Abstract
After two decades of research on money laundering, it seems time to look at what we know and what progress has been made in research. One simple question is whether we know if money laundering has increased, stayed constant, or decreased over these years. This article shows that over the last two decades, money laundering could hardly have decreased. This is largely because the concept of money laundering has broadened. However, there are also some hints that traditional areas of laundering such as fraud and corruption proceeds have increased rather than being effectively combated. There may be potential for money laundering to decrease in the future.
After two decades of research on money laundering, it seems time to look at what we know and what progress has been made in research. One simple question to test this is whether we know if money laundering increased, stayed, constant, or decreased over the last years. With all ambiguity that research on this topic still faces, I argue in this article that over the last two decades money laundering could hardly have decreased. This is largely because that the concept of money laundering has broadened. However, there are also some hints that traditional areas of laundering such as fraud and corruption proceeds have increased rather than being effectively combated. There is, hence, still a potential for money laundering to decrease in the future.
In the following, I first define money laundering and give an overview over diverse ways of measuring money laundering with proxies and with econometric and economic models. I then discuss whether we can give an unambiguous answer regarding the money-laundering trend. I conclude by showing that money laundering progressively merges with other fields of illegal or criminal economic behavior, in particular with tax evasion and the shadow economy. New pathways to research therefore should include tax crime, tax evasion, and tax compliance models as well as some broader aspects of the shadow economy.
What Is Money Laundering?
“Money laundering” is the process of disguising the unlawful source of criminally derived proceeds to make them appear legal. The term money laundering—bringing illicit proceeds from drugs, fraud, and other crime back into the legal economy—owes its name to practices used by Al Capone. He used, literally, launderettes to disguise illegal alcohol revenues during prohibition in the United States. Launderettes, a flourishing cash-intensive business in the 1930s when almost no household had a washing machine, were an ideal location to slip the money from illegal alcohol sales into the cash register. Disguising the origin of such money is even older than Al Capone’s “washing of money.” One of the oldest techniques to circumvent government scrutiny was to use international trade to move money, undetected, from one country to another, by means of fake invoicing or falsely declared merchandise (Zdanowicz 2009).
In earlier time, money laundering was not criminalized. A precondition for criminalizing money laundering is to define the “predicate offenses” understood as the criminal offenses which generated the proceeds, thus making laundering necessary. Hiding or disguising the source of certain proceeds will of course not lead to money laundering unless these proceeds were obtained from criminal activity (i.e., predicate crime). Therefore, what exactly amounts to money laundering, and who can be prosecuted, is largely dependent on what constitutes a predicate crime for the purpose of money laundering.
The history of legislation of money laundering began in the mid-1980s with the criminalization of the proceeds of drug-related offenses in the United States, the United Kingdom, and Australia. Its international reach was provided by the 1988 UN Convention against the illicit traffic in narcotic drugs and psychotropic substances (i.e., Vienna Drug Convention). Initially, therefore, the only offenders that could be prosecuted for money laundering were those attempting to launder the proceeds derived from the production and sale of narcotics.
In 1989, an intergovernmental body called the Financial Action Task Force on Money Laundering (FATF) was established by the G-7 in order to combat money laundering. It established forty recommendations with regard to money laundering. In line with the Financial Action Task Force Recommendations (in particular with recommendation 1) and the provisions of the 1990 Strasbourg Convention on laundering, search, seizure, and confiscation of the proceeds from crime, many states have introduced the criminalization of laundering of proceeds derived from other sources, not only drugs. This has had a profound impact, with the result that serious and highly profitable crimes such as fraud, arms trafficking, and corruption also came under the umbrella of money laundering legislation (Busuoic 2007).
Another broadening of the money laundering definition took place when terrorist financing was included. Terrorist financing came into limelight after the events of terrorism on September 11, 2001. In October 2001, the then US President George W. Bush passed the USA PATRIOT Act, which is a ten-letter acronym for Uniting (and) Strengthening America (by) Providing Appropriate Tools Required (to) Intercept (and) Obstruct Terrorism. Worldwide terrorism became an issue of money laundering as well. The FATF made nine special recommendations for combating terrorist financing (first eight then a year later added a ninth). With this, terrorist financing became part of the money laundering definition, or as Unger (2011) expressed it, the regulation of money laundering expanded from Al Capone to Al Qaeda.
Another move toward broadening the definition was made in 2010, when the FATF established a preliminary draft aiming to qualify tax crime as a predicate offense for money laundering. With this, the FATF wanted to do away with the distinction between tax fraud and tax evasion (FATF 2010). This distinction varies among countries at the moment. Tax evasion can be part of fiscal or financial fraud, which constitutes a predicate crime for money laundering in most countries, but milder forms of evasion can also be excluded from the money laundering definitions or from enforcement depending on national definitions. In the United States, foreign or domestic tax evasion (other than failure to pay US taxes on the proceeds of a crime) does not qualify as a predicate crime (Reuter and Levi 2006). Note that mixing criminal with noncriminal proceeds is a predicate crime in the United States, implying that evading taxes from partly criminal proceeds is part of money laundering (Unger 2007c). In effect, parts of tax evasion are now a predicate crime for money laundering in the United States.
In Europe, the Directive 2005/60/EC of the European Parliament and of the Council, which was formally adopted on September 20, 2005, on the prevention of the use of the financial system for the purpose of money laundering and terrorist financing (the “Third European Union [EU] Money Laundering Directive”), has tried to harmonize several legal definition differences for its member states. However, it did not tackle the problem of tax evasion. One reason is that tax is still perceived strongly as something under national sovereignty (Rawlings 2007).
Unger (2007b) gives an overview over differences in money laundering definitions of different countries and international organizations. For example, the Austrians and the Swiss do not consider tax evasion a crime; the Germans do not consider tax evasion from individuals as a predicate crime, only evasion from business and criminal organizations; the Dutch do not consider tax evasion as a serious crime but only as a misdemeanor, so only when connected to serious fraud can there be an offense related to money laundering.
In February 2012, the FATF revised its standards. One major revision was to formalize tax-related offenses as part of money laundering. Through this, tax evasion became a predicate offense for money laundering (FATF 2012). The reason for this was a Swiss court’s decision to block the turnover by UBS AG of data on 4,450 American tax cheats to the United States. The bank and Swiss authorities agreed to disclose the account information in August, but the Swiss court rejected the deal in January, citing a distinction between tax evasion and tax fraud.
Tax evasion in the United States is punishable by up to five years' imprisonment and a $100,000 fine. A money laundering conviction can mean a prison sentence of up to twenty years and a $500,000 fine, or double the sum of the illicit transactions under question (Association of Certified Fraud Specialists [ACFS] 2010).
When measuring money laundering, the broadening of the money laundering definition clearly poses a problem. The predicate offenses from 1995, when some of the first measurements of laundering were made, are not the same as today (Walker and Unger 2009); hence, measurements are by definition not comparable. In addition, since money laundering is related to predicate offenses, which differ among countries, the definition is not the same in every country. To give an example, hiring illegal workers is a predicate offense for money laundering in the United States, while it is not in Germany and in the Netherlands. Soft drugs like hashish and marijuana are legal in the Netherlands, and hence no predicate offense for laundering there; they remain offenses in many other countries. Comparing money laundering across countries is therefore difficult as the basis for measurement differs largely.
That the US definition includes hiring illegal workers is a typical example of the fact that the money laundering definition also overlaps with definitions of the shadow economy. The shadow economy, however, is defined as a closed border crime (Schneider 2005), while money laundering, especially from organized crime, can be a cross-border crime (Unger 2007a).
A merging of definitions of tax evasion and money laundering (and perhaps parts of the shadow economy) might help overcome problems of measuring different sorts of illicit financial activities that have always been difficult to distinguish neatly. Illicit financial activity comprises
money laundering (= efforts to hide illegally earned funds). tax evasion (= efforts to hide legally earned funds from tax authorities). capital flight (= efforts to evade national capital controls).
Since the definition of money laundering has broadened from Al Capone to Al Qaeda (Unger 2011), we should find an increase of money laundering. Indeed, when only tax evasion is included, laundering more than doubles. Coming back to the infamous gangster Al Capone, he was the most successful bootlegger and liquor smuggler of his time. In 1929, the Bureau of Prohibition began to shut down some of Capone’s breweries, and two years later, he was indicted for income tax evasion and not for money laundering since it did not then exist. Ultimately, Capone was sentenced to eleven years in jail. Including evasion in the money-laundering definition would therefore finally bring full circle the origin of the term money laundering.
Measuring Money Laundering with Proxies
Estimates on global money laundering are still in their infancy, for several reasons: because the underlying crime is unknown and because the proceeds of this unknown crime can only be estimated. Global estimates for money laundering range between $45 and $280 billion by Reuter and Greenfield (2001), $1.5 trillion in the International Monetary Fund (IMF), and Human Development Report 1999 (Transnational Institute [TNI] 2003), and $2.85 billion (Walker 1995, 2002; Walker and Unger 2009).
One approach to measuring money laundering is via proxies. In 1998, then Director of the IMF Michael Camdessus stated at a FATF meeting in Paris that “two to five percent of global [gross domestic product] would probably be a consensus range” (Camdessus 1998). This amounted at that time to about $1.5 trillion. Yet, these numbers are still quoted thirteen years later. If laundering still is $1.5 trillion, then it would have decreased as a percentage of gross domestic product (GDP). Even so, the 2 to 5 percent estimate is also still quoted, which would indicate that money laundering has stayed constant over time as a percentage of GDP and so has increased in absolute dollars. No scientific source of this “wet finger approach” can be traced (Walker and Unger 2009).
Another way of looking at money laundering trends is to take the worldwide proceeds of crime. Usually, it is assumed that a certain percentage of criminal proceeds has to be laundered. For example, it is often estimated that 70 to 80 percent of drugs proceeds need laundering, and the rest is reused in criminal business. The advantage of using proceeds of drug crime data is the fact that they rely upon quite well-developed measurements of drug production. With software similar to Google Earth, the United Nations zooms into the poppy fields of Afghanistan and calculates the average return of poppy. The United Nations also keeps track of the drug sales prices in different locations by investigating caught dealers and drug addicts. The UN Office on Drugs and Crime publishes its World Drug Report annually. According to the World Drug Report 2011 (United Nations Office on Drugs and Crime [UNODC] 2011), the potential opium production in 2009 was 6,900 metric ton (mt) and due to bad harvest this fell to 3,600 mt in 2010. About one-tenth of produced opium is heroin suitable for consumption. In 2009, these numbers suggest that heroin consumption was 375 mt. Potential cocaine production in Columbia amounted to 410 mt in 2010. Figures 1 and 2 show the development of the amounts and proceeds of opium/heroine and of cocaine over time.
Production of heroin and cocaine.
Proceeds of heroin and cocaine.
Overall, we can conclude that there has been a decrease in hard drugs such as heroin, a constant development of cocaine, an increase in soft drugs (e.g., amphetamines, ecstasy), and an increase in fraud (company fraud, credit card fraud, tax fraud). In total, money laundering from these predicate offenses should have increased.
Another way of measuring money laundering by proxies is to use balance of payments discrepancies. Errors and omissions in the balance of payments are used in the “hot money” method. The assumption is that errors and omissions arise primarily because of failure to measure certain movements of private short-term capital and that it is appropriate to add them to the recorded flows of short-term capital in order to get an estimate of total flows of “hot money” (Schneider and Windischbauer 2008). The residual approach measures capital flight by looking at the difference between inflows (sources) of funds and the outflows (uses) of funds, which is unrecorded and is considered the amount of capital flight. The question here is how well the residual reflects capital flight and does not include other discrepancies such as time lags and different calculation conventions. Capital flight also consists of both laundered money and tax evasion. The currency demand approach measures the discrepancy between the regular and excess demand for currency. Tanzi (1999) used this approach for both the shadow economy and money laundering.
The World Bank residual model belongs to this type of approaches. According to this model,
The World Bank model essentially uses errors and omissions (e.g., residuals) in the balance of payments as a proxy for illicit finance. However, it claims that it is capable of explaining these errors rather than using just statistical discrepancies. Evidently, the World Bank model measures all sort of illicit activities not only money laundering. Regardless, from this proxy, we can conclude that illicit activities fluctuate. See figure 3.
Net errors and omissions, adjusted (BoP, current US$ × 1 billion).
Measuring Money Laundering with Economic Models
There are few models that measure the trend of money laundering. Most models such as Walker and Unger (2009), Zdanowicz (2009), and Baker (2005) just measure laundering in points in time. The dynamic model by Schneider (2005, 2007) has the advantage of measuring trends. The dynamic multi indicators multiple causes econometric approach utilizes variables that are causes for money laundering (e.g., crime and bank secrecy) and indicators that go parallel with money laundering (e.g., increases in money demand or the number of convicted launderers). The assumption is that these two sets of variables are statistically independent of each other and both are related to the underlying unobservable variable money laundering. Schneider’s findings indicate an increase in laundering (figure 4).
Dynamic two-sector model.
A different way of measuring trends of laundering is using a dynamic two-sector model as done by Bagella, Busato, and Argentiero (2009). This approach comes originally from the shadow economy and makes several assumptions. First, agents have the option to work partly in the legal economy (wage) and partly in the illegal economy (more liquidity). They face transaction costs in the legal sector and costs of being detected in the illegal sector. Second, two types of firms produce with two different technologies: a legal good and an illegal good. Third, the government sets fines, can influence the probability of detection, and can influence the liquidity (money supply) of the economy. Fourth, the “optimal” money laundered depends on the labor services allocated to the legal and illegal sector and on the prices and quantities of both goods. Finally, the model forecasts the development of the legal and illegal sector. It first compares the development of the actual GDP with the model forecast. If this is accurate, it concludes that the illegal sector forecast also accurately predicts the illegal sector. As can be seen from figure 5, money laundering in the United States is increasing, while it is declining in Europe.
Money laundering in the United States and the European Union (EU) 15.
From all the empirical approaches on money laundering seen so far, we can conclude that money laundering is either fluctuating or increasing, and sometimes even decreasing. This is not a very convincing result after more than one decade of research.
New Pathways to Research
It is very likely that criminals will find new ways to avoid the stricter anti–money laundering regulations. This means that both within the banking sector shifts from the more controlled to the less controlled parts of banking will appear (like, e.g., the over-the-counter derivative trade or new forms of hedge funds) and shifts to other sectors like new electronic payment forms such as eGold and eCash or mobile phone cash payments also have to be expected. It is also likely that ancient forms of abusing trade for laundering will reappear (see Unger and den Hertog 2012). There have been some new ways of measurement which seem promising. One is to use unusual characteristics of variables to identify money-laundering activities. The approach of Zdanowicz (2009) for measuring trade-based money laundering is one of them.
Zdanowicz (2009) analyzes monthly export and import data from the US Merchandise Trade Database on a very detailed ten-digit product classification. This database is produced by the US Department of Commerce, Bureau of Census, and is used to determine the balance of trade. He identifies suspicious merchandise, the share of trade suspect to money laundering for each country, and the amount of money laundering between the United States and countries on the Al Qaeda watchlist. He provides both country risk and merchandise risk indices, helping to identify the countries and products most threatened by money laundering.
Consider a product, say, ketchup, that has an import price that lies below the margins of this country’s usual ketchup prices. All transactions with a price below the 5 percent margin or above the 95 percent margin are classified as trade-based money laundering. Zdanowicz (2009) uses not only country prices but also world prices and variance measures to determine unusual transactions.
A still unresolved weakness of the model is that, no matter how great the price fluctuations are, 10 percent of all transactions are always classified suspicious (e.g., the upper and lower 5 percent). For example, if the ketchup price fluctuation was drastically reduced (e.g., because of less trade-based money laundering in ketchup), so that the distribution of prices became narrower, then 10 percent of the transactions would of necessity still be counted as suspicious. Transactions that under the old distribution were classified unsuspicious would suddenly become suspicious, though the true reason might be a reduction in trade-based money laundering and not an increase.
Unger and Ferwerda (2011) apply the Zdanowicz (2009) model to the real estate sector, a sector which has been experiencing money-laundering problems lately. From the criminological literature on maleficent behavior in real estate, they identify diverse characteristics of a house or business object: that there is a foreign owner, that the owner has unusual number of transactions in real estate, that the object is used for risky exploitation like pubs or night bars, that the owner is a just newly established company, and that the object has an unusual price fluctuation. In total, seventeen characteristics of a house or business object were operationalized and measured for 12,800 objects in two Dutch cities, Utrecht and Maastricht. The Dutch Ministry of Finance, the Tax and Fiscal Fraud Department the Ministry of Justice, and Ministry of Inner Affaires, the Land Register, and other agencies provided access to data. Economists analyzed the objects and finally came up with a list of 150 unusual objects of which 36 were also considered suspicious by the criminologists. Figure 6 shows (with some deviations for safeguarding privacy) a map of the city of Utrecht with the suspected objects.
Map of Utrecht, the Netherlands, with suspected objects.
For estimating trends, seven consecutive years of trade-based money laundering data have been analyzed (Zdanowicz 2009). These estimates show an increasing trend, thus confirming that money laundering is not going down.
Apart from measuring money laundering, more economic modeling should take place in this area. If money laundering is closely related to the tax evasion and to the shadow economy, then tax evasion models can also be applied to money laundering.
For example, figure 7 shows Schneider’s (2005, 2007) estimates of the shadow economy on the x-axis and per-capita GDP on the y-axis. When one groups the countries, one can see the low GDP per capita and high shadow economy countries like Russia, Colombia, and Bolivia. These are countries where drug-based and hard crime money laundering are widely thought to take place. The second group is the high per capita income and low shadow economy countries, like Austria, Switzerland, Ireland, and the United States. This country group can be classified as the tax evading countries, and their proceeds of crime come from tax fraud and tax evasion rather than from producing drugs. The last category is countries that have a lower shadow economy than expected according to the curved fitting line. These are poor and suppressed economies like China, which due to their regime have less opportunities to have the shadow economy flourish (Walker and Unger 2009).
Gross domestic product (GDP) per capita versus shadow economy.
Still ongoing research on money laundering tries to test existing money laundering models. So far, money-laundering trends have been estimated from models that—due to the fact that money laundering cannot be observed—could not be tested with regard to their appropriateness of fit. The actual debate centers on the accuracy of money laundering estimates. One important step forward is to use time series of money laundering estimates generated from one statistically independent source and to look at how well existing models can forecast this time series.
For example, Ferwerda et al. (2013) used this approach to test the Walker and Unger (2009) gravity model. They used the Zdanowicz (2009) data on trade-based money laundering and examined the predictive power of the Walker and Unger gravity model for money laundering. They find some misspecifications in the traditional Walker and Unger model, and opted for a traditional multiplicative pure gravity model that predicts trade-based money laundering better than does the Walker and Unger model. This approach also allows for testing the weights and significance of the variables used in the model. One important and surprising finding of the authors is that in countries with stricter anti–money laundering regimes trade-based money laundering increases. This indicates that there might be a substitution from banks to trade-based money laundering when anti–money laundering policy becomes stricter. This would mean that “water always finds it ways” (Unger and den Hertog 2012), so that criminals discover new ways of laundering once traditional laundering channels are stricter supervised and controlled. In addition, they find that trade and distance are the most important indicators for money laundering. This approach would indicate that money laundering develops in the same direction as trade, and hence increases over time.
Another interesting study is on testing whether money laundering has spillovers to the legal sector, whether it leads to more crime, to an infiltration of the legal sector by criminals. Following this approach, first developed by Masciandaro, Takats, and Unger (2007), criminals invest their illicit proceeds in the legal sector, and they partly reinvest the interest on these proceeds in the illegal sector. More crime necessitates more money laundering, which leads to more criminal investment and crime. As Unger (2007b) showed, the resulting crime multiplier is not stable; indeed, under quite plausible assumptions it very quickly becomes unbelievably large. Money laundering would steadily increase over time.
Newer approaches from network analysis try to model criminal networks and links of honest employees with criminals. The type and number of links differ, and multiple equilibria emerge. With this, the findings of Masciandaro, Takats, and Unger (2007) can be shown to be only a subset of possible reactions of criminals to money-laundering possibilities (Imanpour, n.d.).
Some countries and sectors are more vulnerable to money laundering activities than others. At the moment the IMF and the EU are working on developing a threat analysis for countries and sectors. Variables like the types of crime prevalent in the country that influence the amount of proceeds of crime, the prevalence of sectors with high money-laundering risk such as the transport sector or large financial markets, and anti–money laundering policy efforts count for more or less vulnerability to laundering. It can be shown that two types of countries are most vulnerable: cash-intense poor economies, which allow illicit proceeds to be hidden, and rich economies with a large financial sector, which allow the transfer of large sums of illicit proceeds (Ferwerda et al. 2011).
A still ongoing study done by the EU on the Economic and Legal Effectiveness of Anti–Money Laundering Policy of the twenty-seven member countries analyzes the anti–money laundering process along five steps (ECOLEF 2009): first, to implement EU Directives on anti–money laundering and combating terrorist financing and to define money laundering in the national law; second, to establish anti–money laundering institutions like Financial Intelligence Unites in each country; third, to measure the information flow between the Financial Intelligence Unit and the Public Prosecutor; fourth, to examine the role of the judges that convict money launderers and the number of successfully convicted launderers; and fifth, to measure the costs and benefits of the anti–money laundering and combating terrorist financing system of each country. Large differences in the legal and institutional system among the member countries as well as different economic goals and interests show how difficult it is to harmonize anti–money laundering policy even within Europe.
One interesting research outcome is that countries, under the threat of being blacklisted by the Financial Action Task Force, might increase their efforts to provide money-laundering statistics. Two developments might occur. First, countries threatened with blacklisting may produce “successful” numbers of an effective anti–money laundering policy on paper; that is, official statistics could be rearranged. To give a simple example, a country that listed suspicious transactions in bundles, so that one report contained diverse (say, ten) transactions, might be inclined to count each transaction separately in order to produce ten times more suspicious transaction reports in order to satisfy the requirements of FATF bureaucrats for more reports. Second, however, if the anti-money laundering process gets better documentation, there is the chance that money laundering will increase in the books due to better statistics. Note that better documentation might also shy away launderers who do not like to be traced and who will try to avoid documentation into either other sectors, like trade or into other countries.
At the moment, there are two predicate crimes for money laundering, both of which attract much attention: tax evasion and corruption. In December 2012, an EU project on corruption in public tendering will try to estimate EU money lost through corruption in public tenders in eight EU countries and five sectors. New numbers on corruption imply also new numbers for money laundering from corruption proceeds and will tend to increase the amount of money laundering.
Conclusion
Though research so far does not allow a definite conclusion about money laundering trends, the indicators discussed here suggest that money laundering is increasing rather than declining. The money laundering definition has broadened so that even with a very efficient anti-money laundering strategy, one has to expect that numbers increase. In particular, some of the new categories (e.g., tax crime) are large in volume and add to the amounts. In addition, some of the earlier established predicate crimes for laundering such as fraud and corruption have increased lately, which points at an increase in laundering. Money laundering can be harmful to the economy (e.g., launderers crowding out honest business), to society (e.g., crime and corruption increase), and to politics (e.g., criminals can undermine the government and terrorism can increase). It is a ticking time bomb, and the more laundering the more white-collar people like bank employees and notaries public, real estate agents, and accountants will get drawn into the underworld. It seems therefore important to be able to predict the amounts and trends of laundering and to make sure that it decreases. Research with regard to theoretical modeling of money laundering networks and with regard to empirically testing existing money laundering models is only in its infancy and is done by a small group of researchers. To consider money laundering, tax crime, tax evasion, and the shadow economy as related problems will hopefully attract more economists to research this fascinating and critical problem.
Footnotes
Author's Note
With thanks to the organizers and participants of the conference Shadow 2011 in Muenster, and to the members of the Chair of Public Sector Economics, Utrecht University School of Economics, Andreas Buehn, Ioana Deleanu, Joras Ferwerda, Loek Groot, Daan van der Linde and Victor van Kommer. This paper partly draws on Unger and den Hertog (2012), and on 
.
Declaration of Conflicting Interests
The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.
Funding
The author(s) received no financial support for the research, authorship, and/or publication of this article.