An exploratory case study at the corporate client: The influence of
factors on the accuracy of financial forecasts
Elif Varol Nieuwe Achtergracht 113hs 1018 WS Amsterdam e-mail: e.varol@student.rug.nl tel. nr.: 06 11081761 Student number: 1614398 Amsterdam, March 2013 University of Groningen Faculty of Economics and Business
MSc Business Administration
Specialization in Organizational and Management Control Supervisor: prof. dr. ir. P.M.G. van Veen-Dirks
PREFACE
This master thesis completes my study at the University of Groningen. In 2006, I started at the faculty of economics and business with my bachelor Business Administration. After finishing this bachelor I started with the master Organizational and Management Control. After completing this master thesis I will have to pass a final oral examination to complete my program and start a new challenge.
I was not able to do this research without the intensive support of some people. First of all, I would like to thank my supervisor of the University of Groningen, Paula van Veen-Dirks for her permanent interest, helpful feedback and confidence during these months. Without her useful insights I was not able to write this master thesis.
In addition, I would like to thank my supervisors from KPMG, Maurits van der Wijk and Arno Bastiaans for providing me with the great opportunity to conduct my thesis at KPMG. They have supported me during the whole process, and provided me with helpful feedback to improve my thesis. I have learned a lot about forecasting during these six months and had a very great time there.
Furthermore, it was not possible to write this master thesis without the experiences and useful insights from five companies. I would like to thank the respondents from the participating companies for their cooperation, openness and time to participate in this research. In addition, I would like to thank André Jansen, Jordi Wardenburg and Frans den Toom for conducting the expert interviews and sharing their valuable opinions. Further, I want to thank Judith Hoogeboom for her support and reviews which helped me to improve this thesis.
Finally, I want to thank my parents, my sister and my boyfriend for their support and confidence in the realization of my study and master thesis.
Elif Varol
SUMMARY
Many companies struggle with the accuracy of their financial forecasts. These organizations need to have confidence in the data that is used to report financial results, set future performance expectations, evaluate risk and understand customer interests. Accurate forecasts are crucial since inaccurate forecasts cost organizations money, stakeholders lose confidence in relationships with the management and allocations of resources for investments are not optimal. The solution for inaccurate financial forecasts is sought in eliminating traditional budgeting, using more flexible budgeting and forecasting techniques. In order to completely understand the discipline of financial forecasting, the factors that have an effect on the accuracy of financial forecasts are examined in this research study. After identifying the possible factors that have an influence on the accuracy, a comprehensive framework for forecast accuracy is provided. This framework serves as a means of helping organizations to create and improve stakeholder confidence by providing more accurate forecasting. In addition, managers are able to improve their resource allocation and create an agile organization that can easily adapt to changing business conditions.
A qualitative research at five different companies in the ‘corporate client’ segment, has tested and validated the factors that are based on the insights from the literature and expert interviews. The interviews with employees of participating companies show the importance of these factors that influence the accuracy of financial forecasts. The main purpose of these interviews is to confirm the importance of these factors that are obtained from the literature review and expert interviews. Subsequently, by answering the interview questions and valuing the factors, participating companies understand which factors need more attention, and require improvements to improve the accuracy of forecasts. In addition, participating companies are compared to assign the importance of certain factors to characteristics of companies and markets. The comparison between participating companies provides additional meaning regarding the effects obtained in this research.
TABLE OF CONTENTS
PREFACE………...………..………...………...….….……...…...2 SUMMARY………...………..…..………...…..3 TABLE OF CONTENTS………...………..……...…….…...4 1. INTRODUCTION...………....………..………..…...6 1.1 KPMG International...………...……….………...………...……...……..…...6 1.2 Management problem………...………….………....…….…...…...7 1.3 Research objective……….…….…….8 1.3.1 Research question………..………….…...8 1.3.2 Research methodology………...…….…..…...9 1.3.3 Research Relevance………10 1.4 Conclusion……….10 2. GENERAL LITERATURE……….……….11 2.1 Introduction………....11 2.2 Financial forecasting……….……..….…..11 2.3 Budgeting………..………13 2.4 Rolling Forecasting………...…....……142.5 Rolling forecasting versus Budgeting………...………...15
2.6 Conclusion………..…...16
3.
THEORETICAL BACKGROUND OF FACTORS……..……….…....
173.1 Introduction………....17
3.2 Perspectives of forecasting………..………...…...17
3.3 Process………..….18
3.4 Information Technology………....20
3.5 Model……….………....23
3.6 People & Organization……….………..25
3.8 Design of forecast accuracy framework……….……….………..30
3.9 Preliminary forecast accuracy framework………...………..………....…32
3.10 Relationships between dimensions ………..………..…….33
3.11 Conclusion…..……….………....34
4. RESEARCH METHODS………..35
4.1 Introduction………....35
4.2 Case study………..35
4.3 Type of case study……….….36
4.4 Number of cases.………....37 4.5 Selecting cases…..……….….…...38 4.6 Data collection………...39 4.7 Expert interviews………...…….…...40 4.8 Conclusion……….43 5. RESULTS………...………44 5.1 Introduction………....44 5.2 Research findings………..………...……….….44
5.2.1 Explanations for significant differences……….46
5.3 Ranking of factors……….……….…49
5.3.1 Highest ranked factors……….…..…...54
5.4 Relationships between dimensions………....59
5.5 Conclusion……….60
6. CONCLUSION...61
REFERENCES…………..………..……….…..64
APPENDIX……….…69
1. INTRODUCTION
Forecasting is the process that organizations use to be prepared through forecast of future outcome regarding various business decisions (Vadasz, 2005). Managers use forecasts because of the uncertainty of future investments of the organization (Armstrong, 2001). Currently managers fail to provide accurate forecasts due to the continuing financial crisis and, therefore, uncertainty in the market (Ryan, 2008). Ryan (2008) states that almost 70% of respondents were unable to forecast accurately and to see more than one quarter ahead. This is supported by another study covering Europe, US and Asia where the ability to forecast results was found the number one internal concern for CFO’s (Karaian, 2009). Accurate forecasting is not a recent problem, good forecasting has always been a major concern for operational management. Reason for that is that forecasts are at the heart of the supply chain process. Forecasting is further important for service organizations to ensure that they are able to consistently service customer demand. Furthermore, forecasting and the ability to understand risk, which is how far your forecasts might be inaccurate, is important in order to avoid catastrophes, to manage relationships with stakeholders and customers and to generate cash (Morlidge and Player, 2011). Fritz Roemer, head of the Hackett Groups’ Enterprise Performance Management Practice said: ‘Forecasting is broken in many,
many companies that we see, and this is largely the result of ignorance. Nine of ten companies simply do not really have a proper understanding of what forecasting is and how to do it well’ (Morlidge and
Player, 2011). By improving organization’s understanding of forecasting and their ability to anticipate, organizations are better prepared for future outcomes. As a result of better forecasting, managers deliver more reliable performances and are in better shape to exploit opportunities and avoid potential disasters. This research study is focused on the ‘accuracy of financial forecasts’, which is investigated by finding possible factors that have an influence on the accuracy of financial forecasts. Also, aspects concerning budgeting and rolling forecasting in relation to financial forecasts are discussed.
1.1 KPMG International
countries (www.kpmg.nl). Clients in the forecasting field have to deal with the accuracy of financial forecasts. They need to have confidence in data that is used to report financial results, set future performance expectations, evaluate risk and understand customer interests (www.kpmg.nl). The solution for inaccurate financial results is sought in eliminating traditional budgeting, using more flexible budgeting and forecasting techniques. In order to completely understand the discipline of financial forecasting, KPMG wants to know which factors have an effect on the accuracy of financial forecasts. They would like to enhance accuracy and confidence of forecasts to create measurable business value.
After identifying factors a comprehensive framework for forecast accuracy is provided. This framework can serve as a means of helping organizations to create and improve stakeholder confidence by providing more accurate forecasting. In addition, managers are able to improve their resource allocation and create an agile organization that can easily adapt to changing business conditions. For example, by using technology effectively to access data in real time, business leaders and managers can rely, with the information they use, on the fact that decisions are as accurate and insightful as possible.
1.2 Management problem
market capital. Without a reliable understanding of the business’ direction, its opportunities, and the potential risks, CFOs struggle to provide stakeholders with the transparency and insights that they demand (Parker, 2007). Considering the importance of a well-functioning forecasting process, the aim of this research study is to identify factors that affect the accuracy of financial forecasts.
1.3 Research objective
The main objective of this study is to examine factors that have an influence on the accuracy of financial forecasts. After identifying factors, a comprehensive framework for forecast accuracy is provided. This framework includes all factors that have an influence on accuracy and are divided over components of forecasting. Such a framework is used to ensure that all possible factors are included, validated by experts in the field of budgeting and forecasting, and draw conclusions for participating companies in this research study. In addition, the framework provides useful information that can be used by managers in the design and implementation of forecasting. For instance, managers can improve their understanding of where to start with the implementation process and what they have to take into account to make each forecast a success. Conducting qualitative research at five different companies in the ‘corporate client’ segment will test and validate the factors that are based on the insights from the literature and expert interviews. Corporate clients are companies that are not operating in the financial sector or governance. These companies are operating in all other sectors. Subsequently, interviews with employees of participating companies show the importance of these factors that influence the accuracy of forecasts. Therefore, participating companies benefit from this research study via increased managers’ knowledge and possibilities to improve forecast accuracy. The results of interviews provide recommendations and suggestions to the management of participating companies. These recommendations and suggestions are important to improve the confidence of financial forecast information that is used in the communication with external stakeholders.
1.3.1 Research question
The following research question is drawn to succeed in the research objective as stated above:
What are the important factors that influence the accuracy of financial forecasts, and how do these factors influence the accuracy of financial forecasts?
The following set of sub-questions has been established that further organize the structure of this report: What does financial forecast and accuracy mean?
Why is rolling forecasting seen as a good alternative for the budget?
What are the similarities and differences between the rolling forecast and budget?
Which factors have an influence on the accuracy of financial forecasts based on the insights in the literature?
How important are the factors for participating companies? Why do participating companies find these factors important?
In which way do the factors influence the accuracy of financial forecasts of participating companies?
How important are the relationships between the dimensions regarding the accuracy of financial forecasts?
The first four questions provide an introduction to the subject of this research study. It explains the different aspects which are related to forecasting. For example, ‘accuracy’, ‘financial forecast’, ‘budgeting’ and ‘rolling forecasting’ aspects are explained. It is important to understand underlying relationships between different aspects, in order to provide answer to the research question. The other questions deliver a comprehensive literature review, which provides a theoretical background of the factors that have an effect on the accuracy of financial forecasts. In addition, a clear overview of the factors in the form of a framework is presented. The design of the framework is presented in chapter 3. In chapter 4, research methods are explained. The results of this study are presented in chapter 5. In chapter 6, conclusions and implications are given. Finally, limitations of this research are described, as well as some recommendations for further research.
1.3.2 Research methodology
conducted to test and validate the final list of factors. In chapter 4, research methods, data collection methods and expert interviews are explained more extensively.
1.3.3 Research relevance
The purpose of this study is to identify factors that might influence the accuracy of financial forecasts. The answer to the research question is relevant for organizations that use forecasting methods or planning to use forecasting. This research study is relevant because the results of this research study provide useful information of factors that influence the accuracy of forecasts. This information increases knowledge of organizations that work with forecasting. Therefore, organizations can improve the discipline and accuracy of financial forecasts. In addition, the obtained information is useful for managers that decide to implement forecasting in their organization. This, because forecasting performance has an influence on organizational performance and, therefore, it is important that organizations understand which factors influence the accuracy of their financial forecasts and how these factors influence the accuracy. As a result, this information is helpful for organizations to better predict and manage their future performance. In scientific literature there are several articles about forecasting and aspects which are related to forecasting. Characteristics of financial forecasting, budgeting and rolling forecasting, and differences and similarities between budgeting and rolling forecasting are shortly explained in chapter 2. This, because of the importance to understand general literature about forecasting. This study adds value to existing literature, since it identifies additional factors that influence the accuracy of financial forecasts, that have not been previously studied. In addition, a forecast accuracy framework is designed that can be used as a checklist for the relevance and completeness of this research study.
1.4 Conclusion
2. GENERAL LITERATURE
2.1 Introduction
The goal of the first chapter is to provide an introduction to this research study. This chapter explains the concept ‘forecasting’ and generates an overview of relating aspects. This overview is provided to understand underlying principles between different aspects that are related to forecasting. The overview is based on articles of forecasting, budgeting and planning.
2.2 Financial forecasting
In order to understand the different parts of the main research question, financial forecast and accuracy aspects are explained in this paragraph. Therefore, the first sub-question of this research is: What does
financial forecast and accuracy mean? According to Morlidge and Player (2011), the forecast is a
description of where the organization thinks that they are heading, based on current assumptions. A financial forecast can be produced for different items, like EBIT (Earnings Before Interest and Taxes), cost, sales and revenues. The reason why organizations forecast is that they need to make informed decisions. These informed decisions lead to action, make a change in the business to respond to a forecast, invalidate a forecast, or communicate something to an outside stakeholder about a forecast. Therefore, a good financial forecast should have five qualities. These qualities are; timely, actionable, reliable, aligned and cost effective.
1. Timely; the first requirement of a financial forecast is that it is timely. That is, in time for appropriate action to be taken (Morlidge and Player, 2011). Useful forecasts needs to be available in good time for meaningful action, before events change. For example, if it takes two months to change prices and product cost is heavily dependent on the price of a certain commodity, then you need at least two months visibility of commodity prices.
2. Actionable; the detail which a forecast should contain depends on the nature of decisions that need to be taken (Morlidge and Player, 2011). For example, a financial forecast which simply says that growth will be 3% could be perfectly accurate, but can be seen in many perspectives by the organization. The organization needs to know why the growth is not 2% or which actions should be taken to make it 4% or 5%. In other words, only forecasts that are based on an explicit set of assumptions are actionable.
The first feature of reliable forecasts is that they have to be unbiased. A forecast will never be perfectly accurate, there will always be errors. While it is unrealistic to expect zero errors, errors have to be balanced, there should be approximately as many positive errors as negative errors. Therefore, on average the error is zero and on average organizations will have correct information on which to base a decision. The second feature of a reliable forecast is that the errors (the level of variation) should be within tolerable limits. This means that the range of errors (the variation) should have no material adverse effect on decision-making. Both features of reliability are presented in figure 1. This factor is important for this research study, since factors that have an effect on the accuracy of financial forecasts are examined.
4. The fourth characteristic of good forecasts is that they are aligned (Morlidge and Player, 2011). This means that different forecasts need to be aligned where coordinated action is called for. It would cost a lot of time arguing about different forecasts in the organization and people might be taking completely incompatible decisions. For example, the disconnect between volume and value forecasting is a common form of misalignment in many businesses. Aligning forecasts of value and volume will create a win for the financial community, because their forecast will be more solidly based on what is really happening in the business. For operational managers the advantage is that their forecasts are made more visible and might be taken more seriously.
5. Cost effective; a forecast should be cost effective. The value derived from a forecast should be greater than the cost of producing it (Morlidge and Player, 2011). Because of less detail, faster process times and eliminations of unnecessary analysis, benefits of forecasts will be higher than the costs.
Figure 1. Forecast error and variation (Morlidge and Player, 2011)
simply involves adding up all the estimates of all the individuals responsible for the forecast (Morlidge and Player, 2011). The second type of a forecast model is the mathematical model. A mathematical model is quick and easy to generate forecasts, since patterns of causality are explicit and, therefore, there is a clear link to action. For example, multiplying sales volume by prices is a simple mathematical model used extensively in business forecasting (Morlidge and Player, 2011). The third type of forecasting model is a statistical model. Statistical models can be run quickly and at low cost, because they only look for the existence of patterns in data that repeat themselves. For example, there are large numbers of software products on the market that apply a range of complex algorithms to historical data in order to predict sales (Morlidge and Player, 2011).
To conclude, financial forecasting is important since it enables management to change operations at the right time in order to create the greatest benefit. Therefore, organizations currently face the need to use more accurate financial forecasts (Parker, 2007). All organizations need and have to apply more accurate forecasts to better predict and manage their future performance. This, because it helps organizations to prevent losses by making proper decisions based on relevant information (Parker, 2007).
2.3 Budgeting
In the previous paragraph, the first sub-question is answered. Another important relationship is that between budgeting and forecasting. This relationship is explained in order to understand the change of organizations in use of these methods. In the following section this relation is explained. The second sub-question is: Why should organizations use other forecasting methods instead of budgeting? According to Merchant and Van der Stede (2007), the budget is used by organizations to plan and communicate actions, coordinate sharing of information across the organization, facilitate top management oversight, control costs, and motivate employees to achieve targets. The budget also serves the role of performance management and compensation, strategy formulation, coordination, and communication (Tanlu, 2007).
organizations have seen problems of budgeting in terms of an ineffective process that is too long, costs too much, reinforce bureaucracy, and fails to provide its users with sufficient value (Player, 2003).
Subsequently, in turbulent times organizations are faced with an alarming level of volatility, uncertainty and complexity (Barrett, 2003). Situations such as globalization of markets and suppliers, increased competition, new regulations and technology, and an uncertain economic outlook have led to shorter product life cycles, cost fluctuations and price pressures. As a result, growing number of organizations have abandoned budgeting. To conclude, many organizations have targeted improvements in forecast accuracy and forecasting techniques by eliminating budget and by introducing ‘rolling forecasts’ (Morlidge and Player, 2011). These improvements might be the solution for inaccurate financial forecasts.
2.4 Rolling forecasting
As mentioned in paragraph 2.3, many organizations have targeted improvements in forecast accuracy by introducing rolling forecasting and eliminating budget. Berland (2011) calls for major structural reforms in an organization to manage the organization without a budget that can lead to developing rolling forecasts. As a result, organizations might adapt better to an ever-changing market. Many organizations have introduced rolling forecasting in order to achieve more accurate forecasts. Therefore, this forecasting method is explained in this paragraph. The third sub-question of this research is: Why is rolling
forecasting seen as a good alternative for the budget? A rolling forecast maintains a constant
forward-looking time horizon, usually between 12 and 18 months (Hansen, 2010). Rolling forecasts provide information that can be used for managing the strategy and making decisions (Fraser and Hope, 1999). According to Tanlu (2007), there are four characteristics of rolling forecasts that ensure that numbers are up to date and not obsolete.
1. First, rolling forecasts are not static, they require regular updating. Companies make forecasts over different horizons and update at different intervals e.g., monthly vs. quarterly (Tanlu, 2007). The output of a rolling forecast contains realized results (the actual) of an organization. These realized results provide the organization information about the past, wherefore the organization needs to take corrective action in their rolling forecast for improvement.
2. Second, rolling forecasts are made for horizons that are not limited to the fiscal year (Tanlu, 2007). Forecasts cover predetermined periods that go beyond the fiscal year.
budgetary slack and gaming behavior. As a result, rolling forecasts are able to be more accurate than tools that have a link with incentives and performance targets.
4. Fourth, rolling forecasts are less detailed than a budget (Tanlu, 2007). This makes it easier and faster to produce a regular updated forecast.
Besides positive aspects of rolling forecasts there are some negative aspects. Firstly, shifting target in the rolling forecasting process provides uncertainty between managers, which reduce motivation in the organization (Groot and Waal, 2007). Secondly, it is difficult to make payment agreements because of varying goals of a rolling forecast (Groot and Waal, 2007). At last, due to the focus of a rolling forecast to the future, it is possible that the strategic plan for the short- and midterm is overlooked (Bertens and Van Aken, 2006). However, considering the importance of an accurate forecasting process and strong positive characteristics of a rolling forecast that makes accurate forecasts realizable, this method will be a good alternative for a budget.
2.5 Rolling forecasting versus Budgeting
In this paragraph, a comparison is made between rolling forecasting and budgeting. This comparison is intended to provide organizations a better view of why they could use more rolling forecasting instead of budgeting. Therefore, the fourth sub-question is: What are the similarities and differences between the
rolling forecast and budget? There are several similarities between rolling forecasting and budgeting. An
important similarity is that both methods are used in planning (Tanlu, 2007). The planning process produces a set of plans and these describe objectives and alternative strategies (Armstrong, 1983). In practice, the planning process is affected by a change in the budgeting and rolling forecasting process. In addition, both methods estimate future outcomes.
forecasts are made for horizons that are not limited to the fiscal year and look further to the future. Despite strong criticism by practitioners, studies have shown that the vast majority of organizations still use budgets (Sivabalan et al., 2009). According to Sivabalan et al., (2009), this raises the question: ‘If
budgets are so problematic, why is it that most organizations continue to use them?’ There are two
reasons why most organizations still use budgets. Firstly, organizations use budgets for planning and control reasons and not for evaluation reasons (Sivabalan et al., 2009). Where there is a high degree of environmental uncertainty, it would be unfair and irrelevant to measure and evaluate the performance of employees against the achievement of budgetary targets (Tanlu, 2007). As a result, employees show less effort in promoting budget gaming and budgetary slack. Consequently, if budgets are used for planning and control as opposed to evaluation, many budget criticism might no longer be relevant (Merchant and Van der Stede, 2007). Secondly, budgeting practices can be improved by using rolling forecasts as a supplement. Rolling forecasts involve more frequent forecasting by organizations in order to generate more accurate financial predictions (Sivabalan et al., 2009). This indicates that in most organizations budgets are still used, for planning and control reasons, and that they are supplemented by rolling forecasts. Therefore, most organizations overcome many of the problems claimed for annual budgets, and the use of improved budgeting practices may explain why budgets persists despite the significant criticisms in the literature (Sivabalan et al., 2009).
2.6 Conclusion
3. THEORETICAL BACKGROUND OF FACTORS
3.1 Introduction
The second chapter provided a short introduction of the concept ‘forecasting’. This chapter discusses a theoretical background of possible factors that have an effect on the accuracy of financial forecasts. The theoretical background is based on articles of forecasting, budgeting and planning. This theoretical background is used for to determine factors of influence on the accuracy of financial forecasts for this research study. Ultimately, the main goal of this chapter is to answer the fifth sub-question: Which factors
have an influence on the accuracy of financial forecasts based on the insights in the literature? The
theoretical background of the factors provide a first impression of the influence of the factors on the accuracy of financial forecasts. Most articles approach forecast accuracy from a process and technical perspective. There are however more factors that have an influence on the accuracy of financial forecasts based on some of the articles. Paragraph 3.2 explains perspectives of forecasting in detail. Subsequently, a clear framework is created to provide a clear view of the factors. This framework is the forecast
accuracy framework, and used to describe factors in conjunction that influence the accuracy of financial
forecasts. Validation of factors is executed by interviews held with employees of participating companies and is discussed in the results of chapter 5. This chapter ends with an overview of possible factors that have an influence and serve as a starting point of this research study framework that is appropriate for forecast accuracy.
3.2 Perspectives of forecasting
There are also authors that talk about helping organizations to design an appropriate forecasting process. According to Zotteri and Kalchsmidt (2007), some familiarity with the forecasting process makes managers a better and more discriminating consumer of forecasts. Apart from describing the accuracy from a process perspective, some articles discuss the accuracy issue of forecasting from a technical perspective. This concerns working out how to achieve the likely future outcome, bearing in mind the forecasting processes. It is important to select the right techniques and methods to sustain a good business forecast (Bunn and Wright, 1991). Therefore, many researchers analyzed how the use of various techniques affect forecast accuracy. Information technology is also important since forecasters sometimes have a poor grasp of the business they are attempting to model. They need help to make the right choices about the techniques to use. Further, most of the researchers are focused on the forecast developers rather than the users of the forecast (Wacker and Sprague, 1995). This means that they are more interested in the techniques and methods the developers need to maintain a forecast. At last, statistical results are used to suggest that organizations with new technology have lower forecast errors (Wacker and Sprague, 1995). Organizations invest enormous amounts in information technology to improve operational performance that will increase the accuracy of financial forecasts.
As mentioned before, articles approach forecast accuracy from a process and technical perspective. There are however more factors in literature that are related to the accuracy of forecasts. These factors are examined as well in order to manage the forecast accuracy issue. According to Winklhofer et al., (1996) and Mentzer et al., (1999), influences from inside the organization are also important to achieve accurate financial forecasts. For example, complexity of organization and support of management are some factors. In addition, in order to produce a good financial forecast, a model has to be used by managers (Morlidge and Player, 2011). Therefore, next to the process and IT factors, a forecasting model, organization and people are examined as factors of influence. To conclude, this research study examines process, IT, a forecasting model, and organization and people perspectives of forecasting to determine the factors of influence on the accuracy of forecasting.
3.3 Process
Policies and guidelines are necessary to guard against errors and efficiencies. This paragraph discussed the following process factors in detail:
• Policies & Guidelines • Protocol
• Throughput time
• Budget alignment with forecasting Policies & Guidelines
Organizations have an elaborate set of policies and guidelines that they expect employees to follow (Merchant and van der Stede, 2007). These policies and guidelines can be helpful to achieve a clear forecasting process. The role of employees and their desired actions have to be well communicated, and they have to understand what is required of them in the forecasting process (Armstrong, 1983). The actions will be noticed and rewarded or punished in a certain manner. As a result, a clear forecasting process can improve the accuracy of financial forecasts.
Protocol
In addition, a protocol is send to employees to provide them with information concerning instructions and assumptions of the process (Mentzer and Kahn, 1997). In addition, deadlines should be given to people that are involved in the forecasting process for uploading their information (Armstrong, 1983). If the deadline is met, the uploaded information is passed to the next step in the forecasting process. If a deadline is not met, the whole forecasting process is destroyed. The forecasting process ends after deadlines that are not met, since the forecasting process is planned and process phases are scheduled, and thereby difficult to restore the whole process. According to Mentzer and Kahn (1997), a significant majority of respondents in their survey indicated that their forecasting process was prepared via formal and routine processes. The precise instructions concerning the forecasting process were clearly formulated.
Throughput time
including lower work-in-process and finished goods inventory levels, lower costs, less forecasting error (because forecasts are for shorter time horizons), and improved quality (Johnson, 2003). More importantly, reductions in throughput time increase flexibility and reduce time required to respond to customer orders. This flexibility and reduced responding time is vital to survival and profitability of numerous firms, especially those experiencing market pressures for shorter delivery lead times of customized product.
Budget alignment with forecasting
Organizations use forecasting as supplement for their budget (Sivabalan et al., 2009). For example, budgets are adjusted on basis of forecasts, and budget targets are derived from these forecasts. Managers have to make the decision to align the budget with forecasting, in order to achieve an efficient forecasting process with use of two different forecasting methods. In addition, guidelines should be specific and managers have to understand what their budget responsibilities are regarding the purchases and expenditures of their organization. Therefore, alignment with the budget can prevent an inefficient forecasting process and increase the accuracy of financial forecasts.
3.4 Information Technology
Information technology (IT) is the second component of our forecasting framework. Databases, technology, technical support, and forecasting tools are all potentially influencing factors on the flow of information from the environment to its point of use in organizational decision making (Fildes and Hastings, 1994). The information technology on which organizations base their decisions have different implications for both the level of achievable accuracy and potential for bias. Any technological factor that might influence the accuracy of forecasts, should therefore be identified and discussed in this paragraph. In current literature, several factors of information technology are distinguished:
• Master data management • IT tools & Computer systems • Technical processing speed
• Information technology general controls • Workflow management
• Revision control
Master data management
Master data management (MDM) is playing an important role in improving the quality, reliability and insightfulness of the data that organizations use to produce forecasts. This system contains information from various sources, needed to obtain accurate forecasts (Jain, 2007). Information that is used as input for a forecast is important since it has influence on the reliability and quality of the forecast outcome (Morlidge and Player, 2011). Therefore, statistical models should use historical data with high quality, since bad quality will not lead to accurate forecasts. Further, if an important source of information is not taking into account, the accuracy of a forecast declines. To conclude, using a good master data management solution ensures consistent use of data throughout the organization.
IT tools & Computer systems
Various scientist have analyzed which IT tools and computer systems are used by organizations to transfer data (Zotteri and Kalchschmidt, 2007). The number of products for which a forecast is developed, number of product variants, and level of customization that the market requires, need to be taken into account when using IT tools and computer systems. Organizations use computer systems to record all data that is needed to produce a forecast. In addition to systems, tools like spreadsheets, Enterprise Resource Planning (e.g., Sap and Oracle), and specialist accounting software are used to produce forecasts (Parker, 2007). A general understanding is that quantitative methods become more efficient than qualitative methods as the number of products increase. However, both quantitative and qualitative approaches tend to be adopted by organizations to produce a forecast.
Technical processing speed
Information technology general controls
Information technology general controls (ITGCs) ensure that data processing, that takes place in systems and IT, occurs in a controlled environment. In addition, ITGCs support data and computer operations integrity and security (Norman et al., 2009). For example, a password protect data from unauthorized access and changes. It is important to understand how ITGCs interact and effect forecasting, to make an optimal use of forecasts.
Workflow management
The accuracy of financial forecasts can also be influenced by workflow management of an organization. Workflow management is a core component of modern information technology infrastructure, that automates execution of business processes (Caverlee et al., 2007). For example, this system supports the business process, since it gives certain users access to the data at demand time. Therefore, this system can allow an organization to identify inefficiencies in the forecasting process.
Revision control
The quality of information technology and thereby the accuracy of forecasts is also determined by revision control. Organizations use revision control to track and manage the complexity of a project (O’Sullivan, 2009). A revision control allows team members to track history of files they work on during the development of a project. For example, people can see who made a change, understand when and why it was made, inspect details of change, and re-create state of the project at time of the change. Therefore, revision control allows people to correct mistakes and keep the project up to date, that is needed to produce an accurate financial forecast.
IT strategy alignment with business strategy
can follow key drivers such as actual performance, in order to perform change actions, that could increase performance and the accuracy of financial forecasts.
3.5 Forecasting Model
The forecasting model is the third component that is discussed in order to find factors that influence the accuracy of forecasts. As previous mentioned in paragraph 3.2, a forecasting model has to be used by managers to ensure a good and accurate financial forecast (Morlidge and Player, 2011). A forecast model is used to transform a set of assumptions about the world (input) into a forecast (output). Based on that, each forecasting model needs a set of inputs about environment, and underlying trends in the business. The inputs are assumptions that may include outside world phenomena such as inflation, market growth or commodity prices (Morlidge and Player, 2011). All forecasting is based on hypotheses that are, conditions assumed to be true or expected to occur, in order to validate the forecast (Elgers et al., 1994). Managers need to agree with the model, since the model is designed on basis of purpose (mission), goals (vision) and values of an organization (Kaplan and Norton, 2008). The factors of a forecasting model that are discussed in detail in this paragraph are:
• Frequency of update • Forecast horizon • Level of detail
• Key performance indicators • Top-down & Bottom-up strategies Frequency of update
Forecast horizon
The right choice of forecast horizon is important for the accuracy of forecasts. The forecast horizon is the period of time in the future covered by a forecast (Morlidge and Player, 2011). The length of the forecast horizon needed is determined by decision-making lead times, and will vary between and within businesses. The forecast horizon should be related to the longest lead time in the business. For example, if the longest lead time to launch a new product is 12 months, then the forecast horizon should be related at least 12 months. Thus, the horizon for forecasting is based on the particular organization and nature of the decisions. Overall, most organizations adopt a forecast horizon longer than a year and updated every quarter, which could increase visibility. For example an 18-month forecast.
Level of detail
The level of detail for which forecasts are prepared has to be decided by an organization. Small (1980) found that the majority of organizations produced forecasts for more than one level of product or market detail. For example, a forecast can be produced on the level of stock keeping unit (SKU), product item, product group, region, and channel of distributions (Jain, 2007). Stock keeping units refer to items of stock that are completely specific to function, style, size and location. The production and inventory policies of different SKUs are influenced by characteristics of an product (Kampen et al., 2011). The level of forecast preparation can also be related to size of a company. Larger retailers are more inclined to develop industry forecast on a higher level of aggregation than smaller retailers (Winklehofer et al., 1996). Based on above, level of detail might influence the accuracy of financial forecasts and is therefore included in this research.
Key performance indicators
Top-down & Bottom-up strategies
Two general approaches have been suggested for developing forecasts. These two approaches can be labeled as top-down and bottom-up forecasting strategies (Schwarzkopf et al., 2001). The bottom-up strategy refers to separately forecasting individual items, and this forecast could be made on the operational level. The top-down strategy refers to forecasting the demand for aggregate items. This forecast could be made on the highest level within an organization. Forecasting should be performed on the level where people can be held responsible for the developed forecast. Further, forecasts should be developed on the level where people have the knowledge and skills to produce an forecast to ensure a positive influence on the accuracy of the financial forecast.
3.6 People & Organization
The factors of people and organization that are discussed in this paragraph are: • Strategy of organization • Structure of organization • Complexity of organization • Incentives • Culture • Tone at the top
• Roles & Responsibilities • Skills & Competences • Management orientation • Support of management Strategy of organization
The strategy of an organization influences the accuracy of forecasts. Montgomery (2002) states that organizations should ensure that their forecasting processes is organized strategically. For example, by encouraging a strategic focus, keeping summarize level of detail and modeling with parameters and metrics, organizations can effectively plan for the short-and long-term. According to Montgomery (2002), it is important that the strategy of an organization is aligned with forecasting in order to improve the outcome of a forecast. Therefore, more accurate financial forecasts are provided if the forecastings process contains strategy.
Structure of organization
accuracy. Departments that make changes to the forecast are marketing, sales planning, and product management, since these functional areas have knowledge of customers and markets
Complexity of organization
The complexity of an organization has impact on the accuracy of forecasts. Increasing complexity of organizations has made it more difficult for decision makers to take all the factors relating to the future development of the organization into account. Complexity is related to the number of products offered, number of submarkets served, number of employees involved, demand variability, information quality, and number of relationships among them (Winklhofer et al., 1996). In his article, Peterson (1993) analyzed that practices and processes of a forecast changed according to firm size. For example, larger firms seem to employ forecasts for more purposes than smaller ones. Further, larger firms produce forecasts for different aggregation levels, tend to use a bottom-up forecasting process, and use quantitative and sophisticated forecasting techniques.
Incentives
Incentives are an important factor which influences forecast accuracy (Merchant and Van der Stede, 2007). Rewarding employees based on their performance relating to a budget or target can result in gaming. In case people game the system they can destroy value in two ways. First, when employees perceive budgets as difficult to achieve, both superiors and subordinates can alter formulation of budgets. For example, employees lower the budget. Second, they game the realization of budgets or targets and in doing so destroy value for their organization (Jensen, 2003). Therefore, producing a forecast should be decoupled from performance targets.
Culture
Organizations can have a culture where people are collectively oriented toward solutions and recognize responsibility for organizational problems (Hofstede, 1983). These organizations have a culture that is open, where people discuss in order to find solutions for problems. As a result, an open culture can improve the accuracy of financial forecasts.
Tone at the top
management cannot say one thing and do another. For example, management sometimes sets the wrong tone by not responding appropriately to matters brought to their attention, which can influence forecasting and, therefore, the performance of an organization.
Roles & Responsibilities
According to Mentzer et al., (1999), it is an advantage for an organization when employees know their roles and responsibilities in the forecasting process and understand the forecasting process thoroughly. Organizations require employees who are responsible for coordination and communication of the forecasting processes. These employees can improve the quality and accuracy of forecasts. At last, it is an advantage for an organization to have a manager who leads the forecast and understands how to achieve high quality forecasting output.
Skills & Competences
Employees involved in forecasting should possess certain skills and competences (Merchant and Van der Stede, 2007). Organizations should offer ongoing training in statistics and business and industry knowledge in which the organization operates (Mentzer et al.,1999). Training helps forecasting personnel to develop certain qualifications, for example experience in computer systems, statistics, and understanding of business environments that influence forecast accuracy.
Management orientation
Management orientation influences the accuracy of forecasts, since the results of the forecasting process are discussed during review meetings with management. In order to improve their ability to anticipate, management should become more forward-looking instead of backward-looking (Barrett, 2003). If management is backward-looking, they spend their time in discussing the historical financial data of the organization. Conversely, if the management of an organization is forward-looking, they spend their time in discussing the output of the forecast. It is not useful to discuss only historical financial data, since results obtained in the past cannot be changed. Therefore, there should be a balance between focus on historical financial data to manage performance, and decision-making based on the output of the forecast.
Support of management
rolling forecast, other parts of the organization will follow. This is the case since managers show how important this change is for determining success of forecasts.
3.7 External factors
External factors that have an influence on the accuracy of forecasts are explained in this paragraph. The factors are:
• Economic developments • Market competition
Both factors influence the accuracy of financial forecasts from outside an organization. Therefore, these external factors are discussed separately opposite to all internal factors which are identified and explained in previous paragraphs. Economic developments is one of the first external factor. The contraction in global economic activity caused by the global financial crisis had a big influence on the accuracy of financial forecasts, because forecasters have not predicted the sudden financial crisis in their calculation (Morlidge and Player, 2011). The financial crisis has an influence on market prices and demand and supply of organizations. For instance, if organizations are not able to adjust forecasts on time regarding developments in the economy, the accuracy of financial forecasts will decrease. According to McHugh and Sparkes (1983) and Sanders and Manrodt (1994), the factors that can be considered as most important in limiting forecast accuracy are outside the control of management, for example the instability in the national and world economy.
Market competition is the second external factor which influences the accuracy of forecasts. With
of market competition and the dynamic and unpredictable nature of the external environment. Without adapting the organization to changes in the market, the external environmental uncertainty increases and, therefore, the accuracy of forecasts will decrease. In table 1 an overview of the factors is presented. These factors are also presented in figure 3, in the preliminary forecast accuracy framework.
Table 1. Overview of factors
3.8 Design of forecast accuracy framework
Previous paragraphs of this chapter explained the four internal components of forecasting. These components are dimensions of the forecast accuracy framework. This framework is used in order to provide classification of components. The forecast accuracy framework enables managers to recognize and understand problems in the field of forecasting. The designed forecast accuracy framework starts out with four components of forecasting. These four components are; process, information technology, model and people and organization. Factors that belong to these four components are structured over dimensions of the framework and, finally, a framework is created that is appropriate for forecast accuracy. Because of issues with accuracy in the forecasting field, this framework is used as a starting point for this research study. The ‘Klaverbladmodel’ made by Noordam and Van der Zalm (2006), is another model that as well brings dimensions together. This model described main four areas of an organization that have to be balanced to achieve desired short- and long-term results (Noordam and Van der Zalm, 2006). With the use of this model information about organizational problems, complexities, and business operational issues are provided.
Figure 2. Forecast accuracy model
3.9 Preliminary forecast accuracy framework
In the previous paragraph, the forecast accuracy model is designed. In this paragraph the preliminary forecast accuracy framework is presented. The conceptual model of all relevant factors of influence on the accuracy of forecasting forms the forecast accuracy framework. In order to build the framework, firstly, choices must be made regarding the design of the forecast model. Factors belonging to the forecasting model dimension are; frequency of update, forecast horizon, level of detail, key performance
indicators (KPI’s) and top-down & bottom-up strategies. Secondly, factors belonging to the process
dimension are; policies & guidelines, protocol, throughput time and budget alignment with forecasting. Thirdly, factors that influence the accuracy of forecasts from inside the organization are; strategy of
organization, structure of organization, complexity of organization, incentives, culture, tone at the top, roles & responsibilities, skills & competences, management orientation and support of management.
Fourthly, factors within the information technology dimension are; master data management, IT tools &
computer systems, technical processing speed, information technology general controls (ITGCs), workflow management, revision control and IT strategy alignment with business strategy. At last, external
Figure 3. Preliminary forecast accuracy framework
3.10 Relationships between dimensions
For organizations that use forecasting, it is important to know how the relationships are between dimensions that influence the accuracy of financial forecasts. In interviews with participating companies the following sub-question is answered: How important are the relationships between the dimensions
regarding the accuracy of financial forecasts? The answers of participating companies show the
importance of relationships between dimensions in more detail. For instance, if there are mutual relationships between dimensions that influence the accuracy of forecasts, then it is insufficient to only look at unilateral relationships between a dimension and accuracy of forecast. Therefore, organizations have to know whether the influence of dimensions on accuracy of forecasts are dependent or independent from each other.
Furthermore, the following question needs to be answered since further professionalization of forecasting can be achieved through an integrated approach: How important are the relationships between the
dimensions regarding the accuracy of financial forecasts? According to Kaplan and Norton (2008), there
should be an integrated set of processes and tools in an organization to prevent an organizations underperformance. Participating companies in this research have to look on an integrated manner to forecasting, which means an analysis of relationships between forecasting model, process, organization & people, and information technology. These dimensions of the framework should be examined in an integrated manner, to focus effectively on accuracy related problems in the organization.
3.11 Conclusion
4. RESEARCH METHODS
4.1 Introduction
The theoretical background of the factors that influence the accuracy of financial forecasts is discussed in chapter three. The purpose of this chapter is to explain decisions concerning the chosen research method, the type of case study, number of cases, selection of companies and the data collection process. This chapter starts with the research method for this research study. Subsequently, the appropriate type of case study is discussed and the number of case studies is provided. In addition, the selection of company cases and data collection methods used in this research are outlined. Furthermore, findings from expert interviews are discussed to validate and complete the list of factors. At the end, a clear view is obtained regarding the research methods and data analysis that starts in the following chapter.
4.2 Case study
According to Yin (2009), there are different research methods that can be used in order to answer a research question. The research methods are collecting and analyzing empirical evidence in a different way. Each research method has its own advantages and disadvantages. The most appropriate research method for this research is the case study. The case study is used for the following reasons: firstly, there is no control over behavioral events. For example, in this research there is no control over employees of participating companies that could influence the accuracy of financial forecasts. This also indicates that relevant behaviors cannot be manipulated directly, precisely and systematically. Secondly, this research is focused on contemporary events as opposed to historical events. This means that only the current influences of the factors on the accuracy of forecasts is examined. Thirdly, the research question of this study is: What are the important factors that influence the accuracy of financial forecasts, and how do
these factors influence the accuracy of financial forecasts?
In order to answer this research question, a set of sub-questions has been established that clarifies this research question and provides information that is needed to answer this research question. These questions also provide better understanding of the factors that influence the accuracy of forecasts. Because of these reasons, most of the sub-questions in this research are ‘how’ and ‘why’ related questions, which is the main reason why a case study method is relevant. For example, How important
The case study method is also relevant, because the sub-questions require an extensive and in-depth research of the phenomenon, the accuracy of financial forecasts. At last, ‘what’ questions are also exploratory and, therefore, appropriate for the use of case studies (Yin, 2009).
4.3 Type of case study
In the previous paragraph, the reason for a case study as most appropriate research method for this research study is explained. Another question concerning the type of case study still needs to be answered. Scapens (2004) has identified different types of case studies; Descriptive cases studies describe accounting systems, techniques and procedures used in practice. Such cases studies are useful in providing information concerning the nature of contemporary accounting practices. Illustrative case
studies are used to illustrate new and possibly innovative practices developed by particular companies. Experimental case studies develop new accounting procedures and techniques that are intended to be
helpful to accounting practitioners. Explanatory case studies attempt to explain the reasons for observed accounting practices. The focus of the research is on the specific case. Exploratory case studies are used to explore the possible reasons for particular accounting practices. These case studies represent preliminary investigations, which means that they are intended to generate ideas and hypotheses for rigorous empirical testing at a later stage. The objective of subsequent research is to produce generalizations about the practices (Scapens, 2004).
produce generalizations about the practices. This research study can however not be generalized, because the cases in this research study represent a small sample. Therefore, this study will be an exploratory case study.
4.4 Number of cases
After the selection of the type of case study, the number of company cases is discussed in detail in this paragraph. In this research study, the case is a company that applies forecasting or has the plan to implement forecasting. The respondents of this research share their knowledge and experience regarding forecasting and the importance of the factors in relation to the accuracy of forecasting. Now, the decision concerning the amount of cases has to be made. In order to determine the choice for a number of cases, the distinction between single or multiple case studies has to be clear. In the following section the number of cases are explained.
Single and Multiple cases
According to Yin (2009), case studies can be either single or multiple-case designs. Firstly, single cases are used to confirm or challenge a theory (Yin, 2009). For example, one rationale for a single case is when it represents the critical case in testing a well-formulated theory. The theory has specified a clear set of propositions as well as the circumstances within which the propositions are believed to be true. A single case, meeting all of the conditions for testing the theory, can confirm, challenge or extend the theory. Secondly, a single case represents a unique or extreme case (Yin, 2009). The researcher will focus on one unique or extreme case, because these cases provide valuable information for the research. Thirdly, single-case studies are also ideal for revelatory cases where an observer has access to a phenomenon that was previously inaccessible (Yin, 2009). At last, single-case designs require careful investigation to avoid misrepresentation and to maximize investigators access to evidence (Yin, 2009). A major step in designing and conducting a single case is defining the unit of analysis or the case itself. An operational definition is needed and some caution must be exercised before a total commitment to a complete case study is made (Yin, 2009). This caution is important to ensure that the case in fact is relevant to the issues and questions of interest. In addition, within the single case may still be incorporated subunits of analyses, so that a more complex design is developed. If, however too much attention is given to these subunits and if the larger aspects of the case begin to be ignored, the case study itself will shift its orientations and changes its nature (Yin, 2009).
There are also some reasons why multiple-case studies are selected concerning the design of a case study. Firstly, multiple-case studies are a powerful means to create theory, because they permit replication and extension among individual cases (Eisenhardt, 1991). Replication means that individual cases can be used for independent corroboration of specific propositions. This corroboration helps researchers to perceive patters more easily and to eliminate chance associations. Secondly, extension refers to use of multiple cases to develop more elaborate theory. Different cases often emphasize complementary aspects of a phenomenon. By piercing together individual patters, a researcher can draw a more complete theoretical picture (Eisenhardt, 1991). Thirdly, evidence from multiple cases is often considered more compelling, and the overall study is therefore regarded as being more robust (Yin, 2009). However, any use of multiple-case design should permit replication and the researcher must choose each case carefully. These requirements require extensive resources and time to conduct multiple-case study. Therefore, these types of designs are more expensive and time-consuming to conduct (Yin, 2009).
Decision between alternatives
With the provided information in the previous paragraph, multiple-case study is selected to design this research study. This indicates that multiple companies are investigated to find an answer for the research question. Because of the following reasons the multiple-case study is selected: firstly, the purpose of this research study is to investigate companies with more general characteristics or aspects and not companies with specific or unique aspects. The reason behind the choice for a multiple-case study is to explore a general view of the importance of factors, and the way these factors influence the accuracy of financial forecasts. The purpose of this research is not to investigate a unique or extreme case (Yin, 2009). Secondly, evidence from multiple cases is often more compelling and robust, and, therefore, multiple-case study is applicable for this research study (Yin, 2009). Multiple companies are investigated for this multiple-case study. The additional time and resources required for each extra case are rewarded with extra valuable information of these participating companies. This means that replication of individual cases is important to achieve extra valuable information from participating cases. Thirdly, a single case study is not appropriate for this research study because there is not a well-formulated theory that has to be tested (Yin, 2009). At last, a single case study in not appropriate, since it is not about revelatory cases where an observer has access to a phenomenon that was previously inaccessible (Yin, 2009).
4.5 Selecting cases
each company to obtain a clear view of the factors. In addition, participating companies are compared to assign the importance of certain factors to characteristics of companies and markets. Nonetheless, the main purpose of this research is to explore the factors and their effects on the accuracy of forecasts at different companies. Further, a comparison between participating companies provides additional meaning regarding the effects obtained in this research. Therefore, more information could be obtained from companies from different industries with different characteristics. As a result, a clear distinction is made between different sectors, and the obtained information can be used to improve the accuracy of forecasts.
According to Swanborn (2010), the targeted domain has to be defined before cases can be selected for which conclusions can be valid. In this research most participating companies are corporate clients of KPMG Netherlands and are selected based on the requirements. First, participating companies are using forecasting or planning to implement (rolling) forecasting, and they have knowledge and experience regarding the use of forecasting. In each company there should be an opportunity to improve the accuracy of forecasting. Second, most participating companies are clients of KPMG. This would make it easier to find enough respondents. Swanborn (2010) stated that cases could be selected based on personal interest and involvement, incidental contacts or the researchers network. Since this research focuses on different companies in different sectors, a comparative analysis is made between the participating companies. In this way, differences in important factors, effects and influences are assigned to characteristics of companies and markets. In table 2 an overview is provided of participating companies and their market characteristics.
Company Market characteristics
Company A. planning to implement rolling forecasting
Dynamic market and market leader
Company B. using forecasting Stable and shrinking market and follower of the market
leader
Company C. using rolling forecasting Dynamic market and market leader
Company D. using forecasting Dynamic market and one of the big leaders
Company E. using forecasting Dynamic market and one of the big leaders
Table 2. Overview of companies and market characteristics
