As people do not check, counterfeiters settle for 50 %

Transcription

1 As people do not check, counterfeiters settle for 50 % Hans de Heij De Nederlandsche Bank NV, Amsterdam, The Netherlands ABSTRACT The Dutch are increasingly confident about the authenticity of their euro banknotes. People trust the banknotes coming from an ATM and the change received from a retailer. Furthermore, they do not encounter counterfeits. As a result, the interest in an authenticity self-check is on the decline. A retrogression is also noticed for the public s knowledge of authenticity features, which came out on 2.0 in New introduced features are not recorded; the watermark remains the best known features, although respondents don t know how it looks like. The authenticity sensitivity of the public features of the euro has been studied, for both series of euro banknotes, ES1 and ES2. Another parameter being measured is the confidence score. When both parameters are plotted in a graph, foil features deliver the best results, while colour changing features stay behind. The quality of counterfeited euro banknotes tends to decline over the period On average, counterfeiters mimic about half of its public features. Rather than the introduction of new public authenticity features, these developments point in a direction of confidence features, supporting the public s confidence in euro banknotes. Furthermore, public authenticity features should be intuitive and self-explaining. Keywords: Banknote design, authenticity sensitivity, confidence score, counterfeits. 1. INTRODUCTION In 1979, the Swiss were the first to introduce a leaflet promoting public features. The policy of public features as a first line of defense needs to be reviewed, as public interest in authenticity features of banknotes is on the decline. Two recent studies conducted in the Netherlands point in this direction: 1) Qualitative study on perception of public features in euro banknotes [18], 2) Biannual study to the knowledge and appraisal of euro banknotes in the Netherlands [13]. A third study reveals that the counterfeiter settles for a medium reproduction quality, a quality also slightly declining: 3) Quantitative study to the quality of counterfeited euro banknotes over the years [3]. 2. PERCEPTION OF PUBLIC FEATURES IN EURO BANKNOTES In 2013, the first model of a new series of euro banknotes was introduced, the 5 euro. Other models of the Europe Series followed later (euro 10 in 2014, euro 20 in 2015 and euro 50 in 2017). The designs of the Europa Series or Euro Series 2 (ES2) are an upgrade of the first series (ES1). The images remained the same, the refreshed series aimed for new public features, the emerald number, a colour changing feature, in all models (rolling bar) and a transparent window in the higher denominations (euro 20 and 50). Other public features have been removed (iridescent band) or their designs have been changed (watermark, foil stripe, raised print) Author information: hans.de.heij@dnb.nl 1

2 FOUR TEST SITUATIONS FOR AN AUTHENTICATION SELF-CHECK Features offered In complete banknote (fantasy or genuine) Isolated Instructed A B Intuitive C D Figure 2.1 Four test situations for an authentication self-check. Isolated authenticity features a) b) Figure 2.2 Examples of 'banknote cards' showing isolated authenticity features. a) Emerald number in the 5 euro of the Europa Series (issued in 2013). b) Foil feature in the 20 euro of the Europa Series (issued in 2015). FEATURE - RESPONSE MATRIX GENUINE OR COUNTERFEIT Response Authenticity feature or complete banknote + Genuine - Counterfeit + Genuine Hit False alarm - Counterfeit Miss Correct rejection Figure 2.3 The feature-response matrix for genuine and counterfeited authenticity features or the complete banknote. Hit = genuine perceived as genuine. Miss = counterfeit perceived as genuine. False alarm = genuine perceived as counterfeit. Correct rejection = counterfeit perceived as counterfeit. 2

3 In order to understand how these changes are perceived by the Dutch public, a qualitative study was conducted by DNB in 2016, a study which included the first three models issued of the Europa Series, the 5, 10 and 20 euro. In 2016, the euro 50 was not yet in circulation. Public features of ES1 and ES2 are compared to each other, on two parameters, the authenticity sensitivity and the confidence score. The authenticity sensitivity is the ability to discriminate faked features from genuine ones. How respondents experience confidence in a banknote, or a banknote feature, is provided by a confidence score. The study is mainly focused on the perception of the man in the street and not that much on the underlying technique of the features. From previous studies it is known that the Dutch have a high confidence in the authenticity of euro banknotes. Previous studies also show that knowledge of public features is poor. On average people recall about two features. Furthermore, they have a poor qualitative knowledge of such features; the Dutch know that there is a watermark in euro banknotes (70 %), but they cannot tell what it represents (2 % correct answers), as will be elaborated on in chapter 3. Based on these findings, the assumption is that the Dutch don t check public authenticity features [4]. Authenticity features are meant to cause difficulties for the counterfeiter and are not intended for me, so seems to be the public s attitude, as will be explained in more detail in chapter 3. The study also aimed for the evaluation of a measurement method for completely new candidate features for the public, e.g. for a new series of ES3 banknotes. Therefore, two candidate features were included, features that the Dutch did not know, as they are not part of the euro banknotes. This chapter describes first the test method (section 2.1), followed by the results (section 2.2). 2.1 Method for Intuitively Testing Isolated Features (MITIF) Features in a banknote can be tested within a complete banknote or as an isolated feature. Furthermore, respondents may receive an instruction on how to verify the feature, or they may not receive any instruction and have to rely on their intuition. These four test situations are presented in figure 2.1, respectively A, B, C and D. An example of category D is published by the Bank of Canada [9]. Perception tests were carried out on isolated authenticity features within Canadian banknotes, which had to be judged intuitively by different respondents. This study was taken as an example by DNB. First, so called banknote cards were prepared, stimulus material to be offered to the respondents. Two examples of banknote cards are provided in figure 2.2, showing respectively the emerald number and the foil. A banknote card shows just one public feature, covering the rest of the banknote. An isolated feature has the advantage that people aren t distracted by any other design elements, either in genuine or in fantasy banknotes. Similar to the Canadian study, DNB opted for testing without an instruction, as normally the public has to rely on their own knowledge and intuition. The provided name to this test method is Method for Intuitively Testing Isolated Features, in short MITIF. New features of the Europa Series (ES2) are compared to the ones they replaced, the public security features in Ages and Styles of Europe (ES1). In this study, the ability of people to detect the counterfeits has been reported with the help of the feature-response matrix, as displayed in figure 2.3, identifying four possible situations. The matrix and the terms are borrowed from the Signal Detection Theory [10]. The hit-rate (h) and the miss rate (m) can be converted into a sensitivity index d' or d-prime. In case of banknotes, d-prime is the authenticity sensitivity, a trade-off between hits and misses [3]: d = Z(h) Z(m). Z-scores are for each respondent calculated, for both the hits and the misses. For example for the first respondent the Z(h1) is: h 1 - µ h Z(h 1) = σ h In a similar way Z(m 1) is calculated. 3

4 AUTHENTICITY SENSITIVITY VALUES Authenticity sensitivity (d-prime or d ) Remarks Negative Counterfeit is perceived as being more genuine then the genuine ones. 0 Chance performance. 1 Many counterfeits are classified as genuine Most counterfeits are classified as counterfeit. 2.5 About 90 % of the counterfeits are correctly detected; circa 10 % of the genuine notes are mistakenly judged as a counterfeit. 4 All originals are detected and no counterfeits are marked as genuine. Figure 2.4 Overview of d-prime values and the corresponding authenticity sensitivity. COUNTERFEIT DETECTION-PERFORMANCE REPORTED IN d Study reported by Currency Number of counterfeits Hits Miss d (%) Klein et al. (2004) CAD 33 % 76 % 13 % 1.83 * Van der Horst et al. (2016) EUR 10 % 79 % 8 % 2.21 Jones and Raymond (2017) GBP 4 % 70 % 10 % 1.38 Van der Horst et al. (2017) EUR Foil Europa 75 % 59 % 5 % 1.88 EUR Emerald number 75 % 32 % 25 % 0.20 Raymond (2017) Fantasy 17 % 73 % 16 % 1.58 *) Not reported in Klein et al (2004); calculated by Raymond. Figure 2.5 Overview of studies on the counterfeit detection-performance of several currencies in d (for the group). D-PRIME AND CONFIDENCE SCORES D-prime (d ) Confidence score (scale 1 10) Foil ES2 euro Colour ES1 euro Intaglio ES1 euro Foil ES1 euro Colour ES2 euro Watermark ES1 euro OVI ES1 euro Intaglio ES2 euro Watermark ES2 euro Candidate Feature 1 (RS) OVI ES2 euro 5 (emerald) Candidate Feature 2 (MVC) Figure 2.6 Overview of the d-prime and confidence scores of public features in euro banknotes [11]. 4

5 Authenticity sensitivity: d-prime (d ) Figure 2.4 presents an overview of d-prime values. When people are not able to make a distinction between a genuine banknote and a counterfeit other than by gambling, the features sensitivity represents chance performance (d = 0). When all originals are detected, and no counterfeits are marked as genuine, the feature would be perfect (d = 4). A negative value means that participants think that the counterfeit is more genuine then the genuine ones. Recently several studies report on d-prime values of public authenticity features, as reported in figure 2.5; the authenticity sensitivity is becoming an accepted method to classify security features. Figure 2.5 includes the study by Van der Horst et al. [18], the subject of this chapter, which will be reported in section 2.2. Confidence score The perceived resistance of a banknote against counterfeiting has been object of study by Masuda et al. [14], a study carried out on complete banknotes (category C in figure 2.1). These researchers concluded that, in general, people have limited knowledge of public features in banknotes, and therefore their design should assist the public during an authenticity check. Confidence scores of banknotes are reported by the Bank of Canada and by De Nederlandsche Bank and are reported in this paper in chapter 3. In the study by Van der Horst et al., people were interviewed about their confidence experience of the features offered. At the end they were invited to provide a confidence score to the feature on a scale from 1 to Results of perception of public features of euro banknotes In 2016, the study, planned and ordered by DNB, was carried out by Motivaction, a well-known Dutch agency providing services on gaining consumer insights. In total, 70 respondents participated. Although data has been obtained, the study is considered to be a qualitative study. The focus of the study was on the public security features of the ES2, respectively a watermark (1), a glossy foil with a window and holographic features (2), a printed relief (3) and a colour changing feature (4). People use the colour of the banknote in their authentication process, a finding of a previous study [19]. Therefore, a fifth feature was added, representing colour and resolution, the print quality (5). Original features were offered next to three different types of counterfeited features, imitations of respectively low, medium and high quality, following the Simple Method (section 4.2). When possible, these pieces were counterfeits taken out of circulation. If that proved to be impossible, they were produced within the Reproduction Research Centre in Copenhagen. Respondents were split in two groups. The first group (34) focussed on the authentication of security features and the second group (36) on the confidence that these features exude. Two candidate features (CF) were added, aiming for the exploration of intuitively understanding of security features that people haven t seen Candidate feature 1 (CF1) was a security thread with a micro mirror structure, known as Rolling Star, and candidate two (CF2) was a feature known as Multi Variable Colour or MVC (CF2). More details on the test settings are provided by [18]. Measurement results on authenticity sensitivity Although by setup a qualitative study, quantitative results for d-prime could be reported and are presented in figure 2.6. Significant differences are measured, ranging from d = 1.88 for the foil feature of the Europa Series, to d = 0.20 for the colour changing feature of the ES2, the emerald number. The study confirmed the limited knowledge of participants of public features. Respondents explained that they do not think it is necessary to know or to check a euro banknote, because they think that ATM s and the change from a retailer can be trusted. Furthermore, they have no experience with counterfeits. Participants hardly used the look-feel-tilt method. And in case of feel, many respondents didn t know that they were supposed to touch specific areas. The watermark is best known, although they don t know how it looks like. Measurement results on confidence in authenticity features Features that provided the impression of difficult to be counterfeited contribute to the confidence, so is one of the main findings. Public security features are not there for me, they are there for the counterfeiter, is what the Dutch seem to think, an assumption confirmed by the study done. When a participant had the perception that a feature is difficult to counterfeit, it added to the confidence experience. The foil feature of the ES1 received the highest confidence score (8.2), on a scale from Participants indicated that the genuine emerald number on the Europa Series can t be real, as it is too shiny and glimmering, delivering a confidence score of

6 Graph: (authenticity sensitivity) x (confidence score) The two perspectives of the study, authenticity sensitivity and the confidence score, can be related to each other, as shown in figure 2.7 (using the data provided by figure 2.6). The authenticity sensitivity is reported on the horizontal axis and the confidence score on the vertical axis. The graph is divided in four parts and each part received a name [18]. Features with a low authenticity sensitivity and a low confidence are Not working. A high confidence, combined with a low authenticity sensitivity provides False security. A high sensitivity that doesn t generate much confidence is named Trigger feature. The name of the most desirable quadrant is Intuitive and trusted, representing authenticity features generating confidence and helpful in determination of the authenticity. Figure 2.6 shows that the foil features of the ES1 and ES2 received the highest scores and are the only features located in the section of Intuitive and trusted; the colour changing feature of the euro 50 (ES1) is on the edge. The high sensitivity is mostly due to the fact that 20 respondents out of 34 identified the genuine features correctly (a hit) and that just 5 respondents out of 102 judged a counterfeited foil as genuine (a miss). The window element in the ES2 foil was hardly noticed. Candidate feature CF1 is obviously not working and CF2 is doing only a slightly better. Participants had no idea how to judge this feature, since they don t know it and weren t instructed. Although participants identified the intaglio of the first series as genuine, they also rated about 50 counterfeits as genuine. This is caused by the fact that respondents often don t know they are supposed to touch this feature. In general, the public features of ES1 received a better score in the diagram of figure 2.7, except for the foil. AUTHENTICITY SENSITIVITY AND CONFIDENCE SCORE OF PUBLIC FEATURES IN EURO BANKNOTES Confidence score Authenticity sensitivity (d ) ES1 = Euro Series 1 = Ages and Styles of Europe (issued in 2002) ES2 = Euro Series 2 = Europa Series (first model issued in 2013) CF = Candidate Feature Colour = area near the centre representing colour and resolution (print quality) OVI = Optically Variable Ink (OVI ES 2 = emerald number or Spark) WM = Watermark Figure 2.7 Authenticity sensitivity (d ) and confidence score for several public features of euro banknotes (using the data provided by figure 2.6). 6

7 3. KNOWLEDGE AND CONFIDENCE OF EURO BANKNOTES (2017) The previous chapter reported on the authenticity sensitivity of public features in euro banknotes (d ) and the confidence in public features as experienced by the public. This chapter continues with reporting on the following three topics: 1) People s knowledge of authenticity features in euro banknotes, 2) People s confidence in the genuineness of euro banknotes, 3) People s changing attitude towards an authenticity self-check. These three categories of measurements are reported as part of DNB s biannual poll on knowledge and appraisal of euro banknotes [13]. People s knowledge of authenticity features in euro banknotes Measurement results on the Dutch public s knowledge of authenticity features in euro banknotes are reported in figure 3.1. The 2017-research confirmed earlier findings on the knowledge of euro banknotes in the Netherlands. The Dutch have limited knowledge of public features; on average they are able to recall two features. The public knowledge of the watermark and holographic features is high, while the knowledge of colour changing inks and a transparent window is low. About 14 % cannot recall any feature. People s confidence in the genuineness of euro banknotes When it comes to the authenticity of euro banknotes, the Dutch are confident that their banknotes are genuine. Since 2005, this confidence has been growing steadily as illustrated by figure 3.2. The Dutch confidence score is a report mark, based on the question: If you should give a report mark for your confidence in the authenticity of euro banknotes that you receive in your hands, what would be your score on a scale from 1 to 10?. Figure 3.2 also displays the trust of the Canadians in their banknotes, which is also a constant. The questions asked are different in the Netherlands, the reason why the Canadians seem to have less confidence in their banknotes as the Dutch [4]. Trust of the Dutch in their euro banknotes shows a slight upward trend (figure 3.2), which is mainly caused by the increase of respondents providing a score of 8 or higher (figure 3.3). The increase in confidence is unrelated to the number of counterfeit euro banknotes, which is over the years on comparable levels. In Canada the public s confidence in banknotes is also rising. People s confidence in the genuineness of euro banknotes, Because of the high and stable confidence in the authenticity of euro banknotes, the Dutch do not feel the need to examine euro banknotes (figure 3.4). In 2013, 52 % of the respondents answered never to check a banknote, a number which increased to 65 % in 2017, a significant rise [13]. The fast majority (53 %) do not expect that their behaviour will change. A large part, 22 %, tell that they do an authenticity self-check out of curiosity or habit (22 %). Checking a banknote is for 13 % of the respondents part of their work. Furthermore, banknotes are easier accepted by lower social classes and Dutch respondents living in cities. 7

8 PUBLIC KNOWLEDGE OF AUTHENTICITY FEATURES OF EURO BANKNOTES IN NL Year Number of respondents 2,002 2,015 1,501 1,506 1,058 1,003 1,020 1,010 1,002 Watermark (in %) Hologram/silver foil (in %) Look-through window/hole in foil n.a n.a. n.a. n.a. n.a. n.a. n.a. n.a 2 Security thread (in %) Special ink: glossy stripe (in %) Special ink: colour changing ink (in %) Special ink: rolling bar (Euro series 2)* n.a. n.a. n.a. n.a. n.a. n.a. n.a. 1 2 See-through register (in %) Raised ink, relief (in %) Type of paper (in %) Micro-text (in %) Ultra Violet (UV), total (in %) dull paper (in %) fluorescent fibres (red, blue, green) (in %) ink brightens up (front: flag, signature) (in %) - ink brightens up (reverse: bridge, map) (in %) - hold under UV light, not specified (in %) Infra-Red (IR) (in %) Average number of correct auth. features PARTLY INCORRECT Signs for the blind (in %) Banknote numbers (in %) Cannot be copied (in %) Effect special pen (in %) INCORRECT Signature (in %) Average number of correct, partly correct and incorrect authenticity features DON T KNOW ANY Cannot recall any auth. feature (in %) n.a. = not applicable *) A colour changing ink, from blue to green, showing a movement effect. Promoted as the esmerald number. = Positive knowledge = Knowledge needing attention Figure 3.1 Public knowledge of the authenticity features of euro banknotes in the Netherlands over the years [13]. The answers are given repeated by heart. All measurements are carried out in order of DNB by TNS NIPO, except for 2015 and 2017 when the measurements were carried out by Motivaction. 8

9 Confidence in authenticity of banknotes CONFIDENCE IN BANKNOTES IN CANADA AND THE NETHERLANDS Euro NL Canada Figure 3.2 Confidence in the authenticity of Canadian banknotes, period [1, 2] and euro banknotes in the Netherlands, period [4, 13]. INCREASED CONFIDENCE IN THE NETHERLANDS % 19% 32% 32% 2% % 15% 36% 37% 5% % 14% 34% 43% 1% % 14% 35% 46% % 11% 33% 49% % 9% 28% 51% 3% % 7% 31% 53% 2% 5 or lower or higher Absolutely don't know Figure 3.3 Increased confidence in the authenticity of euro banknotes in the Netherlands over the years [13]. The question asked is: If you should give a report mark for your confidence in the authenticity of euro banknotes that you receive in your hands, what would be your score on a scale from 1 to 10?. AUTHENTICITY SELF-CHECK Answer categories No, and I did not consider it 52 % 54 % 65 % 2. No, but I did consider it 13 % 7 % 6 % 3. Yes, once 5% 5 % 6 % 4. Yes, 2 to 10 times 13 % 13 % 13 % 5. Yes, more than 10 times 18 % 20 % 9 % Figure 3.4 Answers to the question on an authenticity self-check, Have you ever checked a euro banknote for authenticity in the past five years? [13, 15, 21]. 9

10 4. QUALITY OF COUNTERFEITED EURO BANKNOTES ( ) Chapter 2 and 3 elaborated on public authenticity features in euro banknotes as perceived by the users, reporting respectively knowledge (1), confidence (2) and attitude (3). This chapter presents measurements on the work of the counterfeiter. First, the counterfeit situation in the Netherlands is introduced (section 4.1), followed by an introduction of the measuring method of the quality of counterfeited euro banknotes (section 4.2). The last part presents the measurements of the quality of counterfeited euro banknotes as registered in the Netherlands (section 4.3). 4.1 Counterfeit situation in the Netherlands The total amount of registered counterfeits within the Euro Area is provided by the ECB. For 2009, these figures are reported in figure 4.1a. They are the result of the aggregated figures delivered by each of the 19 central banks of the Eurosystem. The Dutch contribution of counterfeits is displayed in figure 4.1b, figures reported by the National Analyse Centre (NAC), part of DNB. In both graphs, the euro 50 is the most counterfeited euro banknote. The 2009-graph is representative for the following years ; roughly 70 % of all counterfeits registered in the Netherlands are euro 50, the others are mainly euro 20, presented in figure 4.2. This table also reports on the total amount of counterfeited banknotes worldwide, telling that the development of euro counterfeits follows the worldwide trend; when worldwide the number of counterfeits registered goes up, there is also an increase in the Netherlands. As euro banknotes circulate across the borders, the amount of euro banknotes circulating in the Netherlands is unknown and can only be calculated by estimating. In other countries, counterfeiters may focus on other denominations. 4.2 Simple Method Since 2008, the National Analysis Centre of DNB reports the registered counterfeits according to the Simple Method [5]. This Method provides insights in the quality of counterfeited banknotes accepted by the public, revealing the targeted features and their reproduced quality. Instead of monitoring all different counterfeits, the method considers the ten most frequent accepted mimicked notes by retailers and the public. The name of the method is derived from the way of scoring, which is simply done by attributing 0, 1 or 2 points for respectively no imitation, a poor imitation and a good imitation. These scores are attributed by two experts of the NAC. Figure 4.3 shows some examples of counterfeit scoring. Euro banknotes have six public features, bringing the maximum counterfeit quality score on 12 (= 6 x 2). The individual quality of each public feature is one more indicator delivered by the Simple Method, as provided in figure 4.4. This table suggests that counterfeiters focus on the features that people recall, as this ranking is remarkable similar to the public s knowledge of the authenticity features as shown in figure 3.1. The watermark is the best known feature, recalled by about 70 % of the respondents. The silver coloured foil takes the second position, recalled by about 40 %. Both features are well imitated by the counterfeiters, taking respectively the positions 3 and 1 in figure 3.1. Arguing along these line, the strongest public authenticity features seems to be the gold shining band and the colour changing feature. However, these features are on the reverse and therefore the counterfeiter does not seem to pay much attention to these features. Whether counterfeiters follow the ranking of figure 4.4 remains unknown. A counter argument is the seethrough register, which is recalled by less than 10 % of the respondents, but is ranked second in figure

11 Distribution of counterfeits mln Circulation Eurozone (2009) Counterfeits in NL (2009) Figure 4.1 Two graphs representing respectively the circulation of euro denominations in the Euro Area and the counterfeited euro denominations in the Netherlands. a) The number of euro denominations in millions issued per denomination within the Euro Area [8]. b) The euro 50 is the most frequent detected counterfeited banknote in the Netherlands [7] a) b) COUNTERFEITED EURO BANKNOTES Netherlands Worldwide Year Numbers Share euro 20 Share euro 50 Numbers , % 70 % 751, , % 63 % 606, , % 60 % 531, , % 69 % 670, , % 67 % 838, , % 59 % 899, , % 73 % 684,000 Figure 4.2. Statistics of the number of counterfeited euro banknotes in the Netherlands and worldwide over the years [6]. 11

12 No watermark Poor watermark Good watermark only dark tints dark and light tints 0 points 1 point 2 points No imitation = 0 point Poor imitation = 1 point Good imitation = 2 points Poor foil stripe Poor foil stripe Poor foil stripe 1 point 1 point 1 point Figure 4.3 Six examples of counterfeited public security features in euro banknote forgeries and the assignment of points according to the simple method. QUALITY OF COUNTERFEITED PUBLIC AUTHENTICITY FEATURES Public features Score (2 points max) Foil See-through register Watermark Security thread Tactile effect gravure Shiny or colour changing* Total score (maximum 12) Average score (maximum 2) Retail features 1. UV IR *) Gold shining band (euro 5, 10, 20) or colour changing numeral (euro 50, 100, 200 and 500). Figure 4.4 Quality of counterfeited public authenticity features in euro banknotes, based on the 10 most frequently received counterfeit types in the Netherlands, respectively in 2008 and

13 Euro banknote counterfeits in the Netherlands a) Monthly rolling average number of registered counterfeits Quantity Benchmark 4200 (50 c/mnic) 12 Average quality b) Imitation quality scale according to Simple Method c) Imitation quality scale according to Simple Method Weighted quality 500 P P3A 50 P P3A C62 20 K L5 50 K40 50 C97a 100 L Figure 4.5 Quantitative and qualitative overview of counterfeited euro banknotes in the Netherlands over the years a) The rolling average of registered counterfeited euro banknotes each month by the National Analysis Centre (December values). b) The average quality of the reproduction of public authenticity features in fake euro banknotes according to the Simple Method. c) The average weighted quality of the reproduction of public authenticity features in fake euro banknotes. In red the counterfeit indicatives which received the highest score, in purple the lowest scores. 13

14 4.3 Euro banknote counterfeits in the Netherlands The Dutch benchmark for counterfeits is 50 ppm (parts per million) [4, 5], which is about 4,200 counterfeits per month. Over the years , the number of counterfeits registered in the Netherlands oscillates around the benchmark, as illustrated by figure 4.5a. Application of the Simple Method over this period shows that counterfeiters score on average 5.9 points on a scale from 0 to 12 points, close to 50 % (figure 4.5b). Furthermore, over these years the standard deviation is 0.3 points, indicating that this average is stable. However, counterfeiters do not treat all seven euro denominations similar. Low denominations often score 0 points, none of the public features has been imitated. Within the average quality, there are several counterfeits which received just 1 point, while others received 11 and even 12 points. Taking the number of counterfeits of a specific variant into account, the average weighted quality can be calculated, as reported in figure 4.5c, which turns out to be 7.1 points, with a standard deviation of 0.3 points. Over the period reported, the weighted average declined from 7.6 in 2008 to 6.8 in 2015, suggesting that the quality of counterfeits is decreasing. The Simple Method also indicates when a banknote should be replaced, i.e. when all public features are imitated well. Concluding, based on detected counterfeits, the simple method delivers feedback on the quality of imitated authenticity features. Counterfeiters do not mimic all public features, they imitate about half. The watermark is the most imitated feature and is in the Netherlands also best known (chapter 3), while iridescence and OVI are the least imitated features and the least known. The quality of public features in counterfeits is on the decline. 5. CONCLUDING Three different studies on public features in euro banknotes all point in the same direction. Public features in banknotes are unknown and remain unused. The Dutch do not see the need to verify euro banknotes and they don t intend to do so in the future. They have a high and stable trust in euro banknotes. And they are right, as the probability to receive a counterfeit is practically zero. Public features should add to confidence first, and second, they should be usable for an authentication self -check. Authenticity features in banknotes should be self -explaining, helpful in authenticating and adding to the confidence. The way to go are realistic images, easy names, fitting the theme and an unambiguous task and judgement. With the help of the Method for Intuitive Testing of Isolated Features (MITIF), the authenticity sensitivity of the public features in euro banknotes has been measured. Furthermore, the confidence experience has been measured. The proposed Method discriminates between the different public features of the euro banknotes. The holographic foil In the Europa series received the highest score on both criteria. The emerald number feature doesn t comply very well with these criteria. To the Dutch public the public features of the Europa Series are no improvement, when compared to the performance to the features of the first series of Ages and Styles of Europe. The second study reports that, in the Netherlands, the general public is aware of two authenticity features. Over the years this average declined from 2.3 (2002) to 2.0 (2017). Watermark and foil are the two best known features. People are unaware of the details of these features. A significant group is not aware of any of these features. The third study, using the Simple Method, reports that counterfeiters settle for about half the maximum achievable quality of all public authenticity features. The study reports over the years Two interesting conclusions were drawn immediately. The two features known by best by the public, watermark and foil, are in the top three of the best imitated public features. The third is the see-through register. Although counterfeiters are able to produce look-alike reproductions of all public features of the first series of euro banknotes, the trend seems to be that they settle for a lower quality; the weighted average quality declined from 7.6 (2002) to 6.8 (2017). The average quality of reproduced euro banknotes is 5.9 and classified as poor. On average, counterfeiters mimic about 50 % of the features. 14

15 ACKNOWLEDGEMENT I want to thank my colleague Mr. Niels Ducro, working at the National Analysis Centre of De Nederlandsche Bank, for delivering the statistics of counterfeited euro banknotes according to the Simple Method. Credits go also to Mr. Jan Binnekamp (De Nederlandsche Bank, Head of Cash Policy Department) for making the studies possible. REFERENCES 1. Bank of Canada (2017). The Bank Note Confidence Index results. Ottawa. 2. Bank of Canada (2009). The Bank Note Confidence Survey. Ottowa. 3. De Heij, H. (2017). A Model for Use-centered Design of Payment Instruments Applied on Banknotes. Thesis. Tilburg University. 4. De Heij, H. (2016). User Experience Functions (UXF) and User Interface Functions (UIF) of Optical Authenticity Features. Conference on Optical Document Security Reconnaissance International. San Francisco. 5. De Heij, H. (2010). Banknote design for retailers and public. DNB Occasional Studies, 8 (4). 6. De Nederlandsche Bank (2017). Daling aantal valse eurobiljetten. (Decline of false euro notes.). Press release. 7. De Nederlandsche Bank (2010). Increase in number of counterfeit euro banknotes in the Netherlands. Press release European Central Bank (2010b). Annual Report Frankfurt. 9. Firth, A. and Balodis, E. (2011). Designing better banknotes. The role of perception studies. Keesing Journal of Documents & Identity, 36, Heeger, D. (2007). Signal Detection Theory. Department of Psychology. New York University. 11. Heinen, H. and Zondervan, I. (2016). Publieksperceptie vernieuwde security features- kwalitatief onderzoek. (Public perception of renewed security features - qualitative research.). Study ordered by DNB. Motivaction. Amsterdam. Not published. 12. Klein, M.R., Gadbois, S., Christie, J.J. (2004). Perception and Detection of Counterfeit Currency in Canada. Optical Security and Counterfeit Deterrence Techniques V, 16th International Symposium Electronic Imaging: Science and Technology. San Jose. 13. Klöne, E.J. and Zondervan, I. (2017). A biennial study about knowledge and appreciation of euro banknotes among the Dutch. Prepared for De Nederlandsche Bank. Motivaction. Amsterdam. 14. Masuda, O., Pedersen, M. and Hardeberg, J.Y. (2015). Effects of awareness to security features on the confidence in banknotes. J. Print Media Technol. Res., 4(2), Randsdorp, Y. and Zondervan, I. (2015). A biennial study about knowledge and perception of euro banknotes among the Dutch. Prepared for De Nederlandsche Bank. Motivaction. Amsterdam. 16. Raymond, J. (2017). The Importance of Intaglio in the Authentication of Banknotes by the General Public. Prepared by Secure Perception Research for International Banknote Designers Association. Birmingham. 17. Raymond, J. and Jones, S. (2016). Visually Navigating a Banknote. University of Birmingham. Power point presentation to the 4th International Banknote Designers Conference. Paris. 18. Van der Horst, F., De Heij, H., Miedema, J. and Van der Woude, M. (2017). Perception of public security features on euro banknotes. IBDA Insight, 13, Van der Horst, F., Eschelbach, M., Sieber, S. and Miedema, J. (2016). Does banknote quality affect counterfeit detection?experimental evidence from Germany and the Netherlands. DNB Working Paper, Van der Horst, F., Meeter, M., Theeuwes, J. and Van der Woude, M. (2011). What is a fit banknote? The Dutch public responds. DNB Occasional Studies, 9(4). 21. Visser, J. and Dijkers, W. (2013). Euro banknotes. A study about awareness and recognition of euro banknotes among the Dutch. Prepared for De Nederlandsche Bank NV.TNS NIPO. Amsterdam. 15