it increases by 0.07 pe	rcentage 	
points, whereas the chances of committing 	violent and 	other offences increase the least as it 	
increases by 0.0	4 percentage 	point	s. 	When analysing the 	age function further, it is remarkable 	
that 	the 	probability of committing an offence starts to decrease 	at different age	s depending on 	
the offence. 	The probability of being charged with 	other 	off	ences starts to decrease when goin	g 	
from 	age 27 to 28, w	hereas the probability of being charged with 	drug and violent offences 	
starts 	diminishing	 from age 32 to 33. Finally, the probability of being charged with economic 	
offences starts decreasing 	at the latest age as it 	starts declining	 from age 	35 to 36.	 	
Overall, being an immigrant increases the chances of committing an	y offence by 0	.51 	
percentage points, while holding all the other variable	s constant. This coefficient is significant 	
at a 99 % confidence level	, as it is for each category of offence	. However, the coefficient is 	
positive for violent	, economic	 and other offences, whereas it is negative for 	drug	 offences. 	
Among 	violent, 	economic	 and other offences, 	the 	probability of committing 	each of these 	
offences increases	 by 0	.33, 0.13 and 0.02	 percentage points	, respectively,	 while holding all the 	
other va	riables constant.	 When taking into account the dependent mean, the 	positive 	association 	
between each offence and the immigration status is strongest	 for violent offences and weakest 	
for other offences.	 The chances of committing a 	drug	 offence 	are	 however 	decreas	ing 	by 0	.27	 	
percentage points when being a	n immigrant 	and holding all the other variables constant.	 	
Furthermore, living in a city of more than 5	 000 inhabitants increases the probability of 	
committ	ing any crime by 0	.19	 percentage points, while holding the other variables constant. 	
This coefficient is significant at a 99 % confidence level, which is	 also 	the case for all the 	
offences except for other offences for which the coefficient is 	negative but not significant.	 This 	
impl	ies	 that committing other offences does not seem to be associated with	 living in a city. 	
Among 	economic, drug and violent 	offences, the probability of being charged with 	these 	
offences 	increases by 0.28, 0.21 and 0.01	 percentage points, respectively,	 when living in a city

48	 	
 
and holding all the other variables constant. When dividing these coefficients by each 	dependent 	
mean, one can observe that the 	positive 	association between living in a city and each offence is 	
strongest for	 drug offences and weakest for other offences.	 	
A lower level of	 years of schooling	 is associated with an increased probability of 	being charged 	
with any sort 	of crime. The coefficient of the average marginal effect is significant at a 99 % 	
con	fidence level in the 5 different models. 	When spending one extra year at school, the 	
probability of being charged with	 other, economic, violent and drug	 offences increases by 	0.1	1, 	
0.06, 0.	05 and 0.	04 percentage points, respectively, while holding all 	the other variables 	
constant. When taking into accoun	t the dependent means, the association between these 	
offences and years of education is strongest	 regarding drug offences and weakest for other 	
offences.	 	
In terms of	 after	-tax	 income, all the 	estimates	 of the 	average marginal effects of income and 	
square	d income are negative. For 	drug, violent and economic offences, 	these coefficients are 	
significant at a 99 % confidence level. This means that the probability of being 	charged with	 	
the	se three	 offences decreases 	by 0.10, 0.02 and 0.01 percentage points, respectively, 	when the 	
after	-tax 	income increases by one unit, holding all the other variables constant.	 The order of 	
percentage points decreases related to each offence corresponds also to the 	relative degree of 	
correlation when taking into account dependent means.	 The fact that the square	d income is also 	
negative implies that the higher the 	after	-tax income	 is, the more the probability of committing 	
these 	offences will decrease	. 	
The probabilit	y of being convicted of other offences, however, does not seem to be affected by 	
the level of 	after	-tax	 income, as neither the variable of income nor of square	d income is 	
significant. As a result of this, the probability of committing any type of offence decreases when 
the 	after	-tax	 income increases, but this effect does not seem to be stronger when the 	after	-tax 	
income	 becomes higher, as the square	d income varia	ble in 	Column	 1 is not significant. The 	
variable of income is, however, significant at a 95 % confidence level in this model.	 	
Next, being unemployed increases the chances of committing any sort of crime by 1.	67	 	
percentage points	 compared 	to those 	outside t	he labour market	. The average marginal effects 	
in the 	5 models	 are	 positive and significant at a 99	 % confidence level. 	Compared with those 	
being out of the labour market, th	e probability of being charged with other, economic, violent 	
and drug offences 	increases	 by 	1.07, 0.36,	 0.25 and 0.20 percentage points, respectively, when 	
being unemployed and holding all the other variables constant.	 When taking into account the

49	 	
 
dependen	t means,	 the association between each offence and being unemployed compared to 	
those outside the labour market is strongest for other offences and weakest for violent offences.	 	
Being employed	 in the state or the municipal sector is associated with a decrea	se in the 	
probability of being charged with any crime of 3.29 and 2.71 percentage points, respectively, 
while holding all the other variables constant	. All the 	estimates of the	 average marginal effects 	
of working in the state or municipal sector are positive and significant at a 99% confidence 
interval	. When taking into account the dependent means, 	the negative relationship between 	
working in the	 state	 or municipal	 sector is str	ongest for	 drug offences and weakest for other 	
offences.	 	
Overall, when 	being employed 	in the private sector,	 the probability of being charged with any 	
crime	 decrease	s when holding all the other variables constant, 	regardless of the concerned 	
industry.	 All the corresponding coefficients	 among the 5 models	 are negative and 	significant at 	
a 99 % confidence level.	 Among individuals employed in the private sector, the chance	s of 	
being 	charged with	 any	 crime decrease the most when one is working in the sector of finance 	
and real estate, as it decreases the chances by 3.	58	 percentage points	 while holding all the other 	
variables constant	. This observation is also applicable when analysing each offence 	
individually.	 	
On the other side of the scale, the probability of committing an	y offence decreases the le	ast	 	
when one is employed in the industry of hotel and restaurant, as the probability is	 decreased by 	
1.08	 percentage points	 while holding all the other variables constant	. The 	estimated 	average 	
marginal 	effect of working in the construction sector is, however, quite close to the last	-	
mentioned industry, as working in the construction sector 	decreases the probability of 	
committing an	y offence by 1.4	2 percentage points	 while holding everything constant	. The 	
probability of committing 	economic, drug and violent offences decreases also the least when 	
being employed in the hotel and restaurant indu	stry, 	but	 for other offences, 	the probability 	
decreases the 	least	 when employed in the construction industry.	 	
Another observation is that 	among employed workers in the private sector, 	the	 negative	 	
association between each offence and working in 	any 	industr	y is weakest for other offences, 	
regardless of the industry and when taking into account the dependent means. This association 
is however strongest 	in the case of	 drug offences 	when being employed	 in any industry except 	
for other industries for which the relationship is strongest 	in terms of	 economic offences.	 This

50	 	
 
might indicate that 	industry	 is relatively less important for the probability of committing 	other 	
offence	s. 	
A different	 tendency is observed 	in the case of self	-employment, as being self	-employed 	in the 	
private sector 	increases the probability of being charged with any offence for al	l industrie	s 	
except for services and other industries.	 Being self	-employed in 	services and other industries 	
decreases the probability of being charged with any crime by 	0.48 and 1.26 percentage points, 	
respectively	, when holding all other variables constant	. This negative association is also 	
observed	 for each of the offences separat	ely.	 	
The 	estimates of Column 1	 with	 any offence 	as the dependent variable 	are 	all 	significant at a 	
99 % confidence level. 	Among the industries that are positively related to	 the probability of	 	
committing any crime,	 the relationship 	is strongest when 	self	-employed in the hotel and 	
restaurant industry	 as it increases the probability of being charged with a crime 	by	 1 percentage 	
point. The positive relationship is weakest when one is self	-employed in the 	construction 	
industry, as the probability 	increase	s in this case by 0.36 percentage points.	 	
The signs of the average marginal effects on any offence are similar to the ones on economic 
and other offences, but they vary 	across	 drug and violent offences.	 Starting with economic 	
offences,	 among the industries inducing a positive relationship with committing economic 	
offences, the relationship is strongest when being self	-employed in th	e finance and real estate 	
industry	 and weakest when self	-employed in the 	construction industry.	 All the 	estimates	 of the 	
average marginal effects of being self	-employed on economic offences are significant at a 99 	
% level.	 	
Regarding	 the probability of be	ing	 charged with other offences	, all the estimate	s in Column 5	 	
regarding self	-employment are significant at a 99 % confidence level, except for the coefficient 	
estimating the average marginal effect of being self	-employed in the financial industry which 	
is insignificant. This implies that being self	-employed is not a	ssociated with a change in the 	
probability of committing other offences. Among 	the industries inducing a positive significant 	
relationship, 	the correlation is strongest when self	-employed in the manufacturing industry and 	
weakest 	when self	-employed in the 	construction industry.	 	
Next,	 the probability of being charged with drug offences is reduced when 	being self	-employed	 	
in the private sector	 in any industr	y. It is reduced the most when being self	-employed in the 	
services industry as the 	probability has been reduced by 0.8	1 percentage points and it is reduced 	
the le	ast	 in the finance and real estate industry corresponding to a decrease in the probability of

51	 	
 
0.30	 percentage points	 while holding all the other variables constant	. The coeffici	ents are 	
significant at a 99% confidence level for every industry, except for the finance and real estate 
industry for which the coefficient is only significant at a 90% confidence level. 	 	
Lastly, the estimated average marginal effects of being self	-employ	ed	 on 	the probability of 	
being charged with 	violent offences	, described in Column 	4, are negative for all industries 	
except for 	the industry of finance and real estate which is positive	 but only significant at a 90 	
% confidence level. The other estimates a	re significant	 at a 99 % confidence level, except for 	
the estimate of self	-employment in the hotel and restaurant industry which is not significant. 	
Among the	 industries inducing a negative	 and significant 	relationship with committing violent 	
offences, 	the correlation is stronges	t when self	-employed in 	other 	industries	 and it is weakest 	
when self	-employed in th	e manufacturing industry.	 	
Overall, it seems tha	t the probability of 	being charged with any offence	 is highest for an around 	
30	-year	-old 	unemployed 	male with low education.	 A higher income	 and living in a city	 imply	 	
a higher probability of being charged with economic, violent and 	drug	 crimes, 	but 	these factors 	
are not related to any changes in the proba	bility of being charged with other offences.	 This 	
might indicate 	that the profile of an offender of other crimes is more random compared to 	
economic, violent and drug offenders who have a relatively more pronounced profile.	 	
Being employed in the 	state or municipal sector or any industry in the 	private sector reduces 	
the probability of committing any offence 	regardless	 of the industry	. The association between 	
being	 self	-employed 	in the private sector 	and committing an offence varies depending on th	e 	
industry and the offence, but the probability of	 committing any offence is	 reduced when self	-	
employed in the services and other industries	, while it 	increases when self	-employed in the four 	
remaining industries.	 	
 
Table 	1: Logit estimation of 	the 	average marginal effects of being charged with main categories 	
of offences	 	
-------------------------------------------------------------------------------------------------------------	 	 	(1) 	 	(2) (3) (4) (5) 	 	 	Any offence Economic Drug Violent Other 	 	--------------------------------------------------------------------------------	-----------------------------	 	Gender 0.04229*** 0.01027*** 0.00852*** 0.01303*** 0.02757***	 	 	(0.00007) (0.00003) (0.00003) (0.00003) (0.00005) 	 	 Age 0.02401*** 0.01343*** 0.01433*** 0.00845*** 0.00997***	 	 	(0.00020) (0.00005) (0.00005) (0.00004) (0.00007) 	 	 Age*Age 	-0.00408*** 	-0.00196*** 	-0.00226*** 	-0.00134*** 	-0.00190***	 	 	(0.00002) (0.00001) (0.00	001) (0.00001) (0.00001)

52	 	
 	
Immigrant 0.00509*** 0.00130*** 	-0.00273*** 0.00327*** 0.00020***	 	 	(0.00007) (0.00003) (0.00003) (0.00003) (0.00005	) 	 	 City 0.00192*** 0.00277*** 0.00212*** 0.00015*** 	-0.00000 	 	 	(0.00007) (0.00004) (0.00003) (0.00003) (0.00005) 	 	 Education 	 	-0.00198*** 	-0.00062*** 	-0.00042*** 	-0.00047*** 	-0.00113***	 	 	(0.00001) (0.00000) (0.00000) (0.00000) (0.00000) 	 	 	Income 	-0.00018* 	-0.00008*** 	 -0.00101*** 	-0.00017*** 	-0.00000 	 	 	(0.00007) (0.00000) (0.00001) (0.00000) (0.00000) 	 	 Income*Income 	-0.00000 	-0.00000*** 	-0.00001*** 	-0.00000*** 	-0.00000 	 	 	(0.00000) (0.00000) (0.00000) (0.00000) (0.00000) 	 	 Unemployed 0.01670*** 0.00358*** 0.00196*** 0.00247*** 0.01072***	 	 	 	(0.00011) (0.00005) (0.00003) (0.00005) (0.00008) 	 	 Employed state sector 	-0.03285*** 	-0.01897*** 	-0.01220*** 	-0.00856*** 	-0.01499***	 	 	(0.00015) (0.00008) (0.00007) (0.00005) (0.00008) 	 	 Employed municipal sector 	-0.02710*** 	-0.01392*** 	-0.00977*** 	-0.00632*** 	-0.01360***	 	 	(0.00011) (0.00006) (0.00005) (0.00005) (0.00007) 	 	 Employed Private 	-0.01415*** 	-0.01079*** 	-0.00643*** 	-0.00496*** 	-0.00423***	 	Construction (0.00014) (0.00006) (0.00004) (0.00004) (0.00007) 	 	 Employed Private 	-0.01078*** 	-0.00643*** 	-0.00443*** 	-0.00369*** 	-0.00577***	 	Hotel and Restaurant (0	.00015) (0.00007) (0.00006) (0.00006) (0.00011) 	 	 Employed Private 	-0.02030*** 	-0.01229*** 	-0.00786*** 	-0.00624*** 	-0.00844***	 	Manufacturing (0.00012) (0.00004) (0.00	003) (0.00003) (0.00005) 	 	 Employed Private 	-0.03584*** 	-0.01482*** 	-0.01100*** 	-0.01010*** 	-0.01919***	 	Finance and Real Estate (0.00030) (0.00014) (0.00016) (0.00013) (0.00018)	 	 	 Employed Private 	-0.02403*** 	-0.01108*** 	-0.00752*** 	-0.00682*** 	-0.01243***	 	Services (0.00015) (0.00005) (0.00004) (0.00004) (0.00007) 	 	 Employed Private 	 	-0.02315*** 	-0.01460*** 	-0.00794*** 	-0.00613*** 	-0.00986***	 	Other (0.00024) (0.00012) (0.00009) (0.00008) (0.00012) 	 	 Self	-employed Private 0.00357*** 0.00188*** 	-0.00692*** 	-0.00313*** 0.00353***	 	Construction (0.00021) (0.00009) (0.00014) (0.00011) (0.00016) 	 	 Self	-employed Private 0.01001*** 0.00259*** 	-0.00578*** 	-0.00017 	0.00557***	 	Hotel and Restaurant (0.00053) (0.00023) (0.00038) (0.00025) (0.00043) 	 	 Self	-employed Private 0.00789*** 0.00178*** 	-0.00643*** 	-0.00166*** 0.00688***	 	Manufacturing 	 	(0.00018) (0.00008) (0.00012) (0.00009) (0.00013) 	 	 Self	-employed Privat 0.00899*** 0.00362*** 	-0.00297* 0.00211* 0.00158 	 	Finance and Real Estate (0.00160) (0.00070) 	 	(0.00121) (0.00082) (0.00137) 	 	 Self	-employed Private 	-0.00478*** 	-0.00168*** 	-0.00807*** 	-0.00429*** 	-0.00223***	 	Services (0.00024) (0.00009) (0.00014) (0.00011) 	 	(0.00016) 	 	 Self	-employed PrivateO 	-0.01261*** 	-0.00998*** 	-0.00691*** 	-0.00490*** 	-0.00271***	 	Other (0.00021) (0.00013) (0.00012) (0.00010) (0.00014) 	 	-----------------	--------------------------------------------------------------------------------------------	 	Number of observations	 	72995982 72995982 72995982 72995982 72995982 	 	Mean of the outcome 0.03432 0.00980 0.00551 0.00758 0.02029 	 	Pseudo 	������2 	0.09816 0.1036	4 0.18251 0.10605 0.08419 	 	Log pseudolikelihood	 	 	-9838936 	-3605200 	-2036868 	-2908419 	-6629931	 	-------------------------------------------------------------------------------------------------------	------	 	Standard errors in parentheses	 	* p<0.05, ** p<0.01, *** p<0.001	 	All the models are controlled for year	-fixed effects and missing values on education.

53	 	
 
4.2. 	Money	 laundering and receiving proceeds from crime	 	
It is important to first analyse	 the 	association between 	the 	characteristics	 of individuals	 and	 	
the	ir probability of being 	charged with	 money laundering 	or receiving proceeds from crime	 on 	
a general basis	 being	 presented in Table 2	, to thereafter analyse how th	is association	 varies 	
across different criminality behaviour. This will enable establishing possible correlation	s 	
between money laundering and other offences.	 	
A first remark when observing the 	estimation	 results is that all the coefficients of average 	
marginal effects, except for the coefficient of being self	-employed in the finance and real estate 	
industry of the private sector, are significant at a confidence level of 99%. This means that all 
the 	socio	de	mographic and working factors, except for the last	 mentioned, seem to be associated 	
with a change in the probability of committing money laundering or receiving profits from an 
offence.	 First of all, being a man increases the probability of being charged w	ith money 	
laundering or receiving proceeds from offences by 	0.18	 percentage points	 when holding all the 	
other variables constant	. When taking into account the dependent means, this association is 	
stronger compared to 	the association between gender and 	each of the main categories of 	
offences.	 	
Next, the 	estimate of	 the 	average marginal effect of age is positive, whereas the coefficient	 on 	
square	d age	 is negative	. This impl	ies	 that the relationship between age and being 	charged with	 	
money laundering or receiv	ing	 proceeds from crime is concave when holding all the other 	
variables constant. 	The probability of committing money laundering increases	 every year	 when 	
one person is getting one year older until this person reaches the age of 	35. When going from 	
age 35 to 	36 the probability of being charged with money laundering or receiving proceeds 	
from crime starts to decrease.	 This is similar to what is previously observed for economic 	
offenders in general.	 	
As is the case for drug offences but not for any of the other s	orts of offences, 	the probability of 	
being charged with money laundering and receiving proceeds from offences is decreasing when 
a person is an immigrant. 	It decreases	 by 0.02 percentage points when be	ing an immigrant and 	
holding all the other 	variables constant	. When considering the dependent mean, this negative 	
association is weaker 	compared to drug offences.	 	
Next, as is the case for economic, violent and 	drug	 offences, living in a city increases the 	
chances of being 	charged with	 the offences	 of money laundering and receiving profits from 	
crime by 0.07 percentage points, when holding all the other variables constant	. This positive

54	 	
 
association is stronger compared to any of the other categories of offences, taking into account 
the dependent mea	n. 	
The effect of 	schooling years 	is nevertheless persistent among 	all the	 offence	s, as 	the probability 	
of being convicted of money laundering and receiving proceeds from crime decrease	s when 	
one 	become	s more educated. More precisely, this probability decreases by 0.0	07	 percentage 	
points when 	spending	 one extra year 	on	 education while holding the other variables constant. 	
When considering the dependent means, this association is only weaker than 	drug offe	nces	 	
among all the offences.	 	
The 	estimates of the 	average marginal effects of 	after	-tax 	income and square	d income are both 	
negative. The probability of being 	charged with	 money laundering and receiving proceeds from 	
crime is thus decreasing on average by 0.01 percentage points	 with one extra unit of 	after	-tax 	
income	 assuming the relationship is linear	 when holding all the other variables constant. Since 	
the variable of squ	are	d income is also negative, it implies that the higher the 	after	-tax income	 	
is, the more the probability of being 	charged with	 these offences will decrease with one extra 	
unit of 	after	-tax income	. The	 average	 association between income and money launderi	ng is only	 	
weaker than it is for drug offences	 when considering the dependent mean.	 	
Furthermore, in contrast to the main categories of offences, the probability of committing 
money laundering or receiving proceeds from crime decreases when one become	s unemployed	 	
compared to those 	outside of the labour force. The reduction is however quite little, as the	 	
probability decreases by 0.	007 percentage points when being unemployed 	compared to those 	
outside the labour force	 while holding all the other variables	 constant.	 	
However,	 since 	the labour market status of in	dividuals	 and 	the charge	s is updated yearly in the 	
data	set, 	a share of the offenders might be in prison	 during the year their offence is registered. 	
This implies	 that they 	would 	belong to the basis group regrouping 	individuals	 outside of the 	
labour market. 	In this case, it is the offence itself which affect	s the labour market status of a 	
person, and not the	 inverse.	 This is important to have in mind when interpreting the estimates 	
on unemployment	. 	
The chances of being charged with money laundering or receiving proceeds from crime are also 
reduced 	when one person is employed in the state or the 	municipal sector.	 The probability 	
decreases by 0.	30 	and	 0.26 percentage points	, respectively,	 while holding the other variables 	
constan	t. Compared to the dependent mean of the main categories of offences, the association 	
between being employed in the	 state	 or municipal sector 	and the probability to commit an

55	 	
 
offence 	is strongest 	if this offence	 is 	money laundering 	or receiving proceeds from crime 	than 	
any other main offences.	 	
Simi	lar to 	the main categories of offence	s, all the 	estimates	 of the average marginal effects o	f 	
being employed in the private sector are negative regardless of the industry.	 The probability of 	
being 	charged with	 money laundering or receiving proceeds from crime decreases the most 	
when being employed in the finance 	and real estate industry	, as it decreases by 0.26 percentage 	
points	 when holdi	ng all the other variable	s constant	. Right after this industry, it 	is 	the	 other 	
industries which reduce the probability the most, followed by the	 industry	 of manufacturing	, 	
and 	subsequently	, the services 	industry	. 	
Being employed in the construction, hotel and restaurant industry reduce	s the probability of 	
being 	charged with	 mon	ey laundering the least, as it reduces it by 0.18 and 0.	15	 percentage 	
points	, respectively	, while holding everything constant	. 	The association between	 being 	
employed in any industry and committing an offence is strongest if this offence is money 
laundering than for any	 of the	 other main offence	s when considering	 the dependent means.	 	
In contrast to	 the main categories of offence	s, being self	-employed	 in 	any industry	 reduces the 	
probability of committing money laundering or receiving proceeds from crime, except if 
working in 	the finance and real estate industry as th	e average marginal effect 	of 	self	-	
employment in this industry 	is not significant	 and positive	. The following industries are ranked 	
according to	 the 	industry	 that decrease	s the probability of being 	charged with t	he last cited 	
offences 	the most 	when being self	-employed in 	each	 industry	: other 	industries	, services, 	
construction, hotel and restaurant sectors, and finally the 	manufacturing 	industry.	 The 	
probability is	 indeed	 decreased by 0.2	0 percentage points when being self	-employed in other 	
industries	, whereas it is decreased by	 0.0	8 when self	-employed i	n the 	manufacturing 	industry.	 	
Overall, it seems that 	the probability of being charged with money laundering or receiving 	
proceeds from crime 	is the highest for an around 35	-year	-old male 	with low education, low	 	
after	-tax	 income, living in a city and who is not an immigrant.	 Being unemployed, 	self	-	
employed, or 	employed in the private, state 	and	 municipal sector	s, is associated with a lower 	
probability of committing money laundering compared to those outside the labour forc	e. The 	
probability is however decreased relatively less 	when employed in the construction, hotel and 	
restaurant 	industries	, and when self	-employed in the manufacturing, hotel and restaurant 	
industries.

56	 	
 
Table 	2: Logit 	estimation of average marginal effects of being charged with money laundering 	
or receiving proceeds from crime	 	
--------------------------------------------------------------------------------------	--------	 	 	Money laundering 	and receiving proceeds from crime 	 	--------------------------------------------------------------------------------------	--------	 	Gender 0.00175***	 	 	 	(0.00001) 	 	 Age 0.00307***	 	 	(0.00002) 	 	 Age*Age 	 	-0.00045***	 	 	(0.00000) 	 	 Immigrant 	-0.00022***	 	 	 	(0.00001) 	 	 City 0.00071***	 	 	(0.00002) 	 	 Education 	 	-0.00007***	 	 	(0.00000) 	 	 Income 	-0.00012***	 	 	 	(0.00000) 	 	 Income*Income 	-0.00000***	 	 	(0.00000) 	 	 Une	mployed 	-0.00007***	 	 	(0.00001) 	 	 Employed state sector 	-0.00303*	** 	 	(0.00004) 	 	 Employed municipal sector 	-0.00256***	 	 	(0.00003) 	 	 Employed Private 	-0.00176***	 	Construction (0.00002) 	 	 Employed Private 	-0.00148***	 	Hotel and Restaurant (0.0000	3) 	 	 Employed Private 	-0.00210***	 	Manufacturing (0.00002) 	 	 Employed Private 	 -0.00260***	 	Finance and Real Estate (0.00008) 	 	 Employed Private 	-0.00208***	 	Services 	 	(0.00002) 	 	 Employed PrivateO 	-0.00214***	 	Other (0.00005) 	 	 Self	-employed PrivateC 	 	-0.00107***	 	Construction (0.00004) 	 	 Self	-employed PrivateH 	-0.00082***	 	 	 	(0.00011) 	 	 Self	-employed PrivateM 	-0.00075***	 	Manufacturing (0.00003) 	 	 Self	-employed PrivatF 	 	0.00015 	 	Finance and Real Estate (0.00029) 	 	 Self	-employed PrivateS 	-0.00148***	 	Services (0.00005) 	 	 Self	-employed PrivateO 	-0.00200***	 	Other (0.0000	6) 	 	--------------------------------------------------------------------------------------	--------	 	Number of observations 	72995982

57	 	
 
Dependent mean 	 	 	0.00101 	 	Pseudo 	������2 	 	0.17297 	 	Loglikelihood 	-480462 	 	--------------------------------------------------------------------------------------	--------	 	Standard errors in parentheses	 	* p<0.05, ** p<0.01, *** p<0.001	 	All the m	odels are controlled for year	-fixed effects and missing values on education.	 	 
After 	obtaining	 some knowledge 	concerning	 which factors 	are associated with	 one person’s 	
probability to commit money laundering and receiving proceeds from crime, 	the variation in 	
these relationships d	epending on which other offences a money launderer has been charged 	
with	 will be studied.	 Panel	s A and B of Appendix 2	 illustrate 	the 	logit estimations	 when 	
decomposing 	the probability of being charged with money 	laundering or receiving proceeds 	
from crime	 depending on possible	 co	-offences	 committed by the charged individuals, including 	
no other co	-offences committed as one of the categories.	 	
A remark is that in some of these models, some observations have been omitted from the 
regressions 	because	, for some variables, almost all the data ta	kes the value of	 either	 or 1. It is 	
the case for the variables representing whether one is employed or self	-employed in the 	
financial sector, self	-employed in the construction sector and employed in the state sector. As 	
such combinations 	occur rarely,	 there is not enough variation in the data to identify the effects 	
of 	associated 	variables	. This explains the reason for the variation in the number of observations 	
of the different models in the 	estimation	 table. 	It is 	nevertheless	 considered interesting to 	
include these variables as they might be 	significant for the other combinations of outcomes.	 	
First of all, the	 estimated	 average marginal effect of being a man is positive and significant	 at 	
a 99 % confidence level	 across all the combinations of outcomes	. The combination of co	-	
offences which explains the most t	he	 association between	 gender and money laundering	 is the 	
one	 including 	drug, economic, violent and other 	co	-offences.	 The probability of being charged 	
with all these offences	 in addition to money laundering or receiving proceeds from crime	 	
increases on average by 0.03 percentage points when holding all the other variables constant.	 	
The combination	s of co	-offences implying the lowest increase in probability in terms of 	
percenta	ge points 	are	 the one	s including 	either (a) 	economic and 	violent 	offences, or (b) drug 	
and violent offences.	 The probability o	f being charged with these two alternatives of co	-	
offences 	increases by 0.002 percentage points when being a man and 	holding all the other 	
variables constant.	 	
The	 association between age and the probability of committing an offence is similar across all 	
the combinations of 	co	-offences. 	The average marginal 	effects	 of age are positive and of 	
squared 	age are negative. Both	 are	 significant at a 99 % confidence interval.	 The relationship

58	 	
 
between age and	 the probability of being charged with	 each alternative of	 co	-offences is thus 	
concave. 	The age at which the probability of committing any co	-offences varies across 	the 	
different combinations.	 It varies from the age of 17 for which the probability of being charged 	
with violent offences as 	the only	 co	-offence is the hig	hest	, to the age of 46 f	or which the 	
probability of committing 	drug, violent and other offences as co	-offences is the highest.	 	
Moreover, 	similarly to the probability of committing money laundering or receiving proceeds 	
from crime in general, the probability of committing the co	-offences of (a)	 drug and economic 	
offences and (b) drug, economic and other offences, reaches its peak at the age of 35.	 	
The average marginal effect of being an immigrant 	is significant at a 99 % confidence leve	l 	
across	 all the combinations of 	charges	, except for when the 	outcome is to be charged with 	
economic and other offences 	as co	-offences 	for which the coefficient is not significant.	 	
Moreover, the	 sign of	 the 	estimates	 varies across the different	 combinations of 	cha	rges	. 	
In 6 of the 16 models, the correlation is 	negative and significant	. Drug offences 	are	 included	 as 	
co	-offences	 in these 6 models, and economic offences 	are	 included in 4 of them.	 These 	
combinations 	of 	co	-offences drive the 	negative 	association between immigration and 	
committing money laundering in general.	 Among 	these models	, when being an immigrant, 	the	 	
probability of	 being charged with drug, economic and other 	co	-offences in addition to money 	
laundering 	decreases the most in 	terms of percentage points change, as it 	decreases by 0.0	35	 	
percentage points	 when holding everything else constant.	 	
Of	 the 9 remaining models for which the association between immigration and the dependent 	
variable	s is positive and significant, 6 of them include violent offences.	 The probability 	
increases	 the most	 in terms of percentage point changes	 when	 committing no ot	her offences 	
than money laundering or receiving proceeds from crime, as it increases by 	0.011 	percentage 	
points	 when holding all the other variables constant. This might	 indicate that it is more usual to 	
commit a co	-offence than no other offence than money laundering.	 	
Next, 	the assoc	iation between living in a city and any 	alternatives	 of 	co	-offences is 	positive and 	
significant at a 99 % confidence level	 across all the 	alt	ernatives	. The probability of being 	
charged 	with	 all 	4 main	 offences, or 	all of them except for violent offences	 in addition to money 	
laundering	, has the highest percentage points 	increase when living in a city	, as it increases by 	
0.01	1 and 0.	018 	percentage points	, respectively,	 while holding all the other variables constant.	 	
On the other side	 of the scale	, it is the probab	ility o	f being 	charged with	 either (a)	 drug	 and 	
violent 	co	-offence	s, (b)	 only violent 	co	-offence	s, (c)	 violent and other co	-offences and (d)

59	 	
 
violent and economic co	-offences,	 which 	increases	 the least in terms of percentage points 	
changes	 when living in a city	. The probability of being charged with 	one of these 4 alternatives 	
of co	-offence	 in addition to m	oney laundering 	increases by 0.001 percentage points	 each	 when 	
holding all the other variables constant. 	The 	increase in the 	probability of only being charged 	
with money laundering or receiving proceeds from 	an offence, is mod	erate compared to the 	
other combinations of offences, as the probability increases 	by	 0.005 percentage points when 	
living in the city and hold	ing	 other variables constant.	 	
Furthermore, 	the relationship between 	the years of education 	and the 	probability of 	being 	
charged with	 different combinations of money laundering is negative and significant at a 99 % 	
confidence leve	l across	 all the	 outcomes. 	The coefficients of the average marginal effects are 	
however all 	relatively	 small.	 The combination of co	-offences which explains the most the 	
association between education and money laundering is the 	combination	 that 	is composed of	 	
drug, economic and other offences	. The probability of being charged with these offences 	
decreases by 0.002 	percentage points when spending one extra year at school and holding all 	
the other variables constant.	 This group of co	-offences is	 however	 also representing the largest 	
group	 in terms of	 the 	number of offenders.	 	
In 	9 models, the	 decrease in percentage points in the probability of 	being charged with each 	of 	
the outcome	s when spending one extra year on education is 	equal to when rounded to 5 	
decimal places, although it is significant. 	7 of these models include violent offences as	 a co	-	
offence.	 	
Next,	 what is interesting about income, is that	 after	-tax	 income does not seem to be associated 	
with a change in the probability of 	being charged with	 (a) 	no other co	-offences 	and	 the co	-	
offences of	 (b) 	violent and 	other offences 	since the coefficients on the average marginal effect	s 	
of income	 and square	d income	 are	 insignificant in the respective models.	 The fact that the effect 	
of income is insignificant for the probability of only committing money laundering or re	ceiving 	
proceeds from crime might indicate that the total effect of income on being convicted of these 
offences is primarily driven by committing additional offences.	 	
All the 	estimates	 of 	the 	after	-tax	 squared 	income are negative, indicating a concave relationship 	
between income and being 	charged with	 an offence	, which is in line 	with the	 estimations 	
concerning	 money laundering in general	. Th	ese 	estimates	 are significant at a 99 % confidence 	
level	 in 12 models o	ut of 16. 	It implies that the probabilities of being charged with these

60	 	
 
offences decrease even more with an extra unit of 	after	-tax income	 when the 	after	-tax income	 	
is at a relatively higher level	. 	
The negative association between higher income and 	money laundering is mainly driven by	 4 	
combinations of co	-offences as only 	5 models have negative and significant estimates at a 99 	
% level. These outcomes	 are composed of the following co	-offences:	 (a) economic, violent and 	
other offences, (c) economic an	d other offences	, (b) only economics offences	, (d) only other 	
offences	 and (e	) drug, violent and other offences	. Remarkably,	 economic and other offences are 	
the	 most recurrent	 co	-offences 	in these models.	 	
Furthermore, in 8 of the models	, the average marginal effect of 	after	-tax income	 is positive and 	
significant at a 99 % confidence level	, except 	for	 the outcome of violent 	co	-offences which is 	
only confident at a 90 % confidence level.	 The	se combinations of offences tend to 	explain less 	
the probability of committing money laundering in general c	ompared to the 5 last mentioned 	
alternatives of co	-offences because the association between	 income and	 money laundering 	is 	
negative. 	The signs of the estimates	 impl	y that the probabili	ty of being charged with the 	
offences in each model increases relatively more with one extra unit of income when the income 
is low. When the income level is higher, the probability will increase at a slower rate and even 
decrease at a certain level of inco	me.	 	
An interesting observation is 	that	 among these 8 models, 7 of them include 	drug	 offences.	 The 	
coefficients are relatively small, but the probability of being charged with 	additional 	offence	s 	
increases the most 	in terms of percentage points increase 	when the	 co	-offences 	are 	drug	, 	
economic, violent and other	. It	 increases by 0.0	11 percentage points with an extra unit of 	
income while holding all the other variables constant.	 	
Regarding the average 	marginal effect of	 being unemployed	 compared to th	ose outside the 	
labour force	, the significance and the sign is varying a lot across the different outcomes. 	As 	
studied earlier, the relationship between committing money laundering or receiving proceeds 
from offences and being unemployed compared to those out of the labour market is negative. 
This probability decrease is mainly dr	iven	 by the negative and signifi	cant 	estimate	s in the 	
following combinations of outcomes: (a) drug and economic offences, (b) drug, economic, 
violent and other offences, (c) drug, economic and other offences, (d) drug, violent and other 
offences and (e) drug, economic and violent offence	s. Among these outcomes, the probability 	
of being charged with all four offences in addition to money laundering decreases the most	 in 	
terms of percentage points change	 when being unemployed, as it decreases by 0.005 percentage

61	 	
 
points while holding all the	 other variables constant. 	Remarkably, d	rug offences 	are	 included	 	
as co	-offences	 in all the last	-mentioned 	combinations	 and economic 	offences 	are	 included in 4 	
of 5 models.	 	
The models 	including the following co	-offences 	are 	explaining	 the probability of money 	
laundering or receiving proceeds from crime relatively less as their estimate	s regarding 	
unemployment	 are positive	: (a) only money laundering and receiving proceeds from offences, 	
(b)	 economic, violent and other offences, (c) 	othe	r offences and (d) violent offences	. The 	
coefficients are significant at 99 % in the three first	-mentioned models and at a 90 % confidence 	
level in the model only including violent offences	 as co	-offences	. In all the other models which 	
have not been mentioned, the 	correlation between	 unemployment	 and 	given co	-offences	 is not 	
significant.	 	
As is the case for the main variable of interest, money laundering and receiving proceeds from 
crime, 	being	 employed 	in th	e state or municipal sector decreases the probab	ility of being 	
charged with any combination of 	co	-offences	. An e	xcept	ion	 is 	observed regarding	 the outcome 	
of being charged with 	drug	, economic and violent 	co	-offences for which the coefficient of being 	
emp	loyed in the state sector is 	positive and 	not significant. Otherwise, the coefficients are	 	
negative and	 significant at a 99 % confidence level.	 The probability 	de	creases the most 	in terms 	
of percentage points change 	if one is convicted o	f drug	, economic	 and other 	co	-offences,	 when 	
working 	either in the state or the municipal secto	r, implying a 	decrease	 of 0.099	 and	 0.08	0 	
percentage points	, respectively, while holding the other variables constant.	 This might be 	
related to the fact that this combination of co	-offences represent	s the largest share of money 	
launderers.	 	
Regarding the 	estimated 	average marginal effects of 	being employed	 in the different industries 	
of the private sector	 compared to those outside the labour force	, all the coefficients are	 negative 	
and 	significantly different from except for 	the estimate	 of being employed in the financial 	
sector 	when being 	charged with	 violent and other offences meaning that there does n	ot seem to 	
be any relationship.	 The fact that the other estimates are negative i	s in line with t	he previous 	
result on money laundering and receiving proceeds from crime in general.	 	
Another 	observation is that 	the outcome for which 	the 	probability decreases the most 	in terms 	
of percentage points changes 	when 	being employed in any industry i	s the outcome of being 	
charged 	with the co	-offences 	including 	drug	, economic and other offence	s. The probability of	 	
being charged with these 	co	-offenc	es decreases the most when	 being employed in the 	financial

62	 	
 
industry	, by 0.	07	3 percentage points	, and 	it decreases the least when	 being	 employed in the 	
construction 	industry	, meaning by 0.	048	 percentage points	 when holding all the other 	variable	s 	
constant	. 	
On the other hand, the combination of offences for which the probability decre	ases the least	 in 	
terms of percentage points changes	 when being employed in any industry is the outcome 	of 	
only being charged with violent 	co	-offences.	 In this 	co	-offence alternative	, the probability	 	
decreases the most in the 	services 	industry	 and it	 decreases 	the least 	in the con	struction sector	, 	
in terms 	of percentage points changes.	 The	 relative effect of each combination of co	-offences 	
on the	 probability of committing money laundering	 corresponds to their relative size, presented 	
in sub	-section 3.3.1.	 	
There 	is more variation 	in the relationship between being self	-employed in the different 	
industries of the private sector and being charged with different combinations of 	co	-offences.	 	
Fir	st of all, the probability of being charged with no other offences than money laundering and 	
receiving proceeds from offences,	 is increasing when 	self	-employed	 in the	 manufacturing, hotel 	
and restaurant 	industries an	d is decreasing when 	self	-employed	 in the 4 remaining	 industries. 	
The 	estimate	s are	 however	 not significant regarding the 	industry	 of construction and finance	, 	
whereas 	they are	 significant at a 99 % confidence level for the 	remaining	 industries.	 	
The association between being s	elf	-employed and being charged with money laundering or 	
receiving proceeds from crime, in general, was negative when working in any industry except 
for the financ	ial	 one. 	This 	does not correspond to th	e estimates of 	the construction, restaurant 	
and 	manufacturing industries	 on the probability of committing 	no other offences than 	money 	
laundering.	 This might indicate that at least for	 the	se three industries,	 the probability of 	
committing money laundering in general is driven by	 committing	 other co	-offe	nces	. 	
Starting with the construction industry, 	the 	estimated 	average marginal effect of being self	-	
employed in this industry 	is also negative and significant	 at a 99 % confidence in	 7 different 	
models which have in common that they all include 	drug	 offences.	 	
Next, 	in 6 of the models, 	the average marginal effect	 of being self	-employed in the hotel and 	
restaurant industry	 is 	negative and significant	. Drug and economic offences are 	the most 	
recurrent co	-offences in these models. Moreover, 	it is only th	e probability of being charg	ed 	
with 	drug	, economic and other offences 	as co	-offences	 that is significant at a 99 % confidence 	
level	. This combination is therefore	 mainly explaining the 	probability change 	of the 	main 	
offence in interest.	 Furthermore, 	in 9 out of 16 models,	 there is no correlation between being

63	 	
 
self	-employed in this industry and being charged with	 the co	-offences of the respective models, 	
as their estimates are not significantly different from 0.	 Remarkably, 7 of these 9 m	odels include 	
violent offences.	 The same tendencies a	pply	 to the	 manufacturing	 industry	, although there	 are	 	
9 models for which the 	estimates of 	the marginal average effects are negative and significant.	 	
The average marginal effects of being 	self	-employed in the financial and real estate industry 	
are 	not significantly different from 0, except for when 	being charged with	 only economic 	co	-	
offences.	 The coefficient is significant at a 90 % confidence level an	d implies that the 	
probability of being charged with these offences increases by 0.	01	0 percentage points when 	
being self	-employed in the financial sector	 and holding all the other variables constant.	 	
Although all the other coefficients are 	not significant, they are positive, which 	corresponds 	to 	
the 	estimate	 of money laundering	 in general.	 	
Regarding services and other 	industries	, the average marginal effects 	correspond to the 	estimate	 	
of 	being 	charged with	 money laundering or receiving proceeds from crime in general	. Indeed, 	
all the coefficients are negative and significant at a 99 % confidence level	 except 	in the case of	 	
being charged wi	th drug	 an	d violent 	co	-offences	 in 	both	 industries	 and	 only 	violent 	co	-offences 	
in other industries	, for which the coefficients are insignificant	. When being self	-employed in 	
these two industries, the probability of being charged with 	drug	, economic and other offences 	
as additional offences 	decrease	s the most	 in terms of percentage points changes.	 	
Based on the results above,	 it seems that 	laundering money is less associated with 	being 	
convicted of violent offences	 too, whereas it seems that it is relat	ively more linked to 	
committing 	drug	, economic and 	other offences	. This last	-mentioned combination of 	co	-	
offences 	tends	 to drive 	the 	association between 	the different factors 	and	 the probability of being 	
charged with money laundering 	or receiving proceeds from 	crime in 	general.	 	
 
4.3 	Implications for	 Policymakers and	 C	ountermeasures 	 	
This analysis suggests that there is a significant relationship between the probability of being 
charged with most of the offences and the gender, age, immigrant status, resident place, 
education 	level, income and working status of individuals.	 Identifying the characteristics of 	
offenders	 can be used to	 develop more app	ropriate and effective investigation guidelines for 	
the police, but it can also provide som	e proposals	 for policies related to preventing	 white	-collar 	
crime.	 However, 	the 	findings	 of this study	 impl	y only 	correlation and not causality	, meaning 	
that no strong policy recommendations can be made	. The analysis is still of great importance

64	 	
 
as there is little d	ata on criminality and it 	can 	provide 	some valuable information	. Moreover, 	
one	 should be cautious in using 	the results 	in other countr	ies	 than Norway as 	the institutional 	
setting	, as well as the sociodemographic and labour market 	characteristics,	 may differ.	 	
First of all, 	this analysis has	 developed the	 profile of a 	typical offender of not only money 	
laundering and receiving proceeds of crime but also offender of 	drug	, economic, violent and 	
other crimes. 	A person is more 	likely to be charged with money laundering or receiving 	
proceeds from crime if this person	 is an around 35	-year	-old male with low education, low after	-	
tax income, living in a city and who is not an immigrant	. Offender profiling is	 a common 	
practice	 when in	vestigating criminality	. It 	consists of narrowing 	the group of 	potential suspects 	
based 	on information available on passed criminality. It allows 	differentiating offenders from 	
non	-offenders.	 (Dabney et al., 2006)	 The offender profiles developed in this thesis based on 	
past criminality might therefore be used 	during the	 screening 	proce	ss of 	law enforcement agents	 	
in Norway	. 	
Moreover, 	the association between the offence of money laundering or receiving proceeds from 	
crime and other offences might also be beneficial in the proces	s of screening suspects. As shown 	
previously, a person who h	as been charged with money laundering	 or receiving proceeds from 	
crime	 has often also been charged with other 	drug	, economic and 	other offences at the same 	
time.	 If one is investigating 	an individual who has been charged with 	money laundering	 or 	
proceeds from crime	, it might be beneficial 	to use	 more	 resources on investigating	 further	 	
whether this person has also committed other 	dru	g, economic and other offences	, compared to 	
investigating if this person has also committed a violent offence. The reason for this is that a 
smaller proportion of money launderers have been charged with violent offences, compared to 
the three other offence	s. 	
However, 	when 	using experiences from	 past charges 	to influence the direction of the present or 	
future charges,	 one issue arises if these past charges are based on biased 	judgments	 as this bias 	
might	 be reproduced in the future judgment.	 An example 	is that the police arrest	 a larger 	
proportion of immigrants 	compared to the actual 	share	 of 	it in society 	due to 	a though	t or an 	
assumed st	atistic that immigrants are more likely to commit an offence. In this case, 	when 	
profiling offenders based on past charges, 	the police will continue to arrest	 immigrants 	
disproportionally	 based on biased statistics. 	(Dabney et al., 2006)	 	
Secondly, more specific	ally	 in the ca	se of money laundering,	 institutions 	that 	are subject to 	
report	ing	 obligations to FIU are also	 required to 	establish measures preventing 	money

65	 	
 
laundering and terrorist financing based on an assessment of the risk of it, 	cf. Anti	-Money 	
Laundering Act 	§ 6. This includes 	assessing	 the risk of 	their customers 	committing	 the last	-	
mentioned offences.	 However, the Act 	is 	somewhat vague	 as it does not specify how the risk 	
assessment should be conducted, but it provides some guidelines. 	The 	guide to the Money 	
Laundering Act	 mention	s that the risk should be assessed according to the	 institution's own 	
business,	 products, services, customers	 and geographic factors.	 It states that the assessment 	
should also be based on own experience or external sources	 such as reports published by 	
authorities or o	rganisations. 	(Veileder til hvitvaskingsloven	, 2022)	 	
However, when a reporting 	insti	tution is newly established or does not have a lot of customers, 	
it might be challenging 	to establish effective anti	-money laundering measures complying with 	
the Act as they do not have a lot of experience with 	cases of money laundering among their 	
custome	rs. In 	this situation	, it might be useful to 	provide information on the typical profile of a 	
money launderer to the Act or its guide	lines	. Knowing	 which sociodemographic characteristics 	
are associated with a higher probability of committing this crime	 can make it easier to prevent 	
money laundering but	 it also	 increases	 the chances	 to comply with the law if 	the Act were 	more 	
specific than 	currently	. It might also be useful for 	any	 reporting organisation 	because	 creating 	
a common framework 	might	 make their work more targeted and effective.	 	
 	
5.	 CONCLUSION	 	
To summarise the findings 	on the sociodemographic characteristics of offenders	, the probability 	
of being charged with any offence is highest for an 	unemployed 	male with low education	. 	
Regarding age, the probability of being charged with money laundering or receiving proceeds 
from crime is highest for a 	35	-year	-old person, 	whereas the probability of committing any 	
offence is highest for a 30	-year	-old person.	 	
Moreover, it seems that economic, 	drug	 and violent offenders	, including money launderers,	 	
have 	a more specific profile than other offenders 	who seems to be	 relatively	 more random	. The 	
reason 	for	 this statement is that 	the	re seems to be no relationship between the 	probability of 	
committing other offences 	and the 	after	-tax	 income 	or whether one lives in a city	. In contra	st,	 	
the chances of being charged with 	drug	, economic and violent offences, including money 	
laundering or receiving proceeds from crime, 	decrease with higher 	after	-tax	 income 	and 	
increase when living in a city.	 Regarding the immigrant status of a person, being an immigrant

66	 	
 
is associated with a higher probability of being charged with	 economic, violent and other 	
offences, whereas the probability decreases if 	the offences are money laundering, receiving 	
proceeds fr	om crime or 	drug	-related	. 	
Next, regarding th	e labour market	 characteristics,	 being employed in the mu	nicipal, state or 	
private sector is associated with a decrease 	in the 	probabilit	y of being charged with any offence	. 	
This relationship does not vary 	across the different industries in the private sector.	 Being 	
unemploy	ed,	 however,	 is correlated with 	a lower probability of being charged with money 	
laundering 	or receiving proceeds from crime, whereas the relationship is 	positive if the offence 	
belongs to	 any of the four main categories of offences.	 	
Being self	-employed in the private sector is associated with a decrease 	in the	 probability of 	
being charged with 	any offence if working in manufacturing and other industries.	 The 	
relationship varies across 	the	 other four industries, but it is consequently positive for economic 	
offences and negative for 	drug	 offences. 	The	 relationship is also negative for money laundering 	
and receiving proceeds from crime, except for the financial sector which does not seem to 	be 	
correlated	. 	
Finally, being charged with money laundering or receiving proceeds from crime tends to be 
more corr	elated with being charged with economic, 	drug	 or other 	co	-offences, than violent 	co	-	
offences. 	It is important to	 deve	lop more knowledge	 on money laundering as it	 is harming 	
society 	economically and politically. H	aving more information 	on it 	might help to prevent these 	
offences. 	Moreover, the data on criminality is very limited which makes this 	research subject	 	
even more valuable.	 Although these findings contribute to more information on 	criminality 	
profiling which	 might be useful for	 policymakers	, it is important to have in mind th	at this study 	
is only based on reported criminality and 	no causality can	 thus	 be drawn. 	Furthermore	, the 	
results cannot apply to other countries than Norway.

67	 	
 
REFERENCES 	 	
Alalehto, T., & Larsson, D. (2012). Vem är den ekonomiske brottslingen?: En jämförelse 	
mellan länder och brottstyper. 	Sociologisk Forskning	, 49	(1), 25	–44. 	
https://doi.org/10.37062/	sf.49.18390	 	
Benson, M. L., & Kerley, K. R. (2001). 	LIFE COURSE THEORY AND WHITE	-COLLAR 	
CRIME. 	Contemporary Issues in Crime and Criminal Justice: Essays in Honor of Gil 	
Geis	, 121	–136.	 	
Buele, I., Nieves, J., & Cuesta, F. (2019). 	Money Laundering in Ecuador: Profile of the 	
Involved Person and the Socio	-Economic Impact. 	Journal of Legal, Ethical and 	
Regulatory Issues	, 22	(4), 1	–380.	 	
Cohen, L. E., & Felson, M. (1979). 	Social Change and Crime	 Rate Trends: A Routine Activity 	
Approach	. 44	(4).	 	
Dabney, D. A., Dugan, L., Topalli, V., & Hollinger, R. C. (2006). The Impact of Implicit 	
Stereotyping on Offender Profiling: Unexpected Results From an Observational Study 
of Shoplifting. 	Criminal Justice a	nd Behavior	, 33	(5), 646	–674. 	
https://doi.org/10.1177/0093854806288942	 	
Daly, K., & Tonry, M. (1997). Gender, Race, and Sentencing. 	Crime and Justice	, 22	, 201	–	
252.	 	
Dømt til fengsel for heleri og hvitvasking	. (2022, November 4). Økokrim. 	
https://www.okokrim.n	o/doemt	-til	-fengsel	-for	-heleri	-og	-hvitvasking.6567672	-	
549344.html	 	
Endringer i regler på hvitvaskingsområdet	. (2021, June 2). Finanstilsynet. 	
https://www.finanstilsynet.no/nyhetsarkiv/nyheter/2021/endringer	-i-regler	-pa	-	
hvitvaskingsomradet/

68	 	
 
Gilmour, N. (2016	). Understanding the practices behind money laundering 	– A rational choice 	
interpretation. 	International Journal of Law, Crime and Justice	, 44	, 1	–13. 	
https://doi.org/10.1016/j.ijlcj.2015.03.002	 	
Gottschalk, P. (2018). White	-Collar Crime Research. In L. Gunn	esdal (Ed.), 	White	-Collar 	
Crime in the Shadow Economy: Lack of Detection, Investigation and Conviction 
Compared to Social Security Fraud	 (pp. 1	–14). Springer International Publishing. 	
https://doi.org/10.1007/978	-3-319	-75292	-1_1	 	
Hansen, K. (2003). Education	 and the Crime	-Age Profile. 	The British Journal of Criminology	, 	
43	(1), 141	–168.	 	
Hawaii Criminal Justice Information Data Center. (1982). 	Demographic Characteristics of 	
Offender Population	—	Final Report	. 95224	, 1	–147.	 	
Heleri	. (2020, June 22). Jusinfo.no 	- presentert av Oslo Advokat RUV AS. 	
https://jusinfo.no/strafferett/heleri/heleri/	 	
Iversen, M. L. (2021, September 16). 	Hvitvasking	. Økokrim. 	
https://www.okokrim.no/hvitvasking.422268.no.html	 	
Jojarth, C. (2013). Money Laundering: Motives, Methods, Impact, and Countermeasures. In 	
Heinrich	-Böll	-Stiftung & R. Schönenberg (Eds.), 	Transnational Organized Crime	 (pp. 	
17	–34). Transcript Verlag. https://www.jstor.org/stable/j.ctv1fxh0d.5	 	
Ningtyas, V., I	smail, M., & Wahyudi, S. (2022). The Determinants of Money	-Laundering 	
Behavior in Indonesia. 	Journal of Indonesian Applied Economics	, 10	(1), 8	–16. 	
https://doi.org/10.21776/ub.JIAE.2022.010.01.2	 	
Riksadvokaten. (2018). 	Retningslinjer om heleri	-, hvitvaskings	- og selvvaskingssaker	. 	
https://www.riksadvokaten.no/wp	-content/uploads/2018/09/Retningslinjer	-for	-heleri	-	
hvitvasking	-og	-selvvaskingssaker.pdf

69	 	
 
Singh, K., & Best, P. (2019). Anti	-Money Laundering: Using data visualization to identify 	
suspicious activity. 	International Journal of Accounting Information Systems	. 	
https://doi.org/10.1016/j.accinf.2019.06.001	 	
Straffeloven. (1902). 	Almindelig borgerlig Straffelov	 (LOV	-1902	-05	-22	-10). 	Lovdata. 	
https://lovdata.no/NLO/lov/1902	-05	-22	-10	 	
The Anti	-Money Laundering Act. 	(2018). 	Act relating to Measures to Combat Money 	
Laundering and Terrorist Financing	 (Lovdata). LOV	-2018	-06	-01	-23. 	
https://lovdata.no/NLE/lov/2018	-06	-01	-23	 	
The Penal Code. (2005). 	The Penal Code	 (LOV	-2005	-05	-20	-28). 	Lovdata. 	
https://lovdata.no/NLE/lov/2005	-05	-20	-28	 	
Veileder til hvitvaskingsloven	. (2022, November 15). Finanstilsynet. 	
https://www.finanstilsynet.no/nyhetsarkiv/rundskriv/2022/veileder	-til	-	
hvitvaskingsloven/	 	
Verbeek, M. (2017). 	A guide to modern econometrics	 (Fifth edition). Wiley Custom.	 	
Weisbur	d, D., Waring, E., & Chayet, E. F. (2002). White	-Collar Crime and Criminal Careers. 	
Contemporary Sociology	, 31	(5), 599. https://doi.org/10.2307/3090075

70	 	
 
APPENDIX 1	: DECOMPOSITION OF THE OUTCOME VARIABLE	 	
 	
Variable name	 	
Money 
laundering 
or 	proceeds 	
from crime	 	
Drug	 offence	 	Economic 
offence	 	
Violen	t 	
offence	 	
Other 
offence	 	
Only money laundering	 	Yes	 	No	 	No	 	No	 	No	 	
Drug	 	Yes	 	Yes	 	No	 	No	 	No	 	
Drug and economic	 	Yes	 	Yes	 	Yes	 	No	 	No	 	
Drug, economic and violen	t 	Yes	 	Yes	 	Yes	 	Yes	 	No	 	
Drug, economic, violen	t and 	
other	 	Yes	 	Yes	 	Yes	 	Yes	 	Yes	 	
Economic	 	Yes	 	No	 	Yes	 	No	 	No	 	
Economic and violen	t 	Yes	 	No	 	Yes	 	Yes	 	No	 	
Economic, violen	t and other	 	Yes	 	No	 	Yes	 	Yes	 	Yes	 	
Violen	t 	Yes	 	No	 	No	 	Yes	 	No	 	
Violen	t and other	 	Yes	 	No	 	No	 	Yes	 	Yes	 	
Other	 	Yes	 	No	 	No	 	No	 	Yes	 	
Drug and other	 	Yes	 	Yes	 	No	 	No	 	Yes	 	
Drug and violen	t 	Yes	 	Yes	 	No	 	Yes	 	No	 	
Economic and other	 	Yes	 	No	 	Yes	 	No	 	Yes	 	
Drug, violen	t and other	 	Yes	 	Yes	 	No	 	Yes	 	Yes	 	
Drug, economic and other	 	Yes	 	Yes	 	Yes	 	No	 	Yes

71	 	
 
APPENDIX 2	: LOGIT ESTIMATION OF AVERAGE MARGINAL EFFECTS OF THE OUTCOME ALTERNATI	VES WHEN 	
CHARGED WITH MONEY LAUNDERING 	OR	 RECEIVING PROCEEDS FROM CRIME	 	
 	
Panel A	 	-----------------------------------------------------------------------------------------------------------------------------	--------------------------------	 	 	(1) (2) (3) (4) 	 	(5) (6) (7) (8) 	 	 	Only ML D D E D E V D E V O E E V E V O 	 	--------------------------------------	-----------------------------------------------------------------------------------------------------------------------	 	Gender 0.00019*** 0.00008*** 0.00008*** 0.00003*** 0.00034*** 0.00011***	 	0.00002*** 0.00008***	 	 	(0.00000) (0.00000) (0.00000) (0.00000) (0.00001) (0.00000) (0.00000) (0.00000) 	 	 Age 0.00016*** 0.00022*** 0.0	0034*** 0.00007*** 0.00060*** 0.00013*** 0.00002*** 0.00007***	 	 	(0.00001) (0.00001) (0.00001) (0.00000) (0.00001) (0.00000) (0.00000) (0.00000) 	 	 Age*Age 	 	-0.00003*** 	-0.00003*** 	-0.00005*** 	-0.00001*** 	-0.00009*** 	-0.00002*** 	-0.00000*** 	-0.00001***	 	 	(0.00000) (0.00000) (0.00000) (0.00000) (0.00000) 	 	(0.00000) (0.00000) (0.00000) 	 	 Immigrant 0.00011*** 	-0.00001** 	-0.00008*** 	-0.00001*** 	-0.00009*** 0.00004*** 0.00001*** 0.00002***	 	 	(0.00000) 	 	(0.00000) (0.00000) (0.00000) (0.00000) (0.00000) (0.00000) (0.00000) 	 	 City 0.00005*** 0.00006*** 0.00009*** 0.00002*** 0.00011*** 0.00005***	 	0.00001*** 	 	0.00002***	 	 	(0.00001) (0.00000) (0.00001) (0.00000) (0.00001) (0.00000) (0.00000) (0.00000) 	 	 Education 	-0.00001*** 	-0.00000*** 	-0.00001*** 	-0.00000*** 	-0.00001*** 	-0.00001*** 	-0.00000*** 	-0.00000***	 	 	(0.00000) (0.00000) (0.00000) (0.00000) (0.00000) (0.00000) (0.00000) (0.00000) 	 	 Income 0.00000 0.00001*** 0.00006*** 0.00001*** 	 	0.00011*** 	-0.00000*** 	-0.00000*** 	-0.00001***	 	 	(0.00000) (0.00000) (0.00000) (0.00000) (0.00000) (0.00000) (0.00000) (0.00000) 	 	 Income*Income 	-0.00000 	-0.00001*** 	-0.00002*** 	-0.00000*** 	-0.00004*** 	-0.00000 	-0.00000 	-0.00000***	 	 	(0.00000) (0.00000) (0.00000) (0.00000) (0.00000) (0.00000) (0.0	0000) (0.00000) 	 	 Unemployed 0.00004*** 	-0.00000 	-0.00003*** 	-0.00000* 	-0.00005*** 0.00000 	-0.00000 0.00001***	 	 	(0.00001) (0.00000) (0.0000	0) (0.00000) (0.00001) (0.00000) (0.00000) (0.00000) 	 	 Employed state sector 	-0.00023*** 	-0.00017*** 	-0.00030*** 0.00000 	-0.00049*** 	-0.00019*** 	-0.00004*** 	-0.00008***	 	 	 	(0.00001) (0.00001) (0.00002) (.) (0.00003) (0.00001) (0.00001) (0.00001) 	 	 Employed municipal sector 	-0.00019*** 	-0.00015*** 	-0.00026*** 	-0.00007*** 	-0.00045**	* 	-0.00017*** 	-0.00003*** 	-0.00008***	 	 	(0.00001) (0.00001) (0.00001) (0.00001) (0.00002) (0.00001) (0.00000) (0.00001) 	 	 Employed Private 	-0.00009*** 	 	-0.00009*** 	-0.00016*** 	-0.00003*** 	-0.00026*** 	-0.00009*** 	-0.00002*** 	-0.00003***	 	Construction (0.00001) (0.00001) (0.00001) (0.00000) (0.00001) (0.00001) 	 	(0.00000) 	(0.00000)

72	 	
 	
Employed Private 	-0.00008*** 	-0.00007*** 	-0.00017*** 	-0.00003*** 	-0.00032*** 	-0.00008*** 	-0.00002*** 	-0.00004***	 	Hotel and Restaurant (0.00001) (0.00001) (0.00001) (0.	00001) (0.00002) (0.00001) (0.00000) (0.00001) 	 	 Employed Private 	-0.00011*** 	-0.00011*** 	-0.00022*** 	-0.00005*** 	-0.00037*** 	-0.00012*** 	-0.00002*** 	-0.00005***	 	Manufacturing (0.00000) (0.00000) (0.00001) (0.00000) (0.00001) (0.00000) 	 	(0.00000) (0.00000) 	 	 Employed Private 	-0.00022*** 	-0.00011*** 	-0.00021*** 	-0.00004** 	 -0.00037*** 	-0.00016*** 	-0.00002** 	-0.00005***	 	Finance and Real Estate (0.00002) (0.00002) (0.00003) (0.00001) (0.00005) (0.00002) 	 	(0.00001) (0.00001) 	 	 Employed Private 	-0.00014*** 	-0.00010*** 	-0.00020*** 	-0.00004*** 	-0.00032*** 	-0.00012*** 	-0.00002*** 	-0.00004***	 	Services (0.00001) (0.00000) (0.00001) (0.00000) (0.00001) (0.00001) 	 	(0.00	000) (0.00000) 	 	 Employed Private 	-0.00015*** 	-0.00011*** 	-0.00018*** 	-0.00004*** 	-0.00035*** 	-0.00013*** 	-0.00002*** 	-0.00005***	 	Other (0.00001) (0.00001) (0.00002	) (0.00001) (0.00003) (0.00001) (0.00000) (0.00001) 	 	 Self	-employed Private 	-0.00002 	-0.00010*** 	-0.00015*** 	-0.00003*** 	-0.00016*** 	-0.00002* 	-0.00001* 	-0.00002* 	 	Construction	 	(0.00002) (0.00002) (0.00002) (0.00001) (0.00002) (0.00001) (0.00000) (0.00001) 	 	 Self	-employed Private 0.00010*** 	-0.00007* 	-0.00014** 0.00000 	-0.00028** 	 	-0.00007* 	-0.00001 	-0.00003 	 	Hotel and Restaurant (0.00003) (0.00003) (0.00005) (.) (0.00009) (0.00003) 	 	(0.00001) (0.00002) 	 	 Self	-employed Private 0.00006*** 	 -0.00007*** 	-0.00013*** 	-0.00002** 	-0.00016*** 0.00001 	-0.00001* 	-0.00001 	 	Manufacturing (0.00001) (0.00001) (0.00002) (0.00001) (0.00002) (0.00001) 	 	(0.00000) (	0.00001) 	 	 Self	-employed Privat 	-0.00007 0.00005 0.00000 0.00000 0.00010 0.00010* 0.00000 0.00000 	 	Finance and Real Estate (0.00017) (0.00006) (.) 	 (.) (0.00009) (0.00005) (.) (.) 	 	 Self	-employed Private 	-0.00005*** 	-0.00009*** 	-0.00021*** 	-0.00004*** 	-0.00027*** 	-0.00006*** 	-0.00001* 	-0.00003***	 	Services 	(0.00001) (0.00001) (0.00003) (0.00001) (0.00003) (0.00001) (0.00000) (0.00001) 	 	 Self	-employed Private	 	-0.00017*** 	-0.00014*** 	-0.00021*** 	-0.00003*** 	 -0.00025*** 	-0.00014*** 	-0.00003** 	-0.00007***	 	Other	 	(0.00002) (0.00002) (0.00003) (0.00001) (0.00002) (0.00002) (0.00001) (0.00002) 	 	-----------------------------------------------------------------------------------------------------------------------------	------	--------------------------	 	Number of observations	 	72995982 72995982 72987727 66040485 72995982 72995982 72987727 72987727 	 	Dependent mean 	0.000168 0.0000636 0.0000	753 0.0000169 0.000122 0.0000804 0.0000106 0.0000269 	 	
Pseudo 	������2 	0.0621 0.126 0.190 0.168 0.236 0.0905 0.127 0.154	 	Log	-likelihood 	-111438 	-43249 	-46738 	-11151 	-68036 	-55664 	-8416 	-19124	 	 	--------------------------------------------------------------------------------------------------	-----------------------------------------------------------	 	Standard errors in parentheses	 	* p<0.05, ** p<0.01, *** p<0.001	 	ML = Money laundering and receiving proceeds from crime, D = drug offence, V = violent offence, E = economic offence, O = oth	er offe	nce	 	All the models are controlled for year	-fixed effects and missing values on education

73	 	
 	
Panel B	 	-----------------------------------------------------------------------------------------------------------------------------	--------------------------------	 	 	(9) (	10) (	11) (	12) 	 	(13) (	14) (	15) (	16) 	 	 	V V O O D O D V E O D V O D E O 	 	--------------------------------------	-----------------------------------------------------------------------------------------------------------------------	 	Gender 0.00003*** 0.00003*** 0.00011*** 0.00013*** 0.00002*** 0.00014***	 	0.00008*** 0.00043***	 	 	(0.00000) (0.00000) (0.00000) (0.00000) (0.00000) (0.00000) (0.00000) (0.00001) 	 	 Age 0.00002*** 	0.00002*** 0.00007*** 0.00028*** 0.00003*** 0.00016*** 0.00009*** 0.00131***	 	 	(0.00000) (0.00000) (0.00000) (0.00001) (0.00000) (0.00001) (0.00000) (0.0	0002) 	 	 Age*Age 	-0.00000*** 	-0.00000*** 	-0.00001*** 	-0.00004*** 	-0.00000*** 	-0.00002*** 	-0.00001*** 	-0.00019***	 	 	(0.00000) (0.00000) (0.00000) (0.0000	0) (0.00000) (0.00000) (0.00000) (0.00000) 	 	 Immigrant 0.00001*** 0.00001*** 0.00002*** 	-0.00003*** 0.00000*** 	-0.00000 0.00001*** 	-0.00035***	 	 	 	(0.00000) (0.00000) (0.00000) (0.00000) (0.00000) (0.00000) (0.00000) (0.00001) 	 	 City 0.00001*** 0.00001** 0.00002*** 0.00005*** 0.00001*** 0.00004**	* 0.00002*** 0.00018***	 	 	(0.00000) (0.00000) (0.00000) (0.00000) (0.00000) (0.00000) (0.00000) (0.00001) 	 	 Education 	-0.00000*** 	-0.00000*** 	 	-0.00000*** 	-0.00000*** 	-0.00000*** 	-0.00001*** 	-0.00000*** 	-0.00002***	 	 	(0.00000) (0.00000) (0.00000) (0.00000) (0.00000) (0.00000) (0.00000) (0.00000) 	 	 	Income 0.00000* 	-0.00000 	-0.00000*** 0.00001*** 0.00001*** 	-0.00001*** 	-0.00000*** 0.00001***	 	 	(0.00000) (0.00000) (0.00000) (0.00000) (0.	00000) (0.00000) (0.00000) (0.00000) 	 	 Income*Income 	-0.00000*** 	-0.00000 	-0.00000*** 	-0.00001*** 	-0.00000*** 	-0.00000*** 	-0.00000*** 	-0.00003***	 	 	(0.000	00) (0.00000) (0.00000) (0.00000) (0.00000) (0.00000) (0.00000) (0.00000) 	 	 Unemployed 0.00000* 0.00000 0.00001*** 0.00001 	-0.00000 0.00000 	-0.000	01** 	-0.00005***	 	 	(0.00000) (0.00000) (0.00000) (0.00000) (0.00000) (0.00000) (0.00000) (0.00001) 	 	 Employed state sector 	-0.00002*** 	-0.00004*** 	-0.00011*	** 	-0.00019*** 	-0.00003*** 	-0.00021*** 	-0.00008*** 	-0.00099***	 	 	(0.00000) (0.00001) (0.00001) (0.00001) (0.00001) (0.00001) (0.00001) (0.00004) 	 	 Employed municipal sector 	-0.00002*** 	-0.00002*** 	-0.00009*** 	-0.00018*** 	 -0.00002*** 	-0.00017*** 	-0.00007*** 	-0.00080***	 	 	(0.00000) (0.00000) (0.00001) (0.00001) (0.00000) (0.00001) (0.00001) (0.00002) 	 	 Employed Private 	-0.00001*** 	-0.00001*** 	-0.00004*** 	-0.00009*** 	-0.00001*** 	-0.00010*** 	-0.00004*** 	-0.00048***	 	Construction (0.00000) (0.00000) (0.00000) (0.00001) (0.00000) (0.00001) 	 	(0.00	000) (0.00001) 	 	 Employed Private 	-0.00001** 	-0.00001*** 	-0.00004*** 	-0.00008*** 	-0.00001*** 	-0.00009*** 	-0.00004*** 	-0.00057***	 	Hotel and Restaurant (0.00000) (0.00000) (0.00001) (0.00001) (0.00000) (0.00001) 	 	(0.00000) (0.00003) 	 	 Employed Private 	-0.00001*** 	-0.00002*** 	-0.00006*** 	-0.00012*** 	 -0.00002*** 	-0.00012*** 	-0.00006*** 	-0.00064***	 	Manufacturing (0.00000) (0.00000) (0.00000) (0.00000) (0.00000) (0.00000) 	 	(0.00000) (0.00001)

74	 	
 	
Employed Private 	-0.00001* 0.00000 	-0.00009*** 	-0.00016*** 	-0.00002* 	-0.00019*** 	-0.00007*** 	-0.00073***	 	Finance and Real Estate (0.00001) (.) (0.00001) (0.00003) (0.00001) (0.00002) 	 	(0.00	002) (0.00007) 	 	 Employed Private 	-0.00001*** 	-0.00002*** 	-0.00006*** 	-0.00012*** 	-0.00001*** 	-0.00013*** 	-0.00005*** 	-0.00062***	 	Services (0.00000) (0.00000) (0.00000	) (0.00001) (0.00000) (0.00001) (0.00000) (0.00002) 	 	 Employed Private 	-0.00001* 	-0.00002*** 	-0.00005*** 	-0.00011*** 	-0.00001** 	-0.00012*** 	-0.00005*** 	-0.00059***	 	Other 	 	(0.00000) (0.00000) (0.00001) (0.00001) (0.00000) (0.00001) (0.00001) (0.00004) 	 	 Self	-employed Private 	-0.00000 	-0.00001 	-0.00000 	-0.00009*** 0.00000 	 	-0.00002* 	-0.00003*** 	-0.00030***	 	Construction (0.00000) (0.00000) (0.00001) (0.00001) (.) (0.00001) 	 	(0.00001) (0.00002) 	 	 Self	-employed Private 	-0.00000 	 -0.00001 0.00000 	-0.00010* 	-0.00000 	-0.00003 	-0.00001 	-0.00063***	 	Hotel and Restaurant (0.00001) (0.00001) (0.00002) (0.00004) (0.00001) (0.00002) 	 	(0.00001) (0.00016) 	 	 Self	-employed Private 0.00000 0.00000 0.00000 	-0.00006*** 	 -0.00001* 	-0.00001 	-0.00004*** 	-0.00027***	 	Manufacturing (0.00000) (0.00000) (0.00001) (0.00001) (0.00000) (0.00001) 	 	(0.00001) (0.00002) 	 	 Self	-employed Privat 0	.00000 0.00000 0.00003 0.00003 0.00000 0.00006 0.00000 	-0.00014 	 	Finance and Real Estate (.) (.) (0.00005) (0.00006) (.) (0.00005) 	 	 (.) (0.00021) 	 	 Self	-employed Private 	-0.00001* 	-0.00002* 	-0.00003** 	-0.00007*** 	-0.00000 	-0.00006*** 	-0.00004*** 	-0.00051***	 	Services 	(0.00001) (0.00001) (0.00001	) (0.00001) (0.00000) (0.00001) (0.00001) (0.00003) 	 	 Self	-employed Private	 	-0.00001 	-0.00002* 	-0.00006*** 	-0.00013*** 	-0.00001 	-0.00011*** 	-0.00004*** 	-0.00049***	 	Other	 	 	(0.00001) (0.00001) (0.00001) (0.00002) (0.00000) (0.00001) (0.00001) (0.00003) 	 	--------------------------------------------------------------------------------------------------------------	-----------------------------------------------	 	Number of observations 	 	72987727 71635699 72995982 72995982 72622496 72995982 72987727 72995982 	 	Dependent mean 	 	0.0000108 0.0000116 0.0000493 0.0000605 0.00000799 0.000065	6 0.0000226 0.000217 	 	
Pseudo 	������2 	 0.114 0.121 0.101 0.164 0.136 0.133 0.163 0.247 	 	Log	-likelihood 	-8689 	-9021 	 	-35309 	-39546 	-6383 	-44133 	-16185 	-112651	 	-----------------------------------------------------------------------------------------------------------------------------	--------------------------------	 	Standard errors in parentheses	 	* p<0.05, ** p<0.01, *** p<0.001	 	ML = Money laundering and receiving proceeds from crime, D = drug offence, V = violent offence, E = economic offence, O = oth	er offence	 	All the models are controlled for year	-fixed effects and	 missing values on education.

Husholdningers betalingsvillighet for å unngå strømbrudd	 	
 	
Vista Analyse	 | Rapport 2022/23	 	3 	
 
 	
 
 
 
 	
Vista Analyse AS	 	
Meltzers gate 4	 	
0257 Oslo	 	
 
post@vista	-analyse.no	 	
vista	-analyse.no