PART B

CHAPTER 5: FORECASTING COSTS AND REVENUES

High-Low Method
This is a quick and easy approach that estimates fixed and variable costs by comparing the highest and lowest
activity levels.

Advantages Disadvantages
 It is easy to understand.  If only two activity levels are available,
 It is easy to use and can determine the fixed and the estimated fixed and variable costs
variable elements of semi-variable costs even if data may not be very accurate.
is available for only two activity levels.  It assumes that costs are only affected
by the activity level, but this may not be
the case in practice.

Example 1
Electricity costs for the first 6 months of the year are as follows:
Units produced Cost ($)
January 340 2,260
February 300 2,160
March 380 2,320
April 420 2,400
May 400 2,300
June 360 2,266
Calculate the fixed and variable costs using the high-low method.
Solution
Y = a + bx = the general cost equation.
b = VC/unit = CAHAL – CALAL
HAL – LAL
= 2,400 – 2,160
420 – 300 = 2 per unit
a = total FC = CAHAL – (b* HAL)
= 2,400 – (2*420) = 1,560

Linear Regression Analysis

Linear regression analysis analyses past cost data to calculate the values of ‘a’ and ‘b’ in a linear function of a
line of best fit.
The values of ‘a’ and ‘b’ are known as linear regression coefficients.

If there is a reasonable degree of linear correlation between two variables, we can use regression analysis to
calculate the equation of the best fit for the data.
This is known as least squares linear regression.
If the equation relating two variables, × and y, is
y = a + bx
then the values of a and b may be calculated using the following formulae (which are given in the
examination)
Where:
y = dependent variable i.e. the total cost for the period
a = intercept on y-axis i.e. the fixed cost for the period
b = gradient i.e. the variable cost per unit
x = the independent variable i.e. the level of activity

Benefits of simple linear regression

1. Simple and easy to use.
2. Looks at the basic relationship between two sets of data.
3. Can be used to forecast and to produce budgets.
4. Information required to complete the linear regression calculations should be readily available.
5. Computer spreadsheet programmes often have a function that will calculate the relationship between two
sets of data.
6. Simplifies the budgeting process.

Limitations of simple linear regression

1. Assumes a linear relationship between the variables.
2. Only measures the relationship between two variables. In reality the dependent variable is affected by
many independent variables.
3. Only interpolated forecasts tend to be reliable. The equation should not be used for extrapolation.
4. Regression assumes that the historical behaviour of the data continues into the foreseeable future.
5. Interpolated predictions are only reliable if there is a significant correlation between the data.

Example 2
The following table shows the number of units produced each month and the total cost incurred:
Units Cost ($ ‘000)
January 100 40
February 400 65
March 200 45
April 700 80
May 600 70
June 500 70
July 300 50
Calculate the regression line, y = a + bx
Solution
b = n∑xy - ∑x∑y
n∑x2 – (∑x)2
 Units, x Cost($000), y xy x^2
100 40 4,000 10,000
400 65 26,000 160,000
200 45 9,000 40,000
700 80 56,000 490,000
600 70 42,000 360,000
500 70 35,000 250,000
300 50 15,000 90,000
2,800 420 187,000 1,400,000
b = (7*187,000) – (2,800*420)
(7*1,400,000) – (2,800)2 = 0.068 ($000) = 68 per unit
a = ∑y - b∑x
n
a = 420,000 – 68*2,800
7 = 32,800
The specific regression line or equation becomes: y = 32,800 + 68x

Example 3
The following data table is Winston’s machine maintenance costs:
Number of machines Total cost $ Xy x2
serviced y
x
2 110 220 4
4 140 560 16
6 170 1,020 36
8 200 1,600 64
10 230 2,300 100
12 260 3,120 144
Σx = 42 Σy = 1,110 Σxy = 8,820 Σx2 = 364
Calculate the regression line, y = a + bx

Example 4
Using linear regression, calculate for Coastway Café the linear function of the line of best fit between the
number of ice creams sold (y) and average daily temperature (x).
Average daily Number of ice-creams
temperature (y)
(x) xy x2 y2
14 59 826 196 3.481
27 102 2,754 729 10,404
20 84 1,680 400 7,056
22 85 1,870 484 7,225
17 75 1,275 289 5,625
∑ 100 405 8,405 2,098 33,791
Solution
Y = a + bx
b = n∑xy - ∑x∑y
 n∑x2 – (∑x)2
a = ∑y - b∑x
n
b = (5*8,405) – (100*405)
(5*2,098) – (100)2 = 3.11
a = 405 – (3.11*100)
5 = 18.8
The specific function becomes: y = 18.8 + 3.11x

Advantages and Disadvantages of Linear Regression Analysis

Advantages Disadvantages
 It is easy to use.  It assumes that a linear relationship exists
 It can be used to forecast costs and revenues. between two variables, but this may not be
 It only needs a few data sets to establish the true. For example, variable costs per unit
linear relationship between two variables. might decrease at higher output levels.
 If the relationship between two variables is  It only uses two variables to calculate a linear
strong, forecast costs and revenues should be function, but in practice, there may be more
reliable. than one independent variable affecting the
 It uses all available data and provides a dependent variable.
definitive line of best fit. (Unlike the high/low  It assumes that the way costs and revenues
method, which only uses two observations at behave in the past is the way that they will
extremes and is likely to be non-typical.) behave in the future, but this is not always
the case in practice.
 If the relationship between two variables is
not strong, forecast costs and revenues may
not be reliable.

The correlation coefficient, r

Pearson’s correlation coefficient is a measure of how linear the relationship between variables is.
If they are related, a change in one variable will cause a change in the other variable.
A correlation coefficient of +1 indicates perfect positive linear correlation, whereas -1 indicates perfect
negative linear correlation.
The further away from +1 or – 1, the less linear correlation exists.
The correlation coefficient (r) may be calculated using the following formula (which is given to you in the
examination)

Interpreting the Correlation Coefficient

r value Description
1 Perfect positive correlation
-1 Perfect negative correlation
Between 0 and 1 partial positive correlation
Between 0 and -1 partial negative correlation

Example 5
Using the data in example 4 above, calculate the correlation coefficient, r.
Solution
r= n∑xy - ∑x∑y
√{n∑x2 – (∑x)2 * n∑y2 – (∑y)2}
r= (5*8,405) – (100*405)
√{(5*2,098) – (100)2 * (5*33,791) – (405)2} = 0.98

Interpretation: r = 0.98 is a strong positive correlation between x and y. i.e. if x increases, y also increases.

Coefficient of determination, r2
The coefficient of determination is the square of the coefficient of correlation i.e. (r2).
It is a measure of how much of the variation in the dependent variable is ‘explained’ by the variation of the
independent variable.
The coefficient of determination is calculated by squaring the correlation coefficient (r2 = r × r).
From Example 5 above, r2 = 0.982 = 0.9604 or 96.04%.
Interpretation: 96.04% of the variation in y, has been explained/caused by the variation in x.
The rest, i.e.3.96% is caused by other factors not included in the model or error.
This means that, x is a good predictor of y, or the model is good for prediction.

TIME SERIES ANALYSIS

Introduction
A time series is a data set whose independent variable is time.
A time series is a series of figures recorded over time, e.g. unemployment over the last 5 years, output over
the last 12 months, etc.
Time series analysis is a technique used to:
• identify whether there is any underlying historical trend
• use this analysis of the historical trend to forecast the trend into the future
• identify whether there are any seasonal variations around the trend
• apply estimated seasonal variations to a trend line forecast in order to prepare a forecast season by season.

The following are examples of time series.

a) Output at a factory each day for the last month
b) Monthly sales over the last 2 years
c) Total annual costs for the last 10 years
d) Retail Prices Index each month for the last 10 years
e) The number of people employed by a company each year for the last 20 years.
A graph of a time series is called a historigram.

Time Series Components:

Component Description
Trend The underlying long-term movement in values over time.
The regular rise and fall over shorter periods of time.
Seasonal
For example, umbrella sales are likely to be higher than average every winter and lower
variation
than average every summer.
Cyclical medium-term changes in values resulting from factors that repeat in cycles. Cyclical
variation variations are longer-term than seasonal variations.
Random
Fluctuations that are not part of a pattern and are difficult to predict
variation
Identifying Variations from a Graph

 The data line is the plotted time series data on a graph.

Time series data can be plotted onto a graph to identify patterns.
 The trend is the data line's general direction, showing the data values' long-term movement. In this
graph, the trend appears to be increasing.
 Seasonal variations are short-term fluctuations around the trend line. They are the gaps between the
trend and the data line on a graph that regularly recur.

Predicting Future Values

Time series analysis can be expressed as either an additive or multiplicative model:
Y = T + S + C+ R
Or
Y=T×S×C×R
Y = Data point
T = Trend
S = Seasonal variation
C = Cyclical variation
R = Random variation

NB:
For short-term forecasting, cyclical and random variations are ignored as they are difficult to predict. The
model may be simplified to:
Y=T+S
Or
Y=T×S

Illustration
Poin
Label Description
t
A Seasonal Variation – There was a significant increase in sales in Year 2 – this seasonal variation
Year 2 occurred when the World Cup took place (once every four years).
B Trend Line – Relevant Draw the line of best fit on the graph to show a relationship between actual
Range [Solid line only] sales and time.
On a time series graph, a line of best fit is called a trend line as it shows the
trend of the results over the 12 years.
Sales are moving in an upward direction as they increase over time. Linear
regression analysis can also be used to establish the linear function of this
trend line.
C Seasonal Variation – There was a significant increase in sales in Year 6 – this seasonal variation
Year 6 occurred during the World Cup.
D Seasonal Variation – There was a significant increase in sales in Year 10 – this seasonal variation
Year 10 occurred during the World Cup.
E Forecast Sales – Year Forecast the trend in Year 14 by reading the value from the graph. The
14 forecast trend for Year 14 is approximately $45,000.
However, the World Cup is scheduled to take place in Year 14, so the forecast
trend needs to be adjusted by the seasonal variation to make a reliable sales
forecast for Year 14.
F Extrapolated Trend One of the things to do in time series analysis is to predict future trends. This
Line [Dotted line only] is done by extending (extrapolating) the trend line outside the relevant range
to forecast the trend in the coming years.
 Relevant range
The trend line’s relevant range is between Years 1 and 12 because this is where the observed actual sales
figures have been plotted.
To see what the trend line would look like outside this range, it needs to be extrapolated (extended into the
future). This would mean that, after Year 12, any extrapolated trend line would be outside the relevant range
and may not be a reliable estimate of the trend in the future.
 Extrapolation
Extrapolation, in the context of the time series graph, means extending into the future, that is, beyond Year
12.

Time Series Analysis Approach

Time series analysis is a technique that makes budget forecasts using trend and seasonal variations.
1. Identify the trend of a time series (either by moving averages or linear regression analysis).
2. Forecast the times series’ trend into the future
3. Identify seasonal variations (which will cause the trend to alter from its linear function).
4. Adjust the trend (by the estimated seasonal variations to prepare budget forecasts).

Determining the Trend

The trend is the smoothed-out time series line. In other words, the trend line indicates the general direction of
the data line with minimal fluctuations.
The trend of historical data can be established in the following ways:
 By establishing the equation of the line of best fit using linear regression analysis.
 By drawing the line of best fit.
 By using a method called moving averages.

Moving Averages
The moving average method calculates the average of a set of consecutive periods. Averaging the time series
data removes any seasonal variations to estimate the trend.
If the set of periods is odd, the moving average will coincide with the middle data point.
If the set of periods used is an even number, the moving average must be calculated twice to ensure the mid-
point trend figure coincides with a data point.

Example 6
The following revenue data is available:
Year Revenue
$’000
20X1 50
20X2 54
20X3 55
20X4 59
20X5 60
20X6 64
20X7 68
20X8 72
1. Calculate a three-point moving average from the data and plot it on a graph.
2. Calculate a two-point moving average from the data.

Example 7
Year Actual sales
$000s
Year 1 21.00
Year 2 52.00
Year 3 24.00
Year 4 27.00
Year 5 29.00
Year 6 54.00
Year 7 30.00
Year 8 31.00
Year 9 33.00
Year 58.00
 10
Year 36.00
11
Year 39.00
12
Using the above data, calculate a 4-year moving average for Franklyn’s Football Factory.

Average Periodic Increase

The average period increase of the trend can be calculated by taking the difference between the earliest and
latest values and dividing them by the number of periods.
This assumes the trend is consistent.
The formula is:
Average periodic increase = (Trendlatest value – Trendearliest value) / Number of increases

Illustration
Below is a 3-point moving average:
Year Revenue Sum of three years Moving average
$000 $000 $000
20X 50
1
20X 54 159 53.00
2
20X 55 168 56.00
3
20X 59 174 58.00
4
20X 60 183 61.00
5
20X 64 192 64.00
6
20X 68 204a 68.00
7
20X 72
8
Calculate the average periodic increase for the given trend and forecast the trend’s expected value for 20X9 and 20X0.
Solution
Average periodic increase = (Trendlatest value – Trendearliest value) / Number of increases
Average periodic increase = (68 – 53) / 5
Average periodic increase = 15 / 5
Average periodic increase = 3
The average yearly increase in trend is $3,000.

The expected values for 20X9 and 20X0 are:

Last trend
point (20X7) Periods after the last Trend increase Expected value
Year $000 trend point (20X7) $000 $000
20X9 68 2 6 74
 (3 × 2)
20X0 68 3 9 77
(3 × 3)

Using Regression Coefficients to Calculate the Trend

When the linear function of a trend line has been derived (worked out) by linear regression analysis, it can be
used to forecast future costs and revenues.

Seasonal Variations in Time Series Analysis

Seasonal variations are regularly recurring fluctuations on the trend line for a time series. They are caused by
the pattern of demand for a product or service and can occur annually, monthly, weekly or daily.
The seasonal variation in a time series can be calculated and used in conjunction with the trend to predict
future cash flows.
As a reminder, the formula for calculating a time series is:
Y = T + S (Additive model)
Or
Y = T × S (Multiplicative model)
The seasonal variation may be determined as follows:
S = Y – T (Additive model)
Or
S = Y / T (Multiplicative model)

The steps to calculate the seasonal variation are as follows:

1. Calculate the trend.
2. Calculate the variation for each season.
3. Average the variation for each season, and normalise (this step is to ensure any trend effects are
removed)
 To normalise seasonal variation in the additive model, ensure the sum of the average variations is zero.
 To normalise seasonal variation in the additive model, ensure the sum of the average variations equals
the number of seasons.
(i.e. if each season is a quarter in the year, the sum of average variations should be four since there are
four quarters in a year).

Example 8: Calculating Seasonal Variation (Additive Model)

Year Quarter Revenue ($000)
Q1 60
Q2 55
20X1
Q3 25
Q4 70
Q1 61
Q2 57
20X2
Q3 27
Q4 74
Q1 63
Q2 56
20X3
Q3 29
Q4 77
There is a clear indication of seasonality for each quarter.
a) Using the additive model, calculate the quarterly seasonal variation. (use a three-point moving average for
calculating the trend)
b) Using the additive model, forecast the quarterly sales for 20X4.

Example 9: Calculating Seasonal Variation (Multiplicative Model)

Using the data from Example 8,
a) Using the multiplicative model, calculate the quarterly seasonal variation. (use a three-point moving
average for calculating the trend)
b) Using the multiplicative model, forecast the quarterly sales for 20X4.

Illustration
a) The sales trend in Month 28 is forecast to be $86,000, and the seasonal variation in Month 28 is forecast to
be +15,000.
Calculate the sales budget forecast for Month 28 using the additive model.
b) The sales trend in Month 28 is forecast to be $86,000, and the seasonal variation in Month 28 is indicated
to be 0.82 times the trend.
Calculate the sales budget forecast for Month 28 using the multiplicative model.
Solution:
a) Sales budget forecast for Month 28 = Trend + 15,000 = 86,000 + 15,000 = 101,000
b) Sales budget forecast for Month 28 = $86,000 × 0.82 = $70,520.

Budget forecasts
Information about the trend and seasonal variations can be used to make budget forecasts:
 Use regression coefficients to forecast the trend for a future period.
 Adjust forecast trend values by the seasonal variation to forecast budgets for future periods.

Advantages and Disadvantages of Time Series Analysis

Advantages Disadvantages
 It can be used to make forecasts if a straight-line  It assumes that the trend is a linear function
trend is assumed to exist and information about (straight line), but this may not always be
seasonal variations is available. the case in practice.
 If the relationship between two variables is  It assumes that what has happened in the
strong, forecast costs and revenues should be past is likely to occur in future, but this is
reliable. not always the case in practice.
 It assumes that seasonal variations occur
with constant regularity, though this is not
always the case in practice.

INDEX NUMBERS
Index – A measure to compare values over time.

The purpose of index numbers is to compare values of things over time, for example, prices and quantities of
products. They are also used to adjust historical data and make forecasts.

The most common use is as a way of measuring the effect of inflation on prices.

Simple index numbers/Single item index

A simple price index measures the changes in prices of a single item over time.
Simple index numbers are based on a single item.
There are two types: price relative and quantity relative.
A price relative index number shows changes in the price of an item over time.
A quantity relative index number shows changes in quantity over time.
Simple price index = P1 × 100
P0
Simple quantity index = q1× 100
q0
Remarks:
Under the fixed base method, the base year is allocated an index of 100, and subsequent years are measured
against this base.
Under the chain base method, the base year is always the year immediately before the year under
consideration.
Most indexes are base-100, meaning that the base year is given a value of 100. This makes it easier to see
changes.
For example, a year with a 120 index is 20% more than the base year (100).

Illustration
The oil price is $270 a barrel in 20X1, $300 in 20X2, $340 in 20X3 and $380 in 20X4.
Calculate a chain base index and a fixed base index that uses 20X1 as the base year.
Solution
Chain base index
20X1: 100
20X2: 111 (300/270 × 100)
20X3: 113 (340/300 × 100)
20X4: 112 (380/340 × 100)
Fixed base index
20X1: 100
20X2: 111 (300 / 270 x 100)
20X3: 126 (340/270 × 100)
20X4: 141 (380/270 × 100)

Example 10
The price of coffee was $2.40 in 2006, $2.50 in 2007, and $2.60 in 2008
Calculate the price index for 2007 and 2008 using 2006 as base year.

Example 11
Sales of tea were 8,200 packets in 2008, 9,000 packets in 2009 and 9,400 packets in 2010.
Calculate the quantity index for 2009 and 2010 using 2008 as a base year.

Example 12
The following index is available for Furniture Co:
Year Revenue ($000) Index
20X1 10,000 100
20X2 ? 110
20X3 ? 115
a) Which year has been selected as the base year?
b) What is Furniture Co’s revenue in the years 20X2 and 20X3?

Illustration
The following data relates to Product K:
Year Sales (Units)
20X 6,800
1
20X 8,500
8
A simple quantity index is calculated by using 20X as the base year, Q0 and 20X8 as Q1 (quantity at Time 1):
Quantity index = Q1/Q0 × 100 = 8,500/6,800 × 100 = 125
The quantity index calculated shows that the number of units of Product K sold in 20X8 has increased by 25%
compared to the quantity sold in 20X1.

Weighted index
A weighted index must be calculated when multiple items (or variables) are considered.
For example, a country’s inflation rate is commonly measured by calculating a weighted index based on
several products. The products that are selected are those that represent the items that might appear in the
shopping basket of an average household. These indices are often given names, for example (in the UK) the
Retail Price Index (RPI). Movement of this index is used to compare with movement in other areas, for
example, when considering what is ‘reasonable’ when considering wage and salary increases.

Illustration
A company manufactures three products: Product AA, Product BB and Product CC. Information relating to
these products is shown in the following table.
Product Sales price ($) Sales price ($) 20X6
type 20X5
Product AA 10 12
Product BB 16 20
Product CC 20 22
The management accountant of the company wishes to calculate a weighted price index and uses the
following weightings, which are based on the number of units of each product sold.
Product Sales volume (units) Product type
type
Product AA 8,000 Product AA
Product BB 4,000 Product BB
Product CC 2,000 Product CC
Calculate the price index weighted on units sold.
Solution
1. Calculate the simple price index for each product.
Price index – Product AA P1/P0 × 100% = 12/10 × 100 = 120
=
Price index – Product BB = P1/P0 × 100% = 20/16× 100 = 125
Price index – Product CC = P1/P0 × 100% = 22/20× 100 =110
2. Weight the price indices using the number of units sold
Product type Price Weighting (sales units) Price index × weighting
index
 Product AA 120 8,000 960,000
Product BB 125 4,000 500,000
Product CC 110 2,000 220,000
Total 14,000 1,680,000
3. Calculate the weighted price index
Weighted price index = (Price index × weighting) /Weighting
= 1,680,000/14,000
= 120

Example 13
A weighted price index is based on the consumption of three products: bread, cheese and eggs. The following
information has been provided as at 31 December:
20X2 20X5
Commodity Quantity (annual) Unit price P × Q Unit Price P × Q
$ $ $ $
Bread 100 loaves 0.51 0.62
Cheese 25 kilos 1.60 2.00
Eggs 50 half-dozen 0.80 0.90

Illustration
Sanjay’s annual salary over the past three years is shown below, alongside the corresponding Retail Price
Index (RPI).
Year Annual salary ($) RPI
20X2 35,000 160
20X3 35,800 163
20X4 36,400 167
Has Sanjay’s salary increased or decreased in real terms (so after allowing for inflation) on the previous year, in
20X3 and 20X4?
Solution
20X3
Sanjay’s 20X2 salary at 20X3 prices would be: (35,000 / 160) x 163 = $35,656. So Sanjay’s 20X3 salary of
$35,800 has increased in real terms on his 20X2 salary.
20X4
Sanjay’s 20X3 salary at 20X4 prices would be: (35,800 / 163) x 167 = $36,679. So Sanjay’s 20X4 salary of
$36,400 has decreased in real terms to his 20X3 salary.

Laspeyre and Paasche index numbers

In order, for example, to measure the overall effect of inflation, it is more sensible to consider the change in
price of a typical ‘shopping basket’ of goods rather than looking at just one item.
The Laspeyre price index uses base period quantities, whereas the Paasche price index uses current period
quantities.
Example 14
Below are the stated quantities and unit prices for a typical ‘shopping basket’ in each of the year 2008, 2009,
and 2010.

Calculate price index numbers for 2009 and 2010, with 2008 as a base year, using:
(a) Laspeyre
(b) Paasche

Illustration
The following information is available at 31 December:
Component Price Price Quantity Quantity
20X 20X5 20X2 20X5
A 2 $2.5 4,000 2,000
$1.2
B $5.3 $5.8 1,000 800
C $2.3 $2.7 2,000 4,000
Calculate both the Laspeyre and Paasche price indices as at 31 December 20X5. (Assume 31 December 20X2
is the base date.)
Solution
Componen Price Quantity Price Quantity PoQ0 PnQo PoQn PnQn
t
Po Qo Pn Qn $ $ $ $
A $1.20 4,000 $2.50 2,000 4,800 10,000 2,400 5,000
B $5.30 1,000 $5.80 800 5,300 5,800 4,240 4,640
C $2.30 2,000 $2.70 4,000 4,600 5,400 9,200 10,800
14,700 21,200 15,84 20,440
0
Price indices for 31 December 20X5 (31 December 20X2 as the base):
21,200
Laspeyre = × 100 = 144.2
14,700
20,440
Paasche = × 100 = 129.0
15,840