π High-Low Method Calculator
Split any mixed cost in seconds.
Variable cost per unit
Estimates for planning only, not financial advice.
The high-low method splits a mixed cost into its fixed and variable pieces using just two data points: the highest and lowest activity levels. With a high point of $84,000 at 12,000 units and a low point of $60,000 at 8,000 units, the variable cost is ($84,000 - $60,000) / (12,000 - 8,000) = $6.00 per unit, and fixed cost is $84,000 - ($6 x 12,000) = $12,000. The cost formula is y = $12,000 + $6.00x.
How to Use This Calculator
Enter total cost and activity at the highest and lowest observed activity levels. Use units produced, machine hours, or labor hours, whichever drives the cost.
Enter the cost and activity at the highest activity level.
Enter the cost and activity at the lowest activity level.
Read variable cost per unit, fixed cost, and the full cost formula. For break-even on those costs, our margin calculator covers the pricing side.
The High-Low Formula
The logic is simple: fixed cost does not change between the two points, so the entire cost difference must be variable, and dividing by the activity difference gives the per-unit rate. Then fixed cost is whatever is left over. It is the quickest cost-separation technique in managerial accounting, taught everywhere because it needs no statistics. Its weakness is equally simple: it uses only two observations, so one outlier month corrupts the whole answer.
Worked Example
High: $84,000 at 12,000 units. Low: $60,000 at 8,000 units. Variable rate = $24,000 / 4,000 = $6.00 per unit. Fixed = $84,000 - $72,000 = $12,000. Forecast check: at 10,000 units, predicted cost is $12,000 + $6 x 10,000 = $72,000. If actual cost at 10,000 units was $71,500, the method is close, and the $500 gap is the noise the method ignores. Budgeting 15,000 units next quarter? Plan $12,000 + $90,000 = $102,000.
When High-Low Lies to You
The method assumes cost behaves linearly between the two points and that both points are representative. A month with a one-time repair at the high point inflates the variable rate; a strike month at the low point does the opposite. Step costs break the model entirely: if the second shift adds a supervisor, cost jumps in stairs, not lines. Regression using all months is strictly better when you have the data; high-low is for back-of-envelope work and exams.
My practical rule: use high-low for a first pass, then sanity-check the variable rate against engineering reality. If the method says $6 per unit but materials alone cost $8, one of your two points is dirty. Throw out the outlier month and recompute before you budget on it.
Estimates for planning only, not financial advice.
Frequently Asked Questions
What is the high-low method?
A technique that separates a mixed cost into fixed and variable components using the highest and lowest activity levels observed. The cost change between the two points, divided by the activity change, gives the variable rate.
When is the high-low method used?
For quick cost estimates, budgeting, and forecasting when only summary data is available. It is the simplest cost-separation method, which is why textbooks teach it first and practitioners use regression instead.
What are the weaknesses of high-low?
It uses only two data points, so outliers distort it badly. It assumes linear cost behavior and ignores all other observations. Regression analysis with the full dataset is more reliable whenever the data exists.
What is a mixed cost?
A cost with both fixed and variable elements, like a utility bill with a base charge plus usage rates, or a sales rep with salary plus commission. High-low splits the two.
Can the activity base be machine hours?
Yes. Use whatever drives the cost: units produced, machine hours, labor hours, miles driven. The method works with any consistent activity measure, and the variable rate comes out per that unit.
What if fixed cost comes out negative?
Your data points are inconsistent with the model, usually an outlier or a step cost. Do not budget on a negative fixed cost; investigate the months and consider regression or the scattergraph method.
How is this different from regression?
Regression uses every data point to fit the best line statistically; high-low uses only two. Regression also reports how well the line fits. High-low is faster, regression is truer.
Does high-low work for step costs?
Poorly. Step costs jump at thresholds, violating the linearity assumption. If adding a shift adds a fixed supervisor salary, neither high-low nor simple regression captures it without adjustment.
What is the scattergraph method?
Plotting every month's cost against activity and fitting a line by eye. It uses all the data like regression but stays visual and quick, and it exposes outliers that high-low would silently swallow.