Sampling Concepts in Sales and Use Tax Audits

Sampling Concepts in Sales and Use Tax Audits Revised February 23, 2008 Dr. Will Yancey, CPA Independent Consultant Dallas, Texas will@willyancey. com 972. 387. 8558

Objective n Objective is NOT: To stress you n To make you a statistician in one hour n n Objective IS: To stimulate your thinking n To impart practical knowledge for managing audits of sales and use tax, unclaimed property, etc. n n Encourage discussion at any time 2

Copyright © You are welcome to use these concepts and repeat anecdotes. n Before you reproduce or distribute these materials, please let me know. n Dr. Will Yancey retains the copyright on all materials he creates. n 3

Know Your Data! n Availability of supporting documents is your best defense in any audit of sales and use tax, unclaimed property, or other transactions. n Do a preliminary investigation before the auditors arrive. ¨ Conduct a “walk-through” to discover and document what supporting documents are available and retrievable. n ¨ n Include electronic and paper files Identify potential exposure areas Data mining software links at http: //willyancey. com/data-mining. htm ¨ IDEA from www. audimation. com in Houston ¨ 4

Documents Available? Classified by availability of supporting documents Dark Period Cloudy Period Light Period None Partial All available 5

Dark Period n The Dark Period includes the older years where no supporting documentation can be located. n All that may be known about the Dark Period is total revenue by year, total purchases, or some other total $. 6

Light Period The Light Period covers the more recent years where we have access to all data: n Total revenue ¨ Supporting documents ¨ Name and address information ¨ n The Light Period is the basis for estimation to the Dark Period n Depending on the number of items we can: Sample, or ¨ Review the actual detail (also known as 100% sample, census, actual basis). ¨ 7

Cloudy Period n Facts and data are unclear in the Cloudy Period. n Control totals may exist. n Some supporting documents are missing. ¨ BUT, we must make a decision on where to separate the Dark Period from the Light Period. n Know your records before the auditor selects the sample! 8

Dark Period Estimation n Estimate liability for the Dark Period. ¨ Generally involves an estimated ratio of Liability-to- Revenue or Liability-to-Total-Payables from the Light Period. ¨ Project that ratio onto Dark Period. n Justification ¨ All states do estimates for bars, restaurants, and small businesses where records are unavailable. ¨ If there is less than 100% error rate in the Light Period, it is reasonable to assume less than 100% error rate in the Dark Period. 9

Sales Data n n n Billing system versus Accounts Receivable (A/R) system Customer master file Sales tax exemption certificates http: //willyancey. com/tax_sales. htm#exempt-certs n n Sales product description codes Sales tax billed but not paid (“short pay”) Bad debts: billed but no payment On consignment but sale not consummated 10

Location Data n n n Bill-to versus ship-to address Origin vs. Destination sourcing for taxing jurisdiction Local city and county situsing errors ¨ Increasing attention from local government n n n Goods received documentation Services delivered documentation Modification at central facility before deployment to field locations 11

Purchases Data n General Ledger (G/L) versus Accounts Payable (A/P) ¨ n Documenting where and how used ¨ ¨ ¨ n Detail versus aggregate Purchase orders (PO) Requisitions to vendor or company stores Authorization for expenditure (AFE) Contracts Receiving reports Affidavits from operations people Clearing accounts ¨ ¨ Work in progress (WIP) Construction in progress (CIP) Prepaid expenses Installment payments and deposits paid 12

Goods Received/Invoice Received (GR/IR) In the SAP enterprise system the GR/IR system has journal entries for 1. 2. 3. 4. Goods receipt: inventory received at the dock. Invoice verification: vendor invoice received: a three-way match is performed internally between Purchase Order (“PO”), Goods Receipt (“GR”) and vendor invoice. Vendor paid (by check or electronic funds transfer (“EFT”), or cleared. Clearing includes A/P offset against a credit or A/R. If purchaser accepts evaluated receipts from the vendor, then a simultaneous Good Receipt / Invoice Verification document is generated. SAP graphic from tinamccormick@etaxsolutions. com 13

Fruit Salad Parable Task is to estimate the amount of different fruits by sampling a single spoonful. 1. Fruit in four large chunks. 2. Fruit cut in many small pieces 3. Fruit crushed in a smoothie 4. Fruit floating in liquid 14

Triage Different strategies for different sub-populations n Exclude: No likelihood of error n Detail: Actual base exam of every item n Sample: Likely to be some errors; and cost of sampling is less than cost of detail. 15

Exclude Possible exclusions in a sales and use tax audit (but not all situations): n Always pay sales tax to vendor and always taxable: telecom, utilities, etc. n Never pay sales or use tax and always exempt: charitable contributions, government taxes and fees, inventory for resale, employee benefits n No tax consequence: some journal entries, reversals, and reclasses, etc. 16

Detail Possible detail in a sales and use tax audit (but not all situations): n High dollar items. ¨ n n n High threshold somewhere between $5, 000 and $500, 000, depending on population. Select vendors or customers where always in error. Construction in progress (CIP) Information Technology (IT) purchases that include taxable and nontaxable components 17

Reversals (“negative amounts”) occur because of reclassifications, error corrections, returns, voids, etc. n The negative amounts are offset to the positive amounts. n ¨ Matched pair = one positive $ offset one negative $ to the exact amount ¨ Multiple rows or partial offsets 18

Reversal examples Matched reversal on row 8 and 9 Matched reversals on rows 117 and 118; 119 and 120 19

Strategies for Reversals Chose a combination of these: n Match and extract ¨ Sort by descending absolute value $ Detail the largest negative items n Sample some negative items n Apply error rate from positive $ onto negative $ n Ignore the negatives if very small % of total positive population base $ Does the auditor need to see supporting documentation for every reversal? n 20

Sampling in a sales and use tax audit (but not all situations): n Likely to be some error n Cost of sampling is less than detail exam. n Mix of many different vendors. n Taxpayer and auditor can agree on sampling plan. 21

What Pie to Sample? The sample error rate is projected on to what pie? 22

Population and Sample Data Download Universe (all data available anywhere in any system) Target Population (population of audit (e. g. , outstanding checks, AR credits) Sampling Frame Estimate with statistical or nonstatistical evaluation of sampling risk (list of possible sample items) Projection Sample Base (selected items) Sample Results (error ratio or taxability ratio) 23

Drill-down Table Downloa d co u n t $ Excluded revers al pairs Excluded accou - 0 24

What Sample Size? Parable of the Footballs and the Fish n You are asked to determine the weight of 1, 000 footballs. You know they are identical in weight. You can weigh only one ball at a time. How many must you weigh? n You are asked to determine the weight of 1, 000 different fish taken from a lake. They are highly variable in weight. You can weigh only one fish at a time. How many must you weigh? 25

Stratification How could we organize the fish into piles so we could sample and make a reasonable estimate of total weight of all fish? n What criteria should we use? n How can we define criteria so each fish is in one and only one pile? n How many piles should we have? n How precise do we want the estimate? 26

Is stratification better? Stratum A Stratum B Stratum C Combined Unstratified Base $ 50 40 10 100 Error % 0% 10% 80% 100 12% Result $ - 4 8 12 27

Texas Sample Size CAMS (Computer Audit Menu System) Policy for stratified random samples http: //www. cpa. state. tx. us/taxinfo/audit/sampling/samplin 3. htm Minimum 100 per stratum if taxed and nontaxed are not in the same stratum Minimum 125 per stratum if taxed and nontaxed are mixed within that stratum 28

Minimum Sample Size – State Policies State California Multistate Tax Commission Ohio Minimum per stratum 300 or more 300 250 Indiana Texas Tennessee Florida 200, or 40 if small $ 100 or 125 70 Was 30, now 100 Citations at http: //willyancey. com/sampling. htm 29

Minimum Errors for Projection How many is the minimum number of errors in the sample results for the auditor to project from that stratum to population? Usually the number of errors includes both overpayments and underpayments. n IRS: minimum of 3 to 5 per stratum n California BOE: minimum of 3 per stratum n Texas: appears to be a minimum of 3 30

Statistical versus Nonstatistical Statistical sampling requires that: n Every sample item is selected with some known probability of selection. n Results are evaluated using mathematical statistics. n Confidence interval is presented. Nonstatistical sampling does not meet above. n Texas Comptroller uses nonstatistical sampling. n Time period block sampling is nonstatistical. 31

References n May the Force be with you! Statistical Education and Software http: //willyancey. com/statistics. htm n Sampling in Sales and Use Tax Audits http: //willyancey. com/sampling. htm n Sampling SUT Audit Review Articles http: //willyancey. com/sampling-sut-review. html 32