Quantifying
Financial Crime in the
John Walker
Director, John Walker Crime Trends Analysis;
Research Fellow, Centre for Transnational Crime Prevention,
For the Financial Crime Forum Asia Pacific
5-7 June 2006
I’ve grown up in a period that coincides with the development of computers and computer science, from the point where they became “useful”, even though they were the size of a small football stadium. The chief accountant at my first jobs had a computer to do the company’s accounts and print out the payslips. It was difficult to imagine anything else they could be good for, because they were just adding up machines attached to high-volume printers.
I was also lucky enough to be born a number cruncher at the right time, and I learnt to program these machines, using long-forgotten languages like Algol and Fortran, and I soon discovered that you could get them to do far cleverer things than just add up the wage packets. I used matrix algebra to work out how much extra traffic you would get along a certain road if you allowed a certain development to take place. Few of my school mates ever imagined that there was a real-world use for matrix algebra! I even used probability theory to see if space satellites would achieve the scientific experiment they were designed for. This now allows me to say, with at least some authority, “It’s not rocket science” whenever I think a problem is solvable and others think it’s too hard.
From there, I worked in regional government in the UK and Australia, using statistics to try to solve some riddle that needed a town planning solution (e.g. why are there traffic jams at “that” particular place every time our football team plays at home, or what will happen to the jobs situation here when the baby-boomers all leave school and start looking for work? Working for the Melbourne Metropolitan planning authority in the early 1970s, I was given some local area crime statistics, and asked to identify the regional planning implications. I found this stuff so fascinating that I came back to it a few years later, and was offered a research position at the Australian Institute of Criminology.
Most criminologists are not very numerate, so I found myself in demand and learnt on the job. Part of my university training, however, was in economics, and I started to feel that there was a vast unexplored side to criminology. It was clear that there are crimes perpetrated for economic gain, but what about the other aspects of it? I was again blessed with good timing, because this was the time when criminologists were experimenting with randomised surveys as an alternative to police statistics to better measure the extent of crime. I insisted that we ought to ask about the costs to the victims – how much property was stolen? – did you have to go for medical treatment? – did you have to take time off work? Etc. We discovered that this was indeed an important question. People often didn’t even report a crime to the police if it didn’t cost them much, or if the costs of reporting to police outweighed the cost of the crime. At the other end of the scale, politicians, we found, didn’t do much about crime unless there was political mileage for them – but you could make them listen if you showed them how high the costs of crime actually were. They then had a political “angle” to work with – they could tell the electors that their policy would save millions of their dollars, but more importantly they could convince their own Treasurers that their policy would be money well spent..
As well as individuals, I surveyed businesses about the
costs of crime, and discovered that many of them are more worried about the
damage to their reputations than to their immediate bottom line, and prefer not
to report crimes to police – and the bigger the crime the more likely this
appeared to be! Presenting these
findings to politicians in
In the early 1990s, I began working as a consultant, and on
the basis of my costs of crime work I was asked to try to estimate the extent
of money laundering in
- - - - - - -
In this
paper, I’ll begin by giving a very simple definition of financial crime, but
one that has very wide-ranging implications.
I’ll then
present an analysis of the size of the problem – globally and in the Asia
Pacific region.
I’m then
going to have an audience question and answer session, but don’t worry – I’ve
thought of some really good questions that you should ask me, and I’m going to
answer them straight away, so you don’t go home disappointed with me.
Then I’m
going to try to show the implications of these estimates, in terms of
developing evidence-based strategies against financial crime.
I’m going
to end up with a few words about how I would go about it – remembering that it
isn’t rocket science – it IS achievable.
- - - - - - -
I’m going
to start with a short list of what might be included as financial crimes. I’m sure you’ll be able to think of others,
but for the purposes of this presentation, a very short list will be enough.
Financial
fraud
The adverse
impact of financial fraud, not only on individuals and the commercial sector
but even on national economic systems, is increasing rapidly worldwide. Left
unchecked, fraud could lead to the financial ruin of people and commercial
enterprises as well as seriously damage economic systems.
Counterfeit
currency
A major
challenge confronting societies today is the growing integration of the world’s
economies, which has sparked myriad transnational criminal schemes designed to
exploit financial payment systems. Fraud schemes involving credit cards,
identity theft, bank fraud or the most fundamental financial crime – the
production of counterfeit currency – recognise no national boundaries. Although counterfeiting has diminished substantially
since the establishment of special law enforcement units assigned to fight it,
the crime continues to present a potential danger to national economies and
financial losses to consumers. Recent developments in photographic and computer
technology, as well as printing devices, have made the production of
counterfeit money relatively easy, thereby increasing the threat.
Intellectual
property (IP) crime
Intellectual
property (IP) crime is the generic term for a wide range of counterfeiting and
piracy offences. These include trademark, patent and copyright infringement.
The last decade has seen a steady worldwide increase in these types of criminal
offences. One reason is the ready availability of modern technology to
counterfeiters, who systematically use it to infringe trademarks and breach
copyrights. Counterfeiting is so widespread that few legitimately manufactured
goods are not copied in one form or another and the rights of the owners
infringed.
Payment
cards
As the use
of payment cards continues to increase, it is likely that the frequency and
extent of fraud will also continue to expand. The payment card industry and law
enforcement community are constantly working to develop enhanced production and
security measures to deter payment card fraud, but criminals continue to devise
more sophisticated methods to override such features.
Money
laundering
The international police community is aware that there is a need to achieve major results in the struggle against the financial criminal activities related to the organized criminal groups. But now here’s the problem: money laundering is sourced from the proceeds of both financial and non financial types of crimes. So if we’re serious about developing evidence-based policies to combat financial crimes, we’re going to have to collect evidence about the whole range of large-scale income-generating crimes, including all forms of organised crime – prostitution, drugs, people-smuggling, gun-running – the lot! These aren’t of themselves financial crimes, but they generate the income that then requires further – financial – crimes in order to keep the money and the criminals themselves safe from justice.
Conclusion:
Any type of major income-generating crime can therefore
contribute to the problem of financial crime. To understand financial crime, we
have to understand All major crime.
- - - - - - -
Just as in local politics, where properly-directed political action is much more likely if you can identify which problems are the costliest, so it is in international and global politics. The right strategies against financial crime will not be found by merely acting on knee-jerk reactions, simplistic assumptions or by following a dominant viewpoint. We need our policies to be based on evidence. We need them to be focused on the most serious problems – the priority areas. Any other approach is wasteful of scarce crime prevention resources. Therefore we need to be able to identify the most serious problems.
But it is not possible to quantify financial crime by looking only at banking and other finance data because
· Criminal origins can rarely be identified.
· Layering and placement processes make it difficult to separate money obtained from crime from legitimate funds.
We need to know:
· How much crime there is (and where and what sorts of crimes?)
· How much profit is made from the crimes?
· How much of the profit is laundered?
· Where does it go to, and
· How much trouble does it cause?
Financial crime is a whole set of different problems, on different economic scales, in different places. You can’t prioritise a set of problems without knowing what they are, how bad they are and where they hurt most, and you can’t develop effective financial crime prevention programs without knowing what the priority areas are.
Conclusion:
By focussing on financial transactions, we are mostly looking in the
wrong direction. The only
successful approach to quantifying Financial Crime will be to start from
estimates of the profits of criminal enterprises.
- - - - - - -
Unfortunately, this means we have to ask questions that are bigger than most criminologists ever dare to ask.
·
How
much crime is there around the world, and how much of it is based in the Asia
Pacific region?
·
How
big are crime profits around the world, and where are they generated?
·
What
factors make crime more profitable in some countries than others?
·
What
factors make some countries more attractive to ML than others?
·
How
much money is laundered each year around the world, and how much of it is
generated in, or laundered through, the Asia Pacific region?
·
How much harm is caused by crime and ML in the
Asia Pacific region, and who suffers most?
Conclusion:
These are the sorts of questions that economists – not criminologists –
commonly ask and answer.
It is neither rocket science nor is it impossible.
- - - - - - -
There are a number of options traditionally used by economists. The first one is generally used by people who are being harassed by journalists who need “a figure before my deadline”.
Guess - The “Wet finger” approach – you wet your finger and see which way the wind is blowing.
The second approach can be called the Top down approach: - examples include
· What proportion of global GDP is likely to be Proceeds of Crime?
· What proportion of global finance is held in dubious “offshore” accounts?
This approach is an improvement on the first, and has the advantage of only really needing two numbers, one of which is an actual and credible statistic, such as the global GDP estimate, and the other is generated possibly by option 1 but may be derived from real analysis. Either way, since at least half of the calculation has credibility, it is a big step forward.
The third approach is, rather inelegantly called “Bottom up”, and involves the use of the five-step logic.
· How much crime, proceeds, ML generated;
· what offences generate the money;
· how is it laundered;
· where does it go?
· How much trouble does it cause?
A few countries, including the
This approach is far superior, because it promises to identify those problem areas that should be priority targets for action, and can therefore be the basis for a real evidence-based approach to financial crime prevention. The downside is that is it a little more data-hungry that the other approaches. Well – since we are being honest with each other today, it is a LOT more data hungry than the other approaches.
There are a couple of important points that are worth making here.
Firstly, Triangulation – cross-checking with other data – is an important technique. If your data suggest that $X are generated by crime in a given country, but you also know that the economy of that country cannot support such a figure, then you have to reconsider whether $X is correct. This could mean finding a third figure – triangulation – that can help decide where the true fogures are likely to be.
Secondly, and this will offend almost all statisticians in the Forum, Accuracy is (almost) meaningless; credibility is everything. We waste our time trying to produce perfect statistics on financial crime. We will never achieve it, and if we wait for perfect statistics before we take evidence-based action against financial crime, we will be waiting for ever. The official data are mostly – as a famous New York District Attorney once delicately put it – crap, but there are large numbers of knowledgeable people out there in the world, with plenty of evidence and wisdom to analyse it. What is of fundamental importance, however, is whether the estimate we come up with is credible. If its is credible, then politicians of good intent will find them difficult to ignore, and evidence-based political action will (or at least might!) follow.
Conclusion:
Only the bottom-up approach, beginning with estimates of criminal
proceeds, with triangulation across other statistics, shows any promise
of producing meaningful estimates
- - - - - - -
Here are some examples of the top-down approach. They are at the very least useful as points of triangulation, and at best are worthy estimates that can be used to determine at what level to pitch an evidence-based financial crime prevention programme.
The first one is the notorious estimate of money laundering
of 2-5% of global GDP, which comes from Michel Camdessus, then head of the IMF,
in 1996. While the IMF has never
produced any supporting analysis for this figure, we should at least feel sorry
for Camdessus, who was at the time being hounded by journalists for some sort
of figure. His estimate clearly has some
credibility, since it is still repeated frequently by journalists who don’t
have the time to do any better, as well as professionals in the financial
advice industry who are trying to establish the value of their own
services. 2-5% is globally equivalent to
US $1-2.5 trillion; AU$14.7 – 36.7
billion for
The second top-down approach is that of estimating the size
of each country’s “shadow economy”. This
is, effectively, income generated in a country which is not reported to the
taxation authorities. Professor
Friedrich Schneider or the University of Linz, Austria, has done some
interesting work for the World Bank, suggesting that globally US$6.6 trillion
of untaxed income is generated – some of which is probably generated by
crime. His estimates are AU$20 billion
for
A third approach was presented last year in Raymond Baker’s
very interesting book Capitalism’s Achilles Heel. He analysed the offshore holdings of very
high net-worth individuals around the world, based on Merrill Lynch’s World
Wealth Report. His total for the Asia
Pacific region, including north and south America and
Holdings of
High Net-worth Individuals Offshore (Raymond Baker 2005)
|
Region |
Total Holdings (US $trillions) |
% Offshore |
Amount Offshore (US $trillions) |
|
|
5.7 |
30 |
1.7 |
|
|
7.4 |
34 |
2.5 |
|
|
3.6 |
31 |
1.1 |
|
|
8.8 |
31 |
2.7 |
|
|
1.1 |
27 |
0.3 |
|
|
0.6 |
33 |
0.2 |
|
World |
27.2 |
31 |
8.5 |
These latter two estimates can therefore serve as “upper limits” on the extent of criminal proceeds and financial crime – the real figures must be something less than $3.3 US trillion a year.
- - - - - - -
My approach in
|
Crime
Type |
Recorded
Crime 2001 (000s) |
Estimated
Number
of Incidents
2001
(000s) |
Property
Stolen
& Damaged
$/Incident |
Total
Property
Losses $
million |
Estimated
%
Proceeds (1995
Survey) |
Estimated Proceeds $
million |
%
Laundered (a=
1995 Survey b=2005
survey) |
Implied
Money
Laundering
AUD
million |
|
Robbery |
27 |
168 |
800 |
134 |
80.0 |
107 |
30.0a |
32 |
|
Residential
Burglary |
275 |
819 |
1,100 |
901 |
80.0 |
721 |
10.0a |
72 |
|
Non-Res
Burglary |
160 |
176 |
2,400 |
422 |
80.0 |
338 |
10.0a |
34 |
|
Theft
of Motor Vehicles |
140 |
147 |
4,000 |
588 |
80.0 |
470 |
35.0a |
165 |
|
Shoplifting |
73 |
7,304 |
100 |
730 |
80.0 |
584 |
5.0a |
29 |
|
Theft
from Motor vehicle |
266 |
956 |
270 |
258 |
80.0 |
206 |
5.0a |
10 |
|
Other
Theft |
390 |
1,769 |
200 |
354 |
80.0 |
283 |
5.0a |
14 |
|
Criminal
Damage |
319 |
1,914 |
350 |
670 |
1.0 |
7 |
0a |
0 |
|
Arson |
|
|
|
1,350 |
1.0 |
14 |
0a |
0 |
|
Fraud |
|
|
|
5,880 |
75.0 – 90.0 |
4,851 |
69.4b |
3,367 |
|
Drugs |
|
|
|
|
|
3,500 |
83.3b |
2,915 |
|
Total
Offences |
|
|
|
11,287 |
|
11,081 |
|
6,282 |
- - - - - - -
There is no reason why a similar approach would not be successful in other countries. Most countries produce official crime statistics based on police data. Many countries have now conducted crime victims surveys, and their findings can be extended to other, similar, countries that have not conducted surveys of their own. Insurance data exist for many countries, which can be triangulated against the crime and victimisation data. There is now a wealth of economic and demographic data available on virtually every country on the planet, which can be used to moderate or calibrate the estimates. And finally there is a growing body of information on the extent of known money laundering and other financial crimes, that can be used to construct a global model of financial crime and money laundering flows.
Just for fun, some years ago, I actually tried this, using mostly UN data and some heroic assumptions about, for example, the way the economy of a country would impact on the profitability of crime there, or the way the financial regulations and company laws might attract or deter money launderers from using a particular laundering route. The results were quite revealing.
- - - - - - -
In 1998, my model came up with a global figure of almost $US3 trillion per year – rather more than Camdessus’ maximum figure, but in the same ballpark. Because the model works “bottom-up”, the output is presented country-by-country, and can therefore be aggregated to regional level: - the figure for the Asia Pacific Region is around $US1.8 trillion a year , and perfectly consistent with the “upper limit” of $US3.3 trillion mentioned earlier. This suggests that the proceeds of crime amount to just over half of the value of the shadow economies in the Asia Pacific – is this credible? Maybe!
But the model clearly suggested that any fixation the world’s anti-moneylaundering agencies had about the illicit proceeds of drug trafficking were misguided, because fraud and other organised crime is far more important than all the illicit drugs put together, and that the tiny countries like Nauru and the Cayman Islands were simply cashing in on a problem created by the rich countries themselves – not the cause of the problem, per se. At the time, the suggestion that fraud could be far more important a generator of criminal proceeds was not well received in certain countries, but there has been a considerable shift of opinion since the events of September 11 2001. (I could say I told them so, but I will refrain).
Estimates of the Generation of Laundered Money
by Crime Type and Region ($
- - - - - - -
So we can put these estimates for the Asia Pacific Region
together, and we can see that in spite of its shortcomings, the model’s results
are consistent with the top-down estimates. Conclusion: Financial crime in the Asia Pacific is worth
somewhere in the region of $
- - - - - - -
I want to step back a bit now, and look at the concept of triangulation. My money laundering model produces a figure for global money laundering, which can be disaggregated by region or even down to country level. How can we check to see if it is producing credible results?
Again, as so often in my career, I got lucky, and last year I was asked to try to construct an economic model of the global illicit drugs trades. This invitation came after many years of frustration caused by high-powered researchers who said “it couldn’t be done”. Well, we did it, and the results were presented in last year’s World Drug Report, and appear to be credible. The estimated global income to illicit drug producers and traffickers is around a third of a trillion $US, and can be disaggregated to the regional, or even (with care!) country level. This figure is clearly compatible with the results of my money laundering model, but the important thing is that it demonstrates the potential of a complete methodology for analysing transnational crime. This model is based on a blend of official statistics, expert knowledge and triangulation, much of it derived from an annual international expert survey.
- - - - - - -
The UNODC now has a model of the illicit drugs trades around the world, which can be used to identify patterns of trafficking, degrees of harm caused by the trades, in the countries of origin, transit and destination, and priorities for action.
This U.N. Model independently suggests that the global illicit drugs trades are worth around $US 300bn per annum (Income to traffickers). The profit margins must be high, to cover the risk, so perhaps as much as 25% of this $US300bn is launderable profit: - $US75bn. The Walker Moneylaundering Model – even with its “dodgy data and heroic assumptions” - estimated around $US69bn in 1998.
If we can do this for illicit drug trafficking, why can it not be done for the other major economic and financial crimes?
- - - - - - -
Another potential triangulation analysis has been developed
by Professor John Zdanowicz at the
- - - - - - -
Another source of valuable triangulation information is
Schneider’s shadow economy analysis. His
analysis in 2004 produced some interesting – and as-yet unpublished – results
by comparing the estimates of shadow economy against GDP per capita. Unsurprisingly, his analysis suggests that
the poorer countries have higher percentages of shadow economy than the rich
countries, and there appears to be a nice “J”-curve on the graph. But this
analysis seems to identify “excess” shadow economy in some countries, often
those with a reputation for “mafia-type” organised crime, including
- - - - - - -
I now briefly return to Ray Baker’s interesting work on cross-border flow analysis. It is based only on a review of studies of transnational crime, and the data may not be internally consistent, but this is another essential research technique in its own right. Each of these figures can potentially serve as a credibility check on any estimates we are able to generate.
|
Global Flows |
Low ($ |
High ($ |
|
Drugs |
$120 |
$200 |
|
Counterfeit goods |
$80 |
$120 |
|
Counterfeit currency |