Cost Behaviour and Estimation

Question 1

Why do we need to estimate costs?

Answer 1

• assure profitability of a product/service

- evaluate the profit margin

Question 2

Cost behaviour, estimation and prediction

Answer 2

CB: existing relationship between cost and activity
CE: process of estimating relationship of costs and cost driver activities
CP: use of CE to forecast level of costs at particular activity levels

Question 3

Types of cost behaviour

Answer 3

• step fixed costs and step variable costs
• mixed costs
• semi-variable costs

Question 4

Cost estimation methods

Answer 4

• Regression analysis (statistical method)
• account analysis
• engineering method

Question 5

Regression Analysis equation

Answer 5

TC = F + VX or Y = a + bX
TC/Y: total costs, dependent variable
F/a: fixed costs, intercept
V/b: variable cost per unit, slope/gradient
X: cost driver, independent variable

Question 6

Regression Analysis methodology

Answer 6

To create an equation that relates a dependent and one or more independent variables, using data from the past in order to estimate relationships between costs and cost driver activities

Question 7

Line of ‘best fit’ or ‘least squares’

Answer 7

Find the line through data points that minimizes sum of squared distances from the data points to the line

Question 8

Measures of correlation and ‘best fit’

Answer 8

Correlation coefficient (r): strength of relationship between variables
- [r = (-)1]: strong positive or negative relationship
- [r = 0]: no relationship
Coefficient of determination (R^2): measure of goodness of fit
- R^2 close to 1: data points can be strongly fitted into a straight line

Question 9

Considerations of regression analysis

Answer 9

• More data points means more accurate results

- Only estimate costs in the relevant range of activities (e.g. estimation of FC can result in negative)

Question 10

Regression equation may be a poor predictor of future costs if

Answer 10

• cost-activity relationships have changed (e.g. through time, after decisions)
• Costs themselves have changed independently from activity changes (external factors, e.g. supplier prices)

Question 11

Advantages of Regression analysis

Answer 11

• permits the use of more than one cost driver activity to explain costs
• predict costs for new products that are similar to past products with its regression model
• equation can be reused at any times to calculate costs at different activity levels (giving that no external factors or internal changes have taken place)

Question 12

Problems with regression analysis

Answer 12

• these are only estimations, not exact
• can’t be 100% sure that variables have a good relationship (even though logical)
• cannot have more than one Independent variable that are highly correlated
• Historical costs: mismatched periods of highly inflationary periods

Question 13

Account analysis methodology

Answer 13

• review each activity that makes up total cost

- relate costs and activity in the form of general equation (TC = F + VX)

Question 14

Account analysis: process

Answer 14

1. identify cost drivers and their costs
2. sum fixed costs
3. sum VC for each cost driver activity
4. Divide total VC calculated beforehand by total cost driver units to obtain VC per unit
5. Divide FC by number of time periods in data
6. Substitute each number in equation

Question 15

Engineering method process

Answer 15

Direct labour: find labour cost per unit
Direct material: material cost per unit
Overhead costs: estimate total OH costs

Question 16

Advantages of Engineering method

Answer 16

• detailed analysis: accurate cost estimates as company identifies all costs involved
• used to estimate costs of new activities (if past were similar)

Question 17

Disadvantages of Engineering method

Answer 17

high cost, time consuming, more effort required

Question 18

which is better: Regression/account analysis vs engineering method

Answer 18

Regression/account analysis relies on past data, historical data is less accurate, low-cost
Engineering method relies on present data, highly accurate, costly and time consuming
- better results from using several methods