ComparingRegression and the High-Low Method
COMPARINGREGRESSION AND THE HIGH-LOW METHOD
Theregression method offers the best cost-estimation technique inaccounting procedures. It involves a mathematical approach tocalculate the slope as well as the best fit line for all the costs.The intercept represents the fixed costs while the slope depicts thevariable cost for every unit. The regression method is accuratecompared to the high-low method(Horngren,Datar & Rajan, 2012).
TheHigh-Low method utilizes the highest and lowest activities’ levelsover a span of time to make an estimate of the variable and fixedcosts. The High-Low method splits the mixed costs into the fixed andvariable components. However, in comparison to the regression method,it may be misleading if the activity levels-high and low- are not aperfect representative of the normal activity meaning there areoutliers or extremes. It is relatively unreliable but easy tounderstand. This is because it incorporates only two extremes fromthe set of activity levels against the total costs. It is mostaccurate when the costs at the high and low points are a perfectrepresentative of the majority or every other point (Lamarche, 2006).
Usingthe high-low method, the cost function is as follows:
Higheractivity of cost $5,000 $70,000
Lowestactivity of cost 1,000 55,000
Difference $4,000 $15,000
Revenues = a+ (badvert costs)
Slopecoefficient (b) = = 3.7533
Constant(a) = $70,000($4,000 3.75333)
= $70,000$33,332 = $46,668
andConstant (a) = $55,000($1,000 3.333)
= $55,000$3,333 = $51,667
Revenues = $51,667+ (3.333 Advertising costs)
Theincrement in revenues for every $1,000 used on adverts within therange is:
Usingthe regression equation, 3.7533 $1,000 = $3,753
Usingthe high-low equation, 3.333 $1,000 = $3,333. The high-low method is only fair in estimating therelationship between the revenues and advertising costs. However, theregression equation should be used since it utilizes the informationfrom every observation. The high-low method relies only on the datathat has the highest and lowest values for the costs. Theseobservations do not represent all the data.
Horngren,C. T., Datar, S. M., & Rajan, M. V. (2012). Costaccounting: A managerial approach.Upper Saddle River, N.J: Pearson/Prentice Hall.
Lamarche,C. E. (2006). Regressionon panel data.(Dissertation Abstracts International, 67-11.).