Fundamentals of Cost Accounting 3rd Edition by William N. Lanen, Shannon W. Anderson, Michael Maher

Exercise 34

Interpretation of Regression Results

Ross Enterprises maintains a fleet of agricultural equipment for rental to local farmers. Ross maintains all its equipment in a company-owned facility. Data on maintenance costs and operating hours of the equipment have been collected for the past 24 months to help managers plan financial needs.

Managers at Ross were initially excited about having the data and the analysis available for planning, but the initial regression results revealed the following equation: Maintenance costs = $10,564 — $67.13 × Operating hours The coefficient on operating hours was highly significant and the adjusted R2 was 0.89.

Required

How would you explain a negative coefficient; does it seem likely that the more the equipment is operated, the less the company spends on maintenance?

Step 1 of 2

Cost estimation

Cost estimation is an important exercise because it helps managers in decision making. Correct cost estimates result in cost saving and making business successful. Cost estimates helps managers to evaluate and choose the best alternative. It is important for managers to capture the correct cost for each alternative.

There different methods used for cost estimation like engineering, accounting and statistical analysis.

Statistical method

Statistical is more accurate method of cost estimation as compared to engineering and accounting analysis method as they have certain limitations. Under this method, random events are separated from while analyzing relationship between cost and activity. While using statistical method for cost estimation it is important to confirm that past activity levels are related to current estimation.

Simple regression analysis – In simple regression analysis there is only one predictor for the activity and cost is estimated on the basis on one predictor only. For example, overhead cost is estimated on the basis of relationship between total cost and parts used.

Multiple regression analysis – In multiple regression analysis more than one type of predictor is used for the activity and cost is estimated on the basis on multiple predictors. Multiple predictors are used for better and more accurate cost estimation. For example, overhead cost is estimated on the basis of parts used and labor hours used for production purpose. Only parts cost may not provide correct overhead cost estimation, hence labor hours used, must also be included as predictor to calculate overhead cost.

Coefficient correlation, coefficient of determination

R is referred as correlation coefficient which is used to measure the immediacy of data point to regression line. Higher correlation coefficient i.e. close to 1.0 is always better and it shows data points are close to regression line.

is referred as coefficient of determination which is referred as proportion of variation in Y by the X predictors. It shows the percentage change in estimated cost i.e. Y may be directly correlated with the changes in predictor which is X i.e. cost of parts or labor cost. is calculated by multiplying the value of R with value of R i.e. correlation coefficient squared.

Adjusted is same as i.e. correlation coefficient squared but it is also adjusted for different independent variable used for estimation purpose. increases with increase in the number of variables and adjustment to is done which is called adjusted .

Adjusted is calculated when multiple regression is used.