آیا مقالات روزنامه ها درباره کلاهبرداری در کارت های اعتباری بر میزان استفاده از کارت بدهی تأثیر می گذارند؟

This paper investigates the impact of newspaper articles about skimming fraud on debit card usage in the Netherlands using daily transaction data and daily newspaper announcements from January 1st 2005 to December 31st 2008. Key finding is that articles about skimming fraud significantly affect same day debit card usage. The direction and strength of the media effects strongly depend on the specific characteristics of the publications, such as type of fraud addressed and their position in the newspaper, but above all on the frequency with which they come out. The effects, however, are economically small compared to other factors, such as calendar and holiday effects, and do not sustain or accumulate in the long run. Yet, some first cost calculations demonstrate that the impact of media attention on total retail payments efficiency is not to be underestimated.

During the last three decennia, debit cards have rapidly grown into widely used payment instruments at points-of-sale (POS) in the Netherlands. The ongoing increase in its acceptance and usage, however, has made the card increasingly attractive for fraud, and in particular for skimming fraud, where the card data on the magnetic stripe is copied and the PIN is captured at the POS or an automated teller machine (ATM) in order to produce a counterfeit card. Total skimming fraud in the Netherlands increased materially over the past few years, from less than EUR 4 million in 2005 to almost EUR 40 million in 2011 (NVB, 2012 and Currence, 2011). Although the financial damages are still relatively small compared to total debit card sales, the overall societal consequences could be more widespread. Fraud incidents receive a fair amount of media attention in which not only the victims but the entire population is addressed. This may affect overall payment behaviour, as consumers may lose their confidence in the debit card and shift away to other means of payment. Since earlier studies show that the total costs of paying are not negligible and that each individual instrument carries a different cost (e.g. Brits and Winder, 2005, Humphrey, 2010 and Schmiedel et al., 2012), this could eventually affect the overall efficiency of the retail payment system. Clear evidence of safety incidents affecting overall consumer confidence and payment behaviour, however, is lacking. Research into the impact of (perceived or actual) safety on payment choices is scarce and not providing a unanimous answer. Several theories and findings (e.g. Jonker, 2007, Bolt and Chakravorti, 2008, He et al., 2008, Borzekowski et al., 2008, Alvarez and Lippi, 2009 and Kosse, 2010) find that safety is one of the factors considered when choosing a particular instrument. Others, however (such as Yin and DeVaney, 2001, Schuh and Stavins, 2010 and Ching and Hayashi, 2010) find no evidence of safety playing an important role. Therefore, the aim of this paper is to further analyse consumers’ payment behaviour in relation to safety. More precisely, this paper focusses on the impact of newspaper publications about debit card skimming fraud on debit card usage. In using actual high-frequency transaction data and actual newspaper announcements, this work adds to the existing payments literature which is mainly based on consumers’ perceptions and stated behaviour. Moreover, thus far, the impact of media reports on payment choices has not been considered and tested for at all. Therefore, this paper provides new insights into the extent to which payment habits are affected by safety incidents and in particular through the media attention they get. I use a rich set of daily transaction data and newspaper announcements from January 1st 2005 to December 31st 2008. The transaction data was provided by Equens, the Dutch Automated Clearing House (ACH), and covers all daily debit card transactions made by Dutch residents at POS terminals in the Netherlands. Daily newspaper announcements on debit card fraud were extracted from the LexisNexis database, covering both national and regional papers. Key finding is that debit card usage is significantly affected by news reports on skimming fraud. The magnitude and size of the news effects strongly vary with the specific features of the announcements, such as the type of fraud addressed and their position in the paper, but above all with the frequency with which they come out. Overall, skimming fraud news depresses same day card usage, with consumers’ reactions being stronger in periods when more articles are published. A first calculation of the direct social costs associated with the temporary shift in payment behaviour shows that the influence of media attention is not to be neglected when assessing total fraud costs. However, although significant, the media effects are economically small and only last for one day, with consumers reverting back to their regular payment behaviour almost immediately. The paper proceeds as follows: Section 2 describes the development of debit card usage and skimming fraud in the Netherlands. Section 3 presents a selective review of the relevant literature and Section 4 describes the data and methodology. Section 5 reports the results illustrating the impact of newspaper publications on total debit card usage and Section 6 sheds some light on potential social cost implications. Section 7 summarises and concludes.

This paper investigates the impact of newspaper publications about debit card skimming fraud on aggregate debit card usage in the Netherlands using a rich set of daily transaction data and newspaper announcements from January 1st 2005 to December 31st 2008. A first key finding is that news about skimming fraud significantly affects debit card usage. This finding is robust to controlling for a wide range of calendar, holiday and weather variables and a time trend. Moreover, I find that the direction and the size of the news effects strongly vary with the specific characteristics of the publications, such as the type of fraud addressed and the position in the paper, but above all with the frequency with which consumers are confronted with it. Overall, consumers are found to use their card less on days when newspapers report about skimming fraud. The reduction in card usage is strongest in periods with a high frequency of publication days, but weakens again as soon as the media attention decreases. The economic significance of consumers’ response to skimming fraud news provides several meaningful messages. First, it shows that sustained media attention for card fraud has the potential to affect consumers’ confidence and through the induced change in payment behaviour to affect the overall cost efficiency of a payment system. First cost calculations show that during the second half of 2008, for example, the direct social costs associated with the temporary change in payment behaviour in the Netherlands may have amounted up to EUR 4.5 million. This is still small in relative terms and excludes the potential benefits of media attention. Yet it demonstrates that the impact of media attention should not be neglected when assessing the total costs of card fraud. Whereas this paper focuses on skimming fraud only, the results are likely to prevail for other payments-related fraud and safety incidents as well, and underline the importance of all stakeholders being constantly ready to reducing fraud and safety risks to a minimum. Second, similar to the results found in other research fields, the news effects only last for one single day, with consumers reverting back to their regular behaviour almost immediately. Also, the finding that the strength of consumers’ reactions weakens again after periods of radio silence indicates that news effects do not sustain or accumulate in the long run. This might suggest that consumers’ confidence in the debit card is relatively sturdy and not easily affected. However, future research would be valuable, in particular to study more directly the potential long run and cumulative effects of payment fraud news. In this paper, I assess the duration of daily news effects using lagged publication dummies. One limitation of this method is that it passes over the fact that newspapers may report about fraud for a few days in a row. This prolonged fraud news exposure might lead to consumers strengthening their reactions, which may not be correctly captured by the current approach and require a different methodology. In addition, future research on this issue would benefit from combining actual transaction records with micro level survey data. As demonstrated by Wakamori and Welte (2012), consumers may have heterogeneous preferences and hence exhibit different payment behaviour. Some, for instance, may permanently reduce their card usage in response to fraud news, whereas others may only temporarily change their behaviour by postponing their card payments. These differences might not be detected when using aggregate transaction data. In contrast, aggregate transaction data would wrongly point at a temporary fall in card payments that rapidly normalises again. Therefore, with micro survey data on individual choices and preferences one would be better equipped to further explore the long run and cumulative effects of fraud publications. Only then, we are able to assess whether Alsem et al. (2008) were right in stating that “Nothing is as old as yesterday’s news”.