The Shrinking Phantom

     Of all the things that comprise a store's book inventory, one of the more interesting factors to consider is the previous inventory counts. The previous inventory counts can have a bearing on the audit-to-book variances for the current audit. An inventory count is not a seperate entinty all to itself, but rather part of a series of measurements to continually monitor conditions within a store. The first count you do for a client is always going to be the most difficult, not only are you unfamiliar with the store and with their unique merchandise, but also another counter's work will have an impact on that day's audit results, and you have no idea on how accurate that counter was. On an inventory audit the inventory counts will get booked into the store's inventory replacing what they had before. At this point the inventory is "zeroed out". If a count was not particularly accurate and those numbers get "booked", than a highly accurate count on the next audit would produce what's called a "bounce back". A bounce back is essentially the result when one audit comes out over or short, and the subsequent audit results in the exact opposite effect essentially cancelling the two results out. There are actually many things that can create a bounce back, like a badly timed booking of an invoice, but the main suspect is always the inventory counts. But what would happen if one inaccurate count was followed by another inaccurate count, would this produce a bounce back? That may depend on the magnitude and direction of the error.

     To answer this question I'm going to focus on financial inventory counts, primarily for c-stores, and I'll start by discussing the counting error that we already know will occur, error from sales adjustments. Obviously if we count a store while it's still open there will be some level of error in the adjustments made for sales during the count. The size of the error may be small enough to be of little consequence on the audit results, yet the erorr will still be there. It would still throw off the number that get booked, and it should still cause a slight bounce back for the next audit, or would it? If you think about it when won't you have sales adjustment error during a count? When the store gets sold or shut down and they decide to close the doors for a count I suppose, but outside of that probably never. What type of bounce back effects can we expect when we know that error will always be present in every single count?

     Let's say for instance that a sales adjustments inflates the inventory by $20. If the exact same error takes place on the next count, nothing would bounce back at all. The $20 boost on the second count would cancel out the bounce back effect from the first count. If the error on the second count shorted the count by $20, then it would bounce back too much, you would lose the $20 on the second count, and you would be hit with the bounce back from the first count creatign a total shortage of $40. With error taking place every single time lots of scenerios could unfold, you could see inventories padded on consecutive counts, you could see inventories shorted on consecutive counts, you could see inventories go from being over and then to being short and then back again swaying back and forth like a pendulum, only it never quite evens itself out. Unfortunetly when it comes to inventory counting errors there are more things to discuss then the "nickels and dimes" one loses or gains from the half-sales method.

     Inventory error can also appear in cases where there's poor estimating, honest counting mistakes that go undetected, or in cases of inventory fraud. Fraud is the most interesting to consider because it is far more intentional, and the most likely to be repeated on subsequent counts. Talent at estimating can rise and fall, making the exact some honest mistake two times in a row and not fixing it either time is unlikely, but purposely throwing off a count is not always a one-time occurance.

     In most cases of audit fraud an inventory is padded to prevent (or more accurately to conceal) large variances on the audit-to-book comparasions. There are several reasons why a counter would pad an inventory, part of it may be to prevent confrontation with the store managers by having an agreeable audit result, part of it could be to expediate the inventory wrap-up, part of it could be to improve on their level of production. An highly skilled counter I once worked with told me a story of a manager who came up to him one time during a count and told him "count it high or go home". This could easily be a mantra for a lot of inventory companies. Fraud is at its worst when it is the result of collusion between the counter and the store management.

    Let's say that a store manager somehow steals $500 from a store during an audit period, he could cover it up by having a counter pad the next count by $500, thus resulting in decent looking audit results that avoid the scruntiny of loss prevention or accounting. There's only one problem, eventually there's going to be another audit. The store manager has to address the issue that the book inventory still overstates the 'actual' inventory by $500. If a counter pads an inventory, he essentially adds money to the count for merchandise that isn't really in the store at all, something I'd like to refer to as "phantom inventory". How does the manager and maybe even the counter as well, deal with the eventual bounce back effect that this phantom inventory will cause? There are several ways to do this.

     The counter could on the next count pad the inventory again by the same amount. This would offset the bounce back that should occur and may even result in a decent audit result. The problem though is that by offsetting one bounce back effect, you'd be setting up another one to occur on the next count. A counter would then have to pad the count on the next count and the next count, and so on and so on. If the counter pads the inventory by the same amount each time, we could refer to this as the "constant phantom" method. This type of scheme would only delay the bounce back from occuring, it wouldn't eliminate it all together. At some point the padding will stop, and the evitable shortage will occur. If one can delay the bounce back effect long enough it may make it difficult to find the source of the shortage.

     Another approach would be to try to deal with the bounce back effect that gets created when an inventory is padded. In what I'll call the "shrinking phantom" method, consider a scenerio where a counter initially pads the inventory by $500 to conceal a theft. He could pad the inventory by $450 on the next count, then pad the inventory by $400 on the count after that, then $350 on the next count and so on until the counter no longer needs to pad the inventory at all. What's interesting about this method would be that each subsequent count gets more and more accurate to the point where a counter doesn't need to add any phantom inventory at all. What this method does in this case is take the initial loss of $500 that should occur in one audit and spreads it out over 10 audits, where each count has a smaller bounce back effect of $50. However the trick to the shrinking phantom method is in the spread. If you spread it too thick, the audit losses might be big enough to draw some scruntiny from loss prevention, if you spread to too thi, you'll end up employing a system that may take forever to complete. Depending on the aduit frequency 10 audits could take 1 to 2 years to complete, and ideally you'd want the same counter for all the counts. The amount of time that is required to use this method of systematic padding could be a deterant to it's use, another reason for this method not to be used might be greed.

     If the manager and counter successfully cover up the theft during one audit period, they may figure why not do it again. But here's where the bounce back effect works against them. If another theft of $500 takes place during the next audit period, then the counter has to count $1000 of phantom inventory in order to conceal the theft. $500 of it would be needed to counteract the bounce back effect from the last count, and another $500 to conceal the most recent theft. If this trend continues, the phantom inventory will continue to grow and may reach levels to a point where the total inventory counted seems infathomable. The thing about padding an inventory is that the extra money has to go somewhere on the inventory counts. A counter has to stick this phantom inventory somewhere in one or more sections of the store. If phantom inventory continues to increase the current to previous comparasions could start to signs of trouble. Not only that but the book numbers will also rise. In fact they'll show a secular upward trend over the course of several sudits, and this usually is a sign of trouble. Inventory levels don't just steadily and continually grow by natural means alone. From their essay "What Causes Inventory Fraud?", Henry/Jackson detail a story of a counter who used this "growing phantom" method. On the first audit the pad was $2,000, on the second audit it was $4,000, on the third $6,000 and so on, eventually a store that normally carries $30,00 was showing an inventory total of $52,000. Phantom inventory obvisouly was the cause for the sharp rise in inventory and if Jack Henry is writing about this counter then obvisouly he got caught. Constant theft in each audit will leave the counter with few options when it comes to concealing it. Eventually the numbers will get too big and will invite scrunity from Loss Prevention or Accounting. And if the counter decides to stop padding the inventory, then he'll wind up creating the audit results from hell, which is guaranteed to bring scruntiny.

     Then there are situations when a counter misses counting merchandise. This will cause the inventory to be booked short, and create what I call "bonus inventory", where a store actually has merchandise that isn't on book. This would create a bounce back effect on the next count provided that no theft takes place. If someone where to steal bonus inventory, it wouldn't cause a shortage, because they'd be stealing stuff that isn't on book, it'd be as if the merchandise was never theoritcally there to begin with. With bonus inventory you could create the audit shortage before the theft. A collusion scheme with bonus inventory could be interesting to implement. although it would still include purposing counting things short and creating audit shortages. Probably most collusionists would prefer a plan that keeps audit results fairly even not intentionally short.

     Of course trying to identify bonus inventory or phantom inventory is not easy, and just like with counter accuracy, the audit-to-book comparasions offer little insight. When you get right down to it there's very little connection (if any at all) between counter accuracy and the stores' audit results. It's possible for a count that produces a bounce back to be highly accurate, and for a count that doesn't to be somewhat inaccurate. It's also possible to have decent audit results right after a theft, and have a sizeable audit shrink after an audit period with no theft at all. When you think about it all one gets with an audit-to-book comparasion is a series of numbers. One has to dig deeper for the story behind these numbers. All by themselves numbers never speak.