Accounts Receivable: An Audit Simulation ABSTRACT: The project is an easy-to-use simulation allowing students to observe and participate in the accounts receivable confirmation process. The simulation contains 1,000 customer accounts from which students select a random sample. Confirmations are created in viewable *html documents which are electronically mailed to customers. Some customers return their viewable confirmation documents with discrepancies which students must identify. For customers who do not respond, the simulation creates the invoices, bills of lading and purchase orders necessary for students to preform alternative procedures. Although the simulation requires little class time, students develop a familiarity which is beneficial in subsequent lectures. After downloading the Excel spreadsheet, students report that the project requires approximately one hour to complete. Students gain an understanding of (1) the confirmation process; (2) alternative procedures for receivables; (3) evaluation of audit evidence; (4) preparation of work papers; and (5) statistical sampling as a tool to manage risk. “The auditor must obtain sufficient appropriate audit evidence by performing audit procedures to afford a reasonable basis for an opinion regarding the financial statements under audit.” (Auditing Standards Board 2010 AU Section 150.02) 1. Introduction This is your first week on the job as a new hire at the CPA firm, DC&H, LLP and you are participating in training for new hires. Today’s training covers the confirmation of accounts receivable and uses data from one of the firm’s clients, Charles Cabinets. In addition to the confirmation of accounts receivable, you will also cover learn how to vouch from the sales journal to source documents which is often called the three-way match, and how to prepare workpapers. Although there are situations when management may have incentive to understate sales and accounts receivable, for example to minimize income taxes, our primary concern for these accounts is typically overstatement. Consequently, this exercise will focus on overstatement of accounts receivable. When auditing balances, as opposed to transactions, the related financial statement assertion is existence. In 1938, the Securities Exchange Commission (SEC 1940 page 7) investigated the McKesson & Robbins, Inc. fraud and found that the auditors had failed to detect $19 million in fictitious receivables and inventory. At that time, auditing standards did not require the confirmation of receivables. In their findings the SEC stated “The facts of this case, however, demonstrate the utility of circularization and the wisdom of the profession in subsequently adopting confirmation of accounts and notes receivable as a required procedure whenever practicable and reasonable, and where the aggregate amounts of notes and accounts receivable represents a significant proportion of the current assets or of the total assets of a concern.” (Securities and Exchange Commission) This exercise covers auditing standards relating to the confirmation process, audit sampling, evaluation of results, and audit documentation. You will also perform a hands-on training exercise using an Excel simulation. After completing this project you should understand the confirmation process used to audit accounts receivable; one type of alternative audit procedure that can be performed for customers who do not respond to confirmations; how statistical sampling can be used to control detection risk; how to evaluate the audit evidence generated by the audit procedures; and how to prepare audit workpapers for accounts receivable. 2. The Confirmation Process (Auditing Standards Board 2010 AU Section 330) Accounts receivable are addressed in The Confirmation Process (ASB 2010 AU Section 330.34). “Confirmation of accounts receivable is a generally accepted auditing procedure. As discussed in paragraph .06, it is generally presumed that evidence obtained from third parties will provide the auditor with higher-quality evidence than is typically available from within the entity. Thus, there is a presumption that the auditor will request the confirmation of accounts receivable during the audit unless one of the following is true: 2 Accounts receivable are immaterial to the financial statements. The use of confirmations would be ineffective. The auditor’s combined assessed level of inherent risk and control risk is low, and the assessed level, in conjunction with the evidence expected to be provided by analytical procedures or other substantive tests of details, is sufficient to reduce audit risk to an acceptably low level for the applicable financial statement assertions. In many situations, both confirmation of accounts receivable and other substantive tests of details are necessary to reduce audit risk to an acceptably low level for the applicable financial statement assertions.” 3. Audit Sampling (Auditing Standards Board 2010 AU Section 350) The first step in the confirmation process is to determine how many customer accounts must be confirmed in order to provide sufficient evidence and then determine which customers to confirm. Although auditing standards permit the use of non-statistical sampling, statistical sampling provides auditors with an objective method to determine the sufficiency of the evidence. This can be very beneficial if DC&H needs to substantiate its position during peer review, PCAOB inspections or litigation. On most audits, DC&H uses monetary-unit-sampling (MUS) to calculate the sample size, select the sample and analyze the results. MUS is a form of probability-proportional-to-size sampling which may have been discussed in your college statistics course. However, today’s exercise uses mean-per-unit (MPU) sampling, a classical variables approach. We use MPU sampling in this exercise because college curricula typically use classical variables sampling to introduce confidence intervals and hypothesis tests. So, you should be familiar with classical variables sampling. This exercise builds on that knowledge and demonstrates a practical application for confidence intervals and hypothesis tests which are two statistical procedures with which you are familiar. 3 Perhaps the most important reason to use classical variables sampling is to illustrate how sample size can be used in audit planning to manage detection risk. In statistics, alpha (or type I) risk is associated with confidence intervals and beta (or type II) risk is associated with hypothesis testing. In auditing literature, these correspond with the risk of incorrect rejection (alpha risk) and the risk of incorrect acceptance (beta risk). “The risk of incorrect rejection is the risk that the sample supports the conclusion that the recorded account balance is materially misstated when it is not materially misstated. The risk of incorrect acceptance is the risk that the sample supports the conclusion that the recorded account balance is not materially misstated when it is materially misstated.” (ASB 2010 AU Section 350.12) If our primary concern is that accounts receivable may be overstated, there are four possible outcomes. If accounts receivable are in fact fairly presented, we may correctly conclude they are fairly presented or incorrectly conclude they are materially overstated. If accounts receivable are materially overstated, we may correctly conclude accounts receivable are materially overstated or incorrectly conclude they are fairly presented. These outcomes are illustrated in Table 1. Table 1 If accounts receivable are actually fairly presented materially overstated and the auditor concludes accounts receivble are fairly presented materially overstated correct conclusion incorrect acceptance β risk or Type II risk incorrect rejection α risk or Type I risk correct conclusion If the accounts receivable are materially overstated and the sample size is not large enough, then the risk of incorrectly concluding the balance is fairly presented (i.e., the risk of incorrect acceptance) will be unacceptably high. This increases the probability that DC&H might 4 issue an unqualified opinion on financial statements that are materially misstated. Audit risk will be unacceptably high, which exposes DC&H to legal liability. If the accounts receivable are fairly presented and the sample size is not large enough, then the risk of incorrectly concluding the balance is materially overstated (i.e., the risk of incorrect rejection) will be unacceptably high. This increases the probability that DC&H will perceive a need to expand the scope of the audit or even attempt to modify their opinion on financial statements that are fairly presented. Such a conclusion will lead to inefficiency or even damage the firm’s relationship with the client and may result in losing the client. Sample sizes that are larger than necessary reduce the profitability of the engagement and inconvenience the client by performing excessive procedures. At the extreme, firms that consistently perform excessive procedures price themselves out of the market and find it difficult to obtain or retain clients. Charles Cabinet’s accounts receivable has a $2,908,144.44 balance which includes 1,000 customers with active accounts. The standard deviation of these 1,000 accounts is $1,204.33. The standard deviation is a critical element in determining the appropriate sample size. It is important to realize that if any of the customers’ audited balances differ from their recorded balances, the standard deviation will change. If the actual standard deviation differs from that used to calculate the required sample size, sampling risk will differ from the planned risk of incorrect acceptance. Although professional standards do not require auditors to use statistical techniques to determine sample size, those standards do provide the following guidance. “To determine the number of items to be selected in a sample for a particular test of details, the auditor should consider the tolerable misstatement and the expected misstatement, the audit risk, the characteristics of the population, the assessed risk of material misstatement (inherent risk and control risk), and the assessed risk for other substantive procedures related to the same assertion. An auditor who applies statistical 5 sampling uses tables or formulas to compute sample size based on these judgments. An auditor who applies nonstatistical sampling uses professional judgment to relate these factors in determining the appropriate sample size. Ordinarily, this would result in a sample size comparable to the sample size resulting from an efficient and effectively designed statistical sample, considering the same sampling parameters.” (ASB 2010 AU Section 350.23) In this exercise, the audit team has already performed tests of controls and the results of those tests provide evidence that controls over credit sales transactions are very effective. On this basis, control risk was assessed as low for credit sales transactions. Because credit sales are such a significant class of transactions, DC&H’s policies require that inherent risk always be assessed as high for credit sales transactions. The risk of material misstatement (RMM) is the the auditor’s combined assessment of internal risk and control risk. Table 2 provides guidance to determine the acceptable level for risk of incorrect acceptance. In this example, Table 2 indicates the required sample size should be calculated using a 20 percent risk of incorrect acceptance. Any deviation from this firm guidance requires the engagement partner’s approval. Table 2 Appropriate level for Risk of Incorrect Acceptance Inherent Risk High Moderate Low Effectiveness of Controls less effective effective very effective CR = high CR = moderate CR = low RMM = high RMM = high RMM = high RoIA = .05 RoIA = .10 RoIA = .20 RMM = high RMM = mod RMM = mod RoIA = .10 RoIA = .25 RMM = mod RMM = mod RoIA = .20 RoIA = .30 RoIA = .30 RMM = low analytical procedures 4. The Confirmation Process Continued (Auditing Standards Board 2010 AU Section 330) In an actual audit, after you determine the required sample size, an auditor with specialized IT skills would select a random sample of customers from Charles Cabinets' accounts 6 receivable subsidiary ledger. You would then inform the firm’s controller which customers should be confirmed and provide him or her with DC&H's positive confirmation template from which to prepare the confirmation requests. However, in this training session, the Excel simulation will select the sample of customers. Although we have the client prepare the confirmations, we must review the confirmation requests and oversee the confirmations being placed in the envelopes. Confirmation requests must not be mailed from the client’s mail room. It is essential to mail the confirmations from a public post office or from DC&H’s office. The confirmation mailed to customers will include a return envelope that is pre-addressed to DC&H’s office. After ten days, you would go to DC&H’s office and pick up the responses from those customers who have returned their confirmations. The Excel workbook you have been provided contains the simulation. Go to the Summary Tab and enter the required sample size in the designated cell and click the “Generate” button to select a random sample of customers. Next, click the “Create” button and the simulation will prepare confirmations for those customers in the sample. The red “Confirms” link allows you to review these confirmations in your web browser. Next, the simulation will electronically mail the confirmations to the selected customers. In order to retrieve the first round of confirmations returned by customers, click the “First” button. This is equivalent to picking up the mail from DC&H’s office. The green “Firsts” link allows you to observe these responses in your web browser. You will need to agree the customer name, address and balance on the confirmation response with the accounts receivable subsidiary ledger. The information from the subsidiary ledger is included on the “Results” tab of the workbook. The “Results” page of the workbook has a column to record the amount reported 7 by the customer. The “Results” page is designed to be used when you prepare the workpapers and evaluate the evidence obtained. In this simulation you should assume the amount reported by the customer is the correct amount. On actual engagements, you would perform additional procedures to investigate discrepancies between the amount reported by the customer and the balance in the receivables subsidiary ledger. For example, payments from customers which are in transit on December 31st can create discrepancies, as can shipments to customers which are in route as of December 31st. Again, for this simulation assume that the amount reported by the customer is correct. Unfortunately, many customers do not respond to confirmation requests. Last year’s workpapers indicate that only 22 out of 44 customers responded to the confirmations. Confirmations should be scheduled early in the audit process to permit the mailing of a second round of confirmations. One week after mailing the first confirmation requests, you mail a second confirmation request to customers who have not yet responded. In the simulation this is done by clicking the “Second” button, which electronically mails confirmations to those customers who have not yet responded. Responses to the second confirmation request can be retrieved and observed using the green “Seconds” link. The process is the same as used for previous confirmations. The “Results” page of the workbook has a “2nd Balance” column to record the amounts confirmed by these customers. 5. Alternative Procedures In order to achieve the desired level of detection risk, every customer account in the sample must be audited. Alternative procedures are performed for customers who do not respond 8 to the confirmations. One alternative procedure might be to investigate subsequent cash receipts. When a customer pays its account balance in the first few days of January, it provides evidence that the account existed. However, subsequent cash receipts do not establish that the balance existed as of December 31st. In this exercise, the audit program in Table 3 instructs you to vouch from the accounts receivable subsidiary ledger to the invoice, bill of lading and customer’s purchase order. The purchase order will provide external evidence of an agreement between the customer and Charles Cabinets. The bill of lading will provide evidence that the goods were shipped and the revenue earned as of December 31st. The invoice will provide evidence that Charles Cabinets has billed the customer. Table 3 audit program for accounts receivable performed by workpa per da te reference Confirm accounts receivable determine appropriate sample size send first confirmation requests send second confirmation requests Perform alternate procedures for customers who do not respond to confirmations vouch account balance to invoice(s) vouch invoice to bill of lading vouch bill of lading to purchase order Evaluate results of confirmation and alternative procedures Conclude on accounts receivable Click the “Alternative” button in the simulation to retrieve file copies of the invoices, bills of lading and purchase orders for customers who have not responded to the confirmation requests. The green “Invs” link allows you to view these invoices in your web browser. You will agree the customer’s name, address, amount and date on the invoice with the information on the accounts receivable subsidiary ledger. The amount on the invoice should be entered in the 9 appropriate column of the “Results” page of the workbook. The green “BOLs” link enables viewing of bills of lading and the green “POs” link enables viewing of purchase orders. You will agree the information on these source documents with the accounts receivable subsidiary ledger and document your observations in the appropriate columns of the “Results” page of the workbook. 6. Evaluation of Results Auditing standards require auditors to project sample results to the population. The projected misstatement should be compared to tolerable error. Even when the projected misstatement is less than tolerable error, there may be an unacceptable level of risk if the difference between the projected misstatement and tolerable error is small (ASB 2010 AU Section 350.26). In such circumstances additional audit procedures are necessary to reduce the level of audit risk to the acceptable level. Although statistical sampling does not replace auditor judgment, it does provide a valuable tool to help evaluate the results of the confirmation procedures. On actual engagements, unexplained discrepancies would cause us to reconsider our assessment of control risk. In this exercise, control risk was previously assessed as low. Whenever substantive tests reveal a significant number of discrepancies or significantly large discrepancies, the effectiveness of internal controls would need to be reviewed, regardless of the results from previous tests of controls. However, an extensive discussion of internal controls is beyond the scope of this exercise. 10 7. Audit Documentation (Auditing Standards Board 2010 AU Section 339) AU Section 339 Audit Documentation (ASB 2010 AU Section 339.10) states the following: “.10 The auditor should prepare audit documentation that enables an experienced auditor, having no previous connection to the audit, to understand: a. The nature, timing, and extent of auditing procedures performed to comply with SASs and applicable legal and regulatory requirements; b. The results of the audit procedures performed and the audit evidence obtained; c. The conclusions reached on significant matters; and d. That the accounting records agree or reconcile with the audited financial statements or other audited information.” Auditing standards require documentation of the auditor who performed the work and the date the work was completed. Audit documentation must be in sufficient detail to allow auditors who are new to the engagement to understand the procedures performed in the prior audit and must also provide guidance on how to perform the procedure in the current engagement. DC&H policy requires that workpapers utilize the following template. Table 4 Workpaper (reference number) Client name Transaction cycle Class of transactions or Account Nature of test: performed by: date: analytical procedure, test of controls, test of details, substantive analytical procedure Objective: Assertion(s): Tolerable error: Either a dollar amount or a percentage of the account balance Procedure: Description of the procedure performed be in sufficient detail to allow auditors who are new to the engagement to understand the procedures performed in the prior audit and must also provide guidance on how to perform the procedure in the current engagement Conclusion: Evaluation of the results of the procedure or a conclusion regarding the account or class of transactions 11 8. Assignment Hands-On Training 8.1. General Instructions You are required to complete three workpapers similar to those provided from last year’s audit. The simulation uses macros. When Excel opens the file, you will probably need to enable macros. If your computer’s security settings are too restrictive, it may not allow you to enable the macros, in which case you will need to relax the security settings on the computer. The simulation uses four *.html templates to create certain files. You will need to download these templates and place them in the same folder as the spreadsheet. Instructions for the simulation are on the “ReadMe” tab at the bottom of the spreadsheet. You need to include the information from the “Results” tab of the spreadsheet in your workpapers. 8.2. Project Deliverables Use your sample results to perform a hypothesis test evaluating whether the actual balance is materially overstated, given the acceptable level of risk of incorrect acceptance. Tolerable error should be $290,814.44, which is 10 percent of the book value. The risk of incorrect acceptance is 20 percent, and the risk of incorrect rejection is also 20 percent. The hypothesis test needs to evaluate whether the actual balance is less than $2,617,330.00 ($2,908,144.44 - $290,814.44). Calculate the critical value for your test. Compare your sample results to the critical value and determine whether you will (1) accept, or (2) fail to accept, that the recorded book value is not materially overstated. Present your answers in the form of working papers, as required by auditing standards: “The auditor must prepare audit documentation in connection with each engagement in sufficient detail to provide a clear understanding of the work performed (including the 12 nature, timing, and extent, and results of audit procedures performed), the audit evidence obtained and its source, and the conclusions reached.” (ASB 2010 AU Section 339.03) Last year’s working papers are included to assist you in understanding the work to be performed. Use last year’s workpapers (Attachments 1 to 4) as a guide for this year1. There is no need to reinvent the wheel. However, don’t let yourself become mechanical as circumstances change from year to year. It is essential that you (1) understand why you are performing the procedures, (2) objectively evaluate the results your procedures generate, and (3) update the workpapers accordingly. 1 Two examples of workpaper 2 are included. In the first example, a hypothesis test is used to evaluate the sample results. In the alternate presentation of workpaper 2, the sample results are projected to the account balance and the projected overstatement is compared with the tolerable misstatement for accounts receivable. 13 Attachment 1: Workpaper 1 Charles Cabinets Revenue Collection cycle Accounts Receivable performed by: date: John 2/29/12 Nature of test: Test of details Objective: The objective of this procedure is to determine if the accounts receivable account are overstated. Assertion(s): Existence; Valuation and Allocation Tolerable error: For accounts receivable tolerable error has been set at 10% of the account balance Procedure: DC&H, LLP selected a random sample of 44 entries from the accounts receivable sub-ledger. On Feb. 10, 2012, a confirmation letter was sent to each customer in the sample. On Feb. 17, 2012 a second confirmation letter was sent to each customer in the sample who had not responded to the first letter. For each customer who did not respond to either confirmation, we vouched from the account balance on the schedule in Workpaper 3, which was selected from the accounts receivable sub-ledger, to the invoice, bill of lading and sales order. We agreed the date, customer name, address, PO number, and amount from the schedule with the invoice. We then agreed the date, customer name, address, and PO number on the invoice with the bill of lading. Finally, we agreed the date, customer name, address, PO number, and amount from the invoice with customer purchase order. Workpaper 2 shows how the appropriate sample size was calculated and the evaluation of the sample results. Workpaper 3 shows the sample results. Conclusion: Based on the sample results we are unable to conclude that accounts receivable are not materially overstated. More extensive substantive tests of details need to be performed to reduce the risk of incorrect acceptance to the desired level. 14 Attachment 2: Workpaper 2 Charles Cabinets Revenue Collection cycle Accounts Receivable performed by: date: John 2/29/12 Nature of test: Test of details Objective: The objective of this procedure is to determine if the accounts receivable account are overstated. Assertion(s): Existence; Valuation and Allocation Tolerable error: For accounts receivable tolerable error has been set at 10% of the account balance $237,198.36 ( 10% x $2,371.983.60 ) Procedure: Sample size calculation and evaluation of sample results Account balance Tolerable error Standard deviation $2,371,983.60 / N = 930 237,198.36 806.81 Ho: μ > 2,550.52 – 255.05 α = 0.30 Zα/2 = 1.04 β = 0.15 Zβ = 1.04 average = $2,550.52 = 255.05 tolerable error is 10% of the recorded balance risk of incorrect rejection risk of incorrect acceptance TE = Zβ* Sx/√n + Zα/2* Sx/√n 255.05 = 1.04*806.81/√n + 1.04*806.81/√n √n = 2.08*806.81 / 255.05 255.05 = 2.08*806.81/√n n = 43.29 => n = 44 Evaluation of Sample Results Critical Value = μ + Zβ* Sx/√n 2,550.52 + 1.04 * 1,390.09 / √ 44 2,768.47 We are unable to conclude that Accounts Receivable is not materially overstated because the sample mean of $2,425.56 is less than the $2,768.47 critical value. 15 Attachment 3: Workpaper 3 Charles Cabinets Revenue Collection cycle Accounts Receivable performed by: date: John 2/29/12 Nature of test: Test of details Objective: The objective of this procedure is to determine if the accounts receivable account are overstated. Assertion(s): Existence; Valuation and Allocation Tolerable error: For accounts receivable tolerable error has been set at 10% of the account balance Procedure: Results of confirmations and alternative procedures #1 part of the order was backordered and not shipped until Jan. 2012 #2 items were returned prior to 12/31/11 16 Attachment 4: Workpaper 2 alternate Charles Cabinets Revenue Collection cycle Accounts Receivable performed by: date: John 2/29/12 Nature of test: Test of details Objective: The objective of this procedure is to determine if the accounts receivable account are overstated. Assertion(s): Existence; Valuation and Allocation Tolerable error: For accounts receivable tolerable error has been set at 10% of the account balance $237,198.36 ( 10% x $2,371.983.60 ) Procedure: Sample size calculation and evaluation of sample results Account balance Tolerable error Standard deviation $2,371,983.60 / N = 930 237,198.36 806.81 Ho: μ > 2,550.52 – 255.05 α = 0.30 Zα/2 = 1.04 β = 0.15 Zβ = 1.04 average = $2,550.52 = 255.05 tolerable error is 10% of the recorded balance risk of incorrect rejection risk of incorrect acceptance TE = Zβ* Sx/√n + Zα/2* Sx/√n 255.05 = 1.04*806.81/√n + 1.04*806.81/√n √n = 2.08*806.81 / 255.05 255.05 = 2.08*806.81/√n n = 43.29 => n = 44 Projection of Sample Results to the population sample results sample n= mean 106,724.85 44 = 2,425.56 x 930 2,371,983.60 2,255,775.24 116,208.36 = Zβ * Sx /√n = 217.95 x 930 202,690.25 1.04 1,390.09 √44 N= 318,898.62 Book value projected balance projected overstatement allowance for sampling risk projected error plus allowance for sampling risk We are unable to conclude that Accounts Receivable is not materially overstated because the projected overstatement plus the allowance for sampling risk exceeds the $237,198.36 tolerable error. 17 TEACHING NOTE 1. Educational Objectives of the Project This accounts receivable confirmation simulation consists of 1,000 customer accounts. It creates a unique sample for each four-digit student ID. Using customer data, the simulation creates confirmations, responses to confirmations, invoices, bills of lading and purchase orders in the form of *.html files which are readable with most web browsers. The educational objectives of the project are: Allow students to complete the confirmation process Allow students to complete alternative procedures for non-respondents Provide tangible examples of audit evidence Require students to project sample results to the population Require students to form a conclusion from their audit procedures Familiarize students with workpapers 1.1. Availability The following materials are available at the following website: http://ar-simulation.com/. The exercise for distribution to students which is the first portion of this paper The complete exercise including the teaching note Copies of last year’s workpapers The Excel workbook which includes the simulation The four templates which must accompany the simulation Powerpoint introducing the assignment Confirmation responses for n=25 and ID 1234 A list of the discrepancies generated by the simulation 1.1.1. Last Year’s Workpapers Last year’s workpapers are included with the assignment for students to use as templates. We believe this is consistent with practice, where auditors typically have access to the prior year’s workpapers. 18 1.1.2. Excel Simulation and Templates The Excel simulation makes extensive use of macros, which must be enabled. The simulation calls on the following html templates to create document files: TemplatePO.html creates the file containing purchase orders; TemplateBOL.html creates the file containing bills of lading; TemplateINV.html creates the file containing invoices; and TemplateConf.html creates files containing confirmations and responses to confirmations. These template files must be in the same folder as the spreadsheet. The Excel simulation will not look to other folders for these templates. 1.1.3. PowerPoint files and Confirmation Responses (n=25, ID 1234) The PowerPoint file may be useful to explain the assignment to students. The confirmation responses for a sample of 25 using student ID 1234 are also available. The demonstration goes quickly if you already know the confirmation responses. This allows you to focus on your explanation of the process without searching for the amounts on each confirmation. The demonstration can go quickly for two reasons. First, the simulation auto-fills after you enter five responses. Second, you already know that the dollar amount on the purchase order and invoice always equals the book balance and there is a bill of lading supporting each invoice. 1.1.4. Discrepancies The simulation does not have a single solution. While the simulation always generates the same discrepancy for any given customer’s confirmation response, the discrepancies do not flow through to the alternative procedures. None of the purchase orders, bills of lading or invoices contain discrepancies. The purchase orders, bills of lading and invoices are consistent with the 19 recorded book balances. This project design results in a unique solution for each four-digit student ID. 2. Project Efficacy Questionnaire During the spring term of 2011, we asked students to respond to the following questions using SurveyMonkey. Seventy-five out of the 91 students enrolled in our classes responded to the survey. The table following the questions shows the responses to the questions. 1. The Accounts Receivable Project provided a more realistic auditing experience than is typical for accounting classes. 2. I enjoyed completing the AR Project. 3. The AR Project complimented the text and enhanced my understanding of the accounts receivable confirmation process. 4. The AR Project complimented the text and enhanced my understanding of what it means to perform alternate procedures for accounts receivable. 5. The AR Project helped me better understand the nature of audit evidence. 6. The AR Project helped clarify the relationship between audit evidence and auditors’ conclusions regarding the fair presentation of account balances. 7. I am more familiar with auditors’ work papers because of the AR Project. 8. The AR Project helped me better understand how the audit risk model is used. 9. I would learn more and be better prepared for an auditing career if I were exposed to more projects like the AR Project. Responses to Questionnaire number of Strongly repsonses Disagree Strongly Disagree Neutral Agree Agree 00 10 10 20 20 30 30 40 40 50 50 60 60 70 70 80 80 90 90 100 100 average (1) 74 2 0 0 0 0 1 2 3 35 13 18 83.1 (1) (2) 75 2 0 4 1 0 3 6 9 24 16 10 74.8 (2) (3) 75 2 0 0 1 0 1 3 7 27 14 20 82.3 (3) (4) 75 1 0 2 0 1 0 4 10 23 13 21 81.7 (4) (5) 74 1 0 2 1 1 3 3 5 25 16 17 80.1 (5) (6) 75 0 0 4 0 0 8 3 12 23 12 13 76.3 (6) (7) 75 1 1 2 1 0 3 3 2 23 12 27 82.3 (7) (8) 75 1 1 4 3 5 6 5 21 11 10 8 67.1 (8) (9) 75 1 1 0 0 2 3 3 3 11 13 38 86.4 (9) 20 3. Implementation Guidance Many steps in the accounts receivable process are mechanical. The sample size may be given or determined from a matrix, while the preparation and mailing of confirmations is mechanical. Agreeing information from a confirmation or a bill of lading with a subledger can become rote, if not mechanical. Although it may encourage mechanical behavior, we understand that auditors still typically use last year’s workpapers as a template for the current year. Consequently, we warn students that circumstances change from year to year and we encourage them to be vigilant and not blindly follow last year’s workpapers. Prior to using this exercise, we found that confirmation was a word students memorized. They had no basis to reflect on the process. A difference between a recorded balance and confirmation, or any audit difference for that matter, was a vague concept they tried to grasp but had difficulty visualizing. Completion of this exercise enables students to better reflect on the confirmation process and to visualize a discrepancy. Our preference is to assign this project in an auditing course early in the term, after we cover statistical sampling but before we cover risk, internal control, audit documentation, or accounts receivable. Although it may seem ideal to perform the simulation while reading the corresponding material in the text, we have not found this to be the case. By completing this exercise early in the term, students have a referent for many subsequent discussions and are better prepared to engage in the discussions when topics are covered later in the term. We continually refer back to this exercise when we discuss risk, internal control, evidence, audit documentation, vouching and confirmations. We recommend that the simulation be implemented in a manner that minimizes both class time and student time. We spend approximately 20 minutes introducing the assignment in 21 class and students report that the simulation requires an hour or less to complete. 2 It may seem logical to expand the assignment to cover multiple topics such as statistical sampling, accounts receivable, vouching and audit documentation. In fact, we have taught each of those topics as part of the assignment. However, we have enjoyed the most success by covering these topics as they appear on the syllabus and referring back to the simulation as we discuss each of these topics. We are not terribly concerned with how well students complete the assignment. The benefits of the simulation continue to accrue by providing a referent as we discuss each of the topics in subsequent lectures. Having previously covered statistical sampling, we use the simulation to reinforce audit sampling. However, you will note that the objectives relate to the confirmation of receivables, and not statistics. You can assign this simulation prior to coverage of statistical sampling. If so, simply provide students with the required sample size for the simulation which is 78. We email the assignment to students and attach the first portion of this manuscript, the Excel simulation and the four templates used by the Excel simulation. Students are instructed to read the assignment and they are reminded that all four templates must be in the same directory as the Excel simulation. Students can also download the components of the simulation from our website. However, the simulation cannot be run from a website. It must be downloaded to the computer where the Excel program resides. As previously stated, we have already covered statistical sampling and require our students to calculate the sample size necessary to limit the risk of incorrect acceptance and the risk of incorrect rejection. The simulation provides a framework to discuss the risks of incorrect acceptance and incorrect rejection. 2 The PowerPoint slides we use to introduce the assignment are available at http://ar-simulation.com/ 22 In order to add a time dimension to the simulation, you may have students complete the simulation in stages. Students complete the first four steps of the simulation, which takes them through the creation, printing and mailing of the confirmation requests. In order to create a time lag between the mailing of the confirmations and responses, we wait until the next class to assign step 5 of the deliverables. In this step students review the responses and record the confirmed amounts on the results page of the workbook. We create another time lag by waiting until the next class to assign step 6. In this step students review and record the responses to the second confirmations. Finally, during the next class we assign step 7 which covers alternative procedures. Depending on the order in which you cover topics, it may be necessary to explain the vouching procedure used to perform alternative procedures. We spend minimal time explaining the mechanics of the alternative procedures because our students have previously completed the “Vouch and Trace” project where they vouched entries from a sales journal to invoices, shipping documents and sales orders.3 This allows us to focus our discussion on risk. We remind students that the audit planning document has already established the acceptable level of risk and the tolerable misstatement for accounts receivable. In order to achieve the desired level of risk, every account selected in the sample must be audited, not just the accounts of customers who responded to the confirmation. Increasing the required sample size does not increase the time required to complete the project. The results page of the Excel workbook only requires students to complete each procedure five times before the simulation auto-fills the remainder of the entries. For example, once five entries are made in the column for 1st Balances, the simulation auto-fills the remaining 3 Miller, C. and Savage, A. (2009). “Vouch or Trace: A Revenue Recognition Audit Simulation.” Issues in Accounting Education 24(1): 93-103. The vouch and trace project covers revenues and the use of vouching to audit transactions. 23 entries in that column. The same is true for the 2nd Balance, PO Amt, BOL and Inv Amt columns. We did not want students spending significant time mechanically inputting data. We believe five entries allow students to reflect on the documentation of evidence without imposing a significant time burden. We find it necessary to explain the role of the prior year’s workpapers.4 Many students are reluctant to merely update last year’s workpapers. They are concerned that this might be cheating or plagiarism. Students use their sample results to conclude whether or not accounts receivable are materially overstated. The simulation helps students see how audit procedures generate evidence, which become the basis for their opinion. Auditing standards state that “The auditor should project misstatement results of the sample to the items from which the sample was selected. … This total projected misstatement should be compared with the tolerable misstatement for the account balance or class of transactions, and appropriate consideration should be given to sampling risk (ASB 2010 AU Section 350.26).” Having covered statistical sampling, we require students use statistics to evaluate their results and determine if the allowance for sampling risk is adequate. Although standards do not require that auditors evaluate sample results using statistics, statistics can help illustrate the relationship between the sample results and risk. 4 The spreadsheet in Workpaper 3 can be replicated by having the simulation select a sample of 44 using student ID 1123. 24 3.1. Extension of the Project For instructors who wish to extend the project for more coverage and/or post-project discussions of statistics, audit risk, internal control, accounts receivable, audit sampling and audit documentation, we provide additional implementation guidance. 3.1.1. Covering Statistics Concurrently with the Simulation Appendix A is available as a handout if you prefer to cover statistical sampling concurrently with the accounts receivable simulation. Appendix A includes a sampling plan based on the AICPA Audit Guide: Audit Sampling. It also addresses each of the issues raised in the sampling plan. Appendix B is provided as a handout to supplement the statistical calculation in the Workpaper 2. The first portion of Appendix B provides a legend for the terms used in the calculation of the required sample size. The second portion of Appendix B discusses the use of a hypothesis test to evaluate the sample results. The final section of Appendix B evaluates the sample results incorporating the terminology used in auditing standards. The projected error is compared with the tolerable error to determine if the allowance for sampling risk is adequate. 3.1.2. Subsequent Discussions of Audit Risk When teaching the audit risk model, we refer to the simulation. Students are asked (1) to consider the effect on the required sample size if we took a primarily substantive approach, and (2) to consider the effect on the required sample size if we determined internal controls were moderately effective rather than very effective. 3.1.3. Subsequent Discussions of Internal Control Control risk was assessed as low in the simulation because internal controls had previously been found to be very effective. However, the simulation is seeded with a significant number of discrepancies, making it likely that all students will encounter discrepancies. This 25 leads to discussions about the possible need to reassess control risk. The question we pose to students is: “Why are we finding a significant number of discrepancies if the controls are very effective?” We discuss the possible causes of the observed discrepancies. Because discrepancies were only detected on returned confirmations and none were detected by alternative procedures, we can conclude that these are not sales to fictitious customers. Workpaper 3 illustrates two possible discrepancies. The first is an order which was billed even though a portion of the order remained on backorder as of Dec. 31 (see footnote #1 of workpaper 3). In their understanding of the billing process, auditors need to ensure that the client only bills for goods that have been shipped. Otherwise, the auditor may need to extend procedures to include items on backorder. The second discrepancy involves a sales return which was in transit as of year-end (footnote #2 of workpaper 3). The audit will occur after year-end, enabling the auditor to exam returned merchandise. Other examples of overstatements in the simulation could result from: (1) billing customers for sales that did not occur, or at least the sales had not occurred by December 31st; (2) billing customers for more products than were actually shipped or billing at higher prices than those on the approved price list; (3) shipping the customer more goods than ordered; and (4) shipping products to customers who never ordered the products. Next we discuss controls which might prevent such errors, controls such as having someone in the shipping department verify that a customer purchase order exists for every shipment and that the quantity shipped agrees with the quantity ordered. In the billing department, controls should be in place to verify that a bill of lading exists before an invoice is created and also to agree the prices on the invoice with the official price list. 26 3.1.4. Subsequent Discussions of Accounts Receivable By the time we actually cover accounts receivable, students have a better background in auditing. We discuss conditions which might explain discrepancies between the recorded balance and amount confirmed by the customer. The customer may have made a payment which is still in transit to the client on December 31st. The audit client may have recorded a receivable for goods shipped but still in transit to the customer. In an actual audit, auditors would not automatically accept the customer’s response as correct. For discrepancies, the auditor examines source documents to determine which balance is correct. One method of auditing discrepancies is to vouch from the accounts receivable subsidiary ledger to the invoice, bill of lading and purchase order. This is an opportunity to reinforce that vouching to the invoice provides evidence that the audit client billed the customer for the recorded amount; vouching to the bill of lading provides evidence the goods were actually shipped to the customer; and vouching to the purchase order provides evidence the customer actually ordered the goods. This naturally leads to discussion of alternative procedures. Rather than vouching to the supporting documents, some auditors might prefer auditing subsequent cash receipts as an alternative procedure. 3.1.5. Subsequent Discussions of Audit Sampling Although the text we use covers audit sampling after the chapter on accounts receivable confirmations, we cover audit sampling earlier in the class. We do this to provide a framework to discuss the relationship between the quantity of evidence and the risk of an incorrect conclusion. When the text covers account receivable confirmations, we refer to the simulation and again discuss the evaluation of sample results. As this comes toward the end of the audit, we discuss 27 the possible implications if we had detected a material misstatement in accounts receivable. Will it be necessary to reassess control risk and extend our substantive tests? Although Charles Cabinets is a private company, we discuss the implications the detection of a material misstatement would have if we were performing an integrated audit of a public company. The existence of a material misstatement would likely result in an adverse opinion regarding internal controls, even if the client corrects the misstatement. 3.1.6. Subsequent Discussions of Audit Documentation Students appear to have difficulty grasping the necessity for good audit documentation. By including workpapers in this assignment, our students obtain some basic familiarity with the nature of workpapers. We frame discussions regarding audit documentation in terms of litigation. If it becomes necessary for the auditors to defend their position to a jury, we stress the need for persuasive documentation to prove that they actually did perform the procedures and did obtain sufficient evidence to support their opinion. You might also refer to problems noted in PCAOB inspection reports. A word search of one recent inspection report revealed seven instances where the PCAOB found no evidence in the audit documentation that the audit firm had performed a required procedure. 3.1.7. Subsequent Discussions of Fraud If you want to discuss fraud, consider having students read the article “Detecting Circular Cash Flow” in the Journal of Accountancy (Monhemius and Durkin 2009). This article illustrates potential weaknesses of confirmation procedures to detect sophisticated schemes used to inflate receivables. Companies such as the one in the article use receivables as collateral to obtain loans and may have incentive to inflate receivables. In addition, the article points out various red flags that may indicate fraudulent activities. This article helps students understand the need to exercise judgment in the confirmation process. 28 4. Grading Grades in this course are based on 550 total points of which students can earn up to 25 points for successfully completing this project. The rubric in Appendix C was developed from shortcomings we observed in student workpapers. The rubric is segmented into five equally weighted components: timeliness, the material at the top of the workpapers which is common to all three workpapers, and the procedural material on Workpaper 1, Workpaper 2, and Workpaper 3. Students can earn up to 5 points for each segment. We also encourage students to think about what they are doing as they complete the project because many of the procedures and concepts incorporated in the project will be included on the exams. 29 Appendix A Sampling Plan (AICPA Audit Guide: Audit Sampling) The AICPA Audit Guide for audit sampling suggests the following questions be addressed when planning a sampling procedure. “2.30 The following questions apply to planning any audit sampling procedure, whether it is nonstatistical or statistical: • What is the test objective and relevant assertion? (What does the auditor want to learn or be able to infer about the population? What assertions are being tested?) • What is the auditor looking for in the sample? (How is a misstatement or deviation defined?) • What is to be sampled? (How is the population defined?) • How is the population to be sampled? (What is the sampling plan, what is the sampling unit, and what is the method of selection?) • How much is to be sampled? (What is the sample size?) • What do the results mean? (How are the sample results evaluated and interpreted?)” What is the test objective? What assertions are being tested? How is a misstatement defined? How is the population defined? What is the sampling unit? What is the sample size? What is the method of selection? How are the sample results (to be) evaluated? What do the results mean? Determine whether the accounts receivable balance is materially overstated. Materiality is defined as 10% of the accounts receivable balance. Existence; valuation and allocation When the audited value for a customer’s account as determined by confirmation or alternative procedures does not agree with the amount recorded in the accounts receivable subsidiary ledger All customer accounts included in the accounts receivable subledger, which has been agreed to the general ledger balance Customer accounts in the accounts receivable subsidiary ledger Must be determined Random sampling Send positive confirmations to selected customers Perform alternative procedures for customers who do not respond to confirmation requests Perform a hypothesis test to statistically generalize the sample results to the accounts receivable balance. Conclude on whether the accounts receivable balance is materially overstated 30 Appendix B Sample Size calculation using the information in sample workpaper 2 The book value, standard deviation, tolerable error and risk factors in the sample workpapers differ from those in the current assignment. $2,371,983.60 $237,198.36 $2,134,785.24 806.81 930 0.30 0.15 BV book value TE tolerable error BV - TE Sx standard deviation items in population risk of incorrect rejection risk of incorrect acceptance ̅̅̅̅ = Zβ* Sx/√n + Zα/2* Sx/√n 𝑇𝐸 (Zα √𝒏 = [ √𝒏 = 𝟔. 𝟓𝟖𝟎 ⁄2 + Zβ ) × sx ̅̅̅̅ 𝑇𝐸 $2,550.52 $255.05 $2,295.47 Sx/√n 1.04 1.04 ̅̅̅̅ average book value 𝐵𝑉 ̅̅̅̅ 𝑇𝐸 average tolerable error ̅̅̅̅ − 𝑇𝐸 ̅̅̅̅ hypothetical mean μ = 𝐵𝑉 std error of the means Zα/2 number of std dev Zβ number of std dev rearrange the equation to solve for n ] √𝒏 = [ (1.04 + 1.04 ) × $806.81 ] $255.05 𝒏 = 𝟒𝟑. 𝟐 Evaluation of Sample Results using a Hypothesis Test During the planning phase, the standard deviation of the recorded values would be the only value available to calculate the required sample size. If the sample results indicate the standard deviation might actually be greater, we should consider using the standard deviation of the sample results. 1.04 x 209.56 $ 217.95 + 2,295.47 $ 2,513.42 Zβ number of std dev Sx/√n std error of the means allowance for sampling risk μ hypothetical mean Critical value 0.15 1,390.09 44 $ 2,425.57 risk of incorrect acceptance Sx std deviation of sample sample size sample mean The hypothetical mean is the book value less tolerable error. We need evidence indicating there is a low probability the true value is less than the hypothetical mean. The critical value is calculated by adding an allowance for sampling risk to the hypothetical mean. In this example the $2,425 sample mean is less than the critical value indicating there is unacceptable risk that accounts receivable might be is overstated. This indicates that we need to obtain more evidence before we can conclude on the account. 31 If our sample mean exceeded the critical value we could conclude that the account is not materially overstated. There is less than a 15% probability of selecting a sample with a $2,513 mean from population with a $2,134,785 balance. Evaluation by Comparing the Projected Sample Results to Tolerable Error $ 2,550.52 ̅̅̅̅ 𝐵𝑉 ave. book value - 2,425.57 sample average $ 124.95 average error 0.15 1.04 x 209.56 $ 217.95 risk of incorrect acceptance Zβ number of std dev Sx/√n std error of the means allowance for sampling risk $ 2,371,983.60 - 2,255,775.24 $ 116,208.36 237,198.36 $ 120,990.00 202,693.50 book value projected value projected error TE tolerable error allowance for sample risk desired allowance for sample risk Alternatively, we reach the same conclusion by projecting the estimated error from our sample results to population. Although the $116,208 projected error is less than tolerable error, accounting standards require us to add an allowance for sampling risk. The difference between the projected error and tolerable error results in a $120,990 allowance for sampling risk. A $202,693 allowance for sampling risk would be necessary to achieve 15% Beta risk. 32 Appendix C rubric for the simulation workpapers the risk factors, tolerable error and standard deviation in the assignment differ from those in the sample workpapers 1 Workpapers are turned in on time 2 Each workpaper includes the appropriate common information: client name, transaction cycle, account, nature of test, objective, s assertions and tolerable error s sign-off and date should be updated for each work paper 3 Workpaper 3 s s a dollar amount from a confirmation response or the results of the alternative procedures, for each item in the sample sum, mean and standard deviation calculated from sample results 4 Workpaper 2 s s s s the required sample size calculation reflect the current year’s population and risk parameters: Zβ = 0.84, Zα/2 = 1.28, population size = 1,000, tolerable error $290,814.44, and standard deviation = $1,204.33 critical value is calculated from sample results in workpaper 3 the conclusion is appropriate for the critical value 5 Workpaper 1 s sample size and dates in the procedure description are appropriate s the conclusion is consistent with workpaper 2 33 References AICPA Audit Sampling Guide Task Force. 2008 AICPA. Audit Guide: Audit Sampling, 2.30. New York: American Institute of Certified Public Accountants. Auditing Standards Board. 2010a. Codification of Statements on Auditing Standards Generally Accepted Auditing Standards. New York: American Institute of Certified Public Accountants. Miller, C., and Savage, A. 2009. Vouch or Trace: A Revenue Recognition Audit Simulation. Issues in Accounting Education 24(1): 93-103. Monhemius, J., and Durkin, K. 2009. Detecting Circular Cash Flow. Journal of Accountancy Vol. 208(6): 23-30. Retrieved from http://www.journalofaccountancy.com/Issues/2009/Dec/20091793.htm Securities and Exchange Commission, 1940. Accounting Standards Release No. 19. Dec, 5. In the Matter of McKesson & Robbins, Section I, Summary of Findings and Conclusions. Retrieved from http://www.sechistorical.org/museum/papers/1940/ 34