Kế toán, kiểm toán - Chapter nine: Audit sampling: an application to substantive tests of account balances

Audit Sampling: An Application to Substantive Tests of Account BalancesChapter NineSubstantive Tests of Details of Account Balances The statistical concepts we discussed in the last chapter apply to this chapter as well. Three important determinants of sample size are:Desired confidence level.Tolerable misstatement.Expected misstatement. Population plays a bigger role in some of the sampling techniques used for substantive testing. Misstatements discovered in the audit sample must be projected to the population, and there must be an allowance for sampling risk.Substantive Tests of Details of Account BalancesConsider the following information about the inventory account balance of an audit client:The ratio of misstatement in the sample is 2%(€2,000 ÷ €100,000)Applying the ratio to the entire population produces a bestestimate of misstatement of inventory of €60,000.(€3,000,000 × 2%)Book value of inventory account balance3,000,000€ Book value of items sampled100,000€ Audited value of items sampled98,000 Total amount of overstatement observed in audit sample2,000€ Substantive Tests of Details of Account Balances The results of our audit test depend upon the tolerable misstatement associated with the inventory account. If the tolerable misstatement is €50,000, we cannot conclude that the account is fairly stated because our best estimate of the projected misstatement is greater than the tolerable misstatement.Monetary-Unit Sampling (MUS)MUS uses attribute-sampling theory to express a conclusion in monetary amounts (e.g. in euros or other currency) rather than as a rate of occurrence. It is commonly used by auditors to test accounts such as accounts receivable, loans receivable, investment securities and inventory. Monetary-Unit Sampling (MUS)MUS uses attribute-sampling theory (used primarily to test controls) to estimate the percentage of monetary units in a population that might be misstated and then multiplies this percentage by an estimate of how much the euros are misstated.Monetary-Unit Sampling (MUS)Advantages of MUSWhen the auditor expects no misstatement, MUS usually results in a smaller sample size than classical variables sampling.The calculation of the sample size and evaluation of the sample results are not based on the variation between items in the population.When applied using the probability-proportional-to-size procedure, MUS automatically results in a stratified sample. Monetary-Unit Sampling (MUS)Disadvantages of MUSThe selection of zero or negative balances generally requires special design consideration.The general approach to MUS assumes that the audited amount of the sample item is not in error by more than 100%.When more than one or two misstatements are detected, the sample results calculations may overstate the allowance for sampling risk. Steps in MUS SamplingSteps in MUS SamplingSampling may be used for substantive testing to:Test the reasonableness of assertions about a financial statement amount (i.e. is the amount fairly stated). This is the most common use of sampling for substantive testing.Develop an estimate of some amount.Steps in MUS SamplingFor MUS the population is defined as the monetary value of an account balance, such as accounts receivable, investment securities or inventory.Steps in MUS SamplingAn individual euro represents the sampling unit.Steps in MUS SamplingA misstatement is defined as the difference between monetary amounts in the client’s records and amounts supported by audit evidence.Steps in MUS SamplingSteps in MUS SamplingThe auditor selects a sample for MUS by using a systematic selection approach called probability-proportionate-to-size selection. The sampling interval can be determined by dividing the book value of the population by the sample size. Each individual euro in the population has an equal chance of being selected and items or ‘logical units’ greater than the interval will always be selected.Steps in MUS SamplingAssume a client’s book value of accounts receivable is €2,500,000, and the auditor determined a sample size of 93. The sampling interval will be €26,882 (€2,500,000 ÷ 93). The random number selected is €3,977 the auditor would select the following items for testing:CumulativeSampleAccountBalanceEurosItem1001 Ace Emergency Centre2,350€ 2,350€ 1002 Admington Hospital15,495 17,845 3,977€ (1)1003 Jess Base945 18,780 1004 Good Hospital Corp.21,893 40,673 30,859 (2)1005 Jen Mara Corp.3,968 44,641 1006 Axa Corp.32,549 77,190 57,741 (3)1007 Green River Mfg.2,246 79,436 1008 Bead Hospital Centres11,860 91,306 84,623 (4)••••••••1213 Andrew Call Medical- 2,472,032 1214 Lilly Health26,945 2,498,977 2,477,121 (93)1215 Janyne Ann Corp.1,023 2,500,000€ Total Accounts Receivable2,500,000€ 3,977€ 26,882 30,859€ Steps in MUS SamplingAfter the sample items have been selected, the auditor conducts the planned audit procedures on the logical units containing the selected euro sampling units.Steps in MUS SamplingThe misstatements detected in the sample must be projected to the population. Let’s look at the following example:Book value2,500,000€ Tolerable misstatement125,000€ Sample size93 Desired confidence level 95%Expected amount of misstatement25,000€ Sampling interval26,882€ Example InformationSteps in MUS SamplingBasic Precision using the Table If no misstatements are found in the sample, the best estimate of the population misstatement would be zero euros.€26,882 × 3.0 = €80,646 upper misstatement limitSteps in MUS SamplingMisstatements DetectedIn the sample of 93 items the following misstatements were found:€3,284 ÷ €21,893 = 15%Because the Axa balance of €32,549 is greater than the interval of €26,882, no sampling risk is added. Since all the euros in the large accounts are audited, there is no sampling risk associated with large accounts.CustomerBook ValueAudit ValueDifference Tainting Factor Good Hospital21,893€ 18,609€ 3,284€ 15%Marva Medical Supply6,705 4,023 2,682 40%Axa Corp.32,549 30,049 2,500 NALearn Heart Centres15,000 - 15,000 100%Steps in MUS SamplingCompute the Upper Misstatement LimitWe compute the upper misstatement limit by calculating basic precision and ranking the detected misstatements based on the size of the tainting factor from the largest to the smallest.(0.15 × €26,882 × 1.4 = €5,645)Customer Tainting Factor Sample Interval Projected Misstatement 95% Upper Limit Upper Misstatement Basic Precision1.00 26,882€ NA3.080,646€ Learn Heart Centres1.00 26,882 26,882 1.7 (4.7 - 3.0)45,700 Marva Medical0.40 26,882 10,753 1.5 (6.2 - 4.7)16,130 Good Hospital0.15 26,882 4,032 1.4 (7.6 - 6.2)5,645 Add misstatments greaterthat the sampling interval:Axa Corp.NA26,882 NA2,500 Upper Misstatement Limit150,621€ Steps in MUS SamplingWe compare the tolerable misstatement to the upper misstatement limit. If the upper misstatement limit is less than or equal to the tolerable misstatement, we conclude that the balance is not materially misstated. In our example, the final decision is whether the accounts receivable balance is materially misstated or not.Steps in MUS SamplingIn our example the upper misstatement limit of €150,621 is greater than the tolerable misstatement of €125,000, so the auditor concludes that the accounts receivable balance is materially misstated.When faced with this situation, the auditor may:Increase the sample size.Perform other substantive procedures.Request the client adjust the accounts receivable balance.If the client refuses to adjust the account balance, the auditor would consider issuing a qualified or adverse opinion.Risk When Evaluating Account BalancesEffect of Understatement MisstatementsMUS is not particularly effective at detecting understatements. An understated account is less likely to be selected than an overstated account.The most likely error will be reduced by €2,688(– 0.10 × €26,882)Customer Book Value Audit Value Difference Tainting Factor Wayne County Medical2,000€ 2,200€ (200)€ -10%Non-Statistical Sampling for Tests of Account BalancesThe sampling unit for non-statistical sampling is normally a customer account, an individual transaction, or a line item on a transaction. When using non-statistical sampling, the following items must be considered: Identifying individually significant items. Determining the sample size. Selecting sample items. Calculating the sample results.Identifying Individually Significant ItemsThe items to be tested individually are items that may contain potential misstatements that individually exceed the tolerable misstatement. These items are tested 100% because the auditor is not willing to accept any sampling risk.Determining the Sample Size and Selecting the SampleSampleSize=Sampling Population book valueTolerable – Expected misstatement× Confidence factorAuditing standards require that the sample items be selected in such a way that the sample can be expected to represent the population.Calculating the Sample ResultsOne way of projecting the sampling results to the population is to apply the misstatement ratio in the sample to the population. This approach is known as ratio projection.If the population total is €200,000, the projected misstatement would be €20,000 (€200,000 × 10%)Assume the auditor finds €1,500 in misstatements in a sample of €15,000. The misstatement ratio is 10%.Calculating the Sample ResultsA second method is the difference projection. This method projects the average misstatement of each item in the sample to all items in the population.The projected misstatement would be €30,000 (€3 × 10,000). Assume misstatements in a sample of 100 items total €300 (for an average misstatement of €3), and the population contains 10,000 items.Non-Statistical Sampling ExampleThe auditor’s of Calabro Wireless Service have decided to use non-statistical sampling to examine the accounts receivable balance. Calabro has a total of 11,800 (15 + 250 + 11,535) accounts with a balance of €3,717,900. The auditor’s stratify the accounts as follows:Non-Statistical Sampling ExampleThe auditor decides . . .Based on the results of the tests of controls, the risk of material misstatement is assessed as low.The tolerable misstatement is €55,000, and the expected misstatement is €15,000.The desired level of confidence is moderate based on the other audit evidence already gathered.All customer account balances greater than €25,000 are to be audited.Non-Statistical Sampling Example× Confidence factorSampleSize=Sampling population book valueTolerable - Expected misstatementSampleSize=€3,167,900€40,000× 1.2 = 95 (rounded)€3,717,900 – €550,000€55,000 – €15,000Non-Statistical Sampling ExampleThe auditor sent positive confirmations to each of the 110 (95 + 15) accounts selected. Either the confirmations were returned or alternative procedures were successfully used. Four customers indicated that their accounts were overstated and the auditors determined that the misstatements were the result of unintentional error by client personnel. Here are the results of the audit testing: Amount ofBook ValueAudit ValueOver-StratumBook Valueof Sampleof SampleStatement>€25,000550,000€ 550,000€ 549,500€ 500€ >€3,000850,500 425,000 423,000 2,000 €25,000500€ 500€ >€3,0002,000 4,002 <€3,000250 6,298 Total projected misstatement10,800€ €250 ÷ 92,000 × €2,317,400Ratio of Misstatementin Stratum Tested100%€2,000 ÷ 425,000 × €850,500Why Did Statistical Sampling Fall Out Of Favour?Firms found that some auditors were over relying on statistical sampling techniques to the exclusion of good judgement.There appears to be poor linkage between the applied audit setting and traditional statistical sampling applications.Classical Variables SamplingClassical variables sampling uses normal distribution theory to evaluate the characteristics of a population based on sample data. Auditors most commonly use classical variables sampling to estimate the size of misstatement.Sampling distributions are formed by plotting the projected misstatements yielded by an infinite number of audit samples of the same size taken from the same underlying population.Classical Variables SamplingA sampling distribution is useful because it allows us to estimate the probability of observing any single sample result. In classical variables sampling, the sample mean is the best estimate of the population mean.Classical Variables SamplingAdvantagesWhen the auditor expects a relatively large number of differences between book and audited values, this method will normally result in smaller sample size than MUS.The techniques are effective for both overstatements and understatements.The selection of zero balances generally does not require special sample design considerations.Classical Variables SamplingDisadvantagesDoes not work well when little or no misstatement is expected in the population.To determine sample size, the auditor must estimate the standard deviation of the audit differences.If few misstatements are detected in the sample data, the true variance tends to be underestimated, and the resulting projection of the misstatements and the related confidence limits are not likely to be reliable.Applying Classical Variables SamplingDefining the Sampling UnitThe sampling unit can be a customer account, an individual transaction, or a line item. In auditing accounts receivable, the auditor can define the sampling unit to be a customer’s account balance or an individual sales invoice included in the account balance.Applying Classical Variables SamplingDetermining the Sample SizewhereCC = Confidence coefficientSD = Estimated standard deviation of audit differences.SampleSize=Population size (in sampling units) × CC × SDTolerable misstatement – Estimated misstatement2Applying Classical Variables SamplingThe Confidence Coefficient (CC) is associated with the desired level of confidence. The desired level of confidence is the complement of the risk that the auditor will mistakenly accept a population as fairly stated when the true population misstatement is greater than tolerable misstatement.Applying Classical Variables SamplingThe year-end balance for accounts receivable contains 5,500 accounts with a book value of €5,500,000. The tolerable misstatement for accounts receivable is set at €50,000. The expected misstatement has been judged to be €20,000. The desired confidence is 95%. Based on work completed last year, the auditor estimates the standard deviation at €31. Let’s calculate sample size.SampleSize5,500 × 1.96 × €31€50,000 – €20,0002== 125Applying Classical Variables SamplingCalculating the Sample ResultsThe sample selection usually relies on random-selection techniques. Upon completion, 30 of the customer accounts selected contained misstatements that totalled €330.20. Our first calculation is the mean misstatement in an individual account which is calculated as follows:Meanmisstatementper samplingitem=Total audit differenceSample size€330.20125=€2.65Applying Classical Variables SamplingThe mean misstatement must be projected to the population€14,575 = 5,500 × €2.65 Population size × Mean misstatementper sampling itemProjectedpopulationmisstatement=(in sampling units)Applying Classical Variables SamplingThe formula for the standard deviation is . . .SD =Total squaredaudit differences–Mean differenceper sampling item2SampleSize×Sample size – 1=€36,018.32 – (125 × 2.652)125 – 1= €16.83Applying Classical Variables SamplingConfidenceboundPopulationsizeCCSDSample size××==5,500 × 1.96 × €16.83125√ConfidenceintervalProjectedmisstatementConfidencebound±== €14,575 ± €16,228€16,228Applying Classical Variables SamplingIf both limits are within the bounds of tolerable misstatement, the evidence supports the conclusion that the account is not materially misstated.(€50,000)€50,000Lowerlimit(€1,653)Projectedmisstatement€14,575Upperlimit€30,803€0Tolerable MisstatementEnd of Chapter 9