Fraud detection : identifying conflicts of interest with graphs

Procurement departments are supposed to save companies money. What happens when they are hurt by conflicts of interest? Graph technologies like Neo4j and Linkurious can help find suspicious connections in employee and vendor records : let’s see how graphs can be used for fraud detection!

According to wikipedia, a conflict of interest is a set of circumstances that creates a risk that professional judgement or actions regarding a primary interest will be unduly influenced by a secondary interest. Take for example an elected official. His primary interest is to serve the public but he also wants to be wealthy and get reelected. These are secondary interests. There may be conflicts of interest when the elected official’s primary interests are influenced by his secondary interests. Such a situation may occur if a corporation is going to be influenced by a law supported by the elected official and if the corporation offers to pay for the elected official’s campaign. In that case the politician has a conflict of interest between his duty to the public and his desire to get reelected.

The existence of a conflict of interest doesn’t involve any wrongdoing. Just because the politician is in a situation where he is facing conflicting motivations doesn’t mean he is doing anything illegal. What’s reprehensible would be for him to change one of his decisions based on his secondary interest. For this reason, corporations often encourage people to disclaim potential conflicts of interest. This way the organization can manage the risk.

Conflicts of interest can be prevented. The company can remove the person experiencing the conflict of interest of any decisions affecting the third-party involved. The company can simply decide to apply greater scrutiny. Deals involving the person experiencing the conflict of interest and the third party will be watched closely.

The annual cost of workplace fraud has been estimated at over $400 billion by the Association of Certified Fraud Examiners. For John Verver, VP at ACL (an internal audit technology company) procurement fraud is among the top five areas to monitor for employee fraud. Procurement represents a great risk of conflict of interest : in their capacity as employees individuals are responsible to buy products and services at a cheap rate.

Vendors can influence procurement officers by offering them different advantages like gifts or money. A conflict of interest may also arise from a personal relationship (romantic or otherwise).

There are a few techniques fraudsters use. The accounting firm Clark Schaefer Hackett has listed the 6 most common scenarios :

• Phantom vendor : An employee sets up a fictitious vendor and uses it to submit false invoices.
• Conflict of interest : An employee abuses his or her position to award contracts to vendors in exchange for personal gain.
• Kickbacks : A vendor pays kickbacks to employees who facilitate the payment of false or inflated invoices. Often, the price is increased to cover the amount of the kickback.
• Bid rigging : Two or more vendors coordinate their bids to eliminate competition, thus increasing the price, including:
• Duplicate or false invoices : A vendor submits duplicate or false invoices for products or services that weren’t delivered, usually in collusion with an employee.
• Price fixing and defective pricing : Price fixing includes bid rigging and other anti-competitive agreements among vendors to maintain or raise prices. Defective pricing occurs when a vendor uses inaccurate or incomplete pricing data to reach a price agreement or it charges a higher price than the company agreed to.

Conflicts of interest are particularly hard to fight. If someone doesn’t volunteer the information from which the conflict of interest arises, how can the company find out?

Conflicts of interest are essentially about relationships. Two persons who are working together happen to have an unknown relationship that influence their work. Companies can find out about these relationships by accident. Someone who is aware of the relationship between the vendor and the employee may step in. This can be a colleague or even a person outside the company. Their motivations can be jealousy, personnal gain or ethical concerns.

Of course, relying only on that kind of accident is a poor company policy for fighting employee/vendors fraud. It is especially true as there are a lot of information sources available to find potentially dangerous connections. Companies have access to emails, social media, news, financial records, purchase orders, employee and vendors lists and a wide range of public databases. Finding conflicts of interest is a matter of mining these information to identify relationships.

Let’s look at this table of vendors and employees a fictional firm has. We can see that a few employees seem to have close relationships with several vendors. For example, Jim Bristol lives on 30 Dean Wells Ct and uses the phone number “6394565886”. A quick look at the vendors list shows that the company is doing business with Stone Grass Company LLC. That company shares a phone number and an address with Jim Bristol. That might be the company of Jim Bristol’s wife and that company might not be doing anything wrong. But chances are that if Jim Bristol and Stone Grass Company LLC are involved in a transaction, Jim’s employers will want to have a good look at what is going on.

This is a small example of how data can help uncover hidden conflicts of interest. Of course finding such an example in a vast amount of data can be super tricky. That’s where graphs can be very powerful :

• graph databases like Neo4j are perfect to store information about relationships ;
• it’s easy then to search for patterns within these relationships (for example look for people who are involved in a transaction and share the same address) ;
• tools like Linkurious can help analysts investigate suspicious case by looking more closely at the data ;

Fraud is a serious issue. With today’s technologies companies have the tools to fight against rogue employees. Graph technologies like Linkurious make it possible to investigate large amount of data and find hidden connections. This way companies can identify potential conflicts of interest and act on it.