Statistical Sampling and Risk Analysis in Auditing

Peter Jones

• Publisher: Routledge
• Year: 2017
• ISBN:
  9781351898003
• Edition: 1
• Topic: Negocios y empresa
• Subtopic: Gestión

This book is aimed at those with responsibilities for audit, risk and control - auditors of course - but also finance directors, audit committee members, project accountants, systems designers and other professionals too. Working under pressure, these people often need to take account of theory and best practice but strike a balance with the practical demands of their workplace. This book's practical emphasis on meeting the ever-changing needs of clients and auditees will benefit a wide audience by helping readers to: ¢ select a suitable, practical sampling approach ¢ appreciate the statistical implications ¢ evaluate the results of audit testing ¢ take account of risk and control evaluation in targeting valuable audit resources. It does this by laying out the principles behind a concept and then grounding them in 'real life' cases for the reader to work through. These are accompanied by suggested solutions which, while not definitive answers, do provide valuable advice and guidance. Finally the range of appendices, including a complete copy of the statement of auditing standards, SAS 430, make this book an essential resource for everyone concerned about modern auditing.

1	Introduction

The people this book aims to help

This book is for people responsible for managing or undertaking audits. Both internal and external audits can benefit from statistical sampling and both have been borne in mind throughout. A reasonable level of audit experience is assumed, though little or no experience or knowledge of statistics or statistical sampling methods is expected. The book may well be useful to directors, members of audit committees and others who, while they might not be directly involved in auditing, are concerned about various strategic aspects such as the cost of the audit, the amount of audit testing and the overall coverage of important systems.

Although the book makes no attempt to cover any academic or professional syllabus, students of accounting, auditing and others whose courses include auditing will also find the book useful.

We have tried to keep the text to a length suitable for people who are more interested in how to start applying statistical sampling in practical, financial auditing than in knowing every possible sampling technique, every application and the derivation of every formula.

The help on offer

Our main concern is to help professional auditors decide when and how best to use statistical sampling during their ordinary work. Although this is not a book about statistics, and in most situations auditors will not need to become statisticians, some explanation of statistical concepts and theory is unavoidable, although this has been kept as far as possible to a useful minimum. We have sometimes pointed out when further advice might be needed, though readers must decide for themselves, within the limitless variation of possible practical situations, exactly when to seek specialist statistical advice.

Despite the enormous and ever-expanding range of audits faced in modern organizations, certain widely-applicable sampling approaches can be discerned and it is upon these that we concentrate. Statistical sampling owes its steadily increasing popularity among auditors to many sources, but four stand out:

1  Existing publications, some of which are listed in Appendix 6.

2  Reasonably well-established practice among external auditing firms and the National Audit Office.

3  Increasing use of, and training in, statistical sampling among internal auditors.

4  Auditing standards and guidance from professional bodies.

This book has not attempted to follow any particular one of the sources just mentioned above. Rather, we have tried to utilize the best suggestions and influences from each, plus any other ideas and practical points that help auditors plan and manage their work to achieve typical audit objectives.

For example, some typical questions that arise and for which help is often sought include the following, though in no particular order or priority:

•  How many items should we test?

•  What should we do if we are not certain how many items we are sampling from?

•  Do we select samples differently for compliance and substantive tests?

•  What should we take into account during our audit planning?

•  Do we need to test items in a different way, or alter our audit programmes?

•  Can we take into account our risk assessments/models?

•  How do we extrapolate results from samples to the system/account being audited?

•  What are the most convenient formulas?

•  What sort of standard working papers might we use?

•  How confident of our conclusions can we be?

•  Should we involve our clients in our approach?

•  Can we report our statistical results in a straightforward manner?

•  Is this going to cost a lot of money for training, software and such?

These and many more practical queries will arise when auditors introduce, or expand their use of, statistical sampling. While we can not promise to offer the perfect solution to every conceivable problem, our emphasis on the practical difficulties and the most commonly used approaches should help you to work out a solution for your particular audits. The basic principles and direction of thought should be clear to you as the book progresses, and these will be illustrated by case studies to help clarify detailed techniques in realistic situations.

Reading through and getting the most from this book

Chapters 2,3 and 4 are, unavoidably, rather discursive: Chapter 2 sets statistical sampling in a wider audit sampling context for the benefit of the complete newcomer; Chapter 3 looks at the advantages and disadvantages; and Chapter 4 looks at some of the basic theory. If you are already reasonably familiar with these aspects, particularly the theory, you may find it worthwhile to delve straight in to Chapter 5 on attribute sampling.

If, as is often the case whenever one approaches new ideas, some of the principles and techniques seem a little complex on first reading, it may help to work through the cases before rereading the theory. In fact, reading back and forward between the explanatory text and the cases is often the best approach. Some people find it easier to read and reflect upon the former, while others find it easier to work through the latter, but both are recommended.

The case studies are accompanied by suggested solutions. These suggestions are not ‘cast in tablets of stone’. It may well be that you have greater or fewer resources than those assumed in the case and may wish to offer more or less assurance to your auditees. It is often possible to rework a case with your assumptions built in.

Do not be too concerned if at first you feel that some of the more complex aspects could not realistically be adopted for your range of audits. Increasing familiarity with statistical sampling techniques is like increasing familiarity with many other things – it’s surprising what hidden benefits can arise. Perhaps the secret of successfully applying these techniques in your complex professional environment is to proceed cautiously – an incremental approach, gaining confidence yourselves and gradually bringing others, including auditees, on board. A good way to begin gaining experience is often to use attribute sampling when testing internal controls, as we will see in due course. These points are reconsidered in Chapter 10.

Recent developments

The development of statistical sampling techniques in auditing has an interesting history, but one beyond the scope of this book. We have, however, provided a brief table of some of the important landmarks for the benefit of those who feel more comfortable with an historical context. (A list of abbreviations is given in Appendix 3.)

Table 1.1 Recent developments

•  1960s – Professional guidance papers on this topic issued by the AICPA in the USA.

•  1970 – IMTA (now CIPFA) issued a small volume titled Statistical Sampling in Auditing.

•  During the late 1970s and the 1980s a number of published works appeared on both sides of the Atlantic, written mainly by learned academics and professional bodies such as the ICAEW and the AICPA, some of which are mentioned in Appendix 6. During this time the use of statistical sampling appears to have increased among the major accounting firms and been taken on board by the National Audit Office.

•  1981 – AICPA in the USA issued statement on auditing standard SAS 39 entitled Audit Sampling.

•  1987 – The APC (now APB) issued an exposure draft guideline entitled Audit Sampling. This exposure draft was accompanied by an ‘Audit Brief’ offering relatively detailed guidance on techniques and examples of possible applications. Both these documents remained in circulation for consultation until 1995.

•  During the early 1990s a survey done by CIPFA indicated a growing interest and use of statistical sampling methods by internal audit departments in the UK.

•  1995 – The APB in the UK issued its long-awaited guideline SAS 430, entitled Audit Sampling, setting the standard expected of all CCAB auditors during ‘… any audit using sampling ….’ This guideline comments upon both statistical and non-statistical approaches.

•  1996 (December) – The IFA’s IAPC issued an exposure draft entitled Audit Sampling and Other Selective Testing Procedures.

Note: USA and UK SASs are not a single series of standards.

Conclusion

This book is about the value of statistical sampling to the practical side of auditing. It is written for auditors by an experienced auditor, who is familiar with statistical sampling in both external and internal audit situations.

A range of situations and questions will be addressed, but you should not worry at this stage if some of the more complicated scenarios seem to lack immediate relevance to your own work. It is often more efficient to build up your approach to sampling as you gain experience, rather than try to introduce across-the-board changes to all your audits.

In relatively recent times interest in this topic has grown rapidly throughout the profession. Standards have been promulgated by various professional bodies that imply all auditors should be aware of statistical sampling techniques and know when these should be used.

2	Statistical and non-statistical approaches

In this chapter we consider the position of statistical sampling in relation to the wider question of how to go about selecting items to test during an audit.

Is there really a lot of difference between statistical and non-statistical?

At first it might seem so. Opinions are often divided between those who use statistical sampling and those who reject it. In 1991 a survey of a wide variety of organizations undertaken on behalf of CIPFA indicated that there was a growing interest in the use of statistical sampling. Any distinct split between acceptance and rejection was less pronounced than had at first been expected.

It is often convenient to think in terms of a broad division between statistical and non-statistical approaches, which sometimes gives rise to people reserving the term ‘sampling’ exclusively for statistical sampling and referring to non-statistical approaches as ‘selection’. A sharp division between ‘statistical’ and ‘non-statistical’ is understandable in workplace conversation, particularly when talking broadly about, say, your audit strategy for the next three years or your approach to testing a particular system or account. But when we get down to the level of particular audits and test programmes, it is often more helpful to bear in mind that we are usually talking about a range of possible approaches to sampling and that both statistical and non-statistical approaches can actually have a lot in common.

Professional judgment lies behind and underpins all audit sampling. There is a myth perpetuated in some circles that sta...