Even or rounded-dollar amounts do not normally occur at high-frequency rates. Therefore numbers that are rounded to tens, hundreds, and thousands may be considered as anomalies and some attention should be given to them.

While it is intuitive that high, even amounts paid should be reviewed, some low amounts may be subject to fraud or abuse. Take, for example, reimbursement of travel expenses. The organization may set maximum amounts for each category of reimbursement. These maximums would likely be in even-dollar amounts. Meals may be subject to certain maximums, such as lunch being set at a maximum of $20.00 and dinner at $50.00. Daily accommodations may be fixed at a maximum of $200.00. To ensure that the maximums are not abused, the claims should be checked against receipts. Valid claims should have receipts that show the maximum amount or are in excess of the maximum amount.

Even amounts need to be identified. This is simple to accomplish in IDEA. We have a large file consisting of payments that contains 2,066,536 records totaling $15,258,988,474.48. We want to identify amounts that are in the even tens of thousands. Using the criteria equation of (PAYMENT_AMOUNT % 10000) = 0 .AND. PAYMENT_AMOUNT < > 0, it takes mere seconds to obtain 1,799 matches of even tens of thousands totaling $402,920.00.00.

The "%" sign in the equation is not the percentage sign, but rather represents the MOD or modulus. The MOD operator is simply the remainder after a calculation. For example, 10 mod 3 would result in 1, which is the remainder of 10 -h 3 = 3.

In our equation, the payment amount mod by 10,000 will give us a remainder. If the remainder is zero, then we would have even 10,000 amounts. The second condition of the equation of PAYMENT_AMOUNT < > 0 eliminates any amounts that were originally zero in the field. Zero divided by any amount (10,000 in our example) would always return a zero.

To obtain amounts of even tens our equation would be (PAYMENT_AMOUNT % 10) = 0 .AND. PAYMENT_AMOUNT < > 0.

For amounts of even hundreds, our equation would be (PAYMENT_AMOUNT % 100) = 0 .AND. PAYMENT_AMOUNT < > 0.

Finally, to obtain the more popular even thousand amounts, our equations would be (PAYMENT_AMOUNT % 1000) = 0 .AND. PAYMENT_AMOUNT < > 0. The formula is shown in the Equation Editor in Figure 5.42.

It seems that even tens of thousands are signee cant as shown in the results in Figure 5.43. While only 0 .09 percent of the records matched, these records represent 2.64 percent of the total amounts. The results should be visually scanned and the appropriate payments should be reviewed in detail.

Keep in mind that many payment amounts are normally rounded, particularly amounts such as consulting fees or rent. These types of payments would be low-risk anomalies. A sharp eye should be focused on even amounts where rounding is not expected.

FIGURE 5.42 Isolating Even Tens of Thousands Amounts

FIGURE 5.43 Results of Applying the Even Tens of Thousands Amount Equation

Found a mistake? Please highlight the word and press Shift + Enter