Courses Publications Research Home Links

Bd15156_.gif (323 bytes)

Cost -Definition Behaviour and Measurement

 

 

This information serves as a basis and is needed for cost analysis, cost planning and decision making and cost management.

A firm's costs can be categorized by type and behavior such as:

  1. Direct Material

    Direct Labor

    Factory Overhead

    Marketing

    General and Administrative

  2. Direct

    Indirect

  3. Fixed

    Variable

  4. Marginal

    Incremental

  5.  

    Linear with increased   output
    Non Linear " " "
    Independent " " "

  6. Engineered

    Committed

    Discretionary

    Period

  7. Controllable

    Non Controllable

  8. Relevant

    Non Relevant

  9. Unit

    Total

  10. Prime

    Conversion

  11. Value Added

    Non Value Added

        

 

Which category should be used? 

Depends on situation being considered.

A major need is to develop Cost Estimates (Forecasts) for:

  • Budgeting Purposes

  • Cash Need Planning

  • Price Determination

  • Capital Investment Evaluation

  • Operations Cost Management

 

Situation can be:

  1. No historical (experience) data available

  2. Historical data is available.

 

  1. Where no data is available one can:

  • See if Any Relevant Analogies

  • Generate data through:

  •  Statistical Sampling

  •  Simulation                

       Define work required, methods and material used, conditions to develop an Engineered Cost Estimate. 

  1. Historical data is available.     

Where data is available utilize statistical methods to develop a cost function that can be used to predict   cost at a certain output level.

These functions are most accurate if used in the "Relevant Range".

General Cost Function

 

This is linear assumption is often sufficiently precise.


General Approach to Develop a Cost Function
  1. For the cost (dependent variable) you want to predict, identify the most relevant cost drives (independent variances)
  2. Plot these and see if there is a visual fit and which one seems the best
  3. Depending on the accuracy required, these methods are commonly used to develop the cost function
    • Visual Fit
    • High Low
    • Regression - Least Square Fit
    • Multiple Regression - (Parametric Cost Estimate)

 

Least Square Fit provides the most accurate estimate.  How confident of its predictive value is indicated by R2 the coefficient of determination (the closer to one the better). Standard error of the coefficents (should be small relative to actual parameter of coefficient X) 

Top

Bd15156_.gif (323 bytes)

Courses Publications Research Home Links