Category Archives: Analytical Techniques

And the Cash Flows On

As a fraud examiner and information systems auditor, I’ve always been a big fan of the cash flow statement and I think you should be too. For the non-accountant investigators among you, the cash flow statement reveals what happened to the client’s cash during the reporting period. It’s very much like your bank account statement: You have a beginning balance of cash at the start of the month, you deposit your paycheck, you write some checks for your mortgage and groceries, and then you end the month with a new cash balance. This is what a cash flow statement is: simply a beginning balance of cash, plus or minus some cash transactions, to arrive at an ending cash balance.

Another way to view the cash flow statement is as an income statement that is adjusted for non-cash transactions and transactions that have not yet impacted cash. Non-cash transactions are transactions that affect the income statement but will never affect cash. Depreciation is a non-cash transaction that is added back to profits on the cash flow statement since cash is never paid out or collected when an asset is depreciated. The cash flow statement also clarifies transactions that immediately impact cash. A company can make a sale but not collect on it, or incur an expense and not immediately pay for it in cash. These are called accounts receivable and accounts payable, respectively. Revenues that are earned but not received and expenses that are incurred but not paid would show up on the income statement, but not on the cash flow statement. So the formula for the statement is simply …

Beginning Cash Balance
+I- Net Cash Flows from Operating Activities
+I- Net Cash Flows from Investing Activities
+I- Net Cash Flows from Financing Activities
= Ending Cash Balance

There are two methods of reporting cash flows from operations; in the direct method, the sources of operating cash flows are listed along with the uses of operating cash flows, with the difference between them being the net cash flow from operating activities. In contrast, the indirect method reconciles net income per the income statement with net cash flows from operating activities; that is, accrual-basis net income is adjusted for non-cash revenues and expenses to arrive at net cash flows from operations. The net cash flows from operating activities is the same amount regardless of which method is used. The indirect method is usually easier to compute and provides a comparison of the company’s operating results under the accrual and cash methods of accounting. As a result, most companies choose to use the indirect method, but either method is acceptable.

So what does all this provide as a tool for the fraud examiner? Simply, the cash flow statement provides any CFE with lots of neat information for further analysis in a very compact form. First of all, the statement tells you what the company’s cash receipts and cash payments were for the period. Remember that it’s unlike the income statement in that the income statement takes into account all revenue and expense transactions, whether or not they affected cash. The cash flow statement only considers transactions that involve cash.

The cash flow statement divides the company’s cash transactions into three categories:

• Operating activities, which include all cash received and paid out in connection with the company’s normal business operations, such as cash received from customers and funds paid to vendors. This category essentially encompasses any cash transactions that affect items on the income statement.
• Investing activities, which are cash flows related to the sale or purchase of non-current assets, such as fixed assets, intangible assets, and investments. This category generally covers those cash transactions that affect the asset side of the balance sheet.
• Financing activities, which are all cash inflows and outflows pertaining to the company’s debt and equity financing. Inflows include the proceeds received from issuing stocks and bonds and from borrowing money from a bank. Outflows include debt repayments and cash dividends paid to shareholders. In general, this category includes the cash transactions that affect the liabilities and owners’ equity side of the balance sheet.

In a perfect world, a company should only need loans when it has a timing problem between collecting and spending money or when it’s expanding. However, if a company expends more money than it will ever make, it will eventually go out of business. This is where the cash flow statement is so useful to the fraud examiner. You will want to get an idea of the cash flow necessary to run the business so that you will be able to tell whether the company is generating enough cash from operations to continue to do business. The examiner can also evaluate the relationship between total cash generated from financing and investing activities and the amount generated by operating activities.

Some things you will want to note from the cash flow statement in connection with any suspected financial fraud:
• Does the company have heavy demands on its operating cash each period?
• Do the inflows equal or exceed the outflows?
• Is the cash balance increasing or decreasing over time?
• Is the company making smart decisions about sources and uses of cash given its apparent financial condition?

This is information pertinent to the investigation of a wide range of fraud scenarios, the successful investigation of which involves different data than that commonly available in the income statement. The income statement alone does not reveal a complete picture of the company’s financial health, necessary for a full investigation of so many types of fraud. Evaluating income and cash flows includes considering the timing of items, such as collections of accounts receivable. In the end, a company might have a fabulous looking income statement, but might not have any cash available for operations. This may occur because the revenues recorded on the income statement have not been collected. Remember, as part of doing business, companies usually allow customers to make purchases on credit; this means the companies will collect the cash subsequent to the actual recording of the revenues. For example, a small high-tech manufacturer might have a healthy looking profit on its income statement, but not be able to pay its employees’ salaries. However, the entrepreneurial owners of the company expect all is well, since they think the net income on the income statement to be equal to the amount of cash in the company’s bank account. But, as is often the case, there’s a timing difference between when the company records a sale and when it actually receives the cash from its customers. As a result, the cash balance seldom, if ever, will match the income on the income statement. Other transactions – such as accrued or prepaid expenses, depreciation, and inventory purchases – will also cause a disparity between an organization’s net income and its net cash flows.

The statement of cash flows represents a trove of invaluable information that can cast light on virtually every aspect of a client’s financial health and, thus inform any investigation. Use it to your advantage.

Detect and Prevent

I got a call last week from a long term colleague, one of whose smaller client firms recently discovered a long running key-employee initiated fraud. My friend has been asked to assist her client in developing approaches to strengthen controls to, hopefully, prevent such disasters in the future.

ACFE training has consistently told us over the years, and daily experience repeatedly confirmed, that it is simply not possible or economical to stop all fraud before it happens. The only way for a retail concern to absolutely stop shoplifting might be to close and accept orders only over the Internet. Similarly, the only way for a bank to absolutely stop all loan fraud might be for it to stop lending money.

In general, my friend and I agreed during our conversation, that increasing preventive security can reduce fraud losses, but beyond some point, the cost of additional preventive security will exceed the related savings from reduced fraud losses. This is where detection comes in; it may be economical when prevention is not. One way to prevent a salesclerk from stealing from the register would be for the security department to carefully monitor, review, and approve every one of the clerk’s sales. However, it would likely be much more cost effective instead to implement a simple detective control: an end-of-shift reconciliation between the cash in the register and the transactions logged by the cash register during the clerk’s shift. If refunds are not given at the point of sale, the end-of-shift balance of cash in the register should equal the shift’s sales per the transaction logs minus the balance of cash in the register at the beginning of the shift. Any significant failure of these numbers to reconcile would amount to a red flag. Of course, further investigation could show that the clerk simply made an error and so did not commit fraud.

But the cost effectiveness of detective controls, like preventive controls, imposes limits. First, such controls are not cost free to implement, and improving detective controls may cost more than the results they provide. Second, detective controls produce both false positives and false negatives. A false positive occurs when a detective control signals a possible fraud that upon investigation turns up a reasonable explanation for the indicator. A false negative occurs when a detective control fails to signal a possible fraud when one exists. Reducing false negatives means increasing the fraud detection rate.

Similarly, the cost effectiveness of increasing preventive security has a limit as does the benefit of increasing the fraud detection rate. To increase the detection rate, it’s necessary to increase the frequency at which the detective control signals possible fraud. The result is more expensive investigations, and the cost of such additional investigations can exceed the resulting reduction in fraud losses.

As we all learned in undergraduate auditing, controls are essentially policies and procedures designed to minimize losses due to fraud or to other events such as errors or acts of nature. Corrective controls are merely special control types involved once a loss is known to exist. With respect to fraud, an important corrective control involves the investigation of potential frauds and the investigation and recovery process from discovered frauds.

More generally speaking, fraud investigations themselves serve not only a corrective function but also detective and preventive functions. Such investigations are detective of fraud to the extent that they follow up on fraud signals or red flags in order to confirm or disconfirm the presence of fraud. But once fraud is confirmed to exist, fraud examinations shift toward gathering evidence and become corrective by assisting in recovery from the perpetrator and other sources such as from insurance. Fraud investigations are also corrective in that they can lead to the revelation and repair of heretofore unknown weaknesses.

The end result is that the fraud investigation functions to correct the original loss, and the related discovery of the fraud scenario leads to prevention of similar losses in the future. In summary, the fraud examination has served to detect, correct, and prevent fraud. However, fraud investigations are not normally thought of as detective controls. This so is because fraud investigations tend to be much more costly than standard detective controls and therefore are normally used only when there is already some predication in the form of a fraud indicator triggered by a typical detective control. Therefore, the primary functions of fraud investigations are to address existing frauds and help to prevent future ones.

In some cases, the primary benefit of a fraud investigation might be to prevent future frauds. Even when recovery is impossible or impractical (e.g., because the thief has no assets), unwinding the fraud scheme may still have the benefit of leading to the prevention of the same scheme in the future. Furthermore, a company might benefit from spending a very large sum of money to investigate and prosecute a very small theft in order to deter other individuals from defrauding the company in the same way. Many State governments have statutes specifying that every fraud affecting governmental assets, whether large or small, must be fully investigated because taxpayer funds are involved (the assets affected are public property).

There is never a guarantee that investigating a fraud indicator will lead to the discovery of fraud. Depending on the situation, an investigation might lead to nothing at all (i.e., produce a reasonable explanation for the original red flag) or to the discovery of losses due to simple errors, waste, inefficiencies, or even uncontrollable events like acts of nature. If a lender is considering a loan application, a fraud indicator might indicate nothing, fraud, or an error. On the other hand, in regard to the possible theft of raw materials in a production process, a fraud indicator just might indicate undocumented waste or scrap.

Two important factors to consider concerning the general design of a fraud detection process are not only the costs and benefits of detecting, correcting, and preventing a given fraud scenario but also the costs and benefits of detecting, correcting, and preventing errors, waste, uncontrollable events, and inefficiencies in general. Of course, the particular costs that are relevant will vary from one type of business process to another.

As a general rule, we can say that both preventive controls and detective controls cost less than corrective controls. Corrective controls tend to involve hands-on, resource-intensive investigations, and in many cases, such investigations do not result in recovering the loss. On the other hand, preventive controls can also be quite costly. Banks pay armed guards and incur costs to maintain expensive vaults and alarm systems. Companies surround their headquarters with high fences and armed guards, and use security checkpoints and biometric key card systems inside. On the information technology side, firms use sophisticated firewalls and multi-layer access controls. The costs of all these preventive measures can add up to staggering sums in large companies. Of course, losses that are not prevented or corrected in a timely fashion can lead to the ultimate corrective measure: bankruptcy. In fact, some ACFE estimates show that about one-third of all business failures relate to some form of fraudulent activity.

One positive aspect of the cost of preventive controls is that unlike detective controls, they do not generate fraud indicators that lead to costly investigations. In fact, they tend to do their job in complete silence so that management never even knows when they prevent a fraud. The thick door of a bank vault with a time lock prevents bank employees from entering the building at night to steal its contents. Similarly, passwords, pin numbers, and biometric data silently provide access to authorized individuals and prevent access from others.

The problem with preventive controls is that they are always subject to circumvention by determined and cunning fraudsters. There is no perfect solution to preventing acts of fraud, so detection is necessary as a secondary line of defense, and in some cases, as the primary line of defense. Consider a lending company that accepts online loan applications. It may be difficult or impossible to prevent fraudulent applications, but the company can certainly put a sophisticated (and expensive) system in place to analyze applications and provide indicators that suggest when an application may be fraudulent.

In general, the optimal allocation of resources to prevention versus detection depends on the particular business process under consideration. So, there is no general rule that dictates the optimal allocation of resources between prevention versus detection. But there are some general steps that can assist in making the allocation:

1. Analyze the target business process and identify threats and vulnerabilities.
2. Select reasonable preventive controls according to the business process and customs within the client’s industry.
3. Estimate fraud losses given the assumed preventive controls.
4. Identify and add a basic set of detective controls to the system.
5. For a given set of detective controls, identify the optimal mix of false negatives versus false positives. The optimal mix depends on the costs of investigations versus the costs of losses. Large losses and small investigation costs favor relatively low false negatives and high false positives for red flags.
6. Given the assumed mix of false negative and false positive errors, estimate the incremental cost associated with adding the detective (and related corrective) controls, and estimate the resulting reduction in fraud losses.
7. Compare the reduction in fraud losses with the increase in costs associated with adding the optimal mix of detection and correction controls.
8. If increase in costs is significantly lower than the related reduction in fraud losses, consider adding more detective controls. Otherwise, accept the set of detective controls under consideration.

Needles & Haystacks

A long-time acquaintance of mine told me recently that, fresh out of the University of Virginia and new to forensic accounting, his first assignment consisted in searching, at the height of summer, through two unairconditioned trailers full of thousands of savings and loan records for what turned out to be just two documents critical to proving a loan fraud. He told me that he thought then that his job would always consist of finding needles in haystacks. Our profession and our tools have, thankfully, come a long way since then!

Today, digital analysis techniques afford the forensic investigator the ability to perform cost-effective financial forensic investigations. This is achieved through the following:

— The ability to test or analyze 100 percent of a data set, rather than merely sampling the data set.
–Massive amounts of data can be imported into working files, which allows for the processing of complex transactions and the profiling of certain case-specific characteristics.
–Anomalies within databases can be quickly identified, thereby reducing the number of transactions that require review and analysis.
–Digital analysis can be easily customized to address the scope of the engagement.

Overall, digital analysis can streamline investigations that involve a large number of transactions, often turning a needle-in-the-haystack search into a refined and efficient investigation. Digital analysis is not designed to replace the pick-and-shovel aspect of an investigation. However, the proper application of digital analysis will permit the forensic operator to efficiently identify those specific transactions that require further investigation or follow up.

As every CFE knows, there are an ever-growing number of software applications that can assist the forensic investigator with digital analysis. A few such examples are CaseWare International Inc.’s IDEA, ACL Services Ltd.’s ACL Desktop Edition, and the ActiveData plug-in, which can be added to Excel.

So, whether using the Internet in an investigation or using software to analyze data, fraud examiners can today rely heavily on technology to aid them in almost any investigation. More data is stored electronically than ever before; financial data, marketing data, customer data, vendor listings, sales transactions, email correspondence, and more, and evidence of fraud can be located within that data. Unfortunately, fraudulent data often looks like legitimate data when viewed in the raw. Taking a sample and testing it might or might not uncover evidence of fraudulent activity. Fortunately, fraud examiners now have the ability to sort through piles of information by using special software and data analysis techniques. These methods can identify future trends within a certain industry, and they can be configured to identify breaks in audit control programs and anomalies in accounting records.

In general, fraud examiners perform two primary functions to explore and analyze large amounts of data: data mining and data analysis. Data mining is the science of searching large volumes of data for patterns. Data analysis refers to any statistical process used to analyze data and draw conclusions from the findings. These terms are often used interchangeably.

If properly used, data analysis processes and techniques are powerful resources. They can systematically identify red flags and perform predictive modeling, detecting a fraudulent situation long before many traditional fraud investigation techniques would be able to do so.

Big data is now a buzzword in the worlds of business, audit, and fraud investigation. Big data are high volume, high velocity, and/or high variety information assets that require new forms of processing to enable enhanced decision making, insight discovery, and process optimization. Simply put, big data is information of extreme size, diversity, and complexity.

In addition to thinking of big data as a single set of data, fraud investigators should think about the way data grow when different data sets are connected together that might not normally be connected. Big data represents the continuous expansion of data sets, the size, variety, and speed of generation of which makes it difficult to manage and analyze.

Big data can be instrumental to fact gathering during an investigation. Distilled down to its core, how do fraud examiners gather data in an investigation? We look at documents and financial or operational data, and we interview people. The challenge is that people often gravitate to the areas with which they are most comfortable. Attorneys will look at documents and email messages and then interview individuals. Forensic accounting professionals will look at the accounting and financial data (structured data). Some people are strong interviewers. The key is to consider all three data sources in unison. Big data helps to make it all work together to tell the complete picture. With the ever-increasing size of data sets, data analytics has never been more important or useful. Big data requires the use of creative and well-planned analytics due to its size and complexity. One of the main advantages of using data analytics in a big data environment is, as indicated above, that it allows the investigator to analyze an entire population of data rather than having to choose a sample and risk drawing conclusions in the event of a sampling error.

To conduct an effective data analysis, a fraud examiner must take a comprehensive approach. Any direction can (and should) be taken when applying analytical tests to available data. The more creative fraudsters get in hiding their schemes, the more creative the fraud examiner must become in analyzing data to detect these schemes. For this reason, it is essential that fraud investigators consider both structured and unstructured data when planning their engagements.
Data are either structured or unstructured. Structured data is the type of data found in a database, consisting of recognizable and predictable structures. Examples of structured data include sales records, payment or expense details, and financial reports.

Unstructured data, by contrast, is data not found in a traditional spreadsheet or database. Examples of unstructured data include vendor invoices, email and user documents, human resources files, social media activity, corporate document repositories, and news feeds.

When using data analysis to conduct a fraud examination, the fraud examiner might use structured data, unstructured data, or a combination of the two. For example, conducting an analysis on email correspondence (unstructured data) among employees might turn up suspicious activity in the purchasing department. Upon closer inspection of the inventory records (structured data), the fraud examiner might uncover that an employee has been stealing inventory and covering her tracks in the records.

Data mining has roots in statistics, machine learning, data management and databases, pattern recognition, and artificial intelligence. All of these are concerned with certain aspects of data analysis, so they have much in common; yet they each have a distinct and individual flavor, emphasizing particular problems and types of solutions.

Although data mining technologies provide key advantages to marketing and business activities, they can also manipulate financial data that was previously hidden within a company’s database, enabling fraud examiners to detect potential fraud.

Data mining software provides an easy to use process that gives the fraud examiner the ability to get to data at a required level of detail. Data mining combines several different techniques essential to detecting fraud, including the streamlining of raw data into understandable patterns.

Data mining can also help prevent fraud before it happens. For example, computer manufacturers report that some of their customers use data mining tools and applications to develop anti-fraud models that score transactions in real-time. The scoring is customized for each business, involving factors such as locale and frequency of the order, and payment history, among others. Once a transaction is assigned a high-risk score, the merchant can decide whether to accept the transaction, deny it, or investigate further.

Often, companies use data warehouses to manage data for analysis. Data warehouses are repositories of a company’s electronic data designed to facilitate reporting and analysis. By storing data in a data warehouse, data users can query and analyze relevant data stored in a single location. Thus, a company with a data warehouse can perform various types of analytic operations (e.g., identifying red flags, transaction trends, patterns, or anomalies) to assist management with its decision making responsibilities.

In conclusion, after the fraud examiner has identified the data sources, s/he should identify how the information is stored by reviewing the database schema and technical documentation. Fraud examiners must be ready to face a number of pitfalls when attempting to identify how information is stored, from weak or nonexistent documentation to limited collaboration from the IT department.

Moreover, once collected, it’s critical to ensure that the data is complete and appropriate for the analysis to be performed. Depending on how the data was collected and processed, it could require some manual work to make it usable for analysis purposes; it might be necessary to modify certain field formats (e.g., date, time, or currency) to make the information usable.

Fraud Prevention Oriented Data Mining

One of the most useful components of our Chapter’s recently completed two-day seminar on Cyber Fraud & Data Breaches was our speaker, Cary Moore’s, observations on the fraud fighting potential of management’s creative use of data mining. For CFEs and forensic accountants, the benefits of data mining go much deeper than as just a tool to help our clients combat traditional fraud, waste and abuse. In its simplest form, data mining provides automated, continuous feedback to ensure that systems and anti-fraud related internal controls operate as intended and that transactions are processed in accordance with policies, laws and regulations. It can also provide our client managements with timely information that can permit a shift from traditional retrospective/detective activities to the proactive/preventive activities so important to today’s concept of what effective fraud prevention should be. Data mining can put the organization out front of potential fraud vulnerability problems, giving it an opportunity to act to avoid or mitigate the impact of negative events or financial irregularities.

Data mining tests can produce “red flags” that help identify the root cause of problems and allow actionable enhancements to systems, processes and internal controls that address systemic weaknesses. Applied appropriately, data mining tools enable organizations to realize important benefits, such as cost optimization, adoption of less costly business models, improved program, contract and payment management, and process hardening for fraud prevention.

In its most complex, modern form, data mining can be used to:

–Inform decision-making
–Provide predictive intelligence and trend analysis
–Support mission performance
–Improve governance capabilities, especially dynamic risk assessment
–Enhance oversight and transparency by targeting areas of highest value or fraud risk for increased scrutiny
–Reduce costs especially for areas that represent lower risk of irregularities
–Improve operating performance

Cary emphasized that leading, successful organizational implementers have tended to take a measured approach initially when embarking on a fraud prevention-oriented data mining initiative, starting small and focusing on particular “pain points” or areas of opportunity to tackle first, such as whether only eligible recipients are receiving program funds or targeting business processes that have previously experienced actual frauds. Through this approach, organizations can deliver quick wins to demonstrate an early return on investment and then build upon that success as they move to more sophisticated data mining applications.

So, according to ACFE guidance, what are the ingredients of a successful data mining program oriented toward fraud prevention? There are several steps, which should be helpful to any organization in setting up such an effort with fraud, waste, abuse identification/prevention in mind:

–Avoid problems by adopting commonly used data mining approaches and related tools.

This is essentially a cultural transformation for any organization that has either not understood the value these tools can bring or has viewed their implementation as someone else’s responsibility. Given the cyber fraud and breach related challenges faced by all types of organizations today, it should be easier for fraud examiners and forensic accountants to convince management of the need to use these tools to prevent problems and to improve the ability to focus on cost-effective means of better controlling fraud -related vulnerabilities.

–Understand the potential that data mining provides to the organization to support day to day management of fraud risk and strategic fraud prevention.

Understanding, both the value of data mining and how to use the results, is at the heart of effectively leveraging these tools. The CEO and corporate counsel can play an important educational and support role for a program that must ultimately be owned by line managers who have responsibility for their own programs and operations.

–Adopt a version of an enterprise risk management program (ERM) that includes a consideration of fraud risk.

An organization must thoroughly understand its risks and establish a risk appetite across the enterprise. In this way, it can focus on those area of highest value to the organization. An organization should take stock of its risks and ask itself fundamental questions, such as:

-What do we lose sleep over?
-What do we not want to hear about us on the evening news or read about in the print media or on a blog?
-What do we want to make sure happens and happens well?

Data mining can be an integral part of an overall program for enterprise risk management. Both are premised on establishing a risk appetite and incorporating a governance and reporting framework. This framework in turn helps ensure that day-to-day decisions are made in line with the risk appetite, and are supported by data needed to monitor, manage and alleviate risk to an acceptable level. The monitoring capabilities of data mining are fundamental to managing risk and focusing on issues of importance to the organization. The application of ERM concepts can provide a framework within which to anchor a fraud prevention program supported by effective data mining.

–Determine how your client is going to use the data mined information in managing the enterprise and safeguarding enterprise assets from fraud, waste and abuse.

Once an organization is on top of the data, using it effectively becomes paramount and should be considered as the information requirements are being developed. As Cary pointed out, getting the right data has been cited as being the top challenge by 20 percent of ACFE surveyed respondents, whereas 40 percent said the top challenge was the “lack of understanding of how to use analytics”. Developing a shared understanding so that everyone is on the same page is critical to success.

–Keep building and enhancing the application of data mining tools.

As indicated above, a tried and true approach is to begin with the lower hanging fruit, something that will get your client started and will provide an opportunity to learn on a smaller scale. The experience gained will help enable the expansion and the enhancement of data mining tools. While this may be done gradually, it should be a priority and not viewed as the “management reform initiative of the day. There should be a clear game plan for building data mining capabilities into the fiber of management’s fraud and breach prevention effort.

–Use data mining as a tool for accountability and compliance with the fraud prevention program.

It is important to hold managers accountable for not only helping institute robust data mining programs, but for the results of these programs. Has the client developed performance measures that clearly demonstrate the results of using these tools? Do they reward those managers who are in the forefront in implementing these tools? Do they make it clear to those who don’t that their resistance or hesitation are not acceptable?

–View this as a continuous process and not a “one and done” exercise.

Risks change over time. Fraudsters are always adjusting their targets and moving to exploit new and emerging weaknesses. They follow the money. Technology will continue to evolve, and it will both introduce new risks but also new opportunities and tools for management. This client management effort to protect against dangers and rectify errors is one that never ends, but also one that can pay benefits in preventing or managing cyber-attacks and breaches that far outweigh the costs if effectively and efficiently implemented.

In conclusion, the stark realities of today’s cyber related challenges at all levels of business, private and public, and the need to address ever rising service delivery expectations have raised the stakes for managing the cost of doing business and conducting the on-going war against fraud, waste and abuse. Today’s client-managers should want to be on top of problems before they become significant, and the strategic use of data mining tools can help them manage and protect their enterprises whilst saving money…a win/win opportunity for the client and for the CFE.

Finding the Words

I had lunch with a long-time colleague the other day and the topic of conversation having turned to our May training event next week, he commented that when conducting a fraud examination, he had always found it helpful to come up with a list of words specifically associated with the type of fraud scenario on which he was working.  He found the exercise useful when scanning through the piles of textual material he frequently had to plow through during complex examinations.

Data analysis in the traditional sense involves running rule-based queries on structured data, such as that contained in transactional databases or financial accounting systems. This type of analysis can yield valuable insight into potential frauds. But, a more complete analysis requires that fraud examiners (like my friend) also consider unstructured textual data. Data are either structured or unstructured. Structured data is the type of data found in a database, consisting of recognizable and predictable structures. Examples of structured data include sales records, payment or expense details, and financial reports. Unstructured data, by contrast, is data that would not be found in a traditional spreadsheet or database. It is typically text based.

Our client’s employees are sending and receiving more email messages each year, retaining ever more electronic source documents, and using more social media tools. Today, we can anticipate unstructured data to come from numerous sources, including:

• Social media posts
• Instant messages
• Videos
• Voice files
• User documents
• Mobile phone software applications
• News feeds
• Sales and marketing material
• Presentations

Textual analytics is a method of using software to extract usable information from unstructured text data. Through the application of linguistic technologies and statistical techniques, including weighted fraud indicators (e.g., my friend’s fraud keywords) and scoring algorithms, textual analytics software can categorize data to reveal patterns, sentiments, and relationships indicative of fraud. For example, an analysis of email communications might help a fraud examiner gauge the pressures/incentives, opportunities, and rationalizations to commit fraud that exist in a client organization.

According to my colleague, as a prelude to textual analytics (depending on the type of fraud risk present in a fraud examiner’s investigation), the examiner  will frequently profit by coming up with a list of fraud keywords that are likely to point to suspicious activity. This list will depend on the industry of the client, suspected fraud schemes, and the data set the fraud examiner has available. In other words, if s/he is running a search through journal entry detail, s/he will likely search for different fraud keywords than if s/he were running a search of emails. It might be helpful to look at the ACFE’s fraud triangle when coming up with a keyword list. The factors identified in the triangle are helpful when coming up with a fraud keyword list. Consider how someone in the entity under investigation might have the opportunity to commit fraud, be under pressure to commit fraud, or be able to rationalize the commission of fraud.

Many people commit fraud because of something that has happened in their life that motivates them to steal. Maybe they find themselves in debt, or perhaps they must meet a certain goal to qualify for a performance-based bonus. Keywords that might indicate pressure include deadline, quota, trouble, short, problem, and concern. Think of words that would indicate that someone has the opportunity or ability to commit fraud. Examples include override, write-off, recognize revenue, adjust, discount, and reserve/provision.

Since most fraudsters do not have a criminal background, justifying their actions is a key part of committing fraud. Some keywords that might indicate a fraudster is rationalizing his actions include reasonable, deserve, and temporary.

So, even though the concepts embodied in the fraud triangle are a good place to start when developing a keyword list, it’s also important to consider the nature of the client entity’s industry and the types of payments it makes or is suspected of making. Think about the fraud scenarios that are likely to have occurred. Does the entity do a significant amount of work overseas or have many contractors? If so, there might be an elevated risk of bribery. Focus on the payment text descriptions in journal entries or in work delated documentation, since no one calls it “bribe expense.” Some examples of word combinations in payment descriptions that might merit special attention include:

• Goodwill payment
• Consulting fee
• Processing fee
• Incentive payment
• Donation
• Special commission
• One-time payment
• Special payment
• Friend fee
• Volume contract incentive

Any payment descriptions bearing these, or similar terms warrant extra scrutiny to check for reasonableness. Also, examiners should always be wary of large cash disbursements that have a blank journal payment description.

Beyond key word lists, the ACFE tells us that another way to discover fraud clues hidden in text is to consider the emotional tone of employee correspondence. In emails and instant messages, for instance, a fraud examiner should identify derogatory, surprised, secretive, or worried communications. In one example, former Enron CEO Ken Lay’s emails were analyzed, revealing that as the company came closer to filing bankruptcy, his email correspondence grew increasingly derogatory, confused, and angry. This type of analysis provided powerful evidence that he knew something was wrong at the company.

While advanced textual analytics can be extremely revealing and can provide clues for potential frauds that might otherwise go unnoticed, the successful application of such analytics requires the use of sophisticated software, as well as a thorough understanding of the legal environment of employee rights and workplace searches. Consequently, fraud examiners who are considering adding textual analytics to their fraud detection arsenal should consult with technological and legal experts before undertaking such techniques.

Even with sophisticated data analysis techniques, some data are so vast or complex that they remain difficult to analyze using traditional means. Visually representing data via graphs,  link diagrams, time-series charts, and other illustrative representations can bring clarity to a fraud examination. The utility of visual representations is enhanced as data grow in volume and complexity. Visual analytics build on humans’ natural ability to absorb a greater volume of information in visual rather than numeric form and to perceive certain patterns, shapes, and shades more easily than others.

Link analysis software is used by fraud examiners to create visual representations (e.g., charts with lines showing connections) of data from multiple data sources to track the movement of money; demonstrate complex networks; and discover communications, patterns, trends, and relationships. Link analysis is very effective for identifying indirect relationships and relationships with several degrees of separation. For this reason, link analysis is particularly useful when conducting a money laundering investigation because it can track the placement, layering, and integration of money as it moves around unexpected sources. It could also be used to detect a fictitious vendor (shell company) scheme. For instance, the investigator could map visual connections between a variety of entities that share an address and bank account number to reveal a fictitious vendor created to embezzle funds from a company.  The following are some other examples of the analyses and actions fraud examiners can perform using link analysis software:

• Associate communications, such as email, instant messages, and internal phone records, with events and individuals to reveal connections.
• Uncover indirect relationships, including those that are connected through several intermediaries.
• Show connections between entities that share an address, bank account number, government identification number (e.g., Social Security number), or other characteristics.
• Demonstrate complex networks (including social networks).

Imagine a listing of vendors, customers, employees, or financial transactions of a global company. Most of the time, these records will contain a reference to a location, including country, state, city, and possibly specific street address. By visually analyzing the site or frequency of events in different geographical areas, a fraud investigator has yet another variable with which s/he can make inferences.

Finally, timeline analysis software aids fraud examiners in transforming their data into visual timelines. These visual timelines enable fraud examiners to:

• Highlight key times, dates, and facts.
• More readily determine a sequence of events.
• Analyze multiple or concurrent sequences of events.
• Track unaccounted for time.
• Identify inconsistencies or impossibilities in data.

Analytics Confronts the Normal

The Information Audit and Control Association (ISACA) tells us that we produce and store more data in a day now than mankind did altogether in the last 2,000 years. The data that is produced daily is estimated to be one exabyte, which is the computer storage equivalent of one quintillion bytes, which is the same as one million terabytes. Not too long ago, about 15 years, a terabyte of data was considered a huge amount of data; today the latest Swiss Army knife comes with a 1 terabyte flash drive.

When an interaction with a business is complete, the information from the interaction is only as good as the pieces of data that get captured during that interaction. A customer walks into a bank and withdraws cash. The transaction that just happened gets stored as a monetary withdrawal transaction with certain characteristics in the form of associated data. There might be information on the date and time when the withdrawal happened; there may be information on which customer made the withdrawal (if there are multiple customers who operate the same account). The amount of cash that was withdrawn, the account from which the money was extracted, the teller/ATM who facilitated the withdrawal, the balance on the account after the withdrawal, and so forth, are all typically recorded. But these are just a few of the data elements that can get captured in any withdrawal transaction. Just imagine all the different interactions possible on all the assorted products that a bank has to offer: checking accounts, savings accounts, credit cards, debit cards, mortgage loans, home equity lines of credit, brokerage, and so on. The data that gets captured during all these interactions goes through data-checking processes and gets stored somewhere internally or in the cloud.  The data that gets stored this way has been steadily growing over the past few decades, and, most importantly for fraud examiners, most of this data carries tons of information about the nuances of the individual customers’ normal behavior.

In addition to what the customer does, from the same data, by looking at a different dimension of the data, examiners can also understand what is normal for certain other related entities. For example, by looking at all the customer withdrawals at a single ARM, CFEs can gain a good understanding of what is normal for that particular ATM terminal.  Understanding the normal behavior of customers is very useful in detecting fraud since deviation from normal behavior is a such a primary indicator of fraud. Understanding non-fraud or normal behavior is not only important at the main account holder level but also at all the entity levels associated with that individual account. The same data presents completely different information when observed in the context of one entity versus another. In this sense, having all the data saved and then analyzed and understood is a key element in tackling the fraud threat to any organization.

Any systematic, numbers-based system of understanding of the phenomenon of fraud as a past occurring event is dependent on an accurate description of exactly what happened through the data stream that got accumulated before, during, and after the fraud scenario occurred. Allowing the data to speak is the key to the success of any model-based system. This data needs to be saved and interpreted very precisely for the examiner’s models to make sense. The first crucial step to building a model is to define, understand, and interpret fraud scenarios correctly. At first glance, this seems like a very easy problem to solve. In practical terms, it is a lot more complicated process than it seems.

The level of understanding of the fraud episode or scenario itself varies greatly among the different business processes involved with handling the various products and functions within an organization. Typically, fraud can have a significant impact on the bottom line of any organization. Looking at the level of specific information that is systematically stored and analyzed about fraud in financial institutions for example, one would arrive at the conclusion that such storage needs to be a lot more systematic and rigorous than it typically is today. There are several factors influencing this. Unlike some of the other types of risk involved in client organizations, fraud risk is a censored problem. For example, if we are looking at serious delinquency, bankruptcy, or charge-off risk in credit card portfolios, the actual dollars-at-risk quantity is very well understood. Based on past data, it is relatively straightforward to quantify precise credit dollars at risk by looking at how many customers defaulted on a loan or didn’t pay their monthly bill for three or more cycles or declared bankruptcy. Based on this, it is easy to quantify the amount at risk as far as credit risk goes. However, in fraud, it is virtually impossible to quantify the actual amount that would have gone out the door as the fraud is stopped immediately after detection. The problem is censored as soon as some intervention takes place, making it difficult to precisely quantify the potential risk.

Another challenge in the process of quantifying fraud is how well the fraud episode itself gets recorded. Consider the case of a credit card number getting stolen without the physical card getting stolen. During a certain period, both the legitimate cardholder and the fraudster are charging using the card. If the fraud detection system in the issuing institution doesn’t identify the fraudulent transactions as they were happening in real time, typically fraud is identified when the cardholder gets the monthly statement and figures out that some of the charges were not made by him/her. Then the cardholder calls the issuer to report the fraud.  In the not too distant past, all that used to get recorded by the bank was the cardholder’s estimate of when the fraud episode began, even though there were additional details about the fraudulent transactions that were likely shared by the cardholder. If all that gets recorded is the cardholder’s estimate of when the fraud episode began, ambiguity is introduced regarding the granularity of the actual fraud episode. The initial estimate of the fraud amount becomes a rough estimate at best.
In the case in which the bank’s fraud detection system was able to catch the fraud during the actual fraud episode, the fraudulent transactions tended to be recorded by a fraud analyst, and sometimes not too accurately. If the transaction was marked as fraud or non-fraud incorrectly, this problem was typically not corrected even after the correct information flowed in. When eventually the transactions that were actually fraudulent were identified using the actual postings of the transactions, relating this back to the authorization transactions was often not a straightforward process. Sometimes the amounts of the transactions may have varied slightly. For example, the authorization transaction of a restaurant charge is sometimes unlikely to include the tip that the customer added to the bill. The posted amount when this transaction gets reconciled would look slightly different from the authorized amount. All of this poses an interesting challenge when designing a data-driven analytical system to combat fraud.

The level of accuracy associated with recording fraud data also tends to be dependent on whether the fraud loss is a liability for the customer or to the financial institution. To a significant extent, the answer to the question, “Whose loss is it?” really drives how well past fraud data is recorded. In the case of unsecured lending such as credit cards, most of the liability lies with the banks, and the banks tend to care a lot more about this type of loss. Hence systems are put in place to capture this data on a historical basis reasonably accurately.

In the case of secured lending, ID theft, and so on, a significant portion of the liability is really on the customer, and it is up to the customer to prove to the bank that he or she has been defrauded. Interestingly, this shift of liability also tends to have an impact on the quality of the fraud data captured. In the case of fraud associated with automated clearing house (ACH) batches and domestic and international wires, the problem is twofold: The fraud instances are very infrequent, making it impossible for the banks to have a uniform method of recording frauds; and the liability shifts are dependent on the geography.  Most international locations put the onus on the customer, while in the United States there is legislation requiring banks to have fraud detection systems in place.

The extent to which our client organizations take responsibility also tends to depend on how much they care about the customer who has been defrauded. When a very valuable customer complains about fraud on her account, a bank is likely to pay attention.  Given that most such frauds are not large scale, there is less need to establish elaborate systems to focus on and collect the data and keep track of past irregularities. The past fraud information is also influenced heavily by whether the fraud is third-party or first-party fraud. Third-party fraud is where the fraud is committed clearly by a third party, not the two parties involved in a transaction. In first-party fraud, the perpetrator of the fraud is the one who has the relationship with the bank. The fraudster in this case goes to great lengths to prevent the banks from knowing that fraud is happening. In this case, there is no reporting of the fraud by the customer. Until the bank figures out that fraud is going on, there is no data that can be collected. Also, such fraud could go on for quite a while and some of it might never be identified. This poses some interesting problems. Internal fraud where the employee of the institution is committing fraud could also take significantly longer to find. Hence the data on this tends to be scarce as well.

In summary, one of the most significant challenges in fraud analytics is to build a sufficient database of normal client transactions.  The normal transactions of any organization constitute the baseline from which abnormal, fraudulent or irregular transactions, can be identified and analyzed.  The pinpointing of the irregular is thus foundational to the development of the transaction processing edits which prevent the irregular transactions embodying fraud from even being processed and paid on the front end; furnishing the key to modern, analytically based fraud prevention.

New Rules for New Tools

I’ve been struck these last months by several articles in the trade press about CFE’s increasingly applying advanced analytical techniques in support of their work as full-time employees of private and public-sector enterprises.  This is gratifying to learn because CFE’s have been bombarded for some time now about the risks presented by cloud computing, social media, big data analytics, and mobile devices, and told they need to address those risk in their investigative practice.  Now there is mounting evidence of CFEs doing just that by using these new technologies to change the actual practice of fraud investigation and forensic accounting by using these innovative techniques to shape how they understand and monitor fraud risk, plan and manage their work, test transactions against fraud scenarios, and report the results of their assessments and investigations to management; demonstrating what we’ve all known, that CFEs, especially those dually certified as CPAs, CIAs, or CISA’s can bring a unique mix of leveraged skills to any employer’s fraud prevention or detection program.

Some examples …

Social Media — following a fraud involving several of the financial consultants who work in its branches and help customers select accounts and other investments, a large multi-state bank requested that a staff CFE determine ways of identifying disgruntled employees who might be prone to fraud. The effort was important to management not only because of fraud prevention but because when the bank lost an experienced financial consultant for any reason, it also lost the relationships that individual had established with the bank’s customers, affecting revenue adversely. The staff CFE suggested that the bank use social media analytics software to mine employees’ email and posts to its internal social media groups. That enabled the bank to identify accurately (reportedly about 33 percent) the financial consultants who were not currently satisfied with their jobs and were considering leaving. Management was able to talk individually with these employees and address their concerns, with the positive outcome of retaining many of them and rendering them less likely to express their frustration by ethically challenged behavior.  Our CFE’s awareness that many organizations use social media analytics to monitor what their customers say about them, their products, and their services (a technique often referred to as sentiment analysis or text analytics) allowed her to suggest an approach that rendered value. This text analytics effort helped the employer gain the experience to additionally develop routines to identify email and other employee and customer chatter that might be red flags for future fraud or intrusion attempts.

Analytics — A large international bank was concerned about potential money laundering, especially because regulators were not satisfied with the quality of their related internal controls. At a CFE employee’s recommendation, it invested in state-of-the-art business intelligence solutions that run “in-memory”, a new technique that enables analytics and other software to run up to 300,000 times faster, to monitor 100 percent of its transactions, looking for the presence of patterns and fraud scenarios indicating potential problems.

Mobile — In the wake of an identified fraud on which he worked, an employed CFE recommended that a global software company upgrade its enterprise fraud risk management system so senior managers could view real-time strategy and risk dashboards on their mobile devices (tablets and smartphones). The executives can monitor risks to both the corporate and to their personal objectives and strategies and take corrective actions as necessary. In addition, when a risk level rises above a defined target, the managers and the risk officer receive an alert.

Collaboration — The fraud prevention and information security team at a U.S. company wanted to increase the level of employee acceptance and compliance with its fraud prevention – information security policy. The CFE certified Security Officer decided to post a new policy draft to a collaboration area available to every employee and encouraged them to post comments and suggestions for upgrading it. Through this crowd-sourcing technique, the company received multiple comments and ideas, many of which were incorporated into the draft. When the completed policy was published, the company found that its level of acceptance increased significantly, its employees feeling that they had part ownership.

As these examples demonstrate, there is a wonderful opportunity for private and public sector employed CFE’s to join in the use of enterprise applications to enhance both their and their employer’s investigative efficiency and effectiveness.  Since their organizations are already investing heavily in a wide variety of innovative technologies to transform the way in which they deliver products to and communicate with customers, as well as how they operate, manage, and direct the business, there is no reason that CFE’s can’t use these same tools to transform each stage of their examination and fraud prevention work.

A risk-based fraud prevention approach requires staff CFEs to build and maintain the fraud prevention plan, so it addresses the risks that matter to the organization, and then update that plan as risks change. In these turbulent times, dominated by cyber, risks change frequently, and it’s essential that fraud prevention teams understand the changes and ensure their approach for addressing them is updated continuously. This requires monitoring to identify and assess both new risks and changes in previously identified risks.  Some of the recent technologies used by organizations’ financial and operational analysts, marketing and communications professionals, and others to understand both changes within and outside the business can also be used to great advantage by loss prevention staff for risk monitoring. The benefits of leveraging this same software are that the organization has existing experts in place to teach CFE’s how to use it, the IT department already is providing technical support, and the software is currently used against the very data enterprise fraud prevention professionals like staff CFEs want to analyze.  A range of enhanced analytics software such as business intelligence, analytics (including predictive and mobile analytics), visual intelligence, sentiment analysis, and text analytics enable fraud prevention to monitor and assess risk levels. In some cases, the software monitors transactions against predefined rules to identify potential concerns such as heightened fraud risks in any given business process or in a set of business processes (the inventory or financial cycles).  For example, a loss prevention team headed by a staff CFE can monitor credit memos in the first month of each quarter to detect potential revenue accounting fraud. Another use is to identify trends associated with known fraud scenarios, such as changes in profit margins or the level of employee turnover, that might indicate changes in risk levels. For example, the level of emergency changes to enterprise applications can be analyzed to identify a heightened risk of poor testing and implementation protocols associated with a higher vulnerability to cyber penetration.

Finally, innovative staff CFEs have used some interesting techniques to report fraud risk assessments and examination results to management and to boards. Some have adopted a more visually appealing representation in a one-page assessment report; others have moved to the more visual capabilities of PowerPoint from the traditional text presentation of Microsoft Word.  New visualization technology, sometimes called visual analytics when allied with analytics solutions, provides more options for fraud prevention managers seeking to enhance or replace formal reports with pictures, charts, and dashboards.  The executives and boards of their employing organizations are already managing their enterprise with dashboards and trend charts; effective loss prevention communications can make effective use of the same techniques. One CFE used charts and trend lines to illustrate how the time her employing company was taking to process small vendor contracts far exceeded acceptable levels, had contributed to fraud risk and was continuing to increase. The graphic, generated by a combination of a business intelligence analysis and a visual analytics tool to build the chart, was inserted into a standard monthly loss prevention report.

CFE headed loss prevention departments and their allied internal audit and IT departments have a rich selection of technologies that can be used by them individually or in combination to make them all more effective and efficient. It is questionable whether these three functions can remain relevant in an age of cyber, addressing and providing assurance on the risks that matter to the organization, without an ever wider use of modern technology. Technology can enable the an internal CFE to understand the changing business environment and the risks that can affect the organization’s ability to achieve its fraud prevention related objectives.

The world and its risks are evolving and changing all the time, and assurance professionals need to address the issues that matter now. CFEs need to review where the risk is going to be, not where it was when the anti-fraud plan was built. They increasingly need to have the ability to assess cyber fraud risk quickly and to share the results with the board and management in ways that communicate assurance and stimulate necessary change.

Technology must be part of the solution to that need. Technological tools currently utilized by CFEs will continue to improve and will be joined by others over time. For example, solutions for augmented or virtual reality, where a picture or view of the physical world is augmented by data about that picture or view enables loss prevention professionals to point their phones at a warehouse and immediately access operational, personnel, safety, and other useful information; representing that the future is a compound of both challenge and opportunity.

Sock Puppets

The issue of falsely claimed identity in all its myriad forms has shadowed the Internet since the beginning of the medium.  Anyone who has used an on-line dating or auction site is all too familiar with the problem; anyone can claim to be anyone.  Likewise, confidence games, on or off-line, involve a range of fraudulent conduct committed by professional con artists against unsuspecting victims. The victims can be organizations, but more commonly are individuals. Con artists have classically acted alone, but now, especially on the Internet, they usually group together in criminal organizations for increasingly complex criminal endeavors. Con artists are skilled marketers who can develop effective marketing strategies, which include a target audience and an appropriate marketing plan: crafting promotions, product, price, and place to lure their victims. Victimization is achieved when this marketing strategy is successful. And falsely claimed identities are always an integral component of such schemes, especially those carried out on-line.

Such marketing strategies generally involve a specific target market, which is usually made up of affinity groups consisting of individuals grouped around an objective, bond, or association like Facebook or LinkedIn Group users. Affinity groups may, therefore, include those associated through age, gender, religion, social status, geographic location, business or industry, hobbies or activities, or professional status. Perpetrators gain their victims’ trust by affiliating themselves with these groups.  Historically, various mediums of communication have been initially used to lure the victim. In most cases, today’s fraudulent schemes begin with an offer or invitation to connect through the Internet or social network, but the invitation can come by mail, telephone, newspapers and magazines, television, radio, or door-to-door channels.

Once the mark receives and accepts the offer to connect, some sort of response or acceptance is requested. The response will typically include (in the case of Facebook or LinkedIn) clicking on a link included in a fraudulent follow-up post to visit a specified web site or to call a toll-free number.

According to one of Facebook’s own annual reports, up to 11.2 percent of its accounts are fake. Considering the world’s largest social media company has 1.3 billion users, that means up to 140 million Facebook accounts are fraudulent; these users simply don’t exist. With 140 million inhabitants, the fake population of Facebook would be the tenth-largest country in the world. Just as Nielsen ratings on television sets determine different advertising rates for one television program versus another, on-line ad sales are determined by how many eyeballs a Web site or social media service can command.

Let’s say a shyster want 3,000 followers on Twitter to boost the credibility of her scheme? They can be hers for $5. Let’s say she wants 10,000 satisfied customers on Facebook for the same reason? No problem, she can buy them on several websites for around $1,500. A million new friends on Instagram can be had for only $3,700. Whether the con man wants favorites, likes, retweets, up votes, or page views, all are for sale on Web sites like Swenzy, Fiverr, and Craigslist. These fraudulent social media accounts can then be freely used to falsely endorse a product, service, or company, all for just a small fee. Most of the work of fake account set up is carried out in the developing world, in places such as India and Bangladesh, where actual humans may control the accounts. In other locales, such as Russia, Ukraine, and Romania, the entire process has been scripted by computer bots, programs that will carry out pre-encoded automated instructions, such as “click the Like button,” repeatedly, each time using a different fake persona.

Just as horror movie shape-shifters can physically transform themselves from one being into another, these modern screen shifters have their own magical powers, and organizations of men are eager to employ them, studying their techniques and deploying them against easy marks for massive profit. In fact, many of these clicks are done for the purposes of “click fraud.” Businesses pay companies such as Facebook and Google every time a potential customer clicks on one of the ubiquitous banner ads or links online, but organized crime groups have figured out how to game the system to drive profits their way via so-called ad networks, which capitalize on all those extra clicks.

Painfully aware of this, social media companies have attempted to cut back on the number of fake profiles. As a result, thousands and thousands of identities have disappeared over night among the followers of many well know celebrities and popular websites. If Facebook has 140 million fake profiles, there is no way they could have been created manually one by one. The process of creation is called sock puppetry and is a reference to the children’s toy puppet created when a hand is inserted into a sock to bring the sock to life. In the online world, organized crime groups create sock puppets by combining computer scripting, web automation, and social networks to create legions of online personas. This can be done easily and cheaply enough to allow those with deceptive intentions to create hundreds of thousands of fake online citizens. One only needs to consult a readily available on-line directory of the most common names in any country or region. Have a scripted bot merely pick a first name and a last name, then choose a date of birth and let the bot sign up for a free e-mail account. Next, scrape on-line photo sites such as Picasa, Instagram, Facebook, Google, and Flickr to choose an age-appropriate image to represent your new sock puppet.

Armed with an e-mail address, name, date of birth, and photograph, you sign up your fake persona for an account on Facebook, LinkedIn, Twitter, or Instagram. As a last step, you teach your puppets how to talk by scripting them to reach out and send friend requests, repost other people’s tweets, and randomly like things they see Online. Your bots can even communicate and cross-post with one another. Before the fraudster knows it, s/he has thousands of sock puppets at his disposal for use as he sees fit. It is these armies of sock puppets that criminals use as key constituents in their phishing attacks, to fake on-line reviews, to trick users into downloading spyware, and to commit a wide variety of financial frauds, all based on misplaced and falsely claimed identity.

The fraudster’s environment has changed and is changing over time, from a face-to-face physical encounter to an anonymous on-line encounter in the comfort of the victim’s own home. While some consumers are unaware that a weapon is virtually right in front of them, others are victims who struggle with the balance of the many wonderful benefits offered by advanced technology and the painful effects of its consequences. The goal of law enforcement has not changed over the years; to block the roads and close the loopholes of perpetrators even as perpetrators continue to strive to find yet another avenue to commit fraud in an environment in which they can thrive. Today, the challenge for CFEs, law enforcement and government officials is to stay on the cutting edge of technology, which requires access to constantly updated resources and communication between organizations; the ability to gather information; and the capacity to identify and analyze trends, institute effective policies, and detect and deter fraud through restitution and prevention measures.

Now is the time for CFEs and other assurance professionals to continuously reevaluate all we for take for granted in the modern technical world and to increasingly question our ever growing dependence on the whole range of ubiquitous machines whose potential to facilitate fraud so few of our clients and the general public understand.

Bye-Bye Money

Miranda had responsibility for preparing personnel files for new hires, approval of wages, verification of time cards, and distribution of payroll checks. She “hired” fictitious employees, faked their records, and ordered checks through the payroll system. She deposited some checks in several personal bank accounts and cashed others, endorsing all of them with the names of the fictitious employees and her own. Her company’s payroll function created a large paper trail of transactions among which were individual earnings records, W-2 tax forms, payroll deductions for taxes and insurance, and Form 941 payroll tax reports. She mailed all the W-2 forms to the same post office box.

Miranda stole $160,000 by creating some “ghosts,” usually 3 to 5 out of 112 people on the payroll and paying them an average of $650 per week for three years. Sometimes the ghosts quit and were later replaced by others. But she stole “only” about 2 percent of the payroll funds during the period.

A tip from a fellow employee received by the company hotline resulted in the engagement of Tom Hudson, CFE.  Tom’s objective was to obtain evidence of the existence and validity of payroll transactions on the control premise that different people should be responsible for hiring (preparing personnel files), approving wages, and distributing payroll checks. “Thinking like a crook” lead Tom to readily see that Miranda could put people on the payroll and obtain their checks just as the hotline caller alleged. In his test of controls Tom audited for transaction authorization and validity. In this case random sampling was less likely to work because of the small number of alleged ghosts. So, Tom looked for the obvious. He selected several weeks’ check blocks, accounted for numerical sequence (to see whether any checks had been removed), and examined canceled checks for two endorsements.

Tom reasoned that there may be no “balance” to audit for existence/occurrence, other than the accumulated total of payroll transactions, and that the total might not appear out of line with history because the tipster had indicated that the fraud was small in relation to total payroll and had been going on for years.  He decided to conduct a surprise payroll distribution, then followed up by examining prior canceled checks for the missing employees and then scan personnel files for common addresses.

Both the surprise distribution and the scan for common addresses quickly provided the names of 2 or 3 exceptions. Both led to prior canceled checks (which Miranda had not removed and the bank reconciler had not noticed), which carried Miranda’s own name as endorser. Confronted, she confessed.

The major risks in any payroll business cycle are:

•Paying fictitious “employees” (invalid transactions, employees do not exist);

• Overpaying for time or production (inaccurate transactions, improper valuation);

•Incorrect accounting for costs and expenses (incorrect classification, improper or inconsistent presentation and disclosure).

The assessment of payroll system control risk normally takes on added importance because most companies have fairly elaborate and well-controlled personnel and payroll functions. The transactions in this cycle are numerous during the year yet result in lesser amounts in balance sheet accounts at year-end. Therefore, in most routine outside auditor engagements, the review of controls, test of controls and audit of transaction details constitute the major portion of the evidence gathered for these accounts. On most annual audits, the substantive audit procedures devoted to auditing the payroll-related account balances are very limited which enhances fraud risk.

Control procedures for proper segregation of responsibilities should be in place and operating. Proper segregation involves authorization (personnel department hiring and firing, pay rate and deduction authorizations) by persons who do not have payroll preparation, paycheck distribution, or reconciliation duties. Payroll distribution (custody) is in the hands of persons who do not authorize employees’ pay rates or time, nor prepare the payroll checks. Recordkeeping is performed by payroll and cost accounting personnel who do not make authorizations or distribute pay. Combinations of two or more of the duties of authorization, payroll preparation and recordkeeping, and payroll distribution in one person, one office, or one computerized system may open the door for errors and frauds. In addition, the control system should provide for detail control checking activities.  For example: (1) periodic comparison of the payroll register to the personnel department files to check hiring authorizations and for terminated employees not deleted, (2) periodic rechecking of wage rate and deduction authorizations, (3) reconciliation of time and production paid to cost accounting calculations, (4) quarterly reconciliation of YTD earnings records with tax returns, and (5) payroll bank account reconciliation.

Payroll can amount to 40 percent or more of an organization’s total annual expenditures. Payroll taxes, Social Security, Medicare, pensions, and health insurance can add several percentage points in variable costs on top of wages. So, for every payroll dollar saved through forensic identification, bonus savings arise automatically from the on-top costs calculated on base wages. Different industries will exhibit different payroll risk profiles. For example, firms whose culture involves salaried employees who work longer hours may have a lower risk of payroll fraud and may not warrant a full forensic approach. Organizations may present greater opportunity for payroll fraud if their workforce patterns entail night shift work, variable shifts or hours, 24/7 on-call coverage, and employees who are mobile, unsupervised, or work across multiple locations. Payroll-related risks include over-claimed allowances, overused extra pay for weekend or public holiday work, fictitious overtime, vacation and sick leave taken but not deducted from leave balances, continued payment of employees who have left the organization, ghost employees arising from poor segregation of duties, and the vulnerability of data output to the bank for electronic payment, and roster dysfunction. Yet the personnel assigned to administer the complexities of payroll are often qualified by experience than by formal finance, legal, or systems training, thereby creating a competency bias over how payroll is managed. On top
of that, payroll is normally shrouded in secrecy because of the inherently private nature of employee and executive pay. Underpayment errors are less probable than overpayment errors because they are more likely to be corrected when the affected employees complain; they are less likely to be discovered when employees are overpaid. These systemic biases further increase the risk of unnoticed payroll error and fraud.

Payroll data analysis can reveal individuals or entire teams who are unusually well-remunerated because team supervisors turn a blind eye to payroll malpractice, as well as low-remunerated personnel who represent excellent value to the organization. For example, it can identify the night shift worker who is paid extra for weekend or holiday work plus overtime while actually working only half the contracted hours, or workers who claim higher duty or tool allowances to which they are not entitled. In addition to providing management with new insights into payroll behaviors, which may in turn become part of ongoing management reporting, the total payroll cost distribution analysis can point forensic accountants toward urgent payroll control improvements.

The detail inside payroll and personnel databases can reveal hidden information to the forensic examiner. Who are the highest earners of overtime pay and why? Which employees gained the most from weekend and public holiday pay? Who consistently starts late? Finishes early? Who has the most sick leave? Although most employees may perform a fair day’s work, the forensic analysis may point to those who work less, sometimes considerably less, than the time for which they are paid. Joined-up query combinations to search payroll and human resources data can generate powerful insights into the organization’s worst and best outliers, which may be overlooked by the data custodians. An example of a query combination would be: employees with high sick leave + high overtime + low performance appraisal scores + negative disciplinary records. Or, reviewers could invert those factors to find the unrecognized exemplary performers.

Where predication suggests fraud concerns about identified employees, CFEs can add value by triangulating time sheet claims against external data sources such as site access biometric data, company cell phone logs, phone number caller identification, GPS data, company email, Internet usage, company motor fleet vehicle tolls, and vehicle refueling data, most of which contain useful date and time-of-day parameters.  The data buried within these databases can reveal employee behavior, including what they were doing, where they were, and who they were interacting with throughout the work day.

Common findings include:

–Employees who leave work wrongfully during their shift;
–Employees who work fewer hours and take sick time during the week to shift the workload to weekends and public holidays to maximize pay;
–Employees who use company property excessively for personal purposes during working hours;
–Employees who visit vacation destinations while on sick leave;
–Employees who take leave but whose managers do not log the paperwork, thereby not deducting leave taken and overstating leave balances;
–Employees who moonlight in businesses on the side during normal working hours, sometimes using the organization’s equipment to do so.

Well-researched and documented forensic accounting fieldwork can support management action against those who may have defrauded the organization or work teams that may be taking inappropriate advantage of the payroll system. Simultaneously, CFEs and forensic accountants, working proactively, can partner with management to recover historic costs, quantify future savings, reduce reputational and political risk, improve the organization’s anti-fraud policies, and boost the productivity and morale of employees who knew of wrongdoing but felt powerless to stop it.

The Who, the What, the When

CFEs and forensic accountants are seekers. We spend our days searching for the most relevant information about our client requested investigations from an ever-growing and increasingly tangled data sphere and trying to make sense of it. Somewhere hidden in our client’s computers, networks, databases, and spreadsheets are signs of the alleged fraud, accompanying control weaknesses and unforeseen risks, as well as possible opportunities for improvement. And the more data the client organization has, the harder all this is to find.  Although most computer-assisted forensic audit tests focus on the numeric data contained within structured sources, such as financial and transactional databases, unstructured or text based data, such as e-mail, documents, and Web-based content, represents an estimated 8o percent of enterprise data within the typical medium to large-sized organization. When assessing written communications or correspondence about fraud related events, CFEs often find themselves limited to reading large volumes of data, with few automated tools to help synthesize, summarize, and cluster key information points to aid the investigation.

Text analytics is a relatively new investigative tool for CFEs in actual practice although some report having used it extensively for at least the last five or more years. According to the ACFE, the software itself stems from a combination of developments in our sister fields of litigation support and electronic discovery, and from counterterrorism and surveillance technology, as well as from customer relationship management, and research into the life sciences, specifically artificial intelligence. So, the application of text analytics in data review and criminal investigations dates to the mid-1990s.

Generally, CFEs increasingly use text analytics to examine three main elements of investigative data: the who, the what, and the when.

The Who: According to many recent studies, substantially more than a half of business people prefer using e-mail to use of the telephone. Most fraud related business transactions or events, then, will likely have at least some e-mail communication associated with them. Unlike telephone messages, e-mail contains rich metadata, information stored about the data, such as its author, origin, version, and date accessed, and can be documented easily. For example, to monitor who is communicating with whom in a targeted sales department, and conceivably to identify whether any alleged relationships therein might signal anomalous activity, a forensic accountant might wish to analyze metadata in the “to,” “from,” “cc,” or “bcc” fields in departmental e-mails. Many technologies for parsing e-mail with text analytics capabilities are available on the market today, some stemming from civil investigations and related electronic discovery software. These technologies are like the social network diagrams used in law enforcement or in counterterrorism efforts.

The What: The ever-present ambiguity inherent in human language presents significant challenges to the forensic investigator trying to understand the circumstances and actions surrounding the text based aspects of a fraud allegation. This difficulty is compounded by the tendency of people within organizations to invent their own words or to communicate in code. Language ambiguity can be illustrated by examining the word “shred”. A simple keyword search on the word might return not only documents that contain text about shredding a document, but also those where two sports fans are having a conversation about “shredding the defense,” or even e-mails between spouses about eating Chinese “shredded pork” for dinner. Hence, e-mail research analytics seeks to group similar documents according to their semantic context so that documents about shredding as concealment or related to covering up an action would be grouped separately from casual e-mails about sports or dinner, thus markedly reducing the volume of e-mail requiring more thorough ocular review. Concept-based analysis goes beyond traditional search technology by enabling users to group documents according to a statistical inference about the co-occurrence of similar words. In effect, text analytics software allows documents to describe themselves and group themselves by context, as in the shred example. Because text analytics examines document sets and identifies relationships between documents according to their context, it can produce far more relevant results than traditional simple keyword searches.

Using text analytics before filtering with keywords can be a powerful strategy for quickly understanding the content of a large corpus of unstructured, text-based data, and for determining what is relevant to the search. After viewing concepts at an elevated level, subsequent keyword selection becomes more effective by enabling users to better understand the possible code words or company-specific jargon. They can develop the keywords based on actual content, instead of guessing relevant terms, words, or phrases up front.

The When: In striving to understand the time frames in which key events took place, CFEs often need to not only identify the chronological order of documents (e.g., sorted by or limited to dates), but also link related communication threads, such as e-mails, so that similar threads and communications can be identified and plotted over time. A thread comprises a set of messages connected by various relationships; each message consists of either a first message or a reply to or forwarding of some other message in the set. Messages within a thread are connected by relationships that identify notable events, such as a reply vs. a forward, or changes in correspondents. Quite often, e-mails accumulate long threads with similar subject headings, authors, and message content over time. These threads ultimately may lead to a decision, such as approval to proceed with a project or to take some other action. The approval may be critical to understanding business events that led up to a particular journal entry. Seeing those threads mapped over time can be a powerful tool when trying to understand the business logic of a complex financial transaction.

In the context of fraud risk, text analytics can be particularly effective when threads and keyword hits are examined with a view to considering the familiar fraud triangle; the premise that all three components (incentive/pressure, opportunity, and rationalization) are present when fraud exists. This fraud triangle based analysis can be applied in a variety of business contexts where increases in the frequency of certain keywords related to incentive/pressure, opportunity, and rationalization, can indicate an increased level of fraud risk.

Some caveats are in order.  Considering the overwhelming amount of text-based data within any modern enterprise, assurance professionals could never hope to analyze all of it; nor should they. The exercise would prove expensive and provide little value. Just as an external auditor would not reprocess or validate every sales transaction in a sales journal, he or she would not need to look at every related e-mail from every employee. Instead, any professional auditor would take a risk-based approach, identifying areas to test based on a sample of data or on an enterprise risk assessment. For text analytics work, the reviewer may choose data from five or ten individuals to sample from a high-risk department or from a newly acquired business unit. And no matter how sophisticated the search and information retrieval tools used, there is no guarantee that all relevant or high-risk documents will be identified in large data collections. Moreover, different search methods may produce differing results, subject to a measure of statistical variation inherent in probability searches of any type. Just as a statistical sample of accounts receivable or accounts payable in the general ledger may not identify fraud, analytics reviews are similarly limited.

Text analytics can be a powerful fraud examination tool when integrated with traditional forensic data-gathering and analysis techniques such as interviews, independent research, and existing investigative tests involving structured, transactional data. For example, an anomaly identified in the general ledger related to the purchase of certain capital assets may prompt the examiner to review e-mail communication traffic among the key individuals involved, providing context around the circumstances and timing, of events before the entry date. Furthermore, the forensic accountant may conduct interviews or perform additional independent research that may support or conflict with his or her investigative hypothesis. Integrating all three of these components to gain a complete picture of the fraud event can yield valuable information. While text analytics should never replace the traditional rules-based analysis techniques that focus on the client’s financial accounting systems, it’s always equally important to consider the communications surrounding key events typically found in unstructured data, as opposed to that found in the financial systems.