Wirecard, one of Germany’s largest financial services companies, had long been suspected of employing questionable accounting methods before its auditor publicly announced in 2019 that it could not confirm the existence of €1.9 billion in cash. Utilizing a series of internal spreadsheets published by the Financial Times, this article applies data visualization techniques to highlight intriguing trends and relationships in the data. This rare glimpse into internal company data provides a visual peek behind the curtain at an accounting scandal.
Auditors should perpetually be looking to how new technologies can help them do their jobs more efficiently and effectively. The profession has received repeated reminders that data visualization has the ability to significantly enhance audit quality (ICAEW, “Data analytics for external auditors,” 2016, https://bit.ly/2UPlpC4). This article addresses a growing area within the profession by showing how visualizations can help auditors to discover patterns, identify anomalies, and gleam other useful information to assist them in planning and performing an audit (AICPA, “Audit analytics and continuous audit,” 2015, https://bit.ly/2UTK3l3).
In June 2020, Ernst & Young (EY) Germany announced it was unable to issue its 2019 audit report for Wirecard, one of Germany’s largest publicly traded companies, because it could not confirm the existence of €1.9 billion in cash. This followed years of speculation regarding suspicious accounting at the company in the financial press. Utilizing a series of internal spreadsheets published by the Financial Times, detailing monthly sales and profit transactions for customers of some of Wirecard’s largest subsidiaries, this article delves into the scandal using data visualization techniques to highlight intriguing trends and relationships in the data. Applying these techniques after the fraud was disclosed in 2020 provides evidence confirming suspicious financial information, even though other audit and investigative procedures ultimately uncovered the fraud.
Wirecard AG was founded in Germany in 1999 and grew to include subsidiaries around the world. Wirecard started as a payment processor, serving as a “middleman” between online stores and the payment networks of issuing banks or credit card companies, receiving cash and holding it briefly before paying it on to merchants, and retaining some of this amount as commission revenue on the transaction (Dan McCrum, “Wirecard: adjust your perspective,” Financial Times, Jul. 23, 2015, https://on.ft.com/3xd3Iuv). Through a series of acquisitions and growth, the company expanded to include a banking division and purchased a number of small payments companies. Some of these subsidiaries provided merchants with credit card payment terminals, which were supported via Wirecard’s payment system technology. The revenue from these transactions was supposed to be deposited into escrow or trust accounts.
Wirecard records revenues on a gross basis according to IFRS 15. Specifically, Wirecard describes its revenue recognition policy in the notes to its 2018 audited financial statements as follows:
When acquiring partners and/or other platforms are used for processing transactions and Wirecard has control over the transaction, Wirecard defines itself as the principal and the acquiring partner is defined only as a service provider (agent) of Wirecard. Wirecard integrates the various stages for the fulfillment of the performance obligation (processing of the whole payment transaction). Accordingly, all transaction fees incurred by the merchant are recognized as revenue. Fees of other service providers involved in the settlement of the payment transaction, in particular fees for credit-card issuing banks, credit-card companies, payment service providers, and external technical operators are recognized as cost of materials once the service is used or at the time the costs are incurred (Wirecard, Annual Report, 2018, p. 144, https://bit.ly/3tp5ecp).
For example, if an individual pays for a $10 item online with a credit card, the merchant nets $9.70. Wirecard recognizes the $0.30 fee incurred by the merchant as revenue, then deducts $0.20 in fees to banks/credit card companies as cost of materials, resulting in a gross margin of $0.10. The $0.30 fee for the merchant is typically calculated as a percentage of the original sale price plus a flat fee per transaction; for example, payment processor Square states on its website that it charges 2.6% on each sale plus $0.10 per each transaction (https://squ.re/3h8V2zO). The $0.20 cost of materials (e.g., the fee to banks/credit card companies) is typically calculated as a percentage of the original sales price. There may also be other fees charged to the merchant by the payment processor (e.g., Wirecard), such as chargeback fees, which occur when an individual disputes a transaction.
In March 2018, an external legal team investigated Wirecard’s Asian business, headquartered in Singapore, and found evidence of forgery and falsification of revenue through invented and back-dated sales agreements (Dan McCrum & Stefania Palma, “Wirecard’s Law Firm Found Evidence of Forgery and False Accounts,” Financial Times, Feb. 1, 2019, https://on.ft.com/3xbj039). In October 2019, KPMG Germany was hired to conduct a special audit following concerns from whistleblowers over Wirecard’s accounting practices. KPMG published its report in April 2020, which stated that it could not verify the existence of €1 billion in revenue booked from third parties from 2016 to 2018 (Kevin Granville, “Wirecard, a Payments Firm, is Rocked by a Report of a Missing $2 Billion,” New York Times, Jun. 19, 2020, https://nyti.ms/3jvFuaY). After repeated delays, EY Germany announced that it was unable to issue its 2019 audit report of Wirecard because it could not confirm the existence of €1.9 billion in cash balances in trust accounts.
Looking at the Data
In October 2019, the Financial Times published a series of internal correspondence and documents it obtained related to Wirecard’s finance team (Dan McCrum, “The Wirecard documents, explained,” Financial Times, Oct. 15, 2019, https://on.ft.com/3h9WqC6). Within these documents were spreadsheets detailing monthly sales and profit transactions for some of Wirecard’s largest subsidiaries, specifically CardSystems Middle East FZ-LLC (CardSystems, based in Dubai) and Wirecard UK & Ireland Ltd. (WCUKI, based in Dublin). While the data is only related to certain income statement items and certain subsidiaries, these examined entities contributed more than 25% to the €1.0 billion in consolidated revenue in 2016. There are 34 distinct customer names (e.g., merchants that use Wirecard for payment processing) included in the data sets across the two subsidiaries. In its investigation, the Financial Times attempted to contact some of these customers; 15 responded that they had never heard of the Wirecard subsidiaries referenced. The data sets detailing these customers and the underlying sales transactions are from 2016 and 2017, years in which Wirecard received unqualified audit opinions.
Before presenting the author’s data visualization analysis, there are important caveats to note. First, the internal documents utilized were prepared by Wirecard. There is no information as to whether the company’s independent auditors had access to these documents and, if so, which audit procedures were performed on these files in the years prior to the public disclosure of the accounting scandal. This article is not meant to serve as a rebuke of the work performed by the audit team, as the author has no knowledge of conversations between the audit team and Wirecard regarding transactions in these documents. Second, the data is limited: although totals and amounts are consistent across multiple files, its completeness cannot be verified. For the CardSystems subsidiary, the data is for all four quarters of 2017, while for the WCUKI subsidiary, the data is limited to Q3 2016 and Q1 2017. Multiple years of data would strengthen any conclusions to be drawn from the analysis. Third, data analytic and visualization techniques are not the only elements of audit work. Auditors must conduct tests of internal controls, as well as substantive tests and analytical procedures, in order to holistically complete an audit and form an opinion based on all available evidence. Lastly, while professional skepticism is paramount, the purpose of an audit is not to detect fraud. Approaching data analysis after seeing headlines of suspicious accounting likely results in a more extreme view of professional skepticism than that taken by an auditor.
However, e-mail correspondence between the finance director at WCUKI and the head of international finance at Wirecard raises significant red flags as to the occurrence of the transactions underlying the revenue reported in these data sets. In response to an auditor request, the head of international finance sent an e-mail to subsidiaries requesting proper documentation of revenue recognition following the adoption of IFRS 15. The WCUKI finance director responded, attaching the spreadsheets used in this article, and asked, “How would you like to approach this topic as regards documenting how revenue is booked here, as I only get a report usually quarterly to book revenue without the backup data to support the calculations?” (Dan McCrum, “The Wirecard documents, explained,” Financial Times, Oct. 15, 2019, https://on.ft.com/3h9WqC6).
Data visualizations can highlight intriguing trends or relationships that may raise questions as to the assertions regarding proper revenue recognition. If used during an audit, these visualization techniques could assist an audit team in pinpointing specific areas in which to conduct further testing or pursue more extensive client inquiries.
Data visualizations can highlight intriguing trends or relationships that may raise questions as to the assertions regarding proper revenue recognition.
As the PCAOB points out in its Staff Audit Practice Alert 12, “Matters Related to Auditing Revenue in an Audit of Financial Statements,” when auditing revenue, industry averages such as gross margin should be utilized to develop expectations. According to a 2017 McKinsey report on the payments industry (https://mck.co/3waImg0), Wirecard’s two closest competitors, in terms of market capitalization, are Worldpay Group from the United Kingdom and Square from the United States; both also report revenues from payment processing on a gross basis. In Exhibit 1, a bar chart shows the average gross margin ratio for the customers reported by both CardSystems and WCUKI, with reference lines added showing the competitors’ ratios (calculated utilizing publicly available annual reports for 2017), Wirecard’s consolidated gross margin ratio (according to its annual report), and the overall average ratio for each subsidiary. The gross margin ratio is calculated as the commission revenue, which is the resulting profit Wirecard receives from payment processing (e.g., fees charged to customers less fees paid to card processors), divided by sales revenue (e.g., fees charged to customers). From the example above, this would be $0.10 divided by $0.30, or 33%.
To see if there is a correlation with the number of transactions processed by Wirecard for a customer and the gross margin ratio, the bars are color-coded by average transaction count, from highest (darkest) to lowest (lightest). An examination of Exhibit 1 raises questions not only as to why the consolidated gross margin ratio of Wirecard is much higher than its competitors, but also why the two subsidiaries report higher average gross margin than the consolidated group. In addition, there appears to be no relationship between the number of transactions processed by Wirecard for a customer and the gross margin ratio. Thus, it is worth investigating what is driving the large variation in customer gross margin ratio, from a high of 80% to a low of 5%.
CardSystems reported “€58M of commission, equivalent of 4.4% of the €1.3 billion of payments processed …Such lucrative commissions are typically associated with high-risk transactions and raise questions about where Wirecard finds merchants willing to pay such fees” (Dan McCrum, “Wirecard relied on three opaque partners for almost all its profit,” Financial Times, Apr. 24, 2019, https://on.ft.com/3yaUNKj). Square’s 2017 annual report repeatedly refers to its gross payment volume (GPV), which is the total dollar amount of all card payments processed by Square’s customers, net of refunds. Square reported a 1.06% transaction-based margin (e.g., commission revenue) as a percentage of GPV for 2017 (https://bit.ly/364i6JR). Following the example provided above, this would be calculated as $0.10 divided by $10 or 1%.
Exhibit 2 displays a highlighted table of this transaction-based margin calculated for Wirecard’s customers over the periods available. Given the uneven periods in the datasets, Exhibit 2 is split into (2a) and (2b) to separately show CardSystems and WCUKI, respectively. The commission rate for one of CardSystems’ customers, “Firstcag,” jumped from 2.06% to 15.73% from April to May and 14.30% to 3.12% from June to July (Exhibit 2a). Similarly large fluctuations can be seen with other customers, which would be worthy of further exploration and discussion with the client. For WCUKI (Exhibit 2b), “Axxis” reports higher margins than all other customers, while “Shimatomo” reports a considerably lower margin (0.10%), and is the only customer with a consistent transaction-based margin across all the months analyzed. These visualizations show that gaining an understanding of the potential varying policies as they relate to each customer are important. There are large fluctuations in this margin, across time and customer, and Wirecard’s margin appears high for the industry.
Exhibit 3 displays a box-and-whiskers plot, with a reference line to Square’s transaction-based margin, to further examine the spread of this ratio for Wirecard across the customers of both subsidiaries and to detect any outliers in the data. After examining this figure, “Axxis” has a significantly higher transaction-based margin than that of all other customers and the margin for all but one of Wirecard’s customers is greater than that of its competitor Square.
Number of Transactions Processed
A plot of the number of transactions each customer processed with the CardSystems subsidiary throughout 2017 shows a steady trend during the year for some customers, but large fluctuations for others. Filtering this visualization shows that four customers in particular experienced a similar pattern in transaction counts over the year (“BANCBIN”, “FIRSTCAG”, “FOREXT” and “MOBILMAT”). The line graph in Exhibit 4 shows a consistent rise from January to March for these four customers, followed by an approximately 48% decrease for April. Three of the four customers show a similar pattern of insignificant fluctuations from July to December while “BANCBIN” experiences a drastic fluctuation during the last quarter of the year. Given that these customers have a low average transaction count compared to others, they may have escaped scrutiny during audit procedures. However, utilizing visualization and filtering techniques allows an auditor to highlight interesting trends that are worthy of further investigation.
Overall, the data visualizations described here show some patterns, behaviors, and transactions worthy of additional investigation through further discussion between auditor and client as well as investigation of additional supporting evidence. The results presented support the notion that “effective visualization can be a useful tool in identifying areas of financial information that require additional substantive testing” (Lauren Cunningham and Sarah Stein, “Using Visualization Software in the Audit of Revenue Transactions to Identify Anomalies,” Issues in Accounting Education, vol. 33, no. 4, 2018, p. 39). Utilizing real world data sets from a company suspected of major fraud, this article provides insight into a current accounting scandal and contributes to the ongoing discussion of the importance of emerging technologies and techniques in audit procedures.