When should I use regression analysis instead of the high-low method?

Always prefer least-squares regression when you have more than a few data points and need reliable fixed/variable cost estimates for budgeting, pricing, or decision-making. The high-low method is a quick approximation suitable only for introductory analysis or when data is extremely limited.

Can the high-low method produce a negative fixed cost?

Mathematically yes, if the low-activity observation has a higher per-unit total cost that extrapolates to a negative intercept. This signals either an outlier, non-linear cost behavior, or that the chosen observations are inappropriate. Negative fixed costs are economically meaningless and indicate the method has been misapplied.

What is the difference between mixed costs and step costs?

A mixed (semi-variable) cost has both a fixed component and a variable component that change smoothly with activity - like utilities with a base charge plus per-unit consumption. A step cost remains fixed over a range of activity but jumps to a new level at certain thresholds, like hiring an additional supervisor. The high-low method applies to mixed costs, not step costs.

High-Low Method Cost Calculator

Use the high-low method to separate mixed costs into fixed and variable components. Enter the highest and lowest activity levels and their corresponding costs to estimate variable cost per unit and total fixed costs.

By AbraCalc Editorial Team · Reviewed by AbraCalc Methodology Review · Last reviewed: 2026-06-25

Embed this tool on your site

<iframe src="https://abracalc.com/calculator/high-low-method/embed.html" width="340" height="320" style="border:1px solid #e4e7ec;border-radius:12px" loading="lazy"></iframe>

How to use this tool

Enter highest activity level, cost at highest activity, lowest activity level, cost at lowest activity and target activity level (optional) in the fields above.
Results update instantly as you type — or click Calculate.
Read your variable cost per unit and the full breakdown beneath it.

⚠ This tool provides general estimates for education only and is not financial, tax or legal advice. Figures may not reflect your situation — verify with a qualified professional.

Formula

Variable cost per unit: b = (Cost_high − Cost_low) / (Units_high − Units_low)

Fixed cost: a = Cost_high − b × Units_high

Total cost at any level: TC = a + b × Units

How it works

The high-low method is a simple cost-accounting technique that uses only the highest and lowest observed activity levels and their associated costs to estimate the variable and fixed components of a mixed (semi-variable) cost. The variable cost per unit is the slope of the line connecting the two data points, and the fixed cost is the y-intercept. While easy to apply, this method can be distorted by outliers at the extreme data points; regression analysis is more accurate when many data points are available.

Worked example

Separating factory overhead costs

Variable cost per unit = ($30,000 − $18,000) / (5,000 − 2,000) = $12,000 / 3,000 = $4.00/unit
Fixed cost = $30,000 − $4.00 × 5,000 = $30,000 − $20,000 = $10,000
Predicted cost at 3,500 units = $10,000 + $4.00 × 3,500 = $10,000 + $14,000 = $24,000

Variable cost = $4.00/unit; Fixed cost = $10,000; Total cost at 3,500 units = $24,000

Common mistakes to avoid

Selecting the highest and lowest activity levels rather than the highest and lowest cost data points: if the maximum cost observation does not coincide with the maximum activity level, using the wrong pair distorts both the variable rate and fixed cost estimate.
Using outlier data points (a month with an unusual one-time expense or a shutdown period with zero production) as the high or low observation, which violates the method's assumption that both points lie on the same cost line.
Assuming the high-low method gives an accurate cost function: it uses only two observations and ignores all others, making it far less reliable than regression analysis for actual cost estimation and budgeting.

Key terms

What is a mixed cost?: A mixed (semi-variable) cost has both a fixed component (constant regardless of activity) and a variable component that changes proportionally with output or activity level.
What are the limitations of the high-low method?: It uses only two data points, which may be outliers, potentially producing inaccurate estimates. Least-squares regression using all data points is generally more reliable.
What is the cost equation?: The resulting cost equation is TC = Fixed Cost + (Variable Cost per Unit × Activity Level), a linear equation in the form y = a + bx.
What activity measures are used?: Common activity measures include machine hours, direct labor hours, units produced, or miles driven, depending on the cost being analyzed.

Frequently asked questions

When should I use regression analysis instead of the high-low method?: Always prefer least-squares regression when you have more than a few data points and need reliable fixed/variable cost estimates for budgeting, pricing, or decision-making. The high-low method is a quick approximation suitable only for introductory analysis or when data is extremely limited.
Can the high-low method produce a negative fixed cost?: Mathematically yes, if the low-activity observation has a higher per-unit total cost that extrapolates to a negative intercept. This signals either an outlier, non-linear cost behavior, or that the chosen observations are inappropriate. Negative fixed costs are economically meaningless and indicate the method has been misapplied.
What is the difference between mixed costs and step costs?: A mixed (semi-variable) cost has both a fixed component and a variable component that change smoothly with activity - like utilities with a base charge plus per-unit consumption. A step cost remains fixed over a range of activity but jumps to a new level at certain thresholds, like hiring an additional supervisor. The high-low method applies to mixed costs, not step costs.