Chapter 4 Forecasting Cash Flows – Qualitative Methods

Document Sample
Chapter 4 Forecasting Cash Flows – Qualitative Methods Powered By Docstoc
					                            CHAPTER 4


4.1 Under what circumstances is the use of qualitative forecasting techniques

4.2 Outline the steps involved in undertaking a survey to collect forecast estimates
    from individuals. Outline the advantages and disadvantages of using groups, as
    opposed to individuals, to provide forecasts.

4.3 What are the steps normally applied when undertaking a Delphi survey? What
    are some of the variants of this methods?

4.4 How is nominal group technique (NGT) different to the Delphi method?

4.5 Outline the ‘jury of expert opinion’ method. How does this differ from the Delphi

4.6 A company offering eco-tourism adventures wishes to forecast the numbers of
    annual tourists of various types coming to a region five years from now, and
    chooses the Delphi method to develop their forecasts.

(a)   Explain how a panel of appropriate experts could be selected.

(b) Suggest how the information to the panel and questions could be framed.

(c)   How many rounds of data collection would you expect to undertake? Discuss
      how these would proceed.

4.7 The management of a coal-fired electricity generation station is concerned that
    new technology to reduce carbon dioxide emissions will be required to offset
    emissions, and measures such as purchase of carbon credits will be required if the
    national government signs up to the Kyoto protocol. The severity of adjustment
    for the company will depend on the extent of reduction in greenhouse-gas
    emissions agreed to by the national government. Explain how scenario
    forecasting could be used to aid decision making by the company.

4.8 A company engaged in sugar refining and export wishes to make medium term
    projections of international sugar prices. Suggest a method that could be used to
    make these forecasts.


Answer to Q 4.1

Qualitative techniques can be used in a wide range of circumstances. In some cases
quantitative techniques cannot be used, e.g. when past information about the values
being forecast does not exist. For example, for a new product, there are no past data on
sales on which to base estimates of future sales. Similarly, past sales of a product might
not be relevant if a competitor launches a new product with superior features or
performance. In other situations, there is insufficient time to obtain data or use
quantitative techniques, or circumstances are changing so rapidly that a statistically
based forecast would be of little guidance. Even when statistical techniques are
available qualitative techniques involving human judgement are often used by managers
for forecasting. Further, managers appear to be more comfortable dealing with their
own judgements or with those of a colleague, compared with forecasts generated via a
computer package and lacking transparency. Even when quantitative techniques are
used, estimates may be combined with qualitative judgements, or supplemented,
reviewed or screened by subjecting them to qualitative judgements i.e. any forecasts
provided by an analyst are for decision-support, not decision-making.

It has been suggested that the widespread use of human judgement to make business
forecasts can be rationalised in two ways. The first is that people might be better able to
detect changing patterns in time series, which exhibit considerable random variations,
compared with statistical models. The second rationale for the use of human judgement
is that people might be able to integrate external (i.e. non-time series) information into
the forecasting process. A number of qualitative forecasting techniques have been
developed to provide estimates of key parameters for use in financial analysis in such

Answer to Q 4.2

The steps involved in undertaking a survey are outlined in Figure 4.1(of the book). It is
important to note that the decisions made in the early stages affect the choices made in
later stages. For example information needs specified at the start will affect the
sampling design, the way in which the questionnaire is structured and the selection of
data analysis techniques. These ‘forward links’ in the survey process are indicated by
solid arrows in Figure 4.1. It should be noted that if there were only forward links in
the process then the conducting of a survey could be done one step at a time,
completing each step before considering the next. Implicit in this ‘single direction’
approach is the assumption that there are no limiting factors in later steps. This is
seldom, if ever, the case. For instance there are often limitations on data collection or
data processing resources, i.e. a budget constraint. These limitations restrict the
alternatives available at earlier steps; these backward linkages are indicated in Figure
4.1 by dashed lines running upwards. Backward linkages run from the collect data and
analyse data boxes back to the develop questionnaire and sample design phases. This
illustrates that major decisions concerning data collection and analysis should always be
considered before selecting a sample and designing a questionnaire.

Techniques for collecting information from individuals such as surveys and polls can be
easy to implement, done at a low cost and on a timely basis. However evidence
suggests that forecasts produced by groups offer greater forecasting accuracy than those
derived from individuals. Groups also provide more information, although the marginal
increase in information content decreases as group size increases. The use of groups
also provides an opportunity to gain more information about the range of possible
outcome values hence giving an insight into the risk associated with the estimates.
From a behavioural perspective, it is also likely that a group responsible for a
implementing a project will have greater commitment to it if they are also involved in
providing estimates of variables used in the financial analysis leading up to a decision
to proceed with a project. The choice between using individual versus group techniques
really comes down to the particular situation and what is feasible. For instance group
techniques such as Delphi and NGT often require greater skills, resources and time than
collecting information from individuals using some form of survey or poll.

Answer to Q 4.3

In a classic Delphi survey, the first round is unstructured, allowing panellists to identify
freely and elaborate on the issues that they consider important. These are then
consolidated into a single set by the monitors, who then produce a structured
questionnaire designed to elicit the views, opinions and judgements of the panellists in a
quantitative form. The consolidated list of scenarios is presented to the panellists in
round two, at which time they place estimates on key variables such as the time an
event will occur. These responses are then summarised and the summary information is
presented to the panellists, who are invited to reassess their original opinions in light of
anonymous individual responses. In addition, if panellists assessments fall outside the
upper or lower quartiles, they may be asked to provide justifications as to why they
consider their estimates are more accurate than the median values. Further rounds of
collection of estimates, compiling summary information and inviting revisions
continues until there is no further convergence of expert opinion. Experience reveals
this usually occurs after two rounds, or at the most four rounds (Janssen 1978).

There are a number of variants on the classical Delphi method. When the issues are well
defined, a clearly defined scenario can be developed by the monitoring team. In such
circumstances, it is common to replace the unstructured first round with a highly
structured set of questions through which specific estimates of parameters are obtained.
A statistical summary of all responses is then provided to the panel for the second
round, rather than in the third. In such cases, it is common for the Delphi method to
include only one or two iterations. Another variant is the ‘paper’ Delphi (sometimes
also known as a ‘paper and pencil Delphi poll’) that is conducted entirely by mail.
Another variant is the ‘real time’ Delphi whereby feedback is provided by computer
and final results are usually available at the end of the session.

Answer to Q 4.4

The classic Delphi method is conducted through a combination of a polling procedure
and a conference. Communication between conference panellists is however restricted
and undertaken through the monitoring team. Even though panellists are at the same
physical location, there is no face-to-face contact. Even in variants of the Delphi, there
is no face-to-face contact. The nominal group technique (NGT) uses the basic Delphi
structure but in face-to-face meetings which allow discussion among participants. A
meeting with NGT starts without any interaction, with individuals initially writing
down ideas or estimates related to the problem or scenario. Each individual then
presents their ideas or estimates, with no discussion until all participants have spoken.
Then each idea or estimate is discussed. The process is then repeated. For this reason,
NGT is sometimes known as the ‘estimate-talk-estimate’ procedure. In practical terms,
like Delphi, the framing of the questions or the scenario is crucial for the success of the
process. Also, ideally, the leader or moderator of the discussion should come from
outside the group.

Answer to Q 4.5

The jury of expert opinion is one of the simplest and most widely used forecasting
approaches. In its most basic form it involves simply executives meeting and deciding
on the best estimate for the item being forecast. As a precursor to the meeting, it is
common to provide background information to executives. There are a number of
variants of this technique discussed in Chapter 4. One of these variants is when the
estimates of the group are obtained by participants writing their estimates on paper, and
then combined to produce an average. This variant of the jury of expert opinion
approach could almost be considered an informal variant of the Delphi method. The
key difference is that there is no mechanism to prevent interaction amongst group

Answer to Q 4.6

(a)   The criteria for selecting panel members should be determined at the start of the
      exercise. In this case, the criteria may be that the person has a minimum number
      of years of experience in (or knowledge of) the tourism industry in the region and
      be in a senior or management position in which they are exposed to changing
      trends in tourist activity.

(b)   In this case, the issues are likely to be fairly well defined and a clearly developed
      scenario involving the future tourist industry could be developed which sought
      specific estimates of future tourist numbers in five years. As such, the
      unstructured first round of the traditional Delphi might be replaced with a
      structured set of questions through which specific estimates of future tourist
      numbers could be obtained. For example, the Delphi could be framed to present
      the main tourist types and then ask the panelists to estimate the likely number of
      tourists of each group coming to the region in five years time. Alternatively, if a
      traditional Delphi is used, in the first round panelists may be asked to identify the
      key factors influencing tourist numbers coming to the region. For instance,
      respondents might identify that the exchange rate has a major impact on the

      number of international tourists, and that the backpacker numbers are more
      sensitive to these fluctuations compared with the international retiree market
      numbers. Based on this response from the first round, in the second round, a
      specific question might be framed asking what the likely impact on numbers of
      backpacker and retiree respectively if the exchange rate became more or less

(c)   The number of rounds will depend on the panelists and the manner in which the
      Delphi is conducted, i.e. at what stage the structured questionnaire is distributed.
      If the survey was simply asking for estimates of likely tourist numbers in five
      years and these were relatively easy to predict based on past experience then it
      would be likely that estimates from respondents would conver after only one or
      two rounds of the structured survey. The greater the degree of uncertainty, the
      more rounds that will be needed to reach some degree of convergence, or at least
      reach a stage where panelists are not changing their estimates in response to the
      feedback provided. Typically few Delphi surveys however go past two

Answer to Q 4.7

Scenario forecasting could be used in a number of ways in helping the decision making
of the company. For instance, the company could develop three scenarios – a best case,
most likely and worst case – and assess the likely impact of these. The ‘best case’
scenario might be that the no carbon dioxide reductions will be agreed to by the
national government and hence no new technology to reduce carbon emissions will be
needed except to the extent that it would have been replaced with more efficient
technology as part of the normal course of operations. The ‘most likely’ scenario might
be that the national government accepts relatively modest targets for reducing carbon
dioxide emissions by a widely accepted figure of say 5%. The worst case scenario
might be that government targets the electricity industry which results in targets set for
reducing emissions being highly onerous and requiring the company to reduce
emissions to 80% of current levels. Once management identified each of these
scenarios, they could then use these to identify strategies for complying with the
emission requirements and what the likely cost of these would be. The        steps      in
developing these scenarios would follow the steps set out in Chapter 4.

Answer to Q 4.8

Many of the methods outlined in Chapter 4 could be used to forecast medium term
international sugar prices. In fact, it is also possible that one of the quantitative
forecasting methods – regression analysis - discussed in Chapter 3 could also be
applied, if the past trend is likely to be repeated in the future. Which method is most
appropriate ultimately comes down to what level of accuracy in the forecast is required,
the amount of resources available to obtain the forecast and the timeframe in which the
forecast must be obtained. For instance, if accuracy is of paramount importance and
cost not a major consideration, then the NGT would probably be the favoured
technique. However bringing groups of experts together for a face-to-face meeting as is

required for NGT or classic Delphi, may be both difficult and expensive. In such cases,
a postal Delphi may be the most appropriate technique to use. Staticized group
techniques are often even simpler and less costly to apply and may be considered if the
trade-off with slightly reduced accuracy is thought appropriate. Staticized group
estimates can also be compiled quickly if necessary and may be considered when
estimates are needed quickly. Furthermore, when there is the potential for major or
discrete changes, the use of scenarios is a technique that provides a convenient
framework for assessing the potential impact of these.

No matter what technique is selected, it is important to recognise the limitations of the
technique as these will impact on how the technique is applied and the quality of the
forecasts obtained. Furthermore, with any technique that involves the collection of data
it is essential to proceed in an orderly and well thought-out manner. Sometimes there is
a tendency to collect information first and then worry about how it is to be used. The
starting point however should be first to clearly identify what information is needed,
decide on the most appropriate technique to collect the data (in context of the resources,
time and other limitations) and only then commence collection.